article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
it is well known that the baryon spectrum is very rich and much more complex than the meson spectrum .
nevertheless , at the moment , the number of known light - quark mesons is much larger than the number of known baryon resonances @xcite .
this problem has to do with the difficulty of identifying those high - lying baryon resonances that are only weakly coupled to the @xmath2 channel @xcite and thus can not be seen in elastic @xmath2 scattering experiments . since the experimental observations of baryon resonances mainly come from reactions in which the pion is present either in the incoming channel , such as @xmath3 , or in the outgoing one , such as @xmath4 , it would not be surprising if some baryon resonances , very weakly coupled to the single pion , were missing from experimental results .
these baryons may decay mainly into two pion channels ( @xmath5 ) or into channels such as @xmath6 , @xmath7 , @xmath8 and @xmath9 , where the final - state meson is different from the pion @xcite .
although interesting results were provided by cb - elsa @xcite , taps @xcite , graal @xcite , saphir @xcite and clas @xcite , theoretical calculations of strong , electromagnetic and weak decays of baryons may still help experimentalists in their search for those resonances that are still unknown . the qcd mechanism behind the ozi - allowed strong decays @xcite is still not clear . for this reason ,
several phenomenological models have been developed in order to carry out this type of study , including pair - creation models @xcite , elementary meson emission models @xcite and effective lagrangian approaches ( for example , see ref .
attempts at modeling strong decays within the quark model ( qm ) formalism date from micu s suggestion @xcite that hadron decays proceed through @xmath10 pair production with vacuum quantum numbers , i.e. @xmath11 . since the @xmath10 pair corresponds to a @xmath0 quark - antiquark state , this model is known as the @xmath0 pair - creation model @xcite .
a few years after its introduction , le yaouanc @xmath12 @xmath13 used the @xmath0 model to compute meson and baryon open - flavor strong decays @xcite and also evaluated the strong decay widths of @xmath14 , @xmath15 and @xmath16 charmonium states @xcite .
the @xmath0 model , extensively applied to the decays of light mesons and baryons @xcite , has recently been applied to heavy meson strong decays in the charmonium @xcite , bottomonium @xcite , open - charm @xcite and open - bottom @xcite sectors . in the 90 s , capstick and roberts calculated the @xmath2 and the strange decays of non - strange baryons @xcite by using relativized quark model wave functions for baryons and mesons .
the baryon and meson spectra were predicted within the relativized qms of refs .
it is also worthwhile to cite ref .
@xcite , where the authors computed the open - flavor strong decays of @xmath17 baryons up to @xmath18 shell in a chiral quark model . in this paper
, we present our results for the open - flavor strong decay widths of light baryons ( i.e. made up of @xmath19 valence quarks ) into a baryon - pseudoscalar meson pair and a baryon - vector meson pair .
the widths are computed within a modified version of the @xmath0 pair - creation model where , for the first time in the baryon case , a strange quark pair suppression mechanism has been taken into account , analogous to what was done in the meson sector to suppress unphysical heavy quark pair - creation @xcite .
the effects of this mechanism , which breaks the su(3 ) symmetry , can not be re - absorbed in a redefinition of the model parameters or in a different choice of the @xmath0 model vertex factor . in the next section
, we briefly mention the models for the baryon spectrum and structure that we have used for the calculation of the strong decays : the @xmath20 algebraic model @xcite , by bijker , iachello and leviatan , and the hypercentral model ( hqm ) @xcite , developed by giannini and santopinto . in section [ 3p0 model ] , we review the @xmath21 model for the two - body decay of a baryon into a baryon and a meson , including a discussion of the phase space factor .
the results for the strong decays are presented in section [ strong decay widths ] .
finally , we present a summary and conclusions . some details of the calculation of @xmath21 matrix elements are presented in the appendices . of particular interest is appendix [ ap2 ] , where the flavor couplings with @xmath22 breaking induced by the presence of the strange pair suppression mechanism have been derived explicitly for the first time .
the baryon spectrum is computed by means of algebraic methods introduced by bijker , iachello and leviatan @xcite .
the algebraic structure of the model consists of combining the symmetry of the internal spin - flavor - color part , @xmath23 , with that of the spatial part , @xmath20 into @xmath24 the @xmath20 model was introduced @xcite to describe the relative motion of the three constituent parts of the baryon .
the general idea is to introduce a so - called spectrum generating algebra @xmath25 for quantum systems characterized by @xmath26 degrees of freedom . for baryons ,
there are the @xmath27 relevant degrees of freedom of the two relative jacobi vectors @xmath28 and their canonically conjugate momenta , @xmath29 and @xmath30 .
the @xmath20 model is based on a bosonic quantization which consists of introducing two vector boson operators , @xmath31 and @xmath32 , associated to the jacobi vectors , and an additional auxiliary scalar boson , @xmath33 .
the scalar boson does not represent an independent degree of freedom , but is added under the restriction that the total number of bosons @xmath34 is conserved .
the model space consists of harmonic oscillator shells with @xmath35 .
the baryon mass formula is written as the sum of three terms @xmath36 where @xmath37 is a constant , @xmath38 is a function of the spatial degrees of freedom and @xmath39 depends on the internal degrees of freedom . the spin - flavor part is treated in the same way as in ref .
@xcite in terms of a generalized grsey - radicati formula @xcite , which in turn is a generalization of the gell - mann - okubo mass formula @xcite @xmath40 the operators @xmath41 and @xmath42 correspond to the linear and quadratic casimir invariants of the relevant groups for the internal degrees of freedom .
the values of the coefficients @xmath37 , @xmath43 , @xmath44 , @xmath45 , @xmath46 , @xmath47 and @xmath48 are taken from @xcite .
since the space - spin - flavor wave function is symmetric under permutation group @xmath49 of the three identical constituents , the permutation symmetry of the spatial wave function has to be the same as that of the spin - flavor part .
thus , the spatial part of the mass operator @xmath38 has to be invariant under the @xmath49 permutation symmetry .
the dependence of the mass spectrum on the spatial degrees of freedom is given by : @xmath50 in refs .
@xcite , strong decays of baryons were studied in the collective string model , which is a special case of @xmath20 in which the radial excitations are interpreted as rotations and vibrations of an oblate top , in combination with the elementary emission model for the strong decays . here
, we do the same , but calculate the decays in the @xmath0 pair - creation model @xcite . the rotational part of the operator ( [ eqn : u(7)02 ] )
is written in the same form as in @xcite @xmath51 with eigenvalues @xmath52 in this way one gets linear regge trajectories with a slope @xmath53 , as required by the phenomenology @xcite .
the spectrum of the vibrational part is given by @xcite @xmath54 where @xmath55 and @xmath56 are the vibrational quantum numbers , corresponding to the symmetric stretching vibration along the direction of the strings ( breathing mode ) , and two degenerate bending vibrations of the strings .
the spectrum consists of a series of vibrational excitations characterized by the labels @xmath57 , and a tower of rotational excitations built on top of each vibration . @xmath58 and @xmath59 are free parameters fitted to the data . the spectra calculated for the non - strange and strange baryons are shown in tables [ tab : nuc][tab : omega ] , second / third column .
the baryon wave functions are denoted in the standard form as @xmath60 \ , \right > ~ , \label{wf}\ ] ] where @xmath61 and @xmath62 are the spin and total angular momentum @xmath63 . as an example , in this notation the nucleon and delta wave functions are given by @xmath64 \ , \right>$ ] and @xmath65 \ , \right>$
] , respectively .
the starting point of the hqm is the assumption that quark interaction is hypercentral , i.e. it only depends on the hyperradius @xmath66 @xcite , @xmath67 with @xmath68 @xcite .
thus , the spatial part of the three - quark wave function , @xmath69 , is factorized as @xmath70l_{\rho}l_{\lambda}}({\omega}_{\rho},{\omega}_{\lambda},\xi ) \mbox { } , \label{psi}\ ] ] where the hyperradial wave function , @xmath71 , is labeled by the grand angular quantum number @xmath72 and the number of nodes @xmath73 .
@xmath74l_{\rho}l_{\lambda}}({\omega}_{\rho},{\omega}_{\lambda},\xi)$ ] are the hyperspherical harmonics , with angles @xmath75 , @xmath76 and hyperangle @xmath77 @xcite .
the dynamics is contained in @xmath71 , which is a solution of the hyperradial equation @xmath78
\psi_{\gamma \nu}(x ) \\ \hspace{1 cm } = \mbox { } - 2m~[e - v_{3q}(x)]~~\psi_{\gamma \nu}(x )
\mbox { } .
\end{array } \label{hyrad}\ ] ] in the hqm , the quark interaction has the form @xcite @xmath79 where @xmath80 and @xmath53 are free parameters fitted to the reproduction of the experimental data .
( [ h_pot ] ) can be seen as the hypercentral approximation of a cornell - type quark interaction @xcite , whose form can be reproduced by lattice qcd calculations @xcite .
now , to introduce splittings within the su(6 ) multiplets , an su(6)-breaking term must be added . in the case of the hqm , such violation of the su(6 ) symmetry
is provided by the hyperfine interaction @xcite .
the complete hqm hamiltonian is then @xcite @xmath81 where @xmath82 and @xmath83 are the momenta conjugated to the jacobi coordinates @xmath84 and @xmath85 .
in addition to @xmath80 and @xmath53 , there are two more free parameters in the hqm , the constituent quark mass , @xmath86 , and the strength of the hyperfine interaction .
the former is taken , as usual , as @xmath87 of the nucleon mass .
the latter , as in the case of @xmath80 and @xmath53 , is fitted in @xcite to the reproduction of the * * * and * * * * resonances reported in the pdg @xcite .
the hqm has an approximate @xmath88 symmetry @xcite . in ref .
@xcite , the hqm was extended to strange baryons .
specifically , the authors considered a hamiltonian whose su(6 ) invariant part is the same as in the hypercentral model @xcite , while the su(6 ) symmetry is broken by a grsey - radicati - inspired interaction @xcite .
the complete hamiltonian is given by @xmath89 where @xmath90 and @xmath91 the grsey - radicati - inspired interaction , @xmath92 , which describes the splittings within the su(6 ) baryon multiplets @xcite , is written in terms of casimir operators as , @xmath93 where @xmath94 , @xmath95 , @xmath96 , @xmath97 and @xmath98 are free parameters fitted to the data with the values reported in ref .
the @xmath0 pair - creation model is an effective model to compute open - flavor strong decays @xcite . here , a hadron decay takes place in its rest frame and proceeds via the creation of an additional @xmath10 pair .
the quark - antiquark pair is created with the quantum numbers of the vacuum , i.e. @xmath11 ( see fig .
[ fig:3p0decay ] ) , and the decay amplitude can be expressed as @xcite @xmath99 pair - creation model of hadron vertices ; the @xmath10 pair ( 45 ) is created in a @xmath0 flavor - color singlet .
a is the initial - state baryon ; b and c are the final baryon and meson states , respectively . ] in this paper , we focus on the open - flavor strong decays of light baryons ( i.e. made up of @xmath100 quarks ) in the @xmath0 model .
we assume harmonic oscillator wave functions , depending on a single oscillator parameter @xmath101 for the baryons and @xmath102 for the mesons .
the final state is characterized by the relative orbital angular momentum @xmath103 between @xmath95 and @xmath96 and a total angular momentum @xmath104 .
cc state & @xmath105 ( gev ) + @xmath34 & 1.10 + @xmath106 & 1.10 + @xmath107 & 0.72 + @xmath108 & 0.72 + @xmath109 & 0.85 + @xmath110 & 0.85 + the coefficient @xmath111 is the phase space factor for the decay .
we show three possible prescriptions .
the first in the non - relativistic expression , @xmath112 depending on the relative momentum @xmath113 between @xmath95 and @xmath96 and on the masses of the two intermediate - state hadrons , @xmath114 and @xmath115 .
the second option is the standard relativistic form , @xmath116 depending on @xmath113 and on the energies of the two intermediate - state hadrons , @xmath117 and @xmath118 .
the third possibility is to use an effective phase space factor @xcite , @xmath119 where @xmath120 and @xmath121 are the effective baryon and meson masses , respectively , evaluated by means of a spin - independent interaction ( see table [ tab : eff - masses ] ) . according to ref .
@xcite , this is valid in the weak - binding limit , where @xmath108 and @xmath107 are degenerate and @xmath122 . in the case of heavy baryons and mesons ,
whose internal dynamics is almost non - relativistic and the hyperfine interactions are relatively small , the three types of phase space factors provide almost the same results .
the transition operator of the @xmath123 model is given by @xcite : @xmath124^{(0)}_0 \ , b_4^{\dagger}(\vec{p}_4 ) \
, d_5^{\dagger}(\vec{p}_5 ) \mbox { } .
\label{3p0}\end{aligned}\ ] ] here , @xmath125 is the pair - creation strength , and @xmath126 and @xmath127 are the creation operators for a quark and an antiquark with momenta @xmath128 and @xmath129 , respectively .
the @xmath10 pair is characterized by a color - singlet wave function @xmath130 , a flavor - singlet wave function @xmath131 , a spin - triplet wave function @xmath132 with spin @xmath133 and a solid spherical harmonic @xmath134 , since the quark and antiquark are in a relative @xmath135-wave .
@xmath136 is the pair - creation vertex or quark form factor .
if one does not consider the quark form factor , i.e. @xmath137 , the vertex reduces to a constant ( @xmath138 ) , see appendix [ pair - creation vertex ] . in our calculations ,
we introduce a strange quark - pair suppression mechanism in the baryon sector analogously to what was done in charmonium to suppress unphysical heavy quark pair - creation @xcite .
the mechanism consists in introducing a reduction factor @xmath139 in the pair - creation operator which suppresses the creation of strange quark pairs ( @xmath140 ) by a factor of @xmath141 with respect to the nonstrange quark pairs ( @xmath142 and @xmath143 ) .
this particular choice for the pair - creation strength breaks the @xmath22 flavor symmetry and suppresses the creation of @xmath144 pairs .
its effect can not be re - absorbed in a redefinition of the model parameters or in a different choice of the @xmath0 model vertex factor . in appendix [ ap1 ]
we present the derivation of the transition amplitudes in the @xmath21 pair - creation model using the jacobi coordinates of eq .
( [ jacobi ] ) and including the effects of a gaussian smearing of the pair - creation vertex . in appendix [ ap2 ]
we present a derivation of the flavor coupling coefficients including the strangeness suppression mechanism which is valid for both pseudoscalar and vector mesons . to better reproduce the experimental data , we introduce a mixing between @xmath145 and @xmath146 resonances , @xmath147 where @xmath148 is the mixing angle .
this was done in refs .
@xcite , to correct the disagreement between experimental and theoretical results for the helicity amplitudes of the @xmath145 and @xmath146 resonances . in ref .
@xcite this problem was solved with a mixture of pseudoscalar meson - baryon and vector meson - baryon in a coupled - channels scheme .
in this section , we present the results of our calculations of the open - flavor strong decays of non - strange baryons and hyperons into baryon - pseudoscalar and baryon - vector mesons , using the u(7 ) model ( see sec .
[ open - flavor strong decays calculated using the u(7 ) model spectrum ] ) and hqm model ( see sec . [
open - flavor strong decays calculated using the hqm model spectrum ] ) .
the decay amplitudes are computed in the @xmath0 model of refs .
@xcite and sec .
[ 3p0 model ] . here
, we show our results for the open - flavor decays by using the u(7 ) algebraic model of sec .
[ u(7 ) algebraic model ] and refs .
@xcite .
cclcl parameter & & + @xmath125 & 14.3 & & 13.2 & + @xmath149 & 2.99 & gev@xmath150 & 2.69 & gev@xmath150 + @xmath151 & 2.38 & gev@xmath150 & 2.02 & gev@xmath150 + @xmath152 & 0.52 & gev@xmath150 & 0.82 & gev@xmath150 + @xmath153 & 0.33 & gev & & + @xmath154 & 0.55 & gev & & + ccccccccccccc model&resonance & status & @xmath155 [ mev ] & @xmath2 & @xmath6 & @xmath156 & @xmath157 & @xmath158 & @xmath159 & @xmath8 & + & @xmath160 & * * * * & 1430 - 1470 & @xmath161 & @xmath162 & & & @xmath163 & & & exp
. + u(7)&@xmath164 $ ] & & 1444 & 85 & & & & 13 & & & rpsf + u(7)&@xmath164 $ ] & & 1444 & 108 & & & & 22 & & & epsf + hqm&@xmath165 $ ] & & 1550 & 105 & & & & 12 & & & rpsf + + & @xmath166 & * * * * & 1515 - 1530 & 102 & 0 & & & 342 & & & exp .
+ u(7)&@xmath167 $ ] & & 1563 & 134 & 0 & & & 207 & & & rpsf + u(7)&@xmath167 $ ] & & 1563 & 102 & 0 & & & 342 & & & epsf + hqm&@xmath168 $ ] & & 1525 & 111 & 0 & & & 206 & & & rpsf + + & @xmath145 & * * * * & 1520 - 1555 & @xmath169 & @xmath170 & & & @xmath171 & & & exp
. + u(7)&@xmath172 $ ] & & 1563 & 63 & 75 & & & 16 & & & rpsf + u(7)&@xmath172 $ ] & & 1563 & 106 & 86 & & & 14 & & & epsf + hqm&@xmath173 $ ] & & 1525 & 84 & 50 & & & 6 & & & rpsf + + & @xmath146 & * * * * & 1640 - 1680 & @xmath174 & @xmath175 & & @xmath176 & @xmath177 & & & exp .
+ u(7)&@xmath178 $ ] & & 1683 & 41 & 72 & & 0 & 18 & & & rpsf + u(7)&@xmath178 $ ] & & 1683 & 71 & 83 & & & 15 & & & epsf + hqm&@xmath179 $ ] & & 1574 & 51 & 29 & & 0 & 4 & & & rpsf + + & @xmath180 & * * * * & 1670 - 1685 & @xmath181 & @xmath182 & & @xmath183 & @xmath184 & & & exp .
+ u(7)&@xmath185 $ ] & & 1683 & 47 & 11 & & 0 & 108 & & & rpsf + u(7)&@xmath185 $ ] & & 1683 & 29 & 7 & & & 79 & & & epsf + hqm&@xmath186 $ ] & & 1579 & 41 & 9 & & & 85 & & & rpsf + + & @xmath187 & * * * * & 1675 - 1690 & @xmath188 & @xmath189 & & & @xmath190 & & & exp .
+ u(7)&@xmath191 $ ] & & 1737 & 121 & 1 & & 0 & 100 & & & rpsf + u(7)&@xmath191 $ ] & & 1737 & 63 & 0 & & & 99 & & & epsf + hqm&@xmath192 $ ] & & 1798 & 91 & 0 & 0 & 0 & 92 & & & rpsf + + & @xmath193 & * * * & 1650 - 1750 & @xmath194 & @xmath195 & & @xmath196 & @xmath197 ( @xmath61 ) & & & exp .
+ & & & & & & & & @xmath198 ( @xmath97 ) & + u(7)&@xmath199 $ ] & & 1683 & 9 & 3 & & 0 & 561 & & & rpsf + u(7)&@xmath199 $ ] & & 1683 & 5 & 2 & & & 657 & & & epsf + hqm&@xmath200 $ ] & & 1606 & 0 & 0 & 0 & 0 & 0 & & & rpsf + + & @xmath201 & * * * & 1680 - 1740 & @xmath202 & @xmath203 & & @xmath204 & @xmath205 & @xmath204 & & exp . + u(7)&@xmath206 $ ] & & 1683 & 5 & 9 & 0 & 3 & 56 & & & rpsf + u(7)&@xmath206 $ ] & & 1683 & 11 & 9 & & & 58 & & & epsf + hqm&@xmath207 $ ] & & 1808 & 18 & 12 & 0 & 14.1 & 70 & & & rpsf + + & @xmath208 & * * * * & 1650 - 1750 & @xmath209 & @xmath210 & & @xmath211 & @xmath212 & @xmath213 & & exp .
+ u(7)&@xmath214 $ ] & & 1737 & 111 & 7 & 0 & 14 & 36 & 5 & 0 & rpsf + u(7)&@xmath214 $ ] & & 1737 & 123 & 7 & & & 28 & & & epsf + hqm&@xmath215 $ ] & & 1797 & 141 & 8 & 0 & 12 & 30 & 77 & 5 & rpsf + + & @xmath216 & * * * & 1820 - 1920 & @xmath217 & @xmath218 & @xmath219 & & @xmath220 ( @xmath61 ) & @xmath221 & @xmath222 & exp .
+ & & & & & & & & @xmath223 ( @xmath97 ) & & & exp .
+ u(7)&@xmath224 $ ] & & 1975 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & rpsf + u(7)&@xmath224 $ ] & & 1975 & 0 & 0 & & & 0 & & & epsf + hqm&@xmath225 $ ] & & 1899 & 14 & 8 & 2 & 0 & 560 & 80 & 82 & rpsf + + & @xmath226 & * * * & 1875 - 1935 & @xmath227 & @xmath228 & @xmath229 & @xmath230 & & & @xmath231 & exp .
+ u(7)&@xmath232 $ ] & & 1874 & 11 & 12 & 1 & 13 & 63 & 64 & 24 & rpsf + u(7)&@xmath232 $ ] & & 1874 & 17 & 11 & & & 33 & & & epsf + hqm&@xmath233 $ ] & & 1853 & 15 & 12 & 1 & 13 & 70 & 53 & 23 & rpsf + ccccccccccc model&resonance & status & @xmath155 [ mev ] & @xmath2 & @xmath156 & @xmath158 & @xmath234 & @xmath235 & @xmath159 & + & @xmath236 & * * * * & 1230 - 1234 & @xmath237 & & & & & & exp .
+ u(7)&@xmath238 $ ] & & 1246 & 71 & & & & & & rpsf + u(7)&@xmath238 $ ] & & 1246 & 115 & & & & & & epsf + hqm&@xmath238 $ ] & & 1240 & 63 & & & & & & rpsf + + & @xmath239 & * * * & 1550 - 1700 & @xmath240 & & @xmath241 & & & @xmath242 & exp . + u(7)&@xmath243 $ ] & & 1660 & 17 & & 65 & & & & rpsf + u(7)&@xmath243 $ ] & & 1660 & 24 & & 74 & & & & epsf + hqm&@xmath244 $ ] & & 1727 & 31 & & 69 & & & & rpsf + + & @xmath245 & * * * * & 1615 - 1675 & @xmath246 & & @xmath247 & & & @xmath248 & exp .
+ u(7)&@xmath249 $ ] & & 1649 & 5 & & 76 & & & & rpsf + u(7)&@xmath249 $ ] & & 1649 & 10 & & 61 & & & & epsf + hqm&@xmath250 $ ] & & 1584 & 9 & & 59 & & & & rpsf + + & @xmath251 & * * * * & 1670 - 1770 & @xmath252 & & @xmath253 ( @xmath61 ) & @xmath254 & & @xmath255 & exp .
+ & & & & & & @xmath256 ( @xmath97 ) & & & + u(7)&@xmath257 $ ] & & 1649 & 46 & & 311 & & & & rpsf + u(7)&@xmath257 $ ] & & 1649 & 27 & & 343 & & & & epsf + hqm&@xmath258 $ ] & & 1584 & 40 & & 333 & & & & rpsf + + & @xmath259 & * * * * & 1855 - 1910 & @xmath260 & & @xmath261 & & & @xmath262 & exp .
+ u(7)&@xmath263 $ ] & & 1921 & 31 & 1 & 188 & 19 & 0 & 99 & rpsf + u(7)&@xmath263 $ ] & & 1921 & 14 & & 139 & 14 & & & epsf + hqm&@xmath264 $ ] & & 1844 & 26 & 0 & 182 & 15 & & 88 & rpsf + + & @xmath265 & * * * * & 1860 - 1910 & @xmath266 & @xmath267 & @xmath268 & & & & exp .
+ u(7)&@xmath269 $ ] & & 1921 & 26 & 38 & 32 & 4 & & 64 & rpsf + u(7)&@xmath269 $ ] & & 1921 & 39 & & 27 & 3 & & & epsf + hqm&@xmath269 $ ] & & 1871 & 49 & 38 & 34 & 4 & & 60 & rpsf + + & @xmath270 & * * * & 1900 - 1970 & @xmath271 & @xmath272 & @xmath273 ( @xmath135 ) & @xmath274 & & 0 & exp . + & & & & & & @xmath275 ( @xmath276 ) & & + u(7)&@xmath277 $ ] & & 1921 & 7 & 23 & 132 & 22 & 5 & 105 & rpsf + u(7)&@xmath277 $ ] & & 1921 & 14 & & 96 & 15 & & & epsf + hqm&@xmath278 $ ] & & 1856 & 17 & 22 & 137 & 20 & & 102 & rpsf + + & @xmath279 & * * * & 1920 - 1970 & @xmath280 & & & & & & exp .
+ u(7)&@xmath281 $ ] & & 1946 & 0 & 0 & 0 & 0 & 0 & 0 & rpsf + u(7)&@xmath281 $ ] & & 1946 & 0 & & 0 & 0 & & & epsf + + & @xmath282 & * * * * & 1940 - 1960 & @xmath283 & @xmath284 & @xmath285 & & & @xmath286 & exp .
+ u(7)&@xmath287 $ ] & & 1921 & 172 & 5 & 92 & 1 & 0 & 22 & rpsf + u(7)&@xmath287 $ ] & & 1921 & 72 & & 40 & 1 & & & epsf + hqm&@xmath288 $ ] & & 1851 & 146 & 3 & 70 & 1 & & 16 & rpsf + ccccccccccccccccc model&baryon & status & @xmath155 [ mev ] & @xmath289 & @xmath290 & @xmath291 & @xmath292 & @xmath293 & @xmath294 & @xmath295 & @xmath296 & @xmath297 & @xmath298 & @xmath299 & + & @xmath300 & * * * & 1630 - 1690 & @xmath301 & seen & seen & & & & & & & & & exp .
+ @xmath302 @xmath20&@xmath164 $ ] & & 1604 & 3 & 38 & 14 & & & & 7 & & & & & rpsf + & @xmath302@xmath164 $ ] & & @xmath3021704 & @xmath302 4 & @xmath302 45 & @xmath302 18 & & & & @xmath302 6 & & & & & rpsf + & @xmath303 & * * * * & 1665 - 1685 & @xmath304 & @xmath305 & @xmath306 & & & & & & & & & exp .
+ @xmath302 @xmath20&@xmath199 $ ] & & 1711 & 5 & 78 & 8 & & & & 36 & & & & & rpsf + @xmath302 hqm&@xmath302@xmath167 $ ] & & @xmath3021799 & @xmath302 4 & @xmath302 62 & @xmath302 7 & & & & @xmath302 29 & & & & & rpsf + & @xmath307 & * * * & 1730 - 1800 & @xmath308 & @xmath309 & seen & @xmath310 & & & & & & & & exp . + @xmath302 @xmath20&@xmath178 $ ] & & 1711 & 3 & 109 & 3 & 28 & & & 9 & & & & & rpsf + @xmath302 hqm&@xmath302@xmath172 $ ] & & @xmath3021799 & @xmath302 5 & @xmath302 151 & @xmath302 6 & @xmath302 27 & & & @xmath302 7 & & & & & rpsf + & @xmath311 & * * * * & 1770 - 1780 & @xmath312 & @xmath313 & @xmath314 & & & & @xmath315 & & & & & exp .
+ @xmath302 @xmath20&@xmath185 $ ] & & 1822 & 101 & 17 & 38 & 0 & & 4 & 11 & & & & & rpsf + @xmath302 hqm&@xmath302@xmath185 $ ] & & @xmath3021914 & @xmath302 79 & @xmath302 13 & @xmath302 30 & @xmath302 0 & & @xmath302 2 & @xmath302 7 & & & & & rpsf + & @xmath316 & * * * * & 1900 - 1935 & @xmath317 & seen & seen & & & & @xmath318 & & & & & exp .
+ @xmath302 @xmath20&@xmath191 $ ] & & 1872 & 6 & 58 & 33 & 2 & 1 & 96 & 23 & 6 & & 2 & & rpsf + @xmath302 hqm&@xmath302@xmath191 $ ] & & @xmath3021906 & @xmath302 4 & @xmath302 44 & @xmath302 26 & @xmath302 1 & @xmath302 0 & @xmath302 88 & @xmath302 22 & @xmath302 5 & & @xmath302 2 & & rpsf + & @xmath319 & * * * & 1900 - 1950 & @xmath320 & seen & seen & & & seen & seen & seen & & & & exp .
+ @xmath302 @xmath20&@xmath321 $ ] & & 1974 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & & 0 & & rpsf + @xmath302 hqm&@xmath302 @xmath199 $ ] & & @xmath3021914 & @xmath302 31 & @xmath302 6 & @xmath302 11 & @xmath302 0 & @xmath302 0 & @xmath302 554 & @xmath302 99 & @xmath302 251 & & @xmath302 0 & & rpsf + & @xmath322 & * * * * & 1383 - 1385 & & @xmath323 & @xmath324 & & & & & & & & & exp .
+ @xmath302 @xmath20&@xmath325 $ ] & & 1382 & & 3 & 27 & & & & & & & & & rpsf + @xmath302 hqm&@xmath302@xmath325 $ ] & & @xmath3021372 & & @xmath302 3 & @xmath302 24 & & & & & & & & & rpsf + & @xmath326 & * * * * & 2025 - 2040 & @xmath327 & @xmath328 & @xmath327 & & @xmath329 & @xmath330 & @xmath331 & & & & & exp .
+ @xmath302 @xmath20&@xmath332 $ ] & & 2012 & 54 & 37 & 75 & @xmath302 8 & 1 & 30 & 37 & 7 & 0 & 4 & 0 & rpsf + @xmath302 hqm&@xmath302@xmath332 $ ] & & @xmath3022085 & @xmath302 44 & @xmath302 30 & @xmath302 62 & @xmath302 6 & @xmath302 1 & @xmath302 22 & @xmath302 27 & @xmath302 5 & @xmath302 0 & @xmath302 2 & @xmath302 0 & rpsf + ccccccccccccc model&baryon & status & @xmath155 [ mev ] & @xmath289 & @xmath290 & @xmath333 & @xmath293 & @xmath295 & @xmath296 & @xmath297 & @xmath334 & + & @xmath335 & * * * & 1560 - 1700 & @xmath336 & @xmath337 & & & & & & & exp .
+ @xmath302 @xmath20&@xmath165 $ ] & & 1577 & 79 & 40 & & & 19 & & & & rpsf + @xmath302 hqm & @xmath302@xmath165 $ ] & & @xmath3021627 & @xmath30293 & @xmath302 46 & @xmath302 & @xmath302 & @xmath302 18 & @xmath302 & @xmath302 & @xmath302 & @xmath302rpsf + & @xmath338 & * * * * & 1660 - 1680 & @xmath339 & @xmath254 & @xmath340 & & & & & & exp .
+ @xmath302 @xmath20&@xmath173 $ ] & & 1686 & 217 & 33 & 9 & & 7 & & & & rpsf + @xmath302 hqm & @xmath302@xmath173 $ ] & & @xmath3021722 & @xmath302 267 & @xmath302 39 & @xmath302 0 & @xmath302 & @xmath302 5 & @xmath302 & @xmath302 & @xmath302 & @xmath302rpsf + & @xmath341 & * * * * & 1685 - 1690 & @xmath342 & @xmath343 & & & & & & & exp .
+ @xmath302 @xmath20&@xmath168 $ ] & & 1686 & 150 & 16 & 0 & & 168 & & & & rpsf + @xmath302 hqm&@xmath302@xmath168 $ ] & & @xmath3021722 & @xmath302 119 & @xmath302 23 & @xmath302 0 & @xmath302
& @xmath302 168 & @xmath302 & @xmath302 & @xmath302 & @xmath302rpsf + & @xmath344 & * * * & 1720 - 1850 &
@xmath345 & seen & & & seen & & & & exp .
+ @xmath302 @xmath20&@xmath346 $ ] & & 1799 & 0 & 67 & 85 & & 13 & & & & rpsf + @xmath302 hqm & @xmath302@xmath346 $ ] & & @xmath3021837 & @xmath3020 & @xmath302 98 & @xmath302 90 & @xmath302
& @xmath302 10 & @xmath302 &
@xmath302 & @xmath302 & @xmath302rpsf + & @xmath347 & * * * & 1750 - 1850 & @xmath348 & @xmath349 & & & seen & & & & exp .
+ @xmath302 @xmath20 & @xmath207 $ ] & & 1799 & 16 & 4 & 2 & & 40 & & & & rpsf + @xmath302 hqm & @xmath302@xmath350 $ ] & & @xmath3021973 & @xmath302 0 & @xmath302 0 & @xmath302 0 & @xmath302
& @xmath302 0 & @xmath302 & @xmath302 & @xmath302 & @xmath302rpsf + & @xmath351 & * * * * & 1815 - 1825 & @xmath352 & @xmath353 & & & @xmath354 & & & & exp .
+ @xmath302 @xmath20&@xmath192 $ ] & & 1849 & 78 & 31 & 1 & 0 & 73 & & & & rpsf + @xmath302 hqm & @xmath302@xmath192 $ ] & & @xmath3021829 & @xmath302 57 & @xmath302 22 & @xmath302 0 & @xmath302 0 & @xmath302 66 & @xmath302 & @xmath302 & @xmath302 & @xmath302rpsf + & @xmath355 & * * * * & 1810 - 1830 & @xmath356 & @xmath357 & & & @xmath358 & & & & exp .
+ @xmath302 @xmath20&@xmath186 $ ] & & 1799 & 0 & 99 & 9 & 0 & 82 & & & & rpsf + @xmath302 hqm&@xmath302@xmath186 $ ] & & @xmath3021837 & @xmath302 0 & @xmath302 84 & @xmath302 7 & @xmath3020 & @xmath302 64 & @xmath302 & @xmath302 & @xmath302 & @xmath302rpsf + & @xmath359 & * * * * & 1850 - 1910 & @xmath360 & @xmath361 & & & seen & & & & exp .
+ @xmath302 @xmath20&@xmath215 $ ] & & 1849 & 96 & 69 & 31 & 2 & 30 & 28 & & & rpsf + @xmath302 hqm & @xmath302@xmath215 $ ] & & @xmath3021829 & @xmath302 120 & @xmath302 79 & @xmath302 11 & @xmath302 1 & @xmath302 24 & @xmath302 27 & @xmath302
& @xmath302 & @xmath302rpsf + & @xmath362 & * * * * & 2090 - 2140 & @xmath363 & @xmath364 & & & seen & & & & exp .
+ @xmath302 @xmath20&@xmath365 $ ] & & 2074 & 0 & 14 & 3 & 1 & 136 & 0 & 38 & 16 & rpsf + @xmath302 hqm & @xmath302@xmath366 $ ] & & @xmath3021995 & @xmath302 167 & @xmath302 20 & @xmath302 0 & @xmath302 3 & @xmath302 69 & @xmath302 15 & @xmath302 79 & @xmath302 25 & @xmath302rpsf + & @xmath367 & * * * * & 1402 - 1410 & & @xmath368 & & & & & & & exp .
+ @xmath302 @xmath20&@xmath369 $ ] & & 1641 & & 230 & & & & & & & rpsf + @xmath302 hqm&@xmath302@xmath369 $ ] & & @xmath3021658 & @xmath302 & @xmath302 222 & @xmath302 & @xmath302 & @xmath302 & @xmath302 & @xmath302 & @xmath302 & @xmath302rpsf + & @xmath370 & * * * * & 1518 - 1520 & @xmath371 & @xmath372 & & & 1 & & & & exp .
+ @xmath302 @xmath20&@xmath373 $ ] & & 1641 & 10 & 17 & & & & & & & rpsf + @xmath302 hqm&@xmath302@xmath373 $ ] & & @xmath3021658 & @xmath302 8 & @xmath302 13 & @xmath302
& @xmath302 & @xmath302 & @xmath302 & @xmath302 & @xmath302
& @xmath302rpsf + ccccccccc model&baryon & status & @xmath155 [ mev ] & @xmath374 & @xmath375 & @xmath376 & @xmath377 & + & @xmath378 & * * * & 1680 - 1700 & & & & & exp . + @xmath302 @xmath20&@xmath172 $ ] & & 1828 & 58 & 85 & 14 & 0 & rpsf + @xmath302 hqm&@xmath302@xmath173 $ ] & & @xmath3021938 & @xmath302 55 & @xmath30286 & @xmath302 15 & @xmath302 0 & @xmath302rpsf + & @xmath379 & * * * & 1818 - 1828 & @xmath380 & @xmath381 & @xmath382 & @xmath380 & exp .
+ @xmath302 @xmath20&@xmath167 $ ] & & 1828 & 38 & 26 & 6 & 55 & rpsf + @xmath302 hqm&@xmath302@xmath168 $ ] & & @xmath3021938 & @xmath302 29 & @xmath302 21 & @xmath302 5 & @xmath302 55 & @xmath302rpsf + & @xmath383 & * * * * & 1531 - 1532 & & & @xmath384 & & exp .
+ @xmath302 @xmath20&@xmath385 $ ] & & 1524 & & & 11 & & rpsf + @xmath302 hqm&@xmath302@xmath385 $ ] & & @xmath3021511 & @xmath302
& @xmath302 & @xmath302 9 & @xmath302 & @xmath302rpsf + the strong decay widths are computed in the @xmath21 model using eqs .
( [ eqn:3p0-decays - abc ] ) , ( [ eqn : mabc ] ) and ( [ eqn : epsilon ] ) , by considering two possible choices for the phase space factor : the standard relativistic form of eq .
( [ eqn : rel - psf ] ) and the effective phase space factor of eq .
( [ eqn : eff - psf ] ) .
the results obtained with the relativistic phase space factor ( rpsf ) and the model parameters of the second column of table [ tab : parameters ] are reported in tables [ tab : nuc][tab : omega ] ; those obtained with the effective phase space factor ( epsf ) and the model parameters of the third column of table [ tab : parameters ] are reported in tables [ tab : nuc][tab : del ] .
the @xmath0 model parameters of table [ tab : parameters ] are fitted to a sample of 9 transitions , as discussed in app .
[ pair - creation vertex ] . in our calculations ,
whenever available we use the experimental values for the masses of the decaying resonances from the pdg @xcite , otherwise the theoretical predictions of sec .
[ u(7 ) algebraic model ] .
ccccccccccccc @xmath34 & mass & @xmath2 & @xmath6 & @xmath156 & @xmath157 & @xmath158 & @xmath235 & @xmath159 & @xmath8 & @xmath386 & @xmath387 & @xmath388 + + + + @xmath389 $ ] & 1713 & 0 & 0 & 0 & 0 & 0 & & & & & & + @xmath390
$ ] & 1796 & 0 & 3 & 5 & 0 & 65 & & 7 & 7 & & &
+ @xmath391 $ ] & 1874 @xmath392 & 106 & 10 & 0 & 3 & 79 & & 161 & 8 & & & + @xmath393 $ ] & 1874 & 0 & 0 & 0 & 0 & 0 & & 0 & 0 & & & + @xmath394 $ ] & 1975 @xmath392 & 1 & 8 & 23 & 0 & 19 & 1 & 9 & 9 & & & + @xmath395 $ ] & 1975 @xmath392 & 0 & 4 & 11 & 0 & 109 & 5 & 14 & 14 & & &
+ @xmath396 $ ] & 1975 @xmath392 & 6 & 3 & 1 & 0 & 176 & 6 & 16 & 16 & & & + @xmath397 $ ] & 1975 @xmath392 & 25 & 13 & 4 & 0 & 99 & 0 & 5 & 4 & & & + @xmath398 $ ] & 1975 @xmath392 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & & &
+ @xmath399 $ ] & 2094 & 5 & 1 & 1 & 5 & 3 & 2 & 48 & 6 & 2 & 2 & 14 + @xmath400 $ ] & 2094 @xmath392 & 27 & 0 & 0 & 1 & 23 & 1 & 53 & 11 & 0 & 2 & 13 + @xmath401 $ ] & 1829 @xmath392 & 42 & 7 & 0 & 1 & 38 & & 0 & 0 & & & + @xmath402 $ ] & 1933 & 8 & 12 & 3 & 0 & 0 & 0 & 0 & 0 & & & + @xmath403 $ ] & 1933 & 0 & 0 & 3 & 0 & 0 & 0 & 0 & 0 & & &
+ @xmath404 $ ] & 1933 & 0 & 2 & 5 & 0 & 1 & 0 & 0 & 0 & & &
+ + + + @xmath395 $ ] & 1835 & 4 & 8 & 7 & 0 & 97 & - & 8 & 7 & - & - & - + @xmath405 $ ] & 1836 & 0 & 0 & 0 & 0 & 0 & - & 0 & 0 & - & - & - + @xmath406 $ ] & 1836 & 0 & 0 & 0 & 0 & 0 & - & 0 & 0 & - & - & - + @xmath394 $ ] & 1839 @xmath392 & 8 & 16 & 15 & 0 & 27 & - & 6 & 5 & - & - & - + @xmath397 $ ] & 1840 @xmath392 & 12 & 4 & 0 & 0 & 25 & - & 0 & 0 & - & - & - + @xmath396 $ ] & 1844 @xmath392 & 3 & 1 & 0 & 0 & 137 & - & 9 & 8 & - & - & - + @xmath396 $ ] & 1851 @xmath392 & 3 & 1 & 0 & 0 & 137 & - & 9 & 9 & - & - & - + @xmath390 $ ] & 1863 @xmath392 & 0 & 4 & 22 & 0 & 83 & - & 12 & 12 & - & - & - + @xmath407 $ ] & 1887 @xmath392 & 0 & 22&119 & 0 & 87 & - & 32 & 32 & - & - & - + @xmath402 $ ] & 1937 & 0 & 0 & 0 & 0 & 0 & - & 0 & 0 & - & - & - + @xmath404 $ ] & 1942 @xmath392 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & - & - & - + @xmath408 $ ] & 1943 @xmath392 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & - & - & - + @xmath403 $ ] & 1969 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & - & - & - + cccccccc @xmath106 & mass & @xmath2 & @xmath156 & @xmath158 & @xmath234 & @xmath235 & @xmath159 + + + + @xmath409 $ ] & 1764 @xmath392 & 0 & 1 & 70 & & & 23 + @xmath410 $ ] & 1946 & 0 & 0 & 0 & 0 & 0 & 0 + @xmath411 $ ] & 1947 & 1 & 3 & 106 & 5 & 1 & 80 + @xmath412 $ ] & 1947 @xmath392 & 18 & 1 & 107 & 18 & 4 & 32 + @xmath413 $ ] & 1904 @xmath392 & 0 & 0 & 0 & 0 & 0 & 0 + @xmath414 $ ] & 1904 & 0 & 0 & 0 & 0 & 0 & 0 + + + + @xmath409 $ ] & 1832 @xmath392 & 0 & 2 & 89 & 7 & - & 57 + @xmath411 $ ] & 1843 & 4 & 1 & 43 & 1 & - & 51 + @xmath413 $ ] & 1947 @xmath392 & 0 & 0 & 1 & 0 & 0 & 0 + @xmath414 $ ] & 1947 @xmath392 & 0 & 0 & 1 & 0 & 0 & 0 + @xmath412 $ ] & 1859 @xmath392 & 10 & 0 & 97 & 7 & - & 13 + @xmath415 $ ] & 2103 & 0 & 0 & 0 & 0 & 0 & 0 + ccccccccccccccccc @xmath416 & mass & @xmath289 & @xmath290 & @xmath291 & @xmath292 & @xmath293 & @xmath294 & @xmath295 & @xmath417 & @xmath418 & @xmath296 & @xmath297 & @xmath298 & @xmath299&@xmath419 + + + + @xmath407 $ ] & 1822 & 24 & 11 & 6 & 20 & 10 & 5 & 4 & & & & & & &
+ @xmath420 $ ] & 1822 & 22 & 4 & 8 & 0 & 0 & 602 & 98 & & & & & & & + @xmath421 $ ] & 1822 @xmath392 & 1 & 20 & 1 & 1 & 0 & 33 & 9 & & & & & & &
+ @xmath389 $ ] & 1849 @xmath392 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & & & 0 & & & &
+ @xmath422 $ ] & 1872 & 4 & 62 & 19 & 23 & 12 & 17 & 6 & & & 12 & & & & + @xmath390 $ ] & 1926 & 0 & 0 & 0 & 5 & 1 & 65 & 12 & & & 25 & & 3 & & + @xmath423 $ ] & 1999 & 1 & 31 & 1 & 7 & 14 & 28 & 8 & 1 & & 26 & 11 & 6 & 1 & + @xmath391 $ ] & 1999 & 4 & 76 & 6 & 1 & 1 & 60 & 13 & 4 & & 7 & 13 & 21 & 3 & + @xmath424 $ ] & 1999 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & & 0 & 0 & 0 & 0 &
+ @xmath394 $ ] & 2095 & 4 & 2 & 1 & 3 & 4 & 18 & 4 & 4 & 1 & 23 & 5 & 9 & 7 & + @xmath395 $ ] & 2095 & 2 & 1 & 1 & 2 & 2 & 84 & 18 & 13 & 2 & 39 & 7 & 14 & 11 & + @xmath396
$ ] & 2095 @xmath392 & 15 & 3 & 5 & 1 & 0 & 128 & 29 & 18 & 3 & 45 & 8 & 15 & 11 &
+ @xmath397 $ ] & 2095 & 69 & 13 & 24 & 2 & 1 & 54 & 15 & 1 & 0 & 17 & 0 & 3 & 1 & + @xmath398 $ ] & 2095 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & + @xmath401 $ ] & 1957 @xmath392 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & & 0 & & 0 & &
+ @xmath425 $ ] & 1957 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & & 0 & & 0 & &
+ @xmath402 $ ] & 2055 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & & 0 & 0 & 0 & 0 & + @xmath403 $ ] & 2055 & 0 & 0 & 0 & 0 & 0 & 1 & 0 & 0 & & 0 & 0 & 0 & 0 & + @xmath404 $ ] & 2055 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & & 0 & 0 & 0 & 0 &
+ + + + @xmath215 $ ] & 1906 & 4 & 69 & 22 & 19 & 20 & 20 & 7 & & & 21 & & & &
+ @xmath346 $ ] & 1914 & 9 & 7 & 3 & 17 & 22 & 22 & 9 & & & 100 & & & & + @xmath350 $ ] & 2050 @xmath426 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & & & 0 & & & & + @xmath233 $ ] & 2072 & 1 & 31 & 1 & 7 & 16 & 39 & 11 & & & 29 & & & & + @xmath366 $ ] & 2072 & 5 & 89 & 7 & 3 & 3 & 66 & 15 & & & 11 & & & & + @xmath179 $ ] & 2149 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 + @xmath427 $ ] & 2187 & 3 & 2 & 1 & 3 & 4 & 16 & 3 & 5 & 1 & 21 & 6 & 9 & 2 & 16 + @xmath428 $ ] & 2187 & 1 & 1 & 1 & 1 & 2 & 91 & 21 & 17&5 & 40 & 10 & 15 & 3 & 18 + @xmath365 $ ] & 2187 & 19 & 3 & 6 & 1 & 0&147 & 34 & 23&6 & 52 & 10 & 18 & 3 & 49 + @xmath429 $ ] & 2187 & 83 & 15 & 29 & 2 & 1&83 & 21 & 3 & 0 & 30 & 2 & 7 & 1 & 159 + @xmath430 $ ] & 2238 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 + @xmath431 $ ] & 2263 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 + cccccccccccccccc @xmath416 & mass & @xmath289 & @xmath290 & @xmath291 & @xmath292 & @xmath293 & @xmath294 & @xmath295 & @xmath417 & @xmath418 & @xmath296 & @xmath297 & @xmath298 & @xmath299&@xmath419 + + + + @xmath250 $ ] & 1755 & 4 & 5 & 4 & 11 & & 1 & 30 & & & & & & &
+ @xmath258 $ ] & 1755 & 9 & 6 & 14 & 0 & & 181 & 165 & & & & & & &
+ @xmath409 $ ] & 1863 & 0 & 1 & 0 & 1 & 0 & 45 & 39 & & & 5 & & & &
+ @xmath432 $ ] & 2012 & 12 & 18 & 16 & 35 & 14 & 21 & 17 & 0 & & 24 & 6 & 28 & 76 &
+ @xmath278 $ ] & 2012 & 6 & 9 & 8 & 18 & 7 & 79 & 69 & 1 & & 35 & 9 & 40 & 106 &
+ @xmath264 $ ] & 2012 & 11 & 7 & 15 & 1 & 0 & 112 & 101 & 1 & & 37 & 9 & 41 & 106 &
+ @xmath411 $ ] & 2037 @xmath392 & 1 & 1 & 1 & 2 & 1 & 38 & 42 & 0 & 0 & 28 & 12 & 36 & 315 & + @xmath412 $ ] & 2037 & 5 & 4 & 7 & 1 & 0 & 63 & 56 & 1 & 0 & 10 & 2 & 9 & 62 &
+ @xmath433 $ ] & 2037 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 &
+ @xmath244 $ ] & 1765 @xmath392 & 8 & 8 & 9 & 1 & & 13 & 38 & & & & & & &
+ @xmath434 $ ] & 1996 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & & 0 & 0 & 0 & 0 &
+ + + + @xmath435 $ ] & 1883 & 6 & 9 & 8 & 11 & 2 & 68 & 63 & & & 14 & & & &
+ @xmath436 $ ] & 2085 & 6 & 9 & 9 & 18 & 8 & 86 & 76 & 2 & & 39 & 25 & 53 & 19 & + @xmath437
$ ] & 2136 & 0 & 1 & 1 & 2 & 1 & 61 & 63 & 0 & 2 & 29&25 & 43 & 20&1 + @xmath438 $ ] & 2136 & 7 & 5 & 9 & 2 & 0 & 70 & 63 & 1 & 10 & 17 & 7 & 18 & 5 & 4 + @xmath439 $ ] & 2212 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 + ccccccccccc @xmath440 & mass & @xmath289 & @xmath290 & @xmath333 & @xmath293 & @xmath295 & @xmath418 & @xmath296 & @xmath297 & @xmath334 + + + + @xmath420 $ ] & 1799 & 0 & 15 & 1 & & 447 & & & &
+ @xmath389 $ ] & 1826 & 0 & 0 & 0 & 0 & 0 & & & & + @xmath390 $ ] & 1904 & 0 & 3 & 4 & 2 & 54 & & 0 & & 0 + @xmath423 $ ] & 1978 & 27 & 6 & 4 & 10 & 31 & & 56 & 1 & 4 + @xmath391 $ ] & 1978 & 109 & 12 & 2 & 1 & 58 & & 123 & 3 & 16 + @xmath393 $ ] & 1978 & 0 & 0 & 0 & 0 & 0 & & 0 & 0 & 0 + @xmath398 $ ] & 2074 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 + @xmath394 $ ] & 2075 & 0 & 13 & 11 & 12 & 17 & 1 & 0 & 20 & 9 + @xmath395 $ ] & 2075 & 0 & 6 & 6 & 6 & 82 & 4 & 0 & 28 & 13 + @xmath397 $ ] & 2075 & 0 & 51 & 10 & 2 & 57 & 0 & 0 & 1 & 2 + @xmath441 $ ] & 1936 & 0 & 0 & 0 & 0 & 0 & & 0 & & 0
+ @xmath402 $ ] & 2034 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 + @xmath403 $ ] & 2034 & 0 & 0 & 0 & 0 & 1 & 0 & 0 & 0 & 0 + @xmath404
$ ] & 2034 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 + + + + @xmath225 $ ] & 1837 & 0 & 15 & 2 & & 477 & & & &
+ @xmath233 $ ] & 1995 & 38 & 8 & 0 & 10 & 29 & & 55 & 2 & 4 + @xmath442 $ ] & 2072 & 0 & 0 & 0 & 0 & 0 & & 0 & & 0 + @xmath425 $ ] & 2072 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 + @xmath390 $ ] & 2110 & 0 & 0 & 1 & 4 & 35 & 11 & 0 & 41 & 8 + @xmath394 $ ] & 2110 & 0 & 18 & 13 & 12 & 35 & 2 & 0 & 23 & 9 + @xmath395 $ ] & 2110 & 0 & 10 & 6 & 2 & 87 & 7 & 0 & 33 & 14 + @xmath397 $ ] & 2110 & 0 & 50&10 & 2 & 19 & 0 & 0 & 2 & 2 + @xmath443 $ ] & 2160 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 + @xmath402 $ ] & 2186 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 + @xmath403 $ ] & 2186 & 0 & 0 & 0 & 0 & 1 & 0 & 0 & 0 & 0 + @xmath404 $ ] & 2186 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 + ccccccccccc @xmath440 & mass & @xmath289 & @xmath290 & @xmath333 & @xmath293 & @xmath295 & @xmath418 & @xmath296 & @xmath297 & @xmath334 + + + + @xmath444 $ ] & 1756 & 29 & 44 & 14 & & & & & &
+ @xmath445 $ ] & 1891 & 0 & 0 & 0 & 0 & & & 0 & &
+ @xmath446 $ ] & 1939 & 35 & 66 & 36 & 17 & & & 39 & & 6 + @xmath447 $ ] & 1939 & 88 & 85 & 10 & 0 & & & 94 & & 15 + @xmath448 $ ] & 1939 & 0 & 0 & 0 & 0 & & & 0 & & 0 + @xmath449 $ ] & 1896 & 0 & 0 & 0 & 0 & & & 0 & & 0 + + + + @xmath450 $ ] & 2008 & 0 & 1 & 0 & 0 & & & 1 & & 0 + @xmath451 $ ] & 2008 & 0 & 0 & 0 & 0 & & & 0 & & 0 + cccccccccccc @xmath17 & mass & @xmath374 & @xmath375 & @xmath376 & @xmath452 & @xmath453 & @xmath377 & @xmath454&@xmath455 & @xmath456&@xmath457 + + + + @xmath421 $ ] & 1932 & 36 & 6 & 1 & 11 & 7 & 13 & & & &
+ @xmath407 $ ] & 1932 & 43 & 20 & 69 & 0 & 0 & 4 & & & & + @xmath420 $ ] & 1932 & 4 & 7 & 22 & 0 & 216 & 152 & & & &
+ @xmath458 $ ] & 1932 & 23 & 39 & 132 & 0 & 2 & 19 & & & &
+ @xmath389 $ ] & 1957 & 0 & 0 & 0 & 0 & 0 & 0 & & & & + @xmath422 $ ] & 1979 & 198 & 7 & 6 & 47 & 4 & 7 & & & & + @xmath459 $ ] & 1979 & 59 & 5 & 4 & 1 & 20 & 27 & & & &
+ @xmath390 $ ] & 2031 & 2 & 1 & 3 & 0 & 24 & 19 & 2 & & & + @xmath460 $ ] & 1727 & 26 & 4 & 3 & & & 2 & & & &
+ + + + @xmath461 $ ] & 1843 & 125 & 6 & 5 & & & 15 & & & &
+ @xmath462 $ ] & 2053 & 8 & 11 & 37 & 0 & 223 & 154 & & & & + @xmath463 $ ] & 2190 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 + @xmath464 $ ] & 2288 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 + @xmath465 $ ] & 2288 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 + @xmath466 $ ] & 2327 & 3 & 1 & 8 & 1 & 6 & 5 & 8 & 10 & 40 & 1 + @xmath467 $ ] & 2327 & 2 & 1 & 4 & 0 & 35 & 32 & 16 & 16 & 62 & 1 + @xmath468 $ ] & 2327 & 6 & 7 & 24 & 0 & 57 & 53 & 20 & 18 & 69 & 1 + @xmath469 $ ] & 2327 & 26 & 33&108 & 1 & 33 & 33 & 11 & 5 & 16 & 0 + @xmath470 $ ] & 2377 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 + @xmath471 $ ] & 2403 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 + cccccccccccc @xmath17 & mass & @xmath374 & @xmath375 & @xmath376 & @xmath452 & @xmath453 & @xmath377 & @xmath454&@xmath455 & @xmath456&@xmath457 + + + + @xmath250 $ ] & 1869 & 17 & 10 & 7 & 7 & & 7 & & & & + @xmath258 $ ] & 1869 & 5 & 10 & 9 & 0 & & 61 & & & & + @xmath409
$ ] & 1971 & 2 & 1 & 1 & 2 & 51 & 14 & & & &
+ @xmath244 $ ] & 1878 & 19 & 16 & 13 & 1 & & 12 & & & &
+ + + + @xmath244 $ ] & 2022 & 19 & 12 & 13 & 12 & & 24 & &&&
+ @xmath432 $ ] & 2225 & 33 & 19 & 23 & 35 & 33 & 7 & 40 & 34 & 31 & 16 + @xmath278 $ ] & 2225 & 17 & 10 & 12 & 17 & 133 & 29 & 61 & 50 & 44 & 23 + @xmath264 $ ] & 2225 & 14 & 19 & 16 & 3 & 194 & 44 & 66 & 51 & 45 & 24 + @xmath288 $ ] & 2225 & 64 & 87 & 70 & 13 & 56 & 16 & 13 & 3 & 2 & 1 + @xmath413 $ ] & 2352 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 + @xmath414 $ ] & 2352 & 0 & 0 & 0 & 0 & 1 & 0 & 0 & 0 & 0 & 0 + @xmath415 $ ] & 2369 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 + cccccc @xmath472 & mass & @xmath473 & @xmath474&@xmath475 & @xmath476 + + + + @xmath250 $ ] & 1989 & 68 & & &
+ @xmath258 $ ] & 1989 & 20 & & &
+ @xmath409 $ ] & 2085 & 8 & 32 & & + @xmath244 $ ] & 1998 & 79 & & &
+ + + + @xmath250 $ ] & 2142 & 26 & 48 & &
+ @xmath258 $ ] & 2142 & 68 & 403 & & + @xmath244 $ ] & 2162 & 68 & 102 & & + @xmath477 $ ] & 2364 & 109 & 34 & 27 & 155 + @xmath478 $ ] & 2364 & 55 & 137 & 88 & 225 + @xmath479 $ ] & 2364 & 69 & 199 & 117 & 234 + @xmath480 $ ] & 2364 & 308 & 58 & 4 & 23 + @xmath413 $ ] & 2492 & 0 & 0 & 0 & 0 + @xmath414 $ ] & 2492 & 0 & 1 & 0 & 0 + @xmath415 $ ] & 2508 & 0 & 0 & 0 & 0 + below , we provide results for the open - flavor decay widths of strange and non - strange baryons into light baryons plus pseudoscalar or vector mesons in the @xmath0 model formalism of sec .
[ 3p0 model ] , using the hqm results of refs .
@xcite . for the results ,
see tables [ tab : nuc]-[tab : delta - missing ] .
the decays are calculated with the new values of the @xmath0 model parameters of table [ tab : parameters-3p0+hqm ] , which we fitted to a sample of 9 transitions : @xmath481 , @xmath482 , @xmath483 , @xmath484 , @xmath485 , @xmath486 , @xmath487 , @xmath488 and @xmath489 .
although we use the same decay model as in sec .
[ open - flavor strong decays calculated using the u(7 ) model spectrum ] , there are some differences with respect from the previous case : 1 ) as in sec .
[ open - flavor strong decays calculated using the u(7 ) model spectrum ] , for * * * and * * * * states we use the experimental values of the masses , but the quantum number assignments are provided by the hqm and do not always coincide with those of the u(7 ) model , since the two models are different .
for example , this happens for the @xmath146 , @xmath193 and @xmath216 .
thus , in these cases , we expect to obtain quite different results for the decay widths ; 2 ) the hqm and u(7 ) models predict the existence of a few missing states below the energy of 2.1 gev .
the masses , quantum numbers and also quantity of missing states in the two previous models are different .
information on missing states is important to the experimentalists in their search for new baryon resonances .
cc parameter & value + @xmath125 & 13.319 + @xmath149 & 2.758 + @xmath151 & 2.454 + @xmath152 & 0 + @xmath153 & 0.33 + @xmath154 & 0.55 + the quality of our results is comparable to that of refs .
capstick and roberts ( cr ) studied the strong decays of non - strange baryons , nucleon and delta resonances , by using capstick and isgur s relativized baryon model @xcite to describe the masses of unknown resonances , harmonic oscillator wave functions and an effective phase space @xcite ; they did not calculate the open - flavor decays of strange baryons .
our study is more complete , as it includes many more decay channels ( such as , for example , decays into @xmath17 + meson ) and a detailed analysis of the decays of strange baryons .
both calculations , ours and capstick and roberts @xcite , are performed within the @xmath0 model , though there are some differences .
the main one is in the pair - creation mechanism : in refs .
@xcite it does not depend on the flavor of the created @xmath490 pair , while in the present case the strange quark pair - creation is suppressed with respect to the nonstrange pairs .
the effects of this strangeness - suppression mechanism can not be re - absorbed in a redefinition of the model parameters or in a different choice of the @xmath0 model vertex factor .
whether @xmath144 pair - creation has to be suppressed or not may be evaluated by comparing theoretical predictions and experimental data for these particular decay channels .
unfortunately , we think that the large uncertainties on experimental decay widths into channels due to @xmath144 pair - creation do not permit a strong definitive conclusion to be drawn . nevertheless , it is worthwhile to observe that the introduction of a strangeness suppression mechanism may be justified by experimental results concerning the electro - production ratios of @xmath491 , @xmath235 , @xmath492 and @xmath493 baryon meson - states from @xmath494 s @xcite . in our paper , we also compared results obtained with relativistic or effective phase space factors .
our conclusion is that the quality of the results for the amplitudes calculated by using a phase space or the other are similar .
thus , we think that it is preferable to use a relativistic phase space , in order to reduce the number of unnecessary parametrizations . finally , we can say that both studies are characterized by the same problems ( a few results are far from data , like @xmath495 ) and strengths ( a great number of data is fitted with a few parameters ) ; among other things , cr obtained for the decay widths @xmath496 , @xmath497 and @xmath498 , the following theoretical results 216 mev , 778 mev , and 11 mev , respectively , which are to be compared with the experimental data @xmath169 mev , @xmath197 mev and @xmath213 mev , respectively
. we can also compare our results with those of refs .
bijker , iachello and leviatan ( bil ) @xcite computed the open - flavor decay amplitudes within a modified version of the elementary meson emission model ( eme ) , with two parameters , and used the u(7 ) algebraic model to calculate the baryon spectrum .
the eme is an effective model , in which the decay occurs by the emission of a meson from the decaying hadron .
even though the eme and @xmath0 models share some features , there are important differences .
for example , the internal quark dynamics is invisible to the eme vertex , because it does not depend on the meson internal wave function , which brings a non - local character to the @xmath0 matrix elements .
moreover , as in the cr case , bil did not consider a suppression of @xmath499 pair - creation , though it has been shown that this mechanism can be beneficial in meson strong decays @xcite .
it can also be shown that it is impossible to get a correspondence between eme and @xmath0 models , unless the factor in front of the recoil term in the eme is taken equal to @xmath500 . however , this factor has been taken as a free parameter in bil ( and also in other eme studies , refs .
@xcite ) and by fit it came out to be equal 0.04 , which means that the recoil term is practically absent in bil .
this fact , in combination with the non local character of the @xmath0 model , the quark form factor and flavor suppression , explain the differences in the two model predictions .
a more detailed explanation is contained in appendix [ eme vs 3p0 ] .
finally , we can say that the general quality of our results is comparable to those of refs .
@xcite . both of them reproduce the general trend of the data , but , in some cases , they show a few large disagreement with experiments . for example , the results of refs .
@xcite do not agree with the data in the case of the decays @xmath501 and @xmath502 , where they get null amplitudes , while ours do agree . on the contrary , our results for other channels , like @xmath503 and @xmath504
, do not agree with the data , being much larger , while those of refs .
@xcite do agree or they are closer .
there was also an attempt to improve the study of the strong decays using covariant calculations with relativistic constituent quark models ( rcqm ) by melde _
the authors computed the transitions for all @xmath107 , @xmath109 , and @xmath505 strong decay modes of several well - established non - strange and strange baryon states , using the so - called point - form spectator model ( pfsm ) , whose non relativistic limit is the classic eme ; they did not calculate decay amplitudes into baryon - vector meson pairs .
unfortunately , these results still can not provide a satisfactory explanation of the experimental decay widths and , in general , underestimate the available experimental data . finally , it is worthwhile to cite the results of a dynamical coupled - channels study of @xmath506 reactions of ref .
@xcite , where the authors provided a comprehensive analysis of world data of @xmath507 , @xmath508 and @xmath509 reactions , including a coupled channel model for meson production reactions considering all possible final states . as widely discussed in the literature ,
there are several baryons ( mesons ) whose nature may not be a pure three - quark ( quark - antiquark ) one .
some well - known examples are the @xmath510 @xcite and the @xmath511 and @xmath512 @xcite mesons .
some alternative hypotheses ( hadron - hadron molecules , hybrids , tetra / penta - quarks , and so on ) might explain why nave quark models fail to reproduce some of the main properties of these resonances , including mass and decay modes .
let us focus on the @xmath513 . by combining u(7 ) and @xmath0 models ,
we get a mass of 1641 mev and a @xmath290 amplitude of 230 mev ; these numbers can be compared with experimental results , namely @xmath514 mev ( mass ) and @xmath368 mev ( @xmath290 amplitude ) @xcite . such a strong deviation between theoretical predictions and data
might be explained if we interpret the @xmath513 as a baryon - meson molecular state ( for example , see ref .
@xcite ) . given this
, our predictions would refer to the lowest - lying @xmath515 state with the same quantum numbers as the @xmath513 .
the @xmath279 represents another failure of our @xmath0/qm predictions , as we get a null @xmath2 width , while it should be in the range @xmath280 mev @xcite . nevertheless , unlike the @xmath513 case
, here we obtain a prediction for the mass which is compatible with the experimental data @xcite .
possible explanations of this deviation from the data include the eventuality that the @xmath279 is a @xmath388 bound state @xcite or simply the fact that the @xmath2 decay may proceed in a different way
. it would be thus worthwhile to investigate whether the inclusion of baryon - meson higher fock components in @xmath279 s wave function via the unquenched quark model ( uqm ) formalism @xcite may help to solve this problem .
similar issues also occur in the @xmath516 and @xmath517 cases .
@xcite .
another interesting case of departure from experimental data is that of the @xmath251 @xmath158 width @xcite , for which we predict a larger value . on the contrary , in the molecular picture
this decay mode is dynamically suppressed .
it would be worthwhile to investigate this case in the uqm formalism and see if the introduction of continuum components , also determining a renormalization of the @xmath518 wave function , may improve the quality of our result .
it would also be interesting to calculate the couplings for the @xmath234 virtual channel , which is relevant to several reactions .
this coupling has been evaluated in the molecular model in refs .
@xcite .
in addition to a calculation of the decay amplitudes within the uqm , it would also be interesting to investigate baryon - meson - meson decays . a possible way to do that is via the formalism of quasi - two - body decays , discussed in refs .
@xcite . in quasi - two - body decays , the decay of a baryon @xmath94 into a baryon @xmath95 and mesons @xmath519 and @xmath520 proceeds as @xmath521 , where @xmath522 is a baryon resonance . alternately , one may also decide to use the coupled - channel approach .
of particular interest is the @xmath523 decay mode , which is the main one of the @xmath524 ( @xmath525 ) ; it is worthwhile to note that this is larger than the @xmath2 width , even if the latter has more phase space for the decay .
for example , see refs .
we computed the open - flavor strong decays of light baryons ( i.e. made up of @xmath526 , @xmath46 , @xmath527 valence quarks ) into baryon - pseudoscalar and baryon - vector mesons using a modified version of the @xmath0 pair - creation model @xcite , in which we considered a flavor - dependent pair - creation strength to suppress the contributions from heavier @xmath528 pairs , like @xmath144 with respect to @xmath529 ( @xmath530 ) .
the baryon models , which we used in our study to get predictions for missing or higher - lying states , were the @xmath20 @xcite and hypercentral @xcite models .
the possibility of using two models to extract the baryon spectrum makes it possible to give two different points of view , especially in the study of the energy region above @xmath531 gev and the related problem of the missing resonances .
indeed , an important difference between the u(7 ) and hqm models is in the number of missing states that they predict and also in the quantum number predictions for some * * * and * * * * states , like the @xmath216 . in a subsequent paper @xcite , the present results will be extended up to to an energy region ( 2.5 gev ) which will be tested by forthcoming experiments at the jlab .
it is worthwhile to enumerate some of the difficulties and problems connected to this type of calculations .
one problem is related to the difficulty of assigning quantum numbers to resonances within a qm .
sometimes , this can generate strong conflicts between theoretical results and experimental data .
for example , this is the case of the @xmath216 , whose decay amplitudes change significantly whether we use the @xmath0 model in combination with the u(7 ) or hqm models .
see tables [ tab : nuc ] .. another problem has to do with the large quantity of decay thresholds , sometimes lying at similar energies or almost overlapping with one another .
thus , we think that a more complete study would require the introduction of continuum coupling effects ( i.e. higher fock components ) in the baryon wave functions . in some cases , the presence of a threshold can deeply influence the quark structure of a hadron , close in energy , as in the well - known case of the @xmath510 meson @xcite . in this respect ,
it is worthwhile to cite the results of the ebac project , developed by matsuyama @xmath12 @xmath13 @xcite .
this is a dynamical coupled - channel model for investigating the nucleon resonances in the meson production reactions induced by pions and photons .
see also the interesting coupled - channel model results of the bonn - jlich group of refs .
@xcite . a similar formalism , which would make it possible to include meson cloud effects in baryon and meson open- and hidden - flavor decays , will be the subject of a subsequent paper @xcite .
finally , one of the most important points is related to the problem of the missing resonances . is this a matter of degrees of freedom ? in this case , the use of other types of models , characterized by a smaller number of effective degrees of freedoms , such as the quark - diquark one , could , at least partially , solve the problem @xcite .
otherwise , does it have to do with the coupling of these missing states with other types of decay channels , more difficult to observe ?
this is still an open question .
thus , we think that it would be worthwhile to compare the results for spectrum and decays of a three quark qm to those of other type of models , such as the quark - diquark one .
moreover , in the case of higher lying states , it would also be interesting to compare the predictions of three quark models to those for hybrid baryons , where baryons are described as bound states of three constituent quarks and a constituent gluon @xcite .
analogously to what is done in ref .
@xcite , we can study different forms for the pair - creation vertex , to improve the description of the experimental data .
the determination of the best vertex results from a @xmath532-analysis based on a sample of 9 transitions : @xmath481 , @xmath482 , @xmath483 , @xmath484 , @xmath485 , @xmath486 , @xmath487 , @xmath488 and @xmath489 .
the different forms we consider are given by @xmath533 where @xmath534 .
as observed in ref .
@xcite , the forms containing a @xmath535 parameter , such as @xmath536 or @xmath537 $ ] , present a bump around @xmath538 , and thus do not show the expected decreasing behavior .
thus , we do not include them in our analysis .
the quality of the description of the experimental data provided by the four vertices is equivalent ( see table [ tab:9transitions ] ) .
thus , we choose the vertex with the smallest number of free parameters , the first vertex with @xmath539 this is the one used in the calculations of sec .
[ strong decay widths ] and for the analytic derivation of the @xmath0 amplitudes of app .
. the reader may object that we did not consider in our analysis the simplest choice for the pair - creation vertex , namely @xmath138 .
actually , this particular choice is a special case of @xmath540 , i.e. when @xmath137 .
cccccc channel & @xmath541 & @xmath542 & @xmath543 & @xmath544 & exp ( mev ) + @xmath545 & 115 & 118 & 116 & 120 & @xmath237 + @xmath482 & 102 & 98 & 101 & 98 & @xmath546 + @xmath483 & 106 & 108 & 102 & 107 & @xmath547 + @xmath484 & 71 & 72 & 68 & 72 & @xmath548 + @xmath485 & 63 & 55 & 60 & 50 & @xmath549 + @xmath486 & 123 & 114 & 114 & 118 & @xmath550 + @xmath487 & 14 & 14 & 14 & 14 & @xmath551 + @xmath488 & 39 & 42 & 38 & 43 & @xmath552 + @xmath489 & 14 & 16 & 14 & 16 & @xmath553 +
in the following , we list the conventions used for the spin wave functions @xcite : @xmath554 we only show the state with the largest component of the projection @xmath555 .
the other states are obtained by applying the lowering operator in spin space .
the meson and baryon states are written according to the usual prescriptions .
below , we list the conventions used for the flavor wave functions of mesons and baryons . since the mixing angle @xmath556 between @xmath109 and @xmath557 is small , we take @xmath558 .
thus , we identify @xmath559 and @xmath560 . * * the octet mesons * @xmath561 \end{aligned}\ ] ] for the baryon flavor wave functions , @xmath562 , we adopt the convention of ref . @xcite .
we only show the highest charge state @xmath563 with @xmath564 .
the other charge states are obtained by applying the lowering operator in isospin space . * * the octet baryons * @xmath565 \\ & : & \phi_{\lambda}(p ) & = & \frac{1}{\sqrt 6 } [ 2|uud \rangle - |udu \rangle \\ & & & & - |duu \rangle ] \end{array}\ ] ]
@xmath566 \\ & : & \phi_{\lambda}(\sigma^+ ) & = & \frac{1}{\sqrt 6 } [ |suu \rangle + |usu \rangle \\ & & & & - 2|uus \rangle ] \end{array}\ ] ] @xmath567 \\ & : & \phi_{\lambda}(\lambda ) & = & \frac{1}{2 } [ - |dsu \rangle - |sdu \rangle \\ & & & & + |sud \rangle + |usd \rangle ] \end{array}\ ] ] @xmath568 \\ & : & \phi_{\lambda}(\xi^0 ) & = & \frac{1}{\sqrt 6 } [ 2|ssu \rangle \\ & & & & - |sus \rangle - |uss \rangle ] \end{array}\ ] ] * * the decuplet baryons * @xmath569 @xmath570 \end{array}\ ] ] @xmath571 \end{array}\ ] ] @xmath572 * * the singlet baryons * @xmath573 \end{array}\ ] ]
the color , spin and spatial parts of the @xmath21 amplitude @xmath574 , excluding the flavor couplings , were derived in a harmonic oscillator basis by roberts and silvestre - brac ( rsb ) @xcite .
they did not consider a quark form factor ( @xmath137 ) and used the following momenta @xmath575 with conjugate coordinates @xmath576 on the contrary , here we use the standard jacobi coordinates and conjugate momenta for the initial baryon , @xmath94 , and the final state baryon @xmath95 and meson @xmath96 @xcite , @xmath577 where @xmath578 and @xmath579 are the mass and momentum of the quark @xmath580 ( see fig . [ fig:3p0decay ] ) , @xmath581 and @xmath582 the masses of the hadrons @xmath95 and @xmath96 . their center of mass coordinates are defined as @xmath583 the conjugate momenta are given by @xmath584 and @xmath585 the final result is @xcite @xmath586 where the angular momenta of the baryons @xmath94 and @xmath95 , @xmath587 and @xmath588 , are the sum of the @xmath108 and @xmath589 oscillator contributions , @xmath590 and @xmath591 , and the term @xmath592 contains the dependence on the color - spin - flavor part @xmath593 \ , \left [ \begin{array}{ccc } s_b & l_{\lambda_b } & j_b \\ s_c & l_c & j_c \\ s_{bc } & l_{bc } & j_{bc } \end{array } \right ]
\sum_l \hat{l}^2 \nonumber\\ & & \left\ { \begin{array}{ccc } s_a & l_{\lambda_a } & j_a \\ l & s_{bc } & 1 \end{array } \right\ } \left\ { \begin{array}{ccc } s_{bc } & l_{bc } & j_{bc } \\ l & j_a & l \end{array } \right\ } \mbox { } . \end{aligned}\ ] ] the coefficient @xmath594 contains the flavor couplings and is defined as : @xmath595 here , @xmath596 denotes the flavor wave function of the created quark - antiquark pair @xmath597 ~ , \label{new3p0}\ ] ] which , in the limit of equal quark masses , reduces to the usual expression for a flavor singlet .
the coefficients in square brackets are proportional to 9-j coefficients @xmath598
= \hat c \hat f \hat g \hat h \hat i \left\ { \begin{array}{ccc } a & b & c \\ d & e & f \\ g & h & i \end{array } \right\ } \mbox { } , \ ] ] where @xmath599 , finally , @xmath600 represents the spatial contribution @xcite @xmath601 \left [ \begin{array}{ccc } l_3 & l'_3 & 1 \\ l_4 & l'_4 & l_{\lambda_a } \\
l_7 & l_8 & l \end{array } \right ] \left\ { \begin{array}{ccc } l & l_{12 } & l_5 \\ l_6 & l & l_{bc } \end{array } \right\ } \nonumber\\ & & b^{l_{12}}_{l_1l_2 } \ , b^{l_{5}}_{ll_{12 } } \ , b^{l_6}_{l'_1l'_2 } \ , b^{l_7}_{l_3l_4 } \ ,
b^{l_8}_{l'_3l'_4 } \sum_{\lambda , \mu,\nu } d_{\lambda \mu \nu}(x ) \ , i_\nu ( l_5,l_6,l_7,l_8;l ) \nonumber\\ & & \left ( \frac{l'_1+l'_2+l'_3+l'_4 + 2 \mu + \nu + 1}{2 } \right ) !
\frac{q^{l_1+l_2+l_3+l_4 + 2 \lambda + \nu}_0}{g^{2\mu + \nu - l_1-l_2-l_3-l_4 } } \mbox { } , \end{aligned}\ ] ] with @xmath602 ! } } \nonumber\\ b^{l}_{l_1,l_2 } & = & \frac{(-1)^l}{\sqrt{4\pi}}\hat{l_1}\hat{l_2 } \left(\begin{array}{ccc } l_1 & l_2 & l \\ 0 & 0 & 0 \end{array } \right ) \mbox { } .\end{aligned}\ ] ] here @xmath603 , @xmath604 , @xmath605 , @xmath606 and @xmath607 .
@xmath608 is the jacobian for the change of momenta @xmath609 @xmath610 @xmath611 , see eqs .
( [ coord3],[coord4 ] ) , @xmath612 , @xmath613 and @xmath614 , @xmath615 are the normalization coefficients of the harmonic oscillator wave functions of the baryons @xmath94 and @xmath95 , @xmath616 and meson @xmath96 , @xmath617 where @xmath618 is the number of nodes in the harmonic oscillator wave function , @xmath619 a laguerre polynomial and @xmath620 a solid spherical harmonic .
the remaining coefficients are given by : @xmath621 the present results for the @xmath21 amplitudes were obtained in a consistent way using the same jacobi coordinates for the baryon wave functions and the @xmath21 matrix elements .
we observe that the coefficients of eqs .
( [ eqn : g2xf2 ] ) also depend on the parameter @xmath622 of the gaussian quark form factor , @xmath136 .
the case @xmath138 is a subcase of the previous one .
for the special case of ground - state baryons and pseudoscalar mesons , the orbital angular momenta vanish @xmath623 and therefore @xmath624 , @xmath625 , @xmath626 and @xmath627 . due to the conservation of angular momentum and parity ,
the relative orbital angular momentum between the baryon @xmath95 and the meson @xmath96 is equal to @xmath628 . as a result ,
the general expression for the @xmath21 transition amplitude simplifies considerably .
the color - spin - flavor part @xmath592 reduces to @xmath629 where @xmath630 the total angular momentum of the first two quarks and @xmath594 the flavor matrix elements of the @xmath631 transition @xcite .
the spatial contribution simplifies to @xmath632
in the following , we give the flavor coefficients @xmath633 .
these expressions are valid for both pseudoscalar and vector mesons . *
* @xmath634 couplings . *
+ for octet baryons the flavor wave function has two components , labeled by @xmath108 and @xmath589 , both of which give rise to the flavor coupling coefficient @xmath635 where we introduced the superscripts @xmath108 and @xmath589 to distinguish the two contributions . in this case , the flavor couplings are given by @xmath636 for the @xmath108 component and by @xmath637 for the @xmath589 component .
the normalization coefficient is given by @xmath638 * * @xmath639 couplings .
* + the flavor coefficients for octet baryons in combination with a singlet meson are given by @xmath640 for the @xmath108 component and by @xmath641 for the @xmath589 component .
* * @xmath642 couplings . *
+ in this case the final baryon belongs to the decuplet which only couples to the @xmath589 component of the initial octet baryon @xmath643 the flavor coefficients are given by @xmath644 * * @xmath645 couplings .
* + as in the previous case , the decuplet baryon only couples to the @xmath589 component of the octet baryon @xmath646 the flavor coefficients are given by @xmath647 * * @xmath648 couplings .
* + the flavor coefficients for decuplet baryons in combination with a octet meson @xmath649 are given by @xmath650 * * @xmath651 couplings .
* + the flavor coefficients for decuplet baryons in combination with a singlet meson are given by @xmath652
in the following , we shall compare the elementary emission ( eme ) and @xmath0 models . in the eme model
, the decay proceeds via the emission of a meson in terms of an elementary quantum . on the contrary , in the @xmath0 model the decay process
is described in terms of the creation of an additional quark - antiquark pair . in order to do a comparison between eme and @xmath0 models , we consider the simplest form for the eme transition operator @xcite , @xmath653 ~ , \label{hs}\end{aligned}\ ] ] where @xmath654 is the ratio between the meson - emission strength ( emission of a meson @xmath96 by a quark ) and the quark mass , @xmath655 is the flavor operator related to the emission of meson @xmath96 . on the other hand ,
the @xmath0 operator in matrix form is @xmath656 where @xmath657 is the @xmath0 operator @xmath658^{(0)}_0 b_4^{\dagger}(\vec{p}_4 ) \ ,
d_5^{\dagger}(\vec{p}_5 ) \mbox { } .
\label{3p0-t02 } \end{array}\ ] ] in ref .
@xcite , it is shown that equation ( [ 3p0 m ] ) can also be written as @xmath659 \psi_c(2\vec{p}-\vec{k}_c ) \end{array } \mbox { } .
\label{hs3}\ ] ] if we compare eqs .
( [ hs ] ) and ( [ hs3 ] ) , we can observe that in the @xmath0 model there are direct and recoil terms with equal weights ; on the contrary , the recoil term in the eme has a different weight factor @xmath660 . to have a perfect correspondence between eme and @xmath0 models , this factor in the recoil term has to be taken equal to @xmath661
however , this factor has been taken as a free parameter in bil @xcite ( and also in several other studies , like @xcite ) and , by fitting , it turned out to be equal to 0.04 , which means that the recoil term is practically absent in bil .
it is worthwhile noting that in the calculation of open - flavor strong decays for @xmath662 , the final @xmath663 state is characterized by a relative angular momentum @xmath664 , where @xmath587 is the angular momentum of the initial state , @xmath94 . in ref .
@xcite , it is shown that in the eme , the partial wave @xmath665 is essentially given by the direct term with a small correction from the recoil one , while @xmath666 receives its essential contribution from the recoil term .
thus , the presence of the factor @xmath660 in the eme operator can give rise to different results with respect to the @xmath0 model .
another important difference is due to the wave function factor @xmath667 in eq .
( [ hs3 ] ) , related to the composite structure of the meson @xmath96 , which gives rise to a non local character of the operator .
e. eichten , k. gottfried , t. kinoshita , j. b. kogut , k. d. lane , and t .- m .
yan , phys .
lett . * 34 * , 369 ( 1975 ) [ * 36 * , 1276 ( 1976 ) ] ; e. eichten , k. gottfried , t. kinoshita , k. d. lane and t. -m .
yan , phys .
d * 17 * , 3090 ( 1978 ) [ * 21 * , 313 ( 1980 ) ] ; * 21 * , 203 ( 1980 ) .
e. santopinto , ph.d .
thesis , universit degli studi di genova ( 1995 ) ; m. ferraris , m.m .
giannini , m. pizzo , e. santopinto , l. tiator , phys .
b364 * , 231 ( 1995 ) ; e. santopinto , f. iachello and m. m. giannini , nucl . phys .
* a623 * , 100c ( 1997 ) ; e. santopinto , f. iachello and m. m. giannini , eur .
j. a * 1 * ( 1998 ) 307 .
g. morpurgo , nuovo cimento 9 , 461 ( 1952 ) ; j. ballot and m. fabre de la ripelle , ann . of phys .
* 127 * , 62 ( 1980 ) .
g. s. bali _
et al . _ ,
d * 62 * , 054503 ( 2000 ) ; g. s. bali , phys . rep . * 343 * , 1 ( 2001 ) ; c. alexandrou , p. de forcrand , and o. jahn , nucl .
proc . suppl . *
119 * , 667 ( 2003 ) ; h. suganuma , t. t. takahashi , f. okiharu , and h. ichie , nucl .
suppl . * 141 * , 92 ( 2005 ) .
a. ramos _ et al .
_ : dynamical baryon resonances from chiral unitarity , proceedings of the workshop on the physics of excited nucleons ( nstar2004 ) , eds .
bocquet , v. kuznetsov and d. rebreyend , ( world - scientific , singapore , 2004 ) pp 228 - 238 .
r. bijker and e. santopinto , phys .
c * 80 * , 065210 ( 2009 ) ; + e. santopinto and r. bijker , phys .
c * 82 * , 062202 ( 2010 ) ; + r. bijker , j. ferretti and e. santopinto , phys .
c * 85 * , 035204 ( 2012 ) .
e. santopinto , phys .
c * 72 * , 022201 ( 2005 ) ; j. ferretti , a. vassallo and e. santopinto , phys .
c * 83 * , 065204 ( 2011 ) ; m. de sanctis , j. ferretti , e. santopinto and a. vassallo , phys .
c * 84 * , 055201 ( 2011 ) ; eur . phys . j. a * 52 * , no . 5 , 121 ( 2016 ) ; e. santopinto and j. ferretti , phys .
c * 92 * , 025202 ( 2015 ) . | we provide results for the open - flavor strong decays of strange and non - strange baryons into a baryon - vector / pseudoscalar meson pair .
the decay amplitudes are computed in the @xmath0 pair - creation model , where @xmath1 pair - creation suppression is included for the first time in the baryon sector , in combination with the u(7 ) and hypercentral models .
the effects of this @xmath1 suppression mechanism can not be re - absorbed in a redefinition of the model parameters or in a different choice of the @xmath0 model vertex factor .
our results for the decay amplitudes are compared with the existing experimental data and previous @xmath0 and elementary meson emission model calculations . in this respect , we show that distinct quark models differ in the number of missing resonances they predict and also in the quantum numbers of states
. therefore , future experimental results will be important in order to disentangle different models of baryon structure .
finally , in the appendices , we provide some details of our calculations , including the derivation of all relevant flavor couplings with strangeness - suppression
. this derivation may be helpful to calculate the open - flavor decay amplitudes starting from other models of baryons . |
this work is concerned with /s that are written using the meta - language of ellipses : terms such as , for example , the ordered @xmath1-tuple maker : @xmath2 ) .
as the use of ellipses indicates , the syntax for such /s is described _ for any given @xmath1 _ , in the meta - language of the / , i.e. , in the language in which we describe the syntax of /s .
the index @xmath1 is thus a meta - variable .
it is only after we have picked a natural number for @xmath1 , that we can write down an actual / , and it will be `` hard - coded '' for that specific @xmath1 .
for example , the ordered @xmath3-tuple maker is defined as @xmath4 , can be written without ellipses , and is `` hard - coded '' for @xmath5 .
but what if we want @xmath1 , which determines the syntactic structure of the / , to be an argument in the language of the / : how do we go from a / whose syntax is indexed or parameterized by a meta - variable over the natural numbers in the meta - language of the / to a corresponding / parameterized by a church numeral ? in this work ,
we present a systematic approach for encoding terms whose syntax is parameterized by a meta - variable and written using ellipses , to /s that take a church numeral @xmath6 as an argument , and return the corresponding / for that given @xmath1 .
we call such /s _ / _ , following the work of weirich and casinghino on
_ / datatype - generic programming _ @xcite .
when we speak of an / / @xmath7 , we require two things : 1 .
we have in mind an @xmath1-ary term @xmath8 in the meta - language of the / , that is parameterized by a meta - variable @xmath9 . for any specific value of @xmath1 , @xmath8 is a / : @xmath10 , etc .
, are all /s .
2 . for all @xmath11 , @xmath12 . in combinatory logic , bases provide a standard approach to constructing inductively larger combinators from smaller combinators .
we follow this approach by augmenting the standard of @xmath13 basis introduced by / @xcite , curry @xcite , turner @xcite , and many others , with /s @xmath14 of the respective @xmath15 combinators .
we then encode @xmath14 in terms of @xmath13 ( section [ sec : agext ] ) .
@xmath14 can then be used to encode straightforwardly those parts of the term that use ellipses using an / of the / algorithm for the @xmath16 basis . in principle , we could have stopped at this point , since @xmath17 would already be sufficient to encode any / term .
we chose , however , to use @xmath14 to define @xmath18 , which are the / generalizations of @xmath19 , because turner s _ / algorithm _ for the basis @xmath13 extends naturally to the basis @xmath20 .
this extended algorithm ( section [ sec : extending ] ) maintains the simplicity of turner s original algorithm , and generates compact encodings for / /s .
the second part of this work ( section [ sec : induct ] ) demonstrates how the new basis can be used to encode interesting / /s , such as multiple fixed - point combinators .
for background material on the / , we refer the reader to church s original book on the / , _ the calculi of lambda conversion _
@xcite , curry s two volumes _ combinatory logic i , ii _ @xcite , and barendregt s encyclopedic textbook , _ the lambda calculus : its syntax and semantics _ @xcite . here
we briefly list the /s and notation used throughout this work .
[ cols= " < , < , < " , ] to give some intuition as to the rle the different combinators of the basis play in the algorithm , let us analyze just one single case : where @xmath21 : @xmath22 hence we have the rule that @xmath23 the correctness of this algorithm is shown by induction on the _ length _ of the term , rather than by structural induction , because , for example , while @xmath24 , clearly @xmath25 is not a sub - expression of @xmath26 , and the same holds for other cases in the proof .
* example : * we demonstrate the / algorithm by applying it to @xmath27 : @xmath28
in turner s original bracket - abstraction algorithm , the rule for @xmath29 was a base case : @xmath30 the @xmath1-ary and / of the rule for @xmath29 abstracts @xmath1 variables @xmath31 , and is also a base case : @xmath32 note that @xmath34 .
accordingly , @xmath33 is used in the original / algorithm to abstract a variable @xmath35 over an expression @xmath36 , where @xmath37 : @xmath38 the @xmath1-ary and / of the rule for @xmath33 allows for abstracting @xmath1 variables @xmath31 over an expression @xmath36 , where @xmath39 : @xmath40 note that @xmath42 where @xmath43 .
accordingly , @xmath41 is used in the original / algorithm to abstract a variable @xmath35 over an application @xmath44 , where @xmath43 : @xmath45 the @xmath1-ary and / of the rule for @xmath41 allows for abstracting @xmath1 variables @xmath31 over an application @xmath46 , where @xmath39 and @xmath47 : @xmath48 note that @xmath50 where @xmath51 .
accordingly , @xmath49 is used in the original / algorithm to abstract a variable @xmath35 over an application @xmath52 , where @xmath51 : @xmath53 the @xmath1-ary and / of the rule for @xmath49 allows for abstracting @xmath1 variables @xmath31 over an application @xmath54 , where @xmath55 : @xmath56 note that @xmath58 where @xmath59 .
accordingly , @xmath57 is used in the original / algorithm to abstract a variable @xmath35 over an application @xmath60 , where @xmath59 : @xmath61 the @xmath1-ary and / of the rule for @xmath57 allows for abstracting @xmath1 variables @xmath31 over an application @xmath62 , where @xmath63 : @xmath64 in turner s / algorithm , each of the combinators @xmath65 is used to encode an abstraction of a variable over an expression : the @xmath33 combinator is used when the variable does not occur freely in the expression .
the @xmath66 combinators are used when the variable abstracts over an application of two expressions , and correspond to the situations where the given variable occurs freely in one or in both expressions .
we extended turner s / algorithm by introducing four additional rules for @xmath67 , corresponding to the abstraction of a _ sequence of variables _ of an expression .
the extended algorithm shares the simplicity of turner s original algorithm , and generates compact encodings for / /s . in those situations
where @xmath68 , we can use the @xmath33-introduction rule to obtain from @xmath36 a @xmath69-equal expression @xmath70 for which @xmath71 . for any @xmath1-ary / @xmath72 that is written with ellipses ,
a corresponding / / @xmath73 can be defined , such that for any natural number @xmath1 , we have @xmath74
. _ sketch of proof _ : by induction on the length of @xmath72 , a corresponding rule can be applied in the extended algorithm , so that the rewritten expression is / and satisfies the above relation to @xmath72 .
* example : * church @xcite introduces the / @xmath75 , which is encoded via turner s algorithm as @xmath76 .
how would the @xmath1-ary and / extensions be encoded ? _ the @xmath1-ary extension _
: @xmath77
_ the / extension _ : @xmath78 so as we can see , the extended basis @xmath20 provides a natural extension of the original basis for encoding / /s .
the _ selector _ combinators return one of their arguments . for @xmath79
such that @xmath80 , the selector that returns the @xmath81-th of its @xmath82 arguments is defined as follows : @xmath83 an / version of the selector , which we write as @xmath84 , would take /s @xmath85 and return @xmath86 .
we generate @xmath86 in two states : first , we generate a selector in which only the first argument is returned : @xmath87 we then tag on @xmath81 additional abstractions .
suppose we have @xmath88 that are defined as follows : @xmath89 we define the / @xmath90 to map @xmath91 for all @xmath1 .
the relationship between @xmath36 and @xmath92 is given by @xmath93 , and so : @xmath94 we use @xmath90 to generate @xmath95 by applying the @xmath96-th composition of @xmath90 to the identity combinator @xmath29 . from this
we obtain @xmath86 by @xmath97 applications of @xmath33 .
we can now define @xmath84 as follows : @xmath98 this definition satisfies the requirement that @xmath99 .
the _ projection _ combinators take an @xmath1-tuple and return the respective projection : @xmath100 the standard way of defining projections is to take an @xmath1-tuple and apply it to the corresponding selector : @xmath101 the definition of the / extension @xmath102 can be written in terms of @xmath84 : @xmath103 this definition satisfies the requirement that @xmath104 . in his textbook _ the lambda calculus : its syntax and semantics _
@xcite , barendregt introduces one of the standard constructions for @xmath1-tuples : @xmath105 the ordered @xmath1-tuple maker @xmath106 takes @xmath1 /s and returns their ordered tuple . although most texts on the / use it implicitly by using ordered tuples as generalizations to the syntax of the / , it is easily definable : @xmath107 we wish to define the / of the @xmath1-tuple maker @xmath108 , such that : @xmath109 we relate @xmath106 with @xmath110 as follows : @xmath111 using this relation , we define the / @xmath90 to map @xmath112 for all @xmath1 : @xmath113 notice that @xmath114 , so we can obtain @xmath106 by applying the @xmath1-th composition of @xmath90 to @xmath115 . we define @xmath108 as follows : @xmath116 this definition satisfies the requirement that @xmath117 , so for example , @xmath118 .
the task of defining the @xmath106 combinator is given as an exercise in the author s course notes on the / @xcite , where the combinator is referred to as _ malloc _ , in a tongue - in - cheek reference to the c library function for allocating blocks of memory .
a useful property of our representation of ordered @xmath1-tuples , is that it gives us _ left - associated _ applications immediately : @xmath119 this behavior can be used to apply some expression to its arguments , where these arguments are passed in an @xmath1-tuple , in much the same way as the ` apply ` procedure in lisp @xcite , which takes a procedure and a list of arguments , and applies the procedure to these arguments .
one notable difference though , is that the / does not have a notion of / procedures , and so the procedure we wish to apply must `` know '' now many arguments to expect .
we can thus define : @xmath120 because functions in the / are curried , and therefore applications associate to the left , @xmath121 combinator proides for _ left - associated _ applications . for _ right - associated _ applications
, we would like to have an / version of the following @xmath1-ary / : @xmath122 we begin by writing @xmath123 in terms of @xmath124 : @xmath125 the / @xmath90 such that @xmath126 is given by : @xmath127 notice that @xmath128 , so we can obtain @xmath124 by applying the @xmath1-th composition of @xmath90 to @xmath115 .
we define @xmath129 as follows : @xmath130 this definition satisfies the requirement that @xmath131 .
applying @xmath110 to @xmath1 arguments results in a / that takes an argument and returns an @xmath82-tuple , in which the given argument is the @xmath82-st projection .
we use this fact to extend an @xmath1-tuple by an additional @xmath82-st element : @xmath132 we can use it as follows : @xmath133 similarly , we can define the / @xmath134 , for creating an @xmath135-tuple given an @xmath1-tuple and a @xmath81-tuple : @xmath136 for example : @xmath137 when working with indexed expressions , it is convenient to have the _ iota_-function ( written as the greek letter @xmath138 , and pronounced `` yota '' ) , which maps the number @xmath139 to the vector @xmath140 . iota was introduced by kenneth iverson first in the apl notation @xcite , and then in the apl programming language @xcite .
we implement the @xmath138 combinator to take a church numeral @xmath6 and return the ordered @xmath1-tuple @xmath141 . given the standard definition of ordered @xmath1-tuples , it is natural to define @xmath142 .
we know that @xmath143 so the / @xmath90 such that @xmath144 can be characterized as follows : @xmath145 we define @xmath90 as follows : @xmath146 the / @xmath147 such that @xmath148 is defined as follows : @xmath149 we now define @xmath138 as follows : @xmath150 this definition satisfies the requirement that @xmath151 .
it is often useful to be able to reverse the arguments to a function or an @xmath1-tuple .
we can define an @xmath1-ary reversal combinator @xmath152 as follows : @xmath153 @xmath152 can be used in two ways : 1 .
we can use it to reverse an ordered @xmath1 tuple : @xmath154 2 . we can use it to take @xmath1 arguments are return their @xmath1-tuple , in reverse order : @xmath155 we would like to define @xmath156 , the / of @xmath152 , such that @xmath157 .
we start by writing @xmath158 in terms of @xmath152 : @xmath159 the / @xmath90 that takes a church numeral @xmath6 , and maps @xmath160 , is given by : @xmath161 the / @xmath147 such that @xmath162 is given by : @xmath163 notice that @xmath164 , so we can obtain @xmath152 by applying the @xmath1-th composition of @xmath147 to @xmath115 .
we define @xmath156 as follows : @xmath165 note that @xmath157 . below are examples of two slightly different ways of using @xmath156 : @xmath166 we would like to define the combinator @xmath167 , such that : @xmath168 let : @xmath169 we define @xmath170 in terms of @xmath171 : @xmath172 using this relation , we define the / @xmath147 to map @xmath173 for all @xmath1 : @xmath174 notice that @xmath175 , so we can obtain @xmath171 by applying the @xmath1-th composition of @xmath147 to @xmath115 . we define @xmath176 as follows : @xmath177 we now define @xmath167 as follows : @xmath178 this definition satisfies the requirement that @xmath179 .
by now we have the tools nee- ded to construct / , multiple fixed - point combinators in the /.
fixed - point combinators are used to solve fixed - point equations , resulting in a single solution that is the _
least _ in a lattice - theoretic sense .
when moving to @xmath1 multiple fixed - point equations , multiple fixed - point combinators are needed to solve the system , giving a set of @xmath1 solutions , that once again , are the least in the above - mentioned lattice - theoretic sense .
a set of @xmath1 multiple fixed - point combinators are /s @xmath180 , such that for any @xmath1 /s @xmath181 , and @xmath182 we have : @xmath183 brevity is one motivation for the construction of an / fixed - point combinator . using ordinary multiple fixed - point combinators , @xmath1 combinators are needed for _ any _ choice of @xmath1 , which means that if we wish to solve several such systems of equations , we need a great many number of multiple fixed - point combinators .
in contrast , an / fixed - point combinator can be used to find _ any _ multiple fixed - point in a system of _ any _ size : it takes as arguments two church numerals @xmath184 , which specify the size of the system , and the specific multiple fixed - point , and returns the specific multiple fixed - point combinator of interest .
other reasons for using an / fixed - point combinator have to do with the size of the multiple fixed - point combinators and their correctness : the size of the @xmath1-ary extensions of curry s and turing s historical fixed - point combinators is _ quadratic _ to the number of equations , or @xmath185 .
specifying such large terms , be in on paper , in latex , or in a computerized reduction system is unwieldy and prone to errors .
an / fixed - point combinator is surprisingly compact , because the size of the system is specified as an argument .
recall curry s single fixed - point combinator : @xmath186 generalizing curry s single fixed - point combinator to @xmath1 multiple fixed - point equations yields a sequence @xmath187 of @xmath1 multiple fixed - point combinators , where @xmath188 is defined as follows : @xmath189 given the system of fixed - point equations @xmath190 , the @xmath81-th multiple fixed - point is given by @xmath191 .
our inductive definition ( on the syntax of / ) is sufficiently precise and well - defined that we can construct , for any given @xmath11 , a set of multiple fixed - point combinators .
but if @xmath1 is a variable , rather than a constant , then this will not do .
let @xmath192 .
starting with the inner common sub - expression @xmath193 , we note that : @xmath194 the / fixed - point combinator @xmath195 takes @xmath184 , and returns @xmath188 , which is the fixed - point combinator that takes @xmath1 generating functions , and returns the @xmath81-th of @xmath1 multiple fixed - points : @xmath196 abstracting the variables @xmath85 over @xmath197 respectively , we define the / extension of curry s multiple fixed - point combinator : @xmath198 this definition satisfies the requirement that @xmath199 .
recall turing s single fixed - point combinator : @xmath200 generalizing turing s single fixed - point combinator to @xmath1 multiple fixed - point equations yields a sequence @xmath201 of @xmath1 multiple fixed - point combinators , where @xmath202 is defined as follows : @xmath203 our construction follows similar lines as with the @xmath1-ary generalization of @xmath204 . for a given @xmath1 , the ordered @xmath1-tuples @xmath205 are defined as follows : @xmath206 respectively . as before , we begin by encoding a common sub - expression @xmath207 , as follows : @xmath208 the / of turing s multiple fixed - point combinator is given by : @xmath209 we define @xmath210 by abstracting @xmath197 over the above , to get : @xmath211 this definition satisfies the requirement that @xmath212 . in sections [ sssec :
curryfpc ] and [ sssec : turingfpc ] we introduced @xmath1-ary generalizations of curry s and turing s / for solving systems of multiple fixed - point equations .
the goal of this section is to show that these generalizations are , in a precise sense , natural , and obey a well - known relation that holds between the two original , single /s . in his textbook _
the lambda calculus : its syntax and semantics _ @xcite , barendregt mentions , in the proof of proposition 6.5.5 , a result due to / , that relates curry s and turing s /s : let @xmath213 .
we have : @xmath214 to understand whence this / @xmath215 comes , consider the definition of a / : a term @xmath216 , such that for all @xmath35 , @xmath217 is a fixed point of @xmath35 , and so we have : @xmath218 abstracting over @xmath35 , we get a recursive definition for @xmath216 , that can be rewritten as a fixed - point equation : @xmath219 we can solve this fixed - point equation using _ any _ /.
if @xmath216 is a / , then @xmath220 is also a /.
after we prove these to be distinct in the @xmath221 sense , we can define an infinite chain of distinct /s .
furthermore , @xmath215 relates @xmath204 and @xmath222 in an interesting way : @xmath223 , which is a stronger relation than @xmath224 . for the purpose of this work ,
we consider @xmath1-ary generalizations of @xmath204 and @xmath222 to be natural if they satisfy a corresponding @xmath1-ary generalization of the above relation .
we now define @xmath1-ary generalizations of the above term @xmath215 .
if @xmath225 are a set of @xmath1 multiple /s , then for any @xmath31 and @xmath226 , it satisfies : @xmath227 where @xmath228 .
the @xmath1-ary generalizations of @xmath229 are given by @xmath230 , respectively , for all @xmath226 .
[ prop : mnk ] for any @xmath231 and each @xmath232 , we have @xmath233 .
@xmath234 we would like to define the combinator @xmath235 , which is the / of the @xmath236 , such that : @xmath237 we start with @xmath236 : # 1#2 ( _ @ @xmath238 we generate such a repeated application by repeatedly applying the function @xmath90 , defined so that @xmath239
. assuming the variable @xmath1 , which stands for the church numeral @xmath6 in the previous expression , and which occurs free in @xmath90 , we define @xmath90 as follows : @xmath240 we can now use @xmath90 to define @xmath236 : @xmath241 we now define @xmath235 by abstracting @xmath197 over the parameterized expression , to get : @xmath242 this definition satisfies the requirement that @xmath243 .
combined with proposition [ prop : mnk ] , it follows that for @xmath244 and for each @xmath226 , we have : @xmath245 .
the stronger @xmath246 property does not hold when working with encodings , which are by definition , @xmath69-equivalent . finally , just as @xmath215 was used to construct a chain of infinitely - many different /s , so can @xmath235 be used to construct a chain of infinitely - many / /s : if @xmath247 are @xmath1 multiple /s , then so are # 1 @ @xmath248 and so are @xmath249 etc
. we defined @xmath1-ary ( @xmath230 ) and / ( @xmath250 ) generalizations of curry s and turing s /s , and showed that these generalizations maintain the @xmath1-ary and /s of the relationship originally discovered by /.
the significance of / /s is that they are _ single terms _ that parameterize over the number of fixed - point equations and the index of a fixed point , so they can be used to find any fixed point of any number of fixed - point equations : they can be used interchangeably to define mutually - recursive procedures , mutually - recursive data structures , etc .
for example , if @xmath251 are the _ even _ and _ odd _ generating functions given by : @xmath252 then we can use curry s / / to define the /s that compute the _ even _ and _ odd _ functions on church numerals as follows : @xmath253 alternatively , we can use turing s / / to do the same : @xmath254 it might seem intuitive that in order to generate @xmath1 multiple fixed points , we would need @xmath1 generating expressions , and this intuition is responsible for the @xmath185 size of the @xmath1-ary extensions of curry s and turing s /s . a more compact approach , however , is to pass along a single aggregation of the @xmath1 fixed points , which can be done using a single generator function that is applied to itself .
this approach was taken by kiselyov @xcite in his construction of a variadic , multiple / in scheme : .... ( define y * ( lambda s ( ( lambda ( u ) ( u u ) ) ( lambda ( p ) ( map ( lambda ( si ) ( lambda x ( apply ( apply si ( p p ) ) x ) ) ) s ) ) ) ) ) .... a corresponding / version can be encoded in the / in two ways .
first , to emphasize the brevity of this construction , we can write : @xmath255 note that since the @xmath121 combinator reverses its two arguments , we can avoid it altogether by reversing its two arguments _ in situ _ ,
essentially _ inlining _ the @xmath121 combinator .
then for any @xmath11 , let @xmath256 be some /s , and let @xmath247 be a set of @xmath1 multiple /s , @xmath257 satisfies : @xmath258 but to be consistent with how we defined and used other / terms , we should rather define a curried variant @xmath259 : @xmath260 this variant takes a church numeral , followed by @xmath1 /s , and returns the @xmath1-tuple of their multiple fixed points : @xmath261 so it seems that the shortest known multiple fixed - point combinator in scheme translates to a very short multiple fixed - point combinator in the / , perhaps the shortest known as well . in a previous work @xcite
, we have shown that for any @xmath1 /s @xmath262 , which need not even be combinators , it is possible to define a single term @xmath263 that generates @xmath264 .
such a term is known as a _
one - point basis _ ( * ? ? ?
* section 8.1 ) .
it is straightforward to construct a _ dispatcher _
/ @xmath265 , such that @xmath266 , for all @xmath226 .
let @xmath267 , where @xmath268 .
then , for any @xmath226 , we have : @xmath269 notice that a different dispatcher is needed for each @xmath1 , and for each @xmath262 . using our / basis
, we can abstract a church numeral over our construction , and obtain an / one - point basis _ maker_. we define @xmath215 so as to use an / selector to dispatch over @xmath1 expressions : @xmath270 we use @xmath215 to define the / basis maker @xmath271 : @xmath272 we may now define @xmath271 as follows : @xmath273 we can use @xmath271 as follows . for any @xmath274 and @xmath275 , we can define @xmath263 as follows : @xmath276 we now have : @xmath277 notice that we have made no assumptions about @xmath278 , and in particular , have not required that they be combinators .
our one - point basis maker , @xmath271 , provides an abstract mechanism for packaging /s , in a way that they can later be `` unpacked '' .
we used our extended basis and / algorithm to encode useful / /s of increasing complexity .
we took the approach that working with sequences of expressions in an intuitive , modular and systematic way should resemble `` list processing '' known from / and other functional programming languages . in the spirit of list processing ,
the first part of this section introduces / /s for picking elements of sequences , constructing ordered @xmath1-tuples , applying /s to the elements of a tuple , extending and reversing tuples , and constructing new ordered @xmath1-tuples by _ mapping _ over existing tuples .
all these /s correspond to the basic machinery for list processing , e.g. , in /.
once these were defined , we were ready to look at more complex / /s .
our detailed examples include / /s , and an / generator for one - point bases .
we encoded /s of two historical /s by curry and turing .
these /s maintain a relationship discovered by / , so it is natural to wonder whether this relationship is maintained in the /s of these /s , and we have shown this to be the case up to @xmath69-equivalence .
we then encoded an / generator for one - point bases , so that any number of /s can be `` compacted '' into a single expression from which they can be generated .
we tested all the / definitions in this work using a normal - order reducer for the / , and have verified that they behave as expected on an array of examples .
the expressive power of the / has fostered the advent of functional languages .
for example , the algorithmic language scheme @xcite was developed as an interpreter for the / , and offered programmatic support for playing with @xmath0-definability , from church numerals to a call - by - value version of curry s fixed - point combinator @xcite . since scheme provides linguistic support for variadic functions
, it has become a sport to program call - by - value fixed - point operators for variadic functions .
queinnec presented the scheme procedure ` nfixn2 ` , that is a variadic , applicative - order multiple fixed - point combinator ( * ? ? ?
* pages 457458 ) .
the author presented one that directly extends curry s fixed - point combinator @xcite and was a motivation for section [ ssec : mfpcs ] .
the original aim of the combinatory - logic program , as pursued by / @xcite , was the elimination of bound variables @xcite .
to this end , / introduced five constants , each with a conversion rule that described its behavior .
these constants are known today as @xmath279 .
while / did not leave an explicit _ abstraction algorithm _ for translating terms with bound variables to equivalent terms without bound variables @xcite , cardon and hindley claim it extremely likely that he knew of such an algorithm @xcite .
as far as we have been able to verify , the first to have considered the question of how to encode inductive and / /s was curry , first in an extended combinatory logic framework @xcite , where curry first mentions such variables , and refers to them as _ apparent variables _ , and later , for combinatory logic ( * ? ? ?
* section 5e ) . we have not found this terminology used elsewhere , and since the term _ / _ is much more self - explanatory , we have chosen to stick with it .
abdali , in his article _ an abstraction algorithm for combinatory logic _
@xcite , presented a much simpler algorithm for encoding inductive and / /s .
abdali introduces the terms : * @xmath280 , which is an / of @xmath33 , and identical to the @xmath281 combinator used throughout this article . *
@xmath282 , which is an / selector , and is identical to the @xmath84 combinator introduced in section [ ssec : sel ] .
* @xmath283 , which is a _ double _ / of curry s @xmath284 combinator @xcite , generalized for two independent indices .
these combinators can augment any basis , and provide for a straightforward encoding of / /s .
abdali does not explain how he came up with the double generalization of curry s @xmath216 combinator , or how he encoded the definitions for @xmath285 in terms of the basis he chose to use .
/ expressions encoded using @xmath285 , are not as concise as they could be , because the @xmath283 combinator introduces variables even in when they are not needed in parts of an application , and in such cases , a subsequent projection is needed to remove them .
barendregt @xcite seems to have considered this question at least for some special cases , as in exercises 8.5.13 and 8.5.20 , the later of which he attributes to david a. turner .- ary expressions and their encoding in the / , it is plausible that he had really intended to refer to another article by turner , also published in 1979 : _ another algorithm for bracket abstraction _
@xcite . ]
/ s original @xmath279 basis , coupled with turner s / algorithm for that basis , offers several advantages in terms of brevity of the resulting term , simplicity , intuitiveness and ease of application of the algorithm . in the original / algorithm for @xmath279 ,
the length of the encoded / is less than or equal to the length of the original / , because each application is replaced by a combinator , and abstractions are either represented by a single combinator , or are removed altogether through @xmath286-reduction .
the additional / combinators with which we extended the @xmath279 basis maintain this conciseness , because a sequence of left - associated applications to a sequence of variables is replaced by a single / combinator , and a sequence of curried , nested /s is either removed via repeated @xmath286-expansions , or is replaced with by a single / combinator .
the extension of the basis and the corresponding / algorithm to handle / /s is straightforward and intuitive .
the ellipsis ( ` @xmath287 ' ) and its typographical predecessor ` & c ' ( an abbreviation for the latin phrase _ et cetera _ ,
meaning `` and the rest '' ) have been used as meta - mathematical notation , to abbreviate mathematical objects ( numbers , expressions , formulae , structures , etc . ) for hundreds of years , going back to the 17th century and possibly earlier .
such abbreviations permeate the writings of isaac newton , john wallis , leonhard euler , carl friedrich gauss , and up to the present . despite its ubiquity , and perhaps as a paradoxical tribute to this ubiquity , the ellipsis does not appear as an entry in standard texts on the history of mathematical notation , even though the authors of these texts make extensive use of ellipses in their books @xcite . neither is the ellipsis discussed in the kleene s classical text on metamathematics @xcite , nor does it even appear as an entry in the list of symbols and notation at the end of the book , even though kleene makes extensive use of the ellipses both in the main text as well as in the list of symbols and notation .
discussions about the ellipsis and its meanings seem to concentrate in computer literature : roland backhouse refers to the ellipsis as the _ dotdotdot notation _ in one of the more mathematical parts of his book _ program construction : calculating implementations from specifications _
* section 11.1 ) , and suggests that they have many disadvantages , the most important being that `` it puts a major burden on the reader , requiring them to interpolate from a few example values to the general term in a bag of values . ''
some of the examples of ellipses he cites can be rewritten using summations , products , and the like .
others , however involve the meta - language , e.g. , functions that take @xmath1 arguments , where @xmath1 is a meta - variable .
such examples of ellipses can not be removed as easily .
the ellipsis also appears in some programming languages . in some languages ( c , c++ , and java )
it is used to define variadic procedures . in other languages
( ruby , rust , and gnu extensions to c and c++ ) it is used to define a range . in scheme ,
the ellipsis is part of the syntax for writing macros , which can be thought of as a meta - language for scheme .
a formal treatment of ellipses in the macro language for scheme was done by eugene kohlbecker in his phd thesis @xcite .
/ terms are somewhat reminiscent of _ variadic _ procedures in programming languages : the term _ variadic _ , introduced by strachey @xcite , refers to the _ arity _ of a procedure , i.e. , the number of arguments to which it can be applied . a _
dyadic _ procedure can be applied to two arguments .
a _ triadic _ procedure can be applied to three arguments .
variadic _ procedure can be applied to any number of arguments .
programming languages that provide a syntactic facility for defining variadic procedures include c++ and /.
the / has no such syntactic facility , and so it is somewhat of a misnomer to speak of _ variadic /s _ , since the number of arguments is an _ explicit _
parameter in our definitions , whereas in the application of a variadic procedure to some arguments , the _ number _ of arguments is _ implicit _ in an implementation .
nevertheless , within the classical , untyped / , / /s provide an expressivity that comes very close to having variadic /s .
variadic procedures are not just about the procedure _
interface_. when used in combination with ` map ` and ` apply ` , they can provide a kind of generality that is typically deferred to the meta - language or macro system @xcite .
/ / achieves similar generality in the classical / , with some notable differences : variadic procedures are applied to arbitrarily - many arguments , and their parameter is bound to the list of the values of these arguments . by contrast , / expressions take the number of arguments , and return that many curried @xmath0-abstractions . in this work ,
we used ordered @xmath1-tuples , rather than linked lists , as is common in most functional programming languages , in what is perhaps reminiscent of array programming languages . as a result of the choice
to use ordered @xmath1-tuples , the _ apply _ operation became very simple .
it would be straightforward to choose to use linked lists instead , at the cost of having to define _ apply _ as a _
left fold _ operation . in this work
we show how to define , in the language of the / , expressions that contain meta - linguistic ellipses , the size of which is indexed by a meta - variable .
for such an indexed / @xmath8 , our goal was to find a term @xmath7 that takes @xmath1 as an explicit parameter , and assuming it to be a church numeral denoting the size of the indexed expression , evaluates to @xmath8 : @xmath288 .
we call @xmath7 an / of @xmath289 .
of course , our choice of using church numerals in this paper is based on their ubiquity .
in fact , any numeral system can be used , and we have also constructed an / basis around scott numerals @xcite . our approach has been to extend the basis @xmath13 with the /s of @xmath65 combinators and to extend turner s / algorithm to handle abstractions of sequences of variables over an expression .
we then used this extended basis and this extended / algorithm to encode / /s .
our goal has not been to remove all abstractions in / terms , but only those abstractions that are over sequences of variables . of course , it is possible to remove all remaining abstractions , but our goal here has been to define indexed expressions in the / , without resorting to meta - linguistic ellipses , for which the removal of all abstractions is unnecessary . in the first part of this work we presented a natural , / to / s @xmath13 basis for the set of combinators in the / , and extended turner s / algorithm to make use of the additional / combinators in the extended basis .
the extended algorithm retains the conciseness and simplicity of turner s original algorithm .
the second part of this work uses the / basis and the corresponding / algorithm to develop tools for / / , and incidentally demonstrates how the / basis can be used : we introduced several / /s that perform a wide variety of computations on ordered @xmath1-tuples .
these computations were inspired by , and resemble to some extent , the facilities for _ list manipulation _ that are native to the / programming language @xcite : terms that compute mappings , reversal , / fixed - point combinators , / one - point bases , etc .
implementing in the / a functional subset of the list processing capabilities of / is a popular exercise . in his textbook on the / ,
barendregt states that there are two ways to define ordered @xmath1-tuples : inductively , using nested ordered pairs , and another way , which barendregt characterizes as being `` more direct '' , as @xmath290 @xcite .
section [ ssec : ntuplemaker ] shows how to make this more direct definition inductive . in a previous work @xcite
, we derived an applicative - order , variadic fixed - point combinator in scheme . in that work
, we relied on scheme s support for writing variadic procedures , and consequently , on the primitive procedure ` apply ` , to apply procedures to lists of their arguments . in the present work
, we had control over the representation of sequences , so we could encode an / version of ` apply ` , as well as / /s , all within the /.
the author is grateful to his anonymous reviewers and to his editor , neil d. jones .
thanks are also due to john franco and albert meyer for comments and questions about a previous work , and to olivier danvy for his encouragement and suggestions .
felice cardone and j. roger hindley .
lambda - calculus and combinators in the 20th century . in dov
m. gabbay and john woods , editors , _ logic from russell to church _ ,
volume 5 of _ handbook of the history of logic _ , pages 723817 .
north - holland , 2009 .
moses schnfinkel . .
, 92:305316 , 1924 . translated by stefan bauer - mengelberg as `` on the building blocks of mathematical logic '' , in jean van heijenoort , 1967 . _ a source book in mathematical logic _ , 18791931 .
harvard university press .
pages 35566 .
christopher strachey .
fundamental concepts in programming languages .
international summer school in computer programming , copenhagen , denmark , august 1967 . reprinted in higher - order and symbolic computation 13(1/2):1149 , 2000 .
gerald j. sussman and guy l. steele jr .
scheme : an interpreter for extended lambda calculus .
ai memo 349 , artificial intelligence laboratory , massachusetts institute of technology , cambridge , massachusetts , december 1975 . reprinted in higher - order and symbolic computation 11(4):405439 , 1998 .
christopher p. wadsworth .
some unusual @xmath0-calculus numeral systems . in jonathan
p. seldin and j. roger hindley , editors , _ to h. b. curry : essays on combinatory logic , lambda calculus and formalism _ , pages 215230 . academic press , london , 1980 .
stephanie weirich and chris casinghino .
arity - generic datatype - generic programming . in _ proceedings of the 4th acm sigplan workshop on programming languages meets program verification _ , plpv 10 , pages 1526 , new york , ny , usa , 2010 . | ellipses are a meta - linguistic notation for denoting terms the size of which are specified by a meta - variable that ranges over the natural numbers . in this work ,
we present a systematic approach for encoding such meta - expressions in the / , without ellipses : terms that are parameterized by meta - variables are replaced with corresponding @xmath0-abstractions over actual variables .
we call such /s _
/_. concrete terms , for particular choices of the parameterizing variable are obtained by applying an / / to the corresponding numeral , obviating the need to use ellipses .
for example , to find the multiple fixed points of @xmath1 equations , @xmath1 different /s are needed , every one of which is indexed by two meta - variables , and defined using three levels of ellipses . a single / / that takes two church numerals ,
one for the number of fixed - point equations , and one for their arity , replaces all these multiple fixed - point combinators .
we show how to define /s of two historical /s , the first by curry , and the second by turing , for defining multiple fixed points .
these historical /s are related by a construction due to / : we show that likewise , their /s are related by an / of / s construction .
we further demonstrate this approach to / / with additional /s that create , project , extend , reverse , and map over ordered @xmath1-tuples , as well as an / generator for one - point bases . |
we represent word - meaning and meaning - meaning relations uncovered by translation dictionaries between each language in the unbiased sample and major modern european languages by constructing a network structure . two meanings ( represented by a set of english words )
are linked if they are translated from one to another and then back , and the link is weighted by the number of paths of the translation , or the number of words that represent both meanings ( see methods for detail ) .
figure [ fig : schematic ] illustrates the construction in the case of two languages , lakhota ( primarily spoken in north and south dakota ) and coast tsimshian ( mostly spoken in northwestern british columbia and southeastern alaska ) .
translation of sun in lakhota results _ w _ and _ @xmath0paw_. while the later picks up no other meaning , _ w _ is a polysemy that possesses additional meanings of moon and month , hence they are linked to sun .
such polysemy is also observed in coast tsimshian where _ gyemk _ , translated from sun , covers additional meanings including , thus additionally linking to , heat .
each language has its own way of partitioning meanings by words , captured in a semantic network of the language .
it is conceivable , however , that a group of languages bear structural resemblance perhaps because the speakers share historical or environmental features .
a link between sun and moon , for example , reoccurs in both languages , but does not appear in many other languages .
sun is instead linked to divinity and time in japanese , and to thirst and day / daytime in !
the question then is the degree to which the observed polysemy patterns are general or sensitive to the environment inhabited by the speech community , phylogenetic history of the languages , and intrinsic linguistic factors such as literary tradition .
we test such question by grouping the individual networks in a number of ways according to properties of their corresponding languages .
we first analyze the networks of the entire languages , and then of sub - groups . in fig .
[ fig : connectance_graph ] , we present the network of the entire languages exhibiting the broad topological structure of polysemies observed in our data .
it reveals three almost - disconnected clusters , groups of concepts that are indeed more prone to polysemy within , that are associated with a natural semantic interpretation .
the semantically most uniform cluster , colored in blue , includes concepts related to water . a second , smaller cluster , colored in yellow , associates solid natural substances ( centered around stone / rock ) with their topographic manifestation ( mountain ) .
the third cluster , in red , is more loosely connected , bridging a terrestrial cluster and a celestial cluster , including less tangible substances such as wind , sky , and fire , and salient time intervals such as day and year . in keeping with many traditional oppositions between earth and sky / heaven , or darkness , and light , the celestial , and terrestrial components form two sub - clusters connected most strongly through cloud , which shares properties of both .
the result reveals a coherent set of relationships among concepts that possibly reflects human cognitive conceptualization of these semantic domains @xcite .
we test whether these relationships are universal rather than particular to properties of linguistic groups such as physical environment that human societies inhabit .
we first categorized languages by nonlinguistic variables such as geography , topography , climate , and the existence or nonexistence of a literary tradition ( table [ tab : groups ] in appendix ) and constructed a network for each group . a spectral algorithm
then clusters swadesh entries into a hierarchical structure or dendrogram for each language group . using standard metrics on trees @xcite , we find that the dendrograms of language groups are much closer to each other than to dendrograms of randomly permuted leaves : thus the hypothesis that languages of different subgroups share no semantic structure in common is rejected ( @xmath1 , see methods)sea / ocean and salt are , for example , more related than either is to sun in every group we tried . in addition , the distances between dendrograms of language groups are statistically indistinguishable from the distances between bootstrapped languages ( @xmath2 ) .
figure 3 shows a summary of the statistical tests of 11 different groups .
thus our data analyses provide consistent evidences that all languages share semantic structure , the way concepts are clustered in fig .
2 , with no significant influence from environmental or cultural factors .
another structural feature apparent in fig .
[ fig : connectance_graph ] is the heterogeneity of the node degrees and link weights .
the numbers of polysemies involving individual meanings are uneven , possibly toward a heavy - tailed distribution ( fig .
[ fig : word_degree_rank ] ) .
this indicates concepts not only form clusters within which they are densely connected , but also exhibit different levels of being polysemous .
for example , earth / soil has more than hundreds of polysemes while salt has only a few .
having shown that some aspects of the semantic network are universal , we next ask whether the observed heterogeneous degrees of polysemy , possibly a manifestation of varying densities of near conceptual neighbors , arise as artifacts of language family structure in our sample , or if they are inherent to the concepts themselves .
simply put , is it an intrinsic property of the concept , earth / soil , to be extensively polysemous , or is it a few languages that happened to call the same concept in so many different ways .
suppose an underlying `` universal space '' relative to which each language @xmath3 randomly draws a subset of polysemies for each concept @xmath4
. the number of polysemies @xmath5 should then be linearly proportional to both the tendency of the concept to be polysemous for being close to many other concepts , and the tendency of the language to distinguish word senses in basic vocabulary .
in our network representation , a proxy for the former is the weighted degree @xmath6 of node @xmath4 , and a proxy for the latter is the total weight of links @xmath7 in language @xmath3 .
then the number of polysemies is expected ( see methods ) : @xmath8 this simple model indeed captures the gross features of the data very well ( fig .
[ fig : productmodel_matrix ] in the appendix ) . nevertheless , the kullback - leibler divergence between the prediction @xmath9 and the empirical data @xmath10 identifies deviations beyond the sampling errors in three concepts moon , sun and ashes that display nonlinear increase in the number of polysemies ( @xmath11 ) with the tendency of the language distinguish word senses as fig .
[ fig : saturating_words ] in the appendix shows . accommodating saturation parameters ( table [ tab : fitvalue ] in the appendix ) enables the random sampling model to reproduce the empirical data in good agreement keeping the two parameters independent , hence retain the universality over language groups .
the similarity relations between word meanings through common polysemies exhibit a universal structure , manifested as intrinsic closeness between concepts , that transcends cultural or environmental factors .
polysemy arises when two or more concepts are fundamental enough to receive distinct vocabulary terms in some languages , yet similar enough to share a common term in others .
the highly variable degree of these polysemies indicates such salient concepts are not homogeneously distributed in the _ conceptual _ space , and the intrinsic parameter that describes the overall propensity of a word to participate in polysemies can then be interpreted as a measure of the local density around such concept .
our model suggests that given the overall semantic ambiguity observed in the languages , such local density determines the degree of polysemies .
universal structures in lexical semantics would greatly aid another subject of broad interest , namely reconstruction of human phylogeny using linguistic data @xcite .
much progress has been made in reconstructing the phylogenies of word forms from known cognates in various languages , thanks to the ability to measure phonetic similarity and our knowledge of the processes of sound change .
however , the relationship between semantic similarity and semantic shift is still poorly understood .
the standard view in historical linguistics is that any meaning can change to any other meaning @xcite , and that no constraint is imposed on what meanings can be compared to detect cognates @xcite .
it is , however , generally accepted among historical linguists that language change is gradual , and that words in transition from having one meaning to being extended to another meaning should be polysemous .
if this is true , then the weights on different links reflect the probabilities that words in transition over these links will be captured in `` snapshots '' by language translation at any time .
such semantic shifts can be modeled as diffusion in the conceptual space , or along a universal polysemy network where our constructed networks can serve an important input to methods of inferring cognates . .
entries from the initial swadesh list are distinguished with capital letters .
( a ) in - strengths of concepts : sum of weighted links to a node .
( b ) out - strengths of swadesh entries : sum of weighted links from a swadesh entry .
( c ) degree of the concepts : sum of unweighted links to a node ( d ) degree of swadesh entries : sum of unweighted links to a node .
a node strength in this context indicates the total number of polysemies associated with the concept in 81 languages while a node degree means the number of other concepts associated with the node regardless of the number of synonymous polysemies associated with it .
heaven , for example , has the largest number of polysemies , but most of them are with sun , so that its degree is only three .
[ fig : word_degree_rank ] ] the absence of significant cladistic correlation with the patterns of polysemy suggests a possibility to extend the constructed conceptual space by utilizing digitally archived dictionaries of the major languages of the world with some confidence that their expression of these features is not strongly biased by correlations due to language family structure .
large - corpus samples could be used to construct the semantic space in as yet unexplored domains using automated means .
high - quality bilingual dictionaries between the object language and the semantic metalanguage for cross - linguistic comparison are used to identify polysemies .
the 81 object languages were selected from a phylogenetically and geographically stratified sample of low - level language families or _ genera _ , listed in tab .
[ tab : languages ] in the appendex @xcite .
translations into the object language of each of the 22 word senses from the swadesh basic vocabulary list were first obtained ( see appendix-[subsec : meanings ] ) ; all translations ( that is , all synonyms ) were retained .
polysemies were identified by looking up the metalanguage translations ( back - translation ) of each object - language term .
the selected swadesh word senses , and the selected languages are listed in the appendix .
we use modern european languages as a semantic metalanguage , _
i.e. , _ bilingual dictionaries between such languages and the other languages in our sample . this could be problematic if these languages themselves display polysemies ; for example , english _ day _ expresses both daytime , and 24hr period .
in many cases , however , the lexicographer is aware of these issues , and annotates the translation of the object language word accordingly . in the lexical domain chosen for our study ,
standard lexicographic practice was sufficient to overcome this problem . a hierarchical spectral algorithm clusters the swadesh word senses .
each sense @xmath12 is assigned to a position in @xmath13 based on the @xmath12th components of the @xmath14 eigenvectors of the weighted adjacency matrix .
each eigenvector is weighted by the square of its eigenvalue , and clustered by a greedy agglomerative algorithm to merge the pair of clusters having the smallest squared euclidean distance between their centers of mass , through which a binary tree or _ dendrogram _ is constructed we construct a dendrogram for each subgroup of languages according to nonlinguistic variables such as geography , topography , climate , and the presence or absence of a literary tradition ( table [ tab : groups ] in appendix ) . the structural distance between the dendrograms of each pair of language subgroups
is measured by two standard tree metrics . the triplet distance @xmath15 @xcite is the fraction of the @xmath16 distinct triplets of senses that are assigned a different topology in the two trees : that is , those for which the trees disagree as to which pair of senses are more closely related to each other than they are to the third .
the robinson - foulds distance @xmath17 @xcite is the number of `` cuts '' on which the two trees disagree , where a cut is a separation of the leaves into two sets resulting from removing an edge of the tree . for each pair of subgroups ,
we perform two types of bootstrap experiments .
first , we compare the distance between their dendrograms to the distribution of distances we would see under a hypothesis that the two subgroups have no shared lexical structure .
were this null hypothesis true , the distribution of distances would be unchanged under the random permutation of the senses at the leaves of each tree ( for simplicity , the topology of the dendrograms are kept fixed . ) comparing the observed distance against the resulting distribution gives a @xmath18-value , called @xmath19 in figure [ fig : bootstrap ] .
these @xmath18-values are small enough to decisively reject the null hypothesis . indeed , for most pairs of groups the robinson - foulds distance is smaller than that observed in any of the 1000 bootstrap trials ( @xmath20 ) marked as @xmath21 in the table .
this gives overwhelming evidence that the semantic network has universal aspects that apply across language subgroups : for instance , in every group we tried , sea / ocean , and salt are more related than either is to sun . in the second bootstrap experiment ,
the null hypothesis is that the nonlinguistic variables have no effect on the semantic network , and that the differences between language groups simply result from random sampling : for instance , the similarity between the americas and eurasia is what one would expect from any disjoint subgroups of the 81 languages of given sizes 29 and 20 respectively . to test this null hypothesis
, we generate random pairs of disjoint language subgroups with the same sizes as the groups in question , and measure the distribution of their distances .
the @xmath18-values , called @xmath22 in figure [ fig : bootstrap ] , are not small enough to reject this null hypothesis .
thus , at least given the current data set , there is no statistical distinction between random sampling and empirical data further supporting our thesis that it is , at least in part , universal .
the model treats all concepts as independent members of an unbiased sample that the aggregate summary statistics of the empirical data reflects the underlying structure .
the simplest model perhaps then assumes no interaction between concept and languages : the number of polysemies of concept @xmath4 in language @xmath3 , that is @xmath23 , is linearly proportional to both the tendency of the concept to be polysemous and the tendency of the language to distinguish word senses ; and these tendencies are estimated from the marginal distribution of the observed data as the fraction of polysemy associated with the concept , @xmath24 , and the fraction of polysemy in the language , @xmath25 , respectively .
the model can , therefore , be expressed as , @xmath26 , a product of the two . to test the model ,
we compare the kullback - leibler ( kl ) divergence of ensembles of the model with the observation @xcite .
ensembles are generated by the multinominal distribution according to the probability @xmath27 .
the kl divergence is an appropriate measure for testing typicality of this random process because it is the leading exponential approximation ( by stirling s formula ) to the log of the multinomial distribution produced by poisson sampling ( see appendix [ sec : model ] ) .
the kl divergence of ensembles is @xmath28 where @xmath29 is the number of polysemies that the model generates divided by @xmath30 , and the kl divergence of the empirical observation is @xmath31 .
note that @xmath32 is @xmath33 and it is a different value from an expected value of the model , @xmath34 .
the @xmath18-value is the cumulative probability of @xmath35 to the right of @xmath36 .
hy acknowledges support from cabdyn complexity centre , and the support of research grants from the national science foundation ( no .
sma-1312294 ) .
wc and ls acknowledge support from the university of new mexico resource allocation committee .
tb , jw , es , cm , and hy acknowledge the santa fe institute , and the evolution of human languages program .
authors thank ilia peiros , george starostin , and petter holme for helpful comments .
w.c . and t.b
. conceived of the project and participated in all methodological decisions .
l.s . and w.c .
collected the data , h.y . , j.w .
, e.s . , and t.b .
did the modeling and statistical analysis .
i.m . and w.c .
provided the cross - linguistic knowledge .
, e.s . , and c.m .
did the network analysis .
the manuscript was written mainly by h.y .
, c.m . , and t.b . , and all authors agreed on the final version .
99 whorf bl , _ language , thought and reality : selected writing .
_ ( mit press , cambridge , 1956 ) .
fodor ja , _ the language of thought . _
( harvard univ . , new york , 1975 ) .
wierzbicka , a. , _ semantics : primes and universals . _
( oxford university press .
1996 ) lucy ja , _ grammatical categories and cognition : a case study of the linguistic relativity hypothesis . _
( cambridge university press , 1992 ) .
levinson sc , _ space in language and cognition : explorations in cognitive diversity .
_ ( cambridge university press , 2003 ) choi s , bowerman m ( 1991 ) learning to express motion events in english and korean : the influence of language - specific lexicalization patterns .
_ cognition _ * 41 * , 83 - 121 .
majid a , boster js , bowerman m ( 2008 ) _ cognition _ * 109 * , 235 - 250 .
croft w ( 2010 ) relativity , linguistic variation and language universals .
_ cognitextes _ * 4 * , 303 .
evans n , levinson sc ( 2009 ) the myth of language universals : language diversity and its importance for cognitive science .
brain sci . _
* 21 * 429 - 492 .
comrie b _ language universals and linguistic typology , 2nd ed .
_ ( university of chicago press . , 1989 ) .
croft w , _ typology and universals , 2nd ed . _ ( cambridge university press .
2003 ) .
henrich j , heine sj , norenzayan a ( 2010 ) the weirdest people in the world ? _ behav .
brain sci . _ * 33 * 1 - 75 ( 2010 ) .
shopen t ( ed . ) , _ language typology and syntactic description _ , 2nd ed .
( 3 volumes ) ( cambridge university press , cambridge , 2007 ) .
croft w , cruse da , _ cognitive linguistics . _
( cambridge university press .
2004 ) .
koptjevskaja - tamm m , vanhove m ( 2012 ) new directions in lexical typology .
_ linguistics _ * 50 * , 3 .
brown ch ( 1976 ) general principles of human anatomical partonomy and speculations on the growth of partonomic nomenclature .
ethnol . _ * 3 * , 400 - 424 ( 1976 ) .
witkowski sr , brown ch ( 1978 ) lexical universals , _ ann . rev . of anthropol . _ * 7 * 427 - 51 .
brown ch ( 1983 ) where do cardinal direction terms come from ?
_ anthropological linguistics _ * 25 * , 121 - 161 .
viberg ( 1983 ) the verbs of perception : a typological study .
_ linguistics _ * 21 * , 123 - 162 .
evans n , multiple semiotic systems , hyperpolysemy , and the reconstruction of semantic change in australian languages . in _
diachrony within synchrony : language history and cognition _
( peter lang .
frankfurt , 1992 ) .
derrig s ( 1978 ) metaphor in the color lexicon .
chicago linguistic society , the parasession on the lexicon _
85 - 96 .
swadesh m ( 1952 ) lexico - statistical dating of prehistoric ethnic contacts .
_ p. am . philos .
soc . _ * 96 * , 452 - 463 .
vygotsky l , thought and language .
( mit press , cambridge , ma , 2002 ) .
critchlow de , pearl dk , qian cl ( 1996 ) the triples distance for rooted bifurcating phylogenetic trees .
biol . _ * 45 * , 323334 .
dobson aj , comparing the shapes of trees , _
combinatorial mathematics iii _ , ( springer - verlag , new york 1975 ) .
robinson df , foulds lr ( 1981 ) comparison of phylogenetic trees .
biosci . _ * 53 * , 131147 .
dunn m , _ et al . _
( 2011 ) evolved structure of language shows lineage - specific trends in word - order universals .
_ nature _ * 473 * , 79 - 82 .
bouckaert r , _ et al . _
( 2012 ) mapping the origins and expansion of the indo - european language family .
_ science _ * 337 * , 957 .
fox a , _ linguistic reconstruction : an introduction to theory and method . _
( oxford university press .
1995 ) .
hock hh _ principles of historical linguistics .
_ ( mouton de gruyter , berlin , 1986 ) .
nichols j , the comparative method as heuristic . in _
the comparative method reviewed : regularity and irregularity in language change _
( oxford university press , 1996 ) .
dryer ms ( 1989 ) large linguistic areas and language sampling .
_ studies in language _ * 13 * , 257 - 292 .
cover tm , and thomas ja , elements of information theory , ( wiley , new york , 1991 ) .
brown ch , a theory of lexical change ( with examples from folk biology , human anatomical partonomy and other domains ) .
_ anthropol . linguist .
_ * 21 * , 257 - 276 ( 1979 ) .
brown ch & witkowski sr , figurative language in a universalist perspective .
_ * 8 * 596 - 615 ( 1981 ) .
our translations use only lexical concepts as opposed to grammatical inflections or function words . for the purpose of universality and stability of meanings across cultures ,
we chose entries from the swadesh 200-word list of basic vocabulary . among these , we have selected categories that are likely to have single - word representation for meanings , and for which the referents are material entities or natural settings rather than social or conceptual abstractions .
we have selected 22 words in domains concerning natural and geographic features , so that the web of polysemy will produce a connected graph whose structure we can analyze , rather than having an excess of disconnected singletons .
we have omitted body parts which by the same criteria would provide a similarly appropriate connected domain because these have been considered previously @xcite .
the final set of 22 words are as follows : * celestial phenomena and related time units : + star , sun , moon , year , day / daytime , night * landscape features : + sky , cloud(s ) , sea / ocean , lake , river , mountain * natural substances : + stone / rock , earth / soil , sand , ash(es ) , salt , smoke , dust , fire , water , wind the languages included in our study are listed in tab .
[ tab : languages ] .
notes : oceania includes southeast asia ; the papuan languages do not form a single phylogenetic group in the view of most historical linguists ; other families in the table vary in their degree of acceptance by historical linguists . the classification at the genus level , which is of greater importance to our analysis , is generally agreed upon . to accommodate such characteristic , we revise the model eq .
( [ seq : product ] ) to the following function : @xmath37 where degree numbers @xmath38 for each swadesh @xmath4 is proportional to @xmath39 and language size , but is bounded by @xmath40 , the number of proximal concepts .
the corresponding model probability for each language then becomes @xmath41 as all @xmath42 we recover the product model , with @xmath43 and @xmath44 . a first - level approximation to fit parameters @xmath39 and @xmath40
is given by minimizing the weighted mean - square error @xmath45 the function ( [ eq : error_sat ] ) assigns equal penalty to squared error within each language bin @xmath46 , proportional to the variance expected from poisson sampling .
the fit values obtained for @xmath39 and @xmath40 do not depend sensitively on the size of bins except for the swadesh entry moon in the case where all 81 single - language bins are used .
moon has so few polysemies , but the moon / month polysemy is so likely to be found , that the language itelman , with only one link , has this polysemy .
this point leads to instabilities in fitting @xmath47 in single - language bins . for bins of size 39
the instability is removed .
representative fit parameters across this range are shown in table [ tab : fitvalue ] .
examples of the saturation model for two words , plotted against the 9-language binned degree data in fig .
[ fig : saturating_words ] , show the range of behaviors spanned by swadesh entries . ) and the saturation model ( [ eq : sat2 ] ) .
parameters @xmath39 and @xmath40 have been adjusted ( as explained in the text ) to match the word- and language - marginals . from 10,000 random samples
@xmath48 , ( green ) histogram for the product model ; ( blue ) histogram for the saturation model ; ( red dots ) data .
the product model rejects the 9-language joint binned configuration at the at @xmath49 level ( dark shading ) , while the saturation model is typical of the same configuration at @xmath50 ( light shading ) .
[ fig : sample_kl_hist_data_9_joint ] ] the least - squares fits to @xmath39 and @xmath40 do not directly yield a probability model consisent with the marginals for language size that , in our data , are fixed parameters rather than sample variables to be explained .
they closely approximate the marginal n @xmath51 ( deviations @xmath52 link for every @xmath4 ) but lead to mild violations @xmath53 .
we corrected for this by altering the saturation model to suppose that , rather than word properties interacting with the exact value @xmath54 , they interact with a ( word - independent but language - dependent ) multiplier @xmath55 , so that the model for @xmath56 in each language becomes becomes @xmath57 in terms of the least - squares coefficients @xmath39 and @xmath40 of table [ tab : fitvalue ] .
the values of @xmath58 are solved with newton s method to produce @xmath59 , and we checked that they preserve @xmath60 within small fractions of a link .
the resulting adjustment parameters are plotted versus @xmath54 for individual languages in fig .
[ fig : varphis_vs_nls ] .
although they were computed individually for each @xmath3 , they form a smooth function of @xmath54 , possibly suggesting a refinement of the product model , but also perhaps reflecting systematic interaction of small - language degree distributions with the error function ( [ eq : error_sat ] ) . versus @xmath54 for individual languages in the probability model used in text , with parameters
@xmath40 and @xmath61 shown in table [ tab : fitvalue ] .
although @xmath58 values were individually solved with newton s method to ensure that the probability model matched the whole - language link values , the resulting correction factors are a smooth function of @xmath54 .
[ fig : varphis_vs_nls ] ] is now marginally plausible for the joint configuration of 27 three - language bins in the data , at the @xmath62 level ( light shading ) . for reference
, this fine - grained joint configuration rejects the null model of independent sampling from the product model at @xmath63 ( dark shading in the extreme tail ) .
4000 samples were used to generate this test distribution .
the blue histogram is for the saturation model , the green histogram for the product model , and the red dots are generated data .
[ fig : sample_kl_hist_data_27_joint ] ] with the resulting joint distribution @xmath64 , tests of the joint degree counts in our dataset for consistency with multinomial sampling in 9 nine - language bins are shown in fig . [
fig : sample_kl_hist_data_9_joint ] , and results of tests using 27 three - language bins are shown in fig .
[ fig : sample_kl_hist_data_27_joint ] .
binning nine languages clearly averages over enough language - specific variation to make the data strongly typical of a random sample ( @xmath65 ) , while the product model ( which also preserves marginals ) is excluded at the @xmath49 level .
the marginal acceptance of the data even for the joint configuration of three - language bins ( @xmath66 ) suggests that language size @xmath54 is an excellent explanatory variable and that residual language variations cancel to good approximation even in small aggregations .
the preceding subsection showed intermediate scales of aggregation of our language data are sufficiently random that they can be used to refine probability models for mean degree as a function of parameters in the globally - aggregated graph .
the saturation model , with data - consistent marginals and multinomial sampling , is weakly plausible by bins of as few as three languages .
down to this scale , we have therefore not been able to show a requirement for deviations from the independent sampling of links entailed by the use of the aggregate graph as a summary statistic . however , we were unable to find a further refinement of the mean distribution that would reproduce the properties of single language samples . in this section
we characterize the nature of their deviation from independent samples of the saturation model , show that it may be reproduced by models of non - independent ( clumpy ) link sampling , and suggest that these reflect excess synonymous polysemy .
+ * power tests and uneven distribution of single - language @xmath18-values * to evaluate the contribution of individual languages versus language aggregates to the acceptance or rejection of random - sampling models , we computed @xmath18-values for individual languages or language bins , using the kl - divergence ( [ eq : d_kl_l ] ) .
a plot of the single - language @xmath18-values for both the null ( product ) model and the saturation model is shown in fig . [
fig : one_lang_p_vals ] .
histograms for both single languages ( from the values in fig .
[ fig : one_lang_p_vals ] ) and aggregate samples formed by binning consecutive groups of three languages are shown in fig .
[ fig : p_dists_marg_sat ] . for samples from a random model
, @xmath18-values would be uniformly distributed in the unit interval , and histogram counts would have a multinomial distribution with single - bin fluctuations depending on the total sample size and bin width .
therefore , fig .
[ fig : p_dists_marg_sat ] provides a power test of our summary statistics .
the variance of the multinomial may be estimated from the large-@xmath18-value body where the distribution is roughly uniform , and the excess of counts in the small-@xmath18-value tail , more than one standard deviation above the mean , provides an estimate of the number of languages that can be confidently said to violate the random - sampling model . from the upper panel of fig .
[ fig : p_dists_marg_sat ] , with a total sample of 81 languages , we can estimate a number of @xmath67 excess languages at the lowest @xmath18-values of 0.05 and 0.1 , with an additional 23 languages rejected by the product model in the range @xmath18-value @xmath68 .
comparable plots in fig .
[ fig : p_dists_marg_sat ] ( lower panel ) for the 27 three - language aggregate distributions are marginally consistent with random sampling for the saturation model , as expected from fig .
[ fig : sample_kl_hist_data_27_joint ] above .
we will show in the next section that a more systematic trend in language fluctuations with size provides evidence that the cause for these rejections is excess variance due to repeated attachment of links to a subset of nodes .
by kl divergence , relative to 4000 random samples per language , plotted versus language rank in order of increasing @xmath54 . product model ( green )
shows equal or lower @xmath18-values for almost all languages than the saturation model ( blue ) .
three languages basque , haida , and yorb had value @xmath69 consistently across samples in both models , and are removed from subsequent regression estimates .
a trend toward decreasing @xmath18 is seen with increase in @xmath54 .
[ fig : one_lang_p_vals ] ] -values from the 81 languages plotted in fig .
[ fig : one_lang_p_vals ] . the saturation model ( blue )
produces a fraction @xmath70 languages in the lowest @xmath18-values @xmath71 above the roughly - uniform background for the rest of the interval ( shaded area with dashed boundary ) .
a further excess of 23 languages with @xmath18-values in the range @xmath72 $ ] for the product model ( green ) reflects the part of the mismatch corrected through mean values in the saturation model .
( lower panel ) corresponding histogram of @xmath18-values for 27 three - language aggregate degree distributions .
saturation model ( blue ) is now marginally consistent with a uniform distribution , while the product model ( green ) still shows slight excess of low-@xmath18 bins .
coarse histogram bins have been used in both panels to compensate for small sample numbers in the lower panel , while producing comparable histograms .
[ fig : p_dists_marg_sat ] , title="fig : " ] -values from the 81 languages plotted in fig .
[ fig : one_lang_p_vals ] . the saturation model ( blue )
produces a fraction @xmath70 languages in the lowest @xmath18-values @xmath71 above the roughly - uniform background for the rest of the interval ( shaded area with dashed boundary ) .
a further excess of 23 languages with @xmath18-values in the range @xmath72 $ ] for the product model ( green ) reflects the part of the mismatch corrected through mean values in the saturation model .
( lower panel ) corresponding histogram of @xmath18-values for 27 three - language aggregate degree distributions .
saturation model ( blue ) is now marginally consistent with a uniform distribution , while the product model ( green ) still shows slight excess of low-@xmath18 bins .
coarse histogram bins have been used in both panels to compensate for small sample numbers in the lower panel , while producing comparable histograms .
[ fig : p_dists_marg_sat ] , title="fig : " ] + if we define the size - weighted relative variance of a language analogously to the error term in eq .
( [ eq : error_sat ] ) , as @xmath73 fig .
[ fig : one_lang_pval_relvar_corr ] shows that @xmath74 has high rank correlation with @xmath75 and a roughly linear regression over most of the range . )
that the leading quadratic term in the kl - divergence differs from @xmath76 in that it presumes poisson fluctuation with variance @xmath77 at the level of each _ word _ , rather than uniform variance @xmath78 across all words in a language .
the relative variance is thus a less specific error measure . ]
two languages ( itelmen and hindi ) , which appear as large outliers relative to the product model , are within the main dispersion in the saturation model , showing that their discrepency is corrected in the mean link number .
we may therefore understand a large fraction of the improbability of languages as resulting from excess fluctuations of their degree numbers relative to the expectation from poisson sampling .
plotted versus relative variance @xmath79 from eq .
( [ eq : rel_var ] ) for the 78 languages with non - zero @xmath18-values from fig .
[ fig : one_lang_p_vals ] . ( blue ) saturation model ; ( green ) product model . two languages ( circled ) which appear as outliers with anomalously small relative variance in the product model itelman and hindi disappear into the central tendency with the saturation model .
( lower panel : ) an equivalent plot for 26 three - language bins .
notably , the apparent separation of individual large-@xmath54 langauges into two groups has vanished under binning , and a unimodal and smooth dependence of @xmath80 on @xmath79 is seen .
[ fig : one_lang_pval_relvar_corr ] , title="fig : " ] + plotted versus relative variance @xmath79 from eq .
( [ eq : rel_var ] ) for the 78 languages with non - zero @xmath18-values from fig .
[ fig : one_lang_p_vals ] . ( blue ) saturation model ; ( green ) product model .
two languages ( circled ) which appear as outliers with anomalously small relative variance in the product model
itelman and hindi disappear into the central tendency with the saturation model .
( lower panel : ) an equivalent plot for 26 three - language bins .
notably , the apparent separation of individual large-@xmath54 langauges into two groups has vanished under binning , and a unimodal and smooth dependence of @xmath80 on @xmath79 is seen .
[ fig : one_lang_pval_relvar_corr ] , title="fig : " ] fig .
[ fig : variance_test_sat_p_value_excludes ] then shows the relative variance from the saturation model , plotted versus total average link number for both individual languages and three - language bins .
the binned languages show no significant regression of relative variance away from the value unity for poisson sampling , whereas single languages show a systematic trend toward larger variance in larger languages , a pattern that we will show is consistent with `` clumpy '' sampling of a subset of nodes .
the disappearance of this clumping in binned distributions shows that the clumps are uncorrelated among languages at similar @xmath54 . for 78 languages excluding basque , haida , and yorb .
least - squares regression are shown for three - language bins ( green ) and individual languages ( blue ) , with regression coefficients inset .
three - language bins are consistent with poisson sampling at all @xmath54 , whereas single languages show systematic increase of relative variance with increasing @xmath54 .
[ fig : variance_test_sat_p_value_excludes ] ] we may retain the mean degree distributions , while introducing a systematic trend of relative variance with @xmath54 , by modifying our sampling model away from strict poisson sampling to introduce `` clumps '' of links . to remain within the use of minimal models , we modify the sampling procedure by a single parameter which is independent of word @xmath4 , language - size @xmath54 , or particular language @xmath3 .
we introduce the sampling model as a function of two parameters , and show that one function of these is constrained by the regression of excess variance .
( the other may take any interior value , so we have an equivalence class of models . ) in each language , select a number @xmath81 of swadesh entries randomly . let the swadesh indices be denoted @xmath82 .
we will take some fraction of the total links in that language , and assign them only to the swadeshes whose indices are in this privileged set .
introduce a parameter @xmath83 that will determine that fraction .
we require correlated link assignments be consistent with the mean determined by our model fit , since binning of data has shown no systematic effect on mean parameters .
therefore , for the random choice @xmath84 , introduce the normalized density on the privileged links @xmath85
if @xmath86 and @xmath87 otherwise . denote the aggregated weight of the links in the priviledged set by @xmath88 then introduce a modified probability distribution based on the randomly selected links , in the form @xmath89 multinomial sampling of @xmath54 links from the distribution @xmath90 will produce a size - dependent variance of the kind we see in the data .
the expectated degrees given any particular set @xmath91 will not agree with the means in the aggregate graph , but the ensemble mean over random samples of languages will equal @xmath92 , and binned groups of languages will converge toward it according to the central - limit theorem .
the proof that the relative variance increases linearly in @xmath54 comes from the expansion of the expectation of eq .
( [ eq : rel_var ] ) for random samples , denoted @xmath93 } ^2 \right > \nonumber \\ & = & \left < \frac{1}{n^l } \sum_s { \left ( { \hat{n}}_s^l - n^l { \tilde{p}}_{s \mid l } \right ) } ^2 \right > + \nonumber\\ & & \qquad n^l \left < \sum_s { \left ( { \tilde{p}}_{s \mid l } - p_{s \mid l}^{\mbox{\scriptsize model } } \right ) } ^2 \right > .
\label{eq : rel_var_clumpy}\end{aligned}\ ] ] the first expectation over @xmath48 is constant ( of order unity ) for poisson samples , and the second expectation ( over the sets @xmath91 that generate @xmath90 ) does not depend on @xmath54 except in the prefactor .
cross - terms vanish because link samples are not correlated with samples of @xmath94 .
both terms in the third line of eq .
( [ eq : rel_var_clumpy ] ) scale under binning as @xmath95 .
the first term is invariant due to poisson sampling , while in the second term , the central - limit theorem reduction of the variance in samples over @xmath90 cancels growth in the prefactor @xmath54 due to aggregation . because the linear term in eq .
( [ eq : rel_var_clumpy ] ) does not systematically change under binning , we interpret the vanishing of the regression for three - language bins in fig .
[ fig : variance_test_sat_p_value_excludes ] as a consequence of fitting of the mean value to binned data as sample estimators . ) , fitting a saturation model to binned sample configurations using the same algorithms as we applied to our data , and then performing regressions equivalent to those in fig .
[ fig : variance_test_sat_p_value_excludes ] . in about @xmath96 of cases the fitted model showed regression coefficients consistent with zero for three - language bins .
the typical behavior when such models were fit to random sample data was that the three - bin regression coefficient decreased from the single - language regression by @xmath97 .
] therefore , we require to choose parameters @xmath81 and @xmath83 so that regression coefficients in the data are typical in the model of clumpy sampling , while regressions including zero have non - vanishing weight in models of three - bin aggregations . fig .
[ fig : multinom_samp_hists_78_26_words ] compares the range of regression coefficients obtained for random samples of languages with the values @xmath98 in our data , from either the original saturation model @xmath99 , or the clumpy model @xmath90 randomly re - sampled for each language in the joint configuration .
parameters used were ( @xmath100 , @xmath101 ) . ranging from 317 .
@xmath100 was chosen as an intermediate value , consistent with the typical numbers of nodes appearing in our samples by inspection . ] with these parameters , @xmath97 of links were assigned in excess to @xmath97 of words , with the remaining @xmath102 of links assigned according to the mean distribution .
either generated by poisson sampling from the saturation model @xmath103 fitted to the data ( blue ) , or drawn from clumped probabilities @xmath90 defined in eq .
( [ eq : p_tilde_def ] ) , with the set of privileged words @xmath104 independently drawn for each language ( green ) .
solid lines refer to joint configurations of 78 individual languages with the @xmath54 values in fig .
[ fig : variance_test_sat_p_value_excludes ] .
dashed lines are 26 non - overlapping three - language bins .
[ fig : multinom_samp_hists_78_26_words ] ] the important features of the graph are : 1 ) binning does not change the mean regression coefficient , verifying that eq .
( [ eq : rel_var_clumpy ] ) scales homogeneously as @xmath95 .
however , the variance for binned data increases due to reduced number of sample points ; 2 ) the observed regression slope 0.012 seen in the data is far out of the support of multinomial sampling from @xmath99 , whereas with these parameters , it becomes typical under @xmath105 while still leaving significant probability for the three - language binned regression around zero ( even without ex - post fitting ) . | how universal is human conceptual structure ?
the way concepts are organized in the human brain may reflect distinct features of cultural , historical , and environmental background in addition to properties universal to human cognition .
semantics , or meaning expressed through language , provides direct access to the underlying conceptual structure , but meaning is notoriously difficult to measure , let alone parameterize . here
we provide an empirical measure of semantic proximity between concepts using cross - linguistic dictionaries . across languages carefully selected from a phylogenetically and geographically stratified sample of genera ,
translations of words reveal cases where a particular language uses a single polysemous word to express concepts represented by distinct words in another .
we use the frequency of polysemies linking two concepts as a measure of their semantic proximity , and represent the pattern of such linkages by a weighted network .
this network is highly uneven and fragmented : certain concepts are far more prone to polysemy than others , and there emerge naturally interpretable clusters loosely connected to each other .
statistical analysis shows such structural properties are consistent across different language groups , largely independent of geography , environment , and literacy .
it is therefore possible to conclude the conceptual structure connecting basic vocabulary studied is primarily due to universal features of human cognition and language use .
the space of concepts expressible in any language is vast .
this space is covered by individual words representing semantically tight neighborhoods of salient concepts .
there has been much debate about whether semantic similarity of concepts is shared across languages @xcite . on the one hand ,
all human beings belong to a single species characterized by , among other things , a shared set of cognitive abilities . on the other hand ,
the 6000 or so extant human languages spoken by different societies in different environments across the globe are extremely diverse @xcite and may reflect accidents of history as well as adaptations to local environments .
most psychological experiments about this question have been conducted on members of `` weird '' ( western , educated , industrial , rich , democratic ) societies , yet there is reason to question whether the results of such research are valid across all types of societies @xcite .
thus , the question of the degree to which conceptual structures expressed in language are due to universal properties of human cognition , the particulars of cultural history , or the environment inhabited by a society , remains unresolved .
the search for an answer to this question has been hampered by a major methodological difficulty .
linguistic meaning is an abstract construct that needs to be inferred indirectly from observations , and hence is extremely difficult to measure ; this is even more apparent in the field of lexical semantics .
meaning thus contrasts both with phonetics , in which instrumental measurement of physical properties of articulation and acoustics is relatively straightforward , and with grammatical structure , for which there is general agreement on a number of basic units of analysis @xcite .
much lexical semantic analysis relies on linguists introspection , and the multifaceted dimensions of meaning currently lack a formal characterization . to address our primary question ,
it is necessary to develop an empirical method to characterize the space of lexical meanings .
we arrive at such a measure by noting that translations uncover the alternate ways that languages partition meanings into words .
many words have more than one meaning , or sense , to the extent that word senses can be individuated @xcite .
words gain meanings when their use is extended by speakers to similar meanings ; words lose meanings when another word is extended to the first word s meaning , and the first word is replaced in that meaning . to the extent that words in transition across similar , or possibly contiguous , meanings account for
the polysemy ( multiple meanings of a single word form ) revealed in cross - language translations , the frequency of polysemies found across an unbiased sample of languages can provide a measure of semantic similarity among word meanings .
the unbiased sample of languages is carefully chosen in a phylogenetically and geographically stratified way , according to the methods of typology and universals research @xcite .
this large , diverse sample of languages allows us to avoid the pitfalls of research based solely on `` weird '' societies and to separate contributions to the empirically attested patterns in the linguistic data , arising from universal language cognition versus those from artifacts of the speaker - groups history or way of life .
there have been several cross - linguistic surveys of lexical polysemy , and its potential for understanding semantic shift @xcite , in the domains such as body parts @xcite , cardinal directions @xcite , perception verbs @xcite , concepts associated with fire @xcite , and color metaphors @xcite .
we add a new dimension to the existing body of research by providing a comprehensive mathematical method using a systematically stratified global sample of languages to measure degrees of similarity .
our cross - linguistic study takes the swadesh lists as basic concepts @xcite as most languages have words for them . among those concepts
, we chose 22 meanings associated with two domains : celestial objects ( e.g. ` sun , moon , star ` ) and landscape objects ( e.g. ` fire , water , mountain , dust ` ) .
for each word expressing one of these meanings , we examined what other concepts were also expressed by the word .
since the semantic structures of these two domains are very likely to be influenced by the physical environment that human societies inhabit , any claim of universality of lexical semantics needs to be demonstrated here . |
we work over @xmath2 . fix a separated deligne - mumford ( dm ) stack @xmath3 and a compactification @xmath4 . given an open subset @xmath0 of a normal variety @xmath1 together with a regular map @xmath5 , when does this map extend to a regular map @xmath6 ?
let s illustrate why this can be interesting geometrically .
we will denote by @xmath7 the moduli stack of @xmath8-pointed genus @xmath9 curves and by @xmath10 its coarse moduli space .
also , @xmath11 and @xmath12 will denote their respective deligne - mumford compactifications .
now take @xmath13 and @xmath14 ( no marked points ) .
the simplest case is when the dimension of @xmath1 is one .
since extending a map is a local problem we take @xmath15 to be a curve and @xmath16 the complement of a point .
section 2 of @xcite shows that a family of smooth curves over @xmath17 extends to a family of stable curves over @xmath18 if and only if the associated jacobian family extends to a family of semi - abelian varieties . on the other hand , sections 3.53.8 of @xcite
show that a family of abelian varieties over @xmath17 extends to a family of semi - abelian varieties over @xmath18 if and only if the associated monodromy on the @xmath19 homology of a fibre in a small ( analytic ) neighbourhood of @xmath20 is unipotent . combining these results
one obtains : [ thm : dmg ] a family of smooth curves over @xmath17 extends to a family of stable curves over @xmath18 if and only if the induced monodromy on @xmath19 of the fibres around an ( analytic ) neighbourhood of @xmath20 is unipotent . if @xmath21 you may need to blow up the base before you can extend @xmath22 .
a neat example is the family @xmath23 of elliptic curves over @xmath24 .
the fibres are smooth over @xmath25 and stable over @xmath26 .
the induced map @xmath27 does not extend over @xmath28 . to extend it one has to blow up three times to get the surface @xmath29 shown in figure [ f1c ] .
the map @xmath30 now extends to a morphism @xmath31 to the coarse moduli compactification which collapses the exceptional divisors @xmath32 and @xmath33 to the points with @xmath34 invariants @xmath35 and @xmath36 and maps @xmath37 one - to - one onto @xmath38 .
is the discrimant curve @xmath39 .
[ f1c],title="fig : " ] as this example illustrates , it was necessary to resolve @xmath40 to a normal crossing divisor before the map extended .
this is not quite a coincidence . in @xcite
it was shown that ( see remark [ rem : jo ] ) : [ thm : jo ] let @xmath40 be a normal crossing divisor at @xmath20 .
then a morphism @xmath41 extends to a regular map @xmath42 in a zariski neighbourhood of @xmath20 .
theorem [ thm : jo ] holds in arbitrary characteristic ( which is significant ) but if one works over @xmath2 we have the following generalization ( corollary [ cor7 ] ) .
[ specialthm1 ] let @xmath0 be an open subvariety of an irreducible , normal variety @xmath1 .
a morphism @xmath41 extends to a regular map @xmath42 in a zariski neighbourhood of @xmath43 if and only if the local monodromy around @xmath20 is virtually abelian . notice that since the local fundamental group of the complement of a normal crossing divisor is abelian , theorem [ specialthm1 ] implies [ thm : jo ] over @xmath2 .
the result in @xcite actually gives an extension to @xmath11 ( not just @xmath12 ) but only under the added condition that the morphism extends to a regular map @xmath44 over the generic points of @xmath45 .
this condition is sufficient to lift the extension from @xmath12 to @xmath11 ( see corollary [ cor4 ] ) .
one can ask if such extension results exist for other spaces .
another example involves the moduli stack of @xmath9-dimensional principally polarized abelian varieties @xmath46 and its baily - borel compactification @xmath47 .
it was shown in @xcite that [ thm : b ] let @xmath40 be a normal crossing divisor at @xmath20 .
then a morphism @xmath48 extends to a regular map @xmath49 in a zariski neighbourhood of @xmath20 . as an immediate consequence of theorem [ thm5 ] it follows that
: [ specialthm2 ] let @xmath0 be an open subvariety of an irreducible , normal variety @xmath1 .
a morphism @xmath48 extends to a regular map @xmath49 in a zariski neighbourhood of @xmath43 if the local monodromy around @xmath20 is virtually abelian .
notice that in thorem [ specialthm2 ] there is no `` only if '' part .
this is because we do nt know whether the local fundamental group of @xmath46 around a point in the boundary of @xmath47 is virtually abelian .
understanding this local fundamental group is an interesting question in itself .
inspired by these results we introduce the abelian monodromy extension ( ame ) property for a pair @xmath50 consisting of a deligne - mumford stack @xmath3 and a compactification @xmath51 of its coarse scheme ( section [ s1 ] ) .
roughly , we say @xmath51 is an ame compactification of @xmath3 if @xmath5 extends to a regular map @xmath52 whenever the image of the induced map @xmath53 is virtually abelian . here
we think of @xmath54 and @xmath1 as being very small analytic neighbourhoods of a point ( so this is a local condition on the domain but a global condition on the target @xmath50 ) .
section [ s2 ] describes some basic properties of such ame compactifications and shows that if the ame compactification of @xmath3 exists then there is a unique maximal one which we denote @xmath55 ( corollary [ cor6 ] ) .
section [ s3 ] theorem [ thm6 ] shows that @xmath12 is the maximal ame compactification of @xmath7 and derives theorem [ specialthm1 ] as a corollary .
finally , section [ s4 ] theorem [ thm5 ] proves that @xmath47 is the maximal ame compactification of @xmath46 ( among all projective ones ) . in a future paper
we will develop the theory of ame compactifications and give further geometric examples .
we ll also explore a connection between ame compactifications and log canonical models which , among other things , gives a more direct proof of the maximality statement in theorem [ thm5 ] .
i would like to thank samuel grushevsky , joe harris , brendan hassett , klaus hulek and sean keel for many insightful discussions .
johan de jong and jason starr also made helpful suggestions early on while richard hain and jnos kollr were very kind in pointing out the utility of @xcite in proving theorem [ thm5 ] .
i appreciate the support and hospitality of the mathematics departments at harvard and rice university and of the mittag - leffler institute where i was a visitor in the spring of 2007 .
i would also like to thank the referees for helpful , detailed suggestions in particular with respect to defining local monodromy and simplifying some proofs .
in this section we define what it means for a variety ( or stack ) to have the abelian monodromy extension ( ame ) property .
we also make a few remarks about the definition and give an example . all schemes are noetherian . by a variety
we mean an integral , separated ( but possibly singular ) scheme of finite type defined over @xmath2 .
by morphism we will always mean a regular morphism .
a complete scheme is a scheme proper over @xmath56 and a compactification of a variety @xmath57 is a complete variety @xmath51 containing @xmath57 as an open dense subset .
let @xmath57 be an open subvariety of a normal variety @xmath51 .
next , consider an open subvariety @xmath0 of a normal variety @xmath1 together with a morphism @xmath58 . fix a
connected , reduced , complete subscheme @xmath59 .
we can always restrict to a smaller open subset @xmath54 if necessary so that @xmath54 is disjoint from @xmath60 as well as the singular locus of @xmath1 .
let @xmath40 .
since we can find a blowup @xmath61 so that @xmath62 , @xmath63 and @xmath64 are all simple , normal crossing divisors in @xmath29 we can assume that @xmath65 , @xmath45 and @xmath60 are all simple , normal crossing divisors .
this means that around any point @xmath66 there exist local coordinates such that @xmath60 is given by @xmath67 and @xmath45 is given by @xmath68 .
notice that we can choose @xmath0 so that this blowup leaves @xmath54 unchanged .
now choose a small analytic neighbourhood @xmath69 of @xmath60 .
one way to do this is to put a riemannian metric @xmath70 on @xmath1 and take @xmath71 then we can define the _ local monodromy around @xmath60 _ as the image @xmath72 since @xmath60 is connected @xmath73 is connected and so the image of @xmath74 is defined ( up to conjugation ) without having to choose a base point .
there are three things to check to make sure local monodromy is well defined .
the first is that local monodromy does not depend on the choice of @xmath69 . to see this note we identify @xmath75 as sitting inside its normal bundle ( which is a line bundle )
this means that @xmath76 deformation retracts onto @xmath77 for any @xmath78 sufficiently small .
thus @xmath79 and @xmath80 have isomorphic images in @xmath81 .
moreover , since for any other metric @xmath82 we have @xmath83 for @xmath84 we see that this image does not depend on the metric
. now the images of @xmath85 are the same since the map @xmath86 induced by inclusion is surjective .
thus @xmath87 and @xmath88 have the same image for sufficiently small @xmath89 .
the second thing is that local monodromy does not depend on the choice of @xmath54 .
this follows since if @xmath90 is ( in the analytic topology ) a dense , open subset then the map @xmath91 induced by the inclusion is surjective .
similarly , @xmath92 is surjective and hence the images of @xmath93 are the same . the third thing to check is that local monodromy does not depend on the choice of the blowup @xmath29 . to see this
consider a blowup @xmath61 so that @xmath62 , @xmath63 and @xmath64 are also simple , normal crossing divisors .
if you take a small neighbourhood @xmath69 of @xmath60 then @xmath94 is also a small neighbourhood of @xmath64 and then taking @xmath0 so that it avoids the blowup locus we get @xmath95 so that @xmath96 and hence the images of @xmath97 are the same
. most commonly we will take @xmath60 to be a point @xmath43 to obtain the local monodromy around @xmath20 . on the other hand ,
if @xmath1 is complete , we can take @xmath60 to be all of @xmath1 to obtain the _ global monodromy of @xmath1 _ as the image of the induced map @xmath98 .
the reason for this slightly strange terminology is that @xmath57 will often be a moduli space , for instance @xmath99 or @xmath46 , so that the image of @xmath98 can be identified with the usual notion of monodromy .
for instance , @xmath100 ( the orbifold fundamental group ) is the mapping class group @xmath101 so the image of @xmath102 is the monodromy on @xmath103 of a fibre in the family of curves over @xmath54 corresponding to the morphism @xmath22 .
@xmath104 so the image of @xmath105 induced by a morphism @xmath48 is the monodromy on @xmath19 of a fibre in the associated family of abelian varieties . it does not actually matter what dense open @xmath54 we pick in the above definition of monodromy ( as long as the map @xmath58 is regular )
this is because the map @xmath106 is surjective if @xmath107 is dense in the analytic topology so that the images of @xmath93 are the same .
thus sometimes we will take @xmath54 to be the largest open subset where the map @xmath108 is regular .
given an open embedding of normal varieties @xmath109 , the pair @xmath110 has _ the abelian monodromy extension ( ame ) property _ if given any @xmath0 as above the morphism @xmath58 extends to a regular map @xmath52 in a neighbourhood of @xmath20 whenever the local monodromy around @xmath20 is virtually abelian ( a group is virtually abelian if it contains an abelian subgroup of finite index ) . in this case
@xmath51 is complete ( see lemma [ lem1 ] below ) and we say that @xmath51 is an _ ame compactification _ of @xmath57 .
we say @xmath57 has the ame property if it has an ame compactification .
if the pair @xmath110 has the ame property then @xmath51 is complete .
let @xmath111 be a compactification of @xmath51 ( any separated scheme of finite type over @xmath2 has a compactification ) .
consider a blowup @xmath112 so that the boundary @xmath113 is a normal crossing divisor .
then the fundamental group of @xmath114 for a sufficiently small neighbourhood @xmath69 of any point @xmath115 is abelian .
since @xmath116 has the ame property this means that the morphism @xmath117 extends to a morphism @xmath118 .
then @xmath119 is complete and @xmath120 is dominant so @xmath51 must also be complete ( since the image of a complete variety is complete ) . alternatively , one might want to define the ame property by requiring that the morphism extend whenever the local monodromy is abelian instead of virtually abelian .
it turns out these two definitions are the same . to see this suppose the local monodromy is virtually abelian .
the preimage under @xmath98 of a finite index abelian subgroup inside @xmath121 is a finite index subgroup inside @xmath122 and contains a normal finite index subgroup @xmath123 .
take the unramified cover @xmath124 corresponding to @xmath125 and let @xmath126 be the normal closure of @xmath1 in the function field of @xmath127 to get a finite morphism @xmath128 .
then the local monodromy around any point of @xmath129 induced by @xmath130 is abelian and we get a regular extension @xmath131 . by lemma [ lem4 ] below this
means we also get a regular extension @xmath52 .
[ lem4 ] consider the composition @xmath132_{\pi } & \\ s \ar@{-->}[r]^h & x } \end{xy}\ ] ] where @xmath133 , @xmath1 are quasi - projective varieties , @xmath1 is normal and @xmath57 is complete . if @xmath134 is a finite , surjective morphism then the rational map @xmath135 is regular if and only if @xmath136 is regular . notice that this result does not hold if @xmath57 is a stack .
we need to show @xmath135 is regular assuming @xmath136 is regular .
let @xmath29 be the closure of the image of @xmath1 in @xmath137 .
now if @xmath54 is the domain where @xmath135 is regular then @xmath138 is open and dense in @xmath133 .
then the image of @xmath139 is the closure of the image of @xmath54 which is @xmath29 .
thus @xmath140 is finite ( since @xmath141 is finite ) and birational ( since @xmath29 is the closure of the image of @xmath1 ) . since @xmath1 is normal this means @xmath120 is an isomorphism and thus @xmath135 is regular .
consider a smooth , deligne - mumford stable @xmath8-pointed curve @xmath142 .
then @xmath18 is an ame compactification of @xmath143 . to see this
the key observation is that @xmath142 is stable precisely when @xmath144 is not virtually abelian .
more precisely , let @xmath0 be as above and suppose we have a morphism @xmath145 which does not extend to a regular morphism around some point @xmath43 .
we want to show that this implies the local monodromy around @xmath20 is not virtually abelian .
first resolve the map @xmath146 by some blowup @xmath61 to get a morphism @xmath147 . since the morphism @xmath145 did not extend to a regular map @xmath148 around @xmath20
the restricted map @xmath149 is dominant ( since @xmath18 is one - dimensional ) .
now the local monodromy around @xmath20 with respect to @xmath146 is the same as the local monodromy around @xmath129 with respect to @xmath150 .
this local monodromy is the image of @xmath151 where @xmath69 is some analytic open set whose analytic closure contains @xmath129 .
now we can replace @xmath69 by a bigger open subset @xmath152 which contains @xmath153 such that @xmath154 is dense .
since the map @xmath155 is surjective the images of @xmath156 will be the same .
now if the local monodromy around @xmath129 were virtually abelian then the image of @xmath157 would be virtually abelian .
but this means that the image of @xmath158 would be virtually abelian since it is a subgroup of the image in ( [ eq2 ] ) .
now @xmath159 is a dominant map .
so by lemma [ lem3 ] the induced image of fundamental groups is a finite index subgroup of @xmath144 ( which is not virtually abelian by the key observation above ) .
thus the local monodromy around @xmath129 can not be virtually abelian .
[ lem3 ] let @xmath61 be a dominant morphism between irreducible varieties .
if @xmath1 is normal then the image of the induced map @xmath160 is a finite index subgroup .
see lemma 3.3 ( resp .
lemma 11 ) of @xcite .
alternatively , here is a short proof suggested by the referee . by taking general hyperplane sections of @xmath29
we can find @xmath161 such that the composition @xmath162 generically finite .
then we can restrict @xmath120 to some open @xmath0 such that @xmath163 is finite tale .
so the image @xmath164 has finite index . since @xmath165 is surjective ( this is where we use that @xmath1 is normal )
this means the image of @xmath166 has finite index . since this map factors through @xmath160 the result follows .
the definition of the ame property can be extended to stacks as follows .
we denote stacks using caligraphic font and their coarse moduli spaces by ordinary font .
our stacks will be integral , separated , normal deligne - mumford ( dm ) stacks of finite type over @xmath2 .
notice that by keel - mori @xcite a dm stack always has a coarse moduli scheme .
we will need to use the ( topological ) fundamental group of a stack . for a reference see @xcite .
if the stack is a smooth dm stack then this agrees with the orbifold fundamental group introduced by thurston ( see also @xcite ) .
if @xmath3 is a stack and @xmath5 a regular morphism then we define the local monodromy around @xmath167 to be @xmath168 where @xmath69 is a sufficiently small neighbourhood of @xmath20 as before .
then a variety @xmath51 containing the coarse space @xmath57 of @xmath3 as an open subset is an ame compactification of @xmath3 if the morphism @xmath5 extends to a morphism @xmath52 in a neighbourhood of @xmath20 whenever the local monodromy around @xmath20 is virtually abelian . as in lemma [ lem1 ] this implies @xmath51 is complete .
note that we only consider a compactification @xmath51 of @xmath57 rather than @xmath3 .
there are basically two reasons to do this . 1 . under this definition
the pair @xmath169 has the ame property ( theorem [ thm6 ] ) . on the other hand ,
it is not true that every morphism @xmath170 from a punctured , smooth curve extends to a morphism @xmath171 to the fine moduli stack ( the central fibres might not be a stable curves ) .
since the local fundamental group of any puncture in @xmath172 is @xmath173 ( and in particular abelian ) this means that @xmath174 can not have the ame property in any reasonable sense .
+ similarly , the pair @xmath175 has the ame property ( theorem [ thm5 ] ) even though the baily - borel compactification @xmath47 only compactifies @xmath176 .
so in both these cases it makes sense to consider compactifications of @xmath57 rather than @xmath3 .
2 . suppose @xmath4 is a compactification of @xmath3 with @xmath51 its coarse moduli scheme .
further suppose that we have a morphism @xmath52 extending @xmath177 .
on @xmath1 the fibre product @xmath178 is ( tale locally ) of the form @xmath179 \rightarrow s$ ] where @xmath180 is a finite cover and @xmath181 is the group of deck transformations .
thus @xmath52 lifts to a morphism @xmath182 on a finite cover @xmath133 of @xmath1 .
this means that the extra stack structure on @xmath4 can be ignored if you allow finite base changes and tale localization .
one might then be tempted to consider only schemes .
however , the pair @xmath183 does _ not _ have the ame property because @xmath184 is simply connected ( once you loose the stack structure on @xmath99 the fundamental group becomes trivial ) .
so if we ignored stacks altogether we would lose some nice geometric examples of ame pairs like @xmath169 and @xmath175 .
[ prop4 ] let @xmath3 and @xmath185 be two dm stacks whose coarse spaces @xmath57 and @xmath186 have normal compactifications @xmath51 and @xmath111 . let @xmath187 be a morphism which extends to a finite morphism @xmath188 . if @xmath189 has the ame property then @xmath190 has the ame property .
conversely , if @xmath187 is also representable , finite , tale and @xmath190 has the ame property then @xmath189 has the ame property .
suppose @xmath189 has the ame property and consider a morphism @xmath191 from an open subset @xmath0 .
if the local monodromy around @xmath43 with respect to @xmath191 is virtually abelian then the local monodromy around @xmath20 with respect to @xmath5 is the image of the composition @xmath192 and is also virtually abelian .
hence the composition morphism @xmath5 extends to a regular morphism @xmath52 . since @xmath188 is finite the morphism @xmath52 lifts to a morphism @xmath193 by lemma [ lem5 ]
this shows @xmath190 has the ame property .
conversely , suppose that @xmath190 has the ame property and @xmath187 is tale . consider a morphism @xmath5 with virtually abelian local monodromy around @xmath43 .
consider the fibre product @xmath194 and denote by @xmath133 the normal closure of @xmath1 in the function field of @xmath127 .
for any point @xmath195 in the preimage of @xmath20 under @xmath180 the local monodromy around @xmath195 with respect to @xmath196 is virtually abelian .
since @xmath187 is finite , tale the map @xmath197 is injective so the local monodromy around @xmath195 with respect to @xmath198 must also be virtually abelian .
this means the morphism @xmath198 extends to a morphism @xmath199 in a neighbourhood of @xmath195 . since this is true for every such @xmath195 the morphism @xmath196 extends to a morphism @xmath200 in a neighbourhood of the preimage of @xmath20 under @xmath180 .
if , in the notation of lemma [ lem4 ] , we look at @xmath201 then the composition is regular so @xmath52 is regular .
thus @xmath5 also extends to a morphism @xmath52 in a neighbourhood of @xmath20 .
this shows @xmath50 has the ame property .
[ lem5 ] consider the composition @xmath202^h & y \ar[d]^{\pi } \\ &
\sx } \end{xy}\ ] ] where @xmath1 and @xmath186 are varieties , @xmath1 is normal and @xmath3 is a dm stack . if @xmath134 is a finite morphism then the rational morphism @xmath135 is regular if and only if @xmath203 is regular .
as the referee points out , this actually follows from lemma [ lem4 ] .
we need to show that if @xmath203 is regular then so is @xmath135 .
if one takes @xmath204 then we are left with the diagram @xmath132_{\pi ' } \ar[dr ] & \\ s \ar@{-->}[r]^{h } & y } \end{xy}\ ] ] where @xmath205 is a finite morphism .
so @xmath135 is regular by lemma [ lem4 ] .
[ cor8 ] suppose @xmath189 has the ame property and let @xmath206 be a closed , normal dm substack .
denote by @xmath57 and @xmath186 the correponding coarse moduli schemes . then @xmath207 and if we denote by @xmath111 the normalization of the closure of @xmath186 in @xmath51 then the pair @xmath190 also has the ame property .
this follows from proposition [ prop4 ] by considering a closed immersion @xmath208 .
we just need to show that @xmath206 implies @xmath207 .
to do this we use that a dm stack @xmath3 over @xmath2 ( or more generally a tame stack in the sense of @xcite ) is tale locally @xmath3 of the form @xmath209 $ ] for some scheme @xmath60 with an action of a finite group @xmath181 ( see , for example , the introduction of @xcite ) .
more precisely , this means that there exists a finite tale cover @xmath210 of the coarse moduli scheme of @xmath3 such that @xmath211 is locally in the zariski topology isomorphic to @xmath209 \rightarrow ( t / g)$ ] .
the fact that @xmath208 is some closed substack means that @xmath212 \hookrightarrow [ t / g ] = \sx \times_x \tx\ ] ] where @xmath213 is a closed @xmath181-invariant subscheme .
now consider the following commutative diagram @xmath214 & \sy \times_x \tx = [ t'/g ] \ar[d ] \ar[r ] \ar[l ] & ( t'/g ) = y
\times_x \tx \ar[r ] \ar[d ] & y \ar[d ] \\ \sx \ar[d ] & \sx \times_x \tx = [ t / g ] \ar[d ] \ar[r ] \ar[l ] & ( t / g ) = \tx \ar[r ] & x \\ x & \tx = ( t / g ) \ar[l ] & & } \end{xy}\ ] ] looking at the right - most square
we see that @xmath215 because @xmath213 . since @xmath210 is an tale cover it follows @xmath216 .
[ prop6 ] if @xmath217 and @xmath218 are two ame pairs then @xmath219 is also an ame pair .
consider an open subvariety @xmath0 of a normal variety @xmath1 together with a morphism @xmath220 with virtually abelian monodromy around @xmath43 .
the composition with the two projection maps yields morphisms @xmath221 and @xmath222 .
the monodromy around @xmath20 with respect to these maps is also virtually abelian . since @xmath223 and @xmath224 are ame compactifications these maps extend to regular morphisms @xmath225 and @xmath226 in a neighbourhood of @xmath20 .
this gives a regular morphism @xmath227 .
note that the coarse scheme of a product of two dm stacks is the product of the coarse schemes . to see this suppose @xmath228 and @xmath229 are dm stacks with coarse schemes @xmath230 and @xmath231 and let @xmath232 denote the coarse scheme of @xmath233 .
we must show that @xmath234 .
now tale locally @xmath228 and @xmath229 are both quotients of a scheme by a finite group @xcite .
this means that zariski locally on @xmath230 and @xmath231 we can find tale covers @xmath235 and @xmath236 such that @xmath237 is isomorphic to @xmath209 \rightarrow ( t / g)$ ] and @xmath238 is isomorphic to @xmath239 \rightarrow ( t'/g')$ ] for some schemes @xmath240 with finite group actions @xmath241 .
consider the following commutative diagram where every square is a fibre product : @xmath242 & w : = ( \sx_1 \times \sx_2 ) \times_{x_1 \times x_2 } ( \tx_1 \times \tx_2 ) \ar[d ] \ar[l ] \\ z \ar[d]^f & z \times_{x_1 \times x_2 } ( \tx_1 \times \tx_2 ) \ar[d]^g \ar[l ] \\
x_1 \times x_2 & \tx_1 \times \tx_2 .
\ar[l ] } \end{xy}\ ] ] now @xmath243 is the coarse scheme of @xmath244 \times [ t'/g'].\ ] ] on the other hand , since @xmath232 is the coarse scheme of @xmath233 it must be that @xmath245 is also the coarse scheme of @xmath246 ( corollary 3.3 of @xcite ) .
thus @xmath9 must be an isomorphism and since @xmath247 is an tale cover @xmath120 must also be an isomorphism .
we use [ cor8 ] and [ prop6 ] to construct more examples of ame compactifications . choose four general points @xmath248 on @xmath249 and let @xmath250 denote the line between @xmath251 and @xmath252 .
if we take @xmath253 and @xmath254 then @xmath255 is an ame pair . to see this
consider the morphisms @xmath256 for @xmath257 which take a point @xmath258 and map it to the line @xmath259 $ ] .
it is not hard to see that the product of these four maps gives an embedding @xmath260 whose closure we denote @xmath261 .
actually , @xmath261 is nothing but @xmath249 blown up at the four points @xmath251 . since @xmath262 is an ame pair so is @xmath255 by an application of [ cor8 ] and [ prop6 ] . in fact
since @xmath261 projects onto @xmath249 we find that @xmath264 is also an ame pair .
more generally , we may take @xmath265 general points @xmath266 and remove all hyperplanes through any @xmath8 of them to obtain @xmath267 .
as before , we have morphisms @xmath268 for @xmath269 taking a point @xmath270 to the line @xmath259 $ ] .
the product of these maps @xmath271 gives an embedding of @xmath267 .
denote by @xmath272 the closure of @xmath273 inside @xmath274 . by induction
, @xmath275 is an ame pair and so @xmath276 is also an ame pair . on the other hand
, @xmath277 extends to an embedding of @xmath278 while it is undefined at the points @xmath279 .
if one blows up @xmath280 at the points @xmath279 then @xmath277 extends and we find that @xmath272 is precisely this blowup .
thus @xmath272 projects onto @xmath280 and so @xmath281 is an ame pair .
[ cor5 ] let @xmath223 and @xmath224 be two ame compactifications of @xmath3 .
then there exists an ame compactification @xmath51 of @xmath3 which dominates @xmath223 and @xmath224 .
@xmath51 dominates @xmath223 and @xmath224 if it admits regular maps @xmath282 and @xmath283 which extend the identity map on @xmath57 .
take @xmath51 to be the normalization of the closure of @xmath57 embedded inside @xmath284 by the diagonal map .
the two projections give us @xmath103 and @xmath285 so it remains to show @xmath50 has the ame property . by proposition [ prop6 ]
@xmath219 is an ame pair so by corollary [ cor8 ] the closure @xmath51 of @xmath57 in @xmath284 is an ame compactification .
[ cor6 ] if @xmath3 has the ame property then it has a unique , maximal ame compactification @xmath286 in the sense that for any other ame compactification @xmath51 there is a birational , regular map @xmath287 . by corollary ( [ cor5 ] )
any two ame compactifications of @xmath3 are dominated by a larger ame compactification .
if there is no maximal ame compactification there exists a sequence @xmath288 of ame compactifications where each birational map @xmath289 between complete , normal , reduced schemes is not an isomorphism .
take any normal crossing compactification @xmath111 of @xmath57 .
the local fundamental group of a neighbourhood of @xmath290 is free abelian .
this means every open inclusion @xmath291 extends to a regular map @xmath292 . by lemma ( [ lem6 ] )
we know @xmath293 for @xmath294
so this is impossible .
if @xmath51 and @xmath295 are two maximal ame compactifications then there are birational , regular maps @xmath296 and @xmath297 which implies @xmath298 and so a maximal ame compactification is unique .
[ lem6 ] let @xmath299 be a sequence of birational morphisms between complete , normal , reduced schemes .
let @xmath186 be a complete , reduced scheme equipped with dominant morphisms @xmath300 satisfying @xmath301 for each @xmath302 .
then @xmath303 for @xmath294 . here is a short proof suggested by the referee .
we have natural inclusions @xmath304 .
these form a decreasing sequence of closed subsets @xmath305 inside the noetherian scheme @xmath306 .
thus they must stabilize at some point .
now if @xmath307 is not an isomorphism then by zariski s main theorem there must be two points @xmath308 such that @xmath309 . since the @xmath310 are surjective ( they are dominant and all varieties are complete ) we can find points @xmath311 such that @xmath312 .
but then @xmath313 belongs to @xmath305 but not to @xmath314 .
hence @xmath315 .
since the subschemes @xmath316 are the same for @xmath294 this means that @xmath317 is an isomorphism for @xmath294 . by construction ,
any ame compactification @xmath51 of @xmath3 is obtained from @xmath286 by contracting parts of the boundary .
conversely , given a contraction of the boundary @xmath287 , @xmath50 is an ame pair .
so to understand all ame compactifications of @xmath3 it is enough to understand the maximal one @xmath286 and all possible contractions of its boundary .
[ prop7 ] if @xmath3 has the ame property and @xmath318 is a locally closed embedding then there exists a regular morphism @xmath319 which extends @xmath320 .
the closure @xmath111 of @xmath186 in some @xmath286 is an ame compactification . by the defining property of @xmath321 there
exist morphisms @xmath322 whose composition extends @xmath320 the following theorem shows that varieties having the ame property are numerous .
let @xmath109 be a dense , open immersion with @xmath51 normal , complete variety .
then there exists an open @xmath323 such that @xmath324 has the ame property .
suppose @xmath325 is projective .
by example 2 there exists an open subvariety @xmath326 such that @xmath327 is an ame pair .
let @xmath328 .
the closure of @xmath329 is @xmath51 so that @xmath330 is an ame pair by corollary [ cor8 ] .
if @xmath51 is not projective then by chow s lemma there exists a proper , birational morphism @xmath188 where @xmath111 is normal and projective . then by the argument above we can find @xmath331 such that @xmath332 is an ame pair . restricting @xmath333
further we can even assume @xmath334 is an isomorphism with @xmath335 .
this immediately implies @xmath324 is an ame pair where @xmath336 .
we end the section with a lemma , which though not entirely necessary , will simplify some of the subsequent proofs .
[ lem7 ] to show @xmath189 is an ame pair it suffices to check the extension property for normal surfaces @xmath0 .
suppose @xmath5 does not extend to a regular map @xmath52 in a neighbourhood of @xmath43 .
assuming @xmath337 , we construct a codimension one subvariety @xmath59 through @xmath20 with the following property .
denote by @xmath338 the normalization of @xmath60 .
then @xmath339 does not extend to a regular map @xmath340 in a neighbourhood of some point @xmath341 .
if the local monodromy around @xmath20 with respect to @xmath5 is some virtually abelian group @xmath181 then the local monodromy with respect to @xmath339 around @xmath342 is a subgroup of @xmath181 and hence also virtually abelian .
iterating we arrive at a surface which violates the ame property of @xmath50 , thus proving the lemma .
consider the closure @xmath29 of the graph of @xmath108 in @xmath343 .
since @xmath51 is complete the projection @xmath344 is a proper , birational morphism . if the fibre @xmath345 is zero dimensional then , by zariski s main theorem ( ega iii corollary 4.4.9 ) , in a neighbourhood of @xmath20 , @xmath346 implying @xmath52 is regular ( contradiction ) .
let @xmath347 be a curve .
locally around some point in @xmath18 take an irreducible , codimension one normal subscheme containing a generic point of @xmath18 and passing through a point of @xmath29 not on the exceptional locus of @xmath344 .
this is possible if @xmath348 .
denote its closure in @xmath29 by @xmath349 and let @xmath350 .
then @xmath60 is an irreducible , codimension one subscheme of @xmath1 which passes through @xmath20 and intersects @xmath54 .
moreover , @xmath351 does not extend over @xmath20 since the closure of the graph of @xmath352 inside @xmath343 contains @xmath18 in the fibre over @xmath66 .
in fact , for the same reason @xmath353 is not regular in a neighbourhood of @xmath129 and hence in a neighbourhood of some @xmath341 as desired . instead of working with arbitrary varieties one can restrict to the category of projective varieties and every result in this section still holds . instead of a unique maximal ame compactification @xmath286 of @xmath3
there is a unique maximal projective ame compactification @xmath354 and a proper morphism @xmath355 .
find an example where @xmath356 .
denote by @xmath7 the dm moduli stack of stable , smooth , @xmath8-pointed , genus @xmath9 curves and by @xmath11 its deligne - mumford compactification ( as a dm stack ) .
denote by @xmath10 and @xmath12 the corresponding coarse spaces .
see @xcite and @xcite for an introduction to some of the theory involving these spaces .
in this section we prove : [ thm6 ] @xmath12 is the maximal ame compactification of @xmath7 . as before ,
let @xmath1 be a normal variety and @xmath0 an dense , open subset . by a family of stable curves @xmath357 we mean a flat family of stable @xmath8-pointed , genus @xmath9 curve .
we denote the marked points by @xmath358 where @xmath359 .
if the general fibre of @xmath357 is smooth then we get a morphism @xmath360 from some dense , open @xmath90 .
let @xmath59 be a connected , reduced , complete subscheme .
as in section [ s2.1 ] we restrict @xmath361 to a smaller open subset which is disjoint from @xmath60 and any singularities of @xmath1 . after blowing up we can assume the complement of @xmath361 is a normal crossing divisor .
then we define the _ local monodromy around @xmath60 _ as the image @xmath362 where @xmath69 is a small neighbourhood of @xmath60 as in section [ s2.1 ] . similarly , if @xmath1 is complete , the _ global monodromy _ is the image of @xmath363 .
this is precisely the definition from section [ s2.1 ] applied to the morphism @xmath364 when @xmath365 .
if the general fibre of @xmath357 is not smooth then we can find a dense , open @xmath90 over which @xmath366 is of constant topological type .
this gives a map @xmath367 into some boundary stratum .
we then define the _ local monodromy _ to be the image @xmath368 as above .
similarly , the _ global monodromy _ is the image of @xmath369 .
let us study a little more carefully this map @xmath370 .
we restrict our attention to the family @xmath371 .
after possibly having to pull back @xmath372 to a finite cover of @xmath361 , the normalization of @xmath372 is the disjoint union @xmath373 where each @xmath374 is a family of smooth , marked , stable curves . by lemma
[ lem0 ] below @xmath375 is finite , unramified and surjective .
so , after a finite tale base change @xmath376 , the singular points becomes sections ( see for instance ega iv , ( 18.4.7 ) ) .
hence @xmath372 is obtained from @xmath373 by glueing pairs of these `` special '' sections ( `` special '' in order to distinguish them from the sections induced by the marked points ) .
thus we get a map @xmath377 where @xmath378 is the genus of the fibres of @xmath379 , @xmath380 is the number of marked points of @xmath372 lying on @xmath379 and @xmath381 is the number of special sections on @xmath379 .
note that @xmath382 while @xmath383 .
now we could try to define the local monodromy around @xmath59 as @xmath384 where @xmath385 is a small neighbourhood of the preimage of @xmath60 in @xmath133 ( where @xmath133 is the normal closure of @xmath1 in the function field of @xmath386 ) .
there was a choice of cover @xmath386 so this image is only defined up to the equivalence @xmath387 where @xmath388 if they are contained in a common group @xmath181 as subgroups of finite index .
fortunately , this is sufficient for us since we get a well defined concept of the image being virtually abelian .
even better , modulo @xmath387 this image is equal to the local monodromy defined by ( [ eq1 ] ) .
[ lem0 ] let @xmath389 be a family of stable curves over ( an irreducible ) variety @xmath361 .
then @xmath390 has a natural closed subscheme structure which commutes with base change and is finite unramified over @xmath361 .
see @xcite lemma 4.3 or @xcite lemma 3 or @xcite . in what follows
we will often identify the mapping class group @xmath391 with the outer automorphism group of the fundamental group of a genus @xmath9 riemann surface with @xmath8 punctures .
it is well known that @xmath392 is generated by dehn twists .
more generally , we also consider @xmath393 which is the mapping class group of curves with a level @xmath394 structure .
see @xcite for a short introduction to vanishing cycles and dehn twists .
[ prop3 ] let @xmath1 be a normal variety and suppose @xmath395 is a ( semi)stable family of pointed curves .
let @xmath59 be a connected , integral , complete subscheme .
if the local monodromy around @xmath60 is virtually abelian then the global monodromy of @xmath396 is also virtually abelian .
since we only deal with virtually abelian monodromy it suffices to prove the result for the pullback of @xmath397 to any finite branched cover of @xmath1 . from the definition of monodromy above
we can assume the general fibre of @xmath398 is smooth ( even though the general fibre over @xmath60 may still be singular ) and that the monodromy around @xmath60 is abelian and lies in @xmath399 for some @xmath400 .
the local monodromy around @xmath60 is the image of @xmath401 where @xmath69 is a small neighbourhood of @xmath60 .
if the general fibre over @xmath60 is also smooth then we can extend @xmath54 to include an open dense subset of @xmath60 .
but then the global monodromy on @xmath60 is a subgroup of the local monodromy around @xmath60 and thus must be abelian ( and we are done ) .
if the general fibre over @xmath60 is singular consider an arc @xmath402 joining a general point of @xmath1 to a general point @xmath403 .
the restriction of @xmath397 to @xmath402 is a family of smooth , pointed curves degenerating to a semistable curve over @xmath404 .
consider the vanishing cycles @xmath405 of this one dimensional family and choose simple disjoint loops ( which we also call @xmath405 ) representing them .
thus these @xmath406 shrink to a point as we approach @xmath404
. then if we cut @xmath407 along these loops we are left with a family of ( possibly disconnected ) riemann surfaces with boundary except now all the fibres have the same topological type while the fibre over @xmath404 is a punctured riemann surface .
we can similarly do this in a small analytic neighbourhood of @xmath402 .
now we would like to do this globally .
the local monodromy around @xmath60 is abelian , lies in @xmath399 and contains the dehn twist in the multi - loop @xmath408 ( for some @xmath409 ) . by theorem [ thm2
] this means that the local monodromy around @xmath60 preserves the vanishing cycles ( we do not actually need the full strength of theorem [ thm2 ] ) .
this is what allows us , in a neighbourhood of @xmath60 , to consistently choose simple , disjoint loops representing @xmath405 .
if we now again cut the fibres of @xmath397 along these loops we are left with a family of ( possibly disconnected ) riemann surfaces with boundary but now the general fibre over @xmath1 is of the same topological type to a general fibre over @xmath60 which is a punctured riemann surface . since the monodromy around @xmath60 is abelian the monodromy over @xmath60 must also be abelian ( using the same argument as above ) .
[ prop8 ] a family of pointed , stable curves @xmath398 over a normal variety @xmath1 is isotrivial if the global monodromy is virtually abelian . since any two points in the base
can be connected by a series of irreducible curves it suffices to prove the result when @xmath1 is a curve .
if @xmath1 is not smooth or complete take its normalization and compactify . taking a branched cover of this new curve
the pullback of @xmath397 will extend to a family of stable curves so that we may assume from now on that @xmath1 is a smooth , complete curve . by lemma [ lem0 ] it suffices to prove the result when the general fibre of @xmath397 is smooth .
we first deal with the case when @xmath398 is unpointed .
denote by @xmath0 the largest open subset over which the fibres of @xmath397 are smooth .
the first cohomology group @xmath410 of the fibres of @xmath411 gives a variation of principally polarized hodge structures of weight one .
the induced monodromy map on this variation is the composition @xmath412 so that the monodromy of this variation is virtually abelian . by corollary 4.2.9 of @xcite , the connected component containing the identity of the zariski closure of the image of the map @xmath413 is semi - simple . since the image of @xmath414 is ( virtually )
so is its zariski closure in @xmath415 . since
a connected semisimple abelian group is trivial the monodromy must be finite and hence the variation of hodge structure is isotrivial .
consequently , by torelli s theorem , @xmath411 is isotrivial . to deal with the pointed case
consider the short exact sequence of groups @xmath416 obtained by forgetting the @xmath417st point of the curve @xmath18 ( if @xmath18 is a curve of genus @xmath9 then @xmath418 if @xmath419 and @xmath420 if @xmath421 ) .
we complete the proof by induction on @xmath8 .
the argument above proves the base case .
note that the base case @xmath422 is also covered by the argument above while the base case @xmath423 is obvious since any such family is trivial .
suppose @xmath398 is a stable family with @xmath424 marked points given by sections @xmath425 . since the global monodromy @xmath426 of @xmath397 is abelian the image of @xmath181 in @xmath392 is also abelian so by induction ( and after a finite base change if necessary ) @xmath397 is a trivial family @xmath427 with @xmath8 constant sections corresponding to points @xmath428 .
the @xmath417st section is given by a morphism @xmath429 .
the kernel of @xmath430 is the image of @xmath431 which is therefore also abelian . by lemma [ lem3 ] , if @xmath120 is dominant the image of @xmath432 has finite index in @xmath144 .
since there are no abelian subgroups of finite index inside @xmath144 the map @xmath120 must be constant .
this completes the proof .
it is possible to give an alternate proof of proposition [ prop8 ] by using theorem [ thm2 ] below in place of corollary 4.2.9 from @xcite .
the idea is that there is a nice classification of abelian subgroups of @xmath392 given by ivanov who builds on the paper @xcite . to explain this
let @xmath433 be a surface of genus @xmath9 with @xmath8 punctures .
a _ system of circles _
@xmath18 on @xmath433 is a collection of pairwise non - isotopic , non - intersecting circles on @xmath433 .
denote by @xmath434 the surface obtained by cutting @xmath433 along @xmath18 .
a subgroup of @xmath399 is a _
@xmath18-subgroup _ if it is generated by dehn twists around ( some of the ) circles in @xmath18 and by ( at most ) one pseudo - anosov map on each component of @xmath434 each of which fixes the boundary pointwise .
the following theorem appears in @xcite ( page 205 ) : [ thm2 ] a subgroup of @xmath399 ( @xmath400 ) is abelian if and only if it is a @xmath18-subgroup for some system of circles @xmath18 .
as usual we have a normal variety @xmath1 and a dense , open @xmath0 with a morphism @xmath41 .
suppose the local monodromy around @xmath436 is virtually abelian .
we can assume @xmath1 is a small neighbourhood of @xmath20 .
we want to show that we get a regular map @xmath42 .
now @xmath11 has a finite cover @xmath437 by a projective scheme @xmath232 ( see @xcite or @xcite ) .
if we let @xmath133 be the normalization of the closure of @xmath438 inside @xmath439 then we get a projective , generically finite morphism @xmath180 so that the composition @xmath440 is a regular map resolving the rational map @xmath441 .
this gives us a family of stable curves over @xmath133 .
let @xmath60 be an irreducible component of the fibre @xmath442 over @xmath436 .
the local monodromy around @xmath60 is virtually abelian so by proposition [ prop3 ] the global monodromy on @xmath60 is also virtually abelian .
hence , by proposition [ prop8 ] , the family over @xmath60 must be isotrivial .
let @xmath443 be the stein factorization of @xmath180 where @xmath444 has connected fibres and @xmath445 is finite .
let @xmath446 and consider the fibre @xmath447 . by the above
, the image of each irreducible component of @xmath447 is a point in @xmath10 and since @xmath447 is connected the whole fibre is collapsed to a point .
this means @xmath448 is regular in a neighbourhood of @xmath449 ( @xmath133 and subsequently @xmath450 are normal ) .
since @xmath1 is normal , lemma [ lem4 ] implies @xmath451 is also regular in a neighbourhood of @xmath20 .
this proves that @xmath12 is an ame compactification . to see that @xmath12 is maximal fix any @xmath452 and consider the rational map @xmath453 in a neighbourhood of @xmath20 .
the map @xmath454 tale locally around @xmath20 is of the form @xmath455 \rightarrow y / g$ ] where @xmath181 is a finite group and the boundary of @xmath186 ( i.e. the preimage of @xmath456 ) is normal crossing ( this is by definition what it means for @xmath11 to have normal crossing boundary ) .
this means that the local fundamental group around any point on @xmath186 is abelian so that we have a regular map @xmath457 .
thus we end up with the diagram @xmath458 \ar[drr ] & & \\
\tv \ar[r ] & v \ar@{-->}[r]^h & \sm_{g , n}^{ame } } \end{xy}\ ] ] where @xmath459 is a neighbourhood of @xmath20 , @xmath460 is a finite tale cover and @xmath461 is finite .
it follows by lemma [ lem4 ] that @xmath462 is regular in a neighbourhood of @xmath20 .
on the other hand , by corollary [ cor6 ] we also have the natural regular map @xmath463 , so it follows we have an isomorphism in a neighbourhood of @xmath20 . since @xmath20 was chosen arbitrarily we get @xmath464
. we can tweak theorem [ thm6 ] to give necessary and sufficient criteria for being able to extend families of stable curves
i.e. being able to extend a morphism @xmath41 to a regular map @xmath44 .
[ prop9 ] a family of smooth , pointed curves over @xmath0 extends to a family of stable , pointed curves over a neighbourhood of @xmath43 if and only if the local monodromy around @xmath20 is abelian and generated by dehn twists ( about multi - loops ) .
suppose the family extends over a neighbourhood of @xmath20 to a family of stable curves with the fibre over @xmath20 some pointed curve @xmath18 .
denote by @xmath69 the local versal deformation space of @xmath18 .
the locus in @xmath69 corresponding to singular curves is a divisor @xmath465 with simple normal crossing at @xmath466 \in v$ ] .
each component of @xmath465 corresponds to a node of @xmath18 so that the monodromy associated to it is a dehn twist about the vanishing cycle corresponding to the node .
this means that @xmath467 ( where @xmath468 is the number of nodes in @xmath18 ) is generated by dehn twists about disjoint loops .
in particular , every element of @xmath469 is a dehn twist about some multi - loop .
hence , the local monodromy around @xmath20 , which is a subgroup of @xmath469 , is abelian and generated by dehn twists .
conversely , if the local monodromy around @xmath20 is abelian and generated by dehn twists then by theorem [ thm6 ] @xmath41 extends to a regular map @xmath42 in a neighbourhood of @xmath20 .
we want to show that we can lift @xmath42 to @xmath44 . denote by @xmath18 the curve corresponding to the image of @xmath20 in @xmath12 and let @xmath246 be the local versal deformation space of @xmath18 .
the neighbourhood of @xmath466 \in \som_{g , n}$ ] is isomorphic to the quotient stack @xmath470 $ ] where @xmath471 is the automorphism group of @xmath18 .
after possibly restricting @xmath1 the problem comes down to lifting @xmath472 to a morphism @xmath473 .
if we denote by @xmath181 the fundamental group of the intersection of @xmath7 and a small neighbourhood of @xmath466 \in \som_{g , n}$ ] then we have the short exact sequence @xmath474 where @xmath468 is the number of nodes in @xmath18 . in order to be able to lift @xmath472 to @xmath246 , the monodromy around @xmath20 ( which is the image of @xmath475 where @xmath69 is a small neighbourhood of @xmath20 ) must lie in the kernel of @xmath476 .
but the kernel of @xmath476 is precisely the subgroup of @xmath181 generated by dehn twists around the @xmath468 vanishing cycles corresponding to the nodes of @xmath18 .
so if the monodromy is generated by dehn twists it lies in @xmath477 and we are done .
[ cor7 ] let @xmath1 be a normal variety and @xmath0 a dense open subvariety .
a morphism @xmath41 extends in a neighbourhood of @xmath43 to a regular map @xmath42 if and only if the local monodromy around @xmath20 is virtually abelian .
one direction is implied by theorem [ thm6 ] so we just need to consider what happens if @xmath41 extends around @xmath20 . in this case
there exists a finite cover @xmath478 such that @xmath42 lifts to a morphism @xmath440 .
this follows from the same argument as in the proof of theorem [ thm6 ] in section [ ss4 ] .
namely , @xmath11 has a finite flat cover @xmath437 by a projective scheme @xmath232 ( see @xcite ) and we take @xmath479 .
then , by proposition [ prop9 ] , the local monodromy around any point of @xmath480 is abelian so the local monodromy around @xmath20 must have been virtually abelian .
[ rem : jo ] as a consequence to corollary [ cor7 ] we get the following result first proved in @xcite : [ cor4 ] let @xmath40 be a normal crossing divisor at @xmath20 .
the morphism @xmath41 extends in a zariski neighbourhood of @xmath20 to a regular map @xmath44 if and only if it extends over the generic points of @xmath45 .
if @xmath40 is a normal crossing divisor at @xmath20 then in a small neighbourhood of @xmath20 the fundamental group of @xmath54 is free abelian .
thus @xmath41 extends to @xmath42 .
now look at @xmath454 and consider @xmath481 . locally in the tale topology this map looks like @xmath482 \rightarrow y / g = s$ ] for some scheme @xmath186 .
if @xmath186 is not normal then we replace it by its normalization @xmath483 . the action of @xmath181 on @xmath186 lifts to an action of @xmath181 on @xmath483 and we instead consider the map @xmath484 \rightarrow \ty / g$ ] .
notice that @xmath485 is a birational , quasi - finite map and @xmath1 is smooth so that we get @xmath486 .
we would like to show that @xmath120 is an isomorphism .
the locus where @xmath484 \rightarrow \ty / g = s$ ] is an isomorphism is where @xmath487 is tale .
now @xmath1 is smooth , @xmath483 is normal and @xmath9 is tale over an open subset of @xmath1 whose complement has codimension @xmath488 ( this is where we use that the morphism @xmath41 extends over the generic points of @xmath45 ) . by the purity of branch locus theorem ( see for example p.461 of @xcite ) this implies @xmath9 is tale everywhere .
hence @xmath120 is an isomorphism .
thus we get a regular map @xmath489 \rightarrow \som_{g , n}$ ] which lifts @xmath42 . for much the same reason
, we also get the extension theorem from @xcite which generalized @xcite to the case when the base is log regular ( i.e. schemes with boundaries like those of toric varieties ) . in a slightly different direction , see also @xcite for a discussion about extending to a log - smooth family instead of a stable family . in @xcite stix considers the analogous problem of extending @xmath22 to a morphism @xmath490 ( no compactifying ) .
similarly , in @xcite boggi considers extending @xmath22 to a morphism @xmath491 where @xmath492 is the partial compactification consisting of stable curves of compact type . in both of these cases
the condition for an extension to exist can be expressed in terms of monodromy .
denote by @xmath46 the moduli stack of @xmath9-dimensional principally polarized abelian varieties and by @xmath47 its baily - borel ( satake ) compactification . in this section
we prove : [ thm5 ] @xmath47 is the maximal projective ame compactification of @xmath46 . in the case of
@xmath493 we can show that @xmath494 is actually the maximal ame compactification ( not just the maximal projective ame compactification ) .
it is probably true more generally that @xmath495 but we can not prove this at the moment .
we follow the discussion from @xcite . a ( principally ) polarized hodge structure of weight one is a @xmath173 lattice @xmath496 of rank @xmath497 for some @xmath498 whose complexification @xmath499 is equipped with a decreasing hodge filtration @xmath500 such that @xmath501 ( @xmath502 is often denoted @xmath503 ) .
the ( principal ) polarization is given by a ( unimodular ) non - degenerate , skew - symmetric bilinear form @xmath504 satisfying @xmath505 and @xmath506 for any non - zero @xmath507 .
fixing @xmath125 and @xmath508 , the classifying space @xmath45 of polarized hodge structures is the open subset of @xmath509 subject to the condition @xmath510 for any non - zero @xmath507 ( here @xmath511 denotes the grassmannian of @xmath9-planes in @xmath125 )
. one can realize @xmath45 as the siegel upper half space @xmath512 as follows .
fix an integral symplectic basis @xmath513 of @xmath125 with respect to @xmath508 . because @xmath514 for any @xmath507 one can find a basis @xmath515 of @xmath502 such that @xmath516 where @xmath517 .
the condition @xmath505 implies @xmath518 .
@xmath232 is called the normalized period matrix .
denote by @xmath519 the open unit disk and by @xmath520 the punctured unit disk .
consider a variation of weight one polarized hodge structures over @xmath521 .
taking a point to its normalized period matrix gives a period map @xmath522 where @xmath523 denote by @xmath524 the universal cover of @xmath521 .
if we identify @xmath524 with the upper half plane then the covering map @xmath525 is given by @xmath526 .
let @xmath527 be a lift of @xmath476 to the universal cover of @xmath521 .
denote by @xmath528 and @xmath529 the normalized period matrices of @xmath530 and @xmath531 ( note that @xmath532 so one can think of @xmath533 and @xmath529 as multivalued functions ) .
denote by @xmath534 the image of @xmath535 under the map @xmath536 . by landman
s monodromy theorem @xcite , @xmath537 for some @xmath538 .
so after a finite base change we can assume @xmath539 .
let @xmath540 . since @xmath541 we have @xmath542 .
then the monodromy weight filtration @xmath543 induced by @xmath544 is @xmath545 where @xmath546 and @xmath547 .
an elementary computation shows @xmath548 so , since @xmath508 is non - degenerate , @xmath549 for all @xmath550 if and only if @xmath551 .
this means @xmath552 and so @xmath553 is isotropic .
choose the symplectic basis above such that @xmath554 where @xmath555 .
in this basis @xmath556 and @xmath557 where @xmath558 is an @xmath559 symmetric ( since @xmath560 ) matrix . define @xmath561 by @xmath562 . since @xmath563
we find @xmath564 so @xmath565 descends to a map @xmath566 .
denote by @xmath567 and @xmath568 the normalized period matrices of @xmath569 and @xmath570 . since @xmath571 we get @xmath572 or equivalently @xmath573 .
the nilpotent orbit theorem @xcite shows that @xmath566 has a removable singularity at zero which means @xmath568 is holomorphic on @xmath574 .
write @xmath575 where @xmath576 is a @xmath577 symmetric matrix .
[ lem8 ] let @xmath578 .
the weight one ( principally ) polarized hodge structure on @xmath125 induces a weight one ( principally ) polarized hodge structure on @xmath579 with normalized period matrix @xmath576 .
@xmath580 while @xmath554 which means @xmath581 , \dots , [ e_g ] , [ f_{r+1 } ] , \dots , [ f_g]\}.\ ] ] since @xmath582 we find @xmath583 and @xmath584 . using the fact @xmath517 it is then elementary to see @xmath585 \oplus \overline{[f^1 \cap w_1]}\ ] ] so that @xmath586 defines a weight one hodge filtration on @xmath579 .
since @xmath553 is isotropic @xmath508 descends to give a ( principal ) polarization on @xmath587 . fixing the symplectic basis @xmath588 , \dots , [ e_g ] , [ f_{r+1 } ] , \dots , [ f_g]$ ] for @xmath579 , @xmath589 = \span \ { [ v_j ] = [ f_j ] + \sum_{k > r } z_{kj}(t ) [ e_j ] : j = r+1 , \dots , g\}.\ ] ] so , if we write @xmath590 where @xmath591 is a @xmath577 symmetric matrix , the normalized period matrix for @xmath579 is @xmath592 . but
@xmath591 extends to a holomorphic map @xmath593 because @xmath594 . a priori @xmath595 where @xmath596 .
however , for every non - zero real @xmath597 column matrix @xmath57 we have that @xmath598 lies in @xmath599 if @xmath600 and in @xmath601 if @xmath602 .
so by the open mapping theorem @xmath603 for any @xmath600 .
since this is true for any @xmath57 as above we get @xmath604 . by lemma [ lem7 ] , to show @xmath606 has the ame property it suffices to consider normal surfaces @xmath0 .
suppose @xmath1 is a small neighbourhood of @xmath607 such that the local monodromy around @xmath20 with respect to a given morphism @xmath608 is virtually abelian .
we need to show @xmath609 extends to a regular map @xmath610 .
lemma [ lem4 ] allows us to replace @xmath54 be any finite cover so that we can assume the local monodromy around @xmath20 is abelian .
since @xmath1 is normal it has isolated singularities so , after restricting @xmath1 to a smaller neighbourhood if necessary , one can assume @xmath1 is smooth with @xmath45 a normal crossing divisor except possibly at @xmath20 .
let @xmath61 be a log - resolution of the pair @xmath611 meaning @xmath134 is a proper , birational morphism such that @xmath29 is smooth and @xmath612 is a normal crossing divisor . by blowing up further if necessary
one can assume @xmath613 extends to a regular map @xmath614 .
let @xmath18 be a component in the exceptional divisor above @xmath20 and denote by @xmath615 the intersection points of @xmath18 with all the other divisors in @xmath616 . denote by @xmath617 the intersection of @xmath361 with a small open neighbourhood of @xmath18 .
notice the image of the monodromy map @xmath618 is abelian .
we will use this fact to show that @xmath619 collapses @xmath18 to a point .
this will mean that @xmath619 collapses the whole exceptional divisor of @xmath134 to a point and so @xmath610 must be regular .
the first cohomology group of the family of principally polarized abelian varieties corresponding to @xmath620 gives a variation of principally polarized hodge structures of weight one , which we denote @xmath125 , on @xmath617 .
let @xmath232 be the normalized period matrix of @xmath125 . since @xmath616 is normal crossing the kernel of the map @xmath621 where @xmath622 is isomorphic to @xmath173 and generated by a small loop around @xmath18 .
denote by @xmath60 the image of this generator under the monodromy map @xmath623 and let @xmath624 .
notice that @xmath60 is only defined up to conjugacy since we can move the base point of @xmath617 .
however , because the image of the monodromy map is abelian we get a well defined action @xmath625 .
let @xmath626 and @xmath627 .
after taking an appropriate cover branched along @xmath18 , landman s monodromy theorem says @xmath541 so we have a filtration @xmath628 and a quotient @xmath629 . locally around any point @xmath630 of @xmath18
choose coordinates @xmath631 such that @xmath18 is the locus @xmath632 .
by lemma [ lem8 ] , the normalized period matrix of @xmath125 is @xmath633 where @xmath634 is holomorphic .
then @xmath125 induces a weight one principally polarized hodge structure on @xmath629 with normalized period matrix @xmath635 where @xmath555 . notice that @xmath636 together with its hodge structure is well defined over @xmath637 so we get a morphism @xmath638 . since the monodromy of @xmath125 over @xmath617 is abelian the induced monodromy on @xmath636 over @xmath617 is also abelian .
thus the image of @xmath639 is abelian .
restricting to @xmath640 we get a weight one principally polarized hodge structure @xmath636 whose associated period map @xmath641 has abelian image .
but by corollary 4.2.9 of @xcite , the connected component containing the identity of the zariski closure of this image is semi - simple . arguing as in the proof of [ prop8 ] shows that the monodromy on @xmath642 is finite
hence , the morphism @xmath643 collapses @xmath18 to a point .
but , by the construction of @xmath47 , the morphism @xmath614 maps @xmath642 to the boundary component @xmath644 via @xmath645 .
so @xmath619 collapses @xmath18 to a point .
it remains to show that @xmath47 is maximal among all projective ame compactifications of @xmath46 .
we make use of @xmath646 , the perfect ( 1st voronoi ) toroidal compactification of @xmath176 ( see @xcite ) . by lemma [ lem9 ] below
we have a sequence of regular maps @xmath647 .
the case @xmath421 is trivial since @xmath648 .
we will first deal with the case @xmath649 and then use it to prove @xmath47 is maximal for arbitrary @xmath9 .
recall that @xmath650 while @xmath651 is isomorphic to @xmath652 via the torelli map .
the restriction of @xmath653 over @xmath654 is the map @xmath655 \cong \om_{1,2 } \rightarrow a_1^{bb}$ ] given by forgetting the second marked point .
let @xmath465 be any elliptic curve .
we will now show that the map @xmath656 must collapse the fibre @xmath657)$ ] to a point . to do this
we construct a rational map from a surface to @xmath493 with abelian local mondromy around a point @xmath658 whose total transform inside @xmath652 contains the fibre @xmath657)$ ] .
the key is to find a morphism from a surface to @xmath652 whose image contains @xmath657)$ ] but where the local monodromy around the preimage of @xmath657)$ ] is abelian ( then we blow down the preimage to get @xmath658 ) .
we proceed as follows .
fix a point @xmath659 and denote by @xmath660 the projection onto the first factor .
let @xmath661 where @xmath662 and @xmath574 is the diagonal .
alternatively , @xmath69 is a rank two vector bundle defined as the unique non - trivial extension @xmath663 .
let @xmath1 be the surface @xmath664 and @xmath665 the natural projection .
the line bundle @xmath666 defines a @xmath667 map @xmath668 mapping fibres to fibres .
denote by @xmath669 the branching locus of this map and by @xmath670 the image of @xmath574 ( or equivalently of @xmath671 ) .
notice that @xmath672 is an unramified , degree 4 map while @xmath18 is a section of @xmath134 .
also , @xmath673 and @xmath18 meet only at ( the missing ) point @xmath674 which is the image of @xmath675 .
finally , denote by @xmath676 the fibre @xmath677 . a helpful picture to keep in mind is figure [ f1f ] .
let @xmath678 . to a point @xmath679
we can associate the genus two curve obtained as the double cover of @xmath680 branched over @xmath681 and the five points @xmath682 ( by construction these six points are distinct ) . to do this globally over @xmath54 we consider the @xmath683 bundle @xmath684 and take its double cover branched over the divisor corresponding to @xmath682 . to see why this is possible notice that we need to find a line bundle @xmath685 on @xmath686 such that @xmath687 is isomorphic to the branching divisor .
now @xmath688 and , under this isomorphism , the branching divisor has class @xmath689 for some @xmath690 .
thus to find such a line bundle @xmath685 we just need to find a square root of @xmath691 . now the pullback map @xmath692 is surjective since the fibres of @xmath693 are all isomorphic to @xmath694 . also , @xmath695 is a divisible group .
thus we can find a square root of @xmath691 .
so now we have a family of hyperelliptic genus two curves @xmath696 .
the complement @xmath697 is normal crossing except at @xmath698 .
consider the log - resolution @xmath29 and denote by @xmath616 the preimage of @xmath45 which is now normal crossing ( it s necessary to blow up at least twice since @xmath673 and @xmath18 are tangent at @xmath698 as one can check @xmath699 using the push - pull formula ) .
if we denote by @xmath361 the preimage of @xmath54 then the induced morphism @xmath700 extends to a regular map @xmath701 ( this can be ensured either by theorem [ thm6 ] or by blowing up further ) . what is the image of @xmath702 ( the proper transform of @xmath670 ) in @xmath652 ? if you approach a general point @xmath703 along a general arc the associated genus two curves are double covers of @xmath683 branched along six points , two of which are converging .
the limiting admissible cover is illustrated in figure [ f1e](a ) .
the points @xmath248 are independent of @xmath704 ( they are actually the branching points of @xmath705 ) and only @xmath706 varies as you move @xmath703 .
this shows that @xmath707 maps onto the fibre of @xmath708 over @xmath709 \in a_1 \subset a_2^{bb}$ ] .
now @xmath707 intersects the exceptional divisor at one point which we call @xmath710 .
the image of @xmath710 in @xmath652 corresponds to the curve shown in figure [ f1e](b ) .
the local monodromy around @xmath710 induced by the morphism @xmath700 is generated by the two dehn twists about loops @xmath711 and @xmath712 in figure [ f1e](c ) .
however , as a map @xmath713
the dehn twist about @xmath711 induces the identity and so the local monodromy is generated only by @xmath712 .
consequently , if we denote by @xmath246 a small neighbourhood of @xmath702 then the map @xmath714 factors through @xmath715 . since @xmath246 is smooth and @xmath716 is a genus one curve , @xmath717 is homotopic to an @xmath718 bundle over @xmath719 so that @xmath720 .
hence the local monodromy around @xmath707 induced by @xmath713 is abelian .
this is the family we ve been looking for . by blowing up sufficiently many points on @xmath707
we can assume its proper transform @xmath721 has an ample conormal bundle .
then we can blow down @xmath722 to a point @xmath723 .
the local monodromy around @xmath722 is a subgroup of the local monodromy around @xmath707 and thus is abelian .
consequently , the local monodromy around @xmath658 is abelian and we get a regular map @xmath724 .
we end up with a commutative diagram @xmath725 which implies that @xmath726 must collapse to a point the image of @xmath722 inside @xmath652 .
this image is the same as that of @xmath707 which is the fibre of @xmath708 over @xmath709 \in a_1 \subset a_2^{bb}$ ] . finally , consider the restrictions of @xmath656 and @xmath727 to the preimage of @xmath728 .
we get @xmath729 where @xmath656 collapses to a point the fibre @xmath730)$ ] .
since @xmath731 is irreducible and @xmath732 is a normal curve the morphism @xmath733 is flat hence , by lemma [ lem10 ] , @xmath734 is an isomorphism over @xmath728 ( we use that @xmath494 is normal , projective ) .
this shows @xmath735 ( in fact it shows that @xmath736 ) .
we will now use that @xmath735 together with the fact that @xmath737 is projective to conclude that @xmath738 for @xmath739 .
recall that we have @xmath740 while the morphism @xmath741 restricted to the preimage of @xmath742 is isomorphic ( after taking a @xmath743 quotient ) to the universal family @xmath744 .
the preimage @xmath745 is a partial compactification of @xmath176 present in all toroidal compactifications and commonly denoted @xmath746 .
fix a principally polarized abelian variety @xmath747 of dimension @xmath748 and consider the morphism @xmath749 given by @xmath750 \mapsto [ a ' \times a]$ ] .
this map extends to a regular map @xmath751 whose restriction to @xmath752 is given by @xmath753 \mapsto [ a ' \times e , ( a , p)]$ ] where @xmath754 is the identity .
consider now the commutative diagram @xmath755 since the morphism @xmath756 contracts the fibres of @xmath757 it follows that the image of such a fibre in @xmath646 must be contracted to a point by the map @xmath758 .
so in the fibre @xmath759 over @xmath760 \in a_{g-1}$ ] the locus contracted by @xmath656 includes @xmath761 : p \in e\}$ ] .
this is a fibre of the projection @xmath762 so that by lemma [ lem10 ] the whole fibre @xmath759 is collapsed to @xmath747 .
this means that the dimension of the fibre of @xmath763 over @xmath760 \in a_{g-1}$ ] is at most @xmath748 .
thus @xmath764 , which has dimension @xmath765 , is collapsed by @xmath727 to a variety of dimension at most @xmath766 .
in particular , the divisor @xmath767 is collapsed by @xmath727 to a subvariety of lower dimension .
on the other hand , if @xmath768 then @xmath769 where @xmath685 is the determinant of the hodge bundle ( @xcite p.397 ) .
this means that @xmath770 .
since @xmath685 is pulled back from @xmath47 by @xmath727 this implies @xmath738 because otherwise @xmath737 would fail to be projective .
[ lem9 ] let @xmath189 be an ame pair and @xmath771 a toroidal compactification of @xmath57 .
then there exists a regular map @xmath772 extending the identity map on @xmath57 .
locally , @xmath773 is isomorphic to @xmath774 where @xmath775 is normal affine ( and translations by @xmath776 extend to @xmath775 ) .
this means that the local fundamental group around any point @xmath777 is abelian and hence its image in @xmath81 is abelian .
since @xmath51 is an ame compactification the morphism @xmath772 must be regular .
[ lem10 ] consider morphisms @xmath778 between complete , normal varieties
such that the composition @xmath779 is flat with connected fibres and @xmath120 is surjective .
if @xmath120 collapses a single fibre @xmath780 to a point then @xmath9 is an isomorphism .
since @xmath120 collapses @xmath780 to a point the fibre @xmath449 is a point ( here we use that @xmath120 is surjective ) .
thus the relative dimension of @xmath9 is zero .
this means @xmath781 .
if any fibre @xmath782 is positive dimensional then the dimension of @xmath783 is greater than @xmath784 .
this contradicts the fact that @xmath779 is flat .
since @xmath779 has connected fibres @xmath9 must also have connected fibres . also , since @xmath779 is flat @xmath9 must be surjective .
so @xmath9 is surjective with connected fibres of dimension @xmath35 .
this means @xmath9 is quasi - finite and bijective . since @xmath9 is also proper it is in fact finite .
also @xmath9 is bijective so it is birational .
thus @xmath9 is a finite birational map and @xmath57 is normal so @xmath9 is an isomorphism .
what is @xmath785 ?
it seems very likely that @xmath786 . to show this one could try to extend the argument above by studying more carefully the morphism @xmath741 deeper into the boundary of @xmath47 .
there is no reason to consider only principally polarized abelian varieties . more generally , the baily - borel compactification of an arithmetic quotient of any bounded symmetric domain should be its maximal ( projective ) ame compactification .
the fact that the baily - borel compactification is ame follows in much the same way as in the proof of [ thm5 ] except that you need to deal with higher weight hodge structures .
that it is maximal follows from the fact that the baily - borel compactification is the log canonical model ( a result we hope to explain in future work ) .
our definition of the ame property involves the class @xmath787 of abelian groups
. why not consider another class @xmath788 of finitely generated groups ( say solvable groups ) and look for varieties which have the @xmath788 monodromy extension property ?
this is actually a reasonable thing to try to do .
one could , for instance , try to find pairs @xmath116 having the sme ( the solvable monodromy extension ) property .
but this is another story .
the story in this paper is that there are natural moduli spaces which satisfy the ame property . in the case of
@xmath7 we even get an `` if and only if '' statement ( corollary [ cor7 ] ) . if one tried to find ( say ) @xmath789 there are two possibilities .
either it does not exist or there exists a proper surjective morphism @xmath790 . in the latter case if @xmath691 is an isomorphism then we get nothing new so one might as well stay with the ame property since it is easier to check than the sme property
. if @xmath691 is not an isomorphism then it would indeed be interesting to identify explicitly @xmath789 .
however , as many would agree , @xmath12 is likely the most natural compactification of @xmath10 so it would still be favourable to consider @xmath791 instead of @xmath789 since the former is probably more natural . in order to show that @xmath116 has the ame property one needs to check that any morphism @xmath58 extends in a neighbourhood of a point @xmath20 whenever the local mondromy around @xmath20 is virtually abelian
this is easier to do if you only need to check pairs of varieties @xmath0 rather than pairs of more arbitrary schemes .
furthermore , as this paper shows it suffices to consider only pairs of varieties @xmath0 in order to get natural maximal ame compactifications such as @xmath464 .
even if we keep the current definition of an ame compactification it may still be interesting to ask what happens if @xmath0 is a pair of more general schemes .
one is tempted to consider either @xmath1 an integral , separated scheme which is not necessarily of finite type or a formal scheme .
the main examples of such schemes would be the localization or the completion of a variety along a subvariety .
the primary example of the latter is the algebraic disk @xmath792)$ ] which is the completion of @xmath793)$ ] along the ideal @xmath794 .
let us sketch a possible answer to these questions . in the finite type case ,
another way of describing the monodromy around @xmath20 is to look at the closure of the image of @xmath795 and to take a small neighbourhood @xmath69 of the proper transform of @xmath20 .
then the local monodromy around @xmath20 is the image of @xmath796 .
the map @xmath797 extends if and only if the proper transform of @xmath20 is a point and the proofs in this paper show that if the proper transform is not a point then the monodromy around it is not virtually abelian .
so , as before , the map extends in a neighbourhood of @xmath20 if the monodromy is virtually abelian .
now , if @xmath1 is an integral , separated scheme but not necessarily of finite type then a variant of this definition should still work assuming that @xmath51 is of finite type .
for example , suppose @xmath1 is the localization of @xmath798 along @xmath799 where @xmath240 are varieties
. then a morphism @xmath58 corresponds to a morphism @xmath800 from some open subset @xmath801 .
so the local monodromy around some @xmath66 is defined as the image of @xmath802 where @xmath69 is a small neighbourhood of the proper transform of @xmath20 .
then the same argument as above shows that @xmath120 extends to a regular map in a neighbourhood of @xmath20 if the monodromy is virtually abelian .
a little more subtle is the case when @xmath1 is a formal scheme which is the completion of @xmath798 along @xmath799 . given a map @xmath797 one should be able to find a subvariety @xmath207 ( at least locally ) such that the image of @xmath54 is the completion of @xmath186 along @xmath803 .
then we would define the local monodromy around a point @xmath66 as the image of @xmath804 where @xmath69 is a local neighbourhood of the proper transform of @xmath20 .
there is a counter example to theorem [ thm : b ] in positive characteristic given in @xcite , page 192 .
since the tame algebraic fundamental group of the complement of a normal crossing divisor is abelian ( @xcite ) , this means that theorem [ specialthm2 ] can not hold in positive characteristic if we change to using the tame algebraic fundamental group and consider formal neighbourhoods instead of analytic ones .
tame stacks in positive characteristic ; arxiv:0703310 .
orbifolds and stringy topology , _ cambridge university press _ ( 2007 ) , xii+149pp .
on the purity of the branch locus , _ compositio mathematica _ , vol . * 23
* , fasc . 4 ( 1971 ) , 461465 .
smooth compactification of locally symmetric varieties , _ lie groups : history , frontiers and applications , vol .
press , brookline , mass , ( 1975 ) , iv+335pp . some metric properties of arithmetic quotients of symmetric space and an extension theorem , _ j. diff .
geometry _ * 6 * ( 1972 ) , 543560 .
abelian and solvable subgroups of the mapping class groups , _ duke math .
* 50 * ( 1983 ) , no .
4 , 11071120 . monodromy of stable curves of compact type : rigidity and extension , _ int . math .
res . not .
_ ( 2007 ) , no . 6 , article i d rnm017 , 16 pp .
mixed hodge structures , compactifications and monodromy weight filtration , _ topics in transcendental algebraic geometry _
( princeton , n.j . , 1981/1982 ) , ann . of math .
, 106 , princeton univ . press , princeton , nj , ( 1984 ) 75100 .
, degeneration of abelian varieties , ergebnisse der mathematik und ihrer grenzgebiete 3 .
folge , band 22 , springer verlag ( 1990 ) , xii+316pp .
thorie de hodge ii , _ inst .
hautes tudes sci .
( 1971 ) , 557 .
the irreducibility of the space of curves of given genus , _ publications mathmatiques de li.h..s .
_ tome * 36 * ( 1969 ) , 75109 .
polynomials , vanishing cycles and floer homology , _ mathematics : frontiers and perspectives _ amer . math .
soc . , providence , ri ( 2000 ) , 5564 .
intersection theory of toroidal compactifications of @xmath806 , _ bull .
london math .
_ , * 38 * ( 2006 ) , no .
3 , 396400 .
the tame fundamental group of a formal neighbourhood of a divisor with normal crossings on a scheme , lnm vol . * 208 * , springer ( 1971 ) .
expos ix ; sminaire de gomtrie algbrique 7 .
moduli of curves , _ graduate texts in mathematics _ * 187 * , springer - verlag , new york ( 1998 ) .
fundamental groups of rationally connected varieties , _ michigan math . j. _ , * 48 * ( 2000 ) , 359368 ; arxiv : math.ag/0203174 . quotients by groupoids , _ annals of math .
_ , * 145 * ( 1997 ) , 193213 . on coverings of deligne - mumford stacks and surjectivity of the brauer map , _ bull .
london math .
_ , * 36 * ( 2004 ) , no .
2 , 188192 .
automorphisms of teichmller modular groups , _ lectures notes in math .
1346 springer , berlin ( 1988 ) , 199270 . on extending families of curves ,
_ j. algebraic geom .
_ , * 6 * ( 1997 ) , no .
3 , 545562 . on the picard - lefschetz transformation for algebraic manifolds acquiring general singularities ,
* 181 * ( 1973 ) , 89 - 126 .
smooth deligne - mumford compactifications by means of prym level structures , _ j. algebraic geom .
_ , * 3 * ( 1994 ) , no .
2 , 283293 . un thorme de puret pour les familles de courbes lisses , _ c. r. acad .
paris sr .
_ * 300 * ( 1985 ) , no .
14 , 489492 . extending families of curves over log regular schemes ,
_ j. reine angew .
* 511 * ( 1999 ) 4371
. foundations of topological stacks i , .
log smooth extension of a family of curves and semi - stable reduction , _
j. algebraic geom . _ * 13 * ( 2004 ) , no .
2 , 287321 . abbildungen in arithmetische quotienten hermitesch symmetrischer rume ( german ) , _ lecture notes in math .
412 , springer , berlin ( 1974 ) , 243258 . a monodromy criterion for extending curves , _
res . not .
_ ( 2005 ) , no .
29 , 17871802 .
the moduli space of curves and its tautological ring , _ notices amer .
_ * 50 * ( 2003 ) , no . 6 , 647658 . | necessary and sufficient conditions are given ( in terms of monodromy ) for extending a family of smooth curves over an open subset @xmath0 to a family of stable curves over @xmath1 .
more precisely , we introduce the abelian monodromy extension ( ame ) property and show that the standard deligne - mumford compactification is the unique , maximal ame compactification of the moduli space of curves .
we also show that the baily - borel compactification is the unique , maximal projective ame compactification of the moduli space of abelian varieties . |
since the advent of the theory of superconductivity , developed by bardeen , cooper and schrieffer [ bcs ] @xcite and their contemporaries , bogoliubov @xcite and valatin @xcite , countless investigations with these formalisms have been performed in a variety of fields .
this widespread application of the bcs approach is due not only to its power and utility but also because superior approximate or exact solutions are generally not available .
this is especially true in nuclear @xcite and hadronic structure @xcite where physical insight has been limited by uncertainties due to approximations .
hence , further assessing the validity of the bcs approximation is an important endeavor with broad attending interest .
in particular , there have been several studies @xcite of the bcs approximation within the hubbard model for which exact solutions , such as those provided by the bethe ansatz @xcite , are known but only in one dimension .
the purpose of this paper is to extend these investigations in several new directions : 1 ) assess the bcs approximation for not only the hubbard but also a reduced hubbard - like hamiltonian ; 2 ) in addition to energies and energy gaps also examine propagators and correlation observables ; 3 ) evaluate the bcs accuracy in two dimensions . in discussing the hubbard model
it is natural to restrict ourselves to a fixed filling because the hamiltonian conserves electron number .
therefore we utilize the canonical bcs formulation with a fixed number of electrons .
this is superior to the grand canonical approach which only constrains the average electron number and also produces a higher ground state energy . a previous pioneering study @xcite notably detailed the excellent ground state energy agreement between the exact hubbard model and the approximate canonical bcs solution .
however that work also determined , depending upon electron density , that the hubbard and bcs energy gaps do deviate , again in one dimension .
related , an odd filling bcs application @xcite in one dimension also found somewhat less precise results . because the bcs approximation is conjectured to improve in higher dimension @xcite , it is important to perform a numerical assessment in two dimensions . when computing with the bcs ground state many of the terms in the hubbard hamiltonian , @xmath0 , become irrelevant since only components which are either diagonal in momentum space or create and destroy momentum modes in pairs contribute .
truncation to these remnant terms defines the reduced hamiltonian , @xmath1 , which is specified in the next section where the bcs hubbard ground state energy , @xmath2 , is found degenerate with the bcs reduced hamiltonian value . because @xmath0 contains further terms describing additional dynamics we a priori expect @xmath2 to better approximate the exact reduced hamiltonian ground state energy , @xmath3 , than the exact hubbard ground state energy , @xmath4 .
indeed , as we detail below @xmath2 is an order of magnitude closer to @xmath5 than @xmath4 even though the variational treatment is very good for both .
consequently a more complete bcs assessment , especially for systems modeled by @xmath1 , should address other observables which , in contrast to energies , are more sensitive to wavefunction details .
accordingly , we also calculate electron pair propagators , wavefunction inner products and @xmath6 , @xmath7 and @xmath8-wave correlators which are important for establishing superconductivity .
this paper is divided into four sections . in section
ii the hubbard and reduced hamiltonians are introduced and a new formulation of the canonical bcs method is derived .
this section also details formulas for energy gaps , densities , propagators and @xmath6 , @xmath7 and @xmath8-wave correlators .
section iii compares and discusses numerical results for a variety of electron level fillings and different number of lattice sites .
finally , conclusions are summarized in section iv .
we follow the literature notation , especially refs . and , and recall the full hubbard hamiltonian in momentum space @xmath9 the reduced hamiltonian retains the non - diagonal ( @xmath10 ) terms with @xmath11 and diagonal ( @xmath12 ) terms @xmath13 in the above equations @xmath14 creates ( annihilates ) an electron with momentum @xmath15 and spin @xmath16 @xmath17 or @xmath18 . in one dimension
the electron kinetic energy is @xmath19 , while in two dimensions , @xmath20 , with hopping rate @xmath21 ( set to @xmath22 in this work ) between nearest neighbors .
the momenta are restricted to multiples of @xmath23 , where @xmath24 is the linear extent of the system .
the number of sites @xmath25 is @xmath24 in one dimension or @xmath26 in two dimensions .
note that the last term in eq .
( [ hr ] ) can be expressed in terms of a sum over the dummy index @xmath27 yielding @xmath28 , the product of spin up , @xmath29 , and down , @xmath30 , electron number operators for a system having @xmath25 total lattice sites .
since we restrict this discussion to the attractive hubbard model the potential coupling parameter @xmath31 is negative .
the reduced hamiltonian , @xmath1 , has been chosen to only contain elements of @xmath0 which have non - zero expectation values in the bcs state defined in the next section .
we adopt the formulation of ref . but present an alternate bcs derivation . for a given hamiltonian , the bcs approximation entails using the unnormalized variational wavefunction @xmath32 to minimize the ground state energy .
here @xmath33 are the variational parameters and @xmath34 represents the vacuum state with no electrons .
note that the above hamiltonians conserve particle number , but the state @xmath35 includes a mixture of fock components with different particle number . because we are interested in states with fixed fillings we organize these components by electron number @xmath36 with @xmath37 @xmath38 @xmath39 here and below the cryptic summation notation @xmath40
is used to represent the sum over all unordered sets of momentum coordinates , @xmath41 , with the restriction that all of the momenta are different . for a physical system with @xmath42 electrons the bcs ground state energy , @xmath2 ,
is calculated by minimizing the expectation value @xmath43 @xmath44 i.e. @xmath45 where the minimization is achieved by varying the @xmath33 as detailed below . before proceeding several points
are in order .
first , in the grand canonical bcs formulation the ground state energy is obtained by minimizing @xmath46 . using eq .
( [ bcs ] ) one can show that the grand canonical is greater than the minimum canonical energy .
second , combining eqs .
( [ h],[hr],[bcsn ] ) directly yields the degenerate result @xmath47 verifying , as stated above , that the bcs ground state energy is the same for both hamiltonians .
this is also true for the grand canonical bcs energy .
third , the successive operation of @xmath1 on @xmath34 only generates states having electrons paired just like the state @xmath48 with opposite momentum and spin , called cooper pairs . in general , the exact hubbard ground state will contain additional electron correlations between such pairs and the bcs approach assumes these correlations can be ignored .
now , we compute the norm of the fock state @xmath49 and the unnormalized kinetic energy @xmath50 the modified summation index @xmath51 now denotes the sum over all unordered sets of momentum coordinates , @xmath52 , with the stronger restriction that all momenta in the extended set , @xmath53 , are different . for the potential energy
we first calculate @xmath54 then the unnormalized potential energy matrix element is @xmath55 note well that the sum over the set @xmath56 ensures @xmath57 . now define @xmath58 and @xmath59 which both vanish if any two momenta are equal , to express @xmath60 as @xmath61 next , we evaluate the bcs variational equation for a state with @xmath36 electrons @xmath62 by first calculating @xmath63 then we use the product rule for differentiating , @xmath64 to obtain @xmath65 solving for @xmath66 yields @xmath67 with the denominator , @xmath68 equation ( [ gp ] ) is equivalent to eq . ( 34 ) of ref .
( note their equation has a misprint involving a missing @xmath66 ) . in the case of odd filling , @xmath69
, the variational state has a single specific unpaired , say up , electron with momentum @xmath70 , chosen to minimize the energy : @xmath71 operating with the hamiltonian on this state shifts the energy by @xmath72 and repeating the above analysis yields @xmath73 where @xmath74 is equivalent to @xmath75 defined on a space with the @xmath70 mode removed .
finally , the odd electron system variational equation reduces to @xmath76 where @xmath77 using parpack , a parallel implementation of the lanczos algorithm , we exactly calculated the hubbard and reduced hamiltonian ground states as a function of the electron density .
the dimension of the hilbert space for a 16 site hubbard model is @xmath78 .
since the hamiltonians conserve momentum and particle number we may restrict to subspaces with fixed momentum and filling .
however , the dimension for half - filling at @xmath79 momentum is still greater than @xmath80 .
there are about 2000 non - zero terms for each state , so the hubbard hamiltonian matrix can not be stored in memory even in a sparse format .
instead , matrix elements were generated at each step in the lanczos algorithm .
the diagonalization took about 3 hours using 30 1.2 ghz athlon processors in parallel .
for even @xmath36 , we restrict to the zero momentum subspace which yields the minimum energy eigenvector . for odd @xmath42 , the momentum with lowest ground state energy is dependent on the coupling @xmath31 . for @xmath81 it is obvious that the lowest energy will have the momentum of the @xmath82th lowest energy mode . for high @xmath31 , with less than half filling
, the lowest energy will have @xmath79 momentum , but for the value @xmath83 used throughout this paper , the momentum is the same as for @xmath81 .
the total momentum of the ground state for odd @xmath42 corresponds to occupation of a specific unpaired momentum mode . for the hubbard model , occupation of this mode
is not fixed by the action of the hamiltonian but , for the reduced model , all relevant basis states have the same unpaired mode . in this regard
, the reduced model ground state is the same as the bcs state which has the single unpaired mode @xmath70 ( eq . [ eq : bcsodd ] ) . to properly assess any approximation scheme , as well as investigate the robustness of a particular dynamic model , it is important to calculate and compare a variety of observables and theoretical constructs
in addition to the ground state energy , @xmath4 , and energy density , @xmath84 , another important observable is the single particle energy gap , @xmath85 , @xmath86 \nonumber \\ & = & \frac{1 } { 2 } [ e_{n_e + 1 } - 2e_{n_e } + e_{n_e - 1 } ] \ .\end{aligned}\ ] ] in the grand canonical formulation , a quasi - particle excitation is produced by the action of @xmath87 , or equivalently , @xmath88 on the state @xmath48 @xcite with excitation energy @xmath89 here
@xmath90 is the bcs gap and @xmath91 is dependent on the momentum @xmath15 .
the minimum excitation is non - zero since @xmath92 , thus accounting for the name `` gap '' . in the canonical bcs formulation @xmath93
is no longer proportional to @xmath94 .
therefore , @xmath85 averages the effects of adding and subtracting one electron .
because wavefunctions are more sensitive than energies to approximation schemes , we also address other operator wavefunction overlaps such as propagators and correlators @xcite . before introducing theses constructs
we first specify the generalized one , @xmath95 , and two , @xmath96 , particle density matrices for any model state @xmath97 with unit norm @xmath98 @xmath99 here the indices @xmath100 represent possible electron lattice sites .
the position operators , @xmath101 and @xmath102 , are related to the momentum operators by @xmath103 @xmath104 between states with the same filling @xmath105 , while @xmath106 and @xmath107 are the one electron propagators for spin up or down , respectively .
next we construct a variety of correlators as the difference between a specific two particle density and the product of one particle densities .
for example , the on site @xmath6-wave correlator , @xmath108 , is @xmath109 which is the correlated electron pair propagator minus the product of two ( independent ) one electron propagators .
the extended @xmath6-wave correlator , @xmath110 , is @xmath111 which involves a sum , @xmath112 , @xmath113 , over the four planar sites adjacent to lattice site @xmath114 ( @xmath115 ) with weights @xmath116 given by @xmath117 for @xmath118 or @xmath119 and @xmath120 otherwise .
we also list in table i several other correlators , @xmath121 , with different symmetry weights @xmath122 which are important for investigating superconductivity .
see ref . for illustrations and further discussion .
correlator & @xmath121 & non - zero @xmath123 terms + @xmath6-wave & @xmath124 & @xmath125 + @xmath126 & @xmath127 & @xmath128 + @xmath129 & @xmath130 & @xmath131 + @xmath8-wave & @xmath132 & @xmath133 + @xmath134 & @xmath135 & @xmath136 + since our system has translational symmetry , the correlator , @xmath137 , and propagator , @xmath138 , do not depend on @xmath114 .
therefore we can write @xmath139 as discussed in ref . ,
an effective measure of superconductivity is the concept of off - diagonal long range order ( odlro ) which can be quantitatively characterized by the correlator or odlro function , @xmath140 .
the criterion for demonstrating superconductivity is that @xmath140 remains finite , positive and does not decrease to zero at large @xmath114 . using eqs .
( [ eq : pos1 ] , [ eq : pos2 ] ) , the propagator can be written in momentum space as @xmath141 where @xmath142 thus @xmath143 and evaluating @xmath140 in momentum space yields @xmath144 in particular for d - wave symmetry @xmath145 the other correlators were calculated similarly .
we first compute for a one dimensional lattice the approximate bcs and exact hubbard and reduced hamiltonians ground state energies as a function of level filling . in fig .
[ fig : e1 ] the hubbard ( boxes ) , @xmath84 , reduced ( circles ) , @xmath146 , and bcs ( crosses ) , @xmath147 , ground state energy densities are compared as a function of electron filling , @xmath42 , for @xmath148 , @xmath149 sites and @xmath150 .
the bcs and exact hubbard data reproduce the data in refs . and .
as they reported , the agreement is very good , especially for the case of even filling .
the exact reduced data is indistinguishable from the bcs data in the figure . over 99% of the difference between the bcs and the exact hubbard
is due to the difference between @xmath1 and @xmath0 and only 1% is due to quantum correlations between different bcs pairs which are excluded by the variational wavefunction .
similar results have been obtained for other potential strengths and different number of lattice sites ( not shown ) . the hubbard ( boxes ) , reduced ( circles ) and bcs ( crosses ) energy gaps , @xmath85 , are compared in fig . [ fig : g1 ] . as with the ground state energy ,
the bcs gap is the same for either the hubbard or reduced hamiltonian .
in contrast to energy density comparisons , the bcs - hubbard difference is more pronounced especially for half - filling , @xmath151 .
the bcs is still a good treatment for the reduced hamiltonian .
having reproduced and extended one dimensional studies , we now address two dimensional systems .
figure 3 depicts the hubbard , reduced and bcs energy densities for a two dimensional lattice having @xmath152 sites . again
the bcs energies are close to the exact hubbard eigenvalues and indistinguishable from the reduced eigenvalues .
the corresponding energy gap comparisons are displayed in fig .
contrary to conjecture , the difference between the exact and bcs energy gaps remains similar to the one dimensional case .
we further note that the hubbard and reduced energy eigenvalues are generally in reasonably good agreement for all electron densities in both one and two dimensions .
accordingly , it would appear that the reduced hamiltonian incorporates the dominant physical dynamics .
this is fortuitous because @xmath1 can be solved more easily than @xmath0 , whether by the bcs approximation or by diagonalization , which in turn facilitates investigating more complex systems in multi - dimensions .
in general , a variational calculation reproduces the ground state energy much better than other observables .
it is therefore of interest to examine the overlap of the bcs and exact model wavefunctions .
figure [ fig : ov1 ] plots the inner product of the exact , one dimensional hubbard eigenvector with the bcs state as a function of electron filling .
notice the significant deviation from unity especially with increasing density .
as before , the bcs and hubbard results have the largest difference for systems near half - filling and odd number of electrons . in two dimensions
the overlap behavior is similar but now the values are significantly higher ( see fig . [
fig : ov2 ] ) , consistent with the conjecture that the bcs approximation improves with higher dimensionality .
the inner product of the reduced model eigenvector with the @xmath35 is @xmath153 in both 1 and 2 dimensions ( not shown ) .
= 20.5pc = 20.5pc returning to physical observables , this section compares propagators and correlators for exact and approximate model wavefunctions .
unless stated otherwise , all calculations were performed for @xmath150 , lattice sizes @xmath149 x @xmath22 and @xmath154 x @xmath155 , and @xmath156 and @xmath157 .
results for other fillings were also calculated and found to be qualitatively similar .
figures [ fig : p1_16 ] and [ fig : p2_16 ] compare the hubbard , reduced and bcs electron propagators , @xmath158 , for different lattice ranges @xmath8 and @xmath157 in one and two dimensions , respectively .
the bcs and reduced propagators are essentially identical and also similar to the hubbard result . at lower densities
there are somewhat larger differences between the hubbard and bcs ( or reduced ) propagators .
figures [ fig : p1_15 ] and [ fig : p2_15 ] compare the different model single electron propagators again in one and two dimensions , respectively , but now @xmath156 . notice that , just like the model energy comparisons , the hubbard and reduced propagators are more different for odd @xmath42 , but again the bcs accurately reproduces the reduced results .
although lower @xmath42 hubbard - reduced comparisons yield even larger devations ( not shown ) , overall there is no significant difference in agreement between the two propagators in both one and two dimensions .
figures [ fig : s1_16 ] and [ fig : s2_16 ] similarly compare the hubbard , reduced and bcs electron @xmath6-wave correlators , @xmath159 , for @xmath160 in one and two dimensions , respectively .
now there are significant departures between the hubbard and reduced results .
notice that the reduced correlator has a larger magnitude than the hubbard value .
this is consistent with the energy gap comparison , where the reduced gap was larger than the hubbard gap . also , but not shown , lower @xmath42 yields smaller hubbard - reduced correlator differences .
the bcs again very accurately reproduces the reduced correlators .
the hubbard and reduced @xmath6-wave correlators are closer in two dimensions than in one dimension but the limited range of @xmath8 sampled warrants caution in this conclusion . by the mermin - wagner theorem @xcite , the hubbard model will not display long range correlations at @xmath161 in one or two dimensions .
however , according to the bethe ansatz@xcite , at least for large @xmath162 , the ground state energy of the one - dimensional system becomes part of a continuous spectrum as the system length increases .
the theorem does not apply at @xmath163 under these circumstances and therefore long distance correlations are possible .
the limited data shown in figures [ fig : s1_16 ] and [ fig : s2_16 ] are inconclusive on the question of long distance order .
conversely , the bcs and reduced odlro functions are flat for increasing @xmath8 .
the mermin - wagner theorem does not apply to the reduced hamiltonian , even at @xmath161 , because it has long range interactions . for the ground state of the reduced hamiltonian
, the @xmath6-wave correlator reduces to @xmath164 the second term decreases linearly with @xmath25 for large @xmath165 , but the first term is independent of @xmath165 and approaches a constant as @xmath25 increases . for @xmath166 ,
this constant is @xmath167 .
a precise limit can also be given for the bcs correlator at finite @xmath25 .
resummation of the correlator yields terms independent of @xmath165 , @xmath168 with additional terms decreasing rapidly to @xmath79 with increasing @xmath165 .
the hubbard , reduced and bcs electron @xmath6-wave correlators , @xmath159 , for @xmath169 are compared in figs .
[ fig : s1_15 ] and [ fig : s2_15 ] for one and two dimensions , respectively .
the differences are more dramatic for odd fillings as the hubbard correlator in one dimension falls rapidly to @xmath79 .
the two dimensional hubbard correlator also falls rapidly within the limited range in @xmath8 .
in contrast , the bcs and reduced models , with a single specific unpaired electron mode , still exhibit long range correlations .
the @xmath7-wave correlators , @xmath170 , are plotted and compared in figs .
[ fig : p1_16 ] , [ fig : p2_16 ] , [ fig : p1_15 ] and [ fig : p2_15 ] for different fillings and dimensions . both interesting and somewhat surprising
is that the differences between the bcs and reduced correlators are now discernable , especially in 2 dimensions .
since these two model wavefunctions are identical to within 1 part in @xmath171 , this indicates that the @xmath7-wave correlator is a very sensitive wavefunction probe as this correlator s matrix elements are dominated by a very small portion of the state vector .
this result also reflects the small @xmath7-wave correlator magnitude .
none of the models exhibit strong long range order or significant @xmath172 correlations .
finally , the differences between hubbard and reduced @xmath7-wave correlators are dramatic for both one and two dimensions with even or odd filling . finally , the @xmath8-wave correlators , @xmath132 , were calculated in two dimensions ( figs .
[ fig : d2_16 ] and [ fig : d2_15 ] ) . as with the @xmath7-wave correlators ,
the magnitudes are small for non - zero @xmath8 and the differences between the hubbard and reduced correlators are dramatic .
the differences between the reduced and bcs correlators are again present .
the major thrust of this work has been to further numerically assess the accuracy of the bcs approximation . by applying this approximation to multi - electron systems governed by both the hubbard and a simpler , reduced hamiltonian , each exactly diagonalizable
, it has been documented that the bcs provides an excellent description of ground state energies in both one and two dimensions .
however , the bcs single particle energy gaps are less accurate , in both one and two dimensions , especially near half filling .
this undermines previous assertions that the bcs approximation improves with increasing dimension .
further , propagators , overlaps and correlators have been found to be quite sensitive to both interactions and approximations as the bcs and reduced model predictions are appreciably different than the hubbard results .
the slow decrease of the hubbard and reduced hamiltonian odlro propagators and correlator functions with increasing lattice separation is evidence for superconductivity .
however correlations , and thus the degree of superconductivity , are much stronger in the reduced model . finally , the bcs method extremely accurately reproduces all reduced model observables . in summary ,
our results challenge the validity of the bcs method as a good , comprehensive approximation to the attractive hubbard model .
while the hubbard and bcs correlators both exhibit long - range order , their variations with distance are quite different .
because agreement between correlators and inner - product overlaps improves to some extent with dimension , perhaps in three dimensions the bcs method may yet be a very accurate alternative to exact hubbard model wavefunctions but further study is necessary .
finally , for physical systems where the interactions embodied in the reduced hamiltonian dominate , the bcs approximation will be a very accurate approximation scheme permitting significantly more ambitious numerical investigations .
the authors would like to thank dean lee , frank marsiglio , lubos mitas , greg recine and thomas schaefer for informative and fruitful communications .
computer time provided by applied electronics lab , stevens institute of technology is also appreciated .
this work was supported in part by grants doe de - fg02 - 97er41048 and nsf dms-0209931 .
mermin and h. wagner , phys .
* 17 * , 1133 ( 1966 ) and d. ghosh , phys .
* 27 * , 1584 ( 1971 ) proved that expectation value is zero for spin wave operators in the heisenberg and hubbard models , respectively .
d. jasnow and m.e .
fisher , phys .
rev . b**3 * * , 907 ( 1971 ) proved , by the same methods , that a nonzero odlro is explicitly forbidden in the heisenberg model .
the high u expansion of the bethe ansatz is given by m. takahashi , _ thermodynamics of one - dimensional solvable models _
( cambridge university press , cambridge , uk 1999 ) . by symmetry , @xmath173 , with @xmath174 , is equivalent to @xmath175 , with @xmath176 . | we extend previous studies of the bcs canonical approach for the attractive hubbard model .
a derivation of the bcs formulation is presented for both the hubbard and a simpler reduced hamiltonian . using direct diagonalization , exact one and two dimensional solutions for both hamiltonians
are compared to bcs variational calculations .
approximate and exact ground state energies and energy gaps are computed for different electron number systems as well as correlation observables not previously predicted .
reproducing published one dimensional findings , the bcs is an excellent approximation for the hubbard ground state energy but not energy gap , a finding that remains true in two dimensions .
propagators and correlators are found more sensitive to wavefunctions and appreciable differences are computed with the hubbard model exhibiting a weaker degree of superconductivity than the bcs .
however for the reduced hamiltonian model the bcs is an excellent approximation for all observables in both one and two dimensions .
# 1|#1 # 1#1| # 1#1 # 1#2|#1^#2 # 1#2^#2#1| # 1#2^#2#1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1#2 # 1#2 # 1#2 # 1#2#3 # 1#2 # 1#2#3 # 1#2#3#4 _ #
1#1/ *p * _ |
the fundamental plane ( fp ) is an empirical relation between galaxy half - light radius , @xmath6 , surface brightness , @xmath7 , and central velocity dispersion , @xmath8 for early - type galaxies .
locally , @xmath9 in gunn @xmath10 ( jrgensen 1996 ) . under the assumption of homology , this implies @xmath11 .
the fundamental plane is very thin , with an scatter of @xmath12 in coma in @xmath5 ratio for a given galaxy mass , @xmath13 ( jrgensen 1993 ) . with its low scatter ,
the fp is a very powerful tool for measuring the evolution of galaxy @xmath0 ratios as a function of redshift , @xmath13 , etc . because reasonable stellar populations evolve as @xmath14 ( tinsley & gunn 1976 ) , evolution in the properties of the fundamental plane directly probe the star - formation histories of cluster galaxies .
measurements of the evolution of the fp zero - point , scatter , and slope can be used to constrain the mean epoch of star - formation , the spread in early - type galaxy ages ; and the dependence of galaxy age on , e.g. , @xmath13 .
we exploit high resolution keck spectroscopy for velocity dispersions and line strengths , and wfpc2 imaging for the structural parameters , colors , and morphologies .
details can be found in kelson ( 1999abc ) .
the left - hand panel of figure 1 shows the tight fundamental plane relation in the @xmath1 cluster cl1358 + 62 .
the 30 e / s0s in this sample are fit by the following plane in johnson @xmath15 ( kelson 1999c ) : @xmath16 this sample has comparable depth to the fp samples of coma , ensuring that the comparison of these fp exponents with those determined in nearby clusters is reasonably free from selection effects .
we draw the following three conclusions : @xmath17 mild evolution in the fp zero - point , of @xmath18 ( @xmath19 ) , indicates a mean redshift of formation of @xmath20 for the stars in cluster e / s0s .
@xmath17 the scatter in the color - magnitude relation and fp suggest a 1-@xmath8 scatter in ages of 15% . for @xmath21 and @xmath22
, @xmath23-@xmath8 is equivalent to @xmath24 . @xmath17 no evolution in the fp slope
is seen , indicating that the mean luminosity weighted ages of the stellar populations do not depend significantly on @xmath8 or @xmath13 : @xmath25
our sample contains many galaxies of type s0/a through sb .
these early - type spirals are shown with the e / s0s in the right - hand panel of figure 1 .
the spirals show large scatter about the fp and have systematically higher surface brightnesses for a given @xmath8 and @xmath6 .
the scatter for the early - type spirals is twice as large as that for the e / s0s . for the spirals ,
the fp residuals correlate strongly with residual from the color - magnitude relation ( see kelson 1999c ) . correcting the surface brightnesses of the spirals for their young stellar populations using their residuals from the color - magnitude relation ,
one obtains a new plane for the spirals which follows that of the e / s0s .
we conclude that after several gyr , these spirals , including the low - mass e+as , may evolve into present - day , low - mass s0s and appear in nearby fp samples .
the cl1358 + 62 spectra are of sufficiently high @xmath26 that one can accurately measure absorption line strengths , and use stellar population synthesis models , such as those of vazdekis ( 1996 ) , to constrain the systematic variation of the stellar populations along the fp and color - magnitude relations . in figure 2
, we show that the sequence of e / s0 galaxies in cl1358 + 62 is defined by a tight correlation between c@xmath274668 and restframe @xmath28 color .
the best - fit slope of the correlation is nearly identical to that predicted for a sequence of single - burst stellar population models of constant age , in which only metallicity varies . even with these preliminary measurements ,
this conclusion is reinforced by the correlations of the metal - sensitive c@xmath274668index with the high - order balmer lines . in the first panel of figure 3 ,
the c@xmath274668 index is shown to systematically correlate with the h@xmath29 index in a manner which is also consistent with the line of constant age .
the scatter in these two diagrams is consistent with the small scatter in ages that has been estimated using the fp and color - magnitude relations ( kelson 1999c , van dokkum 1998 ) .
the latter two panels show the correlations of c@xmath274668 and h@xmath29 with central velocity dispersion , indicating that metallicity is strongly correlated with the depth of the galaxy potential .
the solid lines are the best - fit slopes , with the @xmath30-@xmath8 uncertainties indicated by dotted lines . using the vazdekis ( 1996 ) models , these correlations with velocity dispersion imply @xmath31 .
preliminary analysis using monte carlo simulations of real data ( kelson 1999 ) indicate that this correlation can produce an observed fp of the approximate form : @xmath32 , though the precise shape depends on sample selection criteria , and the underlying distribution of bulge - to - disk ratios .
we have found that the slope of the fundamental plane has remained constant with redshift .
together with the absorption line strengths and restframe colors , these data suggest that early - type galaxies in clusters have very uniform stellar population ages , and chemical abundances which are strongly determined by the depths of their potential wells , assuming that the stellar populations of cluster e / s0s can be adequately modeled by simple single - burst models .
the implied correlation of metallicity with @xmath8 appears to be the physical basis behind both the color - magnitude relation and the fundamental plane , though further work is required to fully understand the nature of the selection biases and other systematic effects .
jrgensen i. , franx m. , & kjrgaard p. 1993
, , 411 , 34 jrgensen i. , franx m. , & kjrgaard p. 1996
, , 280 , 167 kelson , d.d . , 1999a , , submitted kelson , d.d . , 1999b , , in press kelson , d.d . , 1999c , , in press kelson , d.d .
1999 , in preparation tinsley , b.m . &
gunn , j.e .
1976 , apj , 203 , 52 van dokkum , p.g . , 1998 , , 500 , 714 vazdekis , a. , 1996 , , 107 , 306 worthey , g. & ottavianni , d.l . 1997 , , | the @xmath0 ratios and absorption line - strengths of distant cluster galaxies can be used to directly study their stellar populations , determine their redshift of formation , their scatter in ages , and any dependence of their ages on such internal properties such as velocity dispersion or mass . comparing the zero - point the fundamental plane in the @xmath1 cluster cl1358 + 62 to that in coma , we conclude that the redshift of formation for the stars in massive cluster e / s0s was @xmath2 . the fundamental plane in cl1358 + 62
has the following form : @xmath3 , indicating that the last epoch of star - formation has very little dependence on galaxy mass .
the scatter in the cl1358 + 62 fundamental plane is also very low , equivalent to a scatter in ages of @xmath4.we have also analyzed the @xmath5 ratios of galaxies of type s0/a and later .
these early - type spirals follow a different plane from the e and s0 galaxies , with a scatter that is twice as large as the scatter for the e / s0s
. these residuals also correlate with the residuals from the color - magnitude relation .
preliminary analysis of the absorption line - strengths of the cl1358 + 62 early - types indicate that they are well - described by uniformly old , single - burst stellar populations with metallicity varying as a function of velocity dispersion .
# 1#2 |
let us recall few definitions . * a _ geodesic space _ is a metric space such that any two points can be connected by a minimizing geodesic . *
a _ length space _ is a metric space such that any two points can be connected by a curve with length arbitrary close to the distance between the points . *
a complete length space is an _ alexandrov space with curvature @xmath0 _ if for any quadruple @xmath1 the _ ( 1 + 3)-point comparison _ holds ; i.e. , if @xmath2 or at least one of the model angles @xmath3 is not defined .
_ let @xmath4 be a geodesic space .
assume that for any point @xmath5 there is a neighborhood @xmath6 such that the @xmath7-comparison holds for any quadruple of points in @xmath8 .
then the completion of @xmath4 is an alexandrov space with curvature @xmath0 . _
the question was asked by victor schroeder around 2009 .
later i learned from stephanie alexander , that this statement has an application .
i want to thank stephanie alexander , richard bishop , urs lang , nan li , victor schroeder for the interesting discussions .
a special thanks to alexander lytchak for correcting few mistakes and suggesting an easier proof of lemma [ lem : xyz ] via his favorite ultrapower . in section [ sec :
start ] i give a reformulation of the above theorem , which will be proved in section [ sec : end ] after auxiliary statements proved in section [ sec : prelim ] .
the distance between points @xmath9 and @xmath10 in a metric space will be denoted as @xmath11 .
a geodesic from @xmath9 to @xmath10 will be denoted as @xmath12 $ ] , once we write @xmath12 $ ] we implicitly make a choice of one such geodesic .
also we use the following short - cut notation : @xmath13 x y \r]&=[xy]\backslash\{x\ } , & \l ] x y \r[&=[xy]\backslash\{x , y\ } , & \l [ xy \r[&=[xy]\backslash\{y\}.\end{aligned}\ ] ] we denote by @xmath14 the model @xmath7-plane and @xmath15 ( so @xmath16 if @xmath17 and @xmath18 otherwise ) .
assume @xmath19 $ ] and @xmath20 $ ] be two geodesics in @xmath4 and @xmath21px\r]$ ] and @xmath22py\r]$ ] . since @xmath7-comparison holds in a neighborhood of @xmath23
the function @xmath24 is nonincreasing in both variables for sufficiently small values of the variables .
it follows that angle @xmath25px\r],\bar y\in \l]py\r],\dist{p}{\bar x}{}\to 0,\dist{p}{\bar y}{}\to 0}\ ] ] is defined for any hinge @xmath26,[py])$ ] in @xmath4 . the same way as for alexandrov space
, one can show that @xmath27 for any three hinges formed by geodesics @xmath19 $ ] , @xmath20 $ ] and @xmath28 $ ] .
it follows that , if @xmath29xy\r[$ ] then @xmath30 to prove the theorem , it is sufficient to show that for any @xmath31 , any point @xmath23 and any geodesics @xmath32 $ ] in @xmath4 we have @xmath33 if @xmath34qs\r]$ ] and @xmath35 is small enough .
indeed , once it is proved , it follows that the inequality [ ( * ) ] holds for all @xmath34qs\r]$ ] .
together with [ bigstar ] , it implies that the ( 1 + 3)-point comparison for all quadruples in @xmath4 ; this can be done exactly the same way as in alexandrov space , see @xcite .
hence the completion @xmath36 of @xmath4 is an alexandrov space with curvature @xmath37 for any @xmath31 . from the standard globalization theorem ( as it is stated in @xcite )
we get the result .
as above we denote by @xmath36 the completion of @xmath4 . note that any point @xmath5 admits an open neighborhood @xmath8 in @xmath36 such that @xmath7-comparison holds for any quadruple of points in @xmath8 .
in particular the following condition holds for any point @xmath38 and geodesic @xmath32 $ ] which lie in @xmath39 .
@xmath40 where @xmath34qs\r]$ ] .
an open domain @xmath8 in @xmath36 which satisfy [ ( * * ) ] will be called a _
@xmath7-domain_. note that to prove that @xmath8 is a @xmath7-domain , it is sufficient to check that [ ( * * ) ] holds only if @xmath41 is sufficiently small .
i.e. , if for any point @xmath23 and geodesic @xmath32 $ ] in @xmath39 , the condition [ ( * * ) ] holds if @xmath41 is small enough then [ ( * * ) ] holds for all @xmath42qs\right]$ ] .
note also that if @xmath43 is a @xmath7-domain in @xmath36 then @xmath7-comparison holds for any quadruple of points in @xmath44 .
the later is proved exactly the same way as in alexandrov space : for a quadruple @xmath45 we choose a geodesic @xmath19 $ ] and apply [ bigstar ] together with [ ( * * ) ] for @xmath21px\r[$ ] such that @xmath46 .
( everything works since geodesics with ends in @xmath44 can not leave @xmath43 . ) in particular , if @xmath8 is a @xmath7-domain in @xmath36 then the curvature at each point of @xmath8 is @xmath0 .
therefore any local construction in alexandrov geometry can be performed inside of @xmath8 .
for example , we can construct so called _ radial curves _ as far as we do not get out of @xmath8 .
the radial curves are formed by trajectories which try to escape from a given point @xmath47 using the greedy algorithm ; these curves parametrized in a special way which makes them behave as geodesics in terms of comparisons .
the following proposition is a local version of _ radial monotonicity _ in @xcite and can be proved exactly the same way .
proposition[prop : rad - loc ] let @xmath48 be a @xmath7-domain and @xmath49 .
assume that @xmath50\subset \omega$ ] and @xmath51 then there is a radial curve @xmath52\to \omega$ ] with respect to @xmath47 such that @xmath53 and the distance @xmath54 satisfies the radial monotonicity for any point @xmath23 in @xmath8 .
i.e. , if @xmath55 $ ] is a triangle in @xmath56 $ ] with sides @xmath57 then the function @xmath58 is a nonincreasing in its domain of definition .
the proposition above is used to prove the following lemma .
key lemma[lem : key ] let @xmath59 and @xmath60 be two @xmath7-domains in @xmath36 .
let @xmath61 consider a triangle @xmath62 $ ] in @xmath14 such that @xmath63 set @xmath64 to be the distance from the side @xmath65 $ ] to @xmath66 .
assume @xmath67 and @xmath50\subset \omega_p\cap\omega_q$ ] .
then @xmath68 note that if @xmath69 , then the lemma follows from the triangle inequality ; further we assume @xmath70 .
let @xmath71 $ ] be a point which minimize the distance to @xmath66 ; so @xmath72 .
fix small @xmath73 ; any value @xmath74 will do .
choose @xmath75wp\r]$ ] and @xmath76wq\r]$ ] so that @xmath77 note that geodesic @xmath78 $ ] lies in @xmath79 $ ] .
by alexandrov s lemma ( stated as in @xcite ) , one can choose a point @xmath80 $ ] so that @xmath81 r50 mm pics / key - lemma(0.7 ) set @xmath82 .
consider the radial curve @xmath83\to \omega$ ] with respect to @xmath47 such that @xmath84 .
set @xmath85 . by proposition [ prop : rad - loc ] , we have @xmath86 hence key lemma follows .
lemma[lem : good+good ] let @xmath59 and @xmath60 be two @xmath7-domains in @xmath36 .
let @xmath87 , @xmath88 and @xmath89\subset x\cap(\omega_p\cup\omega_q)$ ] .
then for any geodesic @xmath32\subset \omega_q\cap x$ ] the condition [ ( * * ) ] holds if @xmath41 is sufficiently small .
choose @xmath90\cap\omega_p\cap\omega_q$ ] .
since @xmath60 is a @xmath7-domain , we have @xmath91 for @xmath34qs\r]$ ] .
therefore @xmath92 note that for small values of @xmath41 , we can apply key lemma ; hence the result .
corollary[cor : good+good ] let @xmath93 and @xmath94 be two @xmath7-domains in @xmath36 .
assume @xmath95 is an open set such that for any two points @xmath96 any geodesic @xmath12 $ ] lies in @xmath97 .
then @xmath98 is a @xmath7-domain .
the following lemma makes possible to produce triple of the domains @xmath93 , @xmath94 and @xmath98 as in the corollary above .
lemma[lem : xyz ] let @xmath89 $ ] be a geodesic in @xmath4 and the points @xmath9 , @xmath10 and @xmath99 appear on @xmath100pq\r[$ ] in the same order .
assume that there are @xmath7-domains @xmath101 $ ] and @xmath102 $ ] in @xmath36 .
then there is an open set @xmath103 which contains @xmath104 $ ] and such that for any two points @xmath105 any geodesic @xmath106 $ ] lies in @xmath97 . in particular , by corollary [ cor : good+good ] , @xmath98 is a @xmath7-domain .
before the proof starts , let us discuss ultralimits of metric spaces briefly ; see @xcite for more details .
fix a nonprinciple ultrafilter @xmath107 on the natural numbers .
denote by @xmath108 the @xmath107-power of @xmath36 .
the space @xmath36 will be considered as a subspace of @xmath108 in the natural way .
given a point @xmath109 , we will denote by @xmath110 and @xmath111 the @xmath112-ball centered at @xmath23 in @xmath36 and in @xmath108 correspondingly . in is straightforward to check the following [ clm : k - domain^o ] if @xmath110 is a @xmath7-domain then so is @xmath111 .
arguing by contradiction , assume that there is a sequence of geodesics @xmath113 $ ] such that @xmath114 $ ] , @xmath115 $ ] and @xmath113\not\subset\omega_1\cup\omega_2 $ ] for each @xmath116 .
the @xmath107-limit of @xmath113 $ ] is a geodesic in @xmath108 from @xmath117 to @xmath118 which does not lie in @xmath89 $ ] .
i.e. , geodesics in @xmath108 bifurcate at some point , say @xmath119 $ ] . according to [ clm : k - domain^o ] , if @xmath120 is small enough , the ball @xmath121 forms a @xmath7-domain , a contradiction .
note that one can split the geodesic @xmath89 $ ] into segments in such a way that each segment lies in a @xmath7-domain .
more precisely , there is a sequence of points @xmath122 on @xmath89 $ ] such that the sequence @xmath123 is increasing and each geodesic @xmath124 $ ] lies in a @xmath7-domain . given @xmath120 , the sequence above can be chosen in such a way that in addition @xmath125 . applying lemma [ lem : xyz ] few times , we get that the segment @xmath126 $ ] belongs to one @xmath7-domain . applying key lemma ( [ lem : key ] ) to three points @xmath127 and @xmath34qs\r]$ ] with small enough @xmath41
, we get that [ ( * ) ] holds for @xmath128 , @xmath32 $ ] and @xmath129 . finally since
@xmath130 can be made arbitrary small , the triangle inequality implies that [ ( * ) ] holds for @xmath23 , @xmath32 $ ] and arbitrary @xmath31 .
an alternative proof of the key lemma can be build on the so called _ cat s cradle construction _ from @xcite . this way you do not have to learn who are the radial curves and what is radial monotonicity .
if you were able to do everything as suggested , you should get a weaker version of key lemma . namely you prove the required estimate if @xmath64 is bigger than bisector of @xmath62 $ ] at @xmath66 .
this is good enough for the rest of the proof . playing a bit with the construction , namely making @xmath112
depend on @xmath116 , you can actually get the key lemma in full generality . | i show that if a geodesic space has curvature bounded below locally in the sense of alexandrov then its completion has the same lower curvature bound globally . |
we thank s. moretti for pointing out to us the cms measurement , and initiating the discussion leading to this letter .
we also thank s. gascon - shotkin for useful discussion .
we acknowledge partial support from the labex - lio ( lyon institute of origins ) under grant anr-10-labx-66 and frama ( fr3127 , fdration de recherche `` andr marie ampre '' ) .
ad is partially supported by institut universitaire de france . | after the discovery of a scalar resonance , resembling the higgs boson , its couplings have been extensively studied via the measurement of various production and decay channels on the invariant mass peak .
recently , it has been suggested the possibility to use off - shell measurements : in particular , cms has published results based on the high - invariant mass cross section of the process @xmath0 , which contains the contribution of the higgs . while this measurement has been interpreted as a constraint on the higgs width after very specific assumptions are taken on the higgs couplings , in this letter we show that a much more model - independent interpretation is possible .
lycen 2014 - 08 .5 cm * higgs couplings : disentangling new physics with off - shell measurements*.1 cm * giacomo cacciapaglia , aldo deandrea , guillaume drieu la rochelle and jean - baptiste flament * _ universit de lyon , f-69622 lyon , france ; universit lyon 1 , villeurbanne ; + cnrs / in2p3 , umr5822 , institut de physique nuclaire de lyon + f-69622 villeurbanne cedex , france
_ + [email protected] , [email protected] , [email protected] , [email protected] since the discovery of a scalar boson at the lhc by the atlas and cms collaborations , much effort has been devoted to the study of its properties .
so far , most of the information has been obtained by extracting its couplings from the cross - section measurements in various production / decay modes on the resonance peak @xcite . interpreting those results in a beyond the standard model ( bsm ) context can be easily done in terms of a handful of parameters encoding the modification of the couplings to standard particles and , among the many proposals , one has been chosen as an official recommendation by experimentalists and theorists together @xcite . recently ,
a novel kind of measurement has been put forward , where higgs couplings are extracted from the cross - section integrated away from the resonance peak @xcite , and the first results from cms on the @xmath1 leptons have been published @xcite .
this new class of measurements is most welcome , since it carries information on the higgs couplings at a different mass scale than its mass shell , and such a dependence on the partonic centre of mass energy @xmath2 is paramount to distinguishing bsm effects .
although the measurement @xcite was initially proposed as a mean of determining the higgs width , it was quickly recognised in @xcite that such an extraction was too model - dependent to be meaningful . in this short letter
, we want to point out that the measurement can be primarily interpreted as a bound on the coupling of the higgs to tops and on the loop contribution of new heavy states to the coupling to gluons , in the limit of heavy new physics coloured states .
the direct sensitivity to the new loop is novel , as it is not possible to disentangle it from on - peak measurements .
+ in the experimental analysis @xcite the signal region is defined in order to enhance the rate of events coming from the @xmath0 process , which consists of a box diagram plus the s - channel higgs , over the @xmath3 background .
the higgs contribution consists mainly of a gluon fusion production followed by the decay to two @xmath4 bosons , which interferes with the box diagram .
off - shell , the cross section does not depend explicitly on the higgs width : however , one can always compensate for a different value of the width in the peak region by rescaling appropriately the couplings of the higgs to tops and @xmath5 , thus it is the rescaling of the couplings that affects the off - shell cross section .
the interpretation in terms of width , therefore , only works under very specific assumptions : a large contribution to the total width coming from new decay channels , and the absence of new particles running in the higgs to gluons loop .
the latter point , already properly stressed in @xcite , arises from the fact that at a partonic centre of mass energy away from the higgs mass peak , the loops participating in the production ( and decay ) of the higgs may be resolved if the mass of the exchanged particle is light .
this is the case for standard model ( sm ) loops . to be specific
, the measurement in @xcite is dominated by the gluon fusion production , which is mainly mediated by a top loop in the sm .
this loop form factor does depend on the scale and is significantly different at the higgs mass shell @xmath6 gev and in an off - peak region @xmath7 gev above the top pair threshold .
if the particle running in the loop were much heavier , say around 1 tev , the scale dependence would be quite different . in bsm scenarios
, one can typically expect both modifications of the tree level couplings to the top and new heavy coloured states running in the loop .
due to direct searches at the lhc , the mass of the new states is bound to be quite heavy , thus the approximation of heavy new states is justified . for an on - shell higgs , a cancellation between the two effects may and often does occur , so that the presence of new physics may not appear in on - peak measurement .
then it is clear that only measurements at different @xmath2 can distinguish the two effects since the top and the new physics contribution have a different scale dependence . in this perspective
, the parametrisation of the higgs couplings we proposed in @xcite is very handy , as it suggests to separate into two independent parameters the tree level modification of the top coupling and the effect of any loop of new physics states . the two parameters , relevant for our discussion , are defined as is defined in the same way , while the coupling to the gluons is defined as @xmath8 on - shell , the parameters in the two sets are simply related : @xmath9 , while this degeneracy is removed off - shell .
the difference between the two choices for the coupling to the gluons is crucial for the study of bsm effects from off - shell measurements . ]
@xmath10 where @xmath11 stands for the on - shell amplitude of the process , @xmath12 stands for the contribution from new particles while @xmath13 is the contribution from the top loop with sm coupling to the higgs .
the contribution of new physics in the loop is normalised to the top loop following the expectation that the main contribution comes from new particles associated with the top mass generation .
the two parameters enter in the calculation of the cross section in two different combinations on - peak and off - shell , so that the two effects can be clearly separated by using this parametrisation . [
cols="^,^ " , ] the form factor associated with @xmath14 depends on the mass of the particle(s ) running in the loop , however this dependence is suppressed for large values of the mass , as the loop can be replaced by a higher order operator . to quantify this statement , we considered the process @xmath15 and study the dependence of the off - shell cross section . following the experimental study @xcite , by off - shell cross section
we define the cross section integrated over a range @xmath16 gev after the basic experimental cuts .
note however that this is an intermediate definition between the off - shell region defined in @xcite ( @xmath17gev ) and the signal - enriched region ( @xmath18 gev plus a cut on a matrix element mela discriminant ) , since our cross sections is estimated before the mela discriminant cut .
our calculation has been performed by using the code ` gg2vv ` @xcite , modified by us in order to include the effect of a heavy particle loop . to be specific ,
all the results in this paper are based on a top - like fermion running in the loop . in the left panel of figure [ fig : chi2 ]
we show the dependence of the off - shell cross section as a function of the new fermion mass , fixing @xmath19 ( so that the top loop is standard ) and @xmath20 . from the plot
we can see that the value starts becoming fairly independent on the mass between @xmath21 gev and @xmath22 tev : considering the typical direct bounds from lhc on coloured new particles , the approximation of infinite mass is very effective . in the right panel of figure
[ fig : chi2 ] we also show iso - cross section contours as a function of @xmath23 and @xmath14 together with the region of parameters preferred by on - peak measurements and for large new physics mass .
the degenerate direction @xmath24 is now resolved , thus an interpretation of the experimental bound in terms of these two parameters can be useful to resolve this degeneracy and directly probe new physics loops in the higgs to gluon coupling . in table 1 of @xcite , the off - shell signal - enriched region after mela discrimination
is shown to contain @xmath25 events against the expected @xmath26 : these numbers , when compared to the @xmath27 expected @xmath28 events , show that at @xmath29% c.l .
the study is sensitive to an approximate doubling of the @xmath30 cross section .
the actual bound is extracted from a signal enhanced region , thanks to a matrix element discriminant .
however , the experimental analysis is performed in absence of the new physics loop , therefore a direct reinterpretation of the bound can not be done because the energy dependence and kinematics of the top and new physics loops can differ substantially .
we leave a full analysis for a future publication , and for the experimental collaborations .
the main advantage of our proposal is that the analysis results will contain much more information that the interpretation in terms of the higgs width .
the higgs width bound can be recovered in the limit @xmath31 : in this case , a rescaling of the width by a factor @xmath32 can be compensated on - peak by @xmath33 .
therefore , the published bound @xmath34 at 95% cl can be re - expressed as an upper bound on @xmath35 ( dot in figure [ fig : chi2 ] , with the excluded line on the right side ) .
we can see that this dot is close to the iso - contour giving a doubled cross section , consistently with the previous remark about numbers of events .
this bound , however , only applies to positive values of @xmath23 and it does not cover the possibility of negative couplings , for which the interference term has different sign . to compare this bound with the constraints from the on - peak measurements , in figure [ fig : chi2 ] we show the result of the fit @xcite where we leave the two parameters @xmath23 and @xmath36 free , profile over the new physics loop in the di - photon coupling in order to fit the di - photon signal , and set all the other couplings to the standard model value .
the plot clearly show that the on - peak measurements have a degeneracy for @xmath24 , which is removed by the off - shell measurement .
an analysis of the off - shell measurement in terms of the two parameters would clearly add important information on the higgs couplings , independently on its width .
it is finally interesting to quote the direct bound on @xmath23 coming from the measurement of the @xmath37 associated production . assuming that the decay rate in @xmath38 is standard model like
, the cross section is simply proportional to @xmath39 : from the cms published bound @xcite based on an integrated luminosity of 5 fb@xmath40 at 8 tev , we can extract @xmath41 , while the atlas measurement with full 8 tev dataset leads to @xmath42 at 95%cl @xcite .
+ the simple analysis in this letter shows that the off - shell measurement of @xmath0 , performed by cms @xcite , can be recast to a model - independent bound on the couplings of the higgs to the top and on the contribution of new states in the higgs to gluon coupling . in this perspective , the parametrisation of the higgs couplings that we proposed in previous publications @xcite , where two independent parameters are used for the tree level top coupling and new physics loops , are very useful for a model - independent interpretation of off - shell measurements . |
semitauonic @xmath7 meson decays of the type @xmath12 @xcite are sensitive probes to search for physics beyond the standard model ( sm ) .
charged higgs bosons , which appear in supersymmetry @xcite and other models with at least two higgs doublets @xcite , may contribute measurably to the decays due to the large mass of the @xmath13 .
similarly , leptoquarks @xcite , which carry both baryon number and lepton number , may also contribute to this process .
the ratio of branching fractions @xmath14 is typically used instead of the absolute branching fraction of @xmath15 to reduce several systematic uncertainties , such as those on the experimental efficiency , the magnitude of the cabibbo - kobayashi - maskawa matrix element @xmath16 and the semileptonic decay form factors .
the sm calculations on these ratios predict @xmath17 @xcite and @xmath18 @xcite with a precision of better than 2% and 6% for @xmath19 and @xmath20 , respectively .
exclusive semitauonic @xmath7 decays were first observed by belle @xcite , with subsequent studies reported by belle @xcite , @xcite , and lhcb @xcite .
all the experimental results are consistent with each other , and the average values of refs .
@xcite are @xmath21 and @xmath22 @xcite , which exceed the sm predictions by @xmath23 and @xmath24 , respectively .
the combined analysis of @xmath19 and @xmath20 , taking into account measurement correlations , finds that the deviation is @xmath25 from the sm prediction @xcite .
so far , measurements of @xmath26 at the @xmath7 factories have been performed using hadronic @xcite or inclusive tagging methods @xcite .
semileptonic tagging methods have been employed in studies of @xmath27 decays and have been shown to have similar experimental precision to that of the hadronic tagging method @xcite . in this paper , we report the first measurement of @xmath19 using the semileptonic tagging method .
we reconstruct signal @xmath28 events in modes where one @xmath7 decays semitauonically , @xmath29 followed by @xmath30 ( referred to hereinafter as @xmath31 ) , and the other @xmath7 decays in a semileptonic channel @xmath32 ( referred to hereinafter as @xmath33 ) .
in order to form @xmath19 , we also reconstruct normalization @xmath28 events in modes where both @xmath7 mesons decay to @xmath34 .
we use the full @xmath4 data sample containing @xmath35 @xmath36 pairs recorded with the belle detector @xcite at the kekb @xmath5 collider @xcite .
the belle detector is a general - purpose magnetic spectrometer , which consists of a silicon vertex detector ( svd ) , a 50-layer central drift chamber ( cdc ) , an array of aerogel threshold cherenkov counters ( acc ) , time - of - flight scintillation counters ( tof ) , and an electromagnetic calorimeter ( ecl ) comprising csi(tl ) crystals .
the devices are located inside a superconducting solenoid coil that provides a 1.5 t magnetic field .
an iron flux - return yoke located outside the coil is instrumented to detect @xmath37 mesons and to identify muons ( klm ) .
the detector is described in detail elsewhere @xcite . to determine the reconstruction efficiency and probability density functions ( pdf ) for signal , normalization , and background processes , we use monte carlo ( mc ) simulated events , which are generated with the evtgen event generator @xcite and simulated with the geant3 package @xcite .
the mc samples for signal events are generated using the decay model based on the heavy quark effective theory ( hqet ) in ref .
the normalization mode is simulated using hqet , and reweighted according to the current world - average form factor values : @xmath38 , @xmath39 , and @xmath40 @xcite .
background @xmath41 events are simulated with the isgw @xcite model and reweighted to match the kinematics predicted by the llsw model @xcite . here , @xmath42 denotes the orbitally excited states @xmath43 , @xmath44 , @xmath45 , and @xmath46 .
radially excited states are neglected .
we consider @xmath42 decays to a @xmath47 and a pion , a @xmath48 or an @xmath49 meson , or a pair of pions , where branching fractions are assumed based on quantum - number , phase - space , and isospin arguments .
the sample sizes of the signal , @xmath3 , and continuum @xmath50 @xmath51 production processes correspond to about 40 , 10 , and 6 times the integrated luminosity of the on - resonance @xmath4 data sample , respectively .
charged particle tracks are reconstructed with the svd and cdc . all tracks other than @xmath52 decay daughters
are required to originate from near the interaction point ( ip ) .
electrons are identified by a combination of the specific ionization ( @xmath53 ) in the cdc , the ratio of the cluster energy in the ecl to the track momentum measured with the svd and cdc , the response of the acc , the shower shape in the ecl , and the match between the positions of the shower and the track at the ecl @xcite .
to recover bremsstrahlung photons from electrons , we add the four - momentum of each photon detected within 0.05 radians of the original track direction to the electron momentum .
muons are identified by the track penetration depth and hit distribution in the klm @xcite .
charged kaons are identified by combining information from the @xmath53 in the cdc , the flight time measured with the tof , and the response of the acc @xcite .
we do not apply any particle identification criteria on charged pions .
candidate @xmath54 mesons are formed by combining two oppositely charged tracks with pion mass hypotheses .
we require their invariant mass to lie within 15 mev/@xmath55 of the nominal @xmath56 mass @xcite , which corresponds to approximately @xmath57 , where @xmath11 denotes the resolution of the @xmath58 invariant mass .
we then impose the following additional requirements : both pion tracks must have a large distance of closest approach to the ip in the plane perpendicular to the electron beam line ; the pion tracks must intersect at a common vertex that is displaced from the ip ; and the momentum vector of the @xmath54 candidate should originate from the ip .
neutral pion candidates are formed from pairs of photons with further criteria specific to whether the @xmath59 is from a @xmath60 decay or @xmath61 decay
. for neutral pions from @xmath61 decays , we require the photon daughter energies to be greater than 50 mev , the cosine of the angle in the laboratory frame between the two photons to be greater than zero , and the @xmath62 invariant mass to be within @xmath63 and @xmath64 mev/@xmath55 of the nominal @xmath59 mass @xcite , which corresponds to approximately @xmath65 .
photons are measured as an energy cluster in the ecl with no associated charged tracks .
a mass - constrained fit is then performed to obtain the @xmath59 momentum . for neutral pions from @xmath60 decays , which have lower energies
, we require one photon to have an energy of at least 50 mev and the other to have an energy of at least 20 mev .
we also require a narrow window around the di - photon invariant mass to compensate for the lower photon - energy requirement : within 10 mev/@xmath55 of the nominal @xmath59 mass , which corresponds to approximately @xmath66 .
neutral @xmath61 mesons are reconstructed in the following decay modes : @xmath67 , @xmath68 , @xmath69 , @xmath58 , @xmath70 , @xmath71 , @xmath72 , @xmath73 , @xmath74 , and @xmath75 .
charged @xmath61 mesons are reconstructed in the following modes : @xmath76 , @xmath77 , @xmath78 , @xmath79 , and @xmath80 . the combined reconstructed branching fractions are 37% and 22% for @xmath81 and @xmath82 , respectively . for @xmath61 decay modes without a @xmath59 in the final state , we require the invariant mass of the @xmath61 candidates to be within 15 mev/@xmath55 of the @xmath81 or @xmath82 mass , which corresponds to a window of approximately @xmath83 . for modes with a @xmath59 in the final state , we require a wider invariant mass window : from @xmath84 to @xmath85 mev/@xmath55 around the nominal @xmath81 mass for @xmath81 candidates , and from @xmath86 to @xmath87 mev/@xmath55 around the nominal @xmath82 mass for @xmath82 candidates .
these windows correspond to approximately [ @xmath88 and [ @xmath89 , respectively , in resolution .
candidate @xmath60 mesons are formed by combining @xmath81 and @xmath90 candidates or @xmath82 and @xmath59 candidates . to improve the resolution of the @xmath91-@xmath61 mass difference , @xmath92 , for the @xmath93 decay mode , the charged pion track from the @xmath60 is refitted to the @xmath81 decay vertex .
we require @xmath92 to be within 2.5 mev/@xmath55 and 2.0 mev/@xmath55 , respectively , around the value of the nominal @xmath91-@xmath61 mass difference for the @xmath93 and @xmath94 decay modes .
these windows correspond to @xmath95 and @xmath96 , respectively , in resolution .
we apply a tighter window in the @xmath94 decay mode to suppress a large contribution to the background arising from falsely reconstructed neutral pions .
to tag semileptonic @xmath7 decays , we combine @xmath60 and lepton candidates of opposite electric charge and calculate the cosine of the angle between the momentum of the @xmath7 meson and the @xmath97 system in the @xmath4 rest frame , under the assumption that only one massless particle is not reconstructed : @xmath98 where @xmath99 is the energy of the beam , and @xmath100 , @xmath101 , and @xmath102 are the energy , momentum , and mass , respectively , of the @xmath97 system .
the variable @xmath103 is the nominal @xmath7 meson mass @xcite , and @xmath104 is the nominal @xmath7 meson momentum .
all variables are defined in the @xmath4 rest frame .
figure [ fig : cos_bdstl ] shows the @xmath105 distribution for signal and normalization decay modes in mc samples .
correctly reconstructed @xmath7 candidates in the normalization decay mode are expected to have a value of @xmath105 between @xmath106 and @xmath107 .
correctly reconstructed @xmath7 candidates in the signal decay mode and falsely reconstructed @xmath7 candidates tend to have values of @xmath105 below the physical region due to contributions from additional particles . in each event
we require two tagged @xmath7 candidates that are opposite in flavor .
signal events may have the same flavor due to @xmath36 mixing ; however , we veto such events as they lead to ambiguous @xmath97 pair assignment and larger combinatorial background .
we require that at most one @xmath7 meson be reconstructed from a @xmath82 mode to avoid large background from fake neutral pions when forming @xmath91 candidates . in each signal event
, we assign the candidate with the lower value of @xmath105 ( referred to hereinafter as @xmath108 ) as @xmath31 . the probability of falsely assigning the @xmath31 as the @xmath33 for signal events is about 3% , according to mc simulation . after the identification of the @xmath31 and @xmath33 candidates , we apply further background suppression criteria . on the tag side ,
we require @xmath109 to select @xmath110 . on the signal side
, we require the @xmath91 momentum in the @xmath4 rest frame to be less than 2.0 gev/@xmath111 , whlie , on the tag side , we require it to be less than 2.5 gev/@xmath111 , which accounts for the lepton mass difference . finally , we require that events contain no extra charged tracks , @xmath54 candidates , or @xmath59 candidates , which are reconstructed with the same criteria as those used for the @xmath61 candidates . at this stage ,
the probability of finding multiple candidates is 7% .
when multiple candidates are found in an event , we select the most signal - like candidates based on the quality of vertex - constrained fits for the @xmath61 mesons . distributions for @xmath29 ( solid red circles ) and @xmath32 ( open black circles ) taken from mc simulation.,width=302 ]
to separate reconstructed signal and normalization events , we employ a neural network using the neurobayes software package @xcite .
the variables used as inputs to the network are @xmath112 , the missing mass squared @xmath113 , and the visible energy @xmath114 , where @xmath115 is the four - momentum of particle @xmath116 in the @xmath4 rest frame .
the most powerful observable in separating signal and normalization is @xmath112 . the neural network is trained using mc samples of signal and normalization events .
the dominant background contributions arise from events with misreconstructed @xmath47 mesons ( denoted fakes ) .
the sub - dominant contributions arise from two sources in which @xmath91 mesons from both @xmath31 and @xmath33 are correctly reconstructed .
one source is @xmath41 , where the @xmath42 meson decays to @xmath47 and other particles .
the other source is @xmath117 events , where one @xmath91 meson is correctly reconstructed and the other charmed meson @xmath118 decays semileptonically . if the hadrons in the semileptonic @xmath118 decay are not identified , such events can mimic signal . similarly , events in which @xmath118 is a @xmath119 meson decaying into @xmath120 can also mimic signal . to separate signal and normalization events from background processes , we place a criterion on the sum of the energies of neutral clusters detected in the ecl that are not associated with reconstructed particles , denoted as @xmath121 . to mitigate the effects of photons related to beam background in the energy sum
, we only include clusters with energies greater than 50 , 100 , and 150 mev , respectively , from the barrel , forward , and backward calorimeter regions , defined in ref .
@xcite . signal and
normalization events peak near zero in @xmath121 , while background events can populate a wider range .
we require @xmath121 to be less than 1.2 gev .
to improve the accuracy of the mc simulation , we apply a series of calibration factors determined from control sample measurements .
the lepton identification efficiencies are separately corrected for electrons and for muons to account for differences between the detector responses in data and mc . correction factors for lepton identification efficiencies are evaluated as a functions of the momentum and direction of the lepton using @xmath122 and @xmath123 decays .
we reweight events to account for differing @xmath47 yields between data and mc .
the differing yields of correctly reconstructed @xmath47 mesons in data and mc affect the @xmath19 measurement , as it biases the determination of the background contribution .
calibration factors for events with both correctly- and falsely - reconstructed @xmath61 mesons are estimated for each @xmath61 meson decay mode using a two - dimensional fit to @xmath124 .
for this calibration , we use samples with all selection criteria other than @xmath61 mass and @xmath92 applied
. a two - dimensional pdf is constructed by taking the product of the one - dimensional functions for @xmath124 .
the pdf in each dimension is the sum of a signal component and a background component modelled with a first - order polynomial .
the signal component is a triple gaussian for @xmath81 decay modes without a @xmath59 and a crystal ball function @xcite plus a gaussian for @xmath81 decay modes with a @xmath59 and @xmath82 decay modes . in this calibration
, we do not distinguish signal and tag sides . to estimate calibration factors for specific @xmath61 decay modes , we fit samples in which one @xmath61 meson is reconstructed in a specific mode while the other is reconstructed in any signal mode . from the ratios of data
to mc samples in signal and background yields , we derive calibration factors of the specific decay mode for events with correctly and falsely reconstructed @xmath61 mesons .
we can not independently determine calibration factors for all @xmath61 meson decay modes as we use other decay modes when we calibrate each specific decay mode of a given @xmath61 meson . to estimate all the calibration factors correctly , we first perform the two - dimensional fit for each decay mode without weighting factors , and then iterate the fits using resultant weighting factors until all calibration factors converge . similarly , we estimate calibration factors for events with correctly and falsely reconstructed @xmath91 mesons from a two - dimensional fit to @xmath92 .
calibration factors for events with correctly and falsely reconstructed @xmath91 mesons are separately estimated for subsequent decay to @xmath81 and @xmath82 mesons .
we extract the yields of the signal and normalization processes from a two - dimensional extended maximum - likelihood fit to neural network classifier output @xmath125 and @xmath121 .
the likelihood function consists of five components : signal , normalization , fake @xmath47 events , @xmath41 , and other backgrounds ( predominantly from @xmath117 ) .
the pdfs of all components are determined from mc simulation .
there are significant correlations between @xmath125 and @xmath121 in the normalization and background components , but not for the signal .
we therefore construct the normalization and background pdfs using two - dimensional histograms and apply a smoothing procedure to account for its limited statistical power @xcite .
the signal pdf is the product of one - dimensional histograms in @xmath125 and @xmath121 .
three parameters are floated in the final fit : the yields of the signal , normalization , and @xmath41 components .
the yield of fake @xmath47 events is fixed to the value estimated from sidebands in the @xmath92 distribution . since the pdf shape of fake @xmath47 events depends on the composition of signal , normalization , @xmath41 , and other backgrounds , the relative contributions of these processes to the fake @xmath47 component
are described as a function of the three fit parameters .
the yields of other backgrounds are fixed to the values expected from mc simulation .
the ratio @xmath19 is given by the formula : @xmath126 where @xmath127 and @xmath128 are the reconstruction efficiency and yields of signal ( normalization ) events .
we use @xmath129 as the average of the world averages for @xmath130 and @xmath131 @xcite .
the ratio of efficiencies , @xmath132 , is estimated to be @xmath133 from mc simulation .
the difference between reconstruction efficiencies of signal and normalization events arises from their distinct lepton momentum distributions , and the different event criteria on the @xmath91 momenta .
we validate the pdfs used in the fitting procedure by analysing various control samples . for fake @xmath47 events , we study the @xmath92 sidebands , where we find good agreement in both @xmath125 and @xmath121 . for @xmath110
decays , we require one @xmath7 meson to be reconstructed with the hadronic tagging method @xcite and the other @xmath7 meson reconstructed with the nominal criteria of this analysis . we find good agreement between data and mc in the @xmath121 , @xmath134 , and @xmath135 distributions , but small discrepancies in the @xmath105 distributions , which we incorporate as a systematic uncertainty .
to estimate the systematic uncertainties on @xmath19 , we vary every fixed parameter in turn by one standard deviation and repeat the fit .
the systematic uncertainties are summarized in table [ tab : systematics ] .
the dominant systematic uncertainty arises from the limited size of the mc samples : to estimate this uncertainty , we recalculate pdfs for signal , normalization , fake @xmath47 events , @xmath41 , and other backgrounds by generating toy mc samples from the nominal pdfs according to poisson statistics and repeat the fit with the new pdfs . small discrepancies between the data and mc are found in the @xmath105 distributions in the hadronic tagged samples .
we estimate it as `` pdf shape of the normalization in @xmath105 '' in table [ tab : systematics ] by correcting the @xmath105 distribution in mc samples according to the observed discrepancies , and repeating the fit .
the branching fractions of the @xmath41 decay modes and the decays of the @xmath42 mesons are not well known and therefore contribute significantly to the total pdf uncertainty for @xmath41 decays .
the branching fraction of each @xmath41 decay is varied within its uncertainty .
the uncertainties are assumed to be @xmath136 for @xmath43 , @xmath137 for @xmath44 , @xmath138 for @xmath45 , and @xmath139 for @xmath46 , including the limited knowledge of the @xmath42 decays .
we also consider the impact of contributions from radially excited @xmath140 and @xmath141 , where we assume the branching fractions of @xmath142 to be as large as @xmath143 .
the yield of fake @xmath144 events is fixed to the value estimated from sidebands in the @xmath92 distribution .
we vary this yield within its uncertainties .
we also vary the calibration factors for @xmath61 meson decay modes within their uncertainties for events with falsely reconstructed @xmath47 events .
the yields of other background processes , predominantly from @xmath117 events , are fixed to the values estimated from mc simulation .
we consider variations on the yield and shape of the pdf of these background processes within their measured uncertainties .
the uncertainties for the @xmath117 channels are assumed to be @xmath145 for @xmath146 , @xmath147 for @xmath148 , @xmath145 for @xmath149 , and @xmath137 for @xmath150 .
furthermore , we add an uncertainty of @xmath151 due to the size of the mc sample . we determine the uncertainty from the branching fraction of @xmath152 decay ( which may peak near the signal in the @xmath121 distribution ) to be negligible . the reconstruction efficiency ratio of signal to normalization events is varied within its uncertainty , which is limited by the size of the mc samples for signal events .
we include other minor systematic uncertainties from two sources : one related to the parameters that are used for the reweighting of the semileptonic @xmath153 decays from the isgw model to the llsw model ; and the other from the branching fraction of @xmath30 decay @xcite .
the total systematic uncertainty is estimated by summing the above uncertainties in quadrature . [ cols="^,^,^,^ " , ] [ tab : p_values ]
in conclusion , we report the first measurement of @xmath19 with a semileptonic tagging method using a data sample containing @xmath154 pairs collected with the belle detector .
the result is @xmath155 which is within @xmath10 of the sm prediction including systematic uncertainties , and is in good agreement with other measurements by belle @xcite , @xcite , and lhcb @xcite .
the result is statistically independent of earlier belle measurements .
we investigate the compatibility of the data samples with new physics in a model - independent method by adding the operators one by one .
we also study two types of leptoquark models .
we find our data allow the additional contributions from scalar and vector operators while disfavoring large additional contributions from a tensor operator with @xmath156 , an @xmath157-type leptoquark model with @xmath158 , or an @xmath159-type leptoquark model with @xmath160 , when considering the impact on the decay kinematics .
we thank y. sakaki , r. watanabe , and m. tanaka for their invaluable suggestions .
this work was supported in part by a grant - in - aid for jsps fellows ( no.13j03438 ) and a grant - in - aid for scientific research ( s ) `` probing new physics with tau - lepton '' ( no.26220706 ) .
we thank the kekb group for the excellent operation of the accelerator ; the kek cryogenics group for the efficient operation of the solenoid ; and the kek computer group , the national institute of informatics , and the pnnl / emsl computing group for valuable computing and sinet4 network support .
we acknowledge support from the ministry of education , culture , sports , science , and technology ( mext ) of japan , the japan society for the promotion of science ( jsps ) , and the tau - lepton physics research center of nagoya university ; the australian research council ; austrian science fund under grant no .
p 22742-n16 and p 26794-n20 ; the national natural science foundation of china under contracts no . 10575109 , no . 10775142 , no .
10875115 , no .
11175187 , no . 11475187 and no . 11575017
; the chinese academy of science center for excellence in particle physics ; the ministry of education , youth and sports of the czech republic under contract no .
lg14034 ; the carl zeiss foundation , the deutsche forschungsgemeinschaft , the excellence cluster universe , and the volkswagenstiftung ; the department of science and technology of india ; the istituto nazionale di fisica nucleare of italy ; the wcu program of the ministry of education , national research foundation ( nrf ) of korea grants no .
2011 - 0029457 , no .
2012 - 0008143 , no . 2012r1a1a2008330 , no .
2013r1a1a3007772 , no .
2014r1a2a2a01005286 , no .
2014r1a2a2a01002734 , no . 2015r1a2a2a01003280 , no .
2015h1a2a1033649 ; the basic research lab program under nrf grant no .
krf-2011 - 0020333 , center for korean j - parc users , no .
nrf-2013k1a3a7a06056592 ; the brain korea 21-plus program and radiation science research institute ; the polish ministry of science and higher education and the national science center ; the ministry of education and science of the russian federation and the russian foundation for basic research ; the slovenian research agency ; ikerbasque , basque foundation for science and the euskal herriko unibertsitatea ( upv / ehu ) under program ufi 11/55 ( spain ) ; the swiss national science foundation ; the ministry of education and the ministry of science and technology of taiwan ; and the u.s .
department of energy and the national science foundation .
this work is supported by a grant - in - aid from mext for science research in a priority area ( `` new development of flavor physics '' ) and from jsps for creative scientific research ( `` evolution of tau - lepton physics '' ) .
99 charge - conjugate decays are implied throughout this paper , unless otherwise stated .
h. itoh , s. komine , and y. okada , prog .
. phys . * 114 * , 179 ( 2005 ) .
m. tanaka , z. phys .
c * 67 * , 321 ( 1995 ) .
y. sakaki , r. watanabe , m. tanaka , and a. tayduganov , phys .
d * 88 * , 094012 ( 2013 ) .
s. fajfer , j.f .
kamenik , and i. nisandzic , phys .
d * 85 * , 094025 ( 2012 ) .
kamenik , and f. mescia , phys .
d * 78 * , 014003 ( 2008 ) .
lees _ et al . _
( collaboration ) , phys .
lett . * 109 * , 101802 ( 2012 ) ; j.p .
lees _ et al . _
( collaboration ) , phys .
d * 88 * , 072012 ( 2013 ) .
a. matyja _ et al .
_ ( belle collaboration ) , phys . rev . lett . * 99 * , 191807 ( 2007 ) .
a. bozek _ et al .
_ ( belle collaboration ) , phys .
d * 82 * , 072005 ( 2010 ) .
m. huschle _ et al . _
( belle collaboration ) , phys .
d * 92 * , 072014 ( 2015 ) .
r. aaij _ et al . _
( lhcb collaboration ) , phys .
lett . * 115 * , 111803 ( 2015 ) .
y. amhis _ et al . _
( heavy flavor averaging group ) , arxiv:1412.7515 and online update at http://www.slac.stanford.edu / xorg / hfag/. b. kronenbitter _ et al . _ ( belle collaboration ) , phys .
d * 92 * , 051102(r ) ( 2015 ) .
b. aubert _ et al .
_ ( collaboration ) , phys . rev .
d * 81 * , 051101(r ) ( 2010 ) .
a. abashian _ et al . _ ( belle collaboration ) , nucl .
methods phys .
res . , sect .
a * 479 * , 117 ( 2002 ) ; also see the detector section in j. brodzicka _ et al . _ , prog .
( 2012 ) 04d001 .
s. kurokawa and e. kikutani , nucl .
methods phys .
res . , sect .
a * 499 * , 1 ( 2003 ) , and other papers included in this volume ; t. abe _ et al . _ ,
( 2013 ) 03a001 and following articles up to 03a011 .
lange , nucl .
methods phys .
a * 462 * , 152 ( 2001 ) .
r. brun _ et al .
_ , geant3.21 , cern report no .
dd / ee/84 - 1 , ( 1984 ) ( unpublished ) . m. tanaka and r. watanabe , phys .
d * 87 * , 034028 ( 2013 ) .
d. scora and n. isgur , phys .
d * 52 * , 2783 ( 1995 ) .
leibovich , z. ligeti , i.w .
stewart , and m.b .
wise , phys .
d * 57 * , 308 ( 1998 ) .
k. hanagaki , h. kakuno , h. ikeda , t. iijima , and t. tsukamoto , nucl .
methods phys .
res . , sect .
a * 485 * , 490 ( 2002 ) .
a. abashian _ et al .
methods phys .
res . , sect .
a * 491 * , 69 ( 2002 ) .
e. nakano , nucl .
methods phys .
res . , sect .
a * 494 * , 402 ( 2002 ) .
olive _ et al _ ( particle data group ) , chin .
c * 38 * , 090001 ( 2014 ) .
m. feindt and u. kerzel , nucl .
methods phys .
res . , sect .
a * 559 * , 190 ( 2006 ) .
t. skwarnicki , ph.d .
thesis , institute for nuclear physics , krakow 1986 ; desy internal report , desy f31 - 86 - 02 ( 1986 ) .
friedman , data analysis techniques for high energy particle physics , in : proc .
1974 cern school of computing , cern 74 - 23 ( 1974 ) .
m. feindt , f. keller , m. kreps , t. kuhr , s. neubauer , d. zander , and a. zunpac , nucl .
methods phys .
res . , sect .
a * 654 * , 432 ( 2011 ) .
see supplemental material for fit figures of the controls samples , efficiency variation , measured value of @xmath19 , and background - subtracted momenta distributions of @xmath91 and lepton . z. z. xing , h. zhang and s. zhou , phys .
d * 77 * , 113016 ( 2008 ) . | we report a measurement of the ratio @xmath0 , where @xmath1 denotes an electron or a muon .
the results are based on a data sample containing @xmath2 @xmath3 pairs recorded at the @xmath4 resonance with the belle detector at the kekb @xmath5 collider .
we select a sample of @xmath6 pairs by reconstructing both @xmath7 mesons in semileptonic decays to @xmath8 .
we measure @xmath9 , which is within @xmath10 of the standard model theoretical expectation , where the standard deviation @xmath11 includes systematic uncertainties .
we use this measurement to constrain several scenarios of new physics in a model - independent approach . |
consider an inextensible string of unit tension placed along the @xmath0 axis with its left endpoint at @xmath1 , and let @xmath2 be the total mass of the string segment from @xmath1 to @xmath0 .
the small vertical oscillations @xmath3 of the string obey the equation @xmath4 where @xmath5 is the square of the oscillation frequency .
the natural frequencies of the string depend on how it is tied at the ends ; suppose @xmath6 where @xmath7 is the supremum of the points of increase of @xmath8 , and @xmath9 is some `` tying constant '' .
now , let @xmath10 , and denote by @xmath11 and @xmath12 the two particular solutions of equation ( [ stringequation ] ) satisfying @xmath13 the _ characteristic function _ of the string is then defined by @xmath14 or , equivalently , @xmath15 where @xmath16 is , up to an unimportant factor , the unique non - negative solution of the string equation that is decreasing and satisfies @xmath17 .
the definition of @xmath18 may be extended to @xmath19 by analytic continuation ; @xmath18 is the analog of the weyl
titchmarsh function in the sturm
liouville theory .
the present paper is devoted to the inverse spectral problem for this vibrating string equation , namely the problem of recovering @xmath8 from @xmath18 .
we call this _ krein s inverse problem_. we describe and illustrate a novel algorithm that is applicable to a class of characteristic functions associated with a certain continued fraction .
the remainder of this introduction provides a summary of the approach and discusses its relationship with other works . in a series of papers published in the soviet union during the 1950 s
, m. g. krein made a detailed study of the existence of spectral expansions associated with the vibrating string equation .
for some useful accounts of this work in the english language , see @xcite ; in particular , @xcite contains a rough outline of the historical development of krein s ideas and their overlap with the work of feller aimed at a unified analytical treatment of some discrete and continuous stochastic processes .
denote by @xmath20 the set of functions @xmath21 \rightarrow [ 0,\infty]$ ] that are non - decreasing , right - continuous and infinite at infinity .
we use the convention @xmath22 so that @xmath23 is a well - defined stieltjes measure .
the tying constant @xmath24 is absorbed in the definition of @xmath8 by setting @xmath25 for such @xmath8 , krein gave a precise meaning to equation ( [ stringequation ] ) and its particular solutions @xmath11 , @xmath12 and @xmath16 .
with a slight abuse of terminology , we shall identify @xmath8 with the string .
krein showed that the characteristic function defined by equation ( [ characteristicfunction ] ) is necessarily of the form @xmath26 where @xmath27 , @xmath28 is a right - continuous non - decreasing function with support in @xmath29 such that @xmath30 and @xmath31 .
the paramater @xmath32 is in fact the infimum of the points of increase of @xmath8 , and the function @xmath28 in this representation is called the _ principal spectral function _ of the string .
we shall denote by @xmath33 the set of _ all _ functions expressible in this form .
hence krein showed that @xmath34 for every @xmath35 .
he also conjectured the converse : _ every _ element of @xmath33 is the characteristic function of a unique string in @xmath20 . to the best of our knowledge ,
the first proof of this conjecture was given in @xcite , chapter 6 ; it uses the theory of hilbert spaces of entire functions , and a deep uniqueness theorem due to de branges .
although there is no known systematic procedure for solving krein s inverse problem in its full generality , effective algorithms do exist if the mass distribution belongs to certain special classes .
for instance , if @xmath8 is smooth , then the change of variables @xmath36 where @xmath37 is the inverse of @xmath38 , transforms the string equation into @xmath39 where @xmath40 is the so - called impedance . the inverse problem for this equation
may be tackled by `` integral operator '' techniques going back to the works of gelfand , krein , levitan and marchenko ; see @xcite for an informal account of these techniques and their historical background , and @xcite for more recent developments .
general methods inspired by the gelfand
levitan approach aimed at solving the inverse problem for the string equation have been considered by dym & kravitsky @xcite , winkler @xcite and boumenir @xcite .
another class for which the inverse problem is tractable arises in connection with _ stieltjes moment problem _ @xcite : given a sequence of numbers @xmath42 , find a measure @xmath43 on @xmath44 such that @xmath45 briefly , stieltjes approach was to construct from the @xmath46 a certain continued fraction , namely @xmath47 stieltjes gave an algorithm that permits the calculation of the @xmath48 from the @xmath49 .
the coefficients @xmath48 thus obtained form a finite or infinite sequence of strictly positive numbers , depending on whether the measure @xmath28 has finitely or infinitely many points of growth .
stieltjes distinguished the _ determinate _ case , in which the moment problem has a unique solution , and its complement , the _ indeterminate _ case .
the determinate case arises if and only if the continued fraction either terminates ( i.e. the sequence of the @xmath48 is finite ) or else is infinite but convergent .
the continued fraction is then a well - defined function of @xmath5 which may be expressed as the stieltjes transform of the sought measure @xmath28 .
stieltjes also found a beautifully simple criterion for the convergence of the continued fraction when the sequence of the @xmath48 is infinite ; he showed that convergence occurs if and only if @xmath50 this work on the moment problem provided the inspiration for krein s own theory of strings ; see supplement ii in @xcite . with remarkable insight , krein was able to relate stieltjes continued fraction to the characteristic function of a _ discrete _ string , i.e. a string made up of a sequence of point masses @xmath51 at the positions @xmath52 . to see how , set @xmath53 the string equation then yields a riccati equation for @xmath54 : @xmath55 for a string consisting of point masses , @xmath56 is a linear combination of dirac deltas ; at a point @xmath52 with concentrated mass @xmath51 , this riccati equation should be interpreted as @xmath57 by using equation ( [ weylfunction ] ) , we find @xmath58 and , by iterating , @xmath59 if the principal spectral function has finitely many points of increase , then the string consists of finitely many , say @xmath60 , point masses with finite mass ( and a further point of infinite mass at @xmath24 ) .
then @xmath61 , the expansion terminates , and we have the finite case of stieltjes continued fraction ( [ stieltjescontinuedfraction ] ) .
if the principal spectral function has infinitely many points of increase , then there are infinitely many point masses ; the continued fraction does not terminate but its coefficients obviously still determine @xmath2 for every @xmath0 such that @xmath62 in particular , if @xmath63 then the continued fraction converges and the string is _ completely _ determined by its coefficients since either @xmath64 or else @xmath65 .
let us now elaborate the significance of the foregoing remarks for krein s inverse problem . define @xmath66 and denote by @xmath41 the corresponding set of strings .
every @xmath67 may be expanded in a continued fraction of the form ( [ kreinexpansion ] ) , and the coefficients in the expansion may be computed from the moments of the principal spectral function @xmath28 .
it will serve our purpose to abuse stieltjes terminology somewhat by saying that @xmath18 itself belongs to the _ determinate subclass _ of @xmath68 if the continued fraction is either finite or convergent , and to the _ indeterminate subclass _ otherwise .
it follows , then , that stieltjes continued fraction provides an effective solution method of the inverse problem for strings in @xmath41 whose characteristic functions are in the determinate subclass .
not every member of @xmath41 is a discrete string ; all that can be said is that such strings must begin with a sequence of point masses ; see @xcite , 5.9 .
if @xmath67 is in the determinate subclass , then @xmath8 is a discrete string .
the converse is of course not true , and for a discrete string whose characteristic function is in the indeterminate subclass , one can not recover the tying constant from the continued fraction coefficients alone ; see @xcite , 13 .
[ discreteremark ] the connection between discrete strings and continued fractions can be exploited to provide a reconstruction algorithm of greater applicability .
beals use this idea in the context of integrable systems @xcite ; borcea use it as the basis of their numerical treatment of the inverse sturm liouville problem in impedance form @xcite .
the practical issue that arises is how to find approximations of the spectral data that satisfy stieltjes moment condition and have explicitly computable moments .
both beals and borcea resolve this issue by restricting their attention to the case of a _ discrete spectrum _
, where the characteristic function has a mittag leffler expansion which may be truncated to furnish the required approximations .
we present an alternative truncation strategy applicable to a class that includes some characteristic functions with an absolutely continuous spectrum .
this class is a proper subset of a class denoted @xmath69 , based on a continued fraction expansion that arises naturally in connection with krein s inverse problem .
the expansion rests on the following two useful `` rules '' relating the characteristic function to its string .
the first of these rules concerns the characteristic function of the right - continuous inverse @xmath70 of the string @xmath8 : @xmath71 [ dualproposition ] the string @xmath70 will be called the _ dual _ of @xmath8 .
the second rule concerns the string obtained from @xmath8 by removing @xmath1 from the spectrum of @xmath28 : let @xmath8 be the string with characteristic function @xmath72 then @xmath73 is the characteristic function of the string @xmath74 defined by @xmath75 ^ 2\,\mbox{\em d } \tau\,.\ ] ] [ zeromassproposition ] two other useful formulae connecting @xmath18 and @xmath8 are : @xmath76 and , in the case @xmath77 , @xmath78 for proofs of these results , see @xcite , 6.8 and 6.9 . using these rules and setting @xmath79 , we may write @xmath80 where @xmath81 is the characteristic function of the dual string , say @xmath82 , of @xmath83 . by the same argument ,
@xmath84 where @xmath85 is the characteristic function of the dual string , say @xmath86 , of @xmath87 .
iterating , we obtain @xmath88 this continued fraction , albeit in a very particular context and in a somewhat disguised form , was introduced and studied in our recent work on the excursions of diffusion processes @xcite .
it is also implicit in some of the calculations carried out by donati martin and yor in their generalisation of lvy s formula for the area enclosed by a planar brownian motion @xcite . in [ katssection ]
, we study the relationship between this continued fraction and that of stieltjes .
this relationship may be described in terms of the map @xmath89 then @xmath90 hence the class @xmath91 is the set of characteristic functions that plays , for the continued fraction ( [ katscontinuedfraction ] ) , the part played by the class @xmath68 for the continued fraction of stieltjes .
we show by elementary means that the corresponding set of strings , denoted @xmath69 , consists of those strings whose _ dual _ principal spectral functions have _ negative _ moments .
then , by a straightforward application of kats results on such spectral functions @xcite , we derive a simple criterion , expressed in terms of the mass distribution @xmath2 , for @xmath8 to be in the class @xmath69 .
the upshot is that this class is large enough to be of interest ; in particular , it contains some strings with an absolutely continuous principal spectral function . just as in the case of the stieltjes continued fraction , @xmath92 may be partitioned into two subclasses : determinate and indeterminate .
the determinate subclass consists of all those characteristic functions for which the continued fraction expansion ( [ katscontinuedfraction ] ) is either finite or else convergent the latter case occuring if and only if equation ( [ stieltjesseries ] ) holds .
[ inversesection ] discusses the finite case .
we show there that the string @xmath8 corresponding to such a characteristic function is discrete with finitely many point masses , and we devise an algorithm to compute it explicitly , given the continued fraction coefficients .
the infinite ( convergent ) case may then be tackled by considering the truncations of the continued fraction after @xmath60 terms ; since krein s correspondence between strings and characteristic functions is a homeomorphism , the discrete string so obtained yields in the limit @xmath93 the solution of the inverse problem . in [ diffusionsection ] , we give some examples that arise in the study of the excursions of one - dimensional diffusion processes . knight @xcite and kotani & watanabe @xcite pointed out simultaneously the relevance of krein s theory to this topic : every string @xmath8 defines a _ generalised diffusion process _ , and the excursions of the process from its starting point may be described by means of the principal sprectral function of the dual string . the probabilistic version of krein s inverse problem is to find the diffusion , given the distribution of the excursion lengths .
donati martin and
yor solved this problem explicitly in a number of interesting cases @xcite , and we use their findings to illustrate the effectiveness of our algorithm .
the paper ends with a few concluding remarks in [ conclusionsection ] .
for every @xmath34 , @xmath94 . furthermore , the map @xmath95 from @xmath33 to itself is an involution .
[ involutionlemma ] let @xmath34 . since @xmath96 , we can write @xmath97 where @xmath98 , @xmath99 and @xmath100 a simple calculation shows that @xmath101 where @xmath102
it is then readily seen that @xmath103 and so @xmath94 .
the fact that @xmath104 then follows easily .
@xmath105 [ momenttheorem ] the fact that @xmath106 is an involution implies @xmath107 note that , by definition of @xmath68 , @xmath108 if and only if @xmath109 the theorem is then a consequence of equation ( [ dualmeasure ] )
. @xmath110 [ dualcorollary ] the dual of a discrete string is a discrete string .
hence @xmath111 furthermore , the maps @xmath112 commute .
next , we seek a simple criterion in terms of the mass distribution for @xmath8 to belong to the class @xmath69
. we shall need the following key result , due to kats @xcite : for the principal spectral function @xmath28 of a string @xmath35 to satisfy @xmath113 it is necessary and sufficient that @xmath114 and that @xmath115^{1 + \varepsilon } = o(1 ) \quad \mbox{as $ x \rightarrow l-$}\,.\ ] ] @xmath116 if and only if either 1 .
@xmath117 and @xmath118^{1+\varepsilon } = 0 $ ] + or 2 .
@xmath119 and @xmath120^{1+\varepsilon } = 0 $ ] .
[ stringtheorem ] we note first that , for @xmath116 , the implication @xmath121 holds because @xmath122 has an expansion of the form ( [ kreinexpansion ] ) . furthermore @xmath123 .
hence we have the trivial identity @xmath124 \;\mbox{or}\ ; \left [ m \in { \mathbb m}^{- } \;\mbox{and}\ ; m(\infty- ) = \infty \right ] \ ] ] @xmath125 \;\mbox{or}\ ; \left [ m^\ast \in { \mathbb m}^{- } \;\mbox{and}\ ; m^\ast(\infty- ) < \infty \right ] \,.\ ] ] let us show first that @xmath126 \;\iff\ ; \mbox{condition ( ii ) holds}\ , .
\label{firsthalf}\ ] ] suppose that @xmath127 and @xmath128 .
then @xmath129 and so @xmath130 it follows from corollary [ dualcorollary ] that @xmath131 kats theorem then implies that @xmath132^{1 + \varepsilon } = o(1 ) \quad \mbox{as $ x \rightarrow l-$}\,.\ ] ] this is condition ( ii ) .
conversely , if condition ( ii ) holds , we arrive by a similar argument at the conclusion that @xmath127 and @xmath133 .
this proves ( [ firsthalf ] ) . to complete the proof of the theorem
, it suffices to replace @xmath8 by @xmath70 in ( [ firsthalf ] ) . the problem of characterising the _ determinate subclass _ in terms of the mass distribution appears to be much more difficult .
[ criterionremark ] let @xmath134 be the particular solution of equation ( [ impedanceform ] ) in the interval @xmath135 $ ] such that @xmath136 the problem considered by borcea is to recover @xmath137 , normalised by @xmath138 , from @xmath139 @xcite . now
, suppose that @xmath140 then @xmath139 is the characteristic function of a string @xmath8 such that @xmath141 it is an easy consequence of theorem [ stringtheorem ] that this string belongs to @xmath69 .
knowing @xmath8 , @xmath137 can in principle be obtained via equation ( [ impedance ] ) .
given the continued fraction coefficients @xmath48 , one can ascertain whether @xmath18 is in the determinate class by using stieltjes criterion ( [ stieltjesseries ] ) . to give a more concrete example , consider the case @xmath142 then @xmath143 the series @xmath144 diverges , and so @xmath18 is in the determinate class .
[ impedanceexample ]
we proceed to compute the string corresponding to the finite case of the continued fraction expansion on the right - hand side of equation ( [ katscontinuedfraction ] ) . in what follows
, it will always be assumed that @xmath145 we shall use the compact notation @xmath146 : = \frac{s_j}{-z } + \frac{1}{\displaystyle s_{j+1}+ \frac{1}{\displaystyle \frac{s_{j+2}}{-z}+ \cdots + \frac{1}{\displaystyle s_{j+2k-1}+ \frac{1}{\displaystyle \frac{s_{j+2k}}{-z}}}}}\ ] ] and @xmath147 : = \frac{s_j}{-z } + \frac{1}{\displaystyle s_{j+1}+ \frac{1}{\displaystyle \frac{s_{j+2}}{-z}+ \cdots + \frac{1}{\displaystyle s_{j+2k-1}+ \frac{1}{\displaystyle \frac{s_{j+2k}}{-z}+\frac{1}{\displaystyle s_{j+2k+1}}}}}}\ ] ] and an analogous notation for the corresponding strings .
these characteristic functions are in the determinate subclass of @xmath92 because their images under the map @xmath106 are in the determinate subclass of @xmath68 . for every @xmath148 , @xmath149 = \frac{1}{s_k}\ ] ] and , for every @xmath150 , @xmath151 ( x ) = \frac{m^{\ast}[s_{j+1},\,\ldots,\,s_{k}](t)}{1 + s_j m^{\ast}[s_{j+1},\,\ldots,\,s_{k}](t ) } \label{basicrecurrencerelation}\ ] ] where @xmath152(\tau ) \right \}^2\,d \tau\ , .
\label{xintermsoft}\ ] ] [ recurrencetheorem ] consider the case where @xmath153 is _
odd_. then @xmath154 = \frac{s_j}{-z } + \frac{1}{\displaystyle s_{j+1}+\frac{1}{\displaystyle \frac{s_{j+2}}{-z}+ \cdots+ \frac{1}{\displaystyle \frac{s_{k-1}}{-z}+\frac{1}{\displaystyle s_k}}}}\,.\ ] ] an easy calculation shows that @xmath155 = w[s_{j+1},\ , \ldots,\,s_k]\,.\ ] ] by taking the dual on both sides , we obtain @xmath156 = w^\ast [ s_{j+1},\ , \ldots,\,s_k]\,.\ ] ] it then follows from proposition [ zeromassproposition ] that @xmath157 ( x ) = \frac{m[s_j,\,\ldots,\,s_k](t)}{1 - s_j m[s_j,\,\ldots,\,s_k](t)}\,,\ ] ] where @xmath158(\tau ) \right \}^2 d \tau\,.\ ] ] when we `` turn this around '' and express @xmath159 $ ] in terms of @xmath160 $ ] and @xmath161 in terms of @xmath0 , we obtain the desired result . the case where @xmath153 is
even is analogous .
for every @xmath162 and every @xmath163 , @xmath159 $ ] is a discrete string .
let @xmath148 and proceed by induction on @xmath164 , starting with @xmath165 and going backwards .
the claim is obviously true for @xmath165 .
suppose that @xmath166 $ ] is a discrete string . by equation ( [ xintermsoft ] )
, @xmath0 is then an increasing continuous piecewise linear function of @xmath0 , and the fact that @xmath167 $ ] is piecewise constant is an obvious consequence of equation ( [ basicrecurrencerelation ] ) .
this result permits the efficient recovery of the discrete string @xmath168 $ ] via the sequence @xmath169 \rightarrow m [ s_{n-1},\,s_n ] \rightarrow \cdots \rightarrow m[s_0,\,\ldots,\,s_{n}]\ , .
\label{recurrencerelation}\ ] ] each term in that sequence is a right - continuous piecewise constant function with finitely many jumps , so we need only keep track of the location , say @xmath52 , of these jumps , and of the value , say @xmath170 , of the function there .
we always include @xmath171 and @xmath172 , whether or not @xmath173 .
we have @xmath174(x ) = \frac{1}{s_n } \;\;\mbox{for every $ x \ge 0$}\,.\ ] ] so we set @xmath175 also @xmath176(x ) = \cases { 0 & \mbox{if $ 0 \le x < \frac{1}{s_n}$ } \\
\frac{1}{s_{n-1 } } & \mbox{if $ x \ge \frac{1}{s_n}$ } } \ ] ]
so we set @xmath177 then , for @xmath178 , we have the recurrence relations @xmath179
@xmath180 ^ 2 \left [ y_j^{(2k-1)}-y_{j-1}^{(2k-1)}\right ] , \ ; 1 \le j \le k\ , .
\label{firstgeneralformulaforx}\ ] ] @xmath181 @xmath182 @xmath183 @xmath184 ^ 2 \left [ y_{j}^{(2k)}-y_{j-1}^{(2k)}\right ] , \ ; 1 \le j \le k\ , .
\label{secondgeneralformulaforx}\ ] ] @xmath185 @xmath186 @xmath187 the computation terminates when the superscript reaches the value @xmath60 .
in this section , we describe in broad terms the probabilistic version of krein s inverse problem expounded by knight in @xcite , and illustrate our inversion algorithm by means of an example taken from @xcite .
the reader unfamiliar with diffusion processes will find a useful summary of the theory in @xcite . consider a one - dimensional diffusion process @xmath37 started at the origin , with values in some interval @xmath188 .
suppose that the origin is instantaneously reflecting and that @xmath37 is in natural scale .
such a process is completely determined by its infinitesimal generator , i.e. by the differential operator @xmath189 acting on a suitable set of functions . here , @xmath190 is the speed measure , and the domain of @xmath191 consists of twice differentiable functions @xmath192 satisfying the condition @xmath193 ; an additional condition may be imposed on @xmath194 at the right boundary , depending on the behaviour of the diffusion there .
the process @xmath195 is called the _ local time _ ( at the origin ) ; it measures the time spent by the diffusion @xmath37 in the vicinity of the origin , up to time @xmath161 .
its right - continuous inverse @xmath196 is called the _ inverse local time _ ( at the origin ) ; it is a positive , non - decreasing process that jumps at the times when @xmath37 begins an excursion away from the origin .
the height of the jump is the _ length _ of the excursion , i.e. the time that elapses before @xmath37 returns to its starting point .
it may be shown that the inverse local time is in fact a lvy process . from the theory of such processes ,
one deduces the existence of a _
lvy exponent _
@xmath197 defined implicitly by @xmath198\;\;\mbox{for $ \lambda > 0$}\ , .
\label{levyexponent}\ ] ] furthermore , the lvy exponent is necessarily of the form @xmath199 for some numbers @xmath200 and some _ lvy measure _ @xmath201 , i.e. a measure such that @xmath202 knight shows that , if @xmath37 is a so - called `` gap diffusion '' , then the function @xmath8 appearing in equation ( [ infinitesimalgenerator ] ) may be identified uniquely with a string , still denoted @xmath8 .
the lvy exponent @xmath203 is related to the characteristic function @xmath18 of the string via @xmath204 furthermore @xmath205 krein s inverse problem may thus be rephrased as : `` find the gap diffusion , given the lvy exponent of the inverse local time '' . to give an example ,
let the lvy exponent be given by equation ( [ levykhintchineformula ] ) with @xmath206 and @xmath207 donati martin and yor showed that the corresponding diffusion is , up to a homeomorphism , a bessel process with drift @xcite .
the corresponding string may be expressed in terms of the modified bessel functions .
set @xmath208 where @xmath209 by using the well - known continued fraction expansion for the binomial ( see for instance @xcite , p. 343 ) , it is easy to see that @xmath18 is expressible in the form ( [ katscontinuedfraction ] ) with @xmath210 where @xmath211 denotes pochhammer s symbol and we use the convention @xmath212 for @xmath213 .
stieltjes criterion ( [ stieltjesseries ] ) implies that this continued fraction converges .
hence @xmath18 belongs to the determinate subclass of @xmath92 . in what follows ,
we find approximations of the string @xmath8 by truncating the continued fraction after @xmath60 terms , and computing the corresponding discrete string by the algorithm of the previous section .
take @xmath214 then @xmath215 and the diffusion is a brownian motion with drift @xmath216 in natural scale @xcite .
the dots in figure [ brownianwithdriftfigure ] are the points @xmath217 such that @xmath218 corresponding to the discrete string @xmath168 $ ] constructed by the algorithm of [ inversesection ] .
the superimposed continuous curve is a plot of @xmath8 .
( 0,0 ) ( 220,0)@xmath0 ( 75,120)@xmath8 [ brownianmotionwithdriftexample ] by taking @xmath219 and letting @xmath220 , we obtain @xmath221 the string with this characteristic function was found explicitly by donati - martin and yor @xcite : let @xmath222 be the @xmath44-valued diffusion process , reflected at the origin , with infinitesimal generator @xmath223 \frac{d}{dy}\,.\ ] ] the scale function of @xmath222 is @xmath224 the diffusion @xmath225 is in natural scale and its generator is given by equation ( [ infinitesimalgenerator ] ) , where @xmath8 is the string corresponding to ( [ logarithmicfunction ] ) .
figure [ besselwithdriftfigure ] shows a plot of the string for @xmath0 small , together with an approximation obtained by our algorithm .
( 0,0 ) ( 220,0)@xmath0 ( 75,120)@xmath8 [ alphaiszeroexample ]
we end with some comments on the numerical issues arising from the algorithm : by design , the approximation is _ exact _ in the limit @xmath226 .
more generally , our computations suggest that @xmath227 there is also some numerical evidence that greater accuracy may be obtained by averaging at the jumps , namely @xmath228 where @xmath229 as one would expect , in order to compute the discrete string from the truncated continued fraction , it is necessary to have good approximations of the coefficients
. it will seldom be the case that exact formulae are available . instead
, one will need to compute these coefficients by using some numerical algorithm ; see for instance @xcite and the comments therein . finally , the inversion method we have described generates only piecewise constant approximations of @xmath8 , and so the recovery of @xmath56 in cases where the string is absolutely continuous is not entirely straightforward .
it would be of interest to adapt to our case the ingenious recovery technique that borcea @xcite devised in the context of the sturm
liouville problem in impedance form .
it is a pleasure to thank alain comtet for his encouragement and for discussions of the material , and the laboratoire de physique thorique et modles statistiques , universit paris - sud , for its hospitality while some of this work was carried out .
thanks also to the anonymous referees , whose thoughtful comments and suggestions helped improve an earlier version of the paper .
dym h and kravitsky n. 1978 on recovering the mass distribution of a string from its spectral function , in _ topics in functional analysis ( essays dedicated to m. g. krein on the occasion of his 70th birthday ) _ , 45 - 90 , _ adv . math .
stud . _ * 3 * , academic press , new york . | the spectral data of a vibrating string are encoded in its so - called characteristic function .
we consider the problem of recovering the distribution of mass along the string from its characteristic function .
it is well - known that stieltjes continued fraction leads to the solution of this inverse problem in the particular case where the distribution of mass is purely discrete .
we show how to adapt stieltjes method to solve the inverse problem for a related class of strings .
an application to the excursion theory of diffusion processes is presented . |
low surface brightness galaxies ( lsbgs ) have been a subject of increasing interest in the last two decades .
although an initial study by freeman ( 1970 ) suggested that the central surface brightnesses of disk galaxies in the hubble sequence ( sa - sb - sc ) fell within a rather narrow range , @xcite pointed out that the lack of low surface brightness galaxies may in fact be a selection effect due to the difficulty in discovering galaxies of very low surface brightness , as had previously been recognized by @xcite . indeed , many surveys have since turned up large numbers of lsb galaxies . in practice , `` low surface brightness '' has come to mean galaxies whose central surface brightness is fainter than 22.0 mag arcsec@xmath8 in the @xmath9 band - pass ( i.e. more than @xmath10 outside of the narrow range defined by @xcite of @xmath11-mag / arcsec@xmath12 ) .
however , while many galaxies have been found below the `` freeman value '' of surface brightness , disk galaxies with surface brightnesses that are significantly higher than those of normal spirals do not appear to exist .
this is shown most clearly in @xcite , who find that there is a rather well - defined upper cut - off at a central surface brightness in @xmath13 of @xmath14 mag arcsec@xmath8 .
it was originally thought that all galaxies with a low surface brightness were early or late - type dwarfs @xcite .
however , radial velocity observations by @xcite showed that some galaxies with low surface brightness are actually quite large and luminous @xcite , for example lsbgs like malin 1 type .
indeed , most surveys have revealed lsbgs with disk scale lengths comparable to the milky way ( although these angular diameter limited surveys are strongly biased toward finding the physically largest galaxies ) . based upon surveys to date ,
lsbgs do seem to have different properties than those of brighter galaxies ( possibly indicating different evolutionary histories ) .
the observed trends suggest that lsbgs are in large part a continuous extension of the population of normal disk galaxies , reaching to higher angular momenta and lower masses @xcite .
previous studies @xcite showed that lsbgs are much more gas rich ( in terms of @xmath15 ) and bluer than `` normal '' late type galaxies @xcite .
when combined with their observed low gaseous metallicity @xcite , the evidence suggest that lsbgs have had a historically low star formation rate and are relatively unevolved systems compared to the bulk of normal spirals .
hi observations @xcite show that lsbgs have extended disks with low gas surface densities and high m / l ratios , further confirming the unevolved nature of lsbgs .
even external influences seem to have had little effect in speeding up the evolution of lsbgs @xcite . despite the tremendous progress in this field , there are still significant uncertainties in many of the above results .
first , while the data suggests a young stellar age for lsbgs overall , the spread in the integrated optical colors of lsbgs is large , suggesting that they may actually have traveled along diverse evolutionary paths . in particular , because of the low surface brightness of the underlying population , only a small fraction of the total number of stars is needed to make the colors significantly blue , and thus a small change in the star formation rate can drastically change the apparent mean age of the stellar population as inferred from _ optical _ observations .
second , while emission line measurements have revealed the metallicity of the current gas phase , they have not measured the metallicity of the underlying stellar population .
given the large gas reservoirs observed in lsbgs and the possibility of substantial late time gas infall , it is not clear that the current gas phase metallicity is closely related to the actual metallicity of the stars .
third , the total stellar mass in these galaxies ( and their contribution to dynamics ) is rather uncertain , given the large uncertainties in the stellar mass - to - light ratio at optical wavelengths . finally
, recent modeling of lsbgs predict that from the assumption that blue lsbgs are currently undergoing a period of enhanced star formation , there should exist a population of _ red _ , non - bursting , quiescent lsbgs @xcite .
these galaxies should then also be metal - poor and gas - rich , and share many of the properties of the lsbgs observed in the optical . in all of these areas
, we can greatly improve the observational constraints by undertaking a systematic study of the properties of lsbgs and normal spirals in the near - ir .
combined near - ir and ( eventually ) optical observations are a perfect probe to help to clarify the true stellar masses of lsbgs , to disentangle the roles played by the age and metallicity , and to understand their place in the continuum with normal high surface brightness ( hsb ) spirals .
we expect that contrary to initial impressions , lsbgs should be readily detectable in the near - ir .
although much of the data suggests that the mean stellar population of lsbgs are probably young , we know from the very red integrated colors of some lsbgs ( such as the giant malin - type objects ) that many lsbgs contain a significant _ old population of stars _ , and were therefore not formed recently . this old population should dominate the near - ir ( which is a good tracer of red giant and lower mass stars ) and thus should be detectable with nir observations .
we also should be able to easily identify lsbgs with bulges , as these are among the reddest members of the lsbg population @xcite .
there are also indications that the most massive lsbgs , such as the super - giant galaxies malin 1 and malin 2 , are redder than the majority of lsbgs , with luminosities that are comparable to those of normal spirals .
thus , near - ir observations should be particularly sensitive to the most massive end of the lsbg population . while valuable , previous studies of the ir properties of galaxies in the low surface brightness regime have been limited due to the difficulty of the observations .
@xcite observed a preliminary sample of lsbgs in @xmath0 , @xmath16 and @xmath17 , but the sample was very small and biased toward massive galaxies . @xcite
have observed 14 blue lsbgs in @xmath0 , @xmath16 and @xmath17 and combined with optical photometry , arriving to the result that many properties of lsbgs observed in the optical are reproduced in the near - ir .
in particular , they observe the same morphologies in @xmath9 and @xmath0 , measuring weak optical / near - ir color gradients , a fact which they interpret as a low dust content .
near - ir observations have also been used to probe the stellar population differences between lsbgs and normal spirals .
@xcite undertook a combined optical and ir study of 26 lsbgs , including 5 from his earlier thesis work in @xcite , and examples of both normal blue lsbgs and the rarer class of red lsbgs .
they compare lsbg average ages and metallicities with their physical parameters and find that blue lsbgs are well described by models with low , roughly constant star formation rates , whereas red lsbgs are better described by a `` faded disc '' scenario . among the larger population
, strong correlations are seen between an lsbg s star formation history and its @xmath17-band surface brightness , @xmath17-band absolute magnitude and gas fraction .
these correlations are consistent with a scenario in which the star formation history of an lsbg primarily correlates with its surface density and its metallicity correlates with both its mass and its surface density .
another larger study involving ir data of spiral galaxies was made by @xcite , but included only a very small number of true lsbgs .
the lack of much previous work , and the availability of relatively large ir array detectors , both open the possibility to systematically observe a large sample of lsbgs and normal spirals in the near - ir , to derive fundamental parameters related to the origin of these galaxies and to their relation to normal galaxies . based upon the large expected rewards and the limited existing data , a systematic near - ir study of lsbgs is needed in order to complement existing optical data .
we undertake such a study in this paper , which is organized as follows . in [ samplesec ]
we present the sample used in this work .
[ obssec ] briefly describes our observations , data reductions , photometric calibrations , galaxy photometry , and surface photometry . in
[ resultsec ] we analyze and discuss the most important results , including galaxy photometry , central surface brightnesses and scale lengths , the relationship between hi - mass and near - ir luminosities , and the relationship between stellar mass - luminosity ratios and baryonic mass .
finally , we conclude in [ conclusionsec ] .
all galaxies were initially selected from the catalogue of large angular size low surface brightness galaxies compiled by @xcite from the apm plates , consisting of 516 lsbgs with effective radius larger than @xmath18 arcsec .
some constraints were applied to optimize the galaxies observability and their intrinsic stellar and gas content .
first , a surface brightness cut was applied to ensure detectability at the 40-inch swope telescope and at the 100-inch du pont telescope , both located at las campanas observatory ( lco ) . the corresponding central surface brightness cut - off
was set to @xmath19 mag arcsec@xmath8 for the galaxies observed with the 40-inch telescope , and to @xmath20 mag arcsec@xmath8 for the galaxies observed with the 100-inch telescope .
these limits are set considering ( 1 ) the corresponding limiting @xmath1 magnitudes for the 40-inch and the 100-inch telescopes both equipped with a nicmos3 hgcdte array : 19.0 mag and 20.5 mag , respectively , and ( 2 ) the observed average color @xmath21 for lsbgs @xcite .
we applied additional selection criteria in declination , @xmath22 , in order to observe galaxies from lco .
we then randomly sampled from the resulting catalog to generate a sample which would be observable during the one year of allotted telescope time . finally , we restricted the sample to two ranges in hi mass , selecting galaxies with either log@xmath23 \le 9.0 $ ] ( sample a ) or log@xmath23 \ge 9.5 $ ] ( sample b ) .
this last selection criterion allows us to capture lsbgs in the maximum range of evolutionary states ( i.e. high and low gas - to - star ratios ) and total masses .
there are 112 galaxies from the @xcite catalogue satisfying the above constraints . of these
, 65 galaxies were observed with the 40-inch telescope and 47 galaxies with the 100-inch telescope . in order to have an optimal wavelength baseline , we observed all galaxies through both the @xmath0 and @xmath1 near - infrared filters ( the latter defined by @xcite ) .
due to the critical importance of accurate surface brightness measurements , and given the low surface brightness of these galaxies , we chose not to use the @xmath16 band . the strong and variable
oh emission which dominates the sky background at @xmath16 is not easy to remove , and could significantly affect the apparent surface brightness of the galaxies .
in addition , prominent fringing appears in the @xmath16 band , putting additional hurdles to deriving the correct surface brightness flux .
the @xmath0 band , on the other hand , is stable during time scales of @xmath24 minutes , and the @xmath1 bands is stable enough typically for time scales shorter than 5 minutes , when the outside temperature is below 15@xmath25c @xcite . unlike many earlier near - ir studies of nearby galaxies , we have sufficient sky background in our images to do a reasonable job of sky subtraction .
most galaxies are much smaller than the @xmath26 arcmin fov of both telescopes , with just a few having extents larger than 1.5 arcmin in the @xmath9 band .
all galaxies were observed by gg and ez during photometric nights at lco , using the 40-inch swope telescope and the 100-inch du pont telescope .
seven runs of 3 to 7 nights were allocated for this project .
observations started on february 23rd 1999 and finished on october 23rd 1999 .
a nicmos3 hgcdte array was used in both telescopes , but with different pixel scales : 0.60 arcsec / pix at the 40-inch , and 0.348 arcsec / pix at the 100-inch .
the corresponding fov were @xmath27 and @xmath28 .
observation procedures were the same as used in @xcite , and consist of a sequence of short dithered exposures for all the scientific targets , including the standard stars .
due to the low surface brightness which characterizes these galaxies , exposure times were calculated based on the known blue surface brightness , extrapolated to the @xmath0 and @xmath1 band , and corrected for the additional contribution of the thermal noise and sky emission .
typical total exposure times were between 20 min and 45 min in @xmath0 and between 40 min and 75 min in @xmath1 .
considering the low surface brightness of galaxies , special care was taken with flat - fielding , in order to avoid undesirable artificial spatial fluctuations .
a large series of dome - flats were taken for each run , as well as sky - flats every sunset .
additionally , the raw images of each night were used to construct super - flats , by combining dithered images , using a median sky value obtained after a sigma clipping algorithm to discard cosmic rays , bad pixels , and any residual contribution from the galaxies themselves .
usually , no fewer than 80 images per filter were used to construct nightly super - flats , and only images where the galaxies were small compared the chip size were included in the image stack .
afterward , dome - flats , sky - flats and super - flats were compared .
the largest differences between sky and dome - flats were @xmath29 . for image reduction
, we used a modified version of dimsum .
the reduction procedure is completely analogous to that described in @xcite , to which the reader can refer for details . out of the 112 original selected lsbgs ,
we successfully observed 88 ; 39 from sample a ( low hi mass ) and 49 from sample b ( high hi mass ) . in order to derive color information we calibrated our magnitudes system using the standard star system defined by @xcite . for the standard star photometry iraf
s `` daophot '' software was used .
we adopted a 10 arcsec aperture radius for all the standard stars , which corresponds to the same aperture used by @xcite .
typical seeing was 1.0 arcsec .
no less than 6 standards were measured each night , which allow us to derive calibrated @xmath0 and @xmath1 magnitudes .
the transformation equations were @xmath30 where capitals denote standard magnitudes , small characters denote instrumental magnitudes , @xmath31 and @xmath32 are the zero points , @xmath33 and @xmath34 are the color terms , and @xmath35 and @xmath36 are the extinction coefficients in @xmath0 and in @xmath1 respectively .
@xmath37 is the mean airmass of the observation .
we have computed the transformation parameters of equations ( 1 ) and ( 2 ) using `` photcal '' , included in iraf . when measuring the above coefficients , color terms and zero points were initially computed independently for each night , verifying that they changed by only a small amount ( less than 5% ) from night to night
; because these terms depend only upon the telescope / mirror - instrument - filter combination , they should be constant over each run .
we then adopted a single color term and zero point for the entire run ( based upon the night - to - night average ) , and re - computed the extinction coefficients for each night separately .
the values obtained were very close to the expected values @xmath38 and @xmath39 in @xcite .
the resulting average values typical of all our runs are shown in table [ calib ] .
the typical color term is relatively small , and given the observed values of @xmath40 ( between 0.10 and 0.80 ) , the contribution of this term is @xmath41 in @xmath0 , and therefore may be marginally important when computing the spatial distribution of colors through a given galaxy . given the heterogeneous luminosity profiles of the lsbgs , several different techniques were initially explored to optimize measurement of the apparent magnitudes : ( 1 ) circular apparent magnitudes , using daophot and the growth curve method , ( 2 ) total magnitudes within elliptical apertures , computed using sextractor @xcite .
sextractor also computes isophotal areas which allow one to compute mean surface brightness within a specified isophote level .
a comparison between these two techniques allows us to conclude that a more robust measurement of the flux is obtained with sextractor magnitudes , given the elongated shape of many of the lsbgs in the sample ( see figures [ mosaics1 ] and [ mosaics2 ] ) . in lsbgs with small ellipticities , daophot and sextractor magnitudes differ by @xmath42 . in what follows
, we use only magnitudes computed with sextractor on elliptical apertures .
magnitudes were determined within an isophotal level set by the quality of the data , at a 3@xmath43 level above the sky ; the resulting sizes of the elliptical apertures are listed in table [ phot1 ] .
the major uncertainty arises from the sky level , with a mode of @xmath44% in @xmath0 and @xmath45% in @xmath1 .
we have attempted to compare our derived magnitudes with those from the 2mass survey , but found no overlap in the galaxies studied .
table [ phot1 ] shows the @xmath0 and @xmath1 magnitude and color , for the observed lsbgs . the heliocentric radial velocity from the 21-cm line , the size of the elliptical aperture used for photometry , and the hi mass content from @xcite are also shown .
the absolute magnitudes @xmath46 in table [ phot1 ] were computed as @xmath47 - 25 \label{abs_magnitude}\ ] ] where @xmath48 is the apparent magnitude , @xmath6 is the hubble constant expressed in km sec@xmath7 mpc@xmath7 ( we use @xmath49 km sec@xmath7 mpc@xmath7 in this paper ) , and @xmath50 is the virgocentric peculiar velocity in km / s . throughout this paper
we use @xmath51 km / s , as quoted by @xcite .
this corrective term is important for galaxies with @xmath52 km / s .
it is worth noting that this correction _ does not _
depend on passband , and therefore does not affect colors .
equation [ abs_magnitude ] is a good approximation up to a recessional velocity @xmath53 km sec@xmath7 , where cosmological corrections become important and the parameter @xmath54 must be taken into account ( from table [ phot1 ] we see that our largest velocity is @xmath55 km sec@xmath7 ) . in equation [ abs_magnitude ]
we have neglected @xmath56-corrections . at the lsbgs redshifts @xmath56-corrections are smaller than our photometric errors , and therefore are negligible .
note also that heliocentric radial velocity is measured from the hi emission ( table [ phot1 ] ) , slightly different from the optical heliocentric radial velocity .
however , differences are smaller than 100 km / sec , and therefore changes in absolute magnitudes are no larger than 0.05 mag .
as can be seen from table [ phot1 ] , lsbgs do not necessarily have faint absolute magnitudes in the near - ir . instead ,
some of them appear to be as bright in the near - ir as normal high surface brightness galaxies .
this immediately suggests that they do have a significant population of old stars .
measurements of surface brightness and scale lengths give critical information on the light density and physical size of galaxies .
we now use surface brightness fitting to measure the central surface brightness ( as an indicator of the central light concentration ) and scale length of both the disk and bulge components of the galaxies in our sample .
the reader must recall that the surface brightness difference we measure for the same extended source located at different redshifts @xmath57 is given by @xmath58 where @xmath59 is expressed in mag arcsec@xmath8 . in our case ,
our smallest radial velocity is @xmath60850 km sec@xmath7 ( @xmath61 ) , and our farthest galaxy is located at @xmath62 km sec@xmath7 ( @xmath63 ) .
this implies that the surface brightness difference for an object with a given surface brightness at @xmath64 , but measured at @xmath65 is , at most , @xmath66 mag arcsec@xmath8 .
however , at @xmath67 level around the average redshift , we have @xmath68 km sec@xmath7 and @xmath69 km sec@xmath7 , which yields @xmath70 mag arcsec@xmath8 , which is inside the magnitude error bars .
therefore our surface magnitude estimates are not strongly modified by the range of distance included in the sample .
we adopt the same approach as given by @xcite , and fit our data to the following parametric profile @xmath71 or in terms of magnitudes per square arc - second @xmath72 in equations [ profile_int ] and [ profile_mags ] @xmath73 and @xmath74 are the central surface brightness , the latter measured in mag arcsec@xmath8 .
the parameter @xmath75 is the scale length , measured in arcsec and transformed to physical units ( kpc ) .
similar to @xcite , we initially fit bulges using de vaucouleurs profiles @xcite .
however , the profile was not a good match to the bulge light profile , and poor fitting results were obtained . following @xcite , we then switched to fitting the bulge with an exponential light profile as well , yielding much better fits .
an example bulge and disk fit is presented in figure [ lsb059_profiles ] .
the double exponential profile provides an adequate fit to the data over a large range in radius , though slight deviations suggest that an intermediate srsic profile for the bulge might provide a slightly improved fit .
we find that most of the galaxies require a bulge component in the near - ir ( i.e. they show substantial curvature in the inner parts of a @xmath76 vs @xmath77 plot ) , but some do not and represent completely disk dominated galaxies .
a mosaic with reduced and combined @xmath0 images is shown in figures [ mosaics1 ] ( sample a , low hi mass lsbgs ) and [ mosaics2 ] ( sample b , high hi mass lsbgs ) .
typical image sizes are @xmath78 arcmin .
both figures [ mosaics1 ] and [ mosaics2 ] and our surface brightness decompositions reveal a number of significant results for the morphologies and surface brightnesses of our sample .
first , many of the lsbgs classified as late types based upon the optical @xmath9 band image instead show a dramatic bulge in the near - ir ( confirming the presence of prominent near - ir bulges noted by @xcite ) .
this result differs from that obtained by @xcite and @xcite , who found similar morphologies in the @xmath9 and @xmath0 bands , but for a much smaller sample of galaxies .
examples of galaxies with different optical and near - ir morphologies ( such as in the presence of a bulge in the near - ir ) are the cases of lsbg # 100 and lsbg # 336 ( for example , their @xmath0 images in figure [ mosaics1 ] ) .
the fact that some lsbgs have high surface brightness bulges ( @xmath79 mag arcsec@xmath8 ) has been commented by @xcite , who noted that many lsbgs could be regarded as high surface brightness bulges embedded in low surface brightness disks .
our results support such a hypothesis . in other words , while some galaxies have been classified as lsbgs on the basis of their _ disk _ properties , they are not low surface brightness overall .
this difference is particularly striking in the near - ir , where the surface brightness more closely reflects the stellar surface density .
moreover , as discussed by @xcite , many of the galaxies in the catalog from which our sample are drawn are not true lsbgs at all , and are instead indistinguishable from normal spirals in the optical .
we find a similar result for our near - ir data .
second , in figure [ mj_mu0 ] we plot the infrared luminosity of the galaxies vs the central surface brightness of the disk component .
the figure shows that there is no correlation between these quantities and that almost all of the absolute magnitude vs surface brightness parameter space is covered .
the upper left part of the distribution ( large , luminous but lsb galaxies ) is as populated as the the lower right part of the distribution ( small , quite compact galaxies ) . in our sample
there is no bias against large and faint , nor small but luminous galaxies .
we also observe in figure [ mj_mu0 ] a sharp cutoff towards bright surface brightnesses ( at @xmath80 mag arcsec@xmath8 ) .
this cutoff represents a real physical cutoff ( i.e. nature does not make disks with higher surface brightnesses ) , rather than a selection effect propagating from an optical surface brightness cutoff in the @xmath9 band . if it were the latter , then we would not expect the surface brightness of the cutoff to be constant with the near - ir absolute magnitude , given the strong dependence of the optical to near - ir color with galaxy luminosity .
this physical cutoff further supports our belief that our sample does contain examples of galaxies with normal disk surface brightnesses , in addition to examples with disk surface brightnesses over 3.5 magnitudes fainter , thus giving us a true continuum of galaxies from normal disks to extremely low surface brightness systems .
third , we have also included in figure [ mj_mu0 ] a plot of the dependence of the bulge central surface brightness on the near - ir absolute magnitude . unlike for the disks
, we do see a correlation when _ overall _ surface brightness is considered ( with a spearman test suggesting that less than a 10% chance that they are uncorrelated for the low hi mass subsample ) .
in addition , while figure [ mj_mu0 ] shows a cutoff at bright surface brightnesses for the disk component , the bulges in these same galaxies extend to much higher surface brightnesses ( with a possible , but less sharp cutoff at @xmath81 mag arcsec@xmath8 ) . for a comparison with the optical
see figure 4a from @xcite .
fourth , we compare the sizes ( in kpc ) and surface brightnesses of the bulge ( left ) and disk ( right ) components in figure [ lsb_h_mu0_bd_j ] . it is apparent that lsbgs of the high hi mass subsample b ( open circles ) have brighter bulge surface brightnesses but similar bulge sizes , compared with lsbgs of the low hi mass subsample a ( filled circles ) .
a different trend , however , is observed for disks ( right panel ) . in this case , galaxies of sample a have systematically smaller disk scale lengths with respect to those of sample b , but with little change in typical surface brightness .
therefore , galaxies of sample b have large disks and very high surface brightness bulges .
those galaxies with lower hi masses have small disks , and bulges that are structurally similar to those of the high hi mass sample , but that have very low surface brightness .
the same behavior can be seen in the @xmath1 band ( however we have better signal - to - noise in @xmath0 ) . rather than just comparing our two samples against each other , we can also compare morphology with hi content directly , by examining the bulge and disk structural parameters derived above as a function of hi mass . in figure [ lsb_mass_mu0_j ]
the relationship is shown for the bulge and for the disk central surface brightness , in the left and the right hand panels , respectively .
it is clear that the galaxies with low hi masses are almost entirely lacking in high surface brightness bulges , leading to a much smaller dispersion among the bulge central surface brightnesses .
in contrast to the trends in bulge central surface brightness , there are no strong trends seen in the disk surface brightnesses .
instead , there is a very large dispersion among the disk properties , independently of the gas mass of the galaxies .
note that in all the above analyses using the near - ir bands , we are sampling nearly _ the same _ stellar populations both in the bulge and in the disk , namely red giants and old stellar populations .
thus , the relative sizes and surface brightnesses reflect true differences in the distribution of stellar density .
note also that these measurements of the current stellar content are not _ a priori _ related to the remaining hi content after star formation .
however , the strong trends seen in figure [ lsb_h_mu0_bd_j ] and figure [ lsb_mass_mu0_j ] strongly suggest that the stellar surface density of a galaxy is indeed tightly coupled to the amount of neutral gas the galaxy currently has , though probably indirectly through the total galaxy mass , as we now demonstrate below . in the above analyses
we have begun to consider how some properties vary as a function of the total hi mass .
however , it is possible ( if not likely ) that the hi mass is not the sole controlling quantity driving these trends .
it is quite possible that the hi mass ( m@xmath4 ) correlates closely with the overall galaxy mass i.e. that `` bigger '' galaxies have more of everything . in order to begin examining this possibility , figure [ mass_mag ]
shows the strong correlation we observe between the hi mass and the near - ir absolute magnitudes . indeed , to first order , galaxies which have more stars ( as measured by the near - ir luminosity ) have more hi .
note that there are @xmath45 magnitudes between the most hi massive galaxies and the less massive ones .
recall the striking fact that this correlation ( which is well defined in both @xmath0 and @xmath1 ) is between the mass of the _ hi gas _ , and the near - ir luminosity of the _ stellar populations _ ; _ a priori _ there should not be a correlation between the mass in stars a galaxy has formed and the amount of gas it has left . in figure
[ mass_mag ] , straight lines represent constant values of the ratio of gas mass to stellar luminosity for m@xmath4/l = 0.01 ( long - dashed line ) , 0.1 ( solid line ) , 1.0 ( dotted line ) , and 10.0 ( dashed line ) . note that most lsbgs of our total sample have m@xmath4/l @xmath82 1.0 both in @xmath0 and @xmath1 , with an _ average _ of m@xmath4/l @xmath83 and large scatter .
while the scatter is over an order of magnitude in relative proportion of stars and gas , it is much smaller than the overall range of 8 orders of magnitude in overall gas mass , and thus does not wipe out the correlation between hi mass and total stellar luminosity .
the overall trend in figure [ mass_mag ] suggests that the division between our samples a and b is largely one mass , albeit with a wide range of gas mass to stellar luminosity ratios .
we may further examine our sample for trends in how the gas mass to stellar luminosity ratio varies as a function of the overall galaxy luminosity .
figure [ m_mlratio ] shows the logarithmic gas mass to stellar luminosity ratio as a function of @xmath0 and @xmath1 absolute magnitudes ( where we have adopted @xmath84 and @xmath85 ) .
although there is considerable scatter , the lower luminosity galaxies have a higher fractional gas content , on average .
the straight lines in figure [ m_mlratio ] are least square fits , giving slopes of @xmath86 in @xmath0 and @xmath87 in @xmath1 . the gas mass to stellar luminosity ratio as a function of the absolute magnitude ( figure [ m_mlratio ] ) indicates the relative importance of gas and stars in the total baryonic mass budget ( albeit indirectly ) .
galaxies with high @xmath88 in the near - ir have been historically less efficient at converting their gas into stars , due to either young mean age ( i.e. a late start ) , low star - formation efficiency , large late - time gas infall , fewer past episodes of hi - stripping , or some combination of the above .
the fact that we see some correlation between @xmath88 and the galaxies luminosities suggest that overall there are indeed trends between galaxies star formation histories and their stellar mass .
however , the large scatter suggests that it is not a perfect one - to - one relationship .
we will address these issues in more detail in subsequent sections , where we attempt to reduce the scatter by calculating the actual stellar mass , rather than using the near - ir luminosity as a substitute .
another possible quantity which correlates with m@xmath4 is the galaxy morphology .
the study of morphological types shows a modest correlation with hi content . within our sample , a set of lsbgs with similar hi masses
have morphological types spanning the entire hubble sequence from dim to sa . however , despite the large scatter , a mean trend does exist for morphological types over the whole hi mass interval .
galaxies of sample a ( with low hi masses and , via figure [ mass_mag ] , low stellar luminosities ) exhibit morphological types of dim , sm , sd , and only a few earlier type sc and sb .
only three cases are sa .
in contrast , lsbgs of sample b are much earlier type , and are classified in general as sc , sb and sa , although some cases of sm are also present ( see figures [ mosaics1 ] and [ mosaics2 ] , and text below ) .
much stronger trends are seen between the morphology of the galaxies and their gas mass to stellar luminosity ratio .
figures [ sort1_ml ] and [ sort2_ml ] show images of the galaxies sorted by their m@xmath4/l@xmath89 values .
the range in @xmath90 goes from m@xmath4/l@xmath89 = 0.028 - 0.128 @xmath91 in figure [ sort1_ml ] ( from lsbg # 036 with log(@xmath92 ) = 8.74 , to lsbg # 437 with log(@xmath92 ) = 7.92 ) , and continues up to m@xmath4/l@xmath89 = 1.475 @xmath91 in figure [ sort2_ml ] ( galaxy lsbg # 446 with log(@xmath92 ) = 9.63 ) .
it is apparent in these two figures that as m@xmath4/l@xmath89 becomes larger , galaxies extend to much later hubble types , beyond the text - book hubble sequence .
galaxies with small m@xmath4/l@xmath89 are earlier sa / sb / sc types , while galaxies with large @xmath90 tend to be irr galaxies , suffer from tidal distortions , plumes and other peculiarities . in previous sections
we have considered trends as a function of the hi mass to stellar luminosity .
this ratio gives a measure of the evolutionary state of the galaxies , indicating how far the galaxy has proceeded along the path of converting gas into stars .
however , the infrared stellar luminosity is only an approximate indicator of the stellar mass . we may obtain clearer and more physically meaningful results by using stellar population synthesis codes to translate the stellar luminosities into estimates of the actual stellar masses , allowing us to make a direct comparison between the gaseous and stellar mass . to convert from ir luminosity to stellar mass
, we may take advantage of the work of @xcite , which demonstrated that although the stellar mass - to - light ratio varies by up to a factor of 2 in the near - ir , there are clear trends between the stellar m / l and color . using spectrophotometric spiral galaxy evolution codes , @xcite showed that one can use the colors of galaxies to derive their relative stellar mass - to - light ratios ( under the assumption of a universal spiral galaxy initial mass function ) .
they also showed that uncertainties in the dust - reddening estimates do not strongly affect the final derived stellar masses of a stellar population . for testing whether the variations of m / l with @xmath40 color are robust to different stellar population synthesis models ,
we have used @xcite s relations between the stellar m / l and the @xmath93 and @xmath94 colors , to derive the relationship between m / l and our @xmath95 color and @xmath1 are less than @xmath96 mags . ] .
three different models are tested , which are ( 1 ) an instantaneous burst with a scalo imf , from recent bruzual & charlot models @xcite , ( 2 ) a recent version of pegase models @xcite with an imf slope of @xmath97 , and a model with a formation epoch with bursts , adopting a salpeter imf .
all models have solar metallicity .
the key point is that m / l vs color slopes do not depend on the imf , which only affects the zero point .
the stellar masses do not vary more than 7% between the two models .
we have also used sub - solar metallicities and again the variation in the stellar masses are @xmath98 , and consistent with the variations between models . in what follows
, we adopt the salpeter model ( using table 1 from @xcite ) and obtain coefficients for m / l as a function of @xmath40 color . with the adopted relationship between stellar mass to light ratio and @xmath3 color , we can compute the _ actual _ baryonic masses @xmath99 from the derived stellar masses @xmath100 ( @xmath101 ) and the hi mass as @xmath102 .
we are implicitly assuming that the mass contribution from molecular , ionized , and metallic gas is negligible for our sample .
figure [ mstars_relations ] shows the gas - to - star mass fraction as a function of stellar mass , after computing the stellar masses as described above .
figure [ mstars_relations ] clearly shows that lsbgs with larger stellar masses tend to have smaller gas fractions , and lsbgs with small star masses have larger gas fractions .
we have also included in figure [ mstars_relations ] diagonal lines representing constant hi mass ; dashed line for @xmath103 m@xmath104 ( close to our lowest hi mass ) , solid line for 9.0 m@xmath104 ( upper limit of our sample a ) , dotted line for 9.5 m@xmath104 ( lower limit for our sample
b ) , and long - dashed line for 10.0 m@xmath104 ( close to our largest hi mass ) . these lines can be used for assessing the importance of selection criteria .
for example , there seems to be a lower limit for the mass fraction hi - to - stars at @xmath105 .
however , these lsbgs can exist but their hi would not be detectable , and thus they may not be present in the subsample of galaxies with measured hi fluxes that we selected from the @xcite catalogue . figure [ mstars_relations ] also confirms our conclusions ( 1 ) that low hi mass does not imply that a galaxy has `` used up '' its gas reservoir and converted it entirely to stars ( given the existence of galaxies with low @xmath106 and high @xmath107 ) and ( 2 ) that likewise high hi mass does not imply that a galaxy has only begun to convert hi into stars ( given the large numbers of galaxies with large hi masses but large stellar masses as well ) .
in other words , both high hi mass galaxies _ and _ low hi mass galaxies have a large range of stellar masses ( with low hi mass galaxies having stellar masses between @xmath108 and @xmath109 m@xmath104 , and high hi mass galaxies having stellar masses between @xmath110 and @xmath111 m@xmath104 ) , and thus selecting on hi alone is not sufficient to identify galaxies which are gas - rich or gas - poor . we have also plotted similar relations as in figure [ mstars_relations ] , but using the ratio of the hi mass to the total baryonic mass ( @xmath112 ) , shown in figure [ mbaryonic_relations ] . note that the ratio m@xmath4/m@xmath5 has an upper limit , when all the baryonic mass is concentrated in the gas ( ratio = 1 ) .
we see again that increasing baryonic mass implies that a galaxy has converted a much larger fraction of its baryonic mass into stars .
we have also included in figure [ mbaryonic_relations ] the locus separating galaxies according to their dominant component , stars or gas . again ,
note that `` star - dominated '' lsbgs can still have large quantities of hi , and similarly , `` hi - dominated '' lsbgs can have a small hi - mass if the overall baryonic mass is small .
the extreme galaxy at m@xmath113 m@xmath104 is lsbg 224 which , even with a large hi content ( @xmath114 ) , has a baryonic mass represented almost exclusively by the stars .
it is clearly a prominent sa galaxy ( see figure [ sort1_ml ] ) . in general , all of the star - dominated lsbgs are sa / sb galaxies , whereas the hi - dominated lsbgs have morphologies ranging from spirals to distorted or fuzzy irregulars . among the most gas rich systems we find the most irregular and fuzzy lsbgs ( hi - rich galaxies ) in our sample .
note that , although the number of galaxies from sample a and b is roughly similar , most of the galaxies lie below the dotted horizontal line , and therefore are star dominated ones .
if we investigate these morphological trends further , we discover that both the stellar mass and the stellar - to - hi mass fraction are related closely to surface brightness .
figure [ mu0_relations ] shows the bulge central surface brightness as a function of the ratio m@xmath4/m@xmath5 , and as a function of the total baryonic mass . the ratio m@xmath4/m@xmath5 increases with decreasing surface brightness , showing that at higher bulge surface brightnesses , galaxies have produced a much larger fraction of stars over their lifetime .
it may be that the formation of a bulge creates an extremely efficient epoch of star formation , as gas is driven to very high central densities . given the observed correlations between gas density and star formation rate @xcite , then we would expect that galaxies with the highest stellar densities ( i.e. the high surface brightness bulges ) had the highest past star formation rates , and thus the lowest current gas fractions .
in other words , all galaxies are `` born '' at the locus of m@xmath4/m@xmath115 , and then travel downwards and to the left in figure [ mu0_relations ] to m@xmath4/m@xmath116 , at a rate which depends in large part upon the central surface density of the gas .
the end result is a lack of high surface brightness galaxies with large gas mass fractions , and low surface brightness galaxies with low gas mass fractions ( e.g. the region around @xmath117 mag arcsec@xmath8 and m@xmath4/m@xmath118 ) .
we may explore these ideas more directly by converting the observed surface brightnesses to stellar surface densities , using the color - dependent stellar mass to light ratios calculated above . in figure [ stellar_density ]
we have plotted the @xmath40 color and gas to baryonic mass ratio as a function of the central stellar surface density of the bulge @xmath119 ( in m@xmath104/pc@xmath120 ) .
we see a very strong trend ( correlation coefficient of 0.22 ) towards lower gas fractions with increasing central surface densities , as expected for increased star formation efficiency at higher surface densities .
the scatter is also reduced over figure [ mu0_relations ] .
in addition , we see a trend towards redder colors with increasing surface density , but we will defer a discussion of this effect until [ colorsec ] below . while the increase in m@xmath4/m@xmath5 is a steady function of the bulge surface brightness , an equally striking trend is found with the the total baryonic mass .
figure [ mu0_relations ] shows that the formation of a high surface brightness bulge is a very sharp function of the overall baryonic mass .
galaxies with baryonic masses of less than m@xmath121 simply do not form bulges ( using @xmath122 mag arcsec@xmath8 , marked by a dash - dotted vertical line in figure [ mu0_relations ] , as the demarcation of galaxies which do not have bulges significant enough to be detectable in the near - ir , i.e. those which can not be distinguished clearly from the disk ) mag arcsec@xmath8 , we find mostly irregular morphologies , with a larger tendency toward showing spiral morphologies when dominated by stars ( i.e. low m@xmath4/m@xmath123 ) . ] .
the floor in the baryonic mass suggests that there is a tight coupling between the mass of the galaxy and the conditions necessary for bulge formation .
we can envision such a correlation arising as follows .
assuming that there is no large scale segregation between baryons and dark matter , then the baryonic mass of the galaxy should be directly proportional to the total mass of the galaxy ( m@xmath124/m@xmath125 , for @xmath126 and @xmath127 ) . because the total mass of a galaxy is closely coupled to several astrophysical timescales for galaxy formation ( such as the epoch of first collapse , or the end of an epoch of merging )
, one might expect to see a correlation between _ tracers _ of these quantities and the baryonic mass .
in many scenarios , the bulge forms via a by - product of early merging in the galaxy , or via the collapse of low angular momentum material in the galaxy both of which probably depend closely on the above timescales .
thus , it may be that future models of galaxy formation can use this observed floor for bulge formation as a critical test .
the final question arises as to whether or not selection effects play a role in shaping the trends in figure [ mu0_relations ] . for example
, the lack of galaxies with high surface brightness bulges in the near - ir and low baryonic masses ( i.e. the dashed oval ) may be due to the fact that our sample was drawn from a catalog of low surface brightness galaxies .
such a catalog would be biased against bright bulges , unless accompanied by a low surface brightness disk , which would increase the overall mass of the galaxy , possibly pushing it up above the @xmath128 floor .
however , we suspect that the floor in figure [ mu0_relations ] is real , for several reasons .
first of all , the existence of large numbers of galaxies with @xmath129 mag arcsec@xmath8 suggests that the sample by no means excludes galaxies of high near - ir surface brightness galaxies .
indeed , an analysis by @xcite shows that the original @xcite sample includes a large fraction of high _ optical _ surface brightness galaxies , indistinguishable from `` normal '' spirals .
second , if the addition of a low surface brightness disk were necessary for inclusion in the original impey catalog , there is no reason why the lower left of the plot should not be populated with low mass galaxies that would otherwise have dropped off the bottom of the plot . finally , the observational selection effects would be most apparent in a plot of purely observational quantities ( for example , @xmath130 vs @xmath131 ) , but in such a plot , we see the plane is evenly populated ( e.g. figure [ mj_mu0 ] ) , and we can identify no steps in the process of converting from @xmath130 to m@xmath5 that would lead to the creation of an artificial floor in the baryonic mass associated with high surface brightness bulges .
we stress that the fact that no gas poor but low surface brightness galaxies with very low values of m@xmath4/m@xmath5 and faint surface brightness appear in the figure ( @xmath132 mag arcsec@xmath8 and m@xmath4/m@xmath118 ) , which could be a combined effect of our initial selection on surface brightness and gas content ( @xmath133 mag arcsec@xmath8 ) , even though we have not imposed a low cut for the hi mass . we know that large galaxies with bright total magnitudes but extremely low surface brightness and large disk scale lengths do exist , like malin 1 ( with @xmath134 , @xmath135 mag arcsec@xmath8 and @xmath136 , @xcite ) and
are placed at such a locus in the diagram , where spirals with faint bulges and large stellar masses lie .
however , figure [ mu0_relations ] suggests that these galaxies are rare . in the previous section
we have explored trends involving m@xmath4/m@xmath5 as an indicator of the evolutionary state of a galaxy .
if a galaxy steadily transforms its initial reservoir of gas into stars , then as m@xmath4/m@xmath5 decreases , the mean metallicity of the stellar population should increase as well . in this section
we explore metallicity trends within our sample , after demonstrating that the near - ir @xmath3 color is an excellent indicator of the mean stellar metallicity . within optical wavelengths ,
the variations in the broad band colors of galaxies are thought to be primarily driven by changes either in the mean age of the stellar population or in the dust content @xcite .
however , in the near - ir the situation is quite different .
first of all , reddening due to dust is negligible in all but the most extreme starburst galaxies .
secondly , and most importantly for this analysis , changes in the near - ir color of a galaxy are driven almost entirely by the mean _ metallicity _ of the galaxy ; in the near - ir , the light from a galaxy is dominated by stars on the red giant branch , the location of which shifts to redder or bluer colors depending on metallicity . to demonstrate this effect , in figure [ gridplot ] we plot stellar population synthesis models from bruzual & charlot @xcite of the @xmath40 color as a function of @xmath137 , for different metallicities and exponentially declining or increasing star - formation rates ( i.e. with different mean stellar ages ) , assuming that star - formation started 12 gyr ago .
the @xmath40 color index is clearly _ very _ metal sensitive , and relatively age insensitive ; see @xcite . at a single stellar metallicity ,
the change in @xmath3 with mean stellar age is never more than @xmath138 , even when the mean age of the stellar population changes by 11 gyr .
this implies that whenever one witnesses a change of more than @xmath139 then one must be witnessing a significant change in stellar metallicity .
we now apply these results to observations of the color distribution within our sample . in figure [ abs_mag ]
we show color - magnitude diagrams for galaxies appearing in table [ phot1 ] .
the plot clearly shows that lsbgs span a wide range in ir color ( @xmath140 ) , immediately suggesting a wide range in the stellar metallicities of the galaxies . as argued above , the largest possible shift in color due to age changes alone ( @xmath139 )
is far too small to explain the full range of colors seen in our data . in other words , although changes in mean stellar age may contribute to the range in color , they can not be responsible for all of it .
we may place some limits on the _ maximum _ range of metallicity compatible with the observed spread in @xmath3 color by assuming that the mean stellar age is constant ( i.e. that all the color changes are due to metallicity alone ) .
based upon the grids in figure [ gridplot ] , a factor of 10 change in the metallicity produces changes in color of roughly @xmath141 .
therefore , our observed range in galaxy colors suggests a maximum change in mean stellar metallicity of more than a factor of 100 across the entire sample .
alternatively , we can calculate a _
minimum _ range in metallicity by allowing for the maximum spread in age ( @xmath139 ) .
the remaining color difference still implies a minimum range of over a factor of 20 in mean stellar metallicity .
these two limiting cases suggests that over the mass range in our sample , we have a range in stellar metallicity covering a factor of @xmath142 .
in addition to demonstrating the overall range of near - ir colors , figure [ abs_mag ] shows a statistically significant trend between the ir color and the stellar luminosity of the galaxy , suggesting a trend between metallicity and stellar mass .
( we may also use figure [ abs_mag ] to compare the typical colors of the low gas - mass and high gas - mass subsamples that the lower hi mass galaxies in sample a are bluer that sample b by @xmath143 magnitudes , on average , immediately suggesting a mass dependent trend in galaxy metallicity , given the trends between hi mass , total stellar mass , and baryonic mass between the two samples ) .
we can explore these trends more explicitly in figure [ mass_jk ] , where we plot @xmath40 directly as a function of hi mass .
as expected there is a systematic trend between the hi mass of a galaxy and its near - ir color .
finally we plot @xmath40 as a function of the stellar mass and of the total baryonic mass in figure [ color_mass ] .
again , we see trends in the near - ir color with the stellar and baryonic masses ( with a slope of @xmath144 and a rms of 0.12 for the linear fit of color versus logarithmic stellar mass , and with a slope of @xmath145 with a rms of 0.13 for the linear fit with logarithmic total baryonic mass ) .
the general picture painted by figures [ abs_mag]-[color_mass ] is one where as the stellar mass , gas mass , and/or total baryonic mass within a sample of galaxies increases , the galaxies becomes redder in @xmath3 , and thus more metal rich on average .
however , around all of these relations there is substantial scatter , suggesting that none of these quantities is the principal driver of metallicity .
both our intuition and our data suggest that the true driver of the metallicity is the gas mass fraction m@xmath4/m@xmath5 , which indicates the degree to which a galaxy has completed its lifetime of star formation .
galaxies that have only just begun to create stars have only begun to enrich their gas reservoirs , and thus should have a low mean stellar metallicity and very blue @xmath3 colors .
likewise , galaxies that have completely converted their gas into stars via several generations of star formation will have greatly enriched the gas from which the last half of the stars formed , leading to much more metal rich stellar populations and redder @xmath3 colors . in figure [ ratio_lum_color ] we plot the gas mass fraction m@xmath4/m@xmath5 as a function of the metal sensitive @xmath3 color , and overall the behavior is just as expected .
compared to the previous figures , there is a substantial tightening of the relationship , especially towards the more stellar dominated systems ( i.e. low values of m@xmath4/m@xmath5 ) .
this suggests that the typical stellar metallicity of a galaxy is indeed largely set by the degree to which it has already converted its gas into stars .
we have already seen hints of this effect in figure [ stellar_density ] , where the @xmath3 color is highly correlated with the central surface density of the galaxy , which in turn sets the efficiency of converting gas into stars and thus the subsequent value of m@xmath4/m@xmath5 .
we may also gain additional insight by considering the scatter in figure [ ratio_lum_color ] . in a `` closed box '' model for chemical enrichment , the mean metallicity of the stellar population increases monotonically with decreasing gas mass fraction , leading to a perfect one - to - one relation between the two quantities .
if the closed box model is valid on the scale of galaxies , then the relationship between @xmath3 and m@xmath4/m@xmath5 should have zero scatter ( modulo the small differences in @xmath3 produced by changes in mean age ) .
for the galaxies in figure [ ratio_lum_color ] where star formation has proceeded largely to completion ( m@xmath4/m@xmath146 ) , the scatter is comparable to our observational errors , suggesting that evolution could have proceeded roughly along a closed box pathway , as also concluded by @xcite for normal spirals .
in contrast , for more gas rich galaxies , the scatter in @xmath3 is significantly larger than our uncertainties , suggesting that the closed box model is not a good approximation to the gas and metallicity evolution of these systems . instead
, these systems may still be in a phase of sporadic gas infall , as opposed to the more massive star - dominated galaxies where the gas infall seems to have run to completion .
note also that these gas rich galaxies span a much wider range in stellar luminosity ( and thus stellar mass ) at fixed m@xmath4/m@xmath5 , possibly suggesting a much wider range of mean stellar age , which could contribute to the larger scatter as well . to facilitate comparison with earlier work in the literature
, we have also included in figure [ jk_mass_lum_ratio ] plots of the @xmath3 color as a function of m@xmath4/l , which has traditionally been used as an indicator of the relative fraction of gas and stars . the observed trend toward lower m@xmath4/l and redder colors
is observed in both bandpasses , and while it is more obvious in the @xmath1 band , the relationship is still much weaker overall than the trend with m@xmath4/m@xmath5 in figure [ ratio_lum_color ] .
we fit the trend with straight lines , giving a slope of @xmath147 in @xmath0 ( which is not significant ) , but a slope of @xmath148 in @xmath1 , which is .
note that @xcite found _ no relationship _ between the gas fraction and the @xmath149 color , due to the large scatter ( in that work the gas fraction and color relationship arise from a compilation of different sources of data ) .
our observed tightening of the relationship reflects @xcite s observations that near - ir colors are less scattered than blue colors when plotted against the gas fraction .
we may explore the possible effects of age by considering previous work on the optical colors of lsbg s .
for example , one may compare figure [ mass_jk ] with figure 5a presented by @xcite , where _ no correlation _ exists between the hi mass and the @xmath149 ( optical ) color .
the lack of any correlation between hi mass and the age sensitive @xmath149 color index suggests a stochastic star formation history , as also argued by @xcite for lsb dwarfs .
the measured colors are luminosity weighted , and thus very small changes in the current star formation rate produce large changes in the luminosity - weighted mean age , and thus in @xmath149 . as quoted by @xcite , this clearly suggests the existence of `` dormant '' populations of lsbgs galaxies , which can have large quantities of hi but low current star formation . if we take the increased scatter in figure [ ratio_lum_color ] towards gas rich galaxies as an indication of continuing gas infall , then we may simultaneously have an explanation for the sporadic past star formation history suggested by the optical data , namely gas density dependent star formation thresholds . in this scenario ,
ongoing gas infall occasionally raises the gas surface density above a critical threshold , triggering a burst of star formation and a strong bluing of the optical colors , whiled simultaneously diluting the metallicity of the gas phase and of subsequent generations of stars .
this would yield a larger scatter both in metallicity and in age , leading to the much larger spread in @xmath3 colors .
it would also help to explain why many of the optically faint lsbgs in our sample are sufficiently old to have developed significant red giant branch populations , in spite of what appear to be young luminosity weighted ages implied by their typically blue optical colors .
finally , we note that all of the results above are based upon the color of the entire galaxy .
if we were to consider only the disk component , the relationship between gas - to - star mass - luminosity ratio and color would be loosened with the increased scatter being due to the still low s / n which is obtained for the disk photometry in the near - ir .
the near - ir photometry of low surface brightness galaxies presented in this paper constitutes one of the largest database published until now . because of our original selection of galaxies from the @xcite catalog ,
our sample includes not just a very large number of low surface brightness galaxies , but a number of normal spiral galaxies as well .
the total sample studied provides good indication that an old , red population of low surface brightness galaxies does exist .
many of the lsbgs analyzed exhibit clear bulges and have high surface brightness disks in the near - ir , even though they were originally identified in a catalog of optical lsbgs .
morphologically , the sample spans the full range of spiral hubble types , from sa to i m .
while there are broad correlations between the galaxies optical and near - ir morphologies , we find several cases of galaxies which appear to be late - type in the optical , but have prominent bulges in the near - ir ( e.g. lsbgs # 100 , 338 , 384 , and 446 in figures [ mosaics1 ] and [ mosaics2 ] ) .
in addition to the morphologies , a quantitative analysis of the surface brightness profiles of the galaxies suggests that our sample contains galaxies with disk surface densities comparable to normal spirals , as well as a very high fraction of truly low surface brightness ( and/or low surface density ) disks . on the bright end , we find that the central surface brightnesses of the disks in our sample show a well defined cutoff brighter than @xmath80 mag arcsec@xmath8 ( independent of the @xmath0-band absolute magnitude ) .
we argue that this cutoff is physical , and reflects the maximum surface density at which disks are stable .
on the lsbg end , our sample includes a very high fraction of galaxies that are indeed low surface brightness even in the near - ir , with some reaching roughly 3.5 magnitudes fainter in surface brightness than the bright cutoff .
the disks also have sizes that are systematically smaller for lower mass galaxies .
as part of our analysis , we have calculated the total _ baryonic _
masses ( m@xmath150 ) of galaxies in our sample . for m@xmath151
, we have taken the hi mass from the @xcite catalog , and assumed that other gaseous components are negligible .
for m@xmath123 , we have used the color - dependent stellar mass - to - light ratios of @xcite to translate the observed near - ir luminosity into stellar mass .
the mass - to - light ratios also allow us to convert observed surface brightnesses into stellar mass surface densities .
the resulting masses and surface densities allow us to consider trends involving fundamental physical quantities ( such as the ratio of the gas mass to the total baryonic mass ) .
the total baryonic mass itself is probably a good indicator of the total mass of the galaxies , assuming that baryon and dark matter are evenly mixed on large scales . among the masses of the various components , we see strong correlations , many of which follow trends previously identified in the optical ( for example ,
see compilations in @xcite ) .
galaxies with large stellar masses tend to also have high hi masses , higher central surface densities , but small gas mass fractions ( i.e.are hi - poor ) .
likewise , galaxies with small stellar masses tend to have low hi masses but have a larger fraction of their baryonic mass in the gas phase ( are hi - rich , as for the lsbgs studied by @xcite ) .
we note that even galaxies which have formed large masses of stars can still retain large reservoirs of neutral gas , even after rich episodes of star formation .
we find several strong morphological correlations with the masses of the galaxies .
there are systematic trends towards more irregular , diffuse morphologies and lower surface densities with increasing gas mass fraction ( or alternatively , m@xmath4/l@xmath89 ) .
we also find that the central surface brightness of the bulge component varies strongly with the near - ir luminosity and hi mass , as well as with the stellar and the total baryonic mass ( such that the bulges have higher surface brightness for larger values of the listed quantities ) .
moreover , we find that our sample exhibits a `` floor '' in the baryonic mass required for bulge formation , at @xmath128 ; we have no galaxies with lower baryonic masses that have formed high surface brightness bulges .
we have also explored how the near - ir color varies within our sample . in order to interpret the colors ,
we have used stellar population synthesis models to argue that the @xmath3 color is highly metal sensitive , such that redder colors imply higher metallicities .
we argue that based upon the range of observed colors our sample spans a factor of more than 20 in metallicity ( but less than a factor of 100 ) .
we find that galaxies have redder @xmath3 colors with increasing hi mass , stellar mass , and baryonic mass , suggesting that more massive galaxies have in general have more metal enriched stellar populations .
we find very tightest correlations between color and the gas mass fraction m@xmath4/m@xmath5 in particular .
these correlations are significantly tighter than previous optically determined trends , even in the redder @xmath152 color ( see compilation in figure 6 of @xcite ) . in a closed box model for chemical evolution
, there should be a one - to - one relationship between the metallicity and the gas mass fraction . at small values of m@xmath4/m@xmath5 , this is essentially what we see , in that the observed scatter in the metal - sensitive @xmath3 color is comparable to our observational uncertainties .
this suggests that in the gas poor , bulge dominated massive galaxies , chemical evolution has proceeded along the canonical closed box pathway ( possibly confined to the bulge , which dominates the luminosity and thus the colors of the galaxy ) .
we find a different behavior among the gas rich , disk dominated , lower mass galaxies . in these systems there is large scatter around the relationship between color and m@xmath4/m@xmath5 .
this suggests that not only is the enrichment of these galaxies on - going ( given their high gas mass fraction ) , but that it is not proceding as expected for a `` closed box '' model of enrichment .
instead , it is likely that these gas rich systems are continuing to experience episodes of gas infall .
these episodes of gas accretion first increase the ratio of gas to stars , and then dilute the metallicity of subsequent generations of stars .
episodic accretion would also explain the stochastic star formation suggested by analyses of the optical colors of lsbgs @xcite and theoretical calculations @xcite .
temporary increases above a gas surface density threshold for star formation could dramatically increase the star formation rate @xcite , triggering bursts of star formation and an increased rate of gas consumption .
this change in the luminosity - weighted mean stellar age would produce additional scatter in the color , scatter which will be more dramatic in the optical . within our sample ,
we see strong evidence that the star formation efficiency is indeed surface density dependent .
we find strong correlations between the gas mass fraction m@xmath4/m@xmath5 and the central stellar surface density ( i.e. m@xmath123/pc@xmath120 ) , suggesting that galaxies with the highest initial gas surface density have also had the highest efficiency in turning gas into stars .
this link between surface density and star formation efficiency also suggests that the young mean stellar age of lsbgs ( as deduced from luminosity - weighted optical colors ) results principally from the galaxies being slow to convert gas into stars , rather than from their being slow to assemble ( see also @xcite , @xcite , @xcite ) .
however , the evidence for on - going gas accretion in lower luminosity galaxies suggests that this second effect does play some role as well . for future work , we intend to extend the current work with the addition of optical colors from our on - going database of @xmath9 and @xmath13 photometry for the same galaxies presented here , as well as from other bands provided by the sloan digital sky survey . we shall compute dynamical masses , and current gas phase metallicities using also future spectroscopic data .
this will allow us to perform more critical and accurate tests of the ideas presented in this work .
we acknowledge las campanas observatory for granting the large investment in observing time necessary for this project .
we also thank all of the staff at las campanas .
we would like to acknowledge pablo araya in the reduction of some images appearing in this paper , as well as andreas reisenegger for comments and shaking discussions around the subject presented here .
we also thank the anonymous referee for many helpful suggestions that greatly improved the paper .
part of this research was funded by the andes - carnegie fellowship c-12927 , during a post - doctoral position at las campanas observatory ( gg ) .
gg thanks funding from `` proyecto dipuc 2001/14-e '' .
gg and li thanks `` proyecto fondap '' center for astrophysics .
li acknowledges `` proyecto puente puc '' .
jd was partially supported by nsf grant ast-990862 and by the alfred p. sloan foundation during this work .
this research has made use of the nasa / ipac extragalactic database , which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
the research has also made use of the vizier service from the centre de donnes astronomiques de strasbourg ( cds ) , and the digitized sky survey , which is produced at the space telescope science institute under u.s .
government grant nag w-2166 .
the images of these surveys are based on photographic data obtained using the oschin schmidt telescope on palomar mountain and the uk schmidt telescope .
the plates were processed into the present compressed digital form with the permission of these institutions .
bergvall , n. , rnnback , j. , masegosa , j. , stlin , g. 1999 , , 341 , 697 bertin , e. & arnouts , s. 1996 , , 117 , 393 beijersbergen , m. , de blok , w. , & van der hulst , j. 1999 , , 351 , 903 bell , e. , & de jong , r. , 550 , 212 bell , e. , barnaby , d. , bower , r. 2000 , , 312 , 470 bell , e. & de jong , r. , 2000 , , 312 , 497 bell , e. , bower , r. , de jong , r. , hereld , m. , & rauscher , b. 1999 , , 302 , 55 de blok , w. , & mcgaugh , s. 1997 mnras , 290 , 533 de blok , w. , & van der hulst , j. , , 335 , 421 de blok , w. , mcgaugh , s. , & van der hulst , j. 1996 , mnras , 283 , 18 de blok , w. , van der hulst , j. , & bothun , d. 1995 , mnras , 274 , 235 bothun , g. 1982 , , 50 , 39 bothun , g. , impey , c. , malin , d. , & mould , j. 1987 , aj , 94 , 23 bothun , g. , impey c. , & mcgaugh , s. 1997 , pasp , 109 , 745 courteau , s. 1996 , apjs , 103 , 363 dalcanton , j. , spergel , d. , & summers , f. 1997 , apj , 482 , 659 disney , m. 1976 , nature , 263 , 573 fioc , m. & rocca - volmerange , b. 1997 , a&a , 326 , 950 fioc , m. , & rocca - volmerange , b. , in preparation fisher , j. , & tully , r. 1975 , , 44 , 151 freeman , k. 1970 , , 160 , 811 ( erratum 1970 , , 161 , 802 ) galaz , g. 2000 , aj , 119 , 2118 gerritsen , p. & de blok , w. 1998 , a&a , 342 , 655 giovanelli , r. , haynes , m. , salzer , j. , wegner , g. , da costa , l. , & freudling , w. 1998 , , 116 , 2632 impey , c. & bothun , g. 1989 , apj , 341 , 89 impey , c. , sprayberry d. , irwin m. , & bothun g. 1996 , apjs , 105 , 209 de jong , r. 1996 , , 313 , 45 jimenez , r. , padoan , p. , matteucci , f. , heavens , a. 1998 , , 299 , 123 kennicutt , r. 1998 , knezek , p. , & wroten , j. 1994 , aas , 185 , 77.13 kraan - korteweg , r. 1986 , , 66 , 255 liu , m. , charlot , s. , & graham , j. 2000 , , 543 , 644 longmore , a. , hawarden , t. , goss , w. , mebold , u. , & webster , b. 1982 , mnras , 200 , 325 matthews , l. , van driel , w. , & monnier - ragaigne , d. 2001 , , 365 , 1 mcgaugh , s. , & de blok , w. 1997 , apj , 481 , 689 mcgaugh , s. 1994 , apj , 426 , 135 mcgaugh , s. 1992 , phd thesis , univ .
michigan oneil , k. , bothun , g. , & schombert , j. 2000 , , 119 , 136 oneil , k. , bothun , g. , schombert , j. , cornell , m. , & impey , c. 1997 , , 114 , 2448 persson , e. , murphy , d. , krzeminski , w. , roth , m. , & rieke , m. 1998 , , 116 , 2475 schelegel , d. , finkbeiner , p. & davis , m. 1998 , , 500 , 525 schombert , j. , mcgaugh , s. s. , & eder , j. a. 2001 , , 95 , 1389 schombert , j. & bothun , g. 1988 , , 95 , 1389 sprayberry , d. , bernstein , g. , impey , c. , & bothun , g. 1995 , apj , 438 , 72 sprayberry , d. , impey , c. , & irwin , m. 1996 , , 463 , 535 tully , r. , pierce , m. , huang , j .- s . , saunders , w. , verheijen , m. , & witchalls , p. 1998
, , 115 , 2264 tully , b. & verheijen , m. 1997 , , 484 , 145 tully , r. & fisher , j. 1977 , , 54 , 661 van den bergh , s. 1959 , pub .
david dunlap obs . , 2 , 159 van den bergh , s. 1998 , _ galaxy morphology and classification _ ( cambridge university press ) van den hoek , l. , de blok , w. , van der hulst , j. , & de jong , t. 2000 , , 357 , 397 van der hulst , skillman , e. , smith , t. , bothun , g. , mcgaugh , s. , & de blok , w. 1993 , aj , 106 , 548 de vaucouleurs , g. 1948 , ann . dastrophys .
11 , 247 verheijen , m. 1997 , ph.d .
thesis , univ .
groningen zwicky , f. 1957 , _ morphological astronomy _
( berlin : springer ) cccccccc ( 1 ) & ( 2 ) & ( 3 ) & ( 4 ) & ( 5 ) & ( 6 ) & ( 7 ) & ( 8) + @xmath1535.22 ( 0.15 ) & @xmath1530.21 ( 0.23 ) & @xmath1535.98 ( 0.13 ) & 0.12 ( 0.25 ) & 0.12 ( 0.05 ) & 0.07 ( 0.08 ) & 17 & 9 + @xmath1533.42 ( 0.12 ) & @xmath1530.08 ( 0.26 ) & @xmath1534.00 ( 0.14 ) & 0.07 ( 0.20 ) & 0.11 ( 0.06 ) & 0.09 ( 0.08 ) & 12 & 9 + + 16 & 0027 + 0134 & 3436 & 8.65 & 2250.0 & 2750.0 & 20.0 & 13.119 & 0.03 & 12.817 & 0.06 & -20.186 & -20.488 & 0.302 & 0.07 + 33 & 0101 + 0304 & 4379 & 8.92 & 3600.0 & 1250.0 & 22.0 & 14.742 & 0.04 & 13.901 & 0.07 & -19.089 & -19.930 & 0.841 & 0.08 + 36 & 0104 + 0140 & 4708 & 8.74 & 1800.0 & 2500.0 & 12.0 & 12.706 & 0.03 & 11.858 & 0.04 & -21.282 & -22.130 & 0.848 & 0.05 + 38 & 0105 + 0047 & 632 & 7.48 & 2250.0 & 3750.0 & 20.0 & 13.573 & 0.03 & 13.220 & 0.07 & -16.055 & -16.408 & 0.353 & 0.08 + 295 & 0124 - 0043 & 963 & 8.10 & 2760.0 & 3750.0 & 20.0 & 13.513 & 0.02 & 12.832 & 0.04 & -17.029 & -17.710 & 0.681 & 0.04 + 100 & 0233 + 0012 & 2615 & 8.89 & 2250.0 & 4250.0 & 22.0 & 13.362 & 0.02 & 12.730 & 0.04 & -19.350 & -19.982 & 0.632 & 0.04 + 146 & 0336 + 0212 & 3144 & 8.62 & 3600.0 & 5500.0 & 22.0 & 14.362 & 0.05 & 13.901 & 0.08 & -18.750 & -19.211 & 0.461 & 0.09 + 200 & 0918 - 0028 & 3492 & 8.63 & 2250.0 & 3600.0 & 19.2 & 12.622 & 0.03 & 11.804 & 0.03 & -20.718 & -21.536 & 0.818 & 0.04 + 201 & 0918 + 0147 & 4978 & 8.88 & 3000.0 & 3750.0 & 18.0 & 15.854 & 0.06 & 14.966 & 0.09 & -18.256 & -19.144 & 0.888 & 0.10 + 249 & 1035 + 0238 & 5637 & 8.89 & 3000.0 & 4000.0 & 10.0 & 16.088 & 0.06 & 15.536 & 0.09 & -18.291 & -18.843 & 0.552 & 0.10 + 253 & 1036 + 0158 & 710 & 7.62 & 3150.0 & 3500.0 & 22.0 & 13.118 & 0.03 & 12.620 & 0.05 & -16.763 & -17.261 & 0.498 & 0.06 + 264 & 1043 + 0202 & 1018 & 7.73 & 2880.0 & 4450.0 & 8.0 & 15.898 & 0.04 & 15.151 & 0.08 & -14.765 & -15.512 & 0.747 & 0.09 + 266 & 1047 + 0131 & 1604 & 7.94 & 3000.0 & 3750.0 & 13.9 & 15.653 & 0.05 & 14.910 & 0.07 & -15.997 & -16.740 & 0.743 & 0.09 + 270 & 1050 + 0253 & 1037 & 8.02 & 3150.0 & 4000.0 & 17.0 & 14.504 & 0.04 & 14.036 & 0.07 & -16.199 & -16.667 & 0.468 & 0.08 + 285 & 1108 + 0121 & 993 & 8.00 & 3000.0 & 3950.0 & 26.1 & 14.466 & 0.03 & 13.899 & 0.06 & -16.143 & -16.710 & 0.567 & 0.07 + 308 & 1156 + 0254 & 3233 & 8.72 & 4800.0 & 5500.0 & 17.0 & 15.791 & 0.04 & 15.098 & 0.07 & -17.381 & -18.074 & 0.693 & 0.08 + 312 & 1158 - 0101 & 1533 & 8.45 & 1800.0 & 2750.0 & 20.0 & 12.253 & 0.02 & 11.419 & 0.05 & -19.299 & -20.133 & 0.834 & 0.05 + 314 & 1205 + 0058 & 5870 & 8.97 & 2000.0 & 2750.0 & 21.0 & 14.034 & 0.04 & 13.156 & 0.06 & -20.433 & -21.311 & 0.878 & 0.07 + 319 & 1209 + 0305 & 883 & 7.94 & 3360.0 & 4000.0 & 15.0 & 15.350 & 0.05 & 15.008 & 0.08 & -15.004 & -15.346 & 0.342 & 0.09 + 329 & 1216 + 0029 & 867 & 7.55 & 3000.0 & 3750.0 & 70.0 & 15.411 & 0.05 & 14.940 & 0.07 & -14.903 & -15.374 & 0.471 & 0.09 + 330 & 1217 + 0103 & 2056 & 8.62 & 2400.0 & 2750.0 & 20.0 & 13.392 & 0.03 & 12.656 & 0.05 & -18.797 & -19.533 & 0.736 & 0.06 + 336 & 1223 + 0117 & 1473 & 7.59 & 3000.0 & 3750.0 & 15.0 & 14.788 & 0.03 & 14.101 & 0.06 & -16.677 & -17.364 & 0.687 & 0.07 + 338 & 1223 - 0058 & 2018 & 8.32 & 3120.0 & 3750.0 & 10.4 & 16.128 & 0.07 & 15.334 & 0.08 & -16.021 & -16.815 & 0.794 & 0.10 + 348 & 1227 + 0254 & 1635 & 8.36 & 2400.0 & 3750.0 & 15.0 & 14.074 & 0.06 & 13.612 & 0.07 & -17.618 & -18.080 & 0.462 & 0.09 + 349 & 1228 + 0157 & 1105 & 7.75 & 3000.0 & 3650.0 & 20.9 & 14.421 & 0.06 & 13.611 & 0.07 & -16.420 & -17.230 & 0.810 & 0.09 + 365 & 1257 + 0219 & 874 & 7.61 & 3000.0 & 4000.0 & 18.0 & 13.588 & 0.04 & 12.604 & 0.05 & -16.744 & -17.728 & 0.984 & 0.06 + 384 & 1326 + 0109 & 3267 & 8.74 & 3000.0 & 3750.0 & 25.0 & 14.313 & 0.03 & 13.625 & 0.06 & -18.882 & -19.570 & 0.688 & 0.07 + 385 & 1327 + 0148 & 1049 & 7.49 & 2400.0 & 3650.0 & 20.8 & 14.285 & 0.03 & 13.734 & 0.06 & -16.443 & -16.994 & 0.551 & 0.07 + 393 & 1350 + 0022 & 4605 & 8.84 & 3150.0 & 3500.0 & 18.0 & 13.797 & 0.04 & 13.190 & 0.05 & -20.143 & -20.750 & 0.607 & 0.06 + 398 & 1353 + 020 & 4738 & 8.96 & 3000.0 & 3750.0 & 17.5 & 15.088 & 0.06 & 14.706 & 0.06 & -18.914 & -19.296 & 0.382 & 0.08 + 400 & 1357 - 0017 & 4255 & 8.74 & 4800.0 & 5500.0 & 13.0 & 16.210 & 0.06 & 15.664 & 0.09 & -17.559 & -18.105 & 0.546 & 0.10 + 424 & 1433 + 0249 & 1557 & 8.03 & 4800.0 & 5250.0 & 17.0 & 15.273 & 0.05 & 14.767 & 0.08 & -16.313 & -16.819 & 0.506 & 0.09 + 433 & 1438 + 0049 & 2837 & 8.30 & 3000.0 & 5500.0 & 20.0 & 14.463 & 0.04 & 14.077 & 0.08 & -18.426 & -18.812 & 0.386 & 0.09 + 435 & 1439 + 0053 & 1891 & 8.53 & 1680.0 & 2250.0 & 27.8 & 13.291 & 0.02 & 12.661 & 0.06 & -18.717 & -19.347 & 0.630 & 0.06 + 437 & 1440 - 0010 & 1744 & 7.92 & 2400.0 & 3500.0 & 20.8 & 14.349 & 0.04 & 13.775 & 0.06 & -17.483 & -18.057 & 0.574 & 0.07 + 436 & 1440 - 0008 & 1703 & 8.46 & 2250.0 & 3000.0 & 28.0 & 12.639 & 0.03 & 12.100 & 0.07 & -19.141 & -19.680 & 0.539 & 0.08 + 468 & 2311 - 000 & 4392 & 8.95 & 1800.0 & 2750.0 & 20.0 & 12.364 & 0.03 & 11.721 & 0.05 & -21.474 & -22.117 & 0.643 & 0.06 + 483 & 2319 + 0112 & 8816 & 9.73 & 2700.0 & 4000.0 & 20.0 & 13.810 & 0.04 & 13.240 & 0.07 & -21.541 & -22.111 & 0.570 & 0.08 + 488 & 2327 - 0007 & 5207 & 8.99 & 2700.0 & 3750.0 & 30.0 & 12.776 & 0.03 & 12.145 & 0.06 & -21.431 & -22.062 & 0.631 & 0.07 + 492 & 2329 + 0203 & 5176 & 8.99 & 2700.0 & 3750.0 & 20.0 & 13.831 & 0.03 & 13.168 & 0.06 & -20.363 & -21.026 & 0.663 & 0.07 + + 4 & 0013 - 0034 & 11806 & 10.18 & 1650.0 & 2750.0 & 20.0 & 12.625 & 0.03 & 11.907 & 0.06 & -23.360 & -24.078 & 0.718 & 0.07 + 23 & 0050 + 0005 & 10283 & 9.83 & 800.0 & 3000.0 & 13.0 & 14.300 & 0.03 & 13.540 & 0.06 & -21.385 & -22.145 & 0.760 & 0.07 + 30 & 0056 + 0019 & 5437 & 9.62 & 1800.0 & 2750.0 & 23.0 & 11.915 & 0.02 & 11.169 & 0.05 & -22.386 & -23.132 & 0.746 & 0.05 + 59 & 0121 + 0128 & 5028 & 9.64 & 1350.0 & 1500.0 & 25.0 & 10.726 & 0.02 & 9.985 & 0.05 & -23.405 & -24.146 & 0.741 & 0.05 + 126 & 0311 + 0241 & 9533 & 9.57 & 2250.0 & 2750.0 & 12.0 & 13.986 & 0.04 & 13.238 & 0.07 & -21.534 & -22.282 & 0.748 & 0.08 + 165 & 0350 + 0041 & 11362 & 9.60 & 2250.0 & 3750.0 & 20.0 & 13.156 & 0.03 & 12.298 & 0.05 & -22.746 & -23.604 & 0.858 & 0.06 + 196 & 0913 + 0054 & 11401 & 9.68 & 2700.0 & 2750.0 & 21.0 & 13.050 & 0.04 & 12.096 & 0.05 & -22.859 & -23.813 & 0.954 & 0.06 + 207 & 0929 + 0147 & 17324 & 10.03 & 1800.0 & 2750.0 & 27.0 & 13.855 & 0.05 & 12.105 & 0.06 & -22.962 & -24.712 & 1.750 & 0.08 + 213 & 0954 + 020 & 9592 & 9.59 & 1800.0 & 2750.0 & 24.0 & 13.686 & 0.05 & 12.768 & 0.06 & -21.848 & -22.766 & 0.918 & 0.08 + 224 & 1007 + 0121 & 29213 & 10.69 & 1800.0 & 2750.0 & 24.0 & 13.217 & 0.03 & 12.119 & 0.07 & -24.735 & -25.833 & 1.098 & 0.08 + 225 & 1008 + 0128 & 9915 & 9.87 & 1800.0 & 2750.0 & 24.0 & 13.080 & 0.05 & 12.167 & 0.05 & -22.526 & -23.439 & 0.913 & 0.07 + 230 & 1015 + 0148 & 13660 & 9.84 & 3000.0 & 3750.0 & 12.0 & 14.965 & 0.03 & 13.913 & 0.04 & -21.336 & -22.388 & 1.052 & 0.05 + 242 & 1030 + 0252 & 8825 & 9.65 & 1800.0 & 2750.0 & 16.0 & 12.581 & 0.04 & 11.650 & 0.04 & -22.772 & -23.703 & 0.931 & 0.06 + 262 & 1043 + 0018 & 13966 & 9.50 & 1800.0 & 2750.0 & 10.0 & 13.911 & 0.03 & 13.118 & 0.06 & -22.439 & -23.232 & 0.793 & 0.07 + 265 & 1045 + 0014 & 11644 & 9.57 & 1800.0 & 2500.0 & 12.0 & 13.747 & 0.05 & 12.898 & 0.06 & -22.208 & -23.057 & 0.849 & 0.08 + 267 & 1048 + 0125 & 11560 & 9.62 & 1800.0 & 2750.0 & 18.0 & 12.819 & 0.03 & 11.787 & 0.07 & -23.120 & -24.152 & 1.032 & 0.08 + 281 & 1106 + 0032 & 7639 & 9.69 & 1800.0 & 2750.0 & 10.8 & 12.674 & 0.02 & 11.743 & 0.06 & -22.365 & -23.296 & 0.931 & 0.06 + 291 & 1116 + 0053 & 8059 & 9.56 & 1117.0 & 2750.0 & 18.0 & 13.128 & 0.02 & 12.211 & 0.07 & -22.028 & -22.945 & 0.917 & 0.07 + 294 & 1121 + 0058 & 7784 & 9.68 & 1800.0 & 2750.0 & 24.0 & 12.791 & 0.02 & 11.940 & 0.06 & -22.289 & -23.140 & 0.851 & 0.06 + 296 & 1125 + 0255 & 6842 & 9.81 & 1800.0 & 2000.0 & 25.0 & 12.155 & 0.04 & 11.253 & 0.06 & -22.645 & -23.547 & 0.902 & 0.07 + 298 & 1129 + 0013 & 11899 & 9.99 & 2400.0 & 3000.0 & 14.0 & 13.975 & 0.02 & 13.029 & 0.06 & -22.027 & -22.973 & 0.946 & 0.06 + 299 & 1130 - 0027 & 6688 & 9.50 & 3550.0 & 4900.0 & 14.0 & 14.053 & 0.02 & 13.134 & 0.05 & -20.698 & -21.617 & 0.919 & 0.05 + 310 & 1156 + 0142 & 14131 & 9.77 & 2250.0 & 2850.0 & 19.2 & 13.586 & 0.03 & 12.623 & 0.05 & -22.789 & -23.752 & 0.963 & 0.06 + 316 & 1208 + 0120 & 14100 & 9.72 & 1800.0 & 2750.0 & 18.0 & 14.077 & 0.05 & 13.172 & 0.06 & -22.293 & -23.198 & 0.905 & 0.08 + 324 & 1211 + 0226 & 22259 & 10.19 & 2250.0 & 2750.0 & 17.0 & 14.187 & 0.03 & 13.515 & 0.05 & -23.175 & -23.847 & 0.672 & 0.06 + 326 & 1213 + 0221 & 23252 & 10.09 & 2880.0 & 4500.0 & 5.0 & 16.700 & 0.04 & 15.762 & 0.04 & -20.757 & -21.695 & 0.938 & 0.06 + 327 & 1213 + 0127 & 14976 & 9.76 & 1200.0 & 2000.0 & 12.0 & 13.039 & 0.04 & 12.112 & 0.05 & -23.462 & -24.389 & 0.927 & 0.06 + 345 & 1226 + 0105 & 23655 & 10.26 & 2250.0 & 2750.0 & 17.0 & 13.526 & 0.03 & 12.788 & 0.05 & -23.968 & -24.706 & 0.738 & 0.06 + 346 & 1226 + 0119 & 6958 & 9.62 & 1200.0 & 1500.0 & 15.0 & 12.701 & 0.04 & 12.012 & 0.04 & -22.136 & -22.825 & 0.689 & 0.06 + 370 & 1300 + 0144 & 12264 & 9.98 & 1800.0 & 2500.0 & 25.0 & 13.171 & 0.04 & 12.089 & 0.06 & -22.896 & -23.978 & 1.082 & 0.07 + 377 & 1310 - 0019 & 12040 & 9.84 & 3600.0 & 7500.0 & 17.0 & 14.502 & 0.02 & 13.905 & 0.05 & -21.525 & -22.122 & 0.597 & 0.05 + 378 & 1315 + 0029 & 9497 & 9.94 & 3150.0 & 4000.0 & 16.0 & 14.259 & 0.03 & 13.498 & 0.05 & -21.253 & -22.014 & 0.761 & 0.06 + 380 & 1321 + 0137 & 16954 & 9.64 & 2250.0 & 3750.0 & 17.0 & 13.204 & 0.03 & 12.583 & 0.05 & -23.567 & -24.188 & 0.621 & 0.06 + 407 & 1401 + 0108 & 12901 & 9.66 & 2700.0 & 2750.0 & 17.0 & 13.697 & 0.04 & 12.965 & 0.05 & -22.480 & -23.212 & 0.732 & 0.06 + 410 & 1405 + 0006 & 7518 & 9.73 & 1800.0 & 2750.0 & 17.0 & 13.608 & 0.04 & 13.011 & 0.04 & -21.397 & -21.994 & 0.597 & 0.06 + 446 & 1446 + 0231 & 10261 & 9.63 & 3000.0 & 3750.0 & 10.4 & 16.251 & 0.05 & 15.481 & 0.04 & -19.429 & -20.199 & 0.770 & 0.06 + 447 & 1446 + 0238 & 10281 & 9.73 & 1800.0 & 1500.0 & 17.0 & 13.291 & 0.04 & 12.620 & 0.07 & -22.393 & -23.064 & 0.671 & 0.08 + 462 & 2303 - 0006 & 7468 & 9.71 & 1800.0 & 2750.0 & 30.0 & 12.163 & 0.04 & 11.465 & 0.05 & -22.827 & -23.525 & 0.698 & 0.06 + 463 & 2304 + 0155 & 5244 & 9.57 & 1800.0 & 3750.0 & 25.0 & 13.034 & 0.03 & 12.412 & 0.05 & -21.189 & -21.811 & 0.622 & 0.06 + 470 & 2312 - 0011 & 15380 & 9.98 & 1800.0 & 2750.0 & 18.0 & 13.074 & 0.03 & 12.305 & 0.04 & -23.485 & -24.254 & 0.769 & 0.05 + 471 & 2313 + 0008 & 8667 & 9.76 & 1350.0 & 2750.0 & 22.0 & 12.238 & 0.02 & 11.488 & 0.04 & -23.076 & -23.826 & 0.750 & 0.04 + 473 & 2315 - 0000 & 8938 & 9.71 & 1800.0 & 2750.0 & 22.0 & 12.212 & 0.02 & 11.398 & 0.06 & -23.168 & -23.982 & 0.814 & 0.06 + 474 & 2317 + 0112 & 9046 & 9.74 & 1350.0 & 2500.0 & 22.0 & 12.333 & 0.02 & 11.556 & 0.06 & -23.074 & -23.851 & 0.777 & 0.06 + 485 & 2320 + 0107 & 8868 & 9.84 & 1800.0 & 2750.0 & 18.0 & 13.290 & 0.03 & 12.676 & 0.05 & -22.073 & -22.687 & 0.614 & 0.06 + 484 & 2320 + 0110 & 8951 & 9.91 & 1350.0 & 2500.0 & 15.0 & 12.853 & 0.03 & 12.150 & 0.06 & -22.531 & -23.234 & 0.703 & 0.07 + 487 & 2325 - 0000 & 10147 & 10.00 & 2700.0 & 2500.0 & 18.0 & 12.558 & 0.05 & 11.502 & 0.07 & -23.098 & -24.154 & 1.056 & 0.09 + 515 & 2349 + 0248 & 5324 & 9.79 & 1800.0 & 2750.0 & 22.0 & 12.631 & 0.04 & 12.003 & 0.04 & -21.624 & -22.252 & 0.628 & 0.06 + 516 & 2351 + 0045 & 11279 & 9.59 & 3150.0 & 3000.0 & 20.0 & 15.421 & 0.04 & 14.425 & 0.06 & -20.465 & -21.461 & 0.996 & 0.07 | we present results for @xmath0 and @xmath1 near - ir imaging data on a large sample of 88 galaxies drawn from the catalogue of impey et al .
( 1996 ) .
the galaxies span a wide range in optical and ir surface brightness and morphology ( although they were drawn from a catalog constructed to identify low surface brightness galaxies , hereafter lsbgs ) .
they were also selected to include very low and high hi mass galaxies in order to ensure that they span a wide range of evolutionary states .
the near - ir data unveils many features of lsbgs not seen before in the optical .
first , a high fraction of the observed lsbgs are very luminous in the near - ir , indicating that they have a well developed old stellar population , and that older lsbgs are more frequent in the universe than data from optical bands suggested .
second , the near - ir morphologies are often quite different than seen in the optical .
many diffuse lsbgs that are apparently bulgeless when observed in blue bands , instead exhibit nuclei in @xmath0 and @xmath1 bands .
third , we find significant trends between the near - ir morphologies of the galaxies and their ratio of hi mass to near - ir luminosity .
fourth , we find no trend in disk surface brightness with absolute magnitude , but significant correlations when the bulge surface brightness is used . finally , we find that the formation of a bulge requires a galaxy to have a total baryonic mass above @xmath2 .
a wide variety of other correlations are explored for the sample .
we consider correlations among morphologies , surface brightnesses , near - ir colors , absolute magnitudes , and hi masses .
in addition , using previous results by @xcite , we convert the galaxies near - ir luminosities to stellar masses , using color - dependent stellar mass - to - light ratios .
this allows us to consider correlations among more fundamental physical quantities such as the hi mass , the stellar mass , the total baryonic mass , the gas mass fraction , the mass surface density , and the metallicity ( via the highly metal sensitive color index @xmath3 ) .
we find that the strongest of our correlations are with the ratio of hi mass to total baryonic mass , m@xmath4/m@xmath5 , which tracks the evolutionary state of the galaxies as they convert gas into stars , and which ranges from 0.05 up to nearly 1 for the galaxies in our sample .
we find strong systematic trends in how the metallicity sensitive @xmath3 color becomes redder with decreasing m@xmath4/m@xmath5 , as would be expected for `` closed box '' models of chemical enrichment .
however , the increased scatter with increasing gas mass fraction and decreasing galaxy mass suggests that gas infall is increasingly significant in the gas rich lower mass systems .
we argue that the overall range in @xmath3 color argues for at least a factor of 20 change in the mean stellar metallicity across the mass range spanned by our sample .
we also see strong trends between m@xmath4/m@xmath5 and central surface density , suggesting that increased star formation efficiency with increasing gas surface density strongly drives the conversion of gas into stars .
0@xmath6 = 100 _ h _ km s@xmath7 mpc@xmath7 |
physics of spin transport has been a matter of intensive research in recent years .
one of the topics of current interest is the spin hall effect .
this has been originally proposed theoretically @xcite and later followed by various experimental results @xcite .
the research on the spin hall effect opens a new field of hall effects in time - reversal invariant systems .
this has also led us to a new concept of the quantum spin hall effect , which is the natural spin " extension of the quantum hall effect @xcite . in the quantum spin hall effect in two dimensions
, the bulk is gapped while there are gapless edge states carrying spin current . in this case , the external magnetic field is zero , while the spin - orbit coupling acts as a spin - dependent magnetic field " , giving rise to the effect analogous to the quantum hall effect .
such insulators showing the quantum spin hall effect are characterized by the @xmath0 topological number @xcite . we shall call them as @xmath0 quantum spin hall insulators ( qshi ) .
the simplest system for the two - dimensional ( 2-d ) @xmath0 qshi is realized as a superposition of the wavefunctions of two quantum hall subsystems @xcite , one with spin up and the other with spin down , having opposite chern numbers .
the system respects not only the time - reversal @xmath6 invariance but also the spin - conservation .
such an insulator supports the same numbers of right - moving up - spin edge states and left - moving down - spin edges .
this kramers pair of chiral edge states is often called as the helical edge state . by its construction ,
the number of this kramers pairs of edge states correspond to the chern integer associated with its bulk wavefunction @xcite .
the @xmath7 symmetry guarantees the double degeneracy between right - moving up - spin and left moving down - spin states .
thus , the stability of _ each _ kramers pair is supported by this @xmath7-symmetry .
spin non - conserving ( but @xmath7 invariant ) perturbations , however , introduce level repulsions between two _ different _ kramers pairs .
namely , they usually let two pairs annihilate with each other , and open a gap . accordingly , in the presence of generic spin - non - conserving perturbations , those wavefunctions having _ even _ numbers of kramers pairs , reduce to _ trivial
_ insulators , which have no gapless edge states @xcite .
meanwhile , wavefunctions having _ odd _ numbers of pairs still can have one active helical edge mode .
the latter is dubbed as the @xmath0 quantum spin hall ( topological ) insulator .
thus , stability of such a gapless edge state is protected only by the @xmath7-symmetry , while does not require spin - conservations anymore @xcite .
the three - dimensional ( 3-d ) version @xcite of the @xmath0 qshi carries same characters as that of 2-d does .
the 3-d @xmath0 qshi also allows any spin - nonconservering perturbations , while always requires the @xmath7 symmetry .
simultaneously , however , it is not a mere extension of the 2-d @xmath0 qshi , in a sense that 3-d @xmath0 qshi has _ no @xmath8-analogue _ of qshi .
namely , they support a @xmath9 massless dirac fermion as its surface state @xcite , instead of a helical edge state . in the 2-d surface brillouin zone , say @xmath10-@xmath11 plane , this massless dirac fermion has a spin which depends on the ( surface ) crystal momentum .
it is clear that such an insulator can not be adiabatically connected into a composite of two spinless wavefunctions . in such @xmath0 qshi , the @xmath7 symmetry therefore guarantees the massless nature of each @xmath9 surface dirac fermion .
@xmath0 qshi always has a quantum critical point at its phase boundary to any ordinary insulators , in both 2-d and 3-d .
for example , from the 3-d tight - binding model proposed by fu , kane and mele @xcite , we can explicitly see this ; when a certain @xmath7 symmetric parameter is varied in their model , 3d @xmath0 qshi is driven into an ordinary insulator , latter of which does not support any surface states ( see fig .
[ t1 ] ) . observing this
, a following question naturally arises ; during this tuning , the @xmath7 symmetry is _ always _ preserved , so that the massless nature of the surface dirac fermion is supposed to be protected by this . at the same time
, however , this @xmath9 surface fermion should have become `` massive '' , when a system enter an ordinary insulator phase .
thus , one might ask how this single surface dirac fermion could acquire a finite mass , with keeping the @xmath7 symmetry ?
qshi and an ordinary insulator .
@xmath2 is a system parameter driving the phase transition . when @xmath5 , the system is in a critical phase . in the fu - kane - mele model
, it corresponds to the relative strength of one out of the four nn transfer integrals emitting from a single site.,scaledwidth=45.0% ] the answer is simple ; we have _ two _ sample boundaries , say @xmath12 and @xmath13 .
each boundary supports one @xmath9 surface massless dirac fermion respectively .
they are localized at each boundary , when the bulk gap is sufficiently large .
in such a situation , a mixing between these two @xmath9 surface massless dirac fermions is tiny , i.e. @xmath14 with @xmath15 being the localization length .
however , when a system becomes close to the quantum critical point , a mixing between these two surface states becomes substantial , with increasing @xmath15 .
when a bulk eventually reaches the quantum critical point , two surface massless dirac fermions readily communicate via extended bulk states .
thus , they generally annihilate in pairs , just as in those insulators having even number of @xmath9 surface dirac fermions at one boundary .
this simple picture in the clean limit raises the following non - trivial speculations about the disorder effects on the @xmath0 qshi .
suppose that @xmath7-symmetric random potentials are introduced in the topological insulator phase .
when the corresponding bulk gap is sufficiently large , we could begin with two separate bands . the scaling argument in 3-d @xcite
tells us that each band should always have two mobility edges , respectively ( see fig .
[ t2](b ) ) .
namely , there is no delocalized bulk - wavefunction near the zero energy .
starting from this phase , let us change some @xmath7-invariant model - parameters , so that a bulk transits from this topological insulator to an ordinary one . from the argument in the clean limit
, one can then expect that _ a delocalized bulk - wavefunction should emerge at the zero - energy region at the quantum critical point _ , i.e. @xmath16 ( see fig .
[ t2](a ) ) .
if it were not , the two surface states localized at the two sample boundaries could not communicate at all and they could not annihilate with each other . as a result ,
the system was unable to smoothly enter an ordinary band insulator , since the latter one does not support any surface state at all . .
( a ) @xmath5 ; at quantum critical point .
( b ) @xmath17 ; in the topological insulator phase.,scaledwidth=45.0% ] to put this reversely , the existence of the quantum critical point ( qcp ) having extended bulk wavefunctions is always required , whenever this critical point separates an ordinary insulator and the topological insulator .
this is because these two insulating phases support different numbers of kramers pairs of surface states .
moreover , provided that these surface states are stable by itself , this qcp should be also stable , _ however small _ the density of state ( dos ) at the zero energy is and _ however strong _ the disorder strength is .
otherwise , the topological insulator could be adiabatically connected into an ordinary band insulator , which contradicts the different @xmath0 topological numbers for the two phases . in this paper
, we will uncover several novel features associated with the non - magnetic disorder effects onto this topological quantum critical point .
the organization of this paper is summarized as follows . in the next section
, we will briefly review the effective continuum model for the quantum critical point intervening the @xmath0 topological insulator and an ordinary insulator .
the effective model is known to be described by the @xmath1 dirac fermion , whose mass term brings about the topological quantum phase transition .
namely , when the mass term changed from positive to negative , a system transits from the topological insulator to an ordinary insulator . as such
, we call this mass term especially as the topological mass term . based on this effective model
, we will next introduce various types of the on - site random potentials respecting the @xmath7-symmetry .
note that , in this paper , we restrict ourselves to @xmath7-symmetric cases and exclude magnetic impurities , because , in the absence of the @xmath7 symmetry , the two phases are no longer topologically distinct .
based on the self - consistent born approximation , we first work over the single - particle green function in the section iii .
the phase diagram spanned by the ( bare ) chemical potential @xmath3 , ( bare ) mass term @xmath2 and strength of the disorder @xmath18 is derived . in particular , at the critical point , i.e. @xmath5 , we found some critical value of the disorder strength , @xmath19 , above which the zero - energy state , i.e. @xmath20 , acquires a finite life - time @xmath21 ; @xmath22 = 1-\frac{\alpha_{c}}{\alpha } \label{e1}. \end{aligned}\ ] ] since the density of state in our model is always proportional to the inverse of the life - time ( see below ) , non zero @xmath23 simply means that a system is in a compressible phase .
when a finite but small topological mass @xmath2 is introduced for @xmath24 , the life - time @xmath21 and the renormalized mass @xmath25 becomes as follows ; @xmath26 where @xmath27 is given as a function only of @xmath18 via eq .
( [ e1 ] ) .
thus , when the bare topological mass exceeds the critical value @xmath28 , the density of state vanishes , so that a system enters an incompressible phase .
this gapped phase can be adiabatically connected into band insulator phases in the clean limit .
accordingly , we will reach the phase diagram for the zero - energy state as depicted in fig .
[ t3 ] . .
a blue shaded region corresponds to a compressible phase , which separates two gapped phases , i.e. an ordinary insulator and the topological insulator .
this phase boundary for @xmath24 is given by @xmath29$].,scaledwidth=45.0% ] in the section iii , we also describe the behavior of the one - particle green function for a finite @xmath3 ( see below ) , in which we observe that the compressible phase ( not necessarily metallic phase ) always intervenes the topological insulator phase and an ordinary insulator phase as in fig .
[ t3 ] . focusing on this intervening compressible phase , especially for @xmath30
, we will derive in the section .
iv the diffuson , cooperon and the weak localization correction to the electric conductivity .
we will first observe that the diffuson is composed of two quasi - degenerate low - energy modes ; @xmath31 ( see fig .
( [ t12 - 1 - 0 ] ) or eq . ( [ identity ] ) for the definition of @xmath32 ) .
the first term participates in usual charge diffuson mode , and therefore always has the diffusion pole , i.e. @xmath33^{-1}$ ] ( @xmath34 and @xmath35 stand for the frequency and momentum of the density fluctuation respectively ) .
the other low - energy mode , however , becomes massless only in the absence of the topological mass @xmath2 .
namely , its low - energy and long wavelength behavior is generally truncated by the infrared cutoff @xmath36 , while this infrared cutoff reduces to zero at @xmath5 ( but generic @xmath3 ) , i.e. @xmath37 . physically speaking
, this second mode describes the diffusion of the parity density degree of freedom , which becomes a conserved quantity of our effective hamiltonian at @xmath5 .
namely , the parity - density correlation function exhibits the diffusion pole structure at the critical point ( @xmath5 ) , while it becomes massive in the presence of the finite topological mass .
consequently , the diffuson acquires one additional low - energy , i.e. the 2nd term of eq .
( [ punch1 - 0 ] ) , into which the information of this parity - density correlation function is separately encoded .
when the hole - line of the diffuson time - reversed , these two - mode features are transcribed into the cooperon : the cooperon thus obtained is also composed of two quasi - degenerate dominant contributions ; @xmath38 ( see fig .
[ t12 - 1 - 0 ] and its caption for the definition of @xmath39 ) . in section .
iv , we will see that , at the critical point ( @xmath5 ) , these two contributions give rise to the same amplitude of the anti - weak - localization ( awl ) correction to the electric conductivity .
when the finite topological mass is introduced , however , the cooperon associated with the parity mode channel becomes ineffective , since its backward scattering behaviour is truncated by the cutoff @xmath36 .
meanwhile , the cooperon obtained from the charge mode channel remains effective , even in the presence of finite @xmath2 . as a result ,
the awl correction at the critical point becomes precisely halved , on introducing the finite topological mass ( quantum correction doubling ) . in terms of this novel behaviour of the parity diffusion mode and that of the corresponding awl correction
, we will argue in the section .
v the possible microscopic mechanism of how the delocalized bulk - wavefunction emerges at the critical point , i.e. @xmath4 . to be more specific
, we expect that the parity diffusion mode mentioned above generally becomes massless , when a system transits from the topological insulator side to the ordinary insulator side . assuming that this is the case
, we will attribute the emergence of the extended bulk wavefunction to the awl correction obtained from this parity mode channel , i.e. the second term of eq .
( [ punch2 - 0 ] ) . for the systematic understanding
, however , one generally needs to go beyond the theoretical approach employed in this paper .
several open issues will be also discussed in the section v. a number of appendices describe other topics useful in understanding the main text in more detail . for clarity of the explanation , we have presented the results only in the case of chemical - potential type disorder in the text .
the study in the presence of general @xmath7 symmetric disorders becomes more cumbersome .
but the basic feature such as the phase diagram is expected to be same . in the appendix
a , we will describe how the one - particle green function at the zero - energy state behaves in the presence of these general @xmath7-invariant random potentials .
our weak - localization calculation described in the section .
iv is the controlled analysis , when it comes to the weakly disordered region , @xmath40 .
namely , for this parameter region , one can confirm self - consistently the coupling constant @xmath41 to be sufficiently small around @xmath20 ( see eq .
( [ control ] ) ) . for @xmath24
, however , this coupling constant generally diverges toward @xmath20 , only to make the weak localization calculation ( and scb calculation ) an uncontrolled analysis .
thus , as the complementary analysis for this strongly disordered region , @xmath24 , we employed the mode - mode coupling analysis in the appendices .
b - d . by taking into account the quantum interference effect due to the cooperon terms
, this theoretical framework gives us the gap equation for the diffusion constant .
main results in section .
iv such as the quantum correction doubling are also supported by the analysis in the appendices .
we consider a system with both @xmath7-symmetry and the spatial inversion ( @xmath42)-symmetry . under this symmetry requirement ,
murakami _ et al .
_ recently derived the minimal model for an arbitrary quantum critical point intervening the topological insulator and an ordinary insulator on a quite general ground @xcite .
it turns out to be always described by the @xmath1 dirac fermion given as follows ; @xmath43 where @xmath2 corresponds to the topological mass term .
without loss of generality , one can regard the topological insulator phase to be @xmath17 and an ordinary insulator phase to be @xmath44 ( see fig .
[ t1 ] ) . to see that `` @xmath2 '' actually endows this dirac fermion with a mass
, we note that following five @xmath45 @xmath46-matrices are anticommuting with one another ; @xmath47 the matrices @xmath48 and @xmath49 are pauli matrices , representing the ( generalized ) sublattice index , and the spin index , respectively . in terms of these pauli matrices , we will take the @xmath7 operator as @xmath50 with @xmath51 being the complex conjugation .
meanwhile , the @xmath42 operator will be taken as @xmath52 .
it follows from these conventions that @xmath53 are @xmath7 odd and @xmath42 odd , while @xmath54 is @xmath7 even and @xmath42 even ( see table .
i ) . together with the property that @xmath55 is @xmath7 odd and @xmath42 odd , we can easily see that our hamiltonian in the clean case is indeed @xmath7 even and @xmath42 even .
this guarantees the kramers degeneracy at each @xmath56-point , irrespectively of the topological mass @xmath2 .
we also note that eq .
( [ e5-t ] ) is indeed the low - energy effective continuum hamiltonian for various lattice model recently discussed in literatures @xcite .
.dirac operators and their symmetries . [ cols="^,^,^",options="header " , ] by use of this equality , eq . ( [ bs15 ] ) turns out to consist only of those 8 functions defined in eq .
( [ vde ] ) ; @xmath57_{\beta\alpha } u^{\rm 2pir}_{\alpha\delta,\gamma\beta } ( k , k';q,\omega ) \big[\hat{\gamma}^{\rm r}_{b}(k';q,\omega ) \big]_{\delta\gamma}. \label{bs19 } \end{aligned}\ ] ] @xmath58 above are given as follows ; @xmath59 @xmath60 defined in eq .
( [ bs19 ] ) generally appears in the eom for the @xmath61-type relaxation function and plays role of the ` mean - field ' induced by the @xmath62-type relaxation functions .
namely , this @xmath63 matrix is nothing but the ` self - energy ' in the matrix - formed eom s for the @xmath64-wave sector ( see eq .
( [ eom1 ] ) ) .
thus , we will refer to @xmath65 as the _ relaxation kernel _ henceforth . before deriving the remaining 6 constitutive equations ,
let us remark on the general property of this kernel .
observing eq .
( [ bs19 ] ) , notice that each element of this matrix becomes pure imaginary , when its two arguments taken to be zero ; @xmath66 this can be directly seen from ; @xmath67 + the eom s for the other 6 @xmath64-wave modes can be derived in parallel with that for the current relaxation function .
specifically , we will begin with the bethe - salpeter equation applied by the following , instead of eq .
( [ bs33 ] ) ; @xmath68_{+ } , \nn \end{aligned}\ ] ] with @xmath69 taken to be @xmath70 respectively .
going through the same procedure as described so far , we will reach the constitutive equations for these remaining 6 modes .
combined with eq .
( [ cont1 ] ) and eq .
( [ bs18 ] ) , such equations consist of the following @xmath71 by @xmath71 matrix - formed eom s ; @xmath72 \cdot \hat{\phi}(q,\omega ) \equiv
\hat{a}(q,\omega ) . \label{eom1 } \end{aligned}\ ] ] @xmath73 and @xmath74 have the eight components ; @xmath75 , \nn \\ \hat{a}^t & \equiv & \big [ a_{0 } , \
a_{j } , \
a_{5 } , \
a_{5j } , \ a_{4 } , \ a_{4j } , \
a_{45 } , \ a_{45j}\big ] , \nn \end{aligned}\ ] ] latter of which is defined as follows ; @xmath76 .
\label{othera}\end{aligned}\ ] ] @xmath77 and @xmath65 are defined as follows ; @xmath78 , \ \
\hat{\cal m } \equiv \left[\begin{array}{cc } \hat{\cal m}_1 & \hat{0 } \\ \hat{0 } & \hat{\cal m}_2 \\
\end{array}\right ] , \nn \\ \hat{\cal k}_1 & \equiv & \left [ \begin{array}{cccc } % 0 \omega & -q & & \\ % j -q & \omega c^{-1 } & & \\ % 5 \omega d^{-1 } + 2i e^{-1 } & & \omega c^{-1 } & \\
% 5j & & & \omega c^{-1 } \\
\end{array}\right ] , \label{eom2 } \\ \hat{\cal k}_2 & \equiv & \left[\begin{array}{cccc } \omega c^{-1 } & & &
\\ % 4j & \omega c^{-1 } & & \omega d^{-1 } + 2i e^{-1 } \\ % 45 & & \omega c^{-1 } & -q \\ % j \times j & \omega d^{-1 } + 2ie^{-1 } & -q & \omega c^{-1 } \\
\end{array } \right ] , \nn \\
\hat{\cal m}_1 & \equiv & \left [ \begin{array}{cccc } % 0 0 & & 0 & \\ % j & { \cal m}_{j , j } & & { \cal m}_{j,5j } \\ % 5 { \cal m}_{5,0 } & & { \cal m}_{5,5 } & \\
% 5j & { \cal m}_{5j , j } & & { \cal m}_{5j,5j}\\ \end{array}\right ] , \label{eom3 } \\ \hat{\cal m}_2 & \equiv & \left[\begin{array}{cccc } % 4 { \cal m}_{4,4 } & & { \cal m}_{4,45 } & \\ % 4j & { \cal m}_{4j,4j } & & { \cal m}_{4j,45j } \\ % 45 { \cal m}_{45,4 } & & { \cal m}_{45,45 } & \\ % j \times j & { \cal m}_{45j , 4j } & & { \cal m}_{45j,45j } \\ \end{array}\right ] , \nn\end{aligned}\ ] ] with @xmath79 by solving eq .
( [ eom1 ] ) , one can obtain the asymptotic expressions for the relaxation functions for small @xmath34 and @xmath35 . ;
@xmath80 where @xmath81 stands for the density of state ( see eq .
( [ dos ] ) ) .
the ( renormalized ) diffusion constant @xmath82 used above and other coupling constants are expressed only in terms of the relaxation kernels estimated at @xmath83 ; @xmath84 eq . ( [ bs25 ] ) and eq .
( [ bs19 ] ) become the essential bulding blocks of our gap equation ( see below ) . when the non - crossing approximation is employed for the 1-point green function , the diffusion pole in eq .
( [ bs24 ] ) should be attributed to the ladder - type diagram whose long - wavelength expressions were already obtained in sec .
especially , we have observed that in the section .
ivb that the charge diffusion mode and parity diffusion mode equally dominates the diffuson in the massless case ( @xmath85 ) , while the parity mode becomes ineffective in the presence of the relatively large topological mass ( @xmath86 ) ( see eqs .
( [ dc12 - 11],[f4],[f3 ] ) ) . corresponding to these two limiting cases
, we will derive two types of gap equations and their solutions in this section .
let us begin with the zero topological mass case first . in this case , we will sum up eq .
( [ dc12 - 13-a ] ) and eq .
( [ dc12 - 13-b ] ) , since @xmath87 .
with use of eq .
( [ dc10 - 6 ] ) and @xmath88 , such a summand takes on a following form ; @xmath89 in section .
ivb , we have observed that the overall factor , @xmath90 , has the diffusion pole as in eq .
( [ f4 ] ) , where its _ bare _ expression were calculated explicitly .
namely , by keeping track of the small @xmath35 effect in eq .
( [ dc6 - 1 ] ) , we obtained the bare diffusion constant as in eq .
( [ baredif ] ) . instead of such bare expressions , however , we will describe henceforth this @xmath91 in terms of the _ renormalized _ diffusion constant defined by eq .
( [ bs25 ] ) .
namely , we want @xmath91 to be given by the relaxation kernels , only to obtain the _ self - consistent _ equation for the diffusion constant . to do this , notice that relaxation functions for _ small _ @xmath35 and @xmath34 are dominated by the diffuson as in eq .
( [ bs12 ] ) .
thus , by substituting eq .
( [ masslessd ] ) into eqs .
( [ bs12],[e61 ] ) , we will first express the density relaxation function in terms of @xmath91 ; @xmath92 the factor @xmath93 in the right hand side stems from the momentum integral over @xmath56 and @xmath94 in eq.([e61 ] ) ; @xmath95 is the ultraviolet cutoff of the momentum - integral .
then , we will equate this with @xmath96 obtained in the * step-(i ) * , i.e. eq .
( [ bs24 ] ) . by doing this ,
@xmath91 is expressed in terms of relaxation kernels ; @xmath97 namely , the diffusion constant @xmath82 in the right hand side was already given by the relaxation kernels as in eq .
( [ bs25 ] ) . substituting this back into eq .
( [ masslessd ] ) , we obtain the asymptotic form of the diffusion ; @xmath98 when its hole line time - reversed , the corresponding cooperon at small @xmath34 and @xmath99 is also derived ; @xmath100 where we used the following identities ; @xmath101_{\gamma^{\prime}\gamma } \hat{t}_{1,\alpha\delta,\beta\gamma } \big[\hat{1}\otimes \hat{s}_y\big]_{\beta\beta^{\prime } } & \equiv & - \hat{t}_{1,\alpha\delta,\gamma^{\prime}\beta^{\prime } } , \nn \\ \big[\hat{1}\otimes \hat{s}_y\big]_{\gamma^{\prime}\gamma } \hat{s}_{1,\alpha\delta,\beta\gamma } \big[\hat{1}\otimes \hat{s}_y\big]_{\beta\beta^{\prime } } & \equiv & - \hat{s}_{1,\alpha\delta,\gamma^{\prime}\beta^{\prime } } , \nn \\ \big[\hat{1}\otimes \hat{s}_y\big]_{\gamma^{\prime}\gamma } \hat{s}_{2,\alpha\delta,\beta\gamma } \big[\hat{1}\otimes \hat{s}_y\big]_{\beta\beta^{\prime } } & \equiv & \hat{s}_{2,\alpha\delta,\gamma^{\prime}\beta^{\prime}}. \nn \end{aligned}\ ] ] the diffusion constant @xmath82 in eq .
( [ c7 ] ) is now given by the relaxation kernels , via eq .
( [ bs25 ] ) .
these relaxation kernels are in turn defined by the 2pir vertex function , via eq .
( [ bs19 ] ) .
the 2pir vertex function is usually dominated by the cooperon given by eq .
( [ c7 ] ) , at around @xmath102 . as such
, we will replace ( approximate ) the 2pir vertex function in eq .
( [ bs19 ] ) by this asymptotic form of the cooperon , i.e. eq . ( [ c7 ] ) . by way of this
, we obtain closed coupled equations for the ( renormalized ) diffusion constant @xmath82 ; @xmath103 since eq . ( [ c7 ] )
is valid only for small @xmath104 , we have imposed the additional constraint @xmath105 into these integral variables .
one might regard this upper limit as the mean - free path .
we have already taken in eq .
( [ sc2 ] ) both @xmath34 and @xmath35 to be zero .
thus , @xmath106 in the right hand side stands for @xmath107 estimated there ; @xmath108 where @xmath109 was already defined in eq .
( [ vde ] ) .
the normalized real - valued function @xmath110 used above is given only in terms of @xmath111 .
for example , @xmath112 is given as follows , @xmath113 with its normalization factor @xmath114 ; @xmath115 ( see the appendix .
d for its derivation ) .
thus , eqs .
( [ sc1]-[sc2 ] ) constitute closed coupled equations for the diffusion constant . to solve this gap equation ,
notice first that the coupling between the current and the @xmath54-type current is disconnected in the massless case ; @xmath116
. this can be seen directly from @xmath117_{\delta\gamma } & = & 0 .
\label{use1 } \end{aligned}\ ] ] which leads to @xmath118 . as a result of this , eqs .
( [ sc1]-[sc2 ] ) becomes linear in @xmath82 ; @xmath119 using eqs .
( [ g00j]-[ng00j ] ) , we can readily evaluate the momentum integral in the right hand side of eq .
( [ ml1 ] ) . to do this ,
introduce a new integral variable @xmath120 , so that @xmath121 .
moreover , we approximate @xmath122 in the integrand by @xmath123 , since @xmath112 is a slowly varying function in the scale of @xmath124 .
these treatments give us the following expression for @xmath125 ; @xmath126 where we used @xmath127 and @xmath128 . observing this expression ,
notice that the second member of the r.h.s . is nothing but the quantum correction to the diffusion constant , which basically corresponds to the awl correction to the conductivity . as a function of the chemical potential in the massless case.,scaledwidth=40.0% ] in the presence of relatively large topological mass ,
i.e. @xmath130 , the parity diffusion mode becomes the high - energy degree of freedom .
as such , only the first term in eq .
( [ dc12 - 11 ] ) contributes the diffuson . with eq .
( [ dc10 - 5 ] ) , such diffuson is given as follows ; @xmath131 where @xmath64 and @xmath132 are defined by @xmath133 , @xmath134 and @xmath135 ; @xmath136 contrary to the previous subsection , the tensor - part of the diffuson depends on the model - parameters through the `` tensor - form factor '' @xmath132 . as such , we will employ in this case not only the diffuson constant @xmath82 but also this tensor - form factor @xmath132 as the `` mean - field parameters '' , which should be self - consistently determined .
in other words , both of them should be given by the relaxation kernels , as in eq .
( [ bs25 ] ) . to do this
, we will first calculate both the density relaxation function @xmath137 and the sublattice density relaxation function @xmath138 , by the use of eq .
( [ massived1 ] ) .
namely , we will substitute eq .
( [ massived1 ] ) into eqs .
( [ bs12],[vde ] ) , only to obtain these two functions in terms of @xmath139 and @xmath132 first .
the relaxation functions thus calculated read as follows ; @xmath140 then , we will equate eqs .
( [ 0ive],[5ive ] ) with the first two members of eq .
( [ bs24 ] ) respectively . by way of this ,
@xmath141 and @xmath132 can be given in terms of the relaxation kernels ; @xmath142 by substituting these two back into eq .
( [ massived1 ] ) , we can express the diffuson only in terms of the relaxation kernels .
when its hole - line time reversed , the corresponding cooperon is readily derived ; @xmath143 the tensor - form factor @xmath132 and the diffusion constant @xmath82 appearing in the right hand side above are already given by the relaxation kernels , via eq .
( [ dc12 - 5 ] ) and eq .
( [ bs25 ] ) .
such relaxation kernels are given by the 2pir vertex function ( see eq .
( [ bs19 ] ) ) .
thus , as in the previous subsection , we will approximate the 2pir vertex function by eq .
( [ c6 ] ) . in terms of this substitution
, we arrive at a closed coupled equation for the diffusion constant @xmath82 and the tensor - form factor @xmath132 , whose explicit expressions are given in the appendix . c. when solving this gap equation
, we can see how the quantum correction to the diffusion constant behaves as a function of @xmath2 and @xmath3 .
several limiting values are summarized in fig .
[ t23 ] . especially , in the zero mass limit i.e. @xmath144 , the solution of the gap equation reduces to a following simple function of @xmath145 ; @xmath146 comparing this with eq .
( [ masslesssol ] ) , one can easily see that the quantum correction to the diffusion constant is actually _ half _ of that for @xmath5 case . )
corresponds to the green shaded region.,scaledwidth=40.0% ] the discrepancy between eq .
( [ masslesssol ] ) and eq .
( [ r5 ] ) is responsible for the cooperon term associated with the parity diffusion mode , i.e. the second member of the right hand side of eq .
( [ 162d ] ) . to see this explicitly ,
note first that @xmath147 , @xmath148 and @xmath149 appearing in eqs .
( [ dc12 - 13-c]-[dc12 - 13-d ] ) reduce to zero in the limit of @xmath150 .
then , eqs .
( [ dc12 - 13-c],[dc12 - 13-d ] ) in this limit read as follows ; @xmath151 observing these expressions , notice that the second members of both eq .
( [ r25 ] ) and eq .
( [ r26 ] ) are totally ineffective in the current - type relaxation kernels ; @xmath152_{\delta\gamma } = 0 , \label{use2 } \\ & & \big\{\hat{t_4}\cdot\big(\hat{1}+\hat{t}_1-\hat{s}_1-\hat{s}_2\big ) \big\}_{\alpha\delta,\gamma\beta}\hat{q}_{\mu } \big[\hat{\gamma}_{\mu5}\big]_{\delta\gamma } = 0 , \label{use3 } \end{aligned}\ ] ] the consequence of these two equations are two - fold . the second equation in combination with eq .
( [ use1 ] ) leads @xmath153 first .
thus , we have @xmath154 again , which indicates that @xmath155 originated from eq .
( [ r25 ] ) and that from eq .
( [ r26 ] ) contribute to the quantum correction in an _ additive _ way .
( [ use2 ] ) moreover indicates that these two quantum corrections have the _ same magnitude _ and _ sign_. in other words , the quantum correction derived in appendix .
b2a , i.e. eq . ( [ masslesssol ] ) , can be divided into two parts ; @xmath156 each of these two quantum corrections is originated from eq .
( [ r25 ] ) and eq .
( [ r26 ] ) respectively . since we have already ignored eq .
( [ r26 ] ) for the @xmath129 case , the resulting solution has only single @xmath157 , as in eq .
( [ r5 ] ) .
the mean - field equation for the diffusion constant @xmath82 and @xmath132 in the presence of the finite topological mass @xmath2 is given as follows ; @xmath159 with the relaxation kernels @xmath160 being given by @xmath82 and @xmath132 self - consistently ; @xmath161 note that @xmath106 above are previously defined ; @xmath162 with @xmath109 for @xmath163 and @xmath164 given in eq .
( [ vde ] ) .
@xmath110 used in @xmath165 are given in terms of the 1-point green functions ( @xmath111 and @xmath166 ) and the tensor - form factor @xmath132 ; @xmath167 with @xmath168 .
starting from the bethe - salpeter ( bs ) equation for the response function , we have derived in the section b1 the eom s for the various types of relaxation functions .
such coupled eom s have two features ; they are closed and _ linearized _ with respect to the relaxation functions .
because of these two features , we can solve them for the relaxation functions .
out of this solution , we can relate the renormalized diffusion constant with the 2pir ( two - particle irreducible ) vertex function .
this relation in turn becomes an essential building - block of the self - consistent loop of the diffusion constant ( see appendix .
b ) . to obtain such linearized eom s , we need to reduce the convolution part between the 2pir vertex function and the response function into the simple product between relaxation kernels and relaxation functions .
for this purpose , we have introduced the completeness in the space of the integral variable , say @xmath169 or @xmath170 , associated with this convolution ;
@xmath171 namely , by use of this , any convolution in principle can be decomposed into a simple product ; @xmath172 in the current context , @xmath173 corresponds to the 2pir vertex function , while @xmath174 to the response function .
therefore , @xmath175 corresponds to the relaxation kernels , while @xmath176 does to the relaxation functions ( see also appendix . b1 ) .
the trade - off for this decomposition is therefore the sum over _ infinite _ ( but countable ) numbers of modes specified by @xmath177 . to be more specific ,
we did this decomposition systematically , based on the completeness relation of the @xmath46-matrices and the spherical harmonic function @xmath178 ; @xmath179_{\beta\alpha } y^{*}_{lm}(\hat{k } ) \phi_{\alpha\cdots,\cdots\beta}(|k|\hat{k},\cdots ) .
\label{b0}\end{aligned}\ ] ] as such , the momentum integral only over the angle - direction , @xmath180 , is taken , while that over its radial direction , @xmath181 , is _ not _ taken .
as for the convolution with respect to this radial direction , we simply replace the @xmath182-dependence of @xmath183 by some real - valued function @xmath184 .
namely , we rewrite the right hand side of eq .
( [ b0 ] ) as follows ; @xmath185_{\beta\alpha } y^{*}_{lm}(\hat{k } ) \phi_{\alpha\cdots,\cdots\beta}(|k|\hat{k},\cdots ) \nn \\ & & \hspace{-0.5 cm } = g_{lm,\mu}(|k|)\sum_{k } \sum_{\alpha,\beta } \big[\hat{\gamma}_{\mu}\big]_{\beta\alpha } y^{*}_{lm}(\hat{k } ) \phi_{\alpha\cdots,\cdots\beta}(k,\cdots ) .
\label{b1 } \end{aligned}\ ] ] let us justify this treatment of the radial direction . in the response function , @xmath186 , @xmath35 and @xmath34 are associated with the external momentum and frequency for the bosonic degrees of freedom .
we can take these two to be small , as far as the relaxation functions for the long wave - length and low - energy region is concerned .
then , such a response function is usually dominated by the diffuson @xmath32 ( see eq .
( [ b1d ] ) ) . as a result of this , the @xmath181-dependence in the left hand side of eq .
( [ b1 ] ) and its @xmath35 , @xmath34-dependence can be factorized at the leading order in small @xmath35 and @xmath34 ; @xmath187_{\beta\alpha } y^{*}_{lm}(\hat{k } ) \phi_{\alpha\gamma,\gamma\beta}(x\hat{k},k';q,\omega ) \nn \\&&\hspace{-0.4 cm } = g_{lm,\mu}(x)\sum_{k , k ' } \sum_{\alpha,\beta,\gamma } \big[\hat{\gamma}_{\mu}\big]_{\beta\alpha } y^{*}_{lm}(\hat{k } ) \phi_{\alpha\gamma,\gamma\beta}(k , k';,q,\omega ) \nn \\ & & \hspace{-0.4 cm } = g_{lm,\mu}(x ) \times \phi_{lm,\mu}(q,\omega ) .
\label{b3}\end{aligned}\ ] ] to see this factorization more explicitly , one can take the following steps : ( i ) substitute the asymptotic tensor - form of the diffuson into eq .
( [ b1d ] ) and the left hand side of eq .
( [ b3 ] ) , ( ii ) take the integral and the sum over @xmath180 , @xmath188 , @xmath18 , @xmath189 and @xmath46 in eq ( [ b3 ] ) , and ( iii ) retain the leading order in small @xmath35 and @xmath34 . by way of this , one can reach the factorization given in the right hand side of eq .
( [ b3 ] ) with a specific @xmath190 .
for example , let us follow these prescriptions in the case of zero topological mass case .
observing eq .
( [ masslessd ] ) , notice first the following relation ; @xmath191_{\delta,\gamma } \equiv 0 , \label{mlsim}\end{aligned}\ ] ] for @xmath192 . using this
, one can readily checked that the diffuson in this case turns out to be proportional to the unit matrix , when its right hand side is traced out ; @xmath193 as such , to obtain the normalized function @xmath110 in the massless case , we have only to calculate the following quantity up to the leading order in small @xmath34 and @xmath35 ; @xmath194_{\beta\alpha}\phi_{\alpha\delta,\delta\beta}(k , k^{\prime};q,\omega ) \nn \\ & & \hspace{-0.5 cm } \propto \frac{1}{i\omega - dq^2 } \sum_{\hat{k } } \sum_{\alpha,\delta,\rho } \hat{g}^{r}_{\delta\alpha}(k_{+},\mu_{+ } ) \hat{g}^{a}_{\alpha\rho}(k_{-},\mu_{- } ) \big[\hat{v}_{a}(q)\big]_{\rho\delta}. \nn \end{aligned}\ ] ] for example , taking the current component as @xmath195 above , we have ; @xmath196_{\beta\alpha } \phi_{\alpha\delta,\delta\beta}(k , k^{\prime};q,\omega ) \nn \\ & & \hspace{-0.2 cm } \propto - \frac{q}{i\omega - dq^2 } \big\{\frac{1}{|f^2_0-k^2|^2 } + \frac{8}{3 } \frac{{f^{\prime}_0}^2 k^2}{|f^2_0-k^2|^4}\big\ } + { \cal o}(q^2,\omega ) .
\nn\end{aligned}\ ] ] observing the right hand side , one can then convince oneself of eq .
( [ b3 ] ) .
moreover , the normalized real - valued function @xmath112 will be obtained as in eqs .
( [ g00j],[ng00j ] ) .
99 s. murakami , n. nagaosa and s .- c .
zhang , science 301 , 1348 ( 2003 ) .
j. sinova _
et al . _ , phys .
* 92 * , 126603 ( 2004 ) .
y. k. kato , r. c. myers , a. c. gossard , and d. d. awschalom , science * 306 * , 1910 ( 2004 ) j. wunderlich , b. kstner , j. sinova and t. jungwirth , phys .
. lett . * 94 * , 047204 ( 2005 ) . c. l. kane and e. j. mele , phys .
* 95 * , 146802 ( 2005 ) . c. l. kane and e. j. mele , phys .
* 95 * , 226801 ( 2005 ) .
b. a. bernevig and s .- c .
zhang , phys .
96 * , 106802 ( 2006 ) .
y. hatsugai , phys .
* 71 * , 3697 ( 1993 ) . c. wu , b. a. bernevig and s .- c .
zhang , phys .
96 * , 106401 ( 2006 ) . c. xu and j. e. moore , phys .
b * 73 * , 045322 ( 2006 ) .
l. fu and c. l. kane , phys .
b * 74 * , 195312 ( 2006 ) .
t. fukui and y. hatsugai , phys .
b * 75 * , 121403(r ) ( 2007 )
. b. a. bernevig , t. l. hughes and s .- c .
zhang , science * 314 * , 1757 ( 2006 ) .
l. fu , c. l. kane and e. j. mele , phys .
* 98 * , 106803 ( 2007 ) .
r. roy , cond - mat/0607531 .
j. e. moore and l. balents , phys .
b * 75 * , 121306 ( r ) ( 2007 ) .
s. murakami , phys .
lett . * 97 * , 236805 ( 2006 ) .
d. hsieh _ et al .
_ , nature 452 , 970 , ( 2008 ) .
jeffrey c. y. teo , liang fu , and c. l. kane , phys .
b * 78 * , 045426 ( 2008 ) .
a. p. schnyder , s. ryu , a. furusaki , and a. w. w. ludwig , arxiv:0803.2786 e. abrahams , p. w. anderson , d. c. licciardello and t. v. ramakrishnan , phys . rev . lett . * 42 * , 673 ( 1979 ) d. vollhardt and p. wlfle , phys .
lett . * 45 * 842 ( 1980 ) ; phys .
b * 22 * , 4666 ( 1980 ) .
s. murakami , new j. phys .
* 9 * , 356 ( 2007 ) ; ( corrigendum ) _ ibid . _ * 10 * , 029802 ( 2008 ) s. murakami and s. kuga , arxiv:0806.3309 ( 2008 ) .
m. onoda , y. avishai and n. nagaosa , phys .
. lett . * 98 * , 076802 ( 2007 ) . h. obuse , a. furusaki , s. ryu and c. mudry , phys .
b * 76 * , 075301 ( 2007 ) . a. m. essin and j. e. moore , phys . rev .
b * 76 * 165307 ( 2007 ) .
m. ostrovsky , i. v. gornyi , and a. d. mirlin , phys .
rev . lett .
* 98 * , 256801 ( 2007 ) .
s. ryu , c. mudry , h. obuse , and a. furusaki , phys .
* 99 * , 116601 ( 2007 ) k. nomura , m. koshino , s. ryu , phys .
99 * , 146806 ( 2007 ) .
j. h. bardarson , j. tworzydo , p. w. brouwer and c. w. j. beenakker , phys .
lett . * 99 * , 106801 ( 2007 ) .
h. aoki and t. ando , phys .
lett . * 54 * , 831 ( 1985 ) . although the previous numerical studies done by onoda _ et.al . _ @xcite are about the @xmath197-dimensional case
, we expect that this phase diagram holds true both in the 2-d @xmath0 qsh systems and in the 3-d @xmath0 qsh systems .
r. shindou _ et al .
_ , in progress . | in this paper , we address ourselves to the nonmagnetic disorder effects onto the quantum critical point which intervenes the 3-dimensional @xmath0 quantum spin hall insulator ( topological insulator ) and an ordinary insulator .
the minimal model describing this type of the quantum critical point is the single - copy of the @xmath1 dirac fermion , whose topological mass @xmath2 induces the phase transition between the topological insulator and an ordinary one .
we first derive the phase diagram spanned by the mass - term @xmath2 , chemical potential @xmath3 and strength of the disorder within the self - consistent born approximation . by way of this
, we find a finite density of state appears even at the zero - energy and at the phase transition point , i.e. @xmath4 , if the strength of the disorder potential exceeds some critical value . to infer the structure of the low - energy effective theory around these zero - energy states , we further calculated the weak localization ( wl ) correction to the conductivity . to be more specific
, we have found that the diffuson is dominated by the charge diffusion mode and parity diffusion mode .
while the charge diffusion mode always carries the diffusion pole , the parity diffusion mode becomes massless only at @xmath5 , but suffers from the infrared cutoff for non - zero @xmath2 .
corresponding to this feature of the diffuson , the cooperon is also composed of two quasi - degenerate contributions .
we found that these two give rise to the same magnitude of the anti - weak - localization ( awl ) correction with each other at @xmath5 . as a result , when the topological mass @xmath2 is fine - tuned to be zero ( but for generic @xmath3 ) , the awl correction becomes doubled ( quantum correction doubling ) . based on this observation , we will discuss the possible microscopic picture of the `` levitation and pair annihilation '' phenomena , recently discovered by onoda _
et al _
@xcite . |
we started with j. nagy our study of multifractal properties of typical / generic functions in @xcite where multifractal properties of generic monotone functions on @xmath0 $ ] were treated .
the higher dimensional version of this question was considered in @xcite where with s. seuret we investigated the hlder spectrum of functions monotone in several variables . in @xcite
we also showed that typical borel measures on @xmath0^d$ ] satisfy a multifractal formalism .
multifractal properties of typical convex continuous functions defined on @xmath15^d}$ ] are discussed in @xcite . in @xcite the convex hull of typical continuous functions
@xmath22 $ ] is considered by a. m. bruckner and j. haussermann . in this case
the boundary of this convex hull decomposes into two functions ( in our notation ) @xmath20 and @xmath21 see figure [ fig1 ] .
the upper one , @xmath20 is concave the lower one @xmath21 is convex .
it is shown that for the typical @xmath5 these functions are continuously differentiable on @xmath23 and at the endpoints they have infinite derivatives .
the aim of our paper is to describe the multifractal spectrum of these functions in the multidimensional setting .
that is , generic / typical continuous functions @xmath5 in @xmath24^d}$ ] are considered in the sense of baire category .
the topology on @xmath24^d}$ ] is the supremum metric .
we also prove the multidimensional version of the above mentioned results of a. m. bruckner and j. haussermann .
the convex hull of the graph of @xmath25^d}$ ] is denoted by @xmath4 , that is , @xmath26^d } \}.$ ] we are interested in two functions : @xmath6 is the function on the top of @xmath4 , and @xmath7 is the function on the bottom of @xmath4 . on figure [ fig1 ] these functions are illustrated in dimension one .
the points where @xmath5 and @xmath27 coincide are denoted by @xmath28 , @xmath11 .
the hlder exponent and singularity spectrum for a locally bounded function is defined as follows .
let @xmath29^ { d}})$ ] .
for @xmath30 and @xmath31^ { d}}$ ] , the function @xmath5 belongs to @xmath32 if there are a polynomial @xmath33 of degree less than @xmath34 $ ] and a constant @xmath35 such that , for @xmath36 close to @xmath37 , latexmath:[\[\label{defpoint } the pointwise hlder exponent of @xmath5 at @xmath37 is @xmath39 the singularity spectrum of @xmath5 is defined by @xmath40 here @xmath41 denotes the hausdorff dimension , and @xmath42 by convention .
we will also use the sets @xmath43 the faces of @xmath15^d}$ ] are @xmath44^d } \}\ ] ] and @xmath45^d } \}.\ ] ] then @xmath46^d})=\cup_{i=1}^{2}\cup_{j=1}^{d}f_{i , j}$ ] .
the main result of our paper is the following theorem : [ * 30th ] there exists a dense @xmath47 set @xmath9^d}$ ] such that for every @xmath48 for @xmath11 * @xmath12 is continuously differentiable on @xmath49 ; * if @xmath50^d})$ ] then @xmath51 hence @xmath52 and @xmath53 ; * @xmath54 ; * @xmath55 ; * @xmath56 , that is , @xmath57 for @xmath19 ; * for @xmath58 if @xmath59 then @xmath60 if @xmath61 then @xmath62 * @xmath63 .
the open ball with center @xmath64 and of radius @xmath65 is denoted by @xmath66 .
we use similar notation for open neighborhoods of subsets , for example if @xmath67 then @xmath68 in our proofs we use a standard fixed countable dense set @xmath75 in @xmath76^d}$ ] .
we assume that the functions @xmath77 are in @xmath78^d}$ ] .
( taking an arbitrary countable dense set @xmath79 in @xmath24^d}$ ] by using a mollifier function it is easy to obtain such a dense set of @xmath80 functions . )
[ * 1bbdef ] we say that @xmath81^d}\to { \ensuremath { \mathbb r}}$ ] is _ piecewise linear _ if there is a partition @xmath82 , @xmath83 of @xmath15^d}$ ] into simplices such that for each @xmath72 the set @xmath84 is the subset of a hyperplane in @xmath85 . + we say that @xmath5 is _ independent piecewise linear _ if it is piecewise linear and if @xmath86 denotes the collection of the vertices of the simplices @xmath82 , @xmath83 for a suitable partition then * from @xmath87 , @xmath88 , @xmath89 , the points @xmath90 , @xmath91 are on the same @xmath92-dimensional hyperplane in @xmath85 it follows that @xmath93 , * from @xmath94 , @xmath88 , @xmath89 , the points @xmath95 , @xmath91 are on the same @xmath96-dimensional hyperplane in @xmath73 it follows that @xmath97 . by carathodory s theorem
( see for example @xcite ) from @xmath107 it follows that one can find ( not necessarily different ) @xmath108^d}$ ] , @xmath101 , @xmath100 , @xmath102 such that @xmath109 , @xmath110 if for an @xmath111 we had @xmath112 then letting @xmath113 we would obtain @xmath114 contradicting that @xmath115 . without limiting generality
we prove the statement for @xmath20 .
+ suppose @xmath121 .
by lemma [ * 1clem ] there exist @xmath122^d}$ ] , and @xmath101 , @xmath100 , such that @xmath102 , @xmath123 this and @xmath124 imply @xmath125 put @xmath126 for all @xmath127^d}$ ] .
we also let @xmath128 it is clear that @xmath129 is dense and open in @xmath76^d}$ ] and @xmath130 is dense @xmath47 .
suppose @xmath117 .
then there exists a sequence @xmath131 , @xmath132 such that @xmath133 .
since @xmath134 by we have @xmath135 therefore , using @xmath136 on the other hand , since @xmath137 is concave @xmath138 @xmath139 ( using and ) @xmath140 taking difference @xmath141 @xmath142 ( by the mean value theorem and the choice of @xmath143 ) @xmath144 choosing @xmath145 by we obtain @xmath146 hence , @xmath147 this implies .
again we start with @xmath156^d}$ ] , @xmath157 a countable dense set in @xmath76^d}$ ] .
this time we select @xmath158 we also put @xmath159 we put @xmath160 and select @xmath161 then we can choose @xmath131 such that @xmath162 since @xmath167 is a concave function in @xmath168 it is sufficient to verify that its derivative exists at @xmath169 for any choice of @xmath170 .
this will imply that @xmath171 is monotone decreasing in @xmath168 , without any jump discontinuities , hence for a fixed @xmath64 it is continuous as a function of one variable . in the end of this proof
we will provide a standard argument showing that from the concavity and continuity of @xmath137 one can deduce that @xmath172 is continuous on @xmath13 . by the one dimensional taylor s formula one can find @xmath177 such that @xmath178 and @xmath179 @xmath180 this implies @xmath181 @xmath182 ( using ) @xmath183 using @xmath184 , @xmath185 and @xmath186
we can continue by @xmath187 by and we obtain that @xmath188 and similarly @xmath189 and hence , @xmath190 @xmath191 since @xmath20 is concave this implies that @xmath192 exists .
next we verify that @xmath172 is continuous on @xmath49 .
we have seen that @xmath193 is a monotone decreasing continuous function in @xmath168 for a fixed @xmath194 .
we need to show that @xmath195 is continuous as a function of several variables at any @xmath196 .
this is quite standard .
suppose @xmath197 and @xmath198 are fixed .
choose @xmath199 such that @xmath200 the function @xmath201 is continuous as the top part " of the convex hull @xmath4 of the continuous function @xmath5 .
by uniform continuity of @xmath201 choose @xmath202 such that @xmath203 now suppose that @xmath204 we choose again a countable dense set @xmath226^d}$ ] .
the functions @xmath77 are uniformly continuous and for @xmath227 we choose @xmath228 such that @xmath229^d}.\ ] ] we partition @xmath15^d}$ ] into non - overlapping simplices @xmath82 , @xmath230 such that the diameter of each simplex is less than @xmath231 .
we assume that @xmath232 is the set of vertices of these simplices .
we can also assume that these vertices are sufficiently independent , that is , from @xmath233 , @xmath88 , @xmath89 , the points @xmath95 , @xmath91 are on the same @xmath234-dimensional hyperplane in @xmath73 it follows that @xmath97 .
this means that the second assumption in definition [ * 1bbdef ] is satisfied .
we denote by @xmath238 the function which is defined on @xmath232 and for any @xmath239 @xmath240 on figure [ fig0 ] we illustrate the procedure of selecting @xmath236 . on the left half of this figure
there is @xmath77 with the little circles on its graph .
we suppose that on @xmath0 ^ 1 $ ] we used the simplices " , which are equally spaced line segments of length @xmath241 .
the function @xmath238 is defined on these points and its graph is represented by the little circles on the graph of @xmath77 .
now we perturb @xmath238 a bit in order to obtain an independent " function @xmath242 such that @xmath243 and if @xmath244 , @xmath88 if @xmath89 , and @xmath245 , @xmath91 are on the same hyperplane in @xmath85 then @xmath93 .
this implies that the first assumption of definition [ * 1bbdef ] is satisfied for @xmath246 .
if @xmath247^d}\sm { v(n , m)}$ ] and @xmath64 is in the simplex @xmath82 , @xmath248 with vertices @xmath249 we define @xmath250 so that @xmath251 is on the hyperplane determined by the points @xmath252 therefore , @xmath253 is an independent piecewise linear function ( recall definition [ * 1bbdef ] and see the illustration on the left half of figure [ fig0 ] ) .
we denote by @xmath259 the minimum distance among points of @xmath232 and will select a sufficiently small @xmath260 later .
we put @xmath261 on both halves of figure [ fig0 ] one can see @xmath262 which is the function with the equally spaced small peaks at the points which are multiples of @xmath241 . on the right half of figure [ fig0 ]
one can see @xmath236 which is obtained from @xmath253 ( pictured on the left half of figure [ fig0 ] ) after we added the small peaks .
we suppose that @xmath263 and if @xmath264 denotes the maximum of the norms of the gradients of the hyperplanes in the definition of @xmath253 then @xmath265 this way if we take the convex hull of @xmath266 then @xmath267 will be a subset of @xmath232 .
see the right half of figure [ fig0 ] .
we remark that the resolution of our drawings does not make it possible to take into all the above assumptions and hence they are distorted , but we hope that they can help to understand our procedure .
our next aim is to select a sufficiently small @xmath268 it is clear that given @xmath269 if @xmath270 is sufficiently small then @xmath5 can coincide with @xmath20 only close to some points in @xmath271 , that is for any @xmath272 and any @xmath273 there is @xmath274 such that @xmath275 one can see @xmath236 and @xmath5 .
we also graphed the functions @xmath237 and @xmath20 which will be very close to each other .
the latter function is not exactly piecewise linear but a close approximation to such a function , namely to @xmath237 . in the one dimensional case ,
like on figure [ fig2 ] the nonlinear parts ( not pictured ) are very cloes to some elements of @xmath271 .
the higher dimensional case is a bit more complicated and we discuss it below .
we will select a sufficiently small @xmath269 later . at this point
we suppose that @xmath279 this implies that for any @xmath273 there is a unique @xmath274 .
that is , @xmath5 will almost " look like a piecewise linear function .
this implies that if @xmath272 then @xmath280 we can suppose that @xmath281 is chosen so small , that @xmath282 this estimate will imply that @xmath283 for the typical @xmath284^d}$ ] , that is for @xmath223 . by the choice of @xmath285
if we consider @xmath237 then it is an independent piecewise linear function .
there is a system of non - overlapping simplices @xmath286 such that @xmath287 for any @xmath247^d}$ ] is on a hyperplane determined by a simplex @xmath288 containing @xmath64 . on the left half of figure [ fig2 ] the one dimensional case
is illustrated and these simplices are simply the line segments @xmath289 $ ] , @xmath290 $ ] and @xmath291 $ ] . the endpoints @xmath292 and @xmath293 are points where this function `` breaks '' and these points are on two non - parallel lines ( `` hyperplanes '' ) . these breakpoints / folding regions will be used to find those points where the hlder exponent is @xmath294 .
on the right half of figure [ fig2 ] the two dimensional case @xmath278 is illustrated .
this time we have simplices ( triangles ) in @xmath0^{d}$ ] bounded by solid lines on which @xmath237 is linear . on the right half of figure [ fig2 ]
only the domain of @xmath237 is shown .
the system of the simplices ( triangles ) bounded with dashed lines will be simplices corresponding to @xmath20 .
later we will explain this in more detail .
we denote by @xmath295 the set of the vertices of the simplices @xmath288 , @xmath296 clearly , @xmath297 the union of the faces of these simplices will be denoted by @xmath298 if @xmath299 with @xmath300 then @xmath301 and @xmath302 are on different hyperplanes and hence the graph of @xmath237 breaks " at @xmath303 .
this implies that we can choose @xmath304 such that for any @xmath305 and for any hyperplane @xmath306 passing through @xmath307 one can choose a point @xmath308 such that @xmath309 and @xmath310 where we used that implies @xmath311 .
see figure [ fig4 ] .
it is also clear that if @xmath5 is a good approximation of @xmath266 then one can see similar breaking " properties on @xmath312 .
this time there are no folding edges " like in the case of @xmath266 on @xmath313 but there are regions around @xmath314 where we can see similar phenomena .
using that @xmath266 and @xmath237 are both independent piecewise linear functions one can see that @xmath315^d},\ f\in b ( { f _ { n , m}},\ddd ) \}\to 0\ ] ] @xmath316 apart from and we also assume that @xmath269 is chosen so small that @xmath317 and ( using that @xmath318 ) @xmath319 @xmath320 and @xmath321 recall that we started to make assumptions about @xmath270 in the paragraph containing .
the smaller @xmath281 we need to use the smaller @xmath270 .
next we suppose that using we chose a @xmath270 such that in addition to our other assumptions we have @xmath322 now we want to use the folding property in and for functions @xmath5 which approximate @xmath266 .
this time the folding edges " are not any more @xmath96-dimensional surfaces , but some neighborhoods of them . by and we will be able to bound the dimension of these regions .
suppose @xmath323 with @xmath324 with vertices @xmath325 .
since @xmath266 and @xmath237 are both independent piecewise linear functions there is @xmath326 such that if @xmath327 then one can choose vertices @xmath328 , @xmath329 such that @xmath330 moreover if @xmath331 denotes the simplex determined by @xmath332 then @xmath333 is on the surface of @xmath4 inside a hyperplane determined by @xmath334 , that is , @xmath335 is a face " of @xmath20 approximating @xmath336 . on the right half of figure [ fig2 ]
we have the two dimensional illustration . the simplices ( triangles ) @xmath337 are bounded by solid lines . the simplices ( triangles ) @xmath331 are bounded by dashed lines .
suppose @xmath348 .
then there exists a sequence @xmath131 such that @xmath349 . for each @xmath350 we can define the `` folding region '' as in .
since these regions depend on @xmath350 we denote them by @xmath351 .
set @xmath352 .
if @xmath353 then there exists an @xmath350 such that @xmath354 with a simplex @xmath355 and @xmath356 is the subset of a hyperplane in @xmath85 .
this implies that @xmath20 is locally linear in a neighborhood of @xmath64 and @xmath357 . using one
can easily see that @xmath358 on the other hand , from it also follows that if @xmath359 is a simplex such that its vertices are @xmath360 then there exists @xmath361 such that for @xmath362 , @xmath363 . since @xmath364 is a face " of @xmath20 if @xmath365 then @xmath64 can not belong to the interior of any other face " of @xmath20 .
hence @xmath366 since @xmath367 we obtain that @xmath368 . if @xmath369 then and imply that @xmath370 for @xmath58 a version of lemma [ * 2lem ] can provide dense @xmath47 sets @xmath371^d}$ ] such that and hold with @xmath92 replaced by @xmath72 .
if we use the faces @xmath372 instead of @xmath373 in analogous versions of lemma [ * 2lem ] we can obtain dense @xmath47 sets @xmath374^d}$ ] . | we study the singularity ( multifractal ) spectrum of the convex hull of the typical / generic continuous functions defined on @xmath0^{d}$ ] .
we denote by @xmath1 the set of points at which @xmath2^d}\to { \ensuremath { \mathbb r}}$ ] has a pointwise hlder exponent equal to @xmath3 .
let @xmath4 be the convex hull of the graph of @xmath5 , the concave function on the top of @xmath4 is denoted by @xmath6 and @xmath7 denotes the convex function on the bottom of @xmath4 .
we show that there is a dense @xmath8 subset @xmath9^d}$ ] such that for @xmath10 the following properties are satisfied .
for @xmath11 the functions @xmath12 and @xmath5 coincide only on a set of zero hausdorff dimension , the functions @xmath12 are continuously differentiable on @xmath13 , @xmath14 equals the boundary of @xmath15^d}$ ] , @xmath16 , @xmath17 and @xmath18 if @xmath19 .
, @xmath20 and @xmath21 in 1d , width=532,height=306 ] |
under the influence of gravity , disk galaxies are expected to assemble in an `` inside - out '' fashion : stars form first from high - density gas in the central region of the galaxy where the potential is deepest , and subsequently at increasing galacto - centric radii ( e.g. * ? ? ?
an immediate consequence of this formation scenario is that stars born at the same time and in the same region of a galaxy should have similar chemical compositions .
however , observations in our galaxy suggest that these initial conditions are not maintained .
@xcite argued that the sun was substantially more metal rich than nearby solar age stars and the local interstellar medium ( ism ) .
a recent recalibration of the geneva copenhagen survey ( gcs ) using the infrared flux method finds no discrepancy with solar age stars @xcite , but even modern studies confirm that the age - metallicity relationships ( amrs ) of field and solar neighborhood stars are characterized by higher dispersions than expected @xcite . in addition , simple chemical evolution models that divide the galaxy into independently evolving concentric annuli predict many more low metallicity g - dwarfs in our region of the disk compared to those observed , a discrepancy known as `` the local g - dwarf problem '' @xcite . evidence seemingly in contradiction to standard galaxy chemical evolution theory is not limited to our own galaxy .
metallicity gradients in disk galaxies are shallower than predicted by classical models ( e.g. , * ? ? ?
@xcite and @xcite find unexpectedly old stellar populations on nearly circular orbits in the outskirts of m31 and m33 , respectively .
the outermost regions of ngc300 and ngc7739 show flattened or positive abundance gradients with radius @xcite .
these perplexing observations can not be readily explained within the confines of classic galaxy formation models .
a natural explanation for the observational challenges above arises if the present day radii of many stars could be significantly different from their birth radii .
one difficulty in establishing radial migration as a common phenomenon , from a dynamical standpoint , lies in finding a mechanism that can cause a substantial fraction of stars to migrate several kiloparsecs while retaining the observed approximately circular orbits . in a seminal paper , ( *
* hereafter sb02 ) investigated the relationship between changes in stellar angular momentum and disk heating .
they found that radial migration is a ubiquitous process in spiral galaxies ; stars naturally migrate ( change angular momentum ) as they resonantly interact with transient spiral waves .
stars in corotational resonance ( cr ) with said waves are scattered without heating the disk and maintain their nearly circular orbits ( unlike lindblad resonance ( lr ) scattering ) . in the present paper , we examine radial migration in simulations of disk galaxies .
our experiments include galactic disks evolved both in isolation and under the action of infalling satellites of the type expected in the currently favored cold dark matter ( cdm ) paradigm of hierarchical structure formation ( e.g. , * ? ? ?
* ; * ? ? ?
the latter set of experiments were presented in the studies of ( * ? ? ?
* hereafter k08 ) and @xcite and were utilized to investigate the _ generic _ dynamical and morphological signatures of galactic disks subject to bombardment by cdm substructure . inspired by sb02 , several groups have recently investigated the potential role of radial mixing in the chemical and dynamical evolution of disk galaxies .
@xcite presented the first chemical evolution model to incorporate radial migration .
the rate at which stars migrate via the sb02 mechanism is left as a free parameter constrained by the metallicity distribution function ( mdf ) of solar neighborhood stars in the gcs @xcite .
their model successfully reproduced , within systematic uncertainties , the observed age - metallicity distribution of stars in the gcs @xcite and the observed correlation between tangential velocity and abundance pattern described by @xcite .
however , there is partial degeneracy between the magnitude of radial migration and other parameters in the model such as star - formation rates and gas inflow characteristics .
furthermore , it is unclear whether the level of migration required to fit the data is consistent with theoretical expectations .
numerical simulations have confirmed the occurrence of radial migration under a variety of conditions .
@xcite studied the migration of stars in a simulation of an isolated milky way ( mw)-sized stellar disk formed from the cooling of a pressure - supported gas cloud in a @xmath3 dark matter halo . in their simulations ,
some older stars radially migrated to the outskirts of the disk while maintaining nearly circular orbits , forming a population akin to that observed in m31 and m33 @xcite .
@xcite found that @xmath4 of all stars in the solar neighborhood were not born _ in situ _ ; this is a natural explanation for the observed dispersion in the amr and solar neighborhood metallicity distribution function ( mdf ) .
more recently , @xcite investigated radial migration in a stellar disk perturbed by a low - mass ( @xmath5 ) orbiting satellite .
their numerical simulations integrated test particle orbits in a static galactic potential and highlighted the fact that mergers and perturbations from satellite galaxies and subhalos can induce stellar radial mixing .
although informative , test particle simulations in a static isothermal potential will not capture all the relevant physics of the process of stellar radial migration in disk galaxies ; the interactions between the gravitational perturbations and the self - gravity of the disk are essential to a detailed analysis of the phenomenon . in our paper
, we expand on the analysis of @xcite by investigating radial migration using fully self - consistent numerical simulations both with and without satellite bombardment .
there are now several established phenomena that can cause a star to populate a region of the disk different from its birth radii .
stars on elliptical orbits maintain their guiding center and angular momentum ( modulo asymmetries in the potential ) but can be found over the range in galacto - centric radius defined by their pericenter and apocenter . changing a star s angular momentum , and hence its guiding center ,
requires direct scattering or a resonant interaction with transient patterns in the disk .
the local encounters of stars with molecular clouds ( e.g. * ? ? ? * ) or lindblad resonance ( lr ) scattering between stars and spiral waves ( e.g. sb02 ) both change stellar guiding centers ( albeit to a relatively small degree ) and increase the random motions of stars over time , `` blurring '' the disk .
stars scattered at cr with spiral waves can change their guiding centers by several kiloparsecs without increasing the amplitude of their radial motion . for any single spiral wave , sb02 predict that stars are scattered on each side of the cr , `` churning '' the contents of the disk .
stars may undergo several encounters with transient spiral waves throughout their lifetimes . while sb02 investigate all resonant interactions between spiral waves and stars
, we will refer to this special case of cr as the `` sb02 mechanism '' .
as sb02 note , migration due to spiral waves can be described by blurring and churning regardless of how the waves arise ( satellites , e.g. , could induce spiral structure that would lead to migration described by sb02 ) .
simulations have shown that other transient wave patterns internal to a galaxy , such as those resulting from bar propagation , can produce resonance overlap with existing spiral patterns and induce radial migration @xcite .
orbiting satellites , external to the galaxy and discussed above , will have a complex interaction with the disk as they provide a means of direct scattering over a large area and also induce spiral modes in the disk . in this work
, we aim to characterize radial migration induced by satellite bombardment and compare its effects on the stellar disk with those observed in secularly evolved galaxies .
our investigation complements earlier and ongoing radial migration studies .
we perform a simulation campaign , including numerical experiments of isolated disk galaxies with different scale heights and gas fractions , which in turn lead to different levels of spiral structure . for the first time
, we examine the effect of satellite bombardment on radial migration utilizing simulations where galactic disks are subjected to a cosmologically motivated satellite accretion history . via a comparative approach ,
we determine how the magnitude and efficiency of radial migration depend on input physics , establish correlations between orbital parameters and migration , and present evidence that each of the three migration mechanisms is distinct in the examined parameter space .
these characteristics lead to possible observational signatures that may constrain the relative importance of each migration mechanism in the milky way .
we employ the method of @xcite to construct numerical realizations of self - consistent , multicomponent disk galaxies .
these galaxy models consist of an exponential stellar disk , a hernquist bulge @xcite , and a ( * ? ? ?
* hereafter nfw ) dark matter halo .
they are characterized by @xmath6 free parameters that may be tuned to fit a wide range of observational data for actual galaxies including the mw and m31 .
the @xcite models are derived from three - integral , composite distribution functions and thus represent self - consistent equilibrium solutions to the coupled poisson and collisionless boltzmann equations .
owing to their self - consistency , these galaxy models are ideally suited for investigating the complex dynamics involved in the process of radial mixing .
the @xcite method has been recently used in a variety of numerical studies associated with instabilities in disk galaxies , including the dynamics of warps and bars @xcite , the gravitational interaction between galactic disks and infalling satellites ( @xcite ; k08 ; @xcite ; @xcite ) , and the transformation of disky dwarfs to dwarf spheroidal galaxies under the action of tidal forces from a massive host @xcite .
we refer the reader to @xcite for an overview of all relevant parameters and a detailed description of this technique . for the majority of numerical experiments in the present study , we employ model `` mwb '' in @xcite , which satisfies a broad range of observational constraints on the mw galaxy . specifically , the stellar disk has a mass of @xmath7 , a radial scale length of @xmath8 , and a sech@xmath9 scale height of @xmath10 .
we note that the adopted value for the scale height is consistent with that inferred for the old , thin stellar disk of the mw ( e.g. , * ? ? ?
* ; * ? ? ?
the equivalent exponential scale height is approximately @xmath11 pc , but the sech@xmath9 vertical distribution is more accurate .
the bulge has a mass and a scale radius of @xmath12 and @xmath13 , respectively .
the nfw dark matter halo has a tidal radius of @xmath14 to keep the total mass finite ( k08 ) , a mass of @xmath15 , and a scale radius of @xmath16 .
the total circular velocity of the galaxy model at the solar radius , @xmath17 , is @xmath18 , and the toomre disk stability parameter is equal to @xmath19 at @xmath20 .
we note that direct numerical simulations of the evolution of model mwb in isolation confirm its stability against bar formation for @xmath21 gyr .
we wish to address the dependence of radial mixing in isolated disk galaxies upon initial disk thickness and the presence of gas in the galactic disk .
because of their smaller `` effective '' toomre @xmath22 stability parameter ( e.g. , * ? ? ?
* ) , thinner disks are more prone to gravitational instabilities and thus yield stronger and better defined spiral structure
. therefore , we might expect them to cause enhanced radial mixing compared to their thicker counterparts .
similarly , because radiative cooling in the gas damps its random velocities , the presence of gas is associated with long - lived spiral structure in disks ( e.g. , * ? ? ?
* ) , which may act to increase the amount of radial mixing . correspondingly , we initialize several additional disk galaxy models .
the first modified galaxy model was constructed with the same parameter set as mwb but with a scale height @xmath23 times smaller ( @xmath24 ) . except for disk thickness and vertical velocity dispersion ,
all of the other gross properties of the three galactic components of this model are within a few percent of the corresponding ones for mwb .
not surprisingly , being more prone to gravitational instabilities , this model develops a bar inside @xmath25 at time @xmath26 gyr .
the second set of modified galaxy models are the same as mwb except for the fact that a fraction @xmath27 of the mass of the initial stellar disk is replaced by gas .
thus , the resulting gaseous component is constructed with the same initial density distribution as the stellar disk . in the present study
, we employ two values for the gas fraction , @xmath28 and @xmath29 , and the adopted methodology will be described in detail in a forthcoming paper ( kazantzidis et al .
2011 , in preparation ) .
briefly , the construction of the gas disk is done by assuming that its vertical structure is governed by hydrostatic equilibrium and by computing the potential and the resulting force field for the radially varying density structure of the gas component . by specifying the polytropic index and the mean molecular weight of the gas and using a tree structure for the potential calculation ,
the gas azimuthal streaming velocity is determined by the balance between gravity and centrifugal and pressure support . in the standard cdm paradigm of hierarchical structure formation
, galaxies grow via continuous accretion of smaller satellite systems . to investigate the effect of accretion events on radial mixing in galactic disks , we have also analyzed @xmath30-body simulations of the gravitational interaction between a population of dark matter subhalos and disk galaxies .
these simulations were first presented by k08 , and we summarize them briefly here .
k08 performed high - resolution , collisionless @xmath30-body simulations to study the response of the galactic model mwb discussed above to a typical -motivated satellite accretion history .
the specific merger history was derived from cosmological simulations of the formation of galaxy - sized cdm halos and involved six dark matter satellites with masses , orbital pericenters , and tidal radii of @xmath31 , @xmath32 , and @xmath33 , respectively , crossing the disk in the past @xmath34 gyr . as a comparison
, the large magellanic cloud has an estimated _ total _ present mass of @xmath35 ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , so the masses of the perturbing dark matter satellites correspond to the upper limit of the mass function of observed satellites in the local group .
k08 modeled satellite impacts as a sequence of encounters . specifically , starting with the first subhalo , they included subsequent systems at the epoch when they were recorded in the cosmological simulation .
we note that although k08 followed the accretion histories of host halos since @xmath36 , when time intervals between subhalo passages were larger than the timescale needed for the disk to relax after the previous interaction , the next satellite was introduced immediately after the disk had settled from the previous encounter .
each satellite was removed from the simulation once it reached its maximum distance from the disk after crossing .
this approach was dictated by the desire to minimize computational time and resulted in a total simulation time of @xmath37 gyr instead of @xmath34 gyr that would formally correspond to @xmath38 .
k08 and @xcite found that these accretion events severely perturbed the galactic disk of model mwb without destroying it and produced a wealth of distinctive morphological and dynamical signatures on its structure and kinematics . in this paper
, we will investigate the magnitude of radial mixing induced in disk model mwb by these accretion events .
as in the case of isolated disk models , it is worthwhile to examine the dependence of radial mixing upon disk thickness . for this purpose , we employed the thinner galaxy model with a scale height of @xmath24 described above and repeated the satellite - disk encounter simulations of k08 .
all collisionless numerical simulations discussed in this paper were carried out with the multi - stepping , parallel , tree @xmath30-body code pkdgrav @xcite .
the hydrodynamical simulations were performed with the parallel treesph @xmath30-body code gasoline @xcite . in the gasdynamical experiments ,
we include atomic cooling for a primordial mixture of hydrogen and helium , star formation and ( thermal ) feedback from supernovae .
our star formation recipe follows that of @xcite , which is based on that of @xcite .
gas particles in cold and dense regions which are simultaneously parts of converging flows spawn star particles with a given efficiency @xmath39 at a rate proportional to the local dynamical time .
feedback from supernovae is treated using the blast - wave model described in @xcite , which is based on the analytic treatment of blastwaves described in @xcite . in our particular applications ,
gas particles are eligible to form stars if their density exceeds 0.1 atoms/@xmath40 and their temperature drops below @xmath41 k , and the energy deposited by each type - ii supernova into the surrounding gas is @xmath42 erg .
we note that this choice of parameters and numerical techniques is shown to produce realistic disk galaxies in cosmological simulations @xcite .
lastly , in an attempt to investigate the effect of star formation efficiency on radial mixing , for each of the gas fractions @xmath27 above , we adopted two different values for @xmath39 , namely @xmath43 ( which reproduces the slope and normalization of the observed schmidt law in isolated disk galaxies ) and a lower value of @xmath44 .
all @xmath30-body realizations of the disk galaxy models contain @xmath45 particles of @xmath46 in the disk ; @xmath47 , @xmath48 in the bulge ; and @xmath49 , @xmath50 in the dark matter halo .
the gravitational softening lengths for the three components were set to @xmath51 , @xmath52 , and @xmath53 , respectively .
these are `` equivalent plummer '' softenings ; the force softening is a cubic spline . in the hydrodynamical simulations of isolated galaxies ,
gaseous disks were represented with @xmath54 and @xmath55 for gas fractions of @xmath28 and @xmath29 , respectively . in these cases , the gravitational softening length for the gas particles was set to @xmath56 .
all numerical simulations of isolated and perturbed disks were analyzed at @xmath57 gyr .
this time - scale corresponds to approximately @xmath58 orbital times at the disk half - mass radius . for evaluating the impact of accretion events , it is fairest to compare isolated and perturbed disks after the same amount of integration , but we may underestimate the total amount of mixing in the isolated ( and , to a lesser extent the perturbed ) models . in the future ,
we plan to evolve selected simulations for the full @xmath59 gyr interval since @xmath38 , but such simulations will require more than three times the computational resources used here .
exchange of angular and linear momentum between the infalling satellites and the disk tilt the disk plane and cause the disk center of mass to drift from its initial position at the origin of the coordinate frame ( k08 ; @xcite ) .
therefore , we calculate @xmath60 in the satellite - disk encounter experiments after removing the global displacement and tilt of the disk galaxy by determining the principal axes of the total disk inertia tensor and rotating our original coordinate system such that it is aligned with this tensor .
figure [ fig : sims ] presents surface density maps of the stellar distributions of our four collisionless galactic disks with different initial scale - heights @xmath61 .
each simulation exhibits a distinctive combination of morphological features that could affect radial migration .
the initial smooth disks are unstable to spiral instabilities arising from the swing amplification of particle shot noise , an effect that leads to emerging spiral structure as seen in the final isolated disks ( e.g. * ? ? ?
owing to its smaller `` effective '' toomre @xmath22 stability parameter ( e.g. , * ? ? ?
* ) , the @xmath62 disk ( hereafter , disks with this initial scale height will be referred to as `` thin '' ) develops prominent spiral structure and a strong bar .
conversely , the @xmath63 disk ( hereafter , disks with @xmath63 will be referred to as `` thick '' ) has relatively little spiral structure and does not form a bar in isolation .
we emphasize that the `` thick '' , @xmath63 disk is the one in best agreement with the old , thin disk of the mw , while @xmath62 is too thin ( see section [ sec : isolated_disks ] ) .
the radial and vertical morphology of the perturbed disks is distinct from their isolated counterparts .
both perturbed disks develop bars and are characterized by prominent flaring and much larger scale heights than those of the isolated disks ( k08 ; @xcite ) .
there is evidence that some spiral structure evident in the isolated disks has been washed out in the simulations with substructure bombardment : within @xmath64 of the galactic center , local enhancements of the stellar surface density evident in the isolated disks are substantially more diffuse after the action of the infalling satellites . throughout this section
, we will investigate how the growth and dissipation of spiral structure affect radial migration .
while we have investigated similar maps in the four hydrodynamical simulations , we do not show them here as they are qualitatively similar to their collisionless counterparts . as expected , however , the magnitude of spiral structure increases as the gas fraction rises .
we note that the differential effect of gas on the strength of spiral structure we find here is smaller than that reported in previous numerical investigations of isolated disk galaxies ( e.g. , * ? ? ?
this is mainly due to the effect of stellar feedback , which causes the ism in our simulations to become turbulent and multi - phase ( see also * ? ? ?
we first investigate where particles move throughout each simulation .
figure [ fig : isodisks ] shows the distribution of @xmath60@xmath65 in the six isolated disk simulations , where @xmath66 and @xmath67 refer respectively to each particle s final and initial projected distance from the galactic center .
each panel contains the @xmath60 distribution , the median @xmath68 ( denoted @xmath69 ) , and @xmath70 percentile @xmath68 ( denoted @xmath71 ) for two of the six isolated galaxies in our simulation suite . the isolated , collisionless simulations ( left panel ) clearly demonstrate that disk scale height and gas content affect the radial migration process .
the thin disk s @xmath60 distribution ( black line ) is considerably broader than the thick disk s ( gray hatch ) .
both @xmath69@xmath72 and @xmath71@xmath73 of particles in the thin disk are approximately double those found in the thick disk .
the remaining panels of figure [ fig : isodisks ] compare the @xmath60distributions of the hydrodynamical simulations of our sample .
we examine four isolated disks with two initial gas fractions : @xmath74 ( middle panel ) and @xmath75 ( right panel ) .
the two histograms in each panel represent star formation efficiencies ( @xmath39 ) of @xmath76 ( solid line ) or @xmath77 ( gray , hatched histogram ) .
gas has a strong impact on particle radial migration : @xmath69 increases from @xmath78 in the collisionless case ( all hydrodynamical simulations are of `` thick '' disks ) to @xmath79 when @xmath74 and @xmath80 when @xmath75 .
the extent of radial migration is less dependent on star formation efficiency ; there are only minor differences amongst each pair of histograms in the middle and right panels . however , in both panels , @xmath69 and @xmath71 are highest when @xmath39@xmath81 . in these four hydrodynamical simulations , the time - averaged gas content of the disk
is directly correlated with the percentage of particles that migrate and their typical displacement relative to their formation radius .
decreasing the scale height of the stellar disk , increasing the fraction of the initial disk mass in gas , and lowering the star - formation efficiency all increase the midplane density of the disk .
this enhanced density increases the coherence and self - gravity of spiral structure , supporting it against the dissipative nature of individual particles random motions ( as described by , e.g. , * ? ? ?
the correlation of spiral structure strength with the fraction of particles that migrate away from their birth radii and the median migratory distance traversed suggests that the sb02 mechanism , intimately linked with spiral waves , plays a major role in the migration of particles in the isolated systems .
the symmetric shapes ( to within @xmath82 about @xmath83 ) of the @xmath60 distributions are also consistent with migration via the sb02 mechanism
. resonances between the bar and spiral structure could also influence the timescale for migration in the thin disk models @xcite , but the @xmath63 models do not develop bars , regardless of whether they include gas .
figure [ fig : perdisks ] compares the @xmath60 distributions of the isolated collisionless disks to those of their perturbed counterparts .
the @xmath69 and @xmath71 of each distribution are labeled and color - coded to match the corresponding histogram . in each panel ,
the most sharply peaked histogram ( red ) shows the expected @xmath60distribution from epicyclic motion alone , which we compute given the particle s initial phase space coordinates .
we model the epicycle as a simple harmonic oscillator about the particle s guiding center radius ( @xmath84 ) with an amplitude set by its initial radial energy ( @xmath85 ) ( chapter 3 , * ? ? ?
the energy associated with the circular component of the particle s orbit ( @xmath86 ) is set by @xmath84 , which is the radius of a circular orbit with angular momentum equal to the midplane component of the particle s angular momentum .
the radial energy of the orbit is @xmath87 where @xmath88 is the vertical energy component and @xmath89 is the total energy .
as @xmath85 sets the amplitude of the epicycle , we can solve for the position of the particle as function of time with respect to its guiding center radius .
we choose two random phases of the epicycle oscillation , designating the first as @xmath67and the second as @xmath66 , and compute @xmath60 .
the resulting `` baseline '' distribution is shown in red and is a result of observing the two random phases ( akin to our `` initial '' and `` final '' snapshots ) of the initially elliptical orbits .
it does not take into account the potential heating of these initial orbits and the subsequent increase in amplitude of their radial motion ( blurring ) .
the shape of the distribution does not change if we change our random seed or average a larger number of phases to obtain @xmath66 and @xmath67 . for the isolated thick disk ( right panel ) ,
the @xmath60 distribution from full dynamical evolution is only marginally broader than the expected baseline distribution , suggesting that these stars are not heated with time and continue to follow their initial orbits .
however , satellite bombardment broadens the @xmath60 distribution dramatically , nearly doubling both @xmath69 and @xmath71 , necessitating guiding center modification of a substantial fraction of the orbits . for the thin disk , isolated evolution produces a much broader @xmath60distribution than the baseline distribution , demonstrating the impact of the bar and spiral structure in this more unstable disk .
in this case , satellite bombardment only slightly increases the width of the @xmath60 distribution , despite the strong impact on disk structure that is visually evident in figure [ fig : sims ] .
we suspect that the small net difference between these two histograms reflects a cancellation between two competing effects of satellite bombardment .
accretion events heat the stellar disk and thereby suppress the development of spiral structure , thus reducing the level of sb02 migration .
however , the accretion events also induce radial mixing directly via their dynamical perturbations .
the @xmath60 histograms of both perturbed disks , while still approximately symmetric , are noticeably more asymmetric than those of the isolated disks , with @xmath90 ( @xmath91 ) more particles moving outwards than inwards in the thin ( thick ) cases , compared to @xmath92 for the isolated models .
the @xmath60 distribution for all disk particles in the four collisionless simulations .
each histogram show the fraction of particles in non - overlapping @xmath93 bins of @xmath60 .
the isolated ( hatched , gray histogram ) and perturbed ( black line ) @xmath62 ( left panel ) and @xmath63 ( right panel ) disks are shown .
the perturbed disks show greater dispersion in @xmath60 compared to isolated galaxies with the same geometry .
the red histograms denote the expected @xmath60 distribution from epicyclic motion alone , given the initial orbital configurations of the particles in each disk .
@xmath69 and @xmath71 are labeled and color - coded for the three distributions in each panel.,width=326 ] figure [ fig : mass_dr ] presents clear evidence that satellite - induced migration is distinct from that in the isolated experiments , acting in different environments from either epicyclic blurring or spiral - induced churning . in each panel ,
solid , dashed , and dotted lines show the fraction of particles at each birth radius @xmath94 that migrate by more than @xmath95 , @xmath96 , or @xmath97 , respectively .
shaded regions show the scaled radial surface density profile of the initial disk . for both isolated disks ,
the migration probability decreases outwards and approximately traces the surface density profile .
this behavior is similar to that assumed in the chemical evolution models of @xcite , who parametrized the probability that a star migrates as proportional to the mass surrounding it .
however , for both perturbed disks the probability of migration is flat or increasing outwards beyond @xmath94@xmath98 , where the disk potential weakens , and it definitely does not trace the mass distribution .
to better understand this difference between satellite- and spiral - induced radial mixing , we now investigate how changes in angular momentum and energy are correlated with change in radius
. total energy and angular momentum are the classic two - dimensional integrals of motion @xcite .
figure [ fig : lindblad ] shows lindblad diagrams for the two initial and four final states of the collisionless simulations , plotting particle specific angular momentum projected along the axis of symmetry ( @xmath99 ) versus specific binding energy ( @xmath100 ) .
particles are grouped into 50 distinct linear bins of @xmath100 .
we calculate the median @xmath99 in each bin .
red , yellow , and orange regions connect the @xmath101 , @xmath102 , and @xmath103 percentile @xmath99 intervals ( centered on each median @xmath99 ) , respectively , across all energy bins .
the @xmath76 most discrepant particles in @xmath99 for a given @xmath100 are plotted as individual points . using each disk s rotation curve , we plot @xmath99 and @xmath100 for circular orbits in the midplane of the disk ( green line ) . here , @xmath104 , where @xmath105 is the circular velocity at radius @xmath106 and @xmath107 , where @xmath108 is the potential in the disk midplane at radius @xmath106 .
particles on a circular orbit have the maximum @xmath99 allowed given their energy .
the fraction of particles that migrate more than @xmath109 ( solid line ) , @xmath110 ( long dashed line ) , and @xmath97 ( dotted line ) as a function of formation radius .
results are binned such that the migration probability is calculated for particles in non - overlapping @xmath109 wide annuli .
lines connect the migration fraction in each bin ( x - coordinates are bin centers , from @xmath111 to @xmath112 ) .
the gray regions are the radial mass profiles of the initial disks normalized such that the total mass contained in the first annulus equals @xmath113 on this scale .
the migration probability follows the mass distribution in the isolated disks but is anti - correlated with mass in the perturbed disks.,width=326 ] by construction , particles are initially ( left column of figure [ fig : lindblad ] ) on nearly circular orbits with a small radial velocity component .
the red region , falling close to the circular orbit curve in both initial disks , confirms that @xmath114 is small .
particles significantly displaced from the green curve in the initial states either have extreme @xmath115 or are on highly inclined orbits ( @xmath99 is projected along the @xmath116 axis ) . due to this inclination effect
, the initial thick disk has a lower median @xmath99 as a function of @xmath100 than the thin disk despite having the same @xmath115 distribution . in the isolated thin disk ( upper middle panel of figure [ fig : lindblad ] ) , the range of @xmath99 grows relative to the initial values at every energy @xmath100 .
the change is largest at lower energies .
recall from figure [ fig : sims ] that the isolated thin disk develops a bar , which is composed of particles on radial orbits with relatively low @xmath99 , in the most bound region of the disk .
thus , this relatively large change towards less circular orbits at low energies can be associated with bar formation .
in contrast , the distribution of @xmath99 in the isolated thick disk is basically equivalent to its initial state . in the isolated thick disk
there is no bar , relatively little spiral structure develops , and we find the least radial migration among the four simulations . ignoring the region of the lindblad diagram influenced by members of the bar ,
the isolated disk diagrams suggest that the extensive radial migration seen in the thin disk and associated with guiding center modification decreases the angular momentum of particles .
the two perturbed disks ( right column of figure [ fig : lindblad ] ) show larger changes in their lindblad diagrams .
bar formation in both simulations can explain the significantly lower median @xmath99 at lower energies .
at higher energies , corresponding to less bound particles further out in the disk , both disks show a much larger dispersion in @xmath99 than their isolated counterparts .
additionally , there is substructure in the lindblad diagrams ( groups of relatively low @xmath99 in a narrow range of energy ) not seen in the isolated disks .
satellite bombardment in the perturbed disk simulations has a qualitatively discernible impact on the angular momentum distribution of the disk . to quantify changes in @xmath99
, we introduce the circularity ( @xmath117 ) quantity ( e.g. , * ? ? ?
* ) . for a particle of energy @xmath118 and angular momentum @xmath119 in the @xmath116-direction , we define circularity as @xmath120 , the ratio of @xmath119 to the specific angular momentum the particle would have if it were on a circular orbit with energy @xmath118 ( obtained using the circular orbit curve ) .
circular orbits have @xmath121 and radial orbits have @xmath122 .
negative circularities correspond to retrograde orbits .
we note that this method is formally different from that of @xcite .
the circular orbit curve may not represent the highest @xmath99 for a given @xmath100 due to shot noise in the potential or gravitational potential asymmetries . strictly defining the maximum @xmath99 in each energy bin ( as in @xcite ) results in the same qualitative trends discussed later .
our method benefits from the use of a mathematically constructed and reproducible rotation curve and systematically decreases @xmath123@xmath124 at the @xmath125 level .
the inset of each panel in figure [ fig : lindblad ] shows the circularity distribution of each simulation .
figure [ fig : ecc_circ ] plots the relation between circularity and eccentricity for orbits near the solar radius ( @xmath126 ) in the @xmath63 disk .
circularity is related to eccentricity via the particle s radial and vertical energies as well as its orbital inclination ( since @xmath99 is a projected quantity ) .
orbits in the midplane of the disk have the highest circularity for a given eccentricity .
shaded regions in figure [ fig : ecc_circ ] show the @xmath102 and @xmath103 percentile circularity as a function of eccentricity at the solar radius . for reference in interpreting figure [ fig : lindblad ] and subsequent figures ,
it is worth noting that changes of @xmath127 in @xmath117 typically correspond to quite noticeable changes in eccentricity .
the boundaries of these regions are not smooth due to binning and small number statistics for initially high eccentricity particles
. possible values of circularity as a function of eccentricity in the solar annulus ( @xmath128 ) of the @xmath63 disk .
the top edge of the dark region represents orbits in the midplane of the disk .
the dark region encompasses the @xmath102 percentile of particle circularity in the solar annulus of each simulation ; the bottom edge of the light region corresponds to the @xmath103 percentile circularity .
, width=326 ] returning to figure [ fig : lindblad ] , we see that the two isolated disks show markedly different evolution of their circularity distributions . in the isolated thin disk ,
the fraction of stars with @xmath117 @xmath129 ( rightmost bin ) drops from @xmath130 to @xmath131 , and the median @xmath117 drops from @xmath132 to @xmath133 .
particles in the bar predominantly populate the newly formed low circularity tail . in the thick isolated disk , on the other hand ,
the @xmath117 distribution is nearly identical to the initial disk s , with median @xmath117 dropping only @xmath134 .
this lack of evolution in the circularity distribution is fully consistent with cr scattering . as described by sb02
, individual stars may exchange angular momentum across crs ( churning ) while the overall distribution would remain unchanged .
however , our results from section [ sec : radmix ] show that the delta r distribution of this model , which reflects individual particles and their orbits , is nearly identical to that expected from observing the particles initially elliptical orbits . the circularity distribution of the isolated , thick disk combined with its delta r distribution imply that , on average , individual particle guiding centers are not significantly modified .
the circularity distributions of the perturbed disks are demonstrably altered from their isolated counterparts but are similar to one another .
the most common circularities are now in the range @xmath135 in both disks , with a decrease for @xmath117@xmath136 .
the thin disk starts with more @xmath117 @xmath137 particles because of its smaller orbital inclinations , and its circularity distribution evolves more strongly , with median @xmath117 dropping from @xmath132 to @xmath138 vs. @xmath139 to @xmath140 for the perturbed thick disk .
notably , the perturbed thin and thick disks evolve to similar @xmath60(figure [ fig : perdisks ] ) and @xmath117 distributions despite starting with different scale heights .
figure [ fig : lindzoom ] tracks the changes of selected individual particles in the ( @xmath100 , @xmath99 ) space of the lindblad diagram .
while @xmath100 and @xmath99 are not individually conserved in the presence of a non - axisymmetric perturbation , the jacobi invariant @xmath141 is , where @xmath142 is the pattern speed of the perturbation , assumed to be static and small ( sb02 ; @xcite ) . if @xmath143 , then @xmath144 .
the sb02 mechanism operates at the corotation resonance of the star / particle and the spiral wave , requiring that @xmath145 .
thus , in the galaxy , particles should move parallel to the line that is tangent to the circular orbit curve at their binding energy prior to scattering . in other words
, the sb02 mechanism requires that changes in energy be accompanied by changes in angular momentum that preserve the orbital shape , modulo differences in the slope of the circular orbit curve over the range @xmath146 $ ] of a given particle .
the initial and final @xmath100 , @xmath99 pairs for ten randomly selected particles in each simulation with initial energy @xmath147 and angular momentum @xmath148 ( large , filled square ) . for clarity
we require that plotted particles lose at least @xmath149 in energy during the simulation .
the final @xmath100,@xmath99 pairs of the ten particles are plotted as open squares .
lines connect the initial and final @xmath150 points of each particle .
the @xmath150 curve populated by circular orbits in the initial state of each simulation is indicated by the thick black curve .
the dashed line is tangent to the circular orbit curve at the initial energy of all ten particles.,width=326 ] figure [ fig : lindzoom ] zooms in on the area of the lindblad diagram designated by the two small boxes drawn on the initial state diagrams in figure [ fig : lindblad ] .
we randomly select ten particles within the same initial @xmath100 and @xmath99 range ( filled , dark squares in figure [ fig : lindzoom ] ) and plot their final @xmath100 and @xmath99 .
we require @xmath151 to ensure that the individual tracks are visible . in the isolated thin disk (
top left ) , most particles move parallel to the circular orbit curve tangent ( dashed line ) ; this relationship between @xmath152 and @xmath153 is consistent with corotational resonant scattering ( equation 2 , sb02 ) .
when particles lose energy in the isolated thick disk ( bottom left ) , their @xmath99 typically remains closer to the circular orbit curve than in the isolated thin disk .
the changes in energy and angular momentum are relatively small , indicating that the guiding centers are modified to a modest extent @xcite .
note that these particles are in the top @xmath76 of @xmath154 ; most particles in this particular experiment have @xmath155 .
the randomly selected particles in both perturbed disks have slopes @xmath156 that are steeper than the slope of the tangent to the circular orbit curve at corotation .
this distinct coupling of @xmath153 and @xmath152 , combined with our results concerning the migration probability as a function of @xmath94 ( section [ sec : radmix ] ) , offer compelling evidence that migration in the perturbed disks is driven , at least in part , by a mechanism that does not operate in the isolated systems .
figure [ fig : lindzoom ] shows that the orbital characteristics of many particles in the perturbed disks are modified in a fashion inconsistent with either epicyclic motion or a single spiral wave scattering event .
median change in circularity @xmath123 as a function of @xmath60 for the entire disk ( open squares ) and for particles with @xmath157 ( filled squares ) .
particles are sorted into non - overlapping @xmath158 bins of @xmath60 .
error bars mark the @xmath159 and @xmath160 percentile @xmath123 in each bin .
histograms at the bottom of each panel represent the relative fraction of particles in each bin for the case @xmath157 .
, width=326 ] we now examine the correlation between migration and changes in orbital properties . since the metallicity of star - forming gas increases with time and decreases with radius
, any such correlations also imply observable correlations between the present orbital parameters and metallicities of stars as a function of age and galacto - centric radius .
figure [ fig : dr_dcirc ] shows the median change in circularity , @xmath161 , as a function of radial migration distance @xmath60 in the four collisionless simulations ( open squares ) . in the isolated thick disk ; changes in circularity are tiny ( median @xmath162 at any @xmath60 ; figure [ fig : lindzoom ] shows that particles in this simulation stay close to the circular velocity curve even if they change energy . in the other three simulations ,
particles with negative @xmath60 show substantial drops in circularity ( typical median @xmath163 ) , which are almost certainly associated with bar formation .
the bars that form in these three simulations increase the orbital eccentricities of their members . to focus on behavior in the disk proper ,
the filled squares in each panel show the median @xmath123vs .
@xmath60 for those particles that start and end the simulation at @xmath164 , beyond the extent of the bar .
error bars mark the inter - quartile range ( @xmath165 to @xmath166 percentile ) at each @xmath60 . in the isolated thin disk , particles that migrate inwards ( @xmath60@xmath167 )
experience a modest decrease in circularity , stronger for more negative @xmath60 ( and @xmath153 ) , consistent with the tracks shown in figure [ fig : lindzoom ] .
particles that migrate outwards ( @xmath60@xmath168 ) increase their energy and typically traverse areas of the lindblad diagram with relatively little change in the slope of the circular orbit curve . following the tangent to the circular orbit curve
, particles will not significantly change their circularity in such a scenario ( note the relative lack of low circularity particles at @xmath169 in figure [ fig : lindblad ] ) .
the perturbed disks show a decrease in median circularity at every @xmath60 , and a much wider inter - quartile range indicating a greater range of orbital inclinations and eccentricities .
circularity drops more strongly for particles that have experienced strong radial migration , either inward or outward .
the minimum in these curves is slightly offset to positive @xmath60 because dynamical heating slightly puffs up the disk radially , decreasing the potential at a given radius , thereby increasing its total energy , moving particles to the right on the lindblad diagram , and lowering the circularity for particles with @xmath60@xmath170 .
figure [ fig : dr_dcirc ] indicates that stars with anomalous chemistry for their age and current position should have preferentially more eccentric orbits .
the trend is smaller than the inter - quartile range , but similar in magnitude .
like figure [ fig : dr_dcirc ] , but the change in maximum vertical displacement , @xmath171 , is plotted in place of the change in circularity .
open squares show all disk particles , filled squares show those with @xmath66 , @xmath67 @xmath172 , and error bars mark @xmath159 and @xmath160 percentile at a given @xmath60.,width=326 ] @xcite and @xcite discuss the possible role of radial migration in producing thick disks like the ones observed in the milky way @xcite and other edge - on galaxies @xcite .
figure [ fig : dr_dzmax ] is similar to figure [ fig : dr_dcirc ] , but instead of @xmath123 it plots the change in vertical energy , quantified by the maximum distance @xmath173that a particle can reach from the disk plane .
we compute @xmath173approximately from each particle s vertical velocity component assuming that the final potential of each simulation is static , namely @xmath173@xmath174 where @xmath116 is the vertical position of the particle at the final output , @xmath175 is the velocity along the @xmath116 axis , @xmath176 is the gravitational constant , and @xmath177 is the surface density of the disk at radius @xmath106 assuming all the mass of the disk is in the midplane .
the change in @xmath173 is @xmath171@xmath178 where the subscripts @xmath179 and @xmath180 refer to the initial and final simulation snapshots , respectively .
the two isolated simulations show a shallow linear trend between @xmath171and @xmath60 ( open squares for all particles ; filled squares for those with @xmath67,@xmath66@xmath181 ) .
when particles move outwards ( inwards ) through the disk , they experience a weaker ( stronger ) gravitational potential , thus increasing ( decreasing ) @xmath173 .
however , changes in @xmath173are small , less than @xmath182 pc even when we consider the quartile range at the extremes of @xmath60 . the perturbed disks , by contrast , show larger changes in the median @xmath173 and a dramatic increase in the inter - quartile range at all @xmath60 .
k08 show that the perturbed @xmath183 pc disk develops a two - component vertical structure in quantitative agreement with the observed thin / thick disk structure of the mw .
the perturbed @xmath11 pc disk increases its scale height ( to @xmath184 at the solar annulus ) , but it can still be described by a single component model .
this vertical heating by satellite perturbations is evident in figure [ fig : dr_dzmax ] .
particles that have large positive @xmath60 have the largest increase in @xmath173 , which is plausibly a consequence of moving outwards to regions of lower disk surface density and thus weaker vertical restoring force . for @xmath60 @xmath167 ,
the trend of median @xmath171 with @xmath60 is approximately flat , suggestion a cancellation between the effects of increased @xmath177 at smaller @xmath106 and direct satellite - induced heating of those particles with the largest excursions .
figure [ fig : dr_dzmax ] implies that stars with high metallicity for their age and present location should have preferentially larger @xmath173 , though the scatter is larger than the trend .
the distribution of formation radius ( @xmath94 , in ) for all particles that are in the solar annulus ( @xmath185 ) at the end of each simulation .
histograms report the fraction of solar annulus particles emigrating from non - overlapping @xmath93 annuli in @xmath94 .
particles within the dashed lines remained in the solar annulus throughout the simulation .
both thin disks and the perturbed thick disk show a broad range of @xmath94 at the solar annulus .
the gray histogram in each panel is the @xmath94 distribution for those particles that are in the solar annulus and at large heights above the plane ( @xmath186 ) in the final simulation output.,width=326 ] the solar neighborhood is easier to study than other regions of the galaxy , since high - precision spectroscopy is easier for brighter stars and parallax and proper motion measurements are more accurate at smaller distances .
some of the most detailed chemo - dynamic surveys , such as the gcs @xcite and the radial velocity experiment ( rave , * ? ? ?
* ) concentrate on the solar neighborhood . in this section ,
we repeat some of our earlier analysis specifically for stars that reside in the solar annulus ( @xmath187 ) at the end of each simulation .
this focus on the solar annulus also removes much of the impact of the bars that dominate evolution of the inner disk ( @xmath188 ) in three of our simulations , though some particles from the bar region can migrate as far as the solar radius , and resonances between the bar and spiral structure may increase migration frequency @xcite . figure [ fig : rform ] shows the radius of formation ( @xmath94 ) distribution of particles residing in the solar annulus , marked by the vertical dashed lines .
only the isolated thick disk simulation predicts a final solar annulus dominated by stars born in the solar annulus , with tails extending @xmath189@xmath190 on either side .
the broad @xmath60 distributions of the other three simulations show that their stars migrate to the solar annulus from a wide range of formation radii . the global @xmath60 ( figure [ fig : perdisks ] ) and solar annulus @xmath94 distributions of the isolated and perturbed thin disks are remarkably similar , despite the differences in migration mechanisms discussed in section [ sec : orbits ] . in the isolated thin disk , @xmath191 of solar annulus particles originated at @xmath94@xmath192 and @xmath193 at @xmath94@xmath194 .
corresponding numbers for the perturbed thin disk are @xmath195 and @xmath91 .
the @xmath94 distribution of the perturbed thick disk is slightly narrower , but it still broad with respect to the isolated thick disk . the change in circularity @xmath123 as a function of @xmath60 for particles ending the four collisionless simulations in the solar annulus ( @xmath196 ) .
particles are sorted into non - overlapping @xmath158 bins of @xmath60 .
we plot the median ( squares ) and the @xmath159 and @xmath160 percentile ( error bars ) @xmath123 in each bin .
histograms at the bottom of each panel represent the relative fraction of particles in each bin.,width=326 ] figure [ fig : dcirc_dr_soln ] shows the correlations between @xmath123and @xmath60 for particles that end in the solar annulus . consistent with results for the full disk ( figure [ fig : dr_dcirc ] ) , the solar annulus particles in the isolated thick disk show no significant change in median @xmath117 regardless of @xmath60 . in the isolated thin disk ,
the range of @xmath123 is much larger , with a modest decrease in median @xmath117 .
the median @xmath123 drops at large positive @xmath60 because particles move to less inclined orbits as they migrate outwards and disk heating has slightly modified the galaxy s circular velocity curve , allowing outward moving particles to potentially increase circularity . in the perturbed thick disk , there is a strong and nearly linear trend between @xmath123 and @xmath60 .
particles that migrated to the solar annulus from the outer disk have experienced substantial drops in circularity , while particles migrating from the inner disk show only modest decreases .
the range of @xmath123 is large at all @xmath60 .
results for the perturbed thin disk are intermediate between those of the isolated thin and perturbed thick disks : quasi - linear trends at large @xmath68 but a flat plateau at intermediate @xmath68 .
this behavior is plausibly a consequence of two different mechanisms contributing to migration , with spiral - induced mixing dominating at intermediate @xmath60 and satellite - induced mixing dominating at the extremes .
figure [ fig : dzmax_dr_soln ] shows the solar annulus correlations of @xmath171 with @xmath60 , analogous to figure [ fig : dr_dcirc ] for the full disk . for the two isolated disks there is a clear linear trend of median @xmath171 with @xmath60 as expected from the arguments in section [ sec : orbits ] : particles that migrate outward move to a region of lower disk surface density ,
so if their vertical velocities are not systematically changed by the radial migration they will attain higher @xmath173 .
both perturbed disks show signs of the strong vertical heating induced by satellite accretion .
the median @xmath171 is @xmath197 almost independent of @xmath60 , and the range of @xmath171 is much larger than in the isolated disks . since inwardly migrating particles experience a higher vertical potential at @xmath66 than @xmath67 , they must experience more vertical heating than outwardly migrating particles to keep the @xmath171 trend flat . like figure [ fig : dcirc_dr_soln ] , but the change in vertical displacement @xmath171 is plotted in place of the change in circularity.,width=326 ] returning to figure [ fig : rform ] , gray histograms represent the @xmath94 distributions of particles that end the simulations in the solar annulus at high @xmath116 , @xmath186 . in the isolated thick disk and both perturbed disks ,
the @xmath94 distribution of high @xmath116 particles resembles that of all solar annulus particles . in these three models , therefore , selecting high- @xmath116 particles does not isolate a population with atypical radial migration . in the isolated thin disk , on the other hand ,
the @xmath94 distribution of high-@xmath116 particles is strongly skewed towards low formation radii , with a peak at @xmath94@xmath198 . here , only initially hot particles that migrate significantly outwards via churning or interactions with the bar and
experience a weaker potential have enough vertical energy to overcome the local restoring force and reach high @xmath116 .
only @xmath199 of the solar annulus is at high @xmath116 in the isolated thin disk ( truly the tail of the initial vertical velocity dispersion ) ; this number rises to @xmath200 and @xmath201 in the thin and thick perturbed disks , respectively .
the radial migration mechanisms in the perturbed disks ensure that even the high @xmath116 population of the solar annulus comes from a broad range of @xmath94 .
measurements of the age - metallicity relation for high-@xmath116 stars could be a valuable diagnostic for distinguishing models of radial migration and vertical heating .
observations suggest that radial migration plays an important role in the chemical evolution of the mw disk @xcite .
sb02 described a mechanism by which stars at corotational resonance with spiral waves can scatter , producing large changes in guiding center radius while keeping stars on nearly circular orbits .
previous simulations have shown that migration over several kiloparsecs can occur in isolated disks grown by smooth accretion @xcite and that encounters with satellites can also induce migration @xcite . here
we have carried out a systematic investigation of a variety of simulations to characterize the role of stellar disk properties , gas fractions , and satellite perturbations in producing radial migration .
most importantly , our suite of simulations includes experiments with a level of satellite bombardment expected in @xmath202cdm models of galaxy formation , which earlier investigations ( k08 , * ? ? ?
* ) have shown to produce vertical and in - plane structure resembling that seen in the mw .
for disks evolved in isolation , the degree of migration correlates with the degree of disk self - gravity , susceptibility to bar formation , and spiral structure . the collisionless stellar disk with @xmath63 , chosen to match that of the thin disk in the mw , has no bar and minimal spiral structure .
it exhibits limited radial migration ; the median value of radial change @xmath203 is @xmath69@xmath204 , consistent with the level expected from epicyclic motion ( figure [ fig : perdisks ] ) .
the collisionless @xmath62 disk is much more unstable , develops a bar and spiral structure ( figure [ fig : sims ] ) , and has a higher degree of radial migration , with @xmath69@xmath205 and @xmath71@xmath206 .
the presence of gas is a catalyst for radial migration in the @xmath63 case .
both @xmath69 and @xmath71 increase in models with larger gas fractions or lower star formation efficiency ( which consumes gas more slowly ) .
the galactic bar strongly influences , and perhaps dominates , the resultant individual particle dynamics in the inner galaxy .
substantial radial migration , consistent with the sb02 mechanism or being satellite - induced , occurs in the outer portion of the disk in all of our collisionless experiments except for the isolated @xmath63 disk . in the @xmath62 disk , stars that finish the simulation in the solar annulus ( @xmath207 ) have @xmath69@xmath208 , @xmath71@xmath209 . adding satellite perturbations dramatically increases radial migration in the @xmath63 disks , raising @xmath69 from @xmath210 to @xmath211 globally and from @xmath212 to @xmath213 in the solar annulus .
perturbations do not change the @xmath60 distribution so drastically in the @xmath62 disks , but there are other indications that the nature of radial migration is different . in the isolated galaxies , migrating particles in the outer disk follow tracks in @xmath214 space that parallel the tangent to the local circular velocity curve , as expected for the sb02 mechanism ( see section [ sec : orbits ] ) .
however , inwardly migrating particles in the perturbed disks lose more angular momentum for a given change in energy , causing migration associated with heating and thus different from the sb02 mechanism in this respect .
the relationship between @xmath153 and @xmath152 found in the perturbed experiments is broadly more consistent with scattering at lr ( see sb02 ) , but a detailed decomposition of the modes in the disk and satellites ( beyond the scope of this paper ) is necessary to confirm lr scattering on a particle by particle basis .
the radial distribution of migrating particles presents an even clearer distinction , one with important implications for chemical evolution models ( figure [ fig : mass_dr ] ) . in the isolated disks ,
the probability that a particle with a formation radius @xmath94 undergoes significant migration ( @xmath215 ) is approximately proportional to the disk surface density at @xmath94 ; i.e. , migration follows mass . in the perturbed disks ,
the probability of migration is flat or increasing with radius , so a much larger fraction of migration comes from the tenuous outer disk .
particles in the outer disk are overall more likely to migrate relative to the inner disk as 1 ) they feel a weaker potential and are thus more susceptible to direct heating by satellites and 2 ) their circular frequencies make them more likely to resonantly interact with satellites and migrate .
thus , satellite - induced migration patterns are distinct from those produced by purely secular evolution .
satellite bombardment heats the disk both vertically and radially .
compared to the isolated disks , the perturbed disks experience a greater drop in median circularity ( hence growth of median eccentricity ) during their evolution , and individual particles experience a wider range of circularity changes .
the circularity change is moderately correlated with radial migration distance and direction , though the trend is smaller than the inter - quartile range at a given @xmath60 ( figure [ fig : dr_dcirc ] ) . if we restrict our analysis to the solar annulus , the correlation between @xmath60 and @xmath123 is more significant . in particular , particles in the solar annulus of the @xmath63 disk that migrate @xmath23@xmath216 inwards experience substantial drops in circularity , while those that migrate the same distance outwards nearly maintain their initial circularity ( figure [ fig : dcirc_dr_soln ] ) .
the large inward excursions from the outer disk , driven by satellite perturbations , systematically remove angular momentum from particle orbits . as shown by k08 , vertical heating by satellite bombardment
produces ( in the case of the @xmath63 disk ) a two - component vertical structure in quantitative agreement with the thin and thick disk profiles observed in the mw . in the isolated disks , we find the expected trend that as particles migrate outwards , they experience a weaker vertical potential and increase their vertical energy ( characterized by @xmath173 , the maximum distance from the plane that a particle can reach given its current location and velocity )
however , in our simulations , this effect is not sufficient to produce a second , `` thick - disk '' component in the isolated systems even if they have substantial radial migration .
the satellite perturbations increase the median and range of @xmath173 considerably at all @xmath60 .
the resulting overall trend of @xmath171with @xmath60 is fairly flat ( especially at a fixed @xmath66 , such as the solar annulus , see figures [ fig : dr_dzmax ] and [ fig : dzmax_dr_soln ] ) , suggesting that systems with low vertical velocity dispersion and subjected to weaker potentials ( as in the outer disk ) are more susceptible to vertical heating ( similar to the radial heating trends seen in section [ sec : radmix ] ) .
particles with @xmath94 interior _ or _ exterior to a given @xmath66 can be found at significant distances above the plane in the perturbed disks , while increases in @xmath173 are smaller in the isolated systems and rely on particles moving outwards ( figure [ fig : rform ] ) .
our results confirm earlier findings that spiral structure development in isolated disks @xcite and perturbations by satellites @xcite can produce significant radial mixing of stellar populations while retaining reasonable orbital structure for disk stars .
they strongly support the view ( @xcite ; sb02 ; @xcite ) that radial mixing is an essential ingredient in understanding the chemical evolution of the mw and disk galaxies in general .
while a combination of metallicity and age can be used to estimate a star s formation radius given a chemical evolution model , the uncertainties in this approach ( including the difficulty of estimating ages for typical stars in a spectroscopic survey ) will make it difficult to reconstruct formation radius distributions for observed stars , even in the solar neighborhood .
however , our analysis provides theoretical guidance for chemical evolution models that incorporate radial mixing ( e.g. * ? ? ?
* ) and suggestions for the correlations one might search for between chemical abundances and orbital properties ( though the predicted trends are fairly weak ) .
we regard our perturbed @xmath63 disk as the most relevant simulation , as it includes the satellite accretion events expected in and produces a final vertical structure similar to that measured in the mw . in this simulation @xmath217 of particles in the @xmath218 solar annulus
were `` born '' in that annulus , @xmath219 migrated there from @xmath220 , and @xmath221 migrated there from @xmath222 . radial mixing and
orbital dynamics changes are sensitive to several different aspects of disk modeling , as shown by the variety of our results .
robust predictions should therefore come from calculations that self - consistently include star formation , chemical enrichment , gas accretion , and accretion events , and the interplay among these elements .
we will move in this direction with our future simulations .
giant spectroscopic surveys such as segue , rave , apogee , lamost , and hermes offer an extraordinary opportunity to unravel the formation history of the mw , and they offer an exciting new challenge to theoretical models of galaxy formation .
we thank the anonymous referee for a careful and constructive report .
we acknowledge useful discussions with mario abadi , julianne dalcanton , rok rokar , ralph schnrich , scott gaudi , chris kochanek , victor debattista , simone callegari , tom quinn , and james wadsley .
is funded by the center for cosmology and astro - particle physics ( ccapp ) at the ohio state university .
this work was supported by nsf grants ast @xmath223 and ast @xmath224 , and by an allocation of computing time from the ohio supercomputer center ( http://www.osc.edu ) .
this research made use of the nasa astrophysics data system .
d.w . acknowledges the hospitality of the institute for advanced study and the support of an amias membership during part of this work . | we use a suite of numerical simulations to investigate the mechanisms and effects of radial migration of stars in disk galaxies like the milky way ( mw ) . an isolated , collisionless stellar disk with a mw - like
scale - height shows only the radial `` blurring '' expected from epicyclic orbits . reducing the disk thickness or adding gas to the disk substantially increases the level of radial migration , induced by interaction with transient spiral arms and/or a central bar .
we also examine collisionless disks subjected to gravitational perturbations from a cosmologically motivated satellite accretion history . in the perturbed disk that best reproduces the observed properties of the mw , 20% of stars that end up in the solar annulus @xmath0 started at @xmath1 , and 7% started at @xmath2 .
this level of migration would add considerable dispersion to the age - metallicity relation of solar neighborhood stars . in the isolated disk models ,
the probability of migration traces the disk s radial mass profile , but in perturbed disks migration occurs preferentially at large radii , where the disk is more weakly bound .
the orbital dynamics of migrating particles are also different in isolated and perturbed disks : satellite perturbations drive particles to lower angular momentum for a given change in radius .
thus , satellite perturbations appear to be a distinct mechanism for inducing radial migration , which can operate in concert with migration induced by bars and spiral structure .
we investigate correlations between changes in radius and changes in orbital circularity or vertical energy , identifying signatures that might be used to test models and distinguish radial migration mechanisms in chemo - dynamical surveys of the mw disk .
[ firstpage ] methods : numerical ; galaxy : kinematics and dynamics ; galaxy : disc ; galaxies : formation ; galaxies : evolution |
from a consideration of its condensed matter physics properties , crystal structure , and metallurgy , plutonium is probably the most complicated element in the periodic table , @xcite including a phase diagram with six allotropic phases .
the low - temperature monoclinic @xmath4-phase is stable up to 395 k while the face - center - cubic ( fcc ) @xmath0-phase is stable between 592 and 724 k. furthermore , in stabilized alloys the @xmath4 to @xmath0 phase transformation of pu has a significant volume expansion , with the @xmath0-phase 25% larger in volume than the @xmath4-phase .
this behavior is related to the special position of pu in the periodic table , which is at the boundary between the light actinides that have itinerant 5@xmath5 electrons and the heavy actinides with localized 5@xmath5 electrons .
in this situation the electrons are in a very strongly correlated state where even the best conventional lda band - structure calculations can not predict its unique properties , e.g. , the @xmath0-phase volume @xcite in the experimentally observed nonmagnetic state .
this failure has stimulated numerous attempts to include additonal electronic correlation effects .
although lda theory has repeatedly claimed @xcite that the thermal expansion in pu is a consequence of magnetism , this is in striking contradiction with experimental data .
@xcite several research groups @xcite have applied the lda+@xmath2 method to include more @xmath5-@xmath5 correlation energy . by adjusting the on - site coulomb repulsion energy ( @xmath2 ) appropriately ,
it has been possible to fit to the experimental @xmath0-phase volume .
the lda+@xmath2 calculations also indicated an instability of @xmath0-pu toward an antiferromagnetic ground state .
more recent calculations based on the `` around - mean - field '' lsda+@xmath2 including the spin - orbit interaction @xcite or the fully - localized - limit lsda+@xmath2 quenched the spin polarization through an @xmath6 configuration , and obtained the correct @xmath0-phase volume due to a weak exchange interaction .
@xcite to predict the correct volume , a mixed - level model has also been proposed by one of us together with co - workers .
@xcite in this model , four of the five 5_f _ electrons in the @xmath0-phase were constrained to be localized and were not allowed to hop from site to site , but could hybridize with the conduction electrons .
a major advance came when a new calculation scheme was proposed to merge lda - based methods with dmft .
@xcite dmft @xcite is a many - body technique that is able to treat the band- and atomic - like aspect simultaneously when applied to plutonium . within the lda+dmft approach , @xcite the origin of substantial volume expansion
was explained in terms of the competition between coulomb repulsion and kinetic energy .
thus , two dramatically different pictures for the non - magnetism of @xmath0-pu have emerged . in the `` 5@xmath7 '' picture ,
@xcite by including the spin - orbit interaction , one starts with a closed @xmath8 atomic subshell fully filled with six electrons while the @xmath9 subshell is empty , making pu magnetically inert . in the `` 5@xmath10 '' picture ,
@xcite one starts with an open @xmath8 atomic subshell filled with five electrons , resulting in a magnetic moment that should be screened by the @xmath11 valence electrons
. therefore , more studies are necessary to determine which picture will prevail . in this paper , we report a systematic lda+dmft study of the spectral properties of @xmath0-pu with varying @xmath1 occupancy . throughout this work
, the lda part of hamiltonian is determined from a new tb fit to fp - lapw calculations .
the dmft equations are solved using quantum monte carlo ( qmc ) simulations as well as the more approximate hubbard - i method .
these provide an accurate characterization of spectral properties of @xmath0-pu .
it is found that the @xmath1 spectral density of pu is very sensitive to its occupancy .
good agreement is found with photoemission spectroscopy ( pes ) measurements when about five electrons occupy the @xmath8 subshell , in support of the second picture .
other recent work appears also to be in support of this picture .
measurements of the branching ratio in @xmath12 to @xmath1 transitions favor a @xmath1 count closer to @xmath13 than @xmath14 , @xcite while subsequent lda+dmft calculations @xcite using a vertex - corrected one - crossing approximation for the auxilliary impurity problem yield results for the branching ratio in agreement with experiment @xcite and suggest a @xmath1 occupancy of @xmath15 .
similarly , specific heat calculations carried out using an lda+dmft method with a perturbative @xmath16-matrix and fluctuating exchange approach to the auxiliary impurity problem also favor the @xmath17 configuration .
@xcite the remainder of this paper is organized as follows : theoretical methods are given in sec .
[ sec : model ] ; numerical results for the @xmath0-pu are presented in sec . [
sec : spectrum ] ; and then a summary follows in sec . [
sec : summary ]
most lda+dmft methods to date have been implemented in a basis of linear muffin - tin orbitals ( lmto ) within the atomic sphere approximation ( asa ) , or in its full potential version .
here we use an alternative version of the lda+dmft method , which is based on a recently developed tb theory .
@xcite in this representation , the full second - quantized hamiltonian is written as : @xmath18_{ljm_{j},l^{\prime}j^{\prime}m_{j}^{\prime}}\hat{c}^{\dagger}_{\mathbf{k}ljm_{j}}\hat{c}_{\mathbf{k}l^{\prime}j^{\prime}m_{j}^{\prime } } \nonumber \\ & & + \frac{u_{f } } { 2}\sum_{i , jm_{j}\neq j^{\prime}m_{j}^{\prime}}\hat{n}_{ifjm_{j}}\hat{n}_{ifj^{\prime}m_{j}^{\prime}}\;. \label{eq : hamil}\end{aligned}\ ] ] here @xmath19 are the brillouin - zone wavevectors ; @xmath20 are lattice site indices ; @xmath21 is the orbital angular momentum ; @xmath22 is the total angular momentum ; @xmath23 ; @xmath24 .
the relevant orbitals for pu are @xmath25 , @xmath26 , @xmath27 , and @xmath1 , and so the matrices @xmath28 are
they are given by @xmath30 here the matrix @xmath31 is zero except for @xmath32 s along the @xmath33 @xmath5-@xmath5 diagonals , and @xmath34 \ , , \label{eq : eflda}\ ] ] where @xmath35 is the number of @xmath19 points in the brillouin zone .
the matrices @xmath36 are orthogonalized , single - electron hamiltonian matrices obtained from tb fits to the fp - lapw calculations .
@xcite in this study , strict orthogonality is maintained between the tb orbitals ( hence no overlap matrix need be included ) .
all matrices are calculated at the experimental @xmath0-pu volume .
the slater - koster tables for the @xmath37 matrix elements can be found in standard references , @xcite and we have used an extended formalism for a unified treatment including additional matrix elements involving @xmath5-electrons .
@xcite typical tb calculations are then reduced to using tb as an interpolation scheme ; the matrix elements are determined by fitting to _ ab - initio _ calculated quantities such as the total energy and band energies . in this study
tb inter - site parameter values are evaluated at inter - atomic distances out to the fifth nearest neighbor , resulting in 100 inter - site parameters and 4 on - site parameters .
the effect of the spin - orbit interaction is included as a perturbation , with its parameters kept fixed across the bands , although evaluated at the most important energies , namely at the respective centers of gravity of the occupied state density for each orbital type .
@xcite the spin - orbit coupling adds 3 more parameters to the tb fit , resulting in a total of 107 parameters ( parameter values are available on request from the authors ) .
the quality of the present tb fit for @xmath0-pu is comparable to that shown elsewhere for fcc u. @xcite as customary in lda+dmft calculations , eqs .
( [ eq : hamil]-[eq : eflda ] ) presume that the lda results are sufficient for both the off - diagonal elements of @xmath28 ( hybridization ) and the diagonal elements of the assumed uncorrelated @xmath38 electrons . for the correlated @xmath5 electrons one
needs of course the coulomb interaction as seen in eq .
( [ eq : hamil ] ) , but in addition , since the lda site energy defined by eq .
( [ eq : eflda ] ) includes coulomb contributions , it must be replaced by the true or bare site energy @xmath39 as in eq .
( [ eq : h0 ] ) . both @xmath39 and @xmath40
may be estimated from the dependence of the lda site energy @xmath41 on @xmath5 occupation @xmath42 via lda constrained occupation calculations,@xcite and comparison to spectroscopic data for ions suggests uncertainties of @xmath43 ev in such results for both quantities.@xcite typical values of @xmath40 are around 4 ev for @xmath0-pu , as we also find , and this value is assumed throughout the present paper .
given the sensitivity of @xmath42 to @xmath39 , on the other hand , we have chosen to adjust the parameter @xmath39 in the present work as a natural and convenient means of exploring the sensitivity of the @xmath0-pu spectra to @xmath1 occupancy .
@xcite this adjustment amounts to varying the double counting term in conventional lda+@xmath2 or lda+dmft terminology . within the dmft , the lattice problem eq .
( [ eq : hamil ] ) is mapped onto a multi - orbital quantum single impurity problem subject to a self - consistency condition : @xmath44 here @xmath45 is the weiss function , @xmath46 is a @xmath19-independent self - energy , and the local green s function is defined as @xmath47 , where the lattice green s function reads @xmath48^{-1}\;,\ ] ] with @xmath49 the @xmath50 unit matrix and @xmath51 the chemical potential . since we have restricted strong correlation to the @xmath5 orbitals only , it is customary to take @xmath52 to have nonzero elements only within the @xmath53 @xmath5-@xmath5 block , which permits a similar reduction in the quantum single - impurity problem itself .
furthermore , we neglect the off - diagonal elements in the self - energy and , by ignoring the crystal - field effect in the impurity problem , treat the @xmath54 levels as one 6-fold degenerate level and the @xmath55 as another 8-fold degenerate level .
the only nonzero @xmath5-@xmath5 block self - energy reduces to @xmath56 and we need only find functions for the two spin - orbit states @xmath57 and @xmath58 . note that when no spin - orbit coupling is included , @xmath59 .
correspondingly , the local green s functions is dictated by : @xmath60 it should also be noted that we neglect the crystal - field effects only at the auxiliary impurity problem , all such effects are still retained in @xmath28 , where hybridization dominates anyway . to solve the auxiliary impurity problem ,
we have employed both the hubbard - i approximation and an implementation of the qmc method that has been used to study properties of the compressed lanthanides .
@xcite the former method is more approximate but much faster while the latter is more rigorous but computationally more expensive .
the hubbard - i approximation is also useful in various ways : it is more applicable in the large - volume atomic limit , where it gives insight into the atomic aspects of pu ; it provides a good initial guess for the dmft(qmc ) simulations ; and it makes it possible to examine the full @xmath5-@xmath5 coulomb interaction ( all four slater integrals @xmath61 ) and its consequent term structure in dmft(hi ) in contrast to dmft(qmc ) , where it is still difficult to go beyond just the standard hubbard @xmath62 . unless specifically noted otherwise , all dmft(hi ) and dmft(qmc ) results reported in this paper include only @xmath40 in addition to the spin - orbit intraction . for the dmft(qmc ) calculations , we used 6000 sweeps per qmc iteration for the imaginary time segments @xmath63 at a temperature @xmath64 , which is about as low a temperature as is practical at the present time with qmc , and performed at least 100 iterations . to improve the input data needed for the maximum entropy ( maxent ) analytical continuation , @xcite at the last iteration
, we dramatically increased the number of sweeps such that the number of bins , each containing 100 measurements , is larger than @xmath65 .
spectral density calculated from the dmft(hi ) [ left panels ] and dmft(qmc ) [ right panels ] for @xmath0-pu for varying @xmath1 site energy at @xmath66 k. in all panels , the spectral density contributed from the @xmath57 and @xmath58 sub - shells are represented by the black and red lines , respectively .
the total @xmath1 spectral density is represented by the blue line .
the energy @xmath67 denotes the position of the fermi level.,width=302 ] in fig .
[ fig : nfdep ] we show the spectral density of @xmath0-pu at @xmath68k for various values of the @xmath1 occupancy @xmath42 .
the left three panels display underlying atomic properties obtained from the dmft(hi ) calculations , consistent with a ground state @xmath69 , where @xmath70 and @xmath71 . for @xmath72 in fig .
[ fig : nfdep](a ) , one gets a single lower ( @xmath73 ) and upper ( @xmath74 ) hubbard band . for @xmath75 in fig .
[ fig : nfdep](c ) , however , one may still add an electron to the @xmath76 subshell , so that in addition to the lower ( @xmath77 ) one has two , spin - orbit split upper hubbard bands ( @xmath78 and @xmath79 ) .
for the mixed valent , nonintegral @xmath42 in fig .
[ fig : nfdep](b ) , the system behaves as an ensemble of @xmath80 @xmath81 and @xmath82 @xmath83 configurations . in the absence of hybridization , this leads to four hubbard bands in relative position @xmath84 , @xmath85 , @xmath86 , and @xmath87 , where @xmath88 is the spin - orbit splitting , with areas of @xmath89 , @xmath90 , @xmath91 , and @xmath92 , respectively . the fermi level ( energy zero ) should lie within the second hubbard band at @xmath85 splitting the occupied ( @xmath93 ) and empty ( @xmath94 ) parts , where the latter is the small tail of the @xmath76 spectra ( black ) extending above the fermi level in fig . [
fig : nfdep](b ) . for the more rigorous dmft(qmc )
results one sees , in the three right - hand panels of fig .
[ fig : nfdep ] , the expected transfer of spectral weight away from the hubbard bands into the quasiparticle peak at the fermi level , which dmft(hi ) is incapable of describing .
however , in addition there also appears to be a shift of the outlying dmft(hi ) hubbard structure closer towards the fermi level , while at the same time the one hubbard band already overlapping the fermi level [ second from left in fig .
[ fig : nfdep](b ) ] spreads away from the fermi level .
the overall affect is to give a more smooth and systematic evolution of the dmft(qmc ) spectra with growing @xmath42 than seen in dmft(hi ) .
this is confirmed by an independent evaluation of the dmft(qmc ) state density at the fermi level via @xmath95.@xcite @xmath96 smoothly decreases while @xmath97 ( and the total ) smoothly increase for increasing @xmath42 over the range @xmath98 , consistent with the fermi level passing more into the @xmath99 part of a spin - orbit split peak .
indeed , the maxent results in figs . [ fig : nfdep](d f ) exhibit a splitting in the quasiparticle peak .
on the one hand , we suggest that it is due to the spin - orbit interaction and the associated @xmath22-dependence induced into the self - energy .
such splitting appears in the spectra for each @xmath22 as the two channels couple via @xmath28 .
a reanalysis using maxent of earlier dmft(qmc ) calculations @xcite for ce and pr exhibits similar behavior , and since this is then independent of filling , hybridization dips or gaps can be ruled out . such spin - orbit induced splitting has been experimentally observed in photoemission experiments for ce , @xcite although the hund s rules exchange omitted in fig .
[ fig : nfdep ] ( and discussed shortly ) will have larger impact for mutli-@xmath5 electron atoms . on the other hand
, we do not exclude the possibility that the splitting may be simply an artifact of the analytical continuation technique used to obtain the spectral function at the real energy axis .
note that we find the quasiparticle peak and its associated fine structure to both disappear for @xmath0-pu at higher temperature , @xmath100k , consistent with the @xmath101k kondo energy that has been suggested .
@xcite spectral density calculated from the dmft(hi ) with only @xmath102 ( black line ) and full coulomb ( red line ) interactions for @xmath103 .
( c ) the @xmath1 spectral density calculated from the dmft(hi ) with full coulomb interactions for @xmath104 and scaled / shifted ones for @xmath103 and 6.0 . here
@xmath105 and @xmath106 .
, width=302 ] a comparison of the dmft(qmc ) spectra with experimental photoemission data @xcite is shown in fig .
[ fig : exp](a ) .
the @xmath38 valence contribution was added to the calculated @xmath1 spectra from figs .
[ fig : nfdep](d f ) , and the total was broadened to reflect the 60 mev instrument resolution and 15 k measurement temperature , as well as the lorentzian for the photohole lifetime determining the natural linewidth , including quadratic scaling with binding energy . for the experimental comparison , data from the 40.8 ev heii - alpha line were selected to obtain a photon energy range where orbitals of interest have similar cross sections .
@xcite while additional work is certainly needed for an optimal comparison with the data , fig .
[ fig : exp](a ) does reproduce the trend seen in the dmft(qmc ) results of figs .
[ fig : nfdep](d f ) , where the lower hubbard band is seen to move to more negative energies with increasing @xmath42 , which lends experimental support for a value of @xmath42 closer to 5 . the implications of the comparison in fig . [
fig : exp](a ) do not appear to be compromised by omitting hund s rules exchange and the associated term structure in the theoretical results , where only the monopole slater integral @xmath107 has been incorporated .
this can be tested in dmft(hi ) , where the black curve in fig .
[ fig : exp](b ) is the total @xmath108 @xmath1 spectra of fig .
[ fig : nfdep](c ) , while the red curve now treats the full coulomb interaction taking in addition reasonable experimental values for @xmath109 , @xmath110 , and @xmath111 . in the electron removal
spectra one now sees the standard terms ( labelled ) of the @xmath112 final state .
these agree in relative position and intensity with earlier theoretical analysis , @xcite and serve only to broaden the lower hubbard band to the negative - energy side , leaving the positive - energy side of importance to the comparison in fig .
[ fig : exp](a ) relatively unaffected . while this dmft(hi )
test can not probe the impact of hund s rules exchange on the quasiparticle peak near the fermi level , one might anticipate a reduced effect after such fine structure has been appropriately broadened for comparison to photoemission data .
there has been much discussion of a three - peak structure within about 1 ev of the fermi level in the photoemission of pu systems : in thin layers of puse , @xcite thin pu layers on mg , @xcite single crystal pute , @xcite and suggested in thin film of pusi@xmath113 @xcite and pun .
@xcite the presence of the structure in thin pu layers ( a few ml ) was interpreted as a result of localization effects due to low dimensionality overcoming the itinerant character of 5f electrons .
@xcite more recently it has been suggested that such structure may be evident in @xmath0-pu metal itself .
@xcite such structure may also be seen in dmft(hi ) calculations including the full coulomb interaction . in fig .
[ fig : exp](c ) , the black curve corresponds to a dmft(hi ) calculation with the @xmath1 site energy adjusted so @xmath114 . for comparison
the red ( blue / shaded ) curve shows the @xmath108 ( @xmath115 ) result scaled by @xmath116 ( @xmath117 ) and shifted so its lowest unoccupied ( highest occupied ) state is at the fermi level .
the three peaks discussed are shown labelled : @xmath118 essentially at the fermi level , and then @xmath119 and @xmath120 moving below .
given the agreement between the dmft(hi ) spectra @xmath121 for @xmath114 with the composite @xmath122 ( addition not shown ) , it is evident that especially the left two peaks arise from @xmath123 term structure in the final state of @xmath124 electron removal , and that the strength of these peaks might provide a direct measure of the @xmath125 admixture @xmath126 into the @xmath0-pu ground state , and therefore of @xmath42 itself .
there is little doubt that such an analysis is reasonable for some pu compounds , where precise atomic multiplet calculations were found in agreement with photoemission data in both line positions and intensities .
@xcite such agreement includes cases ( cubic pute and puse ) where the three - peak structure is experimentally observed below the fermi level approximately @xmath127mev .
the situation provided by a metallic environment as in @xmath0-pu is substantially different , however , and so the localized multiplet interpretation is not at all clear .
an obvious concern about the results in fig .
[ fig : exp](c ) , for example , is that the self - energy used in dmft(hi ) is too atomic - like ; the differences between the dmft(hi ) and dmft(qmc ) curves in fig .
[ fig : nfdep ] add to this concern .
the nature of the peak right at the fermi level in @xmath0-pu is clearer .
the low - temperature electronic specific heat in bulk @xmath0-pu metal is @xmath128 mj k@xmath129mol@xmath130 .
@xcite this large , heavy - fermion - like value indicates significant @xmath1 state density at the fermi level which is consistent with the kind of kondo - like physics that dmft(qmc ) can yield , albeit possibly at temperatures below the present work .
corroborating evidence comes from the photon energy dependent and temperature dependent photoemission experiments performed on clean @xmath0-pu metal surfaces , @xcite which strongly suggest that the first peak in the @xmath0-pu arises from @xmath1 electrons that are hybridized with the conduction electrons , which is again a typical quasiparticle feature usually found for materials with an enhanced electron mass .
we report a systematic lda - dmft study of the spectral properties of @xmath0-pu with varying @xmath1 occupancy .
the strong sensitivity of the spectral properties to the @xmath1 occupancy and the inherent ambiguity in defining the associated @xmath1 orbitals suggests the need to use @xmath1 occupancy as a fitting parameter in addition to the hubbard @xmath2 for dmft theories of pu and other materials . by comparing with photoemission data ,
we conclude that an `` open shell '' @xmath3 configuration gives better agreement with experiment than the hypothesized `` closed shell '' @xmath131 case . @xcite the dmft(hi ) approach , which is accurate in the localized , atomic limit , gives significantly different results from the more rigorous dmft(qmc ) method , which confirms the presence of a significant itinerant character in @xmath0-pu , especially at the fermi energy .
the `` 3-peak '' structure seen within about 1 ev of the fermi level in the photoemission spectra of some pu compounds as well as metallic thin films , but which remains controversial for the bulk @xmath0-pu metal , may provide a measure of the admixture of @xmath125 character in the ground state of these materials , and therefore an additional indication of @xmath42 .
however , it is not clear from the present work whether this structure should wash out in a correlated calculation for @xmath0-pu , which rigorously treats both extended and intraatomic exchange effects .
we are grateful to g. kotliar , v. oudovenko , s. y. savrasov , k. held , and c. d. batista for helpful discussions , and to j. e. gubernatis and m. jarrell for making available their maximum entropy code .
we acknowledge the support of the us doe at lanl under contract no .
de - ac52 - 06na25396 , and at llnl under contract no .
w-7405-eng-48 .
j. m. wills , o. eriksson , a. delin , p. h. andersson , j. j. joyce , t. durakiewicz , m. t. butterfield , a. j. arko , d. p. moore , and l. a. morales , j. electron spect . and rel .
phenom . * 135 * , 163 ( 2004 ) ; o. eriksson , d. becker , a. v. balatsky , and j. m. wills , j. j. alloys and comp . * 287 * , 1 ( 1999 ) . in dmft(qmc )
calculations using our constrained occupation values of both @xmath39 and @xmath40 along with an lmto - asa @xmath36 , we also find @xmath132 , in agreement with the spectroscopic - based conclusions of the present paper .
t. gouder , f. wastin , j. rebizant , and l. havela phys .
lett . * 84 * , 3378 ( 2000 ) .
t. gouder , l. havela , f. wastin , and j. rebizant europhys . lett . * 55 * , 705 ( 2001 ) .
l. havela , t. gouder , f. wastin , and j. rebizant phys .
b * 65 * , 235118 ( 2002 ) .
t. durakiewicz , j. j. joyce , g. h. lander , c. g. olson , m. t. butterfield , e. guziewicz , a. j. arko , l. morales , j. rebizant , k. mattenberger , and o.vogt , phys .
b * 70 * , 205103 ( 2004 ) .
t. gouder , r. eloirdi , j. rebizant , p. boulet , and f. huber , phys .
b * 71 * , 165101 ( 2005 ) .
l. havela , f. wastin , j. rebizant , and t. gouder phys .
b * 68 * , 085101 ( 2003 ) .
j. c. lashley , j. singleton , a. , migliori , j.b .
betts , r.a .
fisher , j.l .
smith , and r.j .
mcqueeney , phys . rev
. lett . * 91 * , 205901 ( 2003 ) .
j. j. joyce , t. durakiewicz , c. g. olson , e. guziewicz , k. s. graham , d. p. moore , l. a. morales , mat .
. proc . * 986 * , 35 ( 2007 ) . | by combining the local density approximation ( lda ) with dynamical mean field theory ( dmft ) , we report a systematic analysis of the spectral properties of @xmath0-plutonium with varying @xmath1 occupancy .
the lda hamiltonian is extracted from a tight - binding ( tb ) fit to full - potential linearized augmented plane - wave ( fp - lapw ) calculations .
the dmft equations are solved by the exact quantum monte carlo ( qmc ) method and by the hubbard - i approximation .
we demonstrate strong sensitivity of the spectral properties to the @xmath1 occupancy , which suggests using this occupancy as a fitting parameter in addition to the hubbard @xmath2 . by comparing with photoemission data ,
we conclude that the `` open shell '' @xmath3 configuration gives the best agreement , resolving the controversy over @xmath1 `` open shell '' versus `` close shell '' atomic configurations in @xmath0-pu . |
as general relativity is a purely metrical ( local ) theory it clearly leaves unsettled the global structure ( topology ) of space - time .
however , in cosmology perhaps the most important problems are related to the global structure of space - time , where the topological degrees of freedom ought to play an essential role .
geometry constrains , but does not determine the topology of a space - time .
consider , for example , the friedmann - robertson - walker ( frw ) space - times , whose line element can be given by @xmath3\ , , \label{frwds2}\ ] ] where @xmath4 is the scale factor , @xmath5 is the cosmic time , and the constant spatial curvature @xmath6 specifies the type of geometry ( flat , elliptic or hyperbolic ) of the @xmath7 spacelike section @xmath8 .
clearly frw space - time manifolds @xmath9 can be splitted into @xmath10 .
the number of three - dimensional spacelike manifolds @xmath8 which can be endowed with the three possible geometries of the @xmath0 3-spaces of frw space - time is quite large @xcite : for @xmath11 there are 18 topologically distinct 3-spaces , while for both @xmath12 an infinite number of 3-spaces exist @xcite . even if we restrict ourselves to orientable and compact manifolds we still have an infinite number of 3-spaces @xmath8 for the elliptic and hyperbolic cases , and six families of topologically different spacelike manifolds for the flat 3-spaces @xcite
@xcite . since physical laws are usually expressed in terms of local differential equations , in order to be confident about the physical results one derives it is often necessary to have some degree of control over the topological structure of the space - time manifold so as to include constraints imposed by the topology @xcite .
one is then confronted with the question of what topologies are physically acceptable for a given space - time geometry .
an approach to this problem is to study the possible observational ( or physical ) consequences of adopting particular topologies for the space - time @xcite @xcite ( see also @xcite and references therein ) . in this work we study the role played by the compactness and the degree of connectedness in the time evolution of the energy of a radiating system in the flat frw space - times whose spacelike @xmath0 sections are endowed with seven different topologies , namely the simply connected 3-space @xmath1 , and six multiply connected orientable compact 3-manifolds obtained by suitable identifications of opposite faces of cubes ( four ) and of hexagonal prisms ( two ) after suitable turns @xcite . the radiating system we shall be concerned with is represented by a point - like harmonic oscillator ( energy source ) coupled with a relativistic massless scalar field @xcite
similar radiating systems have been used to study a wide class of radiation phenomena as , for example , gravitational waves @xcite or the radiated energy of oscillating electromagnetic dipoles @xcite .
our radiating system was also used in the study of dynamic and geometric constraints on the radiation in elliptic frw expanding universes @xcite
@xcite . in the next section
the radiating system is described : the action integral is presented , its variation is performed and the corresponding evolution equations are obtained .
we also outline there a proposal for solving these equations . in section [ greenfuns ]
the green functions of the wave operator are derived for each specific space - time manifold with arbitrary scale factor @xmath4 , and combined with the evolution equations to give the corresponding radiation reaction equations .
a misinterpretation @xcite @xcite of wolf s @xcite results ( theorem 3.5.5 ) regarding one of the six orientable compact flat 3-manifolds is also indicated and rectified therein . in section [ numana ] , the radiation reaction equations are numerically integrated .
graphs that show how the system energy varies with the time for each topologically different flat frw manifold are presented for static and non - static cases .
our main conclusions are discussed in section [ conclusions ] .
we show that when the @xmath7 sections are the simply connected 3-space @xmath1 the radiation damping phenomenon is present , whereas for all compact flat frw space - time manifolds we have investigated , the energy @xmath2 exhibits a few relative minima and maxima followed by a growth of the energy with the time .
possible reasons for this divergent behavior of @xmath2 in these compact flat 3-manifolds are examined .
we also discuss the role played by the compactness and by the degree of connectedness in the energy patterns of our system in these space - time manifolds .
further developments are indicated therein .
radiation waves produced by an oscillating energy source as , e.g. , an oscillating electromagnetic dipole , have been studied by using a theoretical model represented by a classical harmonic oscillator ( energy source ) coupled with a relativistic massless scalar field ( scalar radiation waves propagating at speed of light ) @xcite @xcite . in our model
the gravitational field is treated as external , but a suitable conformal coupling with the scalar field is considered @xcite .
the dynamics of our system can be described by an action integral , which contains a term for the scalar field @xmath13 , another associated to the oscillation amplitude @xmath14 of the point - like harmonic oscillator , and a coupling term between the scalar field and the harmonic oscillator according to @xmath15 + \frac{1}{2 } \int dt \,[\,\dot{q}^2 -\omega_{\varepsilon}^{2}\ , q^2\ , ] \nonumber \\ { } & + & \lambda \int d^4x \ , \sqrt{-g } \,\,\rho(t,\vec{x } ) \,\,q(t)\,\ , \phi(t,\vec{x})\ : , \end{aligned}\ ] ] where @xmath16 , @xmath17 is the metric tensor on @xmath9 , @xmath18 , @xmath19 is the scalar curvature of @xmath9 , overdot means derivative with respect to @xmath5 , and @xmath20 is a coupling constant . the function @xmath21 is the normalized density function , which accounts for the coupling between the harmonic oscillator and the scalar field . similarly to the coupling between charges and electromagnetic fields in classical electrodynamics , we shall consider in this work a point - like coupling between the harmonic oscillator and the scalar field , namely the one in which @xmath22 , where @xmath23 is the three - dimensional dirac delta function .
this type of coupling requires a renormalization of the frequencies , and to this end we need an @xmath24-family of uncoupled frequencies @xmath25 ( see section [ greenfuns ] ) . here and in what follows units in which @xmath26 are used .
varying the action ( [ action ] ) with respect to @xmath27 and @xmath28 one obtains the coupled evolution equations of the system , namely @xmath29\ , \phi(t,\vec{x } ) & = & \lambda \ , \rho(t,\vec{x } ) \ , q(t ) \
, , \\
\label{motion2 } \ddot{q}(t ) + \omega_{\varepsilon}^{2}(t ) \ , q(t ) & = & \lambda \int d^3x \,\sqrt{-g } \,\rho(t,\vec{x})\,\phi(t,\vec{x}),\end{aligned}\ ] ] where @xmath30 is the dalembertian operator , and @xmath31 is the wave operator defined on @xmath9 , hereafter simply called wave operator @xcite . before proceeding to the discussion of the green functions for the wave operator in flat frw space - time manifolds ( for any smooth @xmath4 ) we shall consider how one can obtain the radiation reaction equation from the above equations ( [ motion1 ] ) and ( [ motion2 ] ) .
one first solves equation ( [ motion1 ] ) as an initial value problem , by writing the solution in the form @xcite @xmath32 , where @xmath33 satisfies the corresponding homogeneous equation , and where the solution of the inhomogeneous equation is given by @xmath34 with @xmath35 , \ : \vec{x } ' \in { \cal m}_3 $ ] . in ( [ inhoeq ] ) @xmath36 is the retarded green function , often referred to as fundamental solution of the wave operator @xcite . since the scalar field of our system may be thought of as the propagation medium for the radiating energy , we assume the initial condition @xmath37 , which means that the scalar field carries no energy at @xmath38 .
this condition implies that @xmath39 , and therefore @xmath40 .
finally , the radiation reaction equation of the energy source @xmath14 can be found by using the relation @xmath41 $ ] in equation ( [ motion2 ] ) .
we shall return to this point later in the next section . in this work
, the green function plays an essential role in that it incorporates both the geometrical and topological features of the frw @xmath0 3-spaces .
they will be obtained , in the next section , through the study of null geodesics of the space - times for each distinct 3-space ( see table [ flattops ] ) .
in this section we shall derive the green functions of the wave operator for the flat frw space - times whose spacelike @xmath0 sections are the multiply connected compact orientable 3-spaces , obtained by suitable identification of faces of a basic cell ( see refs .
@xcite ) according to the table [ flattops ] , and the simply connected space @xmath1 .
clearly the metric tensor we shall be concerned is given in cartesian coordinates by @xmath42 regarding the manifold @xmath43 one often encounters , in the literature on compact orientable flat 3-spaces , reference to a cube in which each pair of opposite faces is identified after half a turn @xcite
however , if such a cube is endowed with the euclidean geometry then the resulting three - space is not a manifold , but an orbifold @xcite .
this 3-space fails to be a manifold along the edges of the cube .
only when the cube is endowed with the elliptic geometry the resulting space is indeed a manifold , namely the well known real projective space @xmath44 .
recently , using theorem 3.5.5 by wolf @xcite , it has been shown that a basic cell for the euclidean manifold @xmath43 is the cube shown in figure 1 @xcite .
it should be stressed that the basic cell for @xmath43 shown in figure 1 is a standard basic cell ; in general , however , the basic cell for @xmath43 needs not to be a cube .
.__the six compact orientable topologies for the flat 3-manifolds can be obtained by identifying opposite faces of a basic cell as shown in this table . _ _ [ cols="^,<,<",options="header " , ] * the cubes and @xmath45 cases * we shall discuss in this part the green functions for the wave operator on 3-spaces @xmath46 ( @xmath47 ) shown in table [ flattops ] .
the manifold @xmath48 , referred to in the literature as the three - torus @xmath49 @xcite , is a compact multiply connected riemannian manifold , which can be obtained by identifying the opposite faces of a cube of side @xmath50 . to build the green function in @xmath51 we first use the conformal properties of frw space - times . defining the conformal time @xmath52 for @xmath35 $ ]
, we transform the problem of finding the green function @xmath36 of the operator @xmath53 into the problem of finding the green function @xmath54 of the operator @xmath55 @xcite . moreover , a useful way of thinking about @xmath48 in terms of the simply connected 3-manifold @xmath1 is to imagine the cube repeated endlessly in a three - dimensional grid ( basic cell and its images ) , where each repetition consists of the same physical region of space .
the space @xmath1 containing an infinite grid of copies of the basic cell is called the universal covering space of @xmath48 . in this way we now transform the problem of calculating the green function of @xmath55 at @xmath56 due to a point - like source located at @xmath57 and emitting at @xmath58 , to the equivalent problem of finding the field at @xmath59 due to an infinite set of point - like sources , each one located at the center of a cube of the grid , and all of them irradiating simultaneously at @xmath58 . in this infinite grid picture
the green function of @xmath60 is calculated summing up the contributions of each one of the point - like sources . since
a scalar ray is a null geodesic , the distance travelled by each scalar wave front is a measure of the corresponding time travel , so the conformal time @xmath61 of arrival at @xmath62 will depend on the position at which each source is located .
null geodesics corresponding to retarded waves connecting @xmath63 to the observation point @xmath64 are then such that @xmath65 , so the green function for @xmath66 is given by @xmath67 where @xmath68 , @xmath69 . here
@xmath70 is the usual orthonormal basis vectors in @xmath1 and @xmath71 are integers , and @xmath72 .
we shall consider now the radiation reaction equation of our energy source in @xmath48 .
using the green function @xmath73 and ( [ inhoeq ] ) one finds @xmath74 $ ] , which can be used in ( [ motion2 ] ) to furnish the radiation reaction equation of our system .
note that in these calculations the first term ( @xmath75 ) of the green function gives rise to the term @xmath76 , that formally diverges .
a suitable renormalization procedure is therefore necessary .
this resembles the need for renormalization one finds when dealing with accelerated point charges in classical electrodynamics @xcite .
we learn from @xcite that a renormalization procedure can always be made in this case .
this amounts to saying that for any @xmath77 , and any @xmath78 , one can always define an @xmath24-parameter family of frequencies @xmath79 by @xmath80 such that @xmath81 is a @xmath82-sequence that in the limit @xmath83 converges to the dirac delta @xmath84 .
using now ( [ epsfam ] ) , in which we have made @xmath85 , one finds that , for arbitrary initial data @xmath86 , the renormalized radiation reaction equation of the source can be written in the form @xmath87 for any @xmath88 , and where @xmath89^{1/2}$ ] for all @xmath71 integers not all zero , @xmath90 and the step distribution @xmath91 for any @xmath92 , @xmath93 for all @xmath94 .
note that the radiation reaction equation ( [ t1radeq ] ) contains an infinite , but countable , number of retarded terms .
note also that , for a given pair of initial data ( @xmath95 ) , the continuity of @xmath28 and @xmath96 for all @xmath5 is sufficient to ensure that ( [ t1radeq ] ) can be integrated to give a unique solution .
the next 3-manifolds we shall consider are @xmath97 , @xmath98 and @xmath43 ( see table [ flattops ] ) .
the 3-space @xmath97 is a compact multiply connected flat three - dimensional riemannian manifold , obtained by identifying opposites faces of a cube , with a pair being identified after a rotation of @xmath99 of one face relative to the opposite face .
the 3-manifold @xmath98 is obtained by identification of opposites faces of a cubic cell , but now a pair of faces are identified after a turn of @xmath100 ( see table [ flattops ] ) .
finally , the 3-manifold @xmath43 is obtained by pairwise identifying the faces of the cube according to figure 1 .
for these 3-spaces the green functions for the wave operator can again be built by using the infinite grid of cubes picture .
it should be noticed , however , that the grids of cubes corresponding to @xmath101 , @xmath102 , @xmath103 and @xmath104 are different .
( @xmath105 ) , where for each case @xmath106 is a different discrete group of isometries of the covering space @xmath1 acting properly discontinuously , without fixed points @xcite . ] nevertheless , as the center of a basic cell in each case represents the same infinite set of points ( images of the point - like source at ( @xmath107 ) ) of the covering space @xmath1 , the green functions for the wave operator in these 3-spaces are equal to the one we have calculated for @xmath48 , i.e. , are given by ( [ greent1 ] ) .
thus , for arbitrary initial data @xmath86 the radiation reaction equation in @xmath97 , @xmath98 and @xmath43 reduces to ( [ t1radeq ] ) obtained for @xmath48 .
a word of clarification is in order here : the above green functions for the wave operator for the 3-spaces @xmath108 ( @xmath109 ) coincide only because in each case the point - like source is located at the center of the basic cell @xmath110 .
this is so because although all these 3-spaces are locally homogeneous , the 3-spaces @xmath97 , @xmath98 and @xmath43 are not globally homogeneous , thus the corresponding green functions of @xmath55 for @xmath111 are more involved and do not depend simply on the relative position between the source and the observer . regarding the simply connected flat 3-dimensional riemannian manifold @xmath112 ( @xmath0 section of the minkowski space - time ) , the green function for the wave operator @xmath113 is equal to the first term ( @xmath114 ) of @xmath73 appearing in ( [ greent1 ] ) .
thus the radiation reaction equation in this manifold reduces to ( [ t1radeq ] ) with right hand side equal to zero , for all @xmath115 . * the hexagonal prism cases
* we shall discuss now the green functions for the wave operator on the flat hexagonal prism manifolds @xmath116 and @xmath117 described in table [ flattops ] .
the 3-space @xmath116 is a compact multiply connected flat three - dimensional riemannian manifold , which can be obtained by identifying the top and bottom faces of a regular hexagonal prism after a rotation of @xmath118 , while the lateral faces are pairwise identified in the usual manner .
the 3-manifold @xmath117 can be similarly defined , but now the identification of the top and bottom faces ( regular hexagons ) is made after a turn of @xmath119 . in both cases
we denote by @xmath50 the shortest distance between two opposite sides of the regular hexagon and by @xmath120 the height of the hexagonal prism cell . for a point - like source located at the center of the basic cell ,
the green functions for the wave operator in the 3-manifolds @xmath116 and @xmath117 are obviously the same .
moreover they turn out to be similar to the green function ( [ greent1 ] ) .
but now instead of @xmath121 one has @xmath122 , where @xmath123 are integers , to locate the centers of the hexagonal prism cells in which @xmath1 has been tessellated .
note that the number of images at a given distance from the point - like source clearly depends upon the ratio @xmath124 .
the radiation reaction equation in the present cases is similar to ( [ t1radeq ] ) but with the obvious change @xmath125^{1/2}$ ] and again the integers @xmath123 are not all zero . it should be emphasized that as the 3-manifolds @xmath116 and @xmath117 are locally but not globally homogeneous , the green function for the wave operator in these manifolds is again given by ( [ greent1 ] ) , with @xmath126 instead of @xmath121 , only because the point - like source is at the center of the basic cell @xmath127 . to close this section
we emphasize that the green functions , obtained for flat frw space - time manifolds @xmath128 , contain the topological constraints of @xmath8 , information that makes possible to find out the exact radiation reaction equation in each case .
in this section we shall discuss the time behavior of the energy of the harmonic oscillator @xmath129 $ ] , where the function @xmath14 is the solution of the radiation reaction equation corresponding to each flat manifold we have discussed in the previous section . without loss of generality , in the integration of the radiation reaction equations and in the plotting of the energy function we have taken specific values for the constants .
we have also assumed the continuity of @xmath28 and @xmath96 , and chosen suitable values for the initial data @xmath130 and @xmath131 . for
a neat comparison between the simply and multiply connected cases we have chosen @xmath132 , @xmath133 , the length @xmath134 , and for the heights of the hexagonal prism and three - torus we have taken @xmath135 and @xmath136 . as a matter of fact , our three - torus @xmath48 was obtained from a parallelepiped with edges @xmath137 , so @xmath138^{1/2}$ ] was used in equation ( [ t1radeq ] ) .
further , we have also chosen as initial data @xmath139 , @xmath140 , which means that the initial energy of the source has been normalized , i.e. , @xmath141 .
figures 2 , 3 and 4 correspond to the static case @xmath142 , which we have normalized to 1 ( @xmath143 ) , while dynamic situations ( @xmath144 ) are considered in figure 5 .
taking into account the above choices of values and using the computer algebra systems _ mathematica _
@xcite and _ maple _ @xcite the numerical integrations of the radiation reaction equations as well as the corresponding graphs for the energy function were obtained ( see figures 2 to 5 ) .
figure 2 shows the behavior of the energy with the time for the flat frw space - times , in which the @xmath0 sections are any of the orientable compact multiply connected 3-manifolds @xmath48 , @xmath97 and @xmath98 and @xmath43 for different ratios @xmath124 .
note that @xmath120 is the height of the parallelepiped of square basis of side @xmath50 .
the curves exhibit basically the same pattern , namely relative minima and maxima followed by a predominant growth of the energy with the time ( @xmath145 when @xmath146 ) .
these extrema are related to the contribution of the discrete retarded terms ( the right hand side of the radiation reaction equation ) demanded by the compactness and the corresponding connectedness of these manifolds .
the fact that the relative extrema occur at different times and are of different amplitudes ( intensities ) for distinct tessellations ( different ratios @xmath124 ) of the covering manifold , basically reveals the differences in their degree of connectedness ( returning rays take different times to return to the origin ) .
this growth of the energy is discussed in the next section . regarding the behavior of the energy function for the cases in which the @xmath7 sections are @xmath116 and @xmath117 , shown in figure 3
, we again note that due to the compactness and connectedness we have relative minima and maxima with distinct intensities , which take place at different instants for distinct ratios @xmath124 .
the curves again display an eventual growth of the energy with the time for these manifolds .
figure 4 compares the variation of the energy @xmath147 with the time for the cases where the @xmath0 section @xmath8 are @xmath1 , @xmath48 ( with @xmath148 ) and @xmath116 ( also with @xmath148 ) .
this figure shows for the minkowski space - time , as expected , an exponential decay of the energy with the time , whereas for the manifolds @xmath48 and @xmath116 it shows basically the same pattern , i.e. relative minima and maxima occurring at different times , depending on the degree of connectedness , followed by a growth of the energy with the time . although the net role played by the degree of connectedness as well as compactness can be singled out in the static cases @xmath149 , for the sake of completeness we have examined three instances where dynamic expansion takes place .
figure 5 corresponds to the plot of the energy function for the @xmath48 manifold ( with @xmath148 ) in the dynamic expanding cases : ( i ) linear expansion @xmath150 , ( ii ) square root expansion @xmath151 , ( iii ) inflationary expansion @xmath152 , that is , an expansion with future event horizon .
although for the case ( iii ) one clearly has radiation damping , we have not been able to find out so far a closed formal proof of the asymptotical behavior for the other two cases .
we emphasize , nevertheless , that the net role played by the connectedness and compactness can be better singled out in the static cases , where the dynamical degrees of freedom are frozen .
the study of that role for the static cases is in fact the major aim of the present work .
in this work we have studied the role played by the topological compactness and connectedness in the time evolution of the energy of an harmonic oscillator in flat frw space - time manifolds , whose @xmath0 sections are ( i ) the orientable simply connected non - compact 3-space @xmath1 and ( ii ) six possible flat orientable multiply connected compact 3-manifolds given in table [ flattops ] . for the @xmath1 case we found that the energy function @xmath2 exhibits an exponential decay with the time the radiation damping ( @xmath153 when @xmath146 ) takes place , as one could have expected in agreement with @xcite
@xcite . for the manifolds @xmath48 , @xmath97 , @xmath98 and @xmath43 as well as for the manifolds @xmath116 and
@xmath117 the behavior of the energy with the time exhibits basically the same pattern : relative minima and maxima occur at different times for distinct ratios @xmath124 ( distinct tessellations ) depending on the degree of connectedness of each 3-manifold , and are followed by a growth of @xmath2 .
this asymptotical divergent behavior of @xmath2 for these compact manifolds contrasts with the radiation damping of the energy we have found for @xmath1 .
there is a quite simple heuristic argument which suports our numerical results , though .
if one ad hoc assumes an exponential asymptotical behavior for @xmath28 , i.e. @xmath154 with @xmath155 and @xmath156 real constants , then for the static cases ( @xmath157 ) , and for each of the above compact manifolds , in the limit @xmath146 equation ( [ t1radeq ] ) reduces to @xmath158 where @xmath159 is the number of images of the point - like source at a distance @xmath160 in the infinite grid picture . to attain our goal
, we will show that ( [ einfty ] ) has only one real solution for @xmath155 , which is positive .
indeed , let @xmath161 be the right hand side of ( [ einfty ] ) , which is a positive monotone decreasing function of @xmath155 , and such that @xmath162 thus @xmath161 lies entirely in the first quadrant of the plane and crosses it from the top - left to the bottom - right .
now , since @xmath163 then for a given pair ( @xmath164 ) the left hand side of ( [ einfty ] ) is a parabola curved upwards with vertex at @xmath165 .
therefore , it always intersects the curve for @xmath161 in just one point , which is in the first quadrant . in other words , there is only one real @xmath155 solution to equation ( [ einfty ] ) , which is positive .
the unexpected ( unphysical ? )
growth of the energy with the time for the above compact flat 3-manifolds cases illustrates that non - trivial topologies can induce rather important dynamic changes in the behavior of a physical system .
this type of sensitivity has been refered to as _
topological fragility _ and can occur without violation of any local physical law @xcite . a rigorous _ non - numerical _ analysis of the reasons for this surprising divergent behavior of @xmath2
when compact flat frw space - times are considered has been carried out , and we hope to publish our results shortly elsewhere .
we anticipate , however , that the causes for such a behavior lie in the compactness of the manifold in at least one direction , on the one hand , and in the type of coupling between @xmath28 and @xmath27 , on the other hand . a possible physical measure of the degree of connectedness in these _ compact _ 3-manifolds can be made through the study of the number of emitted rays that return to the origin within a given lapse of time . according to this concept of degree of connectedness one learns from figures 2 , 3 and 4 that the greater is the degree of connectedness the earlier is the occurrence of the first relative minimum in the energy function .
incidentally , note that the extension of this concept of degree of connectedness to non - compact 3-manifolds implies that @xmath1 has a null degree of connectedness .
this , of course , is indicated in figure 4 , which shows a net exponential decay of the energy with the time ; no relative minima and maxima come about , which means that no ray returns to the origin .
a simple inspection of the graphs for @xmath2 clearly shows that the derivative @xmath166 of the energy function is discontinuous at a few points .
indeed , for the static case , for example , using ( [ t1radeq ] ) one obtains @xmath167 \,.\ ] ] from this equation one sees that the discontinuities occur at @xmath168 , that is , they come about each time a new term @xmath169 is taken into account in the right hand side of ( [ t1radeq ] ) .
a question which naturally arises here is whether the inverse problem , i.e. that of determining the basic cell ( topology ) corresponding to the spacelike @xmath7 sections from the graphs of the energy @xmath2 , can be solved . regarding this problem
it is clear that one can find the distances of the point - like source to its images by using where the discontinuities of @xmath170 take place , and the number of images at a given distance through the magnitude of the corresponding discontinuities .
so , one can probe the topology of the 3-spaces at least in a few cases .
it is not yet clear whether an algorithm for solving the inverse problem for the most general ( euclidean ) setting can be found , though . as far as we are aware
@xcite this is the first work in which a physical consequence of adopting the flat hexagonal prisms @xmath116 and @xmath117 has been studied .
it is of worth to emphasize that when the expansion of the universe is considered the degree of connectedness is less than the ones for the static cases . for the manifold @xmath171
this can be seen by comparing figure 2 ( static case @xmath172 ) and figure 5 ( monotone expansions @xmath173 ) . before closing this article
we should like to stress that our study does not cover all possible spatially compact orientable flat frw manifolds .
thus , for example , we have not considered that for the 3-manifolds @xmath48 and @xmath98 the basis of the basic cell need not to be a square , it can be a parallelogram .
the restriction we have made , however , does not seem to be decisive for the patterns of the behavior of the energy with the time we have found . to conclude we remark that the study of radiation damping in elliptic ( @xmath174 ) frw manifolds in which the @xmath7 sections are endowed with different ( orientable compact ) topologies is being carried out .
we are grateful to an anonymous referee for valuable suggestions .
we also thank the brazilian scientific agencies cnpq and capes for financial support .
is also grateful to claf . | we study the role played by the compactness and the degree of connectedness in the time evolution of the energy of a radiating system in the friedmann - robertson - walker ( frw ) space - times whose @xmath0 spacelike sections are the euclidean 3-manifold @xmath1 and six topologically non - equivalent flat orientable compact multiply connected riemannian 3-manifolds .
an exponential damping of the energy @xmath2 is present in the @xmath1 case , whereas for the six compact flat 3-spaces it is found basically the same pattern for the evolution of the energy , namely relative minima and maxima occurring at different times ( depending on the degree of connectedness ) followed by a growth of @xmath2 .
likely reasons for this divergent behavior of @xmath2 in these compact flat 3-manifolds are discussed and further developments are indicated . a misinterpretation of wolf s results regarding one of the six orientable compact flat 3-manifolds is also indicated and rectified . |
abundance studies of very metal - poor ( vmp ; [ fe / h ] @xmath3 ) = @xmath4 , and @xmath5 for elements a and b. ] stars have been pursued over the past few decades in order to constrain models of nucleosynthesis , stellar evolution , and early chemical enrichment in the galaxy ( e.g. , * ? ? ?
one important result of these studies is the discovery of carbon enhanced metal - poor ( cemp ) stars , which appear with increasing frequency at lower metallicity @xcite .
these stars may be closely related to carbon stars in the galactic halo , known as ch stars @xcite and subgiant ch stars @xcite .
recent chemical abundance studies for cemp stars have revealed that most ( 7080% ) cemp stars also exhibit excesses of s - process elements such as ba ( the cemp - s stars , according to beers & christlieb 2005 ) , indicating that the origin of the carbon excesses in these stars is likely to be the triple-@xmath6 reaction in asymptotic giant branch ( agb ) stars ( e.g. , * ? ? ? * ) .
the cemp stars that are observed at present are likely to have been polluted by the transfer of carbon - enhanced material from a ( former ) agb companion across a binary system , while the agb star itself has now evolved to an unseen white dwarf ( e.g. , * ? ? ?
thus , the abundance patterns of heavy elements in these stars provide useful constraints on models for s - process nucleosynthesis in agb stars . on the order of 20% of cemp stars exhibit no significant enhancement of their neutron - capture elements ( the cemp - no stars , according to beers & christlieb 2005 ) , suggesting the existence of other possible origins for their carbon excesses ( e.g. , * ? ? ?
* ; * ? ? ?
@xcite have shown that the cemp - no stars generally occur at very low [ fe / h ] ; extreme examples of this class of stars include he 01075240 and he 13272326 , two hyper metal - poor ( hmp ) stars with [ fe / h ] below @xmath7 @xcite and having very large carbon excesses ( [ c / fe ] @xmath8 ) , as well as the recently identified ultra metal - poor ( ump ; [ fe / h ] = @xmath9 ) star he 0557 - 4840 , with [ c / fe ] @xmath10 ( norris et al .
2007 ) . among the cemp stars for which chemical compositions have been obtained from high - resolution spectroscopy
, main - sequence turn - off stars are expected to be of particular importance . in the case of mass transfer in binary systems ,
the accreted material from the primary agb star has been mixed at least by the first dredge - up in red giants , while turn - off stars might preserve pure material accreted from the primary at their surfaces . in such cases
, one can investigate the efficiency of the production of carbon and neutron - capture elements in agb stars from abundance measurements of the secondary star .
another interesting view arises from the suggested influence of so - called thermohaline mixing ( charbonnel & zahn 2007 ; stancliffe et al . 2007 ; denissenkov & pinsonneault 2007 ) , which provides the possibility of mixing the accreted surface material while the observed star is still on the main - sequence or only slightly evolved , prior to first dredge - up . in this scenario ,
the contrast of the observed surface abundances of cemp turn - off stars with more evolved cemp stars also provides valuable clues to the nature of this proposed extensive mixing process .
very large new samples of cemp stars have recently become available , discovered during the course of the sloan digital sky survey ( sdss ; york et al .
2000 ; adelman - mccarthy et al .
although originally designed as an extragalactic survey , sdss has also discovered large numbers of vmp stars @xcite .
although some of the cemp stars are the result of directed studies ( margon et al .
2002 ; downes et al . 2004 ) , many of them have appeared among the calibration objects used by sdss to obtain spectrophotometric and telluric corrections for other spectroscopic data . these calibration stars are primarily brighter , metal - poor main - sequence turn - off f- and g - type stars .
the ongoing extension to sdss , sdss - ii ( which includes the program segue : sloan extension for galactic understanding and exploration ) , is expected to provide tens of thousands of additional vmp stars , at least several thousand of which are expected to be cemp stars .
this paper reports the first application of abundance measurements obtained with high - resolution spectroscopy for cemp star candidates found by the sdss / segue surveys . in 2
we discuss the identification of our sample stars and the observations that were carried out .
a description of our analysis techniques and results is provided in 3 . in
4 we present a discussion of our findings .
the interesting kinematics of the sdss / segue cemp turn - off stars are discussed in 5 .
we conclude with a few remarks in 6 .
the sloan digital sky survey uses a ccd camera ( gunn et al .
1998 ) on a dedicated 2.5 m telescope ( gunn et al . 2006 ) at apache point observatory , new mexico , to obtain images in five broad optical bands ( @xmath11 ; fukugita et al . 1996 ) over approximately 10,000 deg@xmath12 of the high galactic latitude sky . the survey data - processing software measures the properties of each detected object in the imaging data in all five bands , and determines and applies both astrometric and photometric calibrations ( pier et al .
2003 ; lupton et al . 2001 ; ivezi et al .
photometric calibration is provided by simultaneous observations with a 20-inch telescope at the same site ( hogg et al .
2001 ; smith et al . 2002
; stoughton et al . 2002 ; tucker et al .
2006 ) . during the development of pipeline software for the determination of atmospheric parameters ( @xmath13 , @xmath14 , [ fe / h ] ) for stars with available photometry and spectroscopy from sdss and segue ( the segue stellar parameter pipeline ; sspp , see lee et al .
2007a , b ) , it was noticed that a rather large number of stars were present in the sdss / segue database with clearly enhanced ch g - band strengths , and which were likely to be cemp stars .
a list of over 1000 candidate cemp stars was assembled , drawing in particular on the calibration stars used by sdss .
the sample formed the basis for a detailed investigation of the frequency of cemp stars in the sdss database ( see marsteller et al . 2006 ; marsteller 2007 ) .
a handful of the brighter examples of the cemp turn - off stars were identified for carrying out a pilot study of their high - resolution spectroscopic abundances , reported on herein .
the imaging procedures used during the course of sdss are tuned for extragalactic observations . as a result
, there exists a bright limit corresponding to @xmath15 .
thus , the stars available for our study are somewhat faint for high - resolution abundance analyses , even with 8 m - class telescopes .
the targets for the present observing program are listed in table [ tab : obs ] .
figures [ fig : sdss1 ] and [ fig : sdss2 ] show the medium - resolution ( @xmath16 ) sdss spectra of the targets .
in addition to our primary objects , we selected two well - known cemp turn - off stars , lp 7067 @xcite and cs 29526110 @xcite , as comparison stars .
the effective temperatures are primarily estimated from adopted @xmath17 colors ( see [ sec : ana ] ) .
the photometric data and reddening corrections used in this work are listed in table [ tab : photo ] . for the sdss / segue stars the optical photometry information ( @xmath18 and @xmath19 ) are obtained from the sdss photometric system , employing the following empirical transformations , obtained by comparison with existing photometry for hk survey stars and subsequently observed by sdss ( zhao & newberg 2006 ) : @xmath20 + @xmath21 + the photometric data for the comparison stars are taken from @xcite .
the @xmath22 photometry is obtained from the 2mass catalogue @xcite .
the interstellar reddening is estimated from the dust map of @xcite ; the extinction in the @xmath19 and @xmath22 bands is obtained from the reddening relation provided in their table 6 .
high - resolution spectroscopy was obtained with the subaru telescope high dispersion spectrograph ( hds ; noguchi et al .
2002 ) in september 2006 and february 2007 .
our spectra cover the wavelength range from 4050 to 6800 , with a gap between 5350 and 5450 due to the separation between the two detectors .
a two - by - two pixel on - chip binning procedure was applied .
the resolving power of the spectra obtained in 2006 is @xmath23 ( using a slit width of 0.6 ) , while that obtained during the 2007 run is somewhat lower ( @xmath24 ) because a wider slit width ( 0.9 ) was applied in order to collect sufficient photons under relatively poor seeing conditions .
the total exposure times are listed in the third column of table [ tab : obs ] .
it was immediately obvious , during the course of the observing run , that sdss 1707 + 58 exhibited a rapid variation in its radial velocity .
the exposure times for individual exposures for this object are listed separately in table [ tab : s0353 ] .
the total exposure time for this object in table [ tab : obs ] is the value for the spectrum used in the abundance analysis ( see [ sec : ana ] ) .
data reduction was carried out using standard procedures within iraf .
cosmic - ray hits were removed using the procedure described by @xcite .
the wavelength scale was determined using th - ar arc spectra obtained during the observing nights .
examples of the spectra around 5170 are shown in figure [ fig : sp ] .
the photon counts per pixel ( 0.031 ) collected at 5100 in the final spectra are listed in table [ tab : obs ] .
the equivalent widths of atomic lines are measured by fitting gaussian profiles to ( apparently ) isolated absorption features .
the list of atomic lines was made using those of our recent studies for cemp stars @xcite ; the values are given in table [ tab : ew ] .
equivalent widths of the interstellar d1 line ( @xmath25 ) are measured from a direct integration of the absorption features .
the values are given in the last line of table [ tab : ew ] .
we estimated @xmath26 from the absorption by applying the correlation found by @xcite .
the values are given in table [ tab : photo ] .
the agreement between the two estimates of @xmath26 is good in general .
an exception is sdss 0126 + 06 , for which a significantly larger @xmath26 is obtained from the line than from the dust map of @xcite .
this is discussed in
[ sec : sdss ] .
radial velocities for our program stars are measured from the wavelengths of clean fe lines ; results are given in table [ tab : obs ] .
the random errors of the measurements are estimated from @xmath27 , where @xmath28 is the standard deviation of the values from individual lines and @xmath29 is the number of lines used for the measurement . note that the reported error for the sdss 0817 + 26 value is substantially larger than for the rest of our targets , because of the low @xmath30 ratio of the spectrum and the small number of fe lines used ( 5 lines ) .
figure [ fig : rv ] shows the radial velocities of lp 7067 and cs 29526110 obtained by the present work , along with additional measurements obtained during recent service observing runs with the subaru telescope .
for lp 7067 , the results of @xcite are also shown .
a clear variation of radial velocity is found for cs 29526110 , indicating that this object belongs to a binary system , although the orbital period is still unclear . on the other hand ,
no evidence of variability is obtained for lp 7067 , as has been discussed by @xcite , even though we have now obtained data extending over a range of some 6000 days .
it was immediately noticed during the first observing night of 2007 that significant variations existed in the observed doppler shift of sdss 1707 + 58 .
figure [ fig : s0353 ] shows the spectra of this object before any shifts have been obtained prior to co - addition for later analysis .
no velocity variation is found for the d emission lines associated with the earth s atmosphere at 5989.9 and 5995.9 , nor for the interstellar absorption feature due to na , found about 0.5 blueward of these emission lines , indicating that the wavelength calibration was correctly carried out over all exposures .
in contrast , absorption due to the stellar features quickly shifted in the spectra obtained in the first observing night ( the upper three spectra in the figure ) , while no significant variation is found for those obtained during the second night .
we note that the exposure times applied to the first and second observing nights are 40 and 20 minutes , respectively .
the shift in the first night is about 18 km s@xmath31 per 40 minutes .
hence , the broad absorption features of the first three spectra are almost certainly due to a radial velocity shift during the exposures rather than stellar rotation or macro - turbulence .
the radial velocities and line widths measured for each spectrum of sdss 1707 + 58 are listed in table [ tab : s0353 ] .
the measurements are made for strong , , and lines , because measurements from the lines for individual exposures were quite uncertain . during the course of this procedure ,
it was decided to exclude the third spectrum obtained during the second night because of significant contamination from the twilight sky .
the ut and jd ( heliocentric julian day ) of the central time of each exposure is listed in the table .
a spectrum of this object , suitable for carrying out the abundance analysis described below , is obtained by combining individual spectra obtained during the first night , after applying the appropriate doppler corrections .
radial velocities are also measured from the sdss spectra , as listed in table [ tab : rv_sdss ] .
the radial velocities of sdss 003610 , sdss 0126 + 06 , and sdss 0924 + 40 agree with the results from the subaru spectra within the measurement errors .
the radial velocity of sdss 2047 + 00 on jd = 2,452,932 is significantly higher than the other three measurements , suggesting a radial velocity variation and binarity of this object .
we note that this object shows large over - abundances of carbon and neutron - capture elements ( [ sec : ana ] ) that are expected from mass transfer in a binary system .
the radial velocity of sdss 1707 + 58 from the sdss spectrum is within the variation found in the subaru spectra given in table [ tab : s0353 ] .
finally , the radial velocity of sdss 0817 + 26 from the sdss spectrum , 45.8 @xmath32 km sec@xmath31 , is much higher than the value obtained from the hds spectrum
. however , the measurement is only once for each instrument , so further measurement is required to derive any firm conclusions on the binarity of this object .
we determine the effective temperatures from the @xmath17 colors using the scale of @xcite ; these are listed in table [ tab : photo ] as @xmath13@xmath33 .
the @xmath17 values of cs 29526110 , sdss 003610 , sdss 2047 + 00 , sdss 0126 + 06 and sdss 1707 + 58 are slightly lower than the range for which the alonso et al . scale ( formula ) is applicable ( @xmath34 for [ fe / h ] @xmath35 ) . for these objects
, we directly estimate the effective temperature from figure 8 of @xcite , in which the correlation between @xmath36 and effective temperatures for their calibration stars is shown .
the effective temperatures obtained from the @xmath37 colors using the alonso et al .
scale are also listed in table [ tab : photo ] . for cemp stars ,
this color is sometimes severely affected by the presence of molecular absorption , and it is not preferable for temperature estimates .
however , the molecular features of warm cemp stars studied here are not as significant as those for cooler stars , and would be expected to have less of an affect on the observed colors . moreover , the errors in the @xmath22-band photometry for some of the fainter objects in our sample are large ( see below ) , and the @xmath36 colors are relatively sensitive to the reddening correction .
hence , the effective temperatures obtained from the @xmath37 colors are useful for comparison purposes .
the @xmath13@xmath38 of sdss 003610 and sdss 1707 + 58 agree well with their @xmath13@xmath33 determinations .
the @xmath13@xmath38 of the two coolest stars in our sample , lp 7067 and sdss 0924 + 40 , are lower than their @xmath13@xmath33 , perhaps as the result of their relatively strong ch molecular bands affecting the @xmath18-band measurement .
a similar discrepancy between the two @xmath13 estimates is found for sdss 0817 + 26 , even though this object exhibits no or perhaps only a modest ( [ c / fe]@xmath39 ) carbon overabundance . however , the error in the @xmath22 photometry for this star is relatively large , which might explain the discrepancy .
the @xmath40 of sdss 2047 + 00 is quite high ( 6800 k ) for a vmp turn - off star .
however , the error of the @xmath22 photometry ( 0.12 magnitudes ) and the reddening correction for this star ( @xmath41 ) are the largest among our sample .
moreover , the @xmath42 is about 200 k lower than the @xmath43 . for this object
we adopt @xmath13 = 6600 k , which is slightly lower than the estimate from the @xmath36 color .
the @xmath13 of cs 29526110 is also quite high ( 6800 k ) .
however , the reported error of the @xmath22 photometry for this object is 0.03 magnitudes , and the reddening correction adopted ( 0.033 magnitudes ) is not large . for this object ,
the @xmath44 and @xmath45 photometry data are also available @xcite .
the @xmath13 from @xmath46 and @xmath47 estimated using the figures of @xcite are 67006800 k , while the @xmath13@xmath38 is 6500 k. we adopt the @xmath13 from @xmath17 with no modification for this object .
we now estimate the uncertainty in the adopted @xmath13 for our program stars , taking the error in the @xmath17 colors and the error in the scale of @xcite into consideration .
the error of the @xmath22 photometry is the dominant source of the uncertainty in the @xmath17 values for most objects .
the sensitivity of @xmath13 to the color is approximately 150 k per 0.1 magnitude in @xmath36 .
we adopt 100 k as the errors arising from the temperature scale for stars with @xmath48 , for which alonso et al.s formula is applicable , and 150 k for other objects , respectively .
the uncertainties are 100 - 150 k for relatively cool or bright objects ( sdss 0817 + 26 , sdss 0924 + 40 , lp 7067 , and cs 29526110 ) , and 150 - 200 k for others .
the adopted errors of @xmath13 in this study are listed in table [ tab : param ] .
the surface gravity , metallicity , and micro - turbulence for our program stars are determined from an analysis of the and lines , using the model atmospheres of @xcite . the micro - turbulence ( @xmath49 ) and gravity ( log @xmath50 )
are determined so that the derived fe abundance is not dependent on the strengths of fe i lines , nor on the ionization stages , respectively .
an exception is sdss 0817 + 26 , for which the number of useful lines is too small to estimate the micro - turbulence , and no line is available to estimate the gravity .
we adopted typical values ( @xmath14 = 4.0 and @xmath51 = 1.5 km s@xmath31 ) found for turn - off stars ) is adopted for the case of a horizontal branch star , the derived iron abundance is only slightly higher , while the derived sr and ba abundances ( see 3.3 ) are about 0.6 dex lower . ] .
we note that this object is excluded in the discussion of cemp stars because its carbon abundance is not determined by our analysis of the hds spectrum ( see below ) ; we only obtain a weak upper limit for [ c / fe ] .
the number of lines used in the analysis of sdss 1707 + 58 is also quite small , due to the rapid changes of the radial velocity ( see below ) . for this object ,
@xmath51=1.5 km s@xmath31 is also adopted . for sdss 0126 + 06
, a correlation between the line strengths and the derived fe abundances is found even if @xmath51@xmath52 km s@xmath31 is assumed .
since such a high value of @xmath51 is not known in turn - off stars , we adopt @xmath51@xmath53 km s@xmath31 for this object .
larger errors in the gravity ( @xmath54[@xmath14 ] ) and the micro - turbulence ( @xmath54[@xmath51 ] ) are adopted for these objects .
the atmospheric parameters adopted in the following abundance analyses and their corresponding errors are listed in table [ tab : param ] .
figure [ fig : teffg ] shows the estimated effective temperatures and surface gravities for our sample ( filled circles ) along with other cemp stars studied in previous work ( open circles ; see below ) .
the lines are the isochrones by @xcite for [ fe / h]@xmath57 and assumed ages of 10 , 12 , and 14 gyrs .
inspection of this figure shows that our objects fall around the turn - off region for old metal - poor stars , although it is difficult to distinguish whether they are main - sequence stars or subgiants .
we note that if we adopt a higher effective temperature ( 6800 k from @xmath36 ) for sdss 2047 + 00 , the surface gravity also becomes quite high ( @xmath14=4.9 ) , far above the expected value based on isochrones of vmp turn - off stars .
the carbon abundance estimates for our program stars are determined from spectrum synthesis of the ch 4323 band , as previously described by @xcite .
the sources of molecular data are reported by @xcite . the oxygen abundance of [ o / fe]=+0.5 is assumed in the analysis .
we confirmed that the effect of assumed oxygen abundance on the derived carbon abundances is negligible for the range 0@xmath58[o
/ fe]@xmath59 for a star with @xmath13@xmath60 k , in which the fraction of carbon bound in the co molecule is very small .
no signature of the ch band is detected in the hds spectrum of sdss 0817 + 26 , so only an upper limit is determined ( note that in the medium - resolution sdss spectrum , there is sufficient strength in this band , and others , to obtain a detection , [ c / fe]@xmath61 ; see [ sec : sdss ] ) .
the determination of carbon abundance for sdss 1707 + 58 , based on the high - resolution spectrum alone , is very uncertain because of the relatively low s / n ratio of the spectrum .
the full set of results is listed in table [ tab : abund ] .
carbon abundances are also measured from the c@xmath62 swan band at 5165 for lp 7067 , sdss 003610 , and sdss 0924 + 40 .
the result for sdss 0924 + 40 agrees well with that obtained from the ch band , while the carbon abundances of lp 7067 and sdss 003610 from the c@xmath62 band are slightly ( 0.10.2 dex ) higher than those from the ch band , as was also found for the cemp subgiant lp 62544 by @xcite .
although there may exist a small systematic error in the determination of carbon abundances from the c@xmath62 band and/or from the ch band , the measurements from the c@xmath62 band confirm the reliability of carbon abundance determination from the other molecular band .
the abundances for most of the other elements are determined by a standard analysis based on measured equivalent widths .
the effects of hyperfine splitting and isotope shifts are included in the analysis , using @xcite for ba , @xcite for la , and @xcite for pb .
solar isotope ratios are assumed for pb . for ba , we first measured the abundances neglecting the effect of hyperfine splitting , and then applied the isotope ratios of the r - process component in solar system material for the two stars having [ ba / fe ] @xmath63 ( sdss 003610 and sdss 0817 + 26 ) , and s - process ratios for the stars that exhibit ba excesses , as was done previously by @xcite .
while sr and ba abundances are measured for all objects in our sample , other neutron - capture elements are detected in only a few stars .
the abundances of pb , which is a key element for investigation of neutron - capture nucleosynthesis , are measured for lp 7067 , cs 29526110 , sdss 0126 + 06 and sdss 0924 + 40 , while an upper limit is estimated for other stars . the upper limit on the pb abundance
is calculated based on the 3@xmath54 error of the equivalent - width measurement , estimated by @xmath64(@xmath44[s / n ] ) , where @xmath44 is the resolving power and @xmath65 is the number of pixels for which equivalent width measurements are carried out @xcite . the results are listed in table [ tab : abund ] .
six of our program stars exhibit large ba over - abundances .
in particular , the over - abundance found for sdss 1707 + 58 is quite striking ( [ ba / fe ] = + 3.4 ) .
this star also exhibits a large excess of sr ( [ sr / fe ] = + 2.25 ) .
although the carbon abundance estimated from the ch band for this object is very uncertain , we include this object in our discussion of cemp stars as an example of a star that is likely affected by agb nucleosynthesis ( see [ sec : disc ] ) .
random errors in our analysis , which include the uncertainty of the equivalent - width measurements and in the adopted transition probabilities , are estimated to be @xmath66 , where @xmath54 is the standard deviation of derived abundances from individual lines and @xmath29 is the number of lines used in the analysis . when the number of lines are smaller than four , the @xmath54 of ( @xmath67 ) is adopted in the estimates .
typical random errors are 0.050.15 dex , depending on the number of lines used in the analysis .
we also estimate the error due to the uncertainty in equivalent - width measurements for the lines of sdss 003610 .
a typical error in equivalent width ( @xmath68 ) is estimated from the above formula . the typical value for the lines of sdss 003610 is obtained to be 3 m , assuming @xmath69 and @xmath70 .
we added this value to the measured equivalent widths and calculated the fe abundance using the same model atmosphere as used in the analysis .
the derived iron abundance is 0.10 dex higher than the original result .
this value is comparable with the @xmath67 of 0.12 dex obtained for sdss 003610 .
this result confirms that the random errors of the abundance measurements are primarily due to the uncertainties in the equivalent width measurements reflecting the quality of the spectrum , although the @xmath67 also includes the errors in the continuum placement and uncertainties of @xmath71 values .
the errors due to the uncertainty of the atmospheric parameters are estimated for lp 7067 and cs 29526110 .
table [ tab : err ] lists the sensitivity of the derived abundances ( @xmath72 values ) to the changes of parameters . for other objects in our program , the uncertainties are estimated by applying the data for the star of this pair with the closest atmospheric parameters to the object under consideration .
total uncertainties are obtained by adding these values , in quadrature , to the random errors , and are listed in table [ tab : abund ] .
the chemical abundances of lp 7067 are also determined using the updated ( newodf ) atlas grid @xcite , and the differences from those based on the @xcite model are given in table [ tab : err ] ( @xmath73 ) .
the abundances using the newodf model are lower by 0.050.14 dex .
the effect of the difference of model atmospheres on the derived abundances is systematic , and that on the abundance ratios is not significant .
we also applied the model including the excesses of @xmath6 elements , and confirmed the effect on the derived abundances is smaller than 0.01 dex .
further systematic errors could exist in our lte analysis based on one - dimensional ( 1d ) model atmospheres .
the non - lte correction for fe abundances derived from the lines might be the order of + 0.2 dex ( * ? ? ?
* ; * ? ? ?
* and references therein ) , although the values estimated are different between authors .
the direction of the correction for the 3d effect is opposite @xcite .
the most significant 3d effect would appear in the carbon abundances determined from ch molecular features , that could reach to @xmath74 dex in the most metal - poor cases @xcite . in order to obtain the corrections for these effect ,
3d analyses based on non - lte calculation are required .
the elemental abundances of lp 7067 (= cs 31062012 ) and cs 29526110 were studied by @xcite and @xcite .
the atmospheric parameters adopted by them for lp 7067 are @xmath13 = 6250 k , @xmath14 = 4.5 , [ fe / h ] = @xmath75 , and @xmath51 = 1.5 km s@xmath31 , which are quite similar to those in the present study .
although the previous studies are based on a spectrum covering only a blue range of wavelengths , and the spectral line set used in the previous analyses is different from that in the present study , the derived abundances of most elements agree within 0.1 dex .
the cr abundance shows the largest discrepancy , on the order of 0.17 dex , which is still within the 2 @xmath54 range of the measurement errors .
the effective temperature of cs 29526110 adopted by @xcite is 300 k lower than that of the present analysis .
the discrepancy of [ fe / h ] between the two measurements ( 0.32 dex ) is well explained by the difference in the adopted effective temperatures .
the [ c / fe ] values , after correction for the difference in effective temperatures , also agree within the measurement errors .
the abundance ratios of other elements ( [ x / fe ] ) are relatively insensitive to the effective temperature ( see table [ tab : err ] ) .
it is clear that the results for cr and ni from the two studies exhibit significant discrepancies .
however , the measurements for these elements are based on only one line for each ; the results might not be expected to be as reliable as those for other elements .
the [ ba / fe ] derived from the present work is 0.28 dex higher than that of @xcite .
this result is not explained by the differences of adopted atmospheric parameters . while the previous measurement is based on only the two very strong resonance lines , our present analysis added two red lines which are suitable for abundance determination , so the new measurement should be more reliable than the previous one .
table [ tab : comp ] provides a comparison of the atmospheric parameters and carbon abundances between the estimates obtained from the sdss spectra and the present measurements . for the sdss spectra , we begin by adopting the stellar parameters obtained by the sspp ( lee et al .
2007a , b ) . based on these
, we generate synthetic spectra for each star in the region between 42004400 .
the model atmospheres used are the newodf models of castelli & kurucz ( 2003 ) .
the synthetic spectra are generated using the turbospectrum synthesis code @xcite , which employs line broadening according to the prescription of @xcite and @xcite .
the atomic line data are taken mainly from the vald compilation ( as of 2002 ) ( kupka et al . 1999 ) , and updated from the literature , whenever possible .
the molecular species ch and cn are provided by b. plez ( plez & cohen 2005 ) .
we adopted the solar abundances by asplund , grevesse & sauval ( 2005a ) .
the synthetic spectra are generated with a initial resolving power @xmath76 , then were smoothed to the sdss resolution and rebinned to 1 pixels .
we find the best match to the region around the g band ( 4323 and 4325 ) by changing the carbon abundance of the synthetic spectra in order to minimize the discrepancy with the observed spectra .
we estimate that the errors in the derived [ c / fe ] arising from errors in the stellar parameters from the sspp is on the order of 0.35 dex .
the effective temperatures derived from the sdss spectra agree with the values adopted by the present work , based on colors , to within about 100 k. an exception is that for sdss 0126 + 06 , for which the sdss estimate is 370 k higher than the value adopted in the above analysis . for this object , larger
interstellar reddening is derived from the absorption than from the dust map that is adopted in the analysis ( [ sec : ew ] ) .
if the @xmath26 from the measurement is adopted , the @xmath13 is as high as 7000 k , and agrees with the estimate from the sdss spectrum .
in contrast to the agreement of effective temperatures , @xmath14 values estimated from sdss spectra are systematically lower than those determined by our analyses from subaru spectra .
although the results for sdss 0126 + 06 from the two estimates appear to agree well , a similar discrepancy would result if the same effective temperature is adopted in the estimate of gravity .
although further study or calibration to resolve the discrepancy is desired , it should be noted that the @xmath14 values from sdss spectra are already useful to estimate the evolutionary status of the targets ( i.e. , in order to distinguish giants and main - sequence turn - off stars ) . the abundance ratios [ fe / h ] and [ c / fe ] from the two measurements agree fairly well , as found in table [ tab : comp ] .
the discrepancy in [ fe / h ] for sdss 0126 + 06 ( 0.4 dex ) is well explained by the difference in the estimated @xmath13 .
no ch absorption feature is detected in the subaru spectrum of sdss 0817 + 26 .
a very weak ch band is found in the sdss spectrum of this object ( see figure 2 ) , and we have derived [ c / fe ] = + 1.19 @xmath77 , based on its strength .
@xcite estimated that the detection limit of the ch feature in sdss spectra can be as high as [ c / h]@xmath78 for stars with @xmath13 @xmath79 k. thus , sdss 0817 + 26 could also be mis - identified as a carbon - enhanced object in our sample selection , which was carried out before detailed investigations for sdss sample were made .
further calibrations , or strict estimates for the detection limits of ch absorption in sdss spectra as a function of @xmath13 , are desired for more efficiently selecting carbon - enhanced objects from sdss / segue spectra .
the present analysis has obtained elemental abundances for five new cemp stars selected from the sdss / segue surveys , as well as for two known cemp stars .
the remaining program star in our sample ( sdss 0817 + 26 ) is excluded from the following discussion , because no clear excess of carbon nor neutron - capture elements has been found in the high - resolution spectrum .
all seven cemp stars are classified as main - sequence turn - off stars , with @xmath13@xmath60 k. in order to better investigate the nature of such stars , we have compiled all known cemp stars from the literature having @xmath13 higher than 6000 k ( table [ tab : cemp_to ] ) . in this section , we discuss the abundance distributions of carbon and the neutron - capture elements for these cemp turn - off stars .
figure [ fig : cbafe ] shows the c and ba abundance ratios as a function of [ fe / h ] .
all objects but one exhibit very large excesses of c ( [ c / fe ] @xmath80 ) .
the exception is cs 29528041 , which has [ c / fe]= + 1.59 at [ fe / h]@xmath81 @xcite .
excluding this object , a clear correlation is found between [ c / fe ] and [ fe / h ] .
the correlation indicates that there exists a constant [ c / h ] among these stars .
figure [ fig : hist_ch ] is a histogram of the [ c / h ] values for the cemp turn - off stars .
the distribution is compared with that of the 31 cemp giants selected from @xcite that have @xmath82/l@xmath83 higher than 1.5 , where @xmath84 is the luminosity estimated from the stellar parameters assuming a constant stellar mass ( see aoki et al .
the ba - enhanced ( cemp - s ) stars are shown by open bars , while ba - normal ( cemp - no ) stars are shown by the hatched bars .
a glance at this figure shows that the [ c / h ] ratios of the cemp - s turn - off stars distribute within a narrow range around [ c / h]@xmath85 , with the exception of the one object mentioned above ( cs 29528041 ) .
the average and standard deviations of the [ c / h ] ratios for these two cemp - s samples are @xmath58[c / h]@xmath86 and @xmath54([c / h])@xmath87 for the turn - off stars , and @xmath58[c / h]@xmath88 and @xmath54([c / h])@xmath89 for the giants .
the two objects having the lowest [ c / h ] among turn - off stars ( cs 29528041 ) and giants ( cs 30322 - 023 ) are excluded in the calculation of these statistics .
the lower portion of the [ c / h ] distribution might be affected by a temperature - related selection bias for cemp stars identified on the basis of the ch molecular bands , which is significantly weaker in turn - off stars than in giants .
our previous investigation for the detection limit showed that the ch band of a turn - off star with @xmath13@xmath90 and [ c / h]@xmath91 has depths of 2% , which is a conservative detection limit in high - resolution spectra .
for the carbon - enhanced stars selected from the lower - resolution spectra , such as the sdss / segue sample , @xcite estimated that the selection of cemp stars from the ch band is complete for stars having [ c / h ] @xmath92 and @xmath93 for stars with @xmath13 = 6000 and 6500 k , respectively .
hence , in order to investigate the complete distribution of [ c / h ] for turn - off stars , abundance studies of candidate metal - poor stars that are selected regardless of their ch band strengths in medium - resolution spectra are required .
note , however , that the sample of cemp turn - off stars in table [ tab : cemp_to ] includes stars with @xmath13 as low as 6000 k , and that several stars were observed on programs that did not focus on carbon - enhanced stars ( e.g. * ? ? ?
the absence of stars with @xmath94 [ c / h ] @xmath95 in the sample of cemp turn - off stars suggests that such stars are rare compared with cemp stars with higher [ c / h ] values .
we note that we can not derive any conclusion for the lower [ c / h ] range ( [ c / h]@xmath96 ) .
there may well exist a number of cemp turn - off stars that have not yet been identified by the surveys to date .
this range is particularly important for studies of the ba - normal cemp stars , as discussed by @xcite .
@xcite showed that the [ c / h ] distribution for 54 cemp stars with ba excesses , including turn - off stars , subgiants , and giants , exhibits a peak in the range @xmath97 [ c / h ] @xmath98 , and a cut - off at [ c / h ] @xmath85 .
this was interpreted as an indication that ( 1 ) the [ c / h ] ratios produced by agb stars are almost constant at [ c / h ] @xmath99 , independent of metallicity , and ( 2 ) the carbon - enhanced material transferred from agb stars to the companion is directly observed , or is only slightly diluted through the evolution from turn - off stars to giants . in this study we confirmed the absence of objects having [ c / h ] @xmath100 , found that the average of the [ c / h ] values for cemp turn - off stars is higher than that of giants , and that their dispersion is smaller .
this result supports the interpretation of @xcite .
@xcite investigated the process of thermohaline mixing in main - sequence stars that received carbon - enhanced material from a companion agb star across a binary system .
they predicted that the accreted material quickly mixes throughout 90% of the star , and that the enhanced carbon is diluted in main - sequence stars as a result .
the c abundance is predicted to change only slightly after the receiving star evolves through first dredge - up .
this is not supported by the comparison of [ c / h ] distributions in figure [ fig : hist_ch ] , although the possible bias in the sample selection could slightly affect the comparison .
our present observational result suggests that the [ c / h ] ratios measured for turn - off stars represents the values produced by the donor agb stars , and the surface carbon abundance decreases in some cemp stars during their evolution after the first dredge - up .
it should be noted that the [ c / h ] values found in cemp turn - off stars ( i.e. [ c / h]@xmath85 ) agree well with predictions from agb models ( e.g. * ? ? ?
* ) , as discussed by @xcite .
this agreement supports the above interpretation .
the carbon excesses of cs 29528041 and cs 30322023 are exceptionally small among the sample of cemp turn - off stars and giant stars , respectively of this system is @xmath101 , and is another example of cemp turn - off stars having relatively low [ c / h ] values . ] .
it is noteworthy that these two stars both exhibit very large excesses of nitrogen the [ n / fe ] of cs 29528041 is + 3.0 @xcite and that of cs 30322023 is + 2.8 @xcite .
their [ ( c+n)/h ] values are @xmath102 and @xmath103 , respectively , which are not by far lower than those of other stars , although the nitrogen abundances are not determined for several stars including our sdss sample . cs 30322023 is a highly evolved giant , and possibly presently in the agb stage ( according to masseron et al . ) ; its surface composition could have been altered significantly during its evolution .
the observed nitrogen excess of cs 29528041 should instead be a direct result of the nucleosynthesis in a donor agb star . as discussed by @xcite , cs 29528041
might have been polluted by an intermediate - mass agb star in which nitrogen is enriched by the hot bottom burning process . based on their
observed [ c / h ] distribution , we have interpreted the surface abundances of cemp turn - off stars to represent the yields of the agb donors .
if this is indeed the case , [ ba / h ] can be used as an indicator of the s - process efficiency in the donor stars .
figure [ fig : hist_bah ] depicts the distribution of [ ba / h ] for the cemp turn - off stars and giants that show excesses of ba ( [ ba / fe ] @xmath104 ) . the [ ba / h ] of the turn - off stars exhibits a wider distribution than was observed for their [ c / h ] .
the average and standard deviation of the [ ba / h ] for the 16 cemp turn - off stars discussed in the context of their [ c / h ] are @xmath58[ba / h]@xmath105 and @xmath54([ba / h ] ) = 0.68 dex , respectively .
the wider distribution of [ ba / h ] of these stars , as compared to [ c / h ] , implies that the ba abundances produced by agb stars are likely to have a significantly larger intrinsic dispersion .
the abundance ratios currently observed are those of the material transferred from the donor agb stars , which would be dependent on the evolutionary phase in which the mass transfer occurred .
the abundances of carbon and ba are expected to increase with increasing the number of thermal pulses .
detailed comparisons with model predictions for surface abundances after each thermal pulse would be required .
in such comparisons , a strong constraint is the almost constant c / h ratios found in most cemp turn - off stars .
figure [ fig : srbapb ] shows the [ sr / ba ] and [ ba / pb ] ratios of cemp turn - off stars as a function of [ ba / h ] . here , the three elements sr , ba , and pb are regarded as representing the yields at the three abundance peaks of the s - process , corresponding to the neutron magic numbers 50 , 82 , and 126 .
the [ sr / ba ] ratios are distributed over a narrow range ( a standard deviation of 0.34 dex ) around [ sr / ba ] @xmath106 .
however , they also exhibit a statistically significant anti - correlation with [ ba / h ] the null hypothesis that no correlation exists between [ sr / ba ] and [ ba / h ] is rejected by a simple t - test at the 98% confidence level .
that is , the production efficiency of the heavy neutron - capture elements , with respect to the light ones , apparently increases slightly with the total production efficiency of neutron - capture elements .
it should be noted that the [ sr / ba ] ratios in these stars are lower than the prediction from models of the s - process in agb stars by @xcite : the [ ls / hs](@xmath107[hs / ls ] ) values of their calculations , where ls and hs mean the elements in the first and second abundance peaks of the s - process , range between 0 to @xmath108 , depending on stellar mass and choice of @xmath109c pocket , for the metallicity range of [ fe / h]@xmath110 .
the observations suggest more efficient production of ba ( hs ) , or less efficient production of sr ( ls ) , than model predictions .
a similar correlation was found from the abundance measurements for carbon - rich post - agb stars by @xcite .
they reported a positive correlation between the enhancement of s - process elements ( [ s / fe ] , which is the mean of the abundance ratios of several s - process elements ) and the abundance ratios of heavy to light neutron - capture elements ( [ hs / ls ] ) .
we note that the efficiency of the neutron - capture nucleosynthesis is represented by [ s / fe ] in @xcite , while it is evaluated by [ ba / h ] in our investigation .
however , a correlation between [ sr / ba ] and [ ba / fe ] is also found in our sample at a similar confidence level as that between [ sr / ba ] and [ ba / h ] .
the lower panel of figure [ fig : srbapb ] suggests that there may exist a correlation for [ ba / pb ] with [ ba / h ] , but this essentially depends on only one object ( he 00242523 ) , with [ ba / h ] @xmath111 and [ ba / pb ] @xmath112 .
it should be kept in mind that the detection of pb lines is much more difficult than those of sr and ba , which could result in a lack of objects having high [ ba / pb ] and low [ ba / h ] .
the [ ba / pb ] ratios observed are higher than the prediction from models of the s - process by @xcite , as has been argued by previous studies ( e.g. * ? ? ? * ) .
further studies of pb abundances , as well as modeling of the s - process to explain the discrepancy between the observations and predictions , are clearly required in order to better understand the overall neutron - capture nucleosynthesis process in agb stars . as discussed in
[ sec : obs ] , sdss 1707 + 58 exhibited a rapid variation of radial velocities on february 10 , 2007 , suggesting that this object belongs to a close binary system .
however , it is interesting that no clear variation was found in the february 11 spectra , where the radial velocity is close to the middle of the three spectra obtained on february 10 .
this result could well imply a large eccentricity of the binary system , although the radial velocity measurements are still too sparse to confidently derive orbital parameters .
further radial velocity monitoring of this object is clearly desirable .
the observed excesses of neutron - capture elements in this object are very large ( e.g. , [ ba / fe ] = + 3.4 ) . indeed , the [ ba / h ] value of this object is the highest among the cemp stars shown in figure [ fig : hist_bah ] . the [ sr / ba ] ratio ( [ sr / ba ] @xmath113 ) is a typical value found for ba - enhanced cemp stars ( figure [ fig : srbapb ] ) , as compared to that of r - process - enhanced stars ( [ sr / ba]@xmath114 ; e.g. sneden et al .
2003 ) , suggesting that the neutron - capture elements of this star originated from operation of the s - process in an agb donor star .
@xcite studied the cemp turn - off star he 00242523 ( [ fe / h]@xmath115 ) , which they showed to be a short - period spectroscopic binary .
this star has a orbital period of 3.4 days , with a very small eccentricity ( 0.01 ) .
@xcite also reported a candidate cemp turn - off close binary , cs 22958042 , for which a significant radial velocity change was found based on two exposures taken during a single observing run .
sdss 1707 + 58 is possibly another example of cemp stars belonging to close binary systems .
@xcite proposed that he 00242523 underwent a past common - envelope phase with its companion that has become an agb star .
a similar past history may be applicable to sdss 1707 + 58 .
the stellar parameters of sdss 1707 + 58 ( @xmath13 = 6700 k and @xmath14 = 4.2 ) are almost the same as those of he 00242523 ( @xmath13 = 6625 k and @xmath14 = 4.3 ) . on the other hand ,
the abundance ratios of the neutron - capture elements are different between these two stars while he 0024 - 2523 exhibits a very large excess of pb ( [ pb / fe ] = + 3.3 ) and moderate over - abundances of ba ( [ ba / fe ] = + 1.46 ) , sdss 1707 + 58 has a very large overabundance of ba ( [ ba / fe ] = + 3.40 ) .
our derived upper limit on the pb abundance ratio of sdss 1707 + 58 is still quite high ( [ pb / fe ] @xmath116 ) , so it does not constrain this comparison at present .
we note that the possible close binary cs 22958042 exhibits no excess of neutron - capture elements , but shows a large over - abundance of na , as found also for sdss 1707 + 58 .
thus , large variations of chemical abundances are found even in these three ( candidate ) close binary systems .
further abundance measurements , as well as radial velocity monitoring , for sdss 1707 + 58 are clearly desired to understand the evolution of such binary systems .
table [ tab : kinematics ] provides kinematic data for our sdss targets
. distances ( expected to be accurate to on the order of 10 - 15% ) for these stars are estimated using the methods described by @xcite , under the assumption that they are main - sequence stars , as suggested by the surface gravities determined in this study .
proper motions are obtained based on the re - calibrated usno - b2 catalog , as described by munn et al .
( 2004 ) , and are expected to be accurate to on the order of 3 mas / year .
the radial velocities are taken from our high - resolution estimates listed in table 1 , with the exception of sdss 1707 + 58 , where we adopt the value measured from the sdss spectrum .
the space motions , errors in the space motions , and the other derived quantities listed in table [ tab : kinematics ] are obtained following the procedures of @xcite .
all six of the sdss cemp stars exhibit a derived @xmath117 ( maximum distance from the galactic center achieved during the course of their orbits ) in excess of 10 kpc from the galactic center .
four of the stars exhibit @xmath118 ( maximum height above or below the galactic plane achieved during the course of their orbits ) values larger than 10 kpc , or significant retrograde motions , indicating that these stars may belong to the outer - halo population of our galaxy , according to the criteria of @xcite .
this is also reminiscent of the apparent excess of cemp stars with increasing distance from the plane reported by frebel et al .
although the sample size is too small to derive any firm conclusions , the fraction of outer - halo stars among the cemp stars appears to be quite high compared with the fraction found for metal - poor stars ( [ fe / h ] @xmath119 ) studied by @xcite .
a comparison of the fraction of cemp stars associated with the inner- and outer - halo populations , at a given metallicity , potentially constrains the initial mass function of early - generation stars , as discussed recently by @xcite and @xcite , and is clearly of interest for additional study , in particular given the very large samples of cemp stars identified in sdss / segue .
chemical compositions of seven cemp turn - off stars are determined .
six stars among them exhibit a large excess of ba , signature of a contribution by the nucleosynthesis in an agb star .
the distribution of carbon abundances in these stars suggest that the surface of such stars preserves the material transferred from the agb star that was the erstwhile primary star in a binary system . if this is the case , the relatively wide distribution of ba abundances ( [ ba / h ] ) indicates a diversity of the efficiency of the s - process in metal - poor agb stars .
further studies to identify the physical mechanism that produces such diversity are clearly desired .
the present study is the first application of high - resolution spectroscopy to candidate cemp stars from the sdss and segue sample .
comparisons of our results on stellar parameters and chemical abundances with the estimates from the sdss spectra confirmed that the selection of metal - poor stars works well in general .
the sdss / segue survey is providing a large sample of candidate metal - poor stars .
high - resolution spectroscopy for such stars in the near future will reveal the chemical abundance trends in the lowest metallicity range , as well as be useful for exploring the possible dependence of their chemical properties on their derived kinematics . funding for the sdss and sdss - ii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , the u.s .
department of energy , the national aeronautics and space administration , the japanese monbukagakusho , the max planck society , and the higher education funding council for england .
the sdss web site is http://www.sdss.org/. the sdss is managed by the astrophysical research consortium for the participating institutions .
the participating institutions are the american museum of natural history , astrophysical institute potsdam , university of basel , university of cambridge , case western reserve university , university of chicago , drexel university , fermilab , the institute for advanced study , the japan participation group , johns hopkins university , the joint institute for nuclear astrophysics , the kavli institute for particle astrophysics and cosmology , the korean scientist group , the chinese academy of sciences ( lamost ) , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , ohio state university , university of pittsburgh , university of portsmouth , princeton university , the united states naval observatory , and the university of washington .
w. a. is supported by a grant - in - aid for science research from jsps ( grant 18104003 ) .
t. c. b. , b. m. , and t. s. acknowledge support by the us national science foundation under grants ast 04 - 06784 and ast 07 - 07776 , as well as from grant phy 02 - 16783 ; physics frontier center / joint institute for nuclear astrophysics ( jina ) .
d. c. is grateful to jina for support of her long - term visitor status at michigan state university , where the kinematical analysis took place . j.e.n . acknowledges support from the australian research council under grant dp0663562 .
lccccc & & 20 & 16500 & 14 sep 2006 ( 2453992.87 ) & @xmath120 +
cs 29526110 & & 40 & 8000 & 10 feb 2007 ( 2454141.72 ) & @xmath121 + sdss 003610 & sdss j003602.17 - 104336.3 & 120 & 5350 & 14 sep 2006 ( 2453992.95 ) & @xmath122 + sdss 0126 + 06 & sdss j012617.95 + 060724.8 & 120 & 5150 & 14 sep 2006 ( 2453993.04 ) & @xmath123 + sdss 0817 + 26 & sdss j081754.93 + 264103.8 & 94 & 1050 & 10 feb 2007 ( 2454141.76 ) & @xmath124 + sdss 0924 + 40 & sdss j092401.85 + 405928.7 & 160 & 5250 & 10 feb 2007 ( 2454141.87 ) & @xmath125 + sdss 1707 + 58 & sdss j170733.93 + 585059.7 & 117 & 2300 & 10 feb 2007 ( 2454142.10 ) & ... + sdss 2047 + 00 & sdss j204728.84 + 001553.8 & 160 & 3000 & 14
sep 2006 ( 2453992.75 ) & @xmath126 + lcccccccccc & 12.098 & 0.001 & 0.467 & 0.003 & 10.771 & 0.017 & 0.022 & 0.007 & 6206 & 5933 + cs 29526110 & 13.352 & 0.004 & 0.356 & 0.007 & 12.367 & 0.023 & 0.033 & 0.025 & 6800 & 6500 + sdss 003610 & 15.540 & 0.004 & 0.320 & 0.006 & 14.388 & 0.080 & 0.027 & 0.010 & 6500 & 6600 + sdss
0126 + 06 & 15.525 & 0.004 & ... & ... & 14.468 & 0.077 & 0.029 & 0.094 & 6600 & ... + sdss 0817 + 26 & 15.990 & 0.004 & 0.43 & 0.006 & 14.707 & 0.077 & 0.024 & 0.028 & 6302 & 6097 + sdss 0924 + 40 & 15.480 & 0.004 & 0.42 & 0.006 & 14.165 & 0.053 & 0.014 & & 6184 & 6097 + sdss 1707 + 58 & 15.810 & 0.004 & 0.33 & 0.006 & 14.776 & 0.108 & 0.035 & 0.091 & 6700 & 6600 + sdss 2047 + 00 & 16.009 & 0.004 & 0.390 & 0.006 & 14.880 & 0.120 & 0.088 & 0.126 & 6800 & 6600 + lccccc 10 feb.2007 , 14:19 & 2454142.097 & 40 & 35.4 & 27.5 & 40.7 + 10 feb.2007 , 15:00 & 2454142.125 & 40 & 16.9 & 35.5 & 22.2 + 10 feb.2007 , 15:41 & 2454142.153 & 37 & @xmath1272.3 & 23.3 & 3.0 + 11 feb.2007 , 15:34 & 2454143.148 & 20 & 15.6 & 16.9 & 20.8 + 11 feb.2007 , 15:49 & 2454143.159 & 20 & 16.5 & 20.0 & 21.7 + lccccccccccc na i & 5889.95 & 0.00 & 0.10 & 164.2 & 98.6 & 116.1 & 75.3 & ... & 147.3 & 313.8 & 69.5 + na i & 5895.92 & 0.00 & -0.20 & 135.5 & 77.0 & 96.4 & 55.5 & ... & 127.3 & 277.6 & ... + mg i & 4057.50 & 4.35 & -0.89 & 13.0 & ... & ... & ... & ... & ... & ... & ... + mg i & 4571.10 & 0.00 & -5.69 & 1.9 & ... & ... & ... & ... & ... & ... & ... + mg i & 5172.69 & 2.71 & -0.38 & 129.6 & 117.6 & 99.8 & 84.9 & 71.8 & 120.2 & 184.2 & 124.9 + mg i & 5183.60 & 2.72 & -0.16 & 147.5 & 135.2 & 122.0 & 101.6 & 77.3 & 141.4 & 196.5 & 147.3 + mg i & 5528.40 & 4.35 & -0.49 & 33.0 & 28.8 & 22.7 & 11.2 & 22.0 & 28.7 & ... & 40.7 + ca i & 4226.73 & 0.00 & 0.24 & ... & ... & ... & ... & ... & ... & 122.1 & ... + ca i & 4435.69 & 1.89 & -0.52 & 13.8 & 17.9 & 7.1 & ... & ... & ... & ... & 27.7 + ca i & 4454.78 & 1.90 & 0.26 & 35.8 & 41.0 & 38.0 & ... & ... & 41.2 & ... & 49.3 + ca i & 4455.89 & 1.90 & -0.53 & 10.3 & 12.4 & 8.7 & ... & ... & 12.9 & ... & 13.0 + ca i & 5265.56 & 2.52 & -0.11 & 6.3 & ... & 6.8 & ... & ... & 7.3 & ... & ... + ca i & 5588.76 & 2.53 & 0.36 & 12.5 & ... & 13.0 & ... & ... & 12.8 & ... & ... + ca i & 5594.47 & 2.52 & 0.10 & 13.5 & 17.4 & 10.6 & ... & ... & ... & ... & 13.0 + ca
i & 5598.49 & 2.52 & -0.09 & 6.3 & ... & 7.5 & ... & ... & ... & ... & 19.7 + ca i & 5857.45 & 2.93 & 0.24 & 5.9 & ... & ... & ... & ... & ... & ... & 18.3 + ca i & 6102.72 & 1.88 & -0.77 & 4.3 & ... & ... & ... & ... & ... & ... & 15.7 + ca i & 6122.22 & 1.89 & -0.32 & 12.7 & 16.5 & ... & ... & ... & ... & ... & 21.1 + ca i & 6162.17 & 1.90 & -0.09 & 19.2 & 22.3 & 21.5 & ... & ... & ... & ... & ... + ca i & 6439.07 & 2.53 & 0.39 & 16.8 & 19.1 & ... & ... & ... & 17.7 & ... & 26.2 + ca i & 6462.57 & 2.52 & 0.26 & 13.0 & 13.1 & 13.1 & ... & ... & 16.5 & ... & ... + ti i & 4981.73 & 0.85 & 0.56 & 9.9 & 9.1 & ... & ... & ... & ... & ... & ... + ti i & 4991.07 & 0.84 & 0.44 & 6.8 & ... & ... & ... & ... & ... & ... & ... + ti i & 4999.50 & 0.83 & 0.31 & 8.2 & 11.9 & ... & ... & ... & ... & ... & ... + ti i & 5007.21 & 0.82 & 0.17 & 8.6 & ... & ... & ... & ... & ... & ... & 14.4 + ti i & 5064.65 & 0.05 & -0.94 & 4.6 & ... & ... & ... & ... & ... & ... & ... + cr i & 4652.16 & 1.00 & -1.03 & 4.1 & ...
& ... & ... & ... & ... & ... & ... + cr i & 5206.04 & 0.94 & 0.02 & 21.0 & ... & 14.6 & ... & ... & 13.4 & ... & 26.2 + cr i & 5208.44 & 0.94 & 0.16 & 23.5 & 31.9 & 22.5 & ... & ... & 25.9 & ... & 37.4 + fe i & 4063.59 & 1.56 & 0.06 & 88.2 & ... & ... & ... & ... & ... & ... & ... + fe i & 4071.74 & 1.61 & -0.02 & 82.1 & 90.4 & 75.8 & ... & 56.5 & 74.8 & ... & 74.5 + fe i & 4107.49 & 2.83 & -0.88 & 10.6 & 20.7 & ... & ... & ... & ... & ... & ... + fe i & 4143.41 & 3.05 & -0.20 & 17.9 & 19.3 & 19.2 & ... & ... & 27.0 & ... & 23.8 + fe i & 4143.87 & 1.56 & -0.51 & 61.5 & 68.6 & 59.2 & 26.9 & ... & 62.1 & ... & 59.6 + fe i & 4202.03 & 1.49 & -0.71 & ... & ... & ... & ... & ... & ... & 43.3 & ... + fe i & 4271.76 & 1.49 & -0.16 & ... & ... & ... & ... & ... & ... & 38.1 & ... + fe i & 4307.90 & 1.56 & -0.07 & ... & ... & ... & ... & ... & ... & 62.1 & ... + fe i & 4325.76 & 1.61 & 0.01 & ... & ... & ... & ... & ... & ... & 65.0 & ... + fe i & 4383.54 & 1.49 & 0.20 & 98.9 & 102.5 & 87.1 & 69.4 & ... & 102.2 & 91.0 & 99.6 + fe i & 4404.75 & 1.56 & -0.14 & 77.6 & 84.7 & 65.4 & 56.4 & 41.5 & 52.6 & 66.7 & 76.9 +
fe i & 4415.12 & 1.61 & -0.62 & 67.5 & 65.4 & 59.5 & 22.7 & 21.0 & 85.3 & 37.2 & 65.9 + fe i & 4427.31 & 0.05 & -2.92 & 36.0 & 19.7 & 29.3 & ... & ... & 68.0 & ... & 39.7 + fe i & 4442.34 & 2.20 & -1.25 & ... & 19.4 & 22.3 & ... & ... & ... & ... & 19.5 + fe i & 4447.72 & 2.22 & -1.34 & ... & 7.1 & ... & ... & ... & ... & ... & ... + fe i & 4459.12 & 2.18 & -1.28 & 14.2 & 15.6 & 11.8 & ... & ... & 14.9 & ... & 22.1 + fe i & 4461.65 & 0.09 & -3.21 & 13.2 & ... & ... & ... & ... & 12.5 & ... & 22.1 + fe i & 4466.55 & 2.83 & -0.60 & 14.5 & 22.9 & 17.4 & 17.8 & ... & 12.6 & ... & ... + fe i & 4476.02 & 2.85 & -0.82 & 13.3 & ... & 7.5 & ... & ... & ... & ... & 11.6 + fe i & 4494.56 & 2.20 & -1.14 & 15.3 & 9.0 & 11.9 & ... & ... & 15.2 & ... & ... + fe i & 4528.61 & 2.18 & -0.82 & 29.6 & 42.7 & 22.6 & ... & ... & 29.8 & ... & 28.7 + fe i & 4531.15 & 1.49 & -2.15 & 10.7 & ... & ... & ... & ... & ... & ... & ... + fe i & 4592.65 & 1.56 & -2.45 & 4.9 & ... & ... & ... & ... & ... & ... & ... + fe i & 4602.94 & 1.49 & -2.21 & 6.8 & ... & ... & ... & ... & ... & ... & 11.5 + fe i & 4871.32 & 2.87 & -0.36 & ... & 17.2 & ... & ... & ... & ... & ... & ... + fe i & 4872.14 & 2.88 & -0.57 & 13.3 & 12.9 & 9.6 & ... & ... & ... & ... & 16.1 + fe i & 4890.75 & 2.88 & -0.39 & 22.2 & 28.9 & 11.1 & ... & ... & 19.5 & ... & 28.4 + fe i & 4891.49 & 2.85 & -0.11 & 31.8 & 34.5 & 21.9 & ... & ... & 36.5 & ... & 31.5 + fe i & 4903.31 & 2.88 & -0.93 & 5.2 & 9.4 & ... & ... & ... & ... & ... & 17.4 + fe i & 4918.99 & 2.87 & -0.34 & 19.7 & 24.4 & ... & ... & ... & 21.1 & ... & 28.8 + fe i & 4920.50 & 2.83 & 0.07 & 40.3 & 54.7 & 31.8 & ... & ... & 40.2 & ... & 46.5 + fe i & 4957.30 & 2.85 & -0.41 & 22.7 & ... & ... & ... & ... & 21.8 & ... & ... + fe i & 4957.60 & 2.81 & 0.23 & 49.7 & 61.9 & 42.2 & ... & 23.8 & 42.2 & ... & ... + fe i & 4966.09 & 3.33 & -0.87 & 5.1 & 6.2 & ... & ... & ...
& ... & ... & ... + fe i & 4994.13 & 0.92 & -2.96 & 4.2 & ... & ... & ... & ... & ... & ... & ... + fe i & 5006.12 & 2.83 & -0.61 & 13.5 & 12.9 & ... & ... & ... & 14.2 & ... & ... + fe i & 5012.07 & 0.86 & -2.64 & 10.7 & ... & 8.6 & ... & ... & ... & ... & 19.9 + fe i & 5041.76 & 1.49 & -2.20 & ... & 28.8 & 11.4 & ... & ... & ... & ... & ... + fe i & 5049.82 & 2.28 & -1.34 & 10.0 & 9.9 & ... & ... & ... & 14.8 & ... & ... + fe i & 5051.63 & 0.92 & -2.80 & 6.5 & ... & ... & ... & ... & 15.2 & ... & 10.0 + fe i & 5151.91 & 1.01 & -3.32 & 7.7 & ... & ... & ... & ... & ... & ... & ... + fe i & 5171.60 & 1.49 & -1.79 & 16.8 & 14.3 & ... & ... & ... & 13.8 & ... & 23.2 + fe i & 5191.46 & 3.04 & -0.55 & 10.7 & 26.5 & 9.0 & ... & ... & 13.1 & ... & 17.2 + fe i & 5192.34 & 3.00 & -0.42 & 14.4 & 15.7 & 11.0 & ... & ... & 16.7 & ... & 20.8 + fe i & 5194.94 & 1.56 & -2.09 & 9.5 & ... & 14.6 & ... & ... & 8.9 & ... & ... + fe i & 5198.71 & 2.22 & -2.13 & 2.2 & ... & ... & ... & ... & ... & ... & ... + fe i & 5202.34 & 2.18 & -1.84 & 4.4 & ... & ... & ... & ... & ... & ... & ... + fe i & 5216.27 & 1.61 & -2.15 & 7.5 & ... & ... & ... & ... & ... & ... & ... + fe i & 5232.94 & 2.94 & -0.06 & 28.7 & 32.6 & 19.5 & ... & ... & 29.6 & ... & 39.2 + fe i & 5266.56 & 3.00 & -0.39 & 14.9 & 20.5 & 14.3 & ... & ... & 15.2 & ... & 16.9 + fe i & 5269.54 & 0.86 & -1.32 & 67.4 & 64.6 & 57.4 & 17.7 & 42.4 & 58.0 & 39.4 & 74.5 + fe i & 5270.36 & 1.61 & -1.34 & 36.0 & 40.9 & 29.8 & ... & ... & 37.3 & ... & 57.5 + fe i & 5281.79 & 3.04 & -0.83 & 4.7 & ... & ... & ... & ... & ... & ... & ... + fe i & 5324.18 & 3.21 & -0.10 & 16.4 & 23.8 & 14.4 & ... & ... & 17.7 & ... & 23.8 + fe i & 5328.04 & 0.92 & -1.47 & 57.4 & 57.9 & 45.9 & 13.4 & ... & 50.4 & 40.6 & 67.7 + fe i & 5328.53 & 1.56 & -1.85 & 14.4 & 13.1 & 11.5 & ... & ... & 10.2 & ... & 19.7 + fe i & 5339.93 & 3.27 & -0.65 & 7.9 & ... & ... & ... & ... & ... & ... & ... + fe i & 5455.61 & 1.01 & -2.10 & 27.5 & 30.7 & 22.1 & ... & ... & ... & ... & 34.9 + fe i & 5497.52 & 1.01 & -2.85 & 4.9 & ... & ... & ... & ... & ... & ... & ... + fe i & 5506.78 & 0.99 & -2.80 & 6.5 & ... & ... & ... & ... & ... & ... & ... + fe i & 5569.62 & 3.42 & -0.54 & 6.2 & 10.7 & ... & ... & ... & 11.1 & ... & ... + fe i & 5572.84 & 3.40 & -0.28 & 8.7 & 13.9 & ... & ... & ... & 13.0 & ... & 15.8 + fe i & 5586.75 & 3.37 & -0.10 & 12.6 & 17.2 & 8.5 & ... & ... & 14.4 & ... & 17.5 + fe i & 5615.64 & 3.33 & 0.05 & 18.2 & 23.5 & 14.2 & ... & ... & 20.9 & ... & 24.1 + fe i & 6136.61 & 2.45 & -1.40 & 6.9 & ... & ... & ... & ... & ... & ... & ... + fe i & 6137.69 & 2.59 & -1.40 & 6.9 & ... & ... & ... & ... & ... & ... & ... + fe i & 6191.56 & 2.43 & -1.42 & 6.4 & ... & ... & ... & ... & ... & ... & ... + fe i & 6230.72 & 2.56 & -1.28 & 6.2 & ... & ... & ... & ... & 9.4 & ... & ... + fe i & 6393.60 & 2.43 & -1.43 & 5.8 & ... & ... & ... & ... & ... & ... & ... + fe i & 6677.99 & 2.69 & -1.42 & 4.8 & ... & ... & ... & ... & ... & ... & ... + co i & 4121.32 & 0.92 & -0.32 & 10.6 & ... & 7.7 & ... & ... & ... & ... & ... + ni i & 5476.91 & 1.83 & -0.89 & 11.9 & 21.0 & 5.7 & ... & ... & 11.9 & ... & 16.5 + zn i & 4722.15 & 0.00 & -0.37 & 3.0 & ... & ... & ... & ... & ... & ... & ... + sc ii & 4320.75 & 0.61 & -0.25 & 11.5 & ... & ... & ... & ... & 19.4 & ... & ... + ti ii & 4395.00 & 1.08 & -0.51 & ... & ... & ... & ... & ... & ... & 42.3 & ... + ti
ii & 4443.77 & 1.08 & -0.70 & 35.8 & 45.6 & 33.3 & 18.1 & ... & 42.2 & ... & 46.1 + ti ii & 4444.54 & 1.12 & -2.21 & ... & ... & 5.1 & ... & ... & ... & ... & ... + ti ii & 4450.50 & 1.08 & -1.51 & 6.9 & 12.2 & ... & ... & ... & ... & ... & 14.3 + ti ii & 4464.46 & 1.16 & -2.08 & 5.4 & ... & ... & ... & ... & 6.6 & ... & ... + ti ii & 4468.52 & 1.13 & -0.60 & 39.4 & 52.7 & 37.6 & 29.3 & ... & 38.1 & ... & 57.5 + ti ii & 4501.27 & 1.12 & -0.76 & 29.8 & 44.4 & 29.2 & 20.1 & ... & 29.8 & ... & 47.0 +
ti ii & 4533.97 & 1.24 & -0.77 & 37.1 & 47.6 & 34.6 & ... & ... & 37.7 & 33.2 & 57.0 + ti ii & 4563.77 & 1.22 & -0.96 & 24.5 & 29.7 & 29.8 & 13.9 & ... & 24.3 & 31.2 & 27.6 + ti ii & 4571.96 & 1.57 & -0.53 & 29.6 & ... & 33.9 & ... & ... & 41.8 & 33.0 & ... + ti ii & 4589.92 & 1.24 & -1.79 & ... & ... & ... & ... & ... & ... & ... & 11.1 + ti ii & 4805.09 & 2.06 & -1.10 & ... & 9.6 & ... & ... & ... & 8.7 & ... & ... + ti ii & 5226.53 & 1.57 & -1.30 & 6.5 & ... & 9.5 & ... & ... & 8.3 & ... & 16.0 + fe ii & 4491.40 & 2.86 & -2.70 & 3.9 & ... & ... & ... & ... & ... & ... & 10.6 + fe ii & 4508.28 & 2.86 & -2.58 & 7.4 & 13.4 & ... & ... & ... & ... & ... & 12.8 + fe ii & 4515.34 & 2.84 & -2.48 & ... & 9.9 & ... & ... & ... & ... & ... & ... + fe ii & 4520.23 & 2.81 & -2.60 & 4.5 & 9.1 & ... & ... & ... & ... & ... & ... + fe ii & 4522.63 & 2.84 & -2.03 & 10.0 & ... & 11.9 & ... & ... & 17.5 & ... & 17.9 + fe ii & 4555.89 & 2.83 & -2.29 & ... & 20.3 & ... & ... & ... & ... & ... & 17.1 + fe ii & 4583.83 & 2.81 & -2.02 & 17.1 & 42.7 & ... & ... & ... & 20.1 & 22.1 & 26.8 + fe ii & 4923.93 & 2.89 & -1.32 & 34.2 & 63.4 & 32.6 & 15.7 & ... & ... & 23.9 & 46.8 + fe ii & 5018.45 & 2.89 & -1.22 & 41.4 & 68.5 & 43.1 & 21.7 & ... & 45.4 & 47.3 & 57.2 + fe ii & 5197.56 & 3.23 & -2.10 & 3.5 & 14.1 & 8.1 & ... & ... & ... & ... & ... + fe ii & 5234.62 & 3.22 & -2.27 & 4.8 & 12.9 & ... & ... & ... & ... & ... & ... + fe ii & 5276.00 & 3.20 & -1.94 & 7.4 & 18.0 & 10.9 & ... & ... & ... & ... & 17.7 + fe ii & 5316.62 & 3.15 & -1.85 & 9.5 & 29.3 & ... & ... & ... & 10.6 & ... & 17.0 + sr ii & 4077.71 & 0.00 & 0.15 & 74.1 & 129.9 & 60.4 & 102.7 & ... & 91.5 & 287.1 & 104.5 + sr ii & 4215.52 & 0.00 & -0.18 & 63.5 & 111.0 & 43.9 & 92.6 & 43.7 & 82.8 & 160.9 & 99.0 + y ii & 4204.69 & 0.00 & -1.76 & 61.5 & ... & 55.5 & 25.2 & ... & 101.2 & ... & ... + y ii & 4883.68 & 1.08 & 0.07 & ... & 15.1 & ... & ... & ... & ... & ... & 12.7 + y ii & 4900.12 & 1.03 & -0.09 & ... & 47.1 & ... & ... & ... & ... & 35.6 & ... + zr ii & 4048.67 & 0.80 & -0.48 & ... & 14.8 & ... & ... & ... & ... & ... & ... + zr ii & 4149.20 & 0.80 & -0.03 & ... & 31.9 & ... & 20.0 & ... & 32.6 & ... & 24.2 + zr ii & 4150.97 & 0.80 & -1.08 & 4.2 & ... & ... & ... & ... & ... & ... & ... + ba ii & 4554.03 & 0.00 & 0.16 & 152.3 & ... & ... & 140.8 & 51.2 & 88.9 & 282.2 & 108.1 + ba ii & 4934.09 & 0.00 & -0.16 & 147.0 & 169.5 & 34.1 & 127.9 & ... & 126.9 & 248.2 & 102.8 + ba ii & 5853.70 & 0.60 & -1.01 & 48.1 & 77.8 & 9.0 & 38.6 & ... & 37.8 & 77.6 & 30.2 + ba ii & 6141.70 & 0.70 & -0.07 & 92.4 & 123.4 & 12.7 & 85.3 & ... & 78.4 & 192.0 & 74.4 + ba ii & 6496.91 & 0.60 & -0.38 & 81.8 & 115.8 & 8.6 & 88.6 & ... & 68.3 & 150.3 & 63.5 + la ii & 4086.71 & 0.00 & -0.07 & 23.1 & 41.5 & ... & 16.5 & ... & ... & ... & 8.0 + la ii & 4123.22 & 0.32 & 0.13 & 21.4 & 43.2 & ... & 16.7 & ... & ... & ... & 12.7 + na i & 5889.95 & 0.00 & 0.10 & 27.6 & 90.6 & 38.9 & 259.0 & 252.0 & 100.0 & ... & 319.0 + llll & 2,452,146 & @xmath128 & sdss + & 2,452,162 & @xmath129 & sdss + & 2,453,993 & @xmath122 & this work + sdss 0126 + 06 & 2,453,712 & @xmath130 & sdss + & 2,453,713 & @xmath131 & sdss + & 2,453,993 & @xmath123 & this work + sdss 0817 + 26 & 2,452,709 & @xmath132 & sdss + & 2,454,142 & @xmath133 & this work + sdss 0924 + 40 & 2,452,708 & @xmath134 & sdss + & 2,452,636 & @xmath135 & sdss + & 2,454,142 & @xmath125 & this work + sdss 1707 + 58 & 2,451,703 & @xmath136 & sdss + sdss 2047 + 00 & 2,452,466 & @xmath137 & sdss + & 2,452,524 & @xmath138 & sdss + & 2,452,932 & @xmath139 & sdss + & 2,453,993 & @xmath126 & this work + lcccccccc & 6200 & 150 & 4.3 & 0.3 & @xmath140 & 0.3 & 1.4 & 0.3 + cs 29526110 & 6800 & 150 & 4.1 & 0.3 & @xmath141 & 0.3 & 2.1 & 0.3 + sdss 003610 & 6500 & 200 & 4.5 & 0.3 & @xmath140 & 0.3 & 1.5 & 0.3 + sdss 0126 + 06 & 6600 & 200 & 4.1 & 0.3 & @xmath142 & 0.3 & 2.0 & 0.5 + sdss
0817 + 26 & 6300 & 150 & 4.0 & 0.5 & @xmath142 & 0.3 & 1.5 & 0.5 + sdss 0924 + 40 & 6200 & 150 & 4.0 & 0.3 & @xmath143 & 0.3 & 1.4 & 0.3 + sdss 1707 + 58 & 6700 & 200 & 4.2 & 0.3 & @xmath140 & 0.3 & 1.5 & 0.5 + sdss 2047 + 00 & 6600 & 200 & 4.5 & 0.3 & @xmath141 & 0.3 & 1.3 & 0.3 + @l@c@c@c@c@c@c@c@c@c@c@c@c@c@c@c@c@c@c@c@c@c sun & & & & & & & & & & & & & & & & & & & & & + log(a ) & 7.45 & 7.45 & 1.05 & 8.39 & 6.17 & 7.53 & 6.31 & 3.05 & 4.90 & 4.90 & 5.64 & 4.92 & 6.23 & 4.60 & 2.92 & 2.21 & 2.59 & 2.17 & 1.13 & 2.00 + & & & & & & & & & & & & & & & & & & & & & + @xmath72(x ) & 4.92 & 4.92 & 2.3 & 8.00 & 4.76 & 5.42 & 4.00 & 0.72 & 2.92 & 2.76 & 3.04 & 2.60 & 3.77 & 2.07 & 0.47 & & & 1.72 & 0.52 & 2.0 + @xmath144x / fe ] & -2.53 & -2.53 & & 2.14 & 1.12 & 0.42 & 0.22 & 0.20 & 0.55 & 0.39 & -0.07 & 0.21 & 0.07 & 0.00 & 0.08 & & & 2.08 & 1.92 & 2.53 + @xmath145 & 59 & 11 & 1 & & 2 & 5 & 13 & 1 & 5 & 9 & 3 & 1 & 1 & 1 & 2 & & & 4 & 2 & 1 + @xmath54 & 0.12 & 0.12 & 0.2 & 0.25 & 0.25 & 0.13 & 0.09 & 0.18 & 0.14 & 0.12 & 0.14 & 0.19 & 0.18 & 0.13 & 0.22 & & & 0.18 & 0.16 & + & & & & & & & & & & & & & & & & & & & & & + @xmath72(x ) & 5.39 & 5.38 & @xmath146 & 8.4 & 4.25 & 5.69 & 4.45 & & 3.39 & 3.08 & 3.53 & & 4.52 & & 1.63 & 1.50 & 1.80 & 2.5 & 1.16 & 3.1 + @xmath144x / fe ] & -2.06 & -2.07 & & 2.07 & 0.14 & 0.22 & 0.19 & & 0.54 & 0.24 & -0.05 & & 0.35 & & 0.77 & 1.34 & 1.26 & 2.39 & 2.09 & 3.16 + @xmath29 & 40 & 11 & & & 2 & 4 & 8 & & 2 & 7 & 1 & & 1 & & 2 & 2 & 2 & 4 & 2 & 1 + @xmath54 & 0.11 & 0.12 & & 0.26 & 0.19 & 0.16 & 0.10 & & 0.11 & 0.14 & 0.11 & & 0.11 & & 0.27 & 0.20 & 0.20 & 0.22 & 0.20 & + & & & & & & & & & & & & & & & & & & & & & + @xmath72(x ) & 5.04 & 5.03 & @xmath147 & 8.30 & 4.54 & 5.37 & 4.16 & & & 3.00 & 3.05 & 2.73 & 3.65 & & 0.24 & & & 0.05 & & @xmath146 + @xmath144x / fe ] & -2.41 & -2.42 & & 2.32 & 0.78 & 0.25 & 0.26 & & & 0.51 & -0.18 & 0.22 & -0.17 & & -0.27 & & & 0.29 & & + @xmath29 & 33 & 5 & & & 2 & 3 & 9 & & & 8 & 2 & 1 & 1 & & 2 & & & 4 & & + @xmath54 & 0.17 & 0.15 & & 0.32 & 0.32 & 0.19 & 0.13 & & & 0.15 & 0.16 & 0.18 & 0.15 & & 0.29 & & & 0.24 & & + & & & & & & & & & & & & & & & & & & & & & + @xmath72(x ) & 4.34 & 4.34 & @xmath148 & 8.2 & 3.75 & 5.03 & & & & 2.45 & & & & & 1.15 & & 1.41 & 1.81 & 0.48 & 2.3 + @xmath144x / fe ] & -3.11 & -3.11 & & 2.92 & 0.69 & 0.61 & & & & 0.66 & & & & & 1.35 & & 1.93 & 2.75 & 2.46 & 3.41 + @xmath29 & 6 & 2 & & & 2 & 4 & & & & 4 & & & & & 2 & & 1 & 5 & 2 & 1 + @xmath54 & 0.16 & 0.14 & & 0.32 & 0.20 & 0.18 & & & & 0.14 & & & & & 0.37 & & 0.19 & 0.30 & 0.16 & + & & & & & & & & & & & & & & & & & & & & & + @xmath72(x ) & 4.29 & & @xmath146 & @xmath149 & & 4.8 & & & & & & & & & -0.1 & & & -0.22 & & + @xmath144x / fe ] & -3.16 & & & @xmath148 & & 0.43 & & & & & & & & & 0.14 & & & 0.77 & & + @xmath29 & 5 & & & & & 3 & & & & & & & & & 1 & & & 1 & & + @xmath54 & 0.20 & & & & & 0.23 & & & & & & & & & 0.40 & & & 0.35 & & + & & & & & & & & & & & & & & & & & & & & & + @xmath72(x ) & 4.94 & 4.91 & @xmath147 & 8.6 & 4.97 & 5.55 & 4.08 & 0.89 & & 2.79 & 2.84 & & 3.78 & & 1.02 & & & 1.48 & & 2.5 + @xmath144x / fe ] & -2.51 & -2.55 & & 2.72 & 1.31 & 0.52 & 0.28 & 0.35 & & 0.40 & -0.29 & & 0.05 & & 0.60 & & & 1.81 & & 3.01 + @xmath29 & 33 & 4 & & & 2 & 4 & 6 & 1 & & 9 & 2 & & 1 & & 2 & & & 5 & & 1 + @xmath54 & 0.16 & 0.13 & & 0.32 & 0.29 & 0.15 & 0.12 & 0.14 & & 0.14 & 0.16 & & 0.15 & & 0.25 & & & 0.22 & & + & & & & & & & & & & & & & & & & & & & & & + @xmath72(x ) & 4.93 & 4.96 & @xmath150 & 8.0 : & 6.36 & 6.14 & 4.58 & & & 2.98 & & & & & 2.65 & & & 3.05 & & @xmath151 + @xmath144x / fe ] & -2.52 & -2.49 & & + 2.1 : & 2.71 & 1.13 & 0.79 & & & 0.60 & & & & & 2.25 & & & 3.40 & & + @xmath29 & 9 & 3 & & & 2 & 2 & 1 & & & 4 & & & & & 2 & & & 5 & & + @xmath54 & 0.16 & 0.15 & & 0.32 & 0.22 & 0.21 & 0.21 & & & 0.15 & & & & & 0.38 & & & 0.31 & & + & & & & & & & & & & & & & & & & & & & & & + @xmath72(x ) & 5.40 & 5.40 & @xmath146 & & 4.45 & 5.75 & 4.59 & & 3.66 & 3.29 & 3.49 & & 4.26 & & 1.56 & 0.95 & 1.5 & 1.62 & & + @xmath144x / fe ] & -2.05 & -2.05 & & 2.00 & 0.33 & 0.27 & 0.32 & & 0.80 & 0.44 & -0.10 & & 0.08 & & 0.68 & 0.79 & 0.96 & 1.50 & & + @xmath29 & 32 & 9 & & & 1 & 4 & 9 & & 1 & 8 & 2 & & 1 & & 2 & 1 & 1 & 5 & & + @xmath54 & 0.15 & 0.13 & & 0.32 & 0.28 & 0.18 & 0.13 & & 0.15 & 0.15 & 0.15 & & 0.15 & & 0.29 & & & 0.24 & & + lcccccccccc & 0.08 & @xmath152 & 0.01 & @xmath153 & @xmath154 & & 0.07 & @xmath152 & 0.01 & @xmath155 + & 0.01 & @xmath156 & 0.00 & @xmath152 & @xmath157 & & 0.02 & 0.10 & 0.00 & @xmath153 + c ( ch ) & 0.15 & @xmath158 & 0.00 & 0.00 & @xmath158 & & 0.15 & @xmath159 & 0.01 & 0.00 + & 0.11 & @xmath160 & 0.00 & @xmath161 & @xmath162 & & 0.07 & @xmath155 & 0.00 & @xmath163 + & 0.04 & @xmath154 & @xmath153 & @xmath161 & @xmath164 & & 0.06 & @xmath165 & 0.00 & @xmath157 + & 0.05 & @xmath166 & @xmath166 & @xmath152 & @xmath154 & & 0.05 & @xmath152 & 0.00 & @xmath152 + & 0.08 & @xmath166 & @xmath167 & @xmath166 & @xmath154 & & 0.07 & @xmath152 & 0.01 & @xmath166 + & 0.04 & @xmath168 & @xmath166 & @xmath153 & @xmath157 & & 0.04 & @xmath169 & 0.00 & @xmath153 + & 0.08 & @xmath166 & @xmath167 & @xmath152 & @xmath154 & & 0.08 & @xmath152 & 0.01 & @xmath153 + & 0.08 & @xmath167 & @xmath167 & @xmath152 & @xmath163 & & 0.07 & @xmath152 & 0.01 & @xmath152 + & 0.08 & @xmath170 & @xmath166 & @xmath171 & @xmath158 & & 0.07 & @xmath172 & @xmath152 & @xmath173 + & 0.09 & @xmath153 & @xmath166 & @xmath158 & @xmath174 & & 0.07 & @xmath172 & @xmath153 & @xmath160 + & 0.06 & @xmath168 & @xmath166 & @xmath153 & @xmath157 & & & & & + & 0.06 & @xmath175 & @xmath166 & @xmath172 & @xmath161 & & & & & + & 0.08 & @xmath167 & @xmath166 & @xmath152 & @xmath158 & & 0.07 & @xmath152 & @xmath167 & @xmath153 lccccccccccc & 6595 & 6500 & & 3.58 & 4.5 & & @xmath176 & @xmath177 & & 2.50 & 2.3 + sdss 0126 + 06 & 6970 & 6600 & & 4.01 & 4.1 & & @xmath178 & @xmath179 & & 2.71 & 2.9 + sdss 0817 + 26 & 6213 & 6300 & & 3.28 & 4.0 : & & @xmath180 & @xmath181 & & 1.19 & @xmath148 + sdss 0924 + 40 & 6264 & 6200 & & 3.67 & 4.0 & & @xmath182 & @xmath183 & & 2.58 & 2.7 + sdss 1707 + 58 & 6656 & 6700 & & 3.19 & 4.2 & & @xmath184 & @xmath185 & & 2.27 & 2.1 : + sdss 2047 + 00 & 6489 & 6600 & & 3.76 & 4.5 & & @xmath186 & @xmath187 & & 1.93 & 2.0 lccccccccccccccccc & @xmath1272.53 & 2.14 & 1.20 & 0.60 & 0.42 & 0.22 & 0.20 & 0.47 & @xmath1270.07 & 0.07 & 0.00 & 0.08 & 2.08 & 2.53 & 6300 & 4.30 & 1 + cs 29526110 & @xmath1272.06 & 2.07 & 1.40 & 0.14 & 0.22 & 0.19 & & 0.39 & @xmath1270.05 & 0.35 & & 0.77 & 2.39 & 3.16 & 6800 & 4.10 & 1 + sdss 003610 & @xmath1272.41 & 2.32 & & 0.78 & 0.25 & 0.26 & & 0.51 & @xmath1270.18 & @xmath1270.17 & & @xmath1270.27 & 0.29 & & 6500 & 4.50 & 1 + sdss 0126 + 06 & @xmath1273.11 & 2.92 & & 0.69 & 0.61 & & & 0.66 & 0.62 & & & 1.35 & 2.75 & 3.41 & 6600 & 4.10 & 1 + sdss 0924 + 40 & @xmath1272.51 & 2.72 & & 1.31 & 0.52 & 0.28 & 0.35 & 0.40 & @xmath1270.29 & 0.05 & & 0.60 & 1.81 & 3.10 & 6200 & 4.00 & 1 + sdss 1707 + 58 & @xmath1272.52 & 2.1 & & 2.71 & 1.13 & 0.79 & & 0.60 & & & & 2.25 & 3.40 & & 6700 & 4.20 & 1 + sdss 2047 + 00 & @xmath1272.05 & 2 .
& & 0.33 & 0.27 & 0.32 & & 0.62 & @xmath1270.10 & 0.08 & & 0.68 & 1.50 & & 6600 & 4.50 & 1 + cs29528 - 028 & @xmath1272.86 & 2.77 & & 2.33 & 1.69 & 0.46 & 0.59 & 0.87 & & 0.26 & & & 3.27 & & 6800 & 4.00 & 2 + cs22898 - 027 & @xmath1272.26 & 2.20 & 0.90 & 0.33 & 0.41 & 0.40 & & 0.41 & @xmath1270.10 & 0.02 & 0.92 & 0.92 & 2.23 & 2.84 & 6250 & 3.70 & 3,4 + cs29497 - 030 & @xmath1272.57 & 2.47 & 2.12 & 0.58 & 0.44 & 0.47 & 0.67 & 0.64 & 0.03 & 0.04 & & 0.84 & 2.32 & 3.55 & 7000 & 4.10 & 5 + he2148 - 1247 & @xmath1272.32 & 1.91 & 1.65 & & 0.50 & 0.45 & 0.59 & 0.55 & @xmath1270.35 & 0.06 & & 0.76 & 2.36 & 3.12 & 6380 & 3.90 & 6 + he0024 - 2523 & @xmath1272.72 & 2.60 & 2.10 & @xmath1270.17 & 0.73 & 0.66 & 0.37 & 0.85 & @xmath1270.41 & & & 0.34 & 1.46 & 3.30 & 6625 & 4.30 & 7 + he22881 - 036 & @xmath1272.06 & 1.96 & 1.00 & 0.16 & 0.40 & 0.62 & & 0.33 & & & & 0.59 & 1.93 & & 6200 & 4.00 & 8 + he0007 - 1832 & @xmath1272.65 & 2.55 & 1.85 & & 0.76 & 0.32 & & 0.39 & & 0.02 & & & 0.16 & & 6515 & 3.80 & 9 + he0338 - 3945 & @xmath1272.42 & 2.13 & 1.55 & 0.36 & 0.30 & 0.38 & 0.53 & 0.37 & @xmath1270.12 & 0.01 & & 0.74 & 2.41 & 3.10 & 6160 & 4.13 & 10 + he1105 + 0027 & @xmath1272.42 & 2.00 & & & 0.47 & 0.47 & 0.28 & 0.32 & 0.05 & @xmath1270.29 & & & 2.45 & & 6132 & 3.50 & 11 + he0143 - 0441 & @xmath1272.31 & 1.98 & 1.73 & & 0.63 & 0.43 & 0.67 & 0.40 & @xmath1270.38 & @xmath1270.31 & 0.46 & 0.86 & 2.32 & 3.11 & 6240 & 3.70 & 12 + cs31080 - 095 & @xmath1272.80 & 2.69 & 0.70 & @xmath1270.48 & 0.65 & 0.17 & @xmath1270.02 & 0.32 & 0.02 & 0.09 & 0.58 & @xmath1270.41 & 0.77 & & 6050 & 4.50 & 13 + cs22958 - 042 & @xmath1272.80 & 3.15 & 2.15 & 2.62 & 0.32 & 0.36 & 0.05 & 0.32 & @xmath1270.15 & @xmath1270.09 & & @xmath1270.20 & & & 6250 & 3.50 & 13 + cs29528 - 041 & @xmath1273.25 & 1.59 & 3.00 & 1.00 & 0.40 & 0.40 & 0.26 & 0.40 & @xmath1270.17 & 0.00 & & @xmath1270.20 & 0.97 & & 6150 & 4.00 & 13 + lrrrrrrrrrrrrrrrrr & 2.27 & @xmath188 & @xmath189 & @xmath190 & & @xmath191 & 48 & @xmath192 & 47 & @xmath193 & 17 & @xmath194 & 43 & 0.92 & 3 & 66 & 22 + sdss 0126 + 06 & 1.75 & @xmath195 & @xmath196 & @xmath197 & & @xmath198 & 23 & @xmath199 & 19 & @xmath200 & 13 & @xmath201 & 19 & 0.41 & 9 & 22 & 20 + sdss 0817 + 26 & 1.46 & @xmath193 & @xmath202 & @xmath203 & & @xmath204 & 13 & @xmath205 & 20 & @xmath206 & 19 & @xmath207 & 20 & 0.38 & 5 & 12 & 5 + sdss 0924 + 40 & 1.39 & @xmath208 & @xmath209 & @xmath210 & & @xmath211 & 18 & @xmath212 & 35 & @xmath213 & 18 & @xmath214 & 35 & 0.74 & 8 & 54 & 53 + sdss 1707 + 58 & 2.43 & @xmath215 & @xmath0 & @xmath216 & & @xmath217 & 33 & @xmath218 & 20 & @xmath219 & 27 & @xmath220 & 20 & 0.24 & 8 & 13 & 2 + sdss 2047 + 00 & 2.23 & @xmath221 & @xmath222 & @xmath223 & & @xmath224 & 22 & @xmath225 & 32 & @xmath226 & 41 & @xmath227 & 32 & 0.66 & 6 & 31 & 9 ) . the positions of intersteller lines are shown by dotted lines , while the line positions of the stellar absorption are shown by solid lines .
the emission from the earth s atmosphere and interstellar absorption show no variation in wavelengths , while the stellar absorption lines show rapid changes in the 10 february ( ut ) spectra.,width=302 ] and [ ba / fe ] as functions of [ fe / h ] .
filled circles are objects studied by the present work , while open ones are from the literature given in table [ tab : cemp_to ] .
the dotted line in the upper panel means the line for [ c / h]=0.,width=302 ] distributions for cemp turn - off stars ( upper ) and giants ( lower ) .
see text for the selection criteria for these samples .
the open histogram with strong lines indicates the objects having excesses of ba ( [ ba / fe ] @xmath228 ) . the hatched one is for ba - normal stars.,width=302 ] | the chemical compositions of seven carbon - enhanced metal - poor ( cemp ) turn - off stars are determined from high - resolution spectroscopy .
five of them are selected from the sdss / segue sample of metal - poor stars .
another star , also chosen from the sdss / segue sample , has only a weak upper limit on its carbon abundance obtained from the high - resolution spectrum .
the effective temperatures of these objects are all higher than 6000 k , while their metallicities , parametrized by [ fe / h ] , are all below @xmath0 ; the star with the lowest iron abundance in this study has [ fe / h ] = @xmath1 six of our program objects exhibit high abundance ratios of barium ( [ ba / h ] @xmath2 ) , suggesting large contributions of the products of former agb companions via mass transfer across binary systems .
one star in our study ( sdss 1707 + 58 ) exhibits a rapid variation in its radial velocity , which is a strong signature that this star belongs to a close binary .
combining our results with previous studies provides a total of 20 cemp main - sequence turn - off stars for which the abundances of carbon and at least some neutron - capture elements are determined .
inspection of the [ c / h ] ratios for this sample of cemp turn - off stars show that they are generally higher than those of cemp giants ; their dispersion in this ratio is also smaller .
we take these results to indicate that the carbon - enhanced material provided from the companion agb star is preserved at the surface of turn - off stars with no significant dilution , which appears counter to expectations if processes such as thermohaline mixing have operated in unevolved cemp stars .
in contrast to the behavior of [ c / h ] , a large dispersion in the observed [ ba / h ] is found for the sample of cemp turn - off stars , suggesting that the efficiency of the s - process in very metal - poor agb stars may differ greatly from star to star .
four of the six stars from the sdss / segue sample exhibit kinematics that are associated with membership in the outer - halo population , a remarkably high fraction . |
calculations were completed on chama and skybridge at sandia national laboratories and sequoia at lawrence livermore national laboratory .
we are grateful for discussions with stephen bond , kieron burke , hardy gross , ryan hatcher , neepa maitra , thomas mattsson , normand modine , jonathan moussa , kai - uwe plagemann , aurora pribram - jones , kenneth rudinger , travis sjostrom , and sam trickey .
a.d.b , l.s .
, m.p.d . , and r.j.m .
were supported by sandia s laboratory directed research and development ( ldrd ) project 165731 .
s.b.h . was supported by the u.s .
department of energy , office of science early career research program , office of fusion energy sciences .
sandia national laboratories is a multi - program laboratory managed and operated by sandia corporation , a wholly owned subsidiary of lockheed martin corporation , for the u.s .
department of energy s national nuclear security administration under contract de - ac04 - 94al85000 .
53ifxundefined [ 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 10.1126/science.286.5437.72 [ * * , ( ) ] link:\doibase 10.1007/s11214 - 009 - 9582-x [ * * , ( ) ] link:\doibase 10.1016/j.epsl.2009.11.002 [ * * , ( ) ] link:\doibase 10.1063/1.1891746 [ * * , ( ) ] link:\doibase 10.1063/1.3078101 [ * * , ( ) ] link:\doibase 10.1103/physreve.89.043105 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.81.1625 [ * * , ( ) ] link:\doibase 10.1103/physreve.67.026412 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.98.065002 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.102.115001 [ * * , ( ) ] http://www.nature.com/nphys/journal/v4/n12/abs/nphys1103.html [ * * , ( ) ] http://www.ncbi.nlm.nih.gov/pubmed/24909903 [ * * ( ) ] http://dx.doi.org/10.1103/physrevlett.112.145004 [ * * , ( ) ] http://dx.doi.org/10.1103/physrevlett.110.065001 [ * * , ( ) ] http://www.nature.com/nphoton/journal/v9/n4/abs/nphoton.2015.41.html [ * * , ( ) ] link:\doibase 10.1088/0305 - 4608/17/2/002 [ * * , ( ) ] link:\doibase 10.1088/0953 - 8984/12/3/303 [ * * , ( ) ] link:\doibase 10.1103/physreve.89.023108 [ * * , ( ) ] http://dx.doi.org/10.1103/physreve.92.013103 [ * * , ( ) ] http://dx.doi.org/10.1103/physreve.91.033112 [ * * , ( ) ] @noop _ _ ( , ) in @noop _ _ , vol .
( ) link:\doibase 10.1103/physrevb.50.17953 [ * * , ( ) ] link:\doibase 10.1103/physrevb.59.1758 [ * * , ( ) ] link:\doibase 10.1103/physrevb.54.11169 [ * * , ( ) ] link:\doibase 10.1016/0927 - 0256(96)00008 - 0 [ * * , ( ) ] http://iopscience.iop.org/0953-8984/24/23/233202 [ * * , ( ) ] http://dx.doi.org/10.1109/mcse.2014.55
[ * * , ( ) ] link:\doibase 10.1103/physrev.137.a1441 [ * * , ( ) ] @noop _ _ ( , ) http://arxiv.org/abs/1509.03071 [ ( ) ] link:\doibase 10.1063/1.3503594 [ * * , ( ) ] link:\doibase 10.1088/1367 - 2630/14/5/055020 [ * * , ( ) ] http://dx.doi.org/10.1103/physreve.86.036410 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.075109 [ * * , ( ) ] http://dx.doi.org/10.1103/physrevb.85.115135 [ * * , ( ) ] link:\doibase 10.1103/physreva.31.1950 [ * * , ( ) ] link:\doibase 10.1103/physreva.31.3970 [ * * , ( ) ] link:\doibase 10.1103/physrevb.88.081102 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.112.076403 [ * * , ( ) ] link:\doibase 10.1063/1.4872255 [ * * , ( ) ] in link:\doibase 10.1007/978 - 3 - 319 - 04912 - 0 [ _ _ ] , , vol . , ( , ) pp . link:\doibase 10.1103/physrevb.73.035408 [ * * , ( ) ] link:\doibase 10.1063/1.3700800 [ * * , ( ) ] @noop _ _ ( , ) @noop _ _ ( , ) http://dx.doi.org/10.1103/physrevb.81.085104 [ * * , ( ) ] http://dx.doi.org/10.1103/physreva.2.614 [ * * , ( ) ] http://dx.doi.org/10.1103/physrevb.46.12947 [ * * , ( ) ] http://dx.doi.org/10.1103/physrevb.61.7353 [ * * , ( ) ] http://dx.doi.org/10.1103/physrevlett.45.566 [ * * , ( ) ] http://dx.doi.org/10.1103/physrevlett.82.5317 [ * * , ( ) ] * supplemental materials : x - ray thomson scattering without the chihara decomposition *
we have implemented real - time tddft using a plane wave basis and the projector augmented wave ( paw ) formalism @xcite within version 5.3.5 of vasp @xcite . as in other implementations using ultrasoft pseudo - potentials @xcite or paw @xcite we chose a crank - nicolson ( cn ) time integration scheme to propagate the time - dependent kohn - sham ( tdks ) equations numerically rather than chosing an integrator to achieve a high - order convergence in the local error .
the unitarity of the discrete propagator was deemed an essential factor to guarantee the satisfaction of charge conservation and other sum rules associated with the dynamic structure factor ( dsf ) itself . at each timestep ,
the cn - discretized tdks equations are solved using the generalized minimal residual method ( gmres ) @xcite . given the perturbation on identity form of the discrete propagator we expect and observe rapid convergence in the number of iterations .
as many details of our implementation have not been published elsewhere , we begin by demonstrating that it is robust .
we first test stability of the time integration .
specifically , we illustrate that our implementation is not susceptible to any significant or uncontrolled errors , given the intrinsic nonlinearity of the tdks equations and the use of an iterative solver on the linearized problem .
to do so , we time propagate a mermin thermal equilibrium set of orbitals in the absence of a probe potential or ionic motion . here
, the time - dependent potential should be constant in time , and any variations in the instantaneous total free energy or supercell charge are due to the accumulation of numerical errors .
results are reported for 32 beryllium atoms under the warm dense conditions in the paper and using the adiabatic zero - temperature local density approximation throughout .
the initial condition was generated from a mermin dft calculation done on a single atomic configuration drawn from a thermalized dft - md run with a four - electron paw potential , 576 thermally occupied orbitals , and a plane wave cutoff of 1400 ev .
the time propagation utilized the same plane wave cutoff and number of orbitals , and a time step ( @xmath53 ) of 1 attoseconds ( as ) carried out for 8000 steps ( 8 fs ) .
we varied the relative tolerance of the iterative solver , and kept the absolute tolerance fixed to be 2 orders of magnitude smaller .
as the right hand side of the cn - discretized tdks system is of order unit norm , the absolute tolerance is effectively irrelevant .
the resultant free energy per atom and total charge density of the supercell are reported in figure [ fig : conservation ] . ] here , we see that the free energy drift per atom is @xmath54ev/(atom@xmath55fs ) and that the drift in the total charge of the unit cell is @xmath56 electrons / fs in the most permissive case ( 8 digits ) . in the least permissive case
( 14 digits ) , these quantities are @xmath57ev/(atom@xmath55fs ) and @xmath58 electrons / fs , with the 12 digit case producing drifts of a similar order of magnitude . going from 8 digits to 14 digits ,
the average cpu time per step varies linearly from 0.85 s / time step to 1.64 s / time step , i.e. , convergence is exponential in the number of gmres iterations .
as the results are practically indistinguishable from 14 digits and the cpu time per time step is @xmath59 shorter , a relative tolerance of 12 digits was used in all subsequent calculations .
pertinent to the satisfaction of sum rules on @xmath28 , in all calculations in this work , we have verified that charge conservation is guaranteed to a relative accuracy of greater than 8 digits .
we next demonstrate convergence of the integrated density response , the observable of interest , in @xmath53 .
to do so , we consider the same 32 beryllium atom initial condition described above , this time applying a time - varying scalar perturbation of the form described in the main body of the paper . here
@xmath60 where @xmath61 with @xmath62 as , @xmath63 as , @xmath64 ev@xmath55fs , and @xmath65 @xmath66 .
@xmath53 is varied from from @xmath67 as to @xmath68 as , and the associated convergence of the real - time density response is illustrated in figure [ fig : timestep_convergence ] . in @xmath53 .
( a ) illustrates the @xmath69 $ ] for the sinusoidal component of the perturbation at different @xmath53 and ( b ) its instantaneous difference relative to @xmath70 as .
the relative @xmath71 distances from the @xmath72 as trajectory over a 2 fs interval are 5.7@xmath73 , 3.5@xmath73 , 1.9@xmath73 , and 0.8@xmath73 , in order of decreasing @xmath53 .
[ fig : timestep_convergence ] ] from these results , it is evident that the density response exhibits first order convergence .
further , the convergence supports the decision that @xmath74 as ( attosecond ) provides a reasonable balance between accuracy and time required per calculation for generating production results .
next , we consider the determination of parameters for @xmath16 .
we chose @xmath33 to take the form of a gaussian envelope to ensure that the exciting potential and its response are approximately band- and time - limited . to this end
, the pulse width ( @xmath75 ) , determines the bandwidth of the response and was chosen to ensure that modes of the density response with energies on the order of @xmath76s of ev are excited with appreciable amplitude .
the delay , @xmath77 , was chosen to ensure that the excitation is approximately quiescent at @xmath10 as .
the remaining parameter , @xmath34 , determines the effective intensity of the probe potential , and must be chosen to be large enough that the response is not dominated by numerical noise , yet small enough to remain in the linear response regime .
varying @xmath34 over 4 orders of magnitude from @xmath78 ev@xmath55fs to @xmath79 ev@xmath55fs , we do not observe numerical noise to be a problem at any value .
the post - processed dsf computed using these probe amplitudes for a single 32 atom configuration at @xmath80@xmath81 are illustrated in figure [ fig : amplitude_variation ] .
the results for @xmath34 ranging from @xmath79 ev@xmath55fs to @xmath82 ev@xmath55fs are indistinguishable , while the distortion of the results at @xmath78 ev@xmath55fs indicate the onset of physics beyond linear response . that we can easily access this regime is one of the benefits of real - time tddft , though we do not explore this further in this work . in ev@xmath55fs ) .
the spectra for small probe amplitude are indistinguishable indicating that we remain within the linear response regime .
[ fig : amplitude_variation ] ]
each tddft calculation requires a set of kohn - sham orbitals and occupancies as initial conditions .
these are generated using dft - md as implemented in the standard version of the vasp software package @xcite . to assess the impact of the supercell shape and the underlying ionic positions on our dsf calculations , for each value of @xmath1 we run 2 separately thermalized dft - md trajectories on different supercells and draw 5 independent sets of initial conditions from each .
each set of initial conditions is used to seed independent tddft calculations of the dsf at a fixed @xmath1 .
separate dft - md configurations are needed for each value of @xmath1 due to the requirements of the probe potential . for the dsf calculations ,
@xmath16 must be commensurate with our supercell , and consequently any realizable value of @xmath1 must be in the reciprocal lattice of our supercell . to precisely specify the value of @xmath1 we work with tetragonal supercells in which the perturbing @xmath1 is directed along the c - axis . then @xmath1 can be set by scaling the c - axis and the a - axes can be adjusted to ensure that the desired mass density is realized . given the liquid - like ordering in the extreme conditions under investigation , we do not anticipate that this biases our results .
table [ table : geometries ] gives the dimensions of the 8 supercells used in this study .
.parameters defining tetragonal supercell dimensions used in this study .
5 independent initial conditions were drawn from each row of this table to seed tddft calculations of the dsf at momentum transfer @xmath1 .
each @xmath1 is equivalent to an xrts scattering angle , @xmath38 , through the relationship @xmath83 and @xmath84 is the 2 probe wavelength from @xcite .
[ table : geometries ] [ cols="^,^,^,^,^,^",options="header " , ] it is worth noting that all dft - md trajectories are generated with the full four - electron paw potential .
however , both the two - electron and four - electron tddft calculations are initialized from this same set of ionic configurations , with the kohn - sham orbitals and occupancies being recomputed for each set of fixed ionic positions in the two - electron case
. this is done to assess the impact of the frozen core approximation on the dsf in isolation .
all dsf results reported are averaged over 10 configurations sampled at each value of @xmath1 .
the sample size is necessarily small due to the significant computational resources required for each tddft calculation , with each production calculation being done on 1,152 cores .
however , because the electronic density of warm dense beryllium is relatively uniform , we do not see much variability in the dsf from configuration to configuration . in an effort to quantify this variability
, we apply jackknife resampling to estimate the variance and indicate single standard deviation intervals . for the scattering angles considered in the manuscript , we show the estimated error intervals for the dsf computed with four - electron paw potentials in figure [ fig : dsf_variation ] .
4@xmath85 about the mean , which is a solid black line .
the @xmath861@xmath85 intervals are barely perceptible , and we simply present the mean values in the manuscript . [
fig : dsf_variation ] ]
the units on the density response and dsf are such that the following form of the f - sum rule @xcite is satisfied : here , @xmath88 is 2 for the two - electron paw , and 4 for the four - electron paw .
similar real - time calculations in the gas phase have reported satisfaction to within @xmath89 @xcite and we report similar results .
forms of the dsf based upon model dielectric functions may or may not be consistent with various sum rules by construction . our dsf is derived strictly from a time - evolved electronic density for which charge conservation is numerically guaranteed to high precision , and makes no assumptions about the form of the equivalent dielectric response . to this end ,
the primary source of error in our satisfaction of sum rules is due to numerical errors in the post - processing , e.g. , the numerical evaluation of fourier integrals with integrands which become increasingly oscillatory at higher energies , and thus more prone to small errors in the response . when checking sum rules produced using linear response tddft , it is common practice to fit the high energy tail of the dsf to force it to go to zero in a way that is consistent with the integrand in the above sum rule @xcite .
this is also critical for real - time simulations where small phase errors between the real and imaginary parts of the time domain density response are amplified at high energies relative to low energies when post - processing to generate the energy domain response . rather than force the tail to fit some form
, we simply impose a high energy cutoff on the integrand once the dsf has decayed to @xmath90 of its peak value .
we simply seek to verify that the majority of the weight of our response is consistent with the sum rule , and could resort to fitting tails if necessary to improve agreement . applying this technique
, we verify that we satisfy the f - sum rule for our four - electron data with a relative error of @xmath91 ( @xmath92 ) , @xmath93 ( @xmath94 ) , @xmath95 ( @xmath96 ) , and @xmath97 ( @xmath98 ) . in all cases ,
we underestimate the value of the f - sum rule , giving us confidence that a fit of the slowly - decaying high energy tail might be used to improve this agreement . applying this same technique to the two - electron data we find relative errors of @xmath99 ( @xmath92 ) , @xmath89 ( @xmath94 ) , @xmath100 ( @xmath96 ) , and @xmath95 ( @xmath98 ) . here
, we do not uniformly underestimate the sum rule as was the case with the four - electron data .
the two - electron data decays much more abruptly at high energies , and does not need to represent the response of slowly - decaying core states , so we do not view this as a deficiency ( i.e. , fitting a tail would not have as strong of an impact here )
. however , this seems to point to the two - electron response as overestimating the free - free peak , consistent with the small negative dips in the bound - free data in fig .
3b for @xmath92 , @xmath101 , and @xmath102 ( the calculations for which the sum rule was over - estimated ) . whether or not this can be improved with a different two - electron paw may be an interesting topic of further study . computing this integral for the effective free - free and bound - free data from tddft
, we can extract structure factors that we can compare against physically intuitive models .
results are presented in fig .
[ fig : static_structure_factor ] . applied to the effective free - free and bound - free responses extracted from tddft .
the bound - free structure factor is compared to that of doubly- and triply - ionized beryllium @xcite.[fig : static_structure_factor ] .
the free - free structure factor is compared to an analytic fit of qmc data for jellium with @xmath104 a.u @xcite . ] for the bound - free component we compare to results obtained using a configuration - interaction expansion by brown for doubly- and triply - ionized beryllium @xcite . for @xmath51 and @xmath52
, the tddft gives results that are closer to doubly - ionized beryllium , consistent with physical intuition for the thermodynamic conditions under consideration . for @xmath49 and @xmath50
, the tddft gives results that are closer to triply - ionized beryllium , though the difference between the doubly- and triply - ionized structure factors are smaller at these angles .
we note that by excluding the negative difference data below 80 ev in our integration ( evident in fig .
3b ) , we can improve agreement with the doubly ionized curve at all angles .
this indicates that the underestimation of the tddft structure factor may be due to differences between the free electron response for the two - electron and four - electron paws , which may or may not be physical .
this highlights the importance of being able to self - consistently compute the four - electron response without the chihara decomposition using our methodology . for the free - free component ( our frozen core result )
we compare to results for the static structure factor of jellium with @xmath104 a.u .
, consistent with the experimentally determined free - electron density @xcite . in evaluating eqn .
[ eq : static_structure ] , we remove the elastic peak in a region of width @xmath105 centered at @xmath106 and apply a simple linear interpolant between the dsf data on either side of the excluded region .
we compare the resultant free - free static structure factor to an analytic fit to qmc data @xcite by gori - giorgi , et .
the tddft free - free structure factor exhibits the same trend as the qmc fit , but is not expected to agree perfectly .
we only expect qualitative agreement because the external potential experienced by the free electrons in warm dense beryllium is _ not _ a uniform neutralizing background as is the case in jellium .
this qualitative agreement between results from tddft and qmc stands in contrast to vorberger and gericke s recent work in which dft - md is used to compute a free electron density , from which the free - free static structure factor of warm dense beryllium is extracted @xcite . in their work ,
the static structure factor computed from dft differs greatly from qmc , namely it does not go from one to zero for momentum transfers less than @xmath107 ( @xmath108 @xmath66 for our conditions ) , but instead oscillates slightly about unity .
the authors postulate that this is due to the mean field nature of the kohn - sham equations , and that this may be beyond the capability of dft .
our results indicate that this basic physics is in fact within the grasp of tddft . to this end
, it is worth noting that tddft has provided us with a convenient means of computing the static structure factor as an exact functional of the time - dependent density . in this case , we mean exact in the sense that if we are given a representation of the exact interacting time - dependent density , we can map it onto the exact dsf and the exact structure factor through eqn .
[ eq : static_structure ] . | x - ray thomson scattering ( xrts ) is an important experimental technique used to measure the temperature , ionization state , structure , and density of warm dense matter ( wdm ) .
the fundamental property probed in these experiments is the electronic dynamic structure factor ( dsf ) . in most models ,
this is decomposed into three terms [ chihara , j. phys .
f : metal phys . *
17 * , 295 ( 1987 ) ] representing the response of tightly bound , loosely bound , and free electrons . accompanying
this decomposition is the classification of electrons as either bound or free , which is useful for gapped and cold systems but becomes increasingly questionable as temperatures and pressures increase into the wdm regime . in this work
we provide unambiguous first principles calculations of the dynamic structure factor of warm dense beryllium , independent of the chihara form , by treating bound and free states under a single formalism .
the computational approach is real - time finite - temperature time - dependent density functional theory ( tddft ) being applied here for the first time to wdm .
we compare results from tddft to chihara - based calculations for experimentally relevant conditions in shock - compressed beryllium .
= 1 warm dense matter ( wdm ) arises in many contexts ranging from planetary science@xcite to the implosion stage of inertial confinement fusion@xcite .
while there are no sharp pressure , temperature and density boundaries for the wdm regime , it is generally viewed as an intermediate state between a condensed phase and an ideal plasma where fermi degeneracy is present , and the coulomb coupling and thermal energy are comparable in magnitude@xcite
. experimental characterization of warm dense matter is challenging due to the difficulty of producing uniform samples at extreme conditions and developing diagnostic techniques that can provide accurate and independent measurements of these conditions for transient samples opaque to optical photons .
x - ray thomson scattering ( xrts)@xcite , one such diagnostic technique , exploits the scattering of hard coherent x - rays to directly probe the system s dynamic structure factor ( dsf ) . through the fluctuation - dissipation theorem ,
the dsf is related to the system s density - density response , and consequently xrts provides direct insight into electron dynamics .
xrts experiments have been performed on a variety of materials including beryllium@xcite , lithium@xcite , carbon@xcite , ch shells@xcite , and aluminum@xcite . with recent improvements in source brightness@xcite producing increasingly high resolution and high signal - to - noise data
, the full dsf is expected to become routinely available . in anticipation of these advances , it is critical and timely to examine the theoretical constructs underpinning the interpretation of these experiments .
the most common model of xrts experiments relies on an additive form of the dsf due to chihara@xcite : @xmath0 the dsf varies with momentum and energy transfers ( @xmath1 and @xmath2 ) and is partitioned into 3 features that can be interpreted in terms of x - ray scattering processes .
these include scattering from electrons bound to and adiabatically following ions ( @xmath3 ) , from @xmath4 free electrons per ion ( @xmath5 ) , and from bound electrons that are photo - ionized ( @xmath6 ) .
while successfully applied to many systems , this model relies on numerous approximations and assumptions .
most critically , the electrons are separated into bound and free populations , a distinction that is often ambiguous in the wdm regime . each term in eqn .
[ eq : chihara_decomposition ] is subject to different models potentially leading to under - constrained fits to experimental data@xcite .
the ionic feature typically relies upon a decomposition into a product of an ion - ion structure factor , @xmath7 , and an average atomic form factor , @xmath8 , with the first term describing the unscreened bound electrons and the latter the screening cloud , which must be treated carefully@xcite .
however , recent work has focused on moving past this decomposition of the elastic peak@xcite . in this work , we transcend the chihara decomposition by explicitly simulating the real - time dynamics of warm dense matter using a finite temperature form of time - dependent density functional theory ( tddft)@xcite and the projector augmented - wave ( paw ) formalism @xcite . in paw , the all - electron kohn - sham orbitals ( and their associated density ) can be accessed via an explicit linear transformation on the smoother pseudo orbitals . by performing calculations with none of the core states frozen , we avoid making any assumptions between bound and free electrons , and treat all electrons similarly .
we work with a real - time implementation of tddft in the vienna ab - initio simulation package ( vasp)@xcite ( see supplemental material ) that provides a number of attractive features .
physically , higher - order response phenomena and ehrenfest molecular dynamics are accessible in this framework .
computationally , the orthogonalization bottleneck that limits standard dft approaches is removed , as it is only explicitly required for the calculation of the initial state of the kohn - sham orbitals .
this leads to excellent strong scaling@xcite , and we have observed near - perfect scaling up to 65,536 cores in our implementation .
we next outline the details of our tddft calculations , noting that we use hartree atomic units ( @xmath9 ) unless otherwise indicated .
time - dependent quantities evaluated at @xmath10 are indicated by the addition of a subscript @xmath11 and the absence of a temporal argument .
fourier transformed quantities are indicated by a diacritic tilde .
all calculations are spin unpolarized .
the equation of motion in real - time tddft is the time - dependent kohn - sham ( tdks ) equation : @xmath12({{\bf r}},t)\right ) \phi_{n,{{\bf k}}}({{\bf r}},t ) \label{eq : tdks_eqn}\ ] ] in which @xmath13({{\bf r}},t ) = v_{ext}({{\bf r}},t ) + v_h\left[\rho\right]({{\bf r}},t ) + v_{xc}\left[\rho \right]({{\bf r}},t)$ ] .
the orbitals , indexed by band and bloch wave number , are such that their weighted sum produces the time - dependent density , @xmath14 .
the external potential , @xmath15 , includes contributions due to the coulomb field of the bare nuclei as well as a model of the x - ray probe , @xmath16 , which is quiescent until @xmath10 .
@xmath17 and @xmath18 are the hartree and exchange - correlation potentials , with accurate and efficiently computable approximations to the latter being a central theoretical concern of tddft . the initial conditions for the @xmath19 from eqn .
[ eq : tdks_eqn ] are the self - consistent solution to a kohn - sham mermin dft calculation@xcite at electron temperature @xmath20 in a supercell of volume @xmath21 .
the initial density is then : @xmath22({{\bf r}})|^2\ ] ] where @xmath23 is a composite weight consisting of the measure of the specific bloch orbital weighting and the mermin weight encoding temperature dependence according to a fermi - dirac distribution .
it is important to note the implicit dependence of these initial orbitals on the equilibrium electron temperature , @xmath20 . evolving these orbitals under eqn .
[ eq : tdks_eqn ] , the time - dependent density becomes : @xmath24 the weights are not time - evolved , as we do not expect them to change in the linear response regime@xcite . for stronger perturbations , additional formalism might be required for the weights . that the mermin formalism is sensible within tddft in the linear response regime
is supported by recent foundational work@xcite .
the action of a probe potential , @xmath16 , turned on at @xmath10 leads to a change in the time - dependent density .
the linear density - density response function , @xmath25 , encodes the relationship between these two quantities : @xmath26 where we use the notation @xmath27 for @xmath28 and @xmath29 . if we fix ionic positions at time @xmath10 , which evolve slowly on the relevant attosecond timescales , then @xmath30 and we can construct a probe potential that can be used to extract the dsf , similar to sakko , et .
al.@xcite .
the real - time density response to such a probe potential is shown in fig .
[ fig : exemplary_calculation ] . and
@xmath31 ev , movie in supplemental material @xcite ) due to ( a ) a perturbing potential with the illustrated envelope observed at times coinciding with ( b ) the peak of the perturbation , ( c ) the peak of the density response , and ( d ) the plasmons continuing to ring around the system after the perturbation has ceased .
red ( blue ) isosurfaces bound volumes of charge accumulation ( depletion ) and yellow isosurfaces indicate the nodal surface .
the amplitude of the perturbation is within the linear response regime , a discussion of which is in the supplemental material @xcite .
[ fig : exemplary_calculation ] ] in principle , any sufficiently weak analytic probe potential will allow us to extract the response function . for convenience ,
we choose @xmath32 , where @xmath33 is a gaussian envelope and @xmath34 is related to the probe intensity .
the fourier transformed response function , @xmath35 , is then related to the dsf through the fluctuation - dissipation theorem : @xmath36}{1-e^{-\omega / k_bt_e}}\ ] ] we are careful to note that fourier transforms are normalized such that @xmath37 has units of inverse frequency . as a proof of principle , we report calculations of the dsf for 3x - compressed beryllium consistent with the conditions reported in @xcite .
we consider the same momentum transfers as in @xcite to study a range of excitations spanning the collective regime to the beginning of non - collective regime .
for convenience of presentation , each @xmath1 value is mapped onto an xrts scattering angle , @xmath38 , relative to the 2 probe wavelength in @xcite ( see supplemental material @xcite for more information ) .
our results come from averaging the response of electronic densities generated from several static uncorrelated ionic configurations sampled from thermally equilibrated dft - md calculations .
these calculations were performed on 32 and 64 atom supercells with a four electron beryllium paw potential within the local density approximation ( lda ) , with electrons and ions thermostatted at @xmath39 ev , and @xmath40-point sampling at @xmath41 , analogous to the baldereschi mean - value point for cubic supercells . for these conditions
a plane wave cutoff of @xmath42 ev was required to converge the pressure to within @xmath43 , and 576 ( 32 atom)/1152 ( 64 atom ) kohn - sham orbitals were needed to represent the thermal occupation ( @xmath44 ) .
each sample configuration is used to seed a tddft calculation of the dsf , utilizing the same cutoff and number of orbitals
. the initial mermin electronic state is recomputed using a denser @xmath40-point sampling on a @xmath45 ( 32 atom ) or @xmath46 ( 64 atom ) monkhorst - pack grid . to assess the effect of the frozen core approximation ( fca ) ,
we consider electronic initial conditions and dynamics generated using both two and four electron beryllium paw potentials . results of our calculations are illustrated in figs .
[ fig : beryllium_dsf ] and [ fig : comparison ] .
details of the averaging procedure used to generate results , and information concerning the satisfaction of sum rules , can be found in the supplemental material @xcite . and @xmath31 ev ) at the scattering angles ( @xmath38 ) considered in @xcite with ( dashed lines ) and without ( marked lines ) the frozen core approximation .
all tddft calculations utilize the adiabatic lda.[fig : beryllium_dsf ] ] fig .
[ fig : beryllium_dsf ] directly compares the dsf computed with and without the fca . while the paw method still includes the proper all - electron density in aggregate , only the orbitals tied to the two outermost valence electrons
are included in the time - evolved response within the fca .
this effectively removes the dynamics of the core states from the density response and the high energy shoulder above 80 ev in the dsf is removed . for the temperatures and densities
being considered , this is roughly equivalent to partitioning the inner two and outer two electrons into bound and free groups in the chihara picture , such that the two - electron response corresponds to @xmath47 .
however , there are important distinctions to keep in mind . first , in the four - electron calculation
, all electrons are being treated identically , whereas it is typical to treat @xmath47 and @xmath6 using different levels of theory and without self - consistency in the chihara framework .
second , even in the two - electron calculation , the response of the outer two electrons is still aware of the two frozen core states tied to each atom through their screening of the nuclear potential .
finally , these calculations are based upon explicit simulations of the real - time electron dynamics of a bulk supercell of warm dense beryllium rather than a phenomenological model of the response based upon a jellium plus average - atom picture . based upon our observation that the two - electron response roughly corresponds to @xmath47
, we can also extract a quantity akin to @xmath6 by differencing the four - electron and two - electron dsfs .
the effective @xmath47 and @xmath6 computed within tddft are illustrated in fig .
[ fig : comparison ] .
here we compare our tddft calculations to calculations done using state - of - the - art models for @xmath47 and @xmath6 .
the former is treated with an rpa - level model dielectric function with lifetime effects taken from four - electron dft - md calculations of the optical conductivity ; the mermin approximation - ab initio collision frequencies ( ma - aicf ) method in @xcite . the latter is calculated with the formalism developed in @xcite and a quantum mechanical average - atom ion - sphere model with slater exchange . as we are interested in studying energies relevant to the electronic response , we ignore @xmath7 , though it is necessarily present in the chihara - independent tddft calculation . .
( b ) illustrates the difference of the four- and two - electron tddft responses compared to an average atom treatment@xcite of @xmath6 .
[ fig : comparison ] ] examining the dispersion of the primary plasmon peak in fig .
[ fig : comparison]a , we see that tddft predicts a slight ( @xmath485 ev ) blue shift relative to the ma - aicf calculation @xmath49 and @xmath50 , whereas it predicts a stronger ( @xmath4810 ev ) red shift relative to ma - aicf at @xmath51 and @xmath52 .
we attribute this shift to exchange / correlation and band structure effects , not present in the ma - aicf dielectric function .
previously , comparisons of inelastic x - ray scattering spectra to rpa and lda in cold free electron metals ( na and al ) indicate a similar trend in which both lda and experiment are red shifted relative to rpa@xcite .
however , these calculations also indicate that the addition of lifetime effects to the lda are necessary to totally reconcile theory and experiment
. while non - adiabatic exchange correlation kernels are available for energy domain linear response tddft , no such time - domain exchange correlation potentials are currently available for real - time tddft .
as warm dense matter requires a large number of thermally occupied states such that energy domain tddft may become computationally prohibitive , warm dense matter may provide compelling motivation for the development and testing of these functionals .
considering the bound - free shoulder in fig .
[ fig : comparison]b , we see that tddft is generally in good agreement with the average - atom model treatment of @xmath6 with some minor differences .
such average atom models agree well with real - space green s function methods for cold solid beryllium @xcite .
we observe a trend opposite the free - free feature , in which there is a red shift of the tddft result at small angles , and a blue shift at large angles .
we expect that lda will do an increasingly poor job of describing the compton scattering limit at large @xmath38 due to its well - established self - interaction error .
applying a functional with some fraction of fock exchange should remedy this behavior and will be the subject of future investigations .
further , the tddft bound - free feature computed by differencing the two- and four - electron dsfs has a small negative dip below the 80 ev onset of the core feature .
it is difficult to determine whether this is due to core - polarization suppressing the response of the valence electrons in the four - electron calculation , or potentially an artifact of the different pseudization procedures used to generate the two paw potentials used for this comparison .
this motivates further investigations of the paw formalism applied to both real - time tddft and specifically , warm dense matter in which thermal effects will start to blur the line between core and valence electrons .
we presented a method for the direct calculation of the dsf for warm dense matter , independent of the chihara decomposition , by applying real - time tddft to configurations drawn from thermal mermin dft - md calculations .
comparison of our results with state - of - the - art models applied within the chihara picture illustrates some subtle differences between the two , though it generally supports the use of the chihara formalism as an inexpensive alternative to the very detailed and computationally intensive tddft calculations .
we anticipate that tddft may provide a powerful discriminating tool for arbitrating disagreements between these more phenomenological theories and experiment , especially as experimental data becomes more highly resolved .
our framework enables future explorations of systems in which the partition between bound and free electrons is more ambiguous .
it also provides a platform for studying the impact of recent foundational developments in dft at non - zero temperature@xcite . |
one of the most important ingredients in the search for `` signals '' of new phenomena in high energy particle physics experiments is a precise knowledge of the expectations from standard physics ; the `` background '' .
usually this involves perturbative calculations of differential cross sections within the standard model .
many such radiative corrections have been carried out and require the evaluation of one - loop integrals which arise directly from a feynman diagrammatic approach .
often these integrals need to be performed in an arbitrary dimension in order to isolate any infrared and ultraviolet divergences that may be present @xcite .
the basic one - loop tensor integral in @xmath3 dimensions for @xmath0 external particles scattering with outgoing momenta @xmath4 , @xmath0 internal propagators with masses @xmath5 and @xmath6 loop momenta in the numerator can be written , @xmath7 = \int \frac{d^{d}\ell}{i\pi^{d/2 } } \frac{\ell^{\mu_1}\ldots\ell^{\mu_m } } { ( \ell^2-m_1 ^ 2 ) ( ( \ell+q_1)^2-m_2 ^ 2 ) \cdots ( ( \ell+q_{n-1})^2-m_n^2)},\ ] ] where @xmath8 and , @xmath9 the scalar integral is denoted @xmath10 $ ] . in the standard approach to such integrals @xcite one utilises the fact that the tensor structure must be carried by the external momenta or the metric tensor @xmath11 . for example , the simplest non - trivial tensor integral in @xmath3 dimensions has a single loop momentum @xmath12 .
it reads , @xmath13 \equiv \sum_{j=1}^{n-1 } c_j p_j^{\mu } , \label{eq : formfactor}\ ] ] using momentum conservation to eliminate one of the momenta .
the formfactors @xmath14 are determined by multiplying both sides by all possible momenta @xmath15 and rewriting @xmath16 as a difference of propagator factors , @xmath17,\ ] ] thereby reducing the tensor integral to a sum of scalar integrals , @xmath18- \i^{d,(i)}_{n-1}[1]+(m_{i+1}^2-m_{i}^2+q_{i-1}^2-q_{i}^2)\i^{d}_{n}[1 ] \right].\ ] ] here , @xmath19 $ ] represents the `` pinched '' loop integral with the @xmath1 propagators remaining after the @xmath20th propagator factor has cancelled .
the formfactors are then obtained by algebraically solving the system of @xmath1 equations .
this introduces the @xmath21 gram determinant , @xmath22 ( where @xmath20 and @xmath23 run over the @xmath1 independent momenta ) , into the denominator .
each formfactor is a sum over the scalar integrals present in the problem multiplied by a kinematic coefficient that may be singular at the boundary of phase space where the gram determinant vanishes .
typically , @xmath24}{\delta_n } + \sum_{m=1}^n \beta_m \frac{\i_{n-1}^{d,(m)}[1]}{\delta_n } , \label{eq : toy}\ ] ] where the sum is running over the @xmath6 possible pinchings and where @xmath25 and @xmath26 are coefficients composed of the kinematic variables .
since , in many cases , the formfactors @xmath14 are actually finite in this limit , there are large cancellations and there may be problems of numerical stability .
the basic approach has been modified in a variety of ways , including the introduction of a system of @xmath1 reciprocal vectors @xmath27 ( and the associated second rank tensor @xmath28 playing the role of @xmath11 ) to carry the tensor structure @xcite where , @xmath29 so that , @xmath30 this simplifies the identification of the formfactor coefficients , but does not eliminate the gram determinants .
in fact , in both approaches , the number of gram determinants generated is equal to the number of loop momenta in the numerator of the original integral .
a different approach has been suggested by davydychev @xcite , who has identified the formfactors directly as loop integrals in differing numbers of dimensions and with the loop propagator factors raised to different powers .
tarasov @xcite has obtained recursion relations for one - loop integrals of this type , so that a complete reduction is possible . however , in relating the formfactor loop integrals to ordinary scalar loop integrals in 4 ( or close to 4 ) dimensions , the gram determinant once again appears in the denominator as in equation ( [ eq : toy ] ) .
finally , bern , kosower and dixon have used the feynman parameter space formulation for loop integrals to derive explicit results for the scalar integrals including the scalar pentagon @xcite .
the formfactors of the momentum space decomposition are directly related to feynman parameter integrals with one or more feynman parameters in the numerator .
one can see this by introducing the auxiliary momentum @xmath31 , @xmath32 so that after integrating out the loop momentum , the tensor integral for a single loop momentum in the numerator can be expressed in terms of the external momenta @xmath33 , @xmath13 \to \i^d_n[\p^{\mu } ] \equiv - \sum_{i=1}^{n-1 } \i^d_n[x_{i+1}]q_i^\mu.\ ] ] here , @xmath34 $ ] represents the scalar integral with a single factor of @xmath35 in the numerator . by comparing with equation ( [ eq : formfactor ] ) ,
we see that , @xmath36.\ ] ] differentiating with respect to the external kinematic variables , yields relations between integrals with polynomials of feynman parameters in the numerator and the usual scalar integrals .
once again , the gram determinant appears in the denominator , and the final result for the formfactor @xmath14 combines @xmath0-point integrals with the pinched @xmath1-point integrals as in equation ( [ eq : toy ] ) .
the presence of the gram determinant is , in some ways , no great surprise . in the limit @xmath37
, the @xmath1 momenta no longer span an @xmath1-dimensional space , the @xmath1 equations of the passarino - veltman approach are no longer independent and the decomposition is invalid .
stuart @xcite has made modifications to the basic approach to account for this , the main observation being that for @xmath38 , the scalar @xmath0-point integral can be written as a sum of scalar @xmath1-point integrals . as a consequence ,
there are large cancellations between scalar integrals with differing numbers of external legs in the kinematic limit of vanishing gram determinants . for loop corrections to processes such as quarkonium decay , where the two heavy quarks are considered to travel collinearly and share the quarkonium momentum
, one can eliminate the gram determinant singularities completely using the method of stuart @xcite .
however , for more general scattering processes where the collinear limit may be approached , but is not exact , the numerical problems as @xmath37 remain . in this paper
, we wish to address the problem by combining the scalar integrals into functions that are well behaved in the @xmath37 limit @xcite , so that the formfactors are given by , @xmath39 + \sum_{m=1}^n \beta^\prime_m
\i_{n-1}^{d,(m)}[1]\right ) + { \rm finite},\ ] ] where `` finite '' represents terms that are manifestly well behaved as @xmath37 , and the grouping @xmath40 vanishes with @xmath41 . such groupings
combine a variety of dilogarithms , logarithms and constants together in a non - trivial way .
in fact , for higher rank tensor integrals , with higher powers of gram determinants in the denominator , it becomes even more desirable to organise the scalar integrals in this way .
it is possible to construct these well behaved groupings by brute force , making a taylor expansion of the scalar integrals in the appropriate limit .
however , as we will show , they systematically and naturally arise by considering the scalar integral in @xmath42 or higher dimensions dimensions . ] and/or by differentiating the scalar integrals with respect to the external kinematic variables .
our approach is therefore to re - express the formfactor coefficients in terms of functions that are finite as @xmath37 , explicitly cancelling off factors of the determinant where possible .
the one - loop matrix elements for physical processes will then depend on these finite combinations , which can themselves be expanded as a taylor series to obtain the required numerical precision .
an additional improvement is that the physical size of the resulting expression is significantly reduced because the scalar integrals have been combined to form new , more natural functions .
of course , one loop amplitudes may also contain spurious singularities other than those directly arising from gram determinants .
such singularities may occur as one or more of the external legs becomes lightlike or as two external momenta become collinear .
our approach has the advantage of avoiding such ` fake ' singularities .
since the new finite functions are obtained by differentiating the scalar integrals , they can not contain additional kinematic singularities beyond those already present in the scalar integral .
this helps to ensure that only genuine poles - those allowed at tree level - are explicitly present in the one - loop matrix elements . once again
, this helps to reduce the size of the expressions for the amplitudes . in this paper , we address how such finite functions might be generated for arbitrary processes .
we will closely follow the notation and approach of bern , kosower and dixon to derive relationships between integrals with polynomials of feynman parameters in the numerators as well as between integrals with fewer parameters but in higher dimensions .
the basic definitions and notations are introduced in section 2 and the recursive relations for integrals with up to four feynman parameters in the numerator are presented along with the dimension shifting relation of @xcite .
these expressions are valid for arbitrary internal and external masses and for general kinematics .
however , making sense of these relations with respect to the singular limit depends on the actual integral itself ; i.e. on @xmath0 and the specific values of the kinematic variables .
the remainder of our paper describes a series of explicit realisations of the three , four and five point integrals relevant for the one - loop corrections for the decay of a virtual gauge boson into four massless partons @xcite . in section 3
, we consider tensor three point integrals for all internal masses equal to zero , but for general external kinematics .
section 4 describes tensor box integrals with one and two external massive legs while the tensor pentagon integral is explicitly worked through in section 5 .
our main results are summarised in section 6 , while some explicit results for three and four point tensor integrals are collected in the appendices .
the basic integral we wish to work with is the rescaled one - loop integral in @xmath3 dimensions differs from that of @xcite by a factor of @xmath43 . ] , @xmath44 = ( -1)^n\gamma(n - d/2)\int_0 ^ 1 d^nx_i \delta(1-\sum_i x_i ) \left [ \sum_{i , j=1}^n s_{ij}x_ix_j \right ] ^{d/2-n}.\ ] ] here , the feynman parameters @xmath35 have been introduced and the loop momentum has been integrated out .
the symmetric matrix @xmath45 contains all the process specific kinematics and reads , @xmath46 following closely the steps of @xcite we perform the projective transformation @xcite , @xmath47 and introduce the constant matrix @xmath48 such that , @xmath49 the parameters @xmath50 can be related to the kinematic variables present in the problem , while @xmath48 is considered independent of the @xmath50 .
provided that all @xmath50 are real and positive we find , @xmath44 = ( -1)^n\gamma(n - d/2)\int_0 ^ 1 d^nu_i \delta(1-\sum_i u_i ) \left ( \prod_{j=1}^n \alpha_j \right ) \left ( \sum_{j=1}^n \alpha_j u_j \right)^{n - d } \left [ \sum_{i , j=1}^n \rho_{ij}u_iu_j \right ] ^{d/2-n}.\ ] ] it is useful to rescale the integral , @xmath51 so that in @xmath52 , the only dependence on the parameters @xmath50 lies in the factor @xmath53 .
differentiating with respect to @xmath50 brings down a factor of the rescaled feynman parameter @xmath54 under the integral , @xmath55 = \frac{1}{(n - d)}\frac{\partial \ihat^d_n[1]}{\partial \alpha_i } , \label{eq : deriv}\ ] ] where the notation is obvious . with repeated differentiation ,
it is possible to generate all integrals with feynman parameters in the numerator .
the second step of bern , kosower and dixon s work @xcite is to relate the @xmath0-point integral with one feynman parameter in the numerator to a collection of scalar @xmath0 and @xmath1-point integrals , @xmath56 = \frac{1}{2n_n } \sum_{m=1}^n \left ( \frac{\ghat_i\ghat_m}{\gnhat } - \eta_{im } \right ) \ihat_{n-1}^{d\,(m)}[1 ] + \frac{\ghat_i}{\gnhat}\ihat_n^d[1 ] , \label{eq : inxibdk}\ ] ] where coincides with @xmath57 of @xcite .
] , @xmath58 and , @xmath59 in sections 3 , 4 and 5 , explicit examples using this notation will be worked through .
equation ( [ eq : inxibdk ] ) is the analogue of the formfactor reduction in momentum space of @xcite and is easily obtained by integration by parts .
the summation over @xmath6 represents all possible pinchings of the @xmath0-point graph to form @xmath1-point integrals .
as expected , we immediately see the appearance of the gram determinants in the denominator .
equations ( [ eq : deriv ] ) and ( [ eq : inxibdk ] ) are equivalent and since , with a few notable exceptions , the scalar integrals have a taylor expansion around @xmath60 , the act of differentiation will not usually introduce a singular behaviour .
therefore , we might expect that the @xmath0-point and @xmath1-point integrals combine in such a way that the @xmath61 limit is well behaved .
we can see how this happens by considering the @xmath0-point integral in @xmath42 dimensions @xcite , @xmath62 & = & \frac{1}{(n-1-d/2)(n - d)(n - d-1 ) } \rho_{ij } \frac{\partial^2 \ihat^d_n[1]}{\partial \alpha_j\partial \alpha_i } \nonumber \\ & = & \ { 1\over ( n - d-1 ) } { 2n_n\over\gnhat } \left(\ihat_n^d[1 ] \ + \ { 1\over 2n_n } \sum_{m=1}^n \ghat_m \ihat_{n-1}^{d \ , ( m)}[1 ] \right ) , \label{eq : d - change}\end{aligned}\ ] ] so that , @xmath56 \ = \ { 1\over 2 n_n } \left ( ( n - d-1 ) \ghat_i { \ihat_n^{d+2}}[1 ] \ - \sum_{m=1}^n \eta_{im } \ihat_{n-1}^{d\,(m)}[1]\right ) .
\label{eq : inai}\ ] ] it is important to note that there are no gram determinants visible in this equation . they have all been collected into the higher dimensional @xmath0-point integral .
it is clear that if @xmath63 $ ] is finite as @xmath61 , then so is @xmath64 $ ] and therefore so is @xmath65 $ ] .
this confirms that the apparent divergence as @xmath61 is fake .
furthermore , @xmath64 $ ] is an excellent candidate for a finite function - it is well behaved as the gram determinant vanishes and is easily related to the feynman parameter integrals via equation ( [ eq : d - change ] ) . of course
, it may still be divergent as @xmath66 and the dimensionally regulated poles remain to be isolated . by applying the derivative approach , we can easily extend this to two or more feynman parameters in the numerator , @xmath67 & = & \frac{1}{(n - d-1 ) } \frac{\partial \ihat^d_n[a_i]}{\partial \alpha_j } \nonumber \\ & = & { 1\over 2 n_n } \left ( \ghat_i \frac{\partial { \ihat_n^{d+2}}[1]}{\partial \alpha_j } + \frac{\partial \ghat_i}{\partial \alpha_j } { \ihat_n^{d+2}}[1 ] \ - \frac{1}{(n - d-1)}\sum_{m=1}^n \eta_{im } \frac{\partial \ihat_{n-1}^{d\,(m)}[1]}{\partial \alpha_j}\right ) .
\nonumber\end{aligned}\ ] ] using equation ( [ eq : deriv ] ) and the identity , @xmath68 we see that , @xmath69 = \frac{1}{2n_n } \left ( ( n - d-2 ) \ghat_i \ihat_n^{d+2}[a_j ] + \eta_{ij } \ihat_n^{d+2}[1 ] - \sum_{m=1}^n \eta_{im}\ihat^{d\,(m)}_{n-1}[a_j ] \right ) .
\label{eq : inaiaj}\ ] ] note that @xmath70 $ ] does not depend on @xmath71 , and therefore , @xmath72
\sim \frac{\partial \ihat^{d\,(j)}_{n-1}[1]}{\partial\alpha_j } \equiv 0.\ ] ] consequently , the @xmath73 term in the summation vanishes .
differentiation has not produced any new gram determinants and we can treat these integrals as new well behaved building blocks , or substitute for them using equation ( [ eq : inai ] ) with @xmath3 replaced by @xmath42 , @xmath74 & = & \frac{1}{4n_n^2 } \biggl ( ( n - d-2)(n - d-3 ) \ghat_i\ghat_j \ihat_n^{d+4}[1 ] + 2n_n\eta_{ij } \ihat_n^{d+2}[1]\nonumber \\ & & -(n - d-2)\sum_{m=1}^n \ghat_i\eta_{jm}\ihat^{d+2\,(m)}_{n-1}[1 ] - 2n_n\sum_{m=1}^n \eta_{im}\ihat^{d\,(m)}_{n-1}[a_j ] \biggr ) .
\nonumber\end{aligned}\ ] ] the scalar integrals for @xmath75 dimensions can be obtained recursively from equation ( [ eq : d - change ] ) . replacing the factors of @xmath50 in equations ( [ eq : inai ] ) and
( [ eq : inaiaj ] ) and the analogous equations for three and four feynman parameters in the numerator , we find , @xmath76 & = & { 1\over 2 n_n } \left ( ( n - d-1 ) \g_i { \i_n^{d+2}}[1 ] \ - \sum_{m=1}^n \eta_{im}\alpha_i\alpha_m \i_{n-1}^{d\,(m)}[1]\right ) , \label{eq : inxi } \\ \i^d_n[x_i x_j ] & = & \frac{1}{2n_n } \biggl ( ( n - d-2 ) \g_i \i_n^{d+2}[x_j ] + \eta_{ij } \alpha_i\alpha_j\i_n^{d+2}[1 ] \nonumber \\ & & \qquad\qquad - \sum_{m=1}^n \eta_{im}\alpha_i\alpha_m\i^{d\,(m)}_{n-1}[x_j ] \biggr ) , \label{eq : inxixj } \\ \i^d_n[x_i x_j x_k ] & = & \frac{1}{2n_n } \biggl ( ( n - d-3 ) \g_i \i_n^{d+2}[x_j x_k ] + \eta_{ij}\alpha_i\alpha_j \i_n^{d+2}[x_k ] + \eta_{ik}\alpha_i\alpha_k \i_n^{d+2}[x_j]\nonumber \\ & & \qquad\qquad - \sum_{m=1}^n \eta_{im}\alpha_i\alpha_m\i^{d\,(m)}_{n-1}[x_jx_k ] \biggr ) , \label{eq : inxixjxk } \\
\i^d_n[x_i x_j x_k x_l ] & = & \frac{1}{2n_n } \biggl ( ( n - d-4 )
\g_i \i_n^{d+2}[x_j x_k x_l ] + \eta_{ij}\alpha_i\alpha_j \i_n^{d+2}[x_kx_l ] + \eta_{ik}\alpha_i\alpha_k \i_n^{d+2}[x_jx_l ] \nonumber \\ & & \qquad\qquad + \eta_{il}\alpha_i\alpha_l \i_n^{d+2}[x_jx_k ] - \sum_{m=1}^n \eta_{im}\alpha_i\alpha_m\i^{d\,(m)}_{n-1}[x_jx_kx_l ] \biggr ) .
\label{eq : inxixjxkxl}\end{aligned}\ ] ] once again , no gram determinants are apparent and these equations may be solved by recursive iteration .
these are our main results and their use will be made clear with the explicit examples in the following sections . before proceeding to the explicit examples , we note that the full tensor structure in momentum space is simply obtained from the feynman parameter integrals by introducing the auxiliary momentum @xmath31 defined in equation ( [ eq : p ] ) . with an obvious notation ( and after integration of the loop momentum ) the tensor integrals can be written , @xmath77 & \to & \i_n^d[\p^\mu],\nonumber \\
\i_n^d[\ell^\mu\ell^\nu ] & \to & \i_n^d[\p^\mu\p^\nu]- \frac{1}{2}\i_n^{d+2}[g^{\mu\nu}],\nonumber \\ \i_n^d[\ell^\mu\ell^\nu\ell^\rho ] & \to & \i_n^d[\p^\mu\p^\nu\p^\rho ] -\frac{1}{2}\i_n^{d+2}[\ { g\p \}^{\mu\nu\rho}],\nonumber \\ \i_n^d[\ell^\mu\ell^\nu\ell^\rho\ell^\sigma ] & \to & \i_n^d[\p^\mu\p^\nu\p^\rho\p^\sigma ] -\frac{1}{2}\i_n^{d+2}[\{g\p\p\}^{\mu\nu\rho\sigma } ] + \frac{1}{4}\i_n^{d+4}[\{gg\}^{\mu\nu\rho\sigma } ] , \nonumber\end{aligned}\ ] ] where @xmath78 is the usual passarino - veltman notation @xcite , and indicates a sum over all possible permutations of lorentz indices carried by @xmath79 .
for example , @xmath80 throughout the next sections , we make the simplifying choice that @xmath81 .
such integrals are relevant for a wide range of qcd processes involving loops of gluons or massless quarks .
the approach can be straightforwardly extended to include non - zero masses @xcite .
the strategy is to isolate the ultraviolet and infrared poles from the tensor integrals , leaving the finite remainder in the form of groups of terms that are well behaved in all of the kinematic limits . in real calculations where groups of tensor integrals are combined ,
this grouping will often cancel as a whole .
alternatively , if the kinematic coefficient allows , the determinant can be cancelled off for all of the terms in the function .
this approach is well suited to treatment by an algebraic manipulation program , once the raw integrals have been massaged to isolate the poles in @xmath82 and to group the terms . as we will show in the explicit examples , this is usually straightforward .
in processes where the internal lines are massless , there are only three types of triangle graph described by the number of massive external legs . for the one loop corrections to five parton scattering @xcite , only the graphs with one and two massless legs occur . for processes involving a gauge boson such as @xmath83 partons @xcite , graphs with all external legs massive or off - shell contribute .
we first consider triangle integrals with exiting momenta @xmath84 , @xmath85 and @xmath86 as shown in fig .
[ fig : tripin ] and all internal masses equal to zero , @xmath81 . throughout , we systematically eliminate @xmath86 ( and the feynman parameter @xmath87 ) using momentum conservation so that @xmath88 , @xmath89 and , @xmath90 the full tensor structure with up to three loop momenta in the numerator can therefore be derived from loop integrals with up to three powers of @xmath91 or @xmath92 in the numerator . as a first step , we consider the general case , @xmath93 , where the scalar integral in four dimensions is known to be finite . here
the @xmath50 parameters can be determined by , @xmath94 while , @xmath95 from the definition of the matrix @xmath96 , we see that , @xmath97 the variables @xmath57 always appear in the following combinations , @xmath98 the scalar triangle integral for all external masses non - zero is finite in four dimensions @xcite , or @xmath99 dimensions , we omit the superscript @xmath3 . ]
@xmath100 = \frac{1}{\sqrt{-\delta_3 } } \left(\log(a^+a^-)\log\left(\frac{1-a^+}{1-a^-}\right ) + 2\li2(a^+)-2\li2(a^-)\right ) , \label{eq : i3}\ ] ] where @xmath101 is the usual dilogarithm function and @xmath102 are two roots of a quadratic equation , @xmath103 although @xmath104 $ ] appears to diverge as @xmath105 , this is not the case .
as noted by stuart @xcite , in this limit , the triangle graph reduces to a sum of bubble graphs , @xmath106 = \frac{2}{s_{12}+p_1 ^ 2-p_2 ^ 2}\log\left(\frac{s_{12}}{p_2 ^ 2}\right ) + \frac{2}{s_{12}+p_2 ^ 2-p_1 ^ 2}\log\left(\frac{s_{12}}{p_1 ^ 2}\right),\ ] ] and there is a well behaved taylor series in @xmath107 .
the tensor integral can be easily written in terms of higher dimensional scalar integrals and bubble scalar integrals using eqs .
( [ eq : inxi][eq : inxixjxk ] ) . for one feynman parameter in the numerator , this gives , @xmath108 = -(1-\e ) \g_i \i_3^{d=6 - 2\e}[1 ] - \frac{1}{2 } \sum_{m=1}^3 \eta_{im } \alpha_i \alpha_m \i_2^{d=4 - 2\e ( m)}[1 ] .\ ] ] immediately a problem is apparent the coefficient of the scalar integral in @xmath109 dimensions , @xmath110 , is singular as one or more of the external momenta become lightlike .
although the divergence as the gram determinant vanishes has been removed , it appears to have been replaced by a divergence as the invariants vanish . ] .
however , these divergences cancel between the triangle and bubble contributions and the tensor integral itself is well behaved and finite in all kinematic limits and is therefore a better choice for a finite function . in fact , since the triangle scalar integral is finite in @xmath111 dimensions , it is convenient to generate the tensor structure directly from derivatives of the scalar integral .
however , in order to use equation ( [ eq : inxibdk ] ) , we also need the two point integral for external momentum @xmath112 ( and internal masses @xmath81 ) , @xmath113 = \frac{\gamma(2-d/2)\gamma^2(d/2 - 1)}{\gamma(d-2 ) } ( -p^2)^{d/2 - 2 } , \label{eq : i2}\ ] ] for each of the three pinchings , @xmath114 and 3 shown in fig . [
fig : tripin ] . for pinching @xmath6 and @xmath115 , @xmath116 = \left(\frac{\alpha_m}{\prod_{i=1}^3
\alpha_i}\right ) \i_2^{(m)}[1 ] = \frac{c_\gamma}{\e ( 1 - 2\e ) } \left(\frac{\alpha_m}{\prod_{i=1}^3 \alpha_i}\right)^{1-\e},\ ] ] where the usual product of gamma functions obtained in one - loop integrals @xmath117 is given by , @xmath118 rewriting equation ( [ eq : inxibdk ] ) for the case @xmath99 , @xmath119 and @xmath120 and adding , @xmath121 = 0,\ ] ] we see , @xmath122 & = & \frac{1}{\g3hat } \left ( \ghat_3 \ihat_3[1 ] + \frac{1}{2 } \sum_{m=1}^3 \left ( \ghat_3\ghat_m - \eta_{3m}\g3hat \right ) \frac{\alpha_m}{\prod \alpha } \left ( \i_{2}^{(m)}[1 ] - \i_{2}^{(2)}[1 ] \right ) \right)\nonumber \\ & = & \frac{1}{\g3hat } \left ( \ghat_3 \ihat_3[1 ] -\frac{\ghat_2}{\alpha_3}\log\left(\frac{\alpha_2}{\alpha_1}\right ) + 2\log\left(\frac{\alpha_2}{\alpha_3}\right ) \right ) + { \cal o}(\e).\end{aligned}\ ] ] alternatively , this could be obtained by differentiation of equation ( [ eq : i3 ] ) . by trivial replacement of factors of @xmath25 , we find , @xmath123 & = & \frac{1}{\delta_3 } \biggl(p_1 ^ 2(s_{12}+p_2 ^ 2-p_1 ^ 2)\i_3[1 ] + ( p_1 ^ 2+p_2 ^ 2-s_{12})\log\left(\frac{s_{12}}{p_2 ^ 2}\right ) -2p_1 ^ 2\log\left(\frac{s_{12}}{p_1 ^ 2}\right)\biggr ) .
\label{eq : i3x3}\end{aligned}\ ] ] integrals with higher powers of feynman parameters can now be generated by direct differentiation of @xmath124 $ ] , @xmath125 = -\frac{1}{(n+1 ) } \left(\prod \alpha \right ) \alpha_i\alpha_3^n \frac{\partial}{\partial \alpha_i } \ihat_3[a_3^n].\ ] ] all of these functions will be finite in the @xmath66 limit , and can be considered as building blocks in constructing the tensor structures for box and pentagon integrals .
in fact , because they are obtained by differentiating a function well behaved as @xmath105 , they are also finite in this limit .
therefore , they tie together the dilogarithms from the triangle integrals and the logarithms from bubble integrals in an economical and numerically very stable way .
we also see that they are directly generated in tensor structures for box graphs ( equations ( [ eq : inxi][eq : inxixjxkxl ] ) with @xmath126 ) and will naturally cancel in feynman diagram calculations involving both triangle and box graphs . for general calculations with @xmath127 and @xmath128 , we introduce the notation , @xmath129 , \qquad \lc{2n-1}(p_1,p_2 ) = \i_3[x_3^n ] , \qquad \lc{2n}(p_1,p_2 ) = \i_3[x_1x_3^n],\ ] ] for @xmath130 .
the symmetry properties of the triangle function imply that the analogous functions for @xmath131 ( or @xmath132 ) are just obtained by exchanging @xmath84 and @xmath85 . in dealing with box graphs , integrals with @xmath87 in the numerator
will naturally arise . in these cases ,
we systematically eliminate them using @xmath133 .
explicitly , we find , @xmath134 & = & \frac{1}{2\delta_3 } \biggl ( 2p_2 ^ 2(s_{12}+p_1 ^ 2-p_2 ^ 2)\i_3[x_3 ] + p_1 ^ 2(s_{12}+p_2 ^ 2-p_1 ^ 2)\i_3[x_1 ] \nonumber \\ & & \qquad\qquad -p_1 ^ 2p_2 ^ 2 \i_3[1 ] - p_2 ^ 2\log\left(\frac{s_{12}}{p_2 ^ 2}\right ) + p_1 ^ 2+p_2 ^ 2-s_{12 } \biggr ) , \label{eq : i3x1x3 } \\
\i_3[x_3 ^ 2 ] & = & \frac{1}{2\delta_3 } \biggl ( 3p_1 ^ 2(s_{12}+p_2 ^ 2-p_1 ^ 2)\i_3[x_3 ] + p_1 ^ 4 \i_3[1 ] - ( s_{12 } -p_2 ^ 2)\log\left(\frac{s_{12}}{p_2 ^ 2}\right ) -2p_1 ^ 2 \biggr ) , \label{eq : i3x3x3 } \\
\i_3[x_1x_3 ^ 2 ] & = & \frac{1}{6\delta_3 } \biggl ( 4p_2 ^ 2(s_{12}+p_1 ^ 2-p_2 ^ 2)\i_3[x_3 ^ 2 ] + 6p_1 ^ 2(s_{12}+p_2 ^ 2-p_1 ^ 2)\i_3[x_1x_3 ] \nonumber \\ & & \qquad\qquad -3p_1 ^ 2p_2 ^ 2 \i_3[x_3 ] + p_1 ^ 4 \i_3[x_1 ] -p_2 ^ 2\log\left(\frac{s_{12}}{p_2 ^ 2}\right ) + p_2 ^ 2-s_{12 } \biggr ) , \label{eq : i3x1x3x3 } \\
\i_3[x_3 ^ 3 ] & = & \frac{1}{3\delta_3 } \biggl ( 5p_1 ^ 2(s_{12}+p_2 ^ 2-p_1 ^ 2)\i_3[x_3 ^ 2 ] + 2p_1 ^ 4 \i_3[x_3 ] - ( s_{12 } -p_2 ^ 2)\log\left(\frac{s_{12}}{p_2 ^ 2}\right ) - p_1 ^ 2 \biggr ) .
\label{eq : i3x3x3x3}\end{aligned}\ ] ] by expanding as a series in @xmath107 , these functions can be evaluated near the singularity with arbitrary precision .
for example , @xmath135 & = & \frac{(s_{12}+p_1 ^ 2-p_2 ^ 2)}{3s_{12}(s_{12}-p_1 ^ 2-p_2 ^ 2 ) } -\frac{2p_1 ^ 4}{3s_{12}(s_{12}-p_1 ^ 2-p_2 ^ 2)^2 } \log\left(\frac{s_{12}}{p_1 ^ 2}\right)\nonumber \\ & & \qquad
+ \frac{(3s_{12}-3p_1 ^
2+p_2 ^ 2)}{6s_{12}(s_{12}-p_1 ^ 2-p_2 ^ 2 ) } \log\left(\frac{s_{12}}{p_2 ^ 2}\right).\nonumber\end{aligned}\ ] ] to illustrate this , fig .
[ fig : lcn ] shows the various functions at a specific phase space point , @xmath136 and letting @xmath137 vary in such a way that @xmath105 .
this corresponds to @xmath138 .
we see that as this limit is approached , the numerical evaluation of the function using double precision ( an intrinsic numerical precision , @xmath139 , of roughly @xmath140 ) becomes uncertain .
for this particular phase space point , functions with a single gram determinant in the denominator ( @xmath141 ) remain stable until @xmath142 while those with more powers of the gram determinant break down correspondingly sooner - at @xmath143 for @xmath144 and @xmath145 and @xmath146 for @xmath147 and @xmath148 . in general ,
numerical problems typically occur when @xmath149 where @xmath150 is the number of gram determinants in the denominator of the function .
other phase space points yield a similar behaviour .
the unstable points represent a rather small proportion of the allowed phase space .
however , problems may arise using adaptive monte carlo methods such as vegas @xcite where the phase space is preferentially sampled where the matrix elements are large . finding an anomalously high value for the matrix elements in a region of instability would cause the monte carlo integration to focus on that region giving unpredictable results .
of course , these instabilities could be handled by a brute force increase in numerical precision .
while possible , this has the disadvantage of producing significantly slower code , and , since in all cases , the approximate form obtained by making a taylor expansion about @xmath151 and keeping only the constant term works well where the numerical instabilities begin , this is not an attractive solution . in fact
, the approximation is reliable for @xmath152 .
explicit forms for the approximations are collected in appendix c. we now turn to the scalar triangle integrals in higher dimensions .
they appear in the @xmath153 part of the general lorentz structure and recursively in the determination of @xmath154 $ ] . unlike the triangle in four dimensions ,
these integrals are ultraviolet divergent due to the presence of the various pinchings - bubble integrals .
a function that can usefully be used as a building block of matrix element calculations , must be finite as both @xmath105 _ and _ @xmath66 and we must first isolate the poles in @xmath82 .
although equation ( [ eq : d - change ] ) suggests that the ultraviolet pole structure involves @xmath107 , this is easily shown not to be the case .
adding terms proportional to , @xmath155 = 0,\ ] ] to equation ( [ eq : d - change ] ) for @xmath119 , @xmath115 we see , @xmath156 = -\frac{1}{(1-\e)}\frac{1}{\g3hat }
\left ( \ihat_3[1 ] + \frac{1}{2 } \sum_{m=1}^3 \frac{\g_m}{\prod \alpha } \left(\i_2^{(m)}[1 ] - \i_2^{(2)}[1]\right ) + \frac{\g3hat}{2\prod \alpha } \i_2^{(2)}[1]\right),\ ] ] where the divergence as @xmath66 lies exclusively in the last term . reinserting the factors of @xmath25 and the definition of @xmath157 $ ] in @xmath115 dimensions we find ,
@xmath158 = \lc{1s}(p_1,p_2 ) -\frac{1}{2}\left(\frac{(-s_{12})^{-\e}}{\e}+3\right ) c_\gamma , \label{eq : i36}\ ] ] where , @xmath159 -p_1 ^ 2(s_{12}+p_2 ^ 2-p_1 ^ 2 ) \log\left(\frac{s_{12}}{p_1 ^ 2}\right)\nonumber \\ & & \qquad -p_2 ^ 2(s_{12}+p_1 ^ 2-p_2 ^ 2 ) \log\left(\frac{s_{12}}{p_2 ^ 2}\right)\biggr)\nonumber \\ & = & \frac{1}{2 } \left ( p_1 ^ 2 \i_3[x_1 ] + p_2 ^ 2 \i_3[x_3 ] \right ) .
\label{eq : lc1s}\end{aligned}\ ] ] in a similar fashion , the @xmath82 pole structure can be removed from the triangle scalar integral in @xmath160 and @xmath161 dimensions yielding two more functions that are finite as both @xmath162 and @xmath66 .
explicitly we find , @xmath163 & = & \lc{2s}(p_1,p_2 ) -\frac{(p_1 ^ 2+p_2 ^ 2+s_{12})}{24 } \left(\frac{(-s_{12})^{-\e}}{\e } + \frac{19}{6}\right ) c_\gamma , \label{eq : i38 } \\
\i^{d=10 - 2\e}_3[1 ] & = & \lc{3s}(p_1,p_2 ) -\frac{(p_1 ^ 4+p_2 ^ 4+s_{12}^2+p_1 ^ 2p_2 ^ 2+p_1 ^ 2s_{12}+p_2 ^ 2s_{12})}{360 } \left(\frac{(-s_{12})^{-\e}}{\e}+\frac{17}{5}\right ) c_\gamma,\nonumber \\ \label{eq : i310}\end{aligned}\ ] ] where the finite functions are defined by , @xmath164 although these functions have been obtained via equation ( [ eq : d - change ] ) , they are still related to derivatives of the basic scalar integral in @xmath115 , and are therefore finite in the @xmath105 limit .
we can see this by examining the same phase space point as before , @xmath136 and varying @xmath107 .
as expected , the @xmath165 show a similar behaviour to the @xmath166 functions - numerically breaking down at larger and larger values of @xmath107 as the number of gram determinants increases , and being well described by the first term in the taylor expansion as this happens . for completeness
, we collect the limiting approximations in appendix c. for triangle loop integrals with three loop momenta in the numerator , it is also necessary to know the @xmath2 integral with a single feynman parameter in the numerator . rather than differentiating the ultraviolet divergent
@xmath167 $ ] , we can evaluate it in terms of the @xmath115 tensor integrals of section 3.1.1 . using ( [ eq : inaiaj ] ) for @xmath115
, we see that , @xmath168+\ihat_3[a_3a_j ] & = & \frac{1}{2}\biggl ( -(3 - 2\e)(\ghat_1+\ghat_3)\ihat_3^{d=6 - 2\e}[a_j ] + ( \eta_{1j}+\eta_{3j})\ihat_3^{d=6 - 2\e}[1]\nonumber \\ & & \qquad -\sum_{m=1}^3 ( \eta_{1m}+\eta_{3m})\ihat_2^{(m)}[a_j ] \biggr ) , \nonumber\end{aligned}\ ] ] which , for @xmath169 is of little interest . ]
simplifies using , @xmath170 the same equation simplifies the sum over bubble pinchings so that only @xmath171 contributes , while @xmath172 . restoring the factors of @xmath25 and using , @xmath173 = \i_2^{(2)}[x_3 ] = \frac{1}{2 } \left ( \frac{(-s_{12})^{-\e } } { \e } + 2 \right)c_\gamma,\ ] ] yields , @xmath174 = \frac{1}{3 } \biggl ( p_1 ^ 2 \i_3[x_1 x_j ] + p_2 ^ 2 \i_3[x_3 x_j ] \biggr ) - \frac{1}{6 } \left ( \frac{(-s_{12})^{-\e } } { \e } + \frac{8}{3 } \right)c_\gamma .\ ] ] later , we will see that constructing the tensor integrals for box graphs can also generate @xmath175 $ ] and higher dimensional triangle integrals with more parameters in the numerator . in each case
, we use similar tricks with equations ( [ eq : inxi][eq : inxixjxkxl ] ) to rewrite them in terms of the four dimensional integrals with the ultraviolet pole made explicit .
we will also be interested in triangle graphs where one or more of the external momenta is lightlike . here ,
we first focus on the case , @xmath176 . in @xmath115 dimensions
, we have the well known result , @xmath177 = \frac{c_\gamma}{\e^2 } \left ( \frac { ( \alpha_1\alpha_3)^\e - ( \alpha_1\alpha_2)^\e}{\alpha_3-\alpha_2 } \right ) , \ ] ] where the superscript indicates that only two of the three legs are massive . for this choice of kinematics , @xmath178 and @xmath179 so that the singular limit is @xmath180 . because @xmath181 makes no reference to @xmath182
, @xmath183 contains a row of zeroes , @xmath184 and therefore @xmath185 .
consequently , care is needed in applying the equations of section 2 .
in addition , since the scalar integral for three massive legs is finite ( and the results in the preceding subsections have been explicitly derived in @xmath99 ) , one can not just set @xmath176 .
in fact , it is easiest just to bypass the problem and generate the whole tensor structure by direct differentiation of the scalar integral with respect to @xmath186 and @xmath182 .
it is easy to see that , @xmath187 \sim a \ihat_3^{2m}[1 ] + b \frac{(\alpha_1\alpha_3)^\e}{\e } + c \log\left(\frac{\alpha_3}{\alpha_2}\right ) + d,\ ] ] where @xmath188 and @xmath3 are polynomials in @xmath189 and @xmath190 . unlike the all massive case discussed before ,
the scalar integral is singular as @xmath66 . as a general rule
it is not necessary to be particularly careful with double poles in @xmath82 , since they must either cancel or form the infrared poles of real matrix elements .
however , it is possible for the integrals to be multiplied by factors of @xmath82 - from expanding factors of dimension - and the resulting logarithms should occur in combinations that are finite as @xmath180 .
it is easy to see that , @xmath191 = \frac{\log\left(\frac{\alpha_3}{\alpha_2}\right ) } { \alpha_3-\alpha_2 } + { \cal o}(\e),\ ] ] is finite .
so , to tie the logarithms and constants together in combinations that are well behaved in the @xmath180 limit , we use the fact that derivatives of this function are also well behaved , and introduce the functions , @xmath192 \right),\ ] ] for @xmath193 . in terms of invariants ,
@xmath194 with , @xmath195 these functions , or functions closely related to them , have appeared in next - to - leading order matrix element calculations @xcite .
the explicit forms for @xmath196 $ ] appearing in the momentum expansion are well known and are collected in appendix a. although these functions are rather simple , they still contain numerical instabilities as @xmath197 .
this can be seen in fig .
[ fig : lcn00 ] where we show @xmath198 for the specific phase space point @xmath199 and let @xmath200 approach @xmath201 . while a single inverse powers of @xmath202 is handled correctly , higher powers cause problems .
as can be seen from the figure , a suitable approximation is obtained by the first term in the taylor expansion , @xmath203 for @xmath204 .
the other configuration of triangle graph that appears is where two of the momenta are lightlike , @xmath205 .
once again , the tensor structure can be generated by differentiation or canonical passarino - veltman reduction .
here , there is only one scale in the problem so there can be no logarithms and it is neither possible nor necessary to introduce well behaved functions .
the explicit forms for the feynman parameter integrals appearing in the momentum expansion are well known and for the sake of completeness are given in appendix a. this concludes our discussion of triangle graphs . for the case of three massive external legs ( and internal masses set equal to zero ) the four dimensional tensor integrals are finite as @xmath66 and are given by the functions @xmath206 defined in ( [ eq : i3],[eq : i3x3],[eq : i3x1x3][eq : i3x3x3x3 ] ) , while the ultraviolet divergent part @xmath153 part , @xmath167 $ ] is expressed in terms of a similar function ( @xmath207 ) with the pole isolated ( [ eq : lc1s ] ) . for the tensor structure in the simpler case with one lightlike leg , it is useful to group the logarithms and constants using the functions @xmath208 ( [ eq : lcmassless1 ] , [ eq : lcmassless ] ) . for the more general case where the internal masses are non - zero , the same procedure can be utilised .
the matrix @xmath96 has slightly more entries and there are more scales in the problem .
however , the grouping together of triangle graphs and bubble integrals into functions well behaved in the @xmath105 limit and the isolation of the ultraviolet singularities can be made explicit in the same way .
for one - loop corrections to five parton scattering , box graphs with at most one external leg occur .
however , for processes involving a massive vector boson and four massless partons , we can obtain box graphs with a second massive external leg by pinching together two of the partons .
there are two distinct configurations according to the positions of the massive legs ; the adjacent box graph and the opposite box graph .
the box graph is shown in fig .
[ fig : boxin ] for outgoing momenta @xmath84 , @xmath85 and @xmath86 . throughout this section
, we will assume that @xmath209 .
in the adjacent two mass case , @xmath210 and @xmath127 , while for the opposite box , @xmath211 and @xmath128 . unfortunately , the raw scalar integrals for these two cases behave rather differently .
the adjacent box is finite in the limit that @xmath212 , while the opposite box diverges as @xmath212 . in this section
, we work through these two configurations and rewrite the tensor integrals in terms of well behaved functions and explicit poles in @xmath82 .
we first consider the adjacent box with @xmath210 and all internal masses equal to zero . as in the triangle case ,
we systematically eliminate one of the momenta , @xmath213 , and one of the feynman parameters , @xmath214 , so that , @xmath215 the related integrals with @xmath216 and @xmath217 are obtained by @xmath218 ( and the indices @xmath20 and @xmath23 in @xmath50 and @xmath183 transform as @xmath219 and @xmath220 ) . for this kinematic configuration ,
the @xmath25 parameters are defined by , @xmath221 while , @xmath222 and , @xmath223 the coefficients @xmath224 always appear in the following combinations which are directly related to the conventional variables , @xmath225 the scalar integral in @xmath115 can be written @xcite , @xmath226 = \frac{c_\gamma}{\e^2}\left((\alpha_2\alpha_4)^\e + 2 ( \alpha_1\alpha_3)^\e -(\alpha_1\alpha_2)^\e -(\alpha_1\alpha_4)^\e\right ) + 2\ld{0}(p_1,p_2,p_3 ) + { \cal o}(\e),\ ] ] where , @xmath227 in constructing the tensor integrals in @xmath115 , we see from equations ( [ eq : inxi][eq : inxixjxkxl ] ) that the box integral in higher dimensions is needed .
in fact , in @xmath2 , the box integral is infrared and ultraviolet finite .
this can be seen by inspection of equation ( [ eq : d - change ] ) and noting that the pinchings with @xmath228 and 4 in the expression , @xmath226 + \sum_{m=1}^4 \ghat_m\ihat_3^{(m)}[1],\ ] ] are proportional to @xmath229 and , when combined with the appropriate @xmath230 factor , precisely cancel with the pole structure of the box integral .
the final pinching ( @xmath231 ) corresponds to the triangle graph with three massive external legs which is itself finite .
altogether , we find that the adjacent box integral in @xmath232 is , @xmath233 & = & -\frac{2s_{12}s_{23}}{\delta_4 } \left ( \ld{0}(p_1,p_2,p_3 ) + \frac{1}{2}\left(s_{123}+p_1 ^ 2-s_{23 } -\frac{2p_1 ^ 2s_{123}}{s_{12}}\right ) \lc{0}(p_1,p_{23})\right ) \nonumber \\ & \equiv & \ld{1s}(p_1,p_2,p_3 ) , \label{eq : ld1sadj}\end{aligned}\ ] ] where @xmath234 $ ] is defined in equation ( [ eq : i3 ] ) . because of the finiteness properties of the three mass triangle , we will find repeatedly that the @xmath231 pinching should be treated differently from the other three . for higher dimensions , we just reuse equation ( [ eq : d - change ] ) , noting that the triangle pinchings in @xmath235 reintroduce ultraviolet poles
. these can easily be isolated by adding and subtracting combinations of scalar integrals as in section 3.1.2 .
explicitly , @xmath236 & = & \ld{2s}(p_1,p_2,p_3 ) + \frac{c_\gamma}{6 } \left(\frac{(-s_{123})^{-\e}}{\e}+ \frac{11}{3}\right ) , \label{eq : i4adj8 } \\
\i_4^{d=10 - 2\e}[1 ] & = & \ld{3s}(p_1,p_2,p_3 ) + \frac{c_\gamma ( s_{123}+s_{12}+s_{23}+p_1 ^ 2)}{120 } \left(\frac{(-s_{123})^{-\e}}{\e}+ \frac{107}{30}\right ) , \label{eq : i4adj10 } \\
\i_4^{d=12 - 2\e}[1 ] & = & \ld{4s}(p_1,p_2,p_3 ) + \frac{c_\gamma p}{2520 } \left(\frac{(-s_{123})^{-\e}}{\e}+ \frac{129}{35}\right ) , \label{eq : i4adj12}\end{aligned}\ ] ] where , @xmath237 the finite parts of the higher dimension boxes are given by , @xmath238 the @xmath2 and @xmath160 box integrals explicitly appear in the momentum space tensor structure with one and two factors of @xmath153 respectively .
all of these integrals appear either directly or indirectly in the tensor box integrals of equations ( [ eq : inxi][eq : inxixjxkxl ] ) . once again ,
all of these functions are well behaved as @xmath239 and group a variety of dilogarithms , logarithms and constants together in a non - trivial way .
this is shown in fig .
[ fig : ldns ] for a particular point in phase space ; @xmath240 , @xmath241 , @xmath242 with @xmath200 varying so the @xmath239 limit is approached .
we see that although @xmath243 , with a single inverse power of the gram determinant , is numerically stable , the functions with more powers of gram determinant in the denominator break down at much larger values of @xmath244 . in all cases ,
the function is well approximated by the first term of the taylor expansion provided @xmath245 .
these approximations are collected in appendix c. armed with the integrals in @xmath246 , we return to the tensor integrals and use equations ( [ eq : inxi][eq : inxixjxkxl ] ) as the starting point . rewriting equation ( [ eq : inxi ] ) for @xmath126 , @xmath115 and noting that we have eliminated @xmath92 so that @xmath247 and 4 only , we have , @xmath248 = -\g_i \i_4^{d=6}[1 ] - \sum_{m=1}^4 \eta_{im}\alpha_i\alpha_m \i_3^{(m)}[1].\ ] ] the factor @xmath249 multiplying the triangle pinchings will always produce a factor of @xmath250 . for triangle graphs with at least one massless leg ( pinchings @xmath228 and 4 ) ,
the contribution is @xmath251 and will combine with similar poles from other feynman diagrams . on the other hand ,
the @xmath231 pinching ( corresponding to the triangle graph with momenta @xmath252 and @xmath253 flowing outwards ) is finite and , provided the value of @xmath254 is related kinematically to that triangle pinching , there may be a possibility of cancellation with other triangle feynman graphs .
however , for the case @xmath255 and @xmath231 , the associated invariant mass is @xmath201 .
this term can not combine with any other naturally generated triangle graph with the same kinematics .
therefore , we group this term with the box integral , by adding and subtracting , @xmath256 = \g_i \i_3^{(3)}[1],\ ] ] so that , @xmath248 = \g_i \ld{1}(p_1,p_2,p_3 ) - \sum_{m=1}^4 \eta_{im}\alpha_i\alpha_m \left(\i_3^{(m)}[1]-\i_3^{(3)}[1]\right ) , \label{eq : i4axi}\ ] ] with , @xmath257 + \i_3^{(3)}[1]\right).\ ] ] the only non - zero entries in @xmath258 for @xmath259 and @xmath260 are @xmath261 or @xmath262 corresponding to @xmath263 which _ is _ appropriate for @xmath264 $ ] . for integrals with more feynman parameters
it is convenient to introduce the following functions , @xmath265 + \i_3^{(3)}[x_{i_1}\ldots x_{i_{n-1}}]\biggr\},\ ] ] for @xmath266 and 4 . suppressing the arguments of @xmath267 , and using equations ( [ eq : inxixj]-[eq : inxixjxkxl ] ) , we find , @xmath268 & = & \g_i \ld{2j } - \eta_{ij}\alpha_i\alpha_j \biggl\ { \ld{1}+\i_3^{(3)}[1]\biggr\ } - \sum_{m=1}^4 \eta_{im}\alpha_i\alpha_m \biggl\{\i_3^{(m)}[x_j]-\i_3^{(3)}[x_j]\biggr\ } , \\
\label{eq : i4axixjxk } \i_4[x_ix_jx_k ] & = & \g_i \ld{3jk } - \frac{1}{2}\eta_{ij}\alpha_i\alpha_j \biggl\{\ld{2k}+\i_3^{(3)}[x_k]\biggr\ } - \frac{1}{2}\eta_{ik}\alpha_i\alpha_k \biggl\{\ld{2j}+\i_3^{(3)}[x_j]\biggr\ } \nonumber\\ & & \quad - \sum_{m=1}^4 \eta_{im}\alpha_i\alpha_m \biggl\{\i_3^{(m)}[x_jx_k]-\i_3^{(3)}[x_jx_k]\biggr\ } , \\ \label{eq : i4axixjxkxl } \i_4[x_ix_jx_kx_l ] & = & \g_i \ld{4jkl } - \frac{1}{3}\eta_{ij}\alpha_i\alpha_j \biggl\{\ld{3kl}+\i_3^{(3)}[x_kx_l]\biggr\ } - \frac{1}{3}\eta_{ik}\alpha_i\alpha_k \biggl\{\ld{3jl}+\i_3^{(3)}[x_jx_l]\biggr\ } \nonumber \\ & & - \frac{1}{3}\eta_{il}\alpha_i\alpha_l \biggl\{\ld{3jk}+\i_3^{(3)}[x_jx_k]\biggr\ } - \sum_{m=1}^4 \eta_{im}\alpha_i\alpha_m \biggl\{\i_3^{(m)}[x_jx_kx_l]-\i_3^{(3)}[x_jx_kx_l]\biggr\}. \nonumber \\\end{aligned}\ ] ] since we have systematically eliminated @xmath92 using the delta function , @xmath269 and @xmath270 run over @xmath271 and 4 .
this guarantees that the coefficients of the form @xmath272 are only non - zero for @xmath273 and @xmath274 ( or vice - versa ) .
in these cases , @xmath275 , which is again appropriate for the @xmath231 pinching to form a completely massive triangle , @xmath264 $ ] .
altogether , equations ( [ eq : i4axi],[eq : i4axixj][eq : i4axixjxkxl ] ) are sufficient to completely describe the tensor structure of the adjacent box . however , in order to determine the @xmath276 combinations , we need tensor integrals for @xmath2 dimensional box graphs with two or more feynman parameters
. these can be obtained from equation ( [ eq : inxixj ] ) once the @xmath2 box integral with a single feynman parameter , @xmath277 $ ] is known .
this can be derived by differentiating ( [ eq : deriv ] ) which indicates that @xmath277 $ ] is also finite as @xmath66 . to see this , we reuse equation ( [ eq : inxi ] ) and our usual trick of adding and subtracting combinations of the @xmath231 pinching in @xmath2 , @xmath278 & = & \g_i \left(-(3 - 2\e)\i_4^{d=8 - 2\e}[1 ] - \i_3^{d=6 - 2\e ( 3)}[1]\right)\nonumber \\ & & \qquad -\sum_{m=1}^4 \eta_{im}\alpha_i\alpha_m \left(\i_3^{d=6 - 2\e ( m)}[1]- \i_3^{d=6 - 2\e ( 3)}[1 ] \right).\nonumber\end{aligned}\ ] ] both brackets are separately finite .
first , the divergent part of @xmath279 $ ] precisely cancels against that of @xmath280 $ ] .
second , all triangles in @xmath2 dimensions have @xmath281 poles and the difference of any two , is either zero or a log . further differentiation does not change the finiteness properties of the @xmath282 tensor integrals .
the @xmath276 combinations are also well behaved in certain kinematic limits .
for example , @xmath283 $ ] and @xmath264 $ ] are finite as @xmath284 or @xmath285 .
just as differentiating functions which are finite as @xmath66 does not introduce poles in @xmath82 , neither can it introduce poles in the kinematic invariants @xmath201 or @xmath286 .
as an example , consider the function @xmath287 given by , @xmath288 which appears to contain a pole in @xmath201 . in the @xmath284 limit , @xmath289 and @xmath290 so that , @xmath291 and therefore , @xmath292 similarly , @xmath287 contains no power - like divergences and @xmath286 .
this is because the @xmath66 limit has already been taken , and the order of taking the two limits does not commute . for next - to - leading order calculations , we only approach the limit @xmath293 and @xmath82 can safely be taken to zero first . ] in the @xmath294 limit and , with a little more work , it can be shown that , @xmath295 once again , these functions combine dilogarithms , logarithms and constants in a highly non - trivial way to form well behaved building blocks . explicit forms for the @xmath276 functions for @xmath296 and 3 are given in appendix b. the higher dimension and feynman parameter integrals for the one - mass box integral obtained by taking @xmath297 can also be constructed in a similar way . for this kinematic configuration ,
the @xmath25 parameters are defined by , @xmath298 while , @xmath299 and , @xmath300 in terms of invariants , @xmath301 the scalar integral can be written , @xmath302 = \frac{c_\gamma}{\e^2}\left(2(\alpha_1\alpha_3)^\e + 2 ( \alpha_2\alpha_4)^\e -2(\alpha_1\alpha_4)^\e\right ) + 2\ld{0}^{1m}(p_1,p_2,p_3 ) + { \cal o}(\e),\ ] ] where , @xmath303 as expected , in @xmath2 dimensions , the scalar integral is finite , @xmath304 = -\frac{2s_{12}s_{23}}{\delta_4^{1m}}\ld{0}^{1m}(p_1,p_2,p_3 ) \equiv \ld{1s}^{1m}(p_1,p_2,p_3 ) .
\label{eq : ld1s1m}\ ] ] in higher dimensions , the scalar integrals satisfy analagous equations to ( [ eq : i4adj8][eq : i4adj12 ] ) with @xmath297 , and finite parts given by , @xmath305 the stability of these functions as @xmath306 is illustrated in fig .
[ fig : ldns1 m ] for a particular point in phase space ; @xmath240 , @xmath307 , with @xmath286 varying so the limit is approached .
the maximum possible value of @xmath308 occurs when @xmath309 ; i.e. @xmath310 . as before , @xmath311 , with a single inverse power of the gram determinant , is numerically stable .
however there are numerical instabilities for the other functions with more powers of gram determinant in the denominator . in all cases , the function is well approximated by the first term of the taylor expansion provided @xmath312 .
unlike the @xmath127 case , the scalar triangle pinchings all contain infrared poles and there is no benefit in absorbing the @xmath231 piece in the tensor integrals .
therefore we introduce , @xmath313,\ ] ] for @xmath314 and 4 . using equations ( [ eq : inxi][eq : inxixjxkxl ] ) with @xmath126 and @xmath115 we find , @xmath315 & = & \g_i \ld{1}^{1 m } - \sum_{m=1}^4 \eta_{im}\alpha_i\alpha_m \i_3^{(m)}[1],\\ \i_4^{1m}[x_ix_j ] & = & \g_i \ld{2j}^{1 m } - \eta_{ij}\alpha_i\alpha_j \ld{1}^{1 m } - \sum_{m=1}^4 \eta_{im}\alpha_i\alpha_m \i_3^{(m)}[x_j],\\ \i_4^{1m}[x_ix_jx_k ] & = & \g_i \ld{3jk}^{1 m } -\frac{1}{2 } \eta_{ij}\alpha_i\alpha_j \ld{2k}^{1 m } -\frac{1}{2 } \eta_{ik}\alpha_i\alpha_k \ld{2j}^{1m}\nonumber \\ & & \qquad - \sum_{m=1}^4 \eta_{im}\alpha_i\alpha_m \i_3^{(m)}[x_jx_k],\\ \i_4^{1m}[x_ix_jx_kx_l ] & = & \g_i \ld{4jkl}^{1 m } -\frac{1}{3 } \eta_{ij}\alpha_i\alpha_j \ld{3kl}^{1 m } -\frac{1}{3 } \eta_{ik}\alpha_i\alpha_k \ld{3jl}^{1 m } -\frac{1}{3 }
\eta_{il}\alpha_i\alpha_l \ld{3jk}^{1m}\nonumber \\ & & \qquad - \sum_{m=1}^4 \eta_{im}\alpha_i\alpha_m \i_3^{(m)}[x_jx_kx_l].\end{aligned}\ ] ] as in the previous section , the @xmath316 functions are finite as @xmath66 and contain no power - like divergences in the @xmath317 , @xmath318 and @xmath306 limits . for convenience , explicit forms are given in appendix b. the two mass box graph where the massive legs sit on opposite sides is a special case because the scalar integral itself is not finite as @xmath319 . we must therefore proceed with care . to make best use of the symmetry under @xmath218 ,
it is convenient to write , @xmath320 under this flip symmetry , @xmath321 , @xmath322 and @xmath323 in this kinematic configuration , the @xmath25 parameters can be defined by , @xmath324 where @xmath325 is an extra kinematic variable that ensures that the @xmath50 are independent , @xmath326 with this choice of @xmath50 , @xmath327 each row of @xmath96 naturally couples together two of the pinchings ( triangles ) of this box , so we might expect such structure to dominate the integrals .
the associated gram determinant is given by , @xmath328 where , @xmath329 and , @xmath330 we note that the presence of the gram determinant is synonymous with a factor of @xmath331 .
the scalar integral for the opposite box in @xmath115 is given by , @xmath332 = \frac{2}{(1-\lambda ) } \left ( \frac{c_\gamma}{\e^2 } \left ( ( \alpha_1\alpha_3)^\e + ( \alpha_2\alpha_4)^\e - ( \alpha_2\alpha_3)^\e - ( \alpha_1\alpha_4)^\e\lambda^{-\e}\right ) + \ld{0}^{opp}(p_1,p_2,p_3 ) \right ) + { \cal o}(\e),\ ] ] where the finite part @xmath333 can be written , @xmath334 as @xmath335 , there is a manifest singularity in @xmath336 $ ] since , @xmath337 this double logarithm can never combine with lower point scalar integrals to form a combination well behaved as @xmath319 .
in fact , it is easy to see from fig .
[ fig : boxin ] that the only scalar integrals which are available by pinching are the triangle integrals with one and two massive legs .
these are pure poles in @xmath82 and can not be combined with the finite parts of the opposite box integral .
there is no appropriate function which can generate the double logarithm as @xmath335 and consequently no finite function can be formed .
since the matrix elements are in general finite in the limit of vanishing gram determinants , all occurences of @xmath338 divided by the determinant must vanish .
on the other hand , in @xmath2 , the opposite box is not only finite as @xmath66 as expected , but also as @xmath335 .
this is because @xmath339 is effectively @xmath340 and its presence in the numerator of equation ( [ eq : d - change ] ) removes the gram determinant from the denominator .
consequently , we see that , @xmath341 = -\frac{1}{s_{13 } } \ld{0}^{opp}(p_1,p_2,p_3),\ ] ] which , since , @xmath342 is also finite as @xmath343 . in dealing with the all - massive triangle and adjacent box in the previous sections ,
we have constructed groups of scalar integrals that are finite both as @xmath66 and @xmath344 . for the opposite box , this is not easy to do since the raw scalar integral is not finite in either of these limits .
therefore , we follow the approach used for the two - mass triangle graph and obtain the tensor integrals by direct differentiation .
however , we note that the factor @xmath345 associated with the differentiation formula ( [ eq : deriv ] ) is itself divergent , implying that the @xmath346 term in @xmath336 $ ] is needed - the so called @xmath82-barrier .
we circumvent this , by using equation ( [ eq : inxi ] ) , to construct the tensor integral with a single feynman parameter .
explicitly , we find , @xmath347 & = & \frac{1}{(1-\lambda ) } \biggl \{\frac{(\alpha_1-\lambda\alpha_2)\ld{0}^{opp}(p_1,p_2,p_3 ) } { \alpha_1\alpha_3+\alpha_2\alpha_4-\alpha_1\alpha_4-\lambda\alpha_2\alpha_3 } \nonumber \\ & & \qquad -\frac{c_\gamma}{\e^2 } \left(\frac{(\alpha_2\alpha_4)^{\e } - ( \alpha_2\alpha_3)^{\e}}{\alpha_4-\alpha_3}\right ) + \frac{\lambda c_\gamma}{\e^2 } \left(\frac{(\alpha_1\alpha_4)^{\e}\lambda^{-\e } - ( \alpha_1\alpha_3)^{\e}}{\alpha_4-\lambda\alpha_3}\right ) \biggr \},\\ \ihat_4^{opp}[a_4 ] & = & \frac{1}{(1-\lambda ) } \biggl \{\frac{(\alpha_2-\alpha_1)\ld{0}^{opp}(p_1,p_2,p_3 ) } { \alpha_1\alpha_3+\alpha_2\alpha_4-\alpha_1\alpha_4-\lambda\alpha_2\alpha_3 } \nonumber \\ & & \qquad + \frac{c_\gamma}{\e^2 } \left(\frac{(\alpha_2\alpha_4)^{\e } - ( \alpha_2\alpha_3)^{\e}}{\alpha_4-\alpha_3}\right ) -\frac{c_\gamma}{\e^2 } \left(\frac{(\alpha_1\alpha_4)^{\e}\lambda^{-\e } - ( \alpha_1\alpha_3)^{\e}}{\alpha_4-\lambda\alpha_3}\right ) \biggr \}.\end{aligned}\ ] ] this is enough to construct all the necessary integrals , since those for @xmath348 and @xmath349 can be obtained by the simultaneous exchanges , @xmath350 , @xmath351 ( and @xmath218 ) , while integrals with higher powers can be obtained by direct differentiation .
the only subtlety is in differentiating @xmath333 , where it is useful to re - express some of the logarithms as single poles in @xmath82 .
for example , @xmath352 this helps to ensure that although the double poles in @xmath82 are still divided by @xmath331 , the single poles are finite as @xmath335 .
in fact , when the tensor integral is multipled by a factor of @xmath82 we also expect that the resulting single logarithms occur in groups that are finite as @xmath335 .
we therefore introduce the auxiliary functions , @xmath353 \right ) \nonumber \\ & = & - \frac{1}{(1-\lambda ) } \lim_{\e \to 0 } \left ( \alpha_4\alpha_1 \i_3^{(1)}[x_4^{n-1 } ] - \alpha_4\alpha_2 \i_3^{(2)}[x_4^{n-1 } ] \right ) \nonumber \\ & = & \frac{2s_{13}}{\delta_4^{opp } } \left ( s_{12 } \lc{n}^{2m}(p_{12},p_3 ) - p_2 ^ 2 \lc{n}^{2m}(p_2,p_3 ) \right ) , \end{aligned}\ ] ] for @xmath193 and , @xmath354 because these functions contain only a single power of the gram determinant , they are not difficult to evaluate numerically . the tensor integrals are straightforward to derive , but can be rather lengthy , for example , @xmath355 & = & -\frac{(\prod_i \alpha_i)\alpha_4 ^ 2}{1 - 2\e}\frac{\partial}{\partial\alpha_4 } \ihat^{opp}_4[a_4]\nonumber \\ & = & \frac{c_\gamma}{\e^2 } \frac{2s_{13}}{\delta_4^{opp } } \biggl ( \frac{s_{12}^2 \left ( ( -s_{12})^{-\e}-(-s_{123})^{-\e } \right)}{(s_{123 } -s_{12})^2 } - \frac{p_2 ^ 4 \left ( ( -p_2 ^ 2)^{-\e}-(-s_{23})^{-\e } \right)}{(s_{23 } - p_2 ^ 2)^2 } \biggr ) \nonumber \\ & & \qquad + \frac{c_\gamma}{\e } \frac{1}{s_{13 } } \biggl ( \frac{(-s_{123})^{-\e}}{(s_{123 } -s_{12 } ) } - \frac{(-s_{23})^{-\e}}{(s_{23}-p_2 ^ 2 ) } \biggr ) + \frac{2(s_{12}-p_2 ^ 2)^2}{s_{13}\delta_4^{opp } } \ld{0}(p_1,p_2,p_3 ) \nonumber \\ & & \qquad - \frac{(s_{12}-p_2 ^ 2)}{s_{13 } } \lcd{1}(p_1,p_2,p_3 ) - 2 \lcd{2}(p_1,p_2,p_3).\end{aligned}\ ] ] most of these terms are poles in @xmath82 .
for physical processes , the @xmath229 poles from different tensor integrals must combine in such a way that the @xmath356 cancels , while the @xmath281 poles do not depend on the determinant and are ready to cancel with contributions from other loop configurations . in all cases , the scalar box function @xmath333 is associated with a single inverse power of @xmath357 .
as noted earlier , for physical processes that are finite as @xmath319 , there must be cancellations amongst the various tensor integrals so that no terms containing @xmath358 remain .
we therefore choose to leave the @xmath333 functions exposed to facilitate the cancellation of these terms . for third and fourth rank tensor integrals , we also need to know the @xmath2 box with one or two feynman parameters in the numerator and the @xmath160 box .
the former are finite as @xmath66 and are easily obtained by direct differentiation of @xmath359 $ ] so that , for example , @xmath360 & = & \frac{1}{2s_{13}}\biggl\ { \frac{(s_{12}-p_2 ^ 2)}{s_{13 } } \ld{0}(p_1,p_2,p_3 ) - s_{12 } \lc{1}^{2m}(p_{12},p_3 ) \nonumber \\ & & \qquad + p_2 ^ 2 \lc{1}^{2m}(p_2,p_3 ) -\log\left(\frac{s_{123}}{s_{23}}\right ) \biggr \},\end{aligned}\ ] ] while the @xmath160 box is established using the dimension changing equation ( [ eq : d - change ] ) , @xmath361 & = & \frac{1}{6s_{13 } } \biggl \ { \frac{s_{12}s_{23}-p_2 ^ 2s_{123}}{s_{13 } } \ld{0}(p_1,p_2,p_3 ) - s_{12 } \log \left ( \frac{s_{123}}{s_{12 } } \right ) - s_{23 } \log \left ( \frac{s_{123}}{s_{23 } } \right ) \nonumber \\ & & \qquad + p_2 ^ 2 \log \left ( \frac{s_{123}}{p_2 ^ 2 } \right)\biggr \ } + \frac{1}{6 } \left ( \frac{(-s_{123})^{-\e}}{\e } + \frac{11}{3 } \right)c_\gamma .\end{aligned}\ ] ] as might be expected , both of these are finite as @xmath343 and this can be seen directly from equation ( [ eq : ldopplim ] ) . in summary ,
the situation for box integrals is very similar to that for triangle graphs .
when the scalar integral is finite as @xmath239 , differentiating - or equivalently adding factors of feynman parameters - does not introduce kinematic singularities .
hence natural groupings of box and triangle integrals arise that are finite as @xmath239 . furthermore
, the infrared and ultraviolet singularities can be isolated easily .
although we have explicitly worked through a subset of kinematic configurations relevant to certain qcd processes , this method is systematic and can be applied to processes with more general kinematics ( and particularly non - zero internal masses ) .
in this section we consider five point integrals with only one external mass .
the outflowing lightlike momenta are denoted @xmath4 , @xmath362 while the fifth @xmath363 is massive , @xmath364 as shown in fig .
[ fig : pentin ] .
the auxiliary momentum is then , @xmath365 we can make the choice , @xmath366 with , @xmath367 as in the opposite box integral , @xmath325 is an extra kinematic variable that ensures the @xmath50 are independent .
it is the same variable that occurs in the third pinching which forms an opposite box configuration .
the matrix @xmath183 is given by , @xmath368 and the normalisation factor is , @xmath369 the @xmath110 are rather lengthy , but can be read off from @xmath183 .
the scalar pentagon integral is by now well - known in @xmath99 @xcite and in @xmath115 @xcite and can be written in terms of these variables as , @xmath370 = -\frac{1}{2n_5 } \sum_{m=1}^5 \ghat_m \ihat_4^{(m)}[1 ] + \cal{o}(\e ) .\ ] ] the five pinchings and the momenta associated with each is illustrated in fig .
[ fig : pentin ] .
the limit @xmath371 corresponds to the vanishing of the gram determinant associated with the @xmath231 pinching .
the scalar integral for this pinching is not well behaved in this limit and so should not be expected to combine with the other pinchings .
therefore , we separate @xmath372 $ ] according to the pole structure in @xmath82 and @xmath373 .
we identify the function @xmath374 which is finite as both @xmath371 and @xmath375 and does not depend on the opposite - box pinching , @xmath231 , plus a combination of scalar box integrals containing all the infrared poles and the remaining @xmath376 singularities , @xmath377 & = & \left(\sum_{i=1}^5 \alpha_3\alpha_i \kappa_i \right ) \le{1 } - \frac{1}{2n_5 } \g_3 \i_4^{(3)}[1 ] - \alpha_2\alpha_4\left ( \i_4^{(2)}[1]+\i_4^{(4)}[1]\right ) \nonumber \\ & & -\frac{1}{2 } \alpha_3 \left ( \alpha_1\i_4^{(1)}[1]-\alpha_2\i_4^{(2)}[1]-\alpha_4\i_4^{(4)}[1 ] + \alpha_5\i_4^{(5)}[1]\right ) , \end{aligned}\ ] ] where @xmath378 and , @xmath379 .\ ] ] applying equation ( [ eq : inxi ] ) and noting that @xmath380 is both infrared and ultraviolet finite , the integrals with one insertion are also determined in terms of box integrals in @xmath115 dimensions , @xmath381 = -\frac{1}{2n_5}\sum_{m=1}^5 \eta_{im } \ihat_4^{(m)}[1 ] + \cal{o}(\e).\ ] ] making the same separation as before , @xmath382 = \left(\alpha_3\alpha_i\kappa_i\right ) \le{1}-\frac{1}{2n_5 } \sum_{m=1}^5 \alpha_i\alpha_m\left(\eta_{im}-\kappa_i\kappa_m\right ) \i_4^{(m)}[1 ] + \cal{o}(\e),\ ] ] where the bracket multiplying the scalar box integral is always proportional to @xmath373 . similarly , using the usual formula for two feynman parameters ( [ eq : inxixj ] ) and concentrating all of the @xmath383 dependence into @xmath384 , where , @xmath385 we find , @xmath386 & = & \chat_{i , j }
\left ( \ihat_5^{d=6}[1 ] + \frac{1}{2n_5 } \sum_{m=1}^5 \ghat_m \ihat_4^{d=6,(m)}[1 ] \right)\nonumber \\ & & + \frac{1}{4n_5 ^ 2 } \sum_{m=1}^5 ( \eta_{im } \ghat_j - \eta_{ij } \ghat_m ) \ihat_4^{d=6,(m)}[1 ] - \frac{1}{2n_5 } \sum_{m=1}^5 \eta_{jm } \ihat_4^{(m)}[a_i ] \nonumber \\ & = & - \chat_{i , j } \frac{\hat\delta_5 \ihat_5^{d=8}[1]}{n_5 } \nonumber \\ & & + \frac{1}{2n_5 } \sum_{m=1}^5 \left ( \left ( \frac{\ghat^{(m)}_j \eta_{im } - \ghat_m \eta_{ij}^{(m)}}{\eta_{mm } } \right ) \ihat_4^{d=6,(m)}[1 ] - \eta_{jm } \ihat_4^{(m)}[a_i]\right ) .
\label{eq : pent-2ins}\end{aligned}\ ] ] to simplify the non-@xmath384 terms ( and eliminate one power of @xmath373 ) , we have rewritten some of the variables appropriate to the pentagon integral in terms of those appropriate to the box pinchings , @xmath387 and @xmath388 , using the relations @xcite , @xmath389 bern , dixon and kosower have shown @xcite that @xmath390 drops out of the calculation of any gauge theory amplitudes by using the identity @xcite , @xmath391 } , \label{eq : trick}\ ] ] where @xmath392}$ ] represents the metric tensor in @xmath99 . since
, @xmath393 & = & \i_5[\p^{\mu_i}\p^{\mu_j } ] - \frac{1}{2}\i_5^{d=6}[1]g^{\mu_i\mu_j } \nonumber \\ & = & \sum_{i , j=1}^5 \i_5[x_{i+1}x_{j+1}]q_i^{\mu_i}q_j^{\mu_j } - \frac{1}{2}\i_5^{d=6}[g^{\mu_i\mu_j}],\end{aligned}\ ] ] _ all _ terms proportional to @xmath384 in @xmath394 $ ] can be moved into the existing @xmath395 piece .
inspection of equation ( [ eq : pent-2ins ] ) indicates that the correct term to reshuffle is @xmath396/n_5 $ ] rather than merely @xmath397 $ ] . retaining only the @xmath398 terms , @xmath399 g^{\mu_i\mu_j } - \frac{\hat\delta_5 \i_5^{d=8}[1]}{n_5 } \sum_{i , j=1}^4 q_i^{\mu_i } q_j^{\mu_j } c_{i+1,j+1 } \nonumber \\ & = & -\frac{1}{2 } \i_5^{d=6}[1 ] g^{\mu_i\mu_j } + \frac{1}{2 } \left ( \i_5^{d=6}[1 ] + \frac{1}{2n_5 } \sum_{m=1}^5 \g_m \i_4^{d=6,(m)}[1 ] \right ) g^{\mu_i\mu_j}_{[4]}\nonumber \\ & = & \frac{1}{4n_5 } \sum_{m=1}^5 \g_m \i_4^{d=6,(m)}[1 ] g^{\mu_i \mu_j } + { \cal o}(\e).\end{aligned}\ ] ] as expected the @xmath400 $ ] terms precisely cancel . here
the finiteness of @xmath400 $ ] and @xmath401 $ ] has been used to ensure that this term generates only @xmath346 corrections when replacing @xmath402}$ ] with the dimensionally regularised @xmath395 .
the remaining piece should not contain any kinematic singularities associated with lower - point gram determinants , in particular @xmath373 , as it originates from the well - behaved @xmath403 .
this is indeed the case and we write , @xmath404.\ ] ] the pentagon integrals with three insertions can be obtained by direct differentiation of equation ( [ eq : pent-2ins ] ) .
we find , @xmath405 & = & \chat_{i , j } \left ( \ihat_5^{d=6}[a_k ] + \frac{1}{6n_5 } \sum_{m=1}^5 \left ( \eta_{km } \ihat_4^{d=6,(m)}[1 ] + 2\ghat_m \ihat_4^{d=6,(m)}[a_k ] \right ) \right ) \nonumber \\ & & + \frac{1}{6n_5 } \sum_{m=1}^5 \biggl ( 2 \left ( \frac{\ghat^{(m)}_j \eta_{im } - \ghat_m \eta_{ij}^{(m)}}{\eta_{mm } } \right ) \ihat_4^{d=6,(m)}[a_k ] \nonumber \\ & & - \left ( \frac{\eta_{im}\eta_{jk}^{(m ) } - \eta_{ij}^{(m)}\eta_{km}}{\eta_{mm } } \right ) \ihat_4^{d=6,(m)}[1 ] - \eta_{jm } \ihat_4^{(m)}[a_ia_k ] \biggr ) + \mbox{cyclic i , j , k}. \label{eq : pent-3ins}\end{aligned}\ ] ] since the @xmath384 term is obtained by differentiating @xmath406 $ ] , it must be finite as both @xmath376 and @xmath66 and we introduce the finite function , @xmath407 + 2\g_m \i_4^{d=6,(m)}[x_k ] \right ) .\ ] ] recalling that , @xmath408 & = & \i_5[\p^{\mu_i}\p^{\mu_j}\p^{\mu_k } ] -\frac{1}{2}\i_5^{d=6}[\ { g\p \}^{\mu_i\mu_j\mu_k } ] \nonumber \\ & = & - \sum_{i , j , k=1}^5 \i_5[x_{i+1}x_{j+1}x_{k+1}]q_i^{\mu_i } q_j^{\mu_j } q_k^{\mu_k } \nonumber \\ & & + \frac{1}{2}\left\ { \sum_{k=1}^5 \i_5^{d=6}[x_{k+1}]g^{\mu_i\mu_j}q_k^{\mu_k } + \mbox{cyclic i , j , k } \right\},\end{aligned}\ ] ] and keeping only the @xmath398 terms in equation ( [ eq : pent-3ins ] ) we find , @xmath409 g^{\mu_i \mu_j}q_k^{\mu_k } - \left ( \i_5^{d=6}[x_{k+1 } ] + \le{3k } \right ) \sum_{i , j=1}^4 c_{i+1,j+1 } q_i^{\mu_i } q_j^{\mu_j}q_k^{\mu_k } \nonumber \\ & = & \frac{1}{2 } \i_5^{d=6}[x_{k+1 } ] g^{\mu_i \mu_j}q_k^{\mu_k } - \frac{1}{2 } \left ( \i_5^{d=6}[x_{k+1 } ] + \le{3k } \right ) g^{\mu_i \mu_j}_{[4]}q_k^{\mu_k } \nonumber \\ & = & -\frac{1}{2 } \le{3k}g^{\mu_i \mu_j}q_k^{\mu_k } + { \cal o}(\e).\end{aligned}\ ] ] as in the previous case , the finiteness of the coefficient of @xmath392}$ ] has been used to promote it to the full @xmath115 metric tensor .
the tensor integrals with four and five insertions may be obtained by further differentiation , and the same trick used to rewrite the @xmath398 terms as a contribution to the metric tensor structure . in this way , all vestiges of the pentagon in @xmath2 and higher dimensions can be removed , along with the inverse powers of @xmath383 .
the functions introduced in this section , @xmath374 , @xmath410 and @xmath411 , only contain a single power of the gram determinant .
consequently , they can be evaluated without numerical problems . to illustrate this
, we choose a representative phase space point , @xmath412 and use the variable @xmath286 to control @xmath325 .
the sixth variable @xmath413 is chosen to lie within the physical region defined by the pentagon gram determinant , @xmath414 .
[ fig : les ] shows the functions @xmath374 and @xmath410 together with @xmath415 . in each case
, we see that the @xmath335 limit is smoothly approached indicating that the function is intrinsically well behaved in that limit .
in this paper we have developed a new strategy for evaluating one - loop tensor integrals .
it avoids the usual problems associated with the presence of gram determinants . such gram determinants arise when the tensor integrals are expressed in terms of the physical momenta and generate false singularities at the edges of phase space .
in addition to creating numerical instabilities , they tend to increase the size of the one - loop matrix elements .
our approach is to construct groups of scalar integrals which are well behaved in the limit of vanishing gram determinant ( @xmath37 ) , and which can be evaluated with arbitrary precision by making a taylor series expansion in @xmath416 .
in fact such combinations arise naturally by either differentiating with respect to the external parameters - essentially yielding scalar integrals with feynman parameters in the numerator - or by developing the scalar integral in @xmath2 or higher dimensions .
evaluating these new integrals is straightforward - they are just linear combinations of the known scalar integrals in @xmath99 or @xmath115 . as such , they combine the dilogarithms , logarithms and constants from different scalar integrals in an extremely non - trivial way . as a bonus other spurious kinematic singularities
are also controlled - they appear in the denominator of the finite functions , which are well behaved in the singular limit .
although the number of basic functions has increased , the number of dilogarithm evaluations has not , since the functions are generated recursively .
furthermore , because the gram determinant singularities are not genuine , by grouping integrals in this way , the expressions for one loop integrals are compactified . to illustrate the approach for specific integrals , we have applied the method to 3- , 4- and 5-point integrals where
the internal masses have been set equal to zero .
these tensor integrals are relevant for a range of qcd processes where the quark and gluon masses are negligible . for more general processes with arbitrary internal masses and external kinematics
, the relevant combinations of scalar integrals can be obtained using equations ( [ eq : d - change],[eq : inxi][eq : inxixjxkxl ] ) . as a by - product
we have shown how all the gram determinants associated with pentagon graphs can be eliminated .
we thank walter giele , eran yehudai , bas tausk and keith ellis for collaboration in the earlier stages of this work .
ewng thanks alan martin for useful discussions and for suggestions concerning the manuscript .
jmc and djm thank the uk particle physics and astronomy research council for the award of research studentships .
in this appendix , we collect together explicit forms for the triangle graphs which appear as building blocks in the box graphs .
we fix the kinematics according to fig . 1 , so that momenta @xmath84 and @xmath85 are exiting , with @xmath86 determined by momentum conservation .
there are two distinct cases , according to whether one or two of the external masses are zero . here
we provide explicit results for the case @xmath205 .
there is an additional symmetry under the exchange @xmath417 and @xmath131 .
insertions involving @xmath87 can be eliminated using @xmath418 .
the @xmath115 scalar and tensor integrals are given by , @xmath419 & = & \frac{c_\gamma}{\epsilon^2 } \frac { ( -s_{12})^{-\e } } { s_{12}},\\ \i^{1m}_3[x_1 ] = \i^{1m}_3[x_3 ] = 2 \i^{1m}_3[x_1 ^ 2 ] = 2 \i^{1m}_3[x_3 ^ 2 ] & = & - \frac{1}{s_{12 } } \left ( \frac { ( -s_{12})^{-\epsilon } } { \e } + 2 \right)c_\gamma,\\ \i^{1m}_3[x_1 x_3 ] = 3\i^{1m}_3[x_1 ^ 2 x_3 ] = 3\i^{1m}_3[x_1 x_3 ^ 2 ] & = & \frac{1}{2s_{12}},\\ \i^{1m}_3[x_1 ^ 3 ] = \i^{1m}_3[x_3 ^ 3 ] & = & - \frac{1}{3s_{12 } } \left ( \frac { ( -s_{12})^{-\epsilon } } { \e } + \frac{13}{6 } \right)c_\gamma,\end{aligned}\ ] ] while , the necessary integrals in @xmath2 dimensions read , @xmath420 & = & -\frac{1}{2 } \left ( \frac { ( -s_{12})^{-\epsilon } } { \e } + 3 \right)c_\gamma,\\ \i^{1m , d=6 - 2\e}_3[x_1 ] = \i^{1m , d=6 - 2\e}_3[x_3 ] & = & -\frac{1}{6 } \left ( \frac { ( -s_{12})^{-\epsilon } } { \e } + \frac{8}{3 } \right)c_\gamma.\end{aligned}\ ] ] when two of the external legs are massive , @xmath421 but @xmath422 , the divergent integrals read , @xmath423 & = & \frac{c_\gamma}{\epsilon^2 } \left ( \frac { ( -s_{12})^{-\e } - ( -p_1 ^ 2)^{-\e } } { s_{12}-p_1 ^ 2 } \right),\\ \i^{2m}_3[x_3 ] & = & -2\lc{2}^{2m}(p_1,p_2 ) - \frac{c_\gamma}{\e } \frac { ( -s_{12})^{-\e } } { ( s_{12}-p_1 ^ 2 ) } - \frac{p_1 ^ 2}{(s_{12}-p_1 ^ 2 ) } \i^{2m}_3[1],\\ \i^{2m}_3[x_3 ^ 2 ] & = & -3\lc{3}^{2m}(p_1,p_2 ) + \frac{1}{2(s_{12}-p_1 ^ 2 ) } - c_\gamma\frac { ( -s_{12})^{-\e } } { \e } \frac{(s_{12}-3p_1 ^ 2)}{2(s_{12}-p_1 ^ 2)^2 } \nonumber \\ & & \qquad + \frac{p_1 ^ 4}{(s_{12}-p_1 ^ 2)^2 } \i^{2m}_3[1],\\ \i^{2m}_3[x_3 ^ 3 ] & = & -\frac{11}{3 } \lc{4}^{2m}(p_1,p_2 ) + \frac{(s_{12}-2p_1 ^ 2 ) } { 2(s_{12}-p_1 ^ 2)^2 } - \frac{p_1 ^ 6}{(s_{12}-p_1 ^ 2)^3 } \i^{2m}_3[1 ] \nonumber \\ & & \qquad - \frac { ( -s_{12})^{-\e } } { \e } \left ( \frac{1 } { 3(s_{12}-p_1 ^ 2 ) } - \frac{p_1 ^ 2}{2(s_{12}-p_1 ^ 2)^2 } + \frac{p_1 ^ 4}{(s_{12}-p_1 ^ 2)^3 } \right)c_\gamma,\end{aligned}\ ] ] while , @xmath424 = 2\i^{2m}_3[x_1 ^ 2 ] = 3\i^{2m}_3[x_1 ^ 3 ] & = & \lc{1}^{2m}(p_1,p_2),\\ \i^{2m}_3[x_1x_3 ] = 3\i^{2m}_3[x_1 ^ 2x_3 ] & = & \frac{1}{2 } \lc{2}^{2m}(p_1,p_2),\\ \i^{2m}_3[x_1 x_3 ^ 2 ] & = & \frac{1}{3 } \lc{3}^{2m}(p_1,p_2).\end{aligned}\ ] ] the functions @xmath425 are defined by equations ( [ eq : lcmassless],[eq : lcmassless1 ] ) .
the @xmath2 dimension integrals read , @xmath426 & = & \frac{1}{2 } \left ( p_1 ^ 2 \lc1^{2m}(p_1,p_2 ) - \left(\frac { ( -s_{12})^{-\epsilon } } { \e } + 3\right ) c_\gamma \right),\\ \i^{2m , d=6 - 2\e}_3[x_1 ] & = & \frac{1}{6 } \left ( p_1 ^ 2\lc1^{2m}(p_1,p_2 ) - \left ( \frac { ( -s_{12})^{-\epsilon } } { \e } + \frac{8}{3 } \right)c_\gamma \right),\\ \i^{2m , d=6 - 2\e}_3[x_3 ] & = & \frac{1}{6 } \left ( p_1 ^ 2\lc2^{2m}(p_1,p_2 ) - \left(\frac { ( -s_{12})^{-\epsilon } } { \e } + \frac{8}{3 } \right)c_\gamma \right).\end{aligned}\ ] ] the corresponding integrals for the case @xmath217 , @xmath128 can be obtained by using the above formulae with the substitutions , @xmath427 the limit @xmath297 may also be safely taken .
for example , using the fact that in this limit @xmath428 and observing that all of the triangle loop integral contributions proportional to @xmath229 trivially drop out in eqs .
( a.8)-(a.10 ) , we see that , @xmath429 \to \i_3^{1m}[x_3^n].\ ] ]
here we collect together explicit forms for the some of the finite functions appearing in the box integrals .
we fix the kinematics according to fig . 2 , so that momenta @xmath84
, @xmath85 and @xmath86 are exiting , with @xmath430 determined by momentum conservation .
we first focus on the case where two of the adjacent external legs are massive , @xmath431 and @xmath127 . for integrals with a single feynman parameter in the numerator , we have , @xmath432 where the box function in @xmath232 , @xmath243 is given by equation ( [ eq : ld1sadj ] ) . when there are two feynman parameters in the numerator , we eliminate @xmath92 using @xmath433 so there are only three relevant functions .
explicitly , we find , @xmath434 the all massive triangle integral function in @xmath2 , @xmath207 is given by equation ( [ eq : lc1s ] ) , the box integral function in @xmath160 is given in equation ( [ eq : ld2sadj ] ) while the remaining triangle functions are given in appendix a and equations ( [ eq : i3x3][eq : i3x3x3x3 ] ) .
the functions for adjacent box integrals with three insertions of feynman parameters contain the box in @xmath161 ( [ eq : ld3sadj ] ) and triangle in @xmath160 ( [ eq : lc2s ] ) .
all integrals can be obtained in terms of the following four functions , @xmath435 for this kinematic configuration , only @xmath431 and there is a ` flip ' symmetry , so that functions related to the parameter @xmath436 are obtained from those related to @xmath91 , by @xmath437 .
the box integrals in higher dimension are given by equations ( [ eq : ld1s1m][eq : ld4s1 m ] ) , and , @xmath438 for two and three insertions , we find , @xmath439 and , @xmath440
in this section , we collect together suitable expansions of the various functions presented in this paper in the limit @xmath441 . these expressions represent the leading term in the expansion of the functions as a taylor series in @xmath41 and should be evaluated for @xmath442 , where @xmath443 can be determined numerically . typically , @xmath444 where @xmath445 is the largest value the gram determinant can achieve . in general , for a given numerical precision , @xmath139 the numerical problems occur when @xmath446 where @xmath150 is the number of gram determinants in the denominator of the function .
99 g. t hooft and m. veltman , nucl . phys . *
b44 * ( 1972 ) 189 ; + c.g .
bollini and j.j .
giambiagi , phys . lett . *
40b * ( 1972 ) 566 ; + j.f .
ashmore , nuovo cim
. lett . * 4 * ( 1972 ) 289 ; + g.m .
cicuta and e. montaldi , nuovo cim . lett . * 4 * ( 1972 ) 329 .
l. m. brown and r. p. feynman , phys .
* 85 * ( 1952 ) 231 ; + g. passarino and m. veltman , nucl . phys . *
b160 * ( 1979 ) 151 .
van oldenborgh and j.a.m .
vermaseren , z. phys .
* c46 * ( 1990 ) 425 .
r. k. ellis , w. t. giele and e. yehudai , private communication .
a. signer , ph.d .
thesis , ` helicity method for next - to - leading order corrections in qcd ' , eth zurich ( 1995 ) .
a. i. davydychev , phys . lett .
* b263 * ( 1991 ) 107 .
o. v. tarasov , phys .
* d54 * ( 1996 ) 6479 .
z. bern , l. dixon and d. a. kosower , nucl .
phys . * b412 * ( 1994 ) 751 .
z. bern , l. dixon and d. a. kosower , phys .
b302 * ( 1993 ) 299 , _ ibid _ * b318 * ( 1993 ) 649 .
r. g. stuart , comp .
. comm . * 56 * ( 1988 ) 367 .
e. w. n. glover and d. j. miller , ` the one - loop corrections for @xmath453 ' , university of durham preprint dtp/96/66 , hep - ph/9609474 .
l. dixon and a. signer , ` electron - positron annihilation into four jets at next - to - leading order in @xmath454 ' , hep - ph/9609460 .
z. bern , l. dixon and d. a. kosower , ` unitarity - based techniques for one - loop calculations in qcd ' , hep - ph/9606378 ; + z. bern , l. dixon , d. a. kosower and s. wienzierl , ` one - loop amplitudes for @xmath455 ' , hep - ph/9610370 .
g. t hooft and m. veltman , nucl . phys .
* b153 * ( 1979 ) 365 .
z. bern , l. dixon and d. a. kosower , phys .
* 70 * ( 1993 ) 2677 .
z. bern , l. dixon and d. a. kosower , nucl .
b437 * ( 1995 ) 259 .
z. kunszt , a. signer and z. trocsanyi , nucl .
* b411 * ( 1994 ) 397 .
d. kreimer , int .
j. mod . phys . * a8 * ( 1993 ) 1797 ; + a. davydychev , j. phys .
* a25 * ( 1992 ) 5587 .
g. p. lepage , j. comp . phys .
* 27 * ( 1978 ) 192 .
melrose , il nuovo cimento , * 40a * ( 1965 ) 181 ; + w. van neerven and j.a.m .
vermaseren , phys . lett .
* b137 * ( 1984 ) 241 . | we show how to evaluate tensor one - loop integrals in momentum space avoiding the usual plague of gram determinants .
we do this by constructing combinations of @xmath0- and @xmath1-point scalar integrals that are finite in the limit of vanishing gram determinant .
these non - trivial combinations of dilogarithms , logarithms and constants are systematically obtained by either differentiating with respect to the external parameters - essentially yielding scalar integrals with feynman parameters in the numerator - or by developing the scalar integral in @xmath2 or higher dimensions .
an additional advantage is that other spurious kinematic singularities are also controlled . as an explicit example
, we develop the tensor integrals and associated scalar integral combinations for processes where the internal particles are massless and where up to five ( four massless and one massive ) external particles are involved . for more general processes ,
we present the equations needed for deriving the relevant combinations of scalar integrals . = -1.5 cm = -1.5 cm = 21.0truecm = 16.5truecm # 1 [ cols="^",options="header " , ] 2ex # 1lc_#1 # 1lcd_#1 # 1ld_#1 # 1le_#1 2li_2 p 3_3 3hat_3 dtp/96/96 + hep - ph/9612413 + * one - loop tensor integrals in dimensional regularisation * + + + |
the mission , launched in march of 2009 , was designed to address the important question of the frequency of earth - size planets around sun - like stars , and to characterize extrasolar transiting planets through a 3.5-year program of very precise photometric monitoring of @xmath1156000 stars @xcite .
science results from the mission have already begun to appear @xcite . as shown already by ground - based surveys for transiting planets ,
considerable effort is required to validate candidates detected photometrically .
this is because false positives usually outnumber true planetary systems by a large factor , which is about 10:1 for the most successful surveys from the ground , but is not yet well characterized for .
the follow - up efforts by the team have been summarized by @xcite .
spectroscopy is often a crucial step in the vetting process , as it allows not only to measure the mass of a planet but also to examine any changes in the line profiles ( bisector spans ) that might indicate a false positive ( see * ? ? ?
* ; * ? ? ?
some of the most challenging false positives to rule out include chance alignments with a background eclipsing binary ( `` blends '' ) .
however , for faint candidates ( @xmath2 ) high - resolution , high signal - to - noise ratio spectroscopy becomes prohibitively expensive in terms of telescope time . even for brighter candidates ,
the reflex motion of the star due to an earth - mass planet can sometimes be below the radial - velocity detection limit , making spectroscopic confirmation very difficult or impossible .
the question then becomes how to validate these candidates , particularly the ones with small planets that are precisely among the most interesting .
a number of other tests have been developed that can aid in understanding the nature of the candidate , and that rely on the long - term and nearly continuous photometric monitoring of , as well as the very high astrometric precision achieved in determining the centroids of the stars ( see also * ? ? ?
these tests include : _ i ) _ verifying that alternating events have the same depth , which they may not if the signal is due to a background eclipsing binary ; _ ii ) _ checking for the presence of shallow secondary eclipses , which are common in eclipsing binaries but are not expected for the smallest planets ; _ iii ) _ checking for ellipsoidal variations , which could be another sign of a blend . _
iv ) _ checking for changes in the centroid positions correlated with the brightness changes , which , if detected , might indicate a blend , or at the very least , a crowded aperture .
this is a powerful diagnostic that is able to disprove many background blends .
in addition to these tests , high - resolution imaging is an important tool to identify neighboring stars that might be eclipsing binaries with the potential to cause the transit - like signals .
the photometric aperture of is large enough ( typically many arc seconds across ) that it usually includes other stars in addition to the candidate , which increases the risk of such blends . in some cases , near - infrared observations with warm _
spitzer _ can allow one to reject the planet hypothesis if the transit depth at 3.6@xmath3 m or 4.5@xmath3 m is significantly different from that in the band .
such a signature might result from a blend with an eclipsing binary of a different spectral type than the candidate .
even with this extensive battery of tests it may still be difficult or impossible to provide validation for many of the most interesting planet candidates discovered by . for example , blend scenarios involving an eclipsing binary or an eclipsing star - planet pair physically associated with the candidate ( hierarchical triple systems ) and in a long - period orbit around their common center of mass would often be spatially unresolved from the ground .
these configurations may also not be detectable spectroscopically , and would likewise not produce any measurable centroid motion .
therefore , it is imperative to take advantage of all the information available in vetting candidates . with this as our motivation
, we describe here the use of the light curves themselves in a different way to help discriminate between true planetary transits and a large variety of possible blend scenarios , on a much more quantitative basis than simple back - of - the - envelope calculations could provide .
the technique tests these scenarios by directly modeling the light curves as blends , and has considerable predictive power that allows the expected properties of the various configurations to be tested against other information that may be available .
both hierarchical triples and background blends can be explored .
a restricted application of this type of modeling to has already been made for the five multi - planet candidates announced recently by @xcite .
for the present paper we have chosen to illustrate the full potential of the method , which we refer to as , by applying it to the unique case of object of interest 377 ( koi-377 , henceforth kepler-9 ) .
this is a multi - planet system reported and described in detail by @xcite , with _ three _ low - amplitude periodic signals in its light curve .
we have selected this system for two main reasons . on the one hand
, it represents the first unambiguous detection of transit timing variations ( ttvs ) in an extrasolar planet , with a pattern of variation seen in two of its signals ( kepler-9b and kepler-9c ) that constitutes irrefutable evidence that the objects producing them are bona - fide planets .
this offers an ideal opportunity to test because their true nature is already known . on the other hand ,
the third signal ( koi-377.03 ) is small enough that it would correspond to a super - earth , but validation of its planetary origin is not yet in hand .
should it be validated , kepler-9 would become an even more remarkable laboratory for the study of the architecture of planetary systems involving small planets .
thus , exploring the wide range of possible blend configurations that might mimic this shallow signal is of the greatest interest for determining its true nature
. is likely to find many other candidate transiting planets similar to koi-377.03 , for which final validation by other means is not currently feasible , either because the expected radial - velocity signal is too small , or because doppler measurements are otherwise complicated due to the star being chromospherically active , rapidly rotating , or too faint . with the application to kepler-9
we show that our light - curve modeling technique is a powerful tool for exploring astrophysical false positive scenarios that is complementary to other diagnostics , and should play an important role in the discovery of earth - size planets around other targets .
in general the detailed shape of a light curve displaying transit - like events can be expected to contain useful constraints on possible blend scenarios that might be responsible for those signals .
with photometry of the quality of that provided by , those constraints can be quite strong , and may be used to exclude many blend configurations and provide support for the planetary hypothesis .
it is thus highly desirable to take advantage of this information , particularly since it relies only on observations already in hand .
the idea behind is to compare the transit photometry of a candidate against synthetic light curves produced by an eclipsing binary that is included within the photometric aperture of , and is contaminating the light of the candidate .
the usually deep eclipses of the binary are attenuated by the light of the candidate , and reduced in depth so that they appear transit - like . in principle
there is an enormous range of possible binary configurations that could mimic all of the features of true planetary transits , including not only their depth , but also the total duration and the length of the ingress and egress phases .
generally it is only with detailed modeling that these can be ruled out .
possible scenarios include not only background eclipsing binaries , but also hierarchical triples , i.e. , an eclipsing binary physically associated with the candidate in a wide orbit around their common center of mass .
the basic procedure for simulating light curves with was described in detail by @xcite , and further changes and enhancements are discussed below .
briefly , the brightness variations of an eclipsing binary are generated with the binary light - curve code ebop @xcite , based on the nelson - davis - etzel model @xcite , and then diluted by the light of the candidate for comparison with the observations .
effects such as limb darkening , gravity brightening , mutual reflection , and oblateness of the binary components are included .
the objects composing the binary are referred to as the ` secondary ' and ` tertiary ' , and the candidate is the ` primary ' . the properties of each object needed to generate the light curves ( brightness and size ) are taken from model isochrones by @xcite , parametrized in terms of their stellar mass . , which we have extrapolated slightly for this application to 0.10@xmath4 , near the brown dwarf limit ) , and because a convenient web tool provided by the authors allows easy interpolation in both age and metallicity ( http://stev.oapd.inaf.it/cgi-bin/cmd ) .
additionally , isochrone magnitudes are available in a variety of passbands including the and _ spitzer _ passbands , as well as sloan and 2mass.[foot : models ] ] for the primary star the appropriate isochrone is selected by using as constraints the effective temperature , surface gravity , and metallicity determined spectroscopically .
we assign also a mass and a radius from this isochrone , although these characteristics are irrelevant for generating the model light curves .
we then read off the intrinsic brightness of the star ( absolute magnitude ) in the passband , which is the only property needed by .
the brightness of the primary is held fixed throughout all simulations .
the parameters of the binary components are allowed to vary freely over wide ranges in order to provide the best match to the photometry in a chi - square sense , subject only to the condition that the two stars lie on the same isochrone , as expected from coeval formation .
to read off their properties ( absolute magnitude and size ) we use the mass as an intermediate variable .
the specific isochrone adopted for the binary depends on the configuration : for hierarchical triple scenarios we adopt the same age and chemical composition as the primary , whereas for background binaries the isochrone can be different . the light curve itself does not provide a useful constraint on the age or metallicity of the binary in the background case , so a typical choice is a model for solar metallicity and a representative age for the field such as 3gyr . for background
binary scenarios the distance between the binary and the main star is parametrized for convenience in terms of the difference in distance modulus , @xmath5 .
the inferred distance between the primary star and the observer will vary from blend to blend because we constrain the combined brightness of all components of the blend to match the measured apparent brightness of the target .
is also able to account for differential extinction between the primary and the binary , which can have a non - negligible effect in some cases given the relatively low galactic latitude of the field .
early versions of have been used occasionally in recent years to examine transiting planet candidates from ground - based surveys such as ogle , tres , and hatnet ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , as well as from corot @xcite .
these studies have exploited the predictive power of to estimate further properties of the blend scenarios , by testing them against complementary information such as color indices , optical / near - infrared spectroscopy , or near - infrared photometry from _
spitzer_. for the application to , several important modifications have been made to , including the following : _
i ) _ the ability to generate light curves integrated over the 29.4-minute effective duration of an observation when using long - cadence data .
this changes the shape of the transits significantly , given the high precision of the photometry and the relatively short timescales of the events ( see * ? ? ?
* ; * ? ? ?
* ) ; _ ii ) _ de - trending of the original light curves with a 1-day running median to remove instrumental effects , and rejection of outliers ; _ iii ) _ the use of model isochrones specific to the passband , kindly computed for us by l. girardi . can now also use proper limb - darkening coefficients for the same band , as opposed to an approximation to the passband such as the johnson @xmath6 filter , which is considerably narrower .
limb - darkening coefficients have been computed by @xcite and also a. prsa ; ] _ iv ) _ extension to any optical or near - infrared passband . in particular , for any scenario explored with , light curves can be computed at other wavelengths such as the 3.6@xmath3 m and 4.5@xmath3 m passbands of the irac instrument on warm _ spitzer _ , in order to further test the blend hypothesis . additionally , can predict the overall color of a blend in any pair of passbands , including the effects of differential reddening for background or foreground scenarios .
these colors may then be compared with the measured colors of a target .
extinction at different wavelengths is computed following the prescription by @xcite ; _ v ) _ the ability to have the tertiary be a ( dark ) planet instead of a star , in which case the corresponding free parameter becomes the radius of the planet rather than the tertiary mass .
the mass of the planet has little effect on the light curves in most cases , but can nevertheless be set to any value ; _ vi ) _ the ability to include extra light from other stars that may be present in the aperture , which further dilutes the intrinsic signatures from the eclipsing binary ; _ vii ) _ the ability to model systems with eccentric orbits .
eccentricity changes the orbital velocities during transit , and can therefore affect the size ( mass ) of stars that allow satisfactory fits to the light curve .
when exploring blend scenarios involving hierarchical triple systems , the free parameters of the problem are the mass of the secondary , the mass of the tertiary ( or its radius , if a planet ) , and the inclination angle .
a fourth variable , the difference in distance modulus , is added for background blends .
these quantities are typically stepped over wide ranges in a grid pattern to fully map the @xmath7 surface . for the application to kepler-9 below ,
stellar masses are allowed to vary along the isochrones between 0.1@xmath4 and 1.4@xmath4 , although at the larger values the observed duration of the transits is already difficult to match unless the events are highly grazing , in which case the shape would be very different . for planetary tertiaries the radii are allowed to be as large as 1.8@xmath8 ; values higher than this have not been observed .
kepler-9 ( kic3323887 , 2mass19021775 + 3824032 ) is a relatively faint star compared to typical ground - based transit hosts ( magnitude @xmath9 ) , which was observed by the mission beginning in the first quarter of operations , and presents three distinct periodic signals in its light curve .
the two with the largest amplitudes have periods of 19.24 days ( kepler-9b ) and 38.91 days ( kepler-9c ) , and brightness decrements of 6.5 and 6.0 mmag , respectively .
the third signal ( koi-377.03 ) is much shallower ( 0.2 mmag ) , and repeats every 1.59 days .
the two longer periods are within 2.5% of being in a 2:1 ratio , and both objects display obvious ttvs that are anti - correlated , clearly indicating they are interacting gravitationally and therefore orbit the same star , and are planetary in nature ( see * ? ? ?
the estimated radii are quite similar to that of saturn , and the masses are somewhat smaller than saturn , based on available radial - velocity measurements constrained by transit times and durations .
the short - period signal has one of the smallest amplitudes detected by , and may well correspond to a third , super - earth - size planet in the system , with an estimated radius of only @xmath11.5@xmath10 @xcite .
however , because it shows no ttvs related to the other two planets ( nor is expected to , on dynamical grounds ) , and is predicted to induce only a very small reflex velocity on the parent star that may be below detection for such a faint object , the true origin of this signal has not yet been established . in the absence of the crucial evidence of ttvs ,
each of the two largest signals and indeed the third signal as well could in principle be due to a different blend .
therefore , as an illustration of the application of , we model the light curve of kepler-9 at each period separately , as we would any candidate with a single period , and we account for possible blends at the other periods by incorporating extra dilution consistent with those other scenarios .
the goal for the two largest signals is to demonstrate , as a sanity check , that would be able to rule out blends in similar cases where confirmation is lacking , which is expected to find in significant numbers . for the third signal of unknown nature
, the application of should provide valuable evidence one way or the other .
kepler-9 is a solar type star .
the spectroscopic properties of the primary are adopted from @xcite : @xmath11k , @xmath12 , and [ fe / h ] @xmath13 . with these parameters ,
a comparison with the stellar evolution models of @xcite yields a stellar mass of @xmath14 , a radius of @xmath15 , and an age of about 1gyr , along with the absolute magnitude in the band . only the latter is used by , and is held fixed in our modeling .
the distance to the star estimated from the same models is about 650pc , ignoring extinction .
uncertainties in the brightness of the primary stemming from errors in @xmath16 , @xmath17 , and [ fe / h ] are small .
for example , the error in @xmath18 , which has the most direct influence on the intrinsic brightness , translates to an uncertainty of little more than 0.1mag in the absolute magnitude .
this has an insignificant impact on our results .
the photometry used here consists of the long - cadence measurements gathered for kepler-9 during quarters 1 , 2 , and 3 , spanning 218 days , and was treated slightly differently than indicated earlier for a generic candidate because of the complications stemming from the ttvs . using the binary fits tables from mast ( multimission archive at stsci , http://archive.stsci.edu/kepler/ ) ,
the `` raw '' aperture photometry for each quarter was first de - trended using a moving cubic polynomial fit robustly to out - of - transit data , with a sliding window of 999 minutes before and after each individual datapoint .
this technique removes long - term trends due to stellar activity or instrumental errors , but retains the properties of each transit light curve .
statistically significant outliers were removed .
for the two long - period signals , simple folding will not create an accurate light curve because of the strong ttvs .
instead , we used a `` shift - and - stack '' technique , in which each transit event is displaced so that it is centered at `` time '' zero using the measured transit times from @xcite .
along with the measurements in transit , nearly a full cycle of out - of - transit data were also shifted .
specifically , we shifted nearly 25% of an orbital cycle before the transit , and nearly 75% after the transit .
this preserves any curvature outside of eclipse , and in principle would also reproduce any secondary eclipses , both of which can provide useful constraints when modeling the light curve with .
we note , however , that the strong ttvs of the transits would be accompanied by shifted secondary eclipses in a way that can only be predicted by full numerical integration .
this shift - and - stack technique would not align secondary eclipses correctly and thus their depth would need to be significant in each individual event to be noticed . there is no sign of secondary eclipses at these periods in the data at the @xmath19 level , as expected from the planetary nature of the objects , and thus the failure of the shift - and - stack technique to correctly add up the secondary eclipses does not affect our results .
after shifting , all the transit and out - of - transit data were `` stacked '' together and each data point was given a time relative to time zero at the center of each transit event .
this was done separately for the 19-day and 39-day signals .
we have been careful not to use a full cycle of out - of - transit data to avoid using any photometric measurements more than once in the input light curve . for the 1.6-day signal that repeats at regular intervals ( since it shows no ttvs )
, we created a light curve by simply masking out the transits at the other two periods .
the photometric aperture of is typically a few pixels across , with a scale of 398 per pixel ( see below ) .
high - resolution imaging of kepler-9 was performed in order to identify neighboring stars that might be eclipsing binaries blended with and contaminating the target photometry .
images were recorded with the guider camera of the high resolution echelle spectrometer ( hires ; * ? ? ?
* ) on the keck i telescope on mauna kea , in unfiltered light .
the nominal sensitivity of the ccd from 400 to 800 nm yields an effective passband similar to the passband .
the field of view was 43@xmath2057 , and the pixel scale was 030 per pixel .
one of these frames appears in figure [ fig : guider ] , and shows at least four stars in the field of view within 15 of the target .
some of these stars are listed in various astrometric and photometric catalogs .
the brightness of these companions relative to the target was measured using aperture photometry on four separate keck images , and ranges from @xmath21 to 5.9 mag .
0.1 in speckle observations of kepler-9 were carried out on 2010 june 18 with the wiyn 3.5 m telescope located on kitt peak .
they were taken with a two - color emccd speckle camera using narrow - band filters 40 nm wide centered at 562 nm and 692 nm .
we refer loosely to these passbands as @xmath22 and @xmath6 .
the native seeing was 07 .
no companions with @xmath23 mag ( @xmath6 band ) are present in the field of view centered on the target out to 18 , at the 5-@xmath24 confidence level . inside of 02
the sensitivity is reduced , but still allows to rule out brighter companions down to the diffraction limit of 004005 ( see figure [ fig : aosensitivity ] ) .
details of the follow - up speckle observations in the context of the mission are described in more detail by @xcite .
lcccccc[t ! ] kepler-9 & 19:02:17.76 + 38:24:03.2 & & & & & + comp 1 & 19:02:17.91 + 38:24:05.4 & 2.85 & 37.9 & 6.84 & 6.84 & 5.5 + comp 2 & 19:02:18.27 + 38:24:02.8 & 6.04 & 91.7 & 4.52 & 4.17 & 3.7 + comp 3 & 19:02:17.29 + 38:23:57.1 & 8.03 & 221.8 & 6.25 & 6.04 & 5.9 + comp 4 & 19:02:17.69 + 38:24:13.4 & 10.21 & 355.6 & 3.59 & 3.01 & 2.6 + [ -1.5ex ] additionally , kepler-9 was observed on 2010 july 2 at the palomar hale 200-inch telescope with the near - infrared adaptive optics ( ao ) pharo instrument @xcite , a @xmath25 rockwell hawaii hgcdte array detector .
observations were made in the @xmath26 ( 1.25@xmath3 m ) and @xmath27 ( 2.145@xmath3 m ) bands .
the field of view was approximately @xmath28 , and the scale was 25.1mas per pixel .
the ao system guided on the primary target itself , and produced strehl ratios of 0.05 at @xmath26 and 0.3 at @xmath27 .
the central cores of the resulting point spread functions had widths of fwhm @xmath29 at @xmath26 and fwhm @xmath30 at @xmath27 .
the closer of the companions seen earlier in the keck images were easily detected , and we list them all in table [ tab : companions ] along with relative positions ( angular separations and position angles ) , relative brightness estimates , and other identifications . the sensitivity to
faint companions was studied by injecting artificial stars into the image at various separations and with a range of @xmath31 , and then attempting to detect them both by eye and with an automated idl procedure based on daophot . for firm detection
we required the artificial stars to be present in both passbands .
the sensitivity curves as a function of angular separation are shown in figure [ fig : aosensitivity ] , along with the @xmath6-band sensitivity estimated from the speckle observations .
much fainter stars with @xmath32 near a target could in principle be detected by examining images from the palomar observatory sky survey , which date back more than 50 years , provided the proper motion of the target is large enough to have shifted it by several arc seconds over that period .
this is not possible for kepler-9 , since its total proper motion as reported in the ucac2 catalog @xcite is only 13.7 mas yr@xmath33 .
the likelihood of such faint close - in companions must therefore be addressed statistically , if need be . while the ao and speckle observations rule out the presence of bright neighboring stars as close as 01 or slightly less , further limits on even tighter companions can be placed by the spectroscopic observations obtained with hires on the keck i telescope , described by @xcite , since those stars would fall well within the 086 slit of the spectrograph .
we performed simulations in which we added the spectrum of a faint star to the original kepler-9 spectra , over a range of relative brightnesses , and attempted to detect these artificial companions by examining the cross - correlation function .
we estimate conservatively that any such stars with relative fluxes larger than about 1015% ( @xmath31 less than 22.5 mag ) would have been seen , unless their spectral lines are blended with those of the target .
the sharp lines of kepler-9 , with a measured rotational broadening of only @xmath34 , make this rather unlikely .
thanks to the very high astrometric precision of , an analysis of the motion of the photocenter of a target provides an effective way of identifying false positives that are caused by background eclipsing binaries falling within the aperture .
the principles have been explained by @xcite ( see also * ? ? ?
* ; * ? ? ?
the centroid measurements described below use data from quarter 3 only . in quarter 1
the kepler-9 aperture was determined to be too small to optimally capture its flux , and was subsequently enlarged . in quarter 2 experienced undesirable pointing drift , which was later resolved .
these problems complicate the centroid analysis for quarters 1 and 2 , although the results are broadly consistent with the more reliable ones from quarter 3 presented here .
we describe first the use of difference image analysis to demonstrate that the transit sources for all three kepler-9 planets and candidates are restricted to being very near the target star .
a difference image is formed by averaging several exposures near , but outside of a transit and subtracting from this the average of all available exposures near transit center .
this results in a typically isolated signal , a positive intensity with the shape of the point spread function ( psf ) at the _ true spatial location _ of the transit source , and an amplitude equal to the photometric transit depth times the direct image intensity for the target . adopting 40 independent transits of koi-377.03 in quarter 3 ( avoiding those shortly after major disturbances such as a safing event , and avoiding any that overlap with ` b ' and ` c ' transits ) ,
each formed with six symmetrically placed exposures outside of transits ( after a two exposure gap ) and three near transit minimum , results in a 14-@xmath24 signal in the difference image .
the corresponding direct image is formed as the average of both in- and out - of - transit sets such that the direct and difference images are sums and differences of precisely the same exposure sets . for kepler-9b four transits were used from quarter 3 with five exposures in transit , a gap of three , then five more exposures on each side for out - of - transit .
kepler-9c used two transits with seven exposures in - transit , a gap of three , and seven symmetrically placed out - of - transit exposure blocks . by using only exposures pulled close in time , and symmetrically with respect to the transits in use to form a difference image , this effectively imposes a de - trending and avoids any complications from drifts on time scales longer than the average spread of the out - of - transit sets , which for kepler-9c ( the widest ) is about 9 hours .
inspection of the difference images in figure [ fig : differenceimages ] shows that the transit sources for the confirmed ` b ' and ` c ' planets @xcite and the candidate koi-377.03 must arise from close to the target star , with offsets approaching one pixel easily ruled out by inspection . a weighted psf fit ( or more properly , a pixel response function ( prf ) fit ; see * ? ? ?
* ) to each of the direct and three difference images of figure [ fig : differenceimages ] is formed using only the central 3@xmath203 pixel area centered on the brightest pixel .
this leads to offsets with respect to ` b ' , ` c ' , and koi-377.03 of 0.007 , 0.035 , and 0.047 pixels , respectively . for koi-377.03
the formal error from a weighted least squares fit is 0.062 pixels .
we have further assessed the errors by generating a large number of independent realizations of a transit signal of the koi-377.03 relative intensity centered on the target coordinates .
this leads to an _ rms _ scatter of 0.062 pixels .
the noise had been increased by a factor of 1.2 in the difference image beyond direct poisson plus readout noise estimates in order to yield this congruence of least - squares errors and scatter in simulations .
the distribution of offsets follows expected gaussian statistics , e.g. , in the 7472 trials for koi-377.03 the extreme offset is 4.2@xmath24 compared to the expected 4.0 .
we have also shown that simulating transit signals at 0.5 and 1.0 pixel offsets from the target results in similar and smaller statistical scatter , respectively , as less poisson noise is under the transit image .
we take the scatter of 0.062 pixels to generate a 3-@xmath24 error circle of 0.186 pixels , or 074 .
this is the minimum radius within which background eclipsing binaries can not be safely ruled out from centroid analysis of the data itself . to place this in perspective :
centroid analysis has ruled out 98.6% of the area within the 8-pixel optimal aperture ( @xmath3599.6% of the 31-pixel mask ) of figure [ fig : guider ] as the location of potential background eclipsing binaries creating the koi-377.03 signal .
the quantitative results for all three transit sets are given in table [ tab : differenceimages ] .
kepler-9b shows an offset of 0.0074 pixels between the difference and direct image relative to a 1-@xmath24 error of 0.0049 pixels .
a 3-@xmath24 error circle in which background binaries can not be excluded from the centroid analysis of data itself is only 006 .
kepler-9c is the only case of the three showing a formal inconsistency with the offset being 5@xmath24 ; however , even if we combine the offset and 3-@xmath24 formal error any background eclipsing binaries outside of a radius of 022 are excluded as the transit signal source . clearly for all three transit sets , with the 3-@xmath24 error circles comfortably under 1 all of the known companions from high - resolution imaging shown in table [ tab : companions ]
are safely ruled out as sources of the photometric transit signal .
it is worth noting that the formal ( and equal to scatter from monte - carlo simulations ) error on radial offsets is approximately equal in pixel units to the inverse photometric signal - to - noise ( s / n ) ratio , as expected ( see , e.g. , * ? ? ?
lcccc + direct & @xmath36 & 1099.6989 & 273.4557 & + difference & @xmath37 & 1099.7058 & 273.4584 & + s / n ; offset & 174 & 0.0069 & 0.0027 & 0.0074 + errors & 2506 & 0.0029 & 0.0039 & 0.0049 + + direct & @xmath38 & 1099.6999 & 273.4654 & + difference & @xmath39 & 1099.7107 & 273.4990 & + s / n ; offset & 128 & 0.0108 & 0.0336 & 0.0353 + errors & 2991 & 0.0039 & 0.0056 & 0.0068 + + direct & @xmath40 & 1099.6946 & 273.4408 & + difference & @xmath41 & 1099.7131 & 273.4836 & + s / n ; offset & 14.3 & 0.0185 & 0.0428 & 0.0047 + errors & 1046 & 0.0351 & 0.0511 & 0.0062 + [ -1.5ex ] further confirmation that at least the two deeper signals seen in kepler-9 are not due to known stars in the scene can be obtained by placing simulated eclipses on the known stars in the aperture , and comparing them with the observations .
the scene in the aperture is modeled using stars in the input catalog ( kic ; * ? ? ?
* ) , supplemented by the stars in table [ tab : companions ] . all stars within a prf size ( 15 pixels ) in row or column of kepler-9 s aperture are included . to generate the modeled out - of - transit image ,
the measured prf is placed at each star s location on the focal plane , scaled by that star s flux .
this provides the contribution of each star to the flux in the aperture s pixels . for each star @xmath42 in the aperture ,
the depth @xmath43 of a transit is computed that reproduces the observed depth in the aperture pixels . an in - transit image for each @xmath42 is created as in the out - of - transit image , but with the flux of @xmath42 suppressed by @xmath44 .
these model images are subject to errors in the prf @xcite , so they will not exactly match the sky .
a flux - weighted centroid is computed for the out - of - transit image and the in - transit - image generated for each star in the aperture .
this produces row and column centroid offsets @xmath45 and @xmath46 , and the centroid offset distance @xmath47 .
lccc @xmath45 & @xmath48 & @xmath49 & @xmath50 + @xmath46 & @xmath51 & @xmath52 & @xmath53 + @xmath54 & @xmath55 & @xmath56 & @xmath53 + @xmath57 & @xmath58 & @xmath59 & @xmath60 + [ -1.5ex ] lcc|lcc|lcc[b ! ] kepler-9b & @xmath61 & 16.0 & kepler-9c & @xmath62 & 11.9 & koi-377.03 & @xmath63 & 0.86 + comp 1 & depth @xmath64 & & comp 1 & depth @xmath64 & & comp 1 & @xmath65 & 2.49 + comp 2 & @xmath66 & 127 & comp 2 & @xmath67 & 94.6 & comp 2 & @xmath68 & 6.89 + comp 3 & depth @xmath64 & & comp 3 & depth @xmath64 & & comp 3 & @xmath69 & 9.16 + comp 4 & @xmath70 & 188 & comp 4 & @xmath71 & 139 & comp 4 & @xmath72 & 10.1 + [ -1.5ex ] to compare these modeling results with observation we must make low - noise centroid measurements from the observed pixel data .
we do this by creating out - of - transit and in - transit images from de - trended , folded pixel time series . for each pixel time series ,
the de - trending operation has three steps : 1 ) removal of a median - filtered time series with a window size equal to the larger of 48 cadences or three times the transit duration ; 2 ) removal of a robust low - order polynomial fit ; and 3 ) the application of a savitzky - golay filtered time series of order 3 with a width of 10 cadences .
the savitzky - golay filter is not applied within 2 cadences of a transit event , so the transits are preserved .
the resulting pixel time series are folded with the transit period .
each pixel in the out - of - transit image is the average of 30 points taken from the folded time series outside the transit , 15 points on either side of the transit event .
each pixel in the in - transit image is the average of as many points in the transit as possible : seven for kepler-9b and kepler-9c , and four for koi-377.03 .
centroids are computed for the in - transit and out - of - transit images in the same way as the modeled images .
uncertainties of these centroids are estimated via monte carlo simulation , where a noise realization is injected into 48-cadence smoothed versions of the pixel time series for each trial .
a total of 2000 trials are performed each for kepler-9b , kepler-9c , and koi-377.03 .
the in- and out - of - transit images are formed using the same de - trending , folding and averaging as the flight data .
the measured uncertainties are in the range of a few times @xmath73 pixels .
table [ tab : measured_centroids ] shows the resulting measurements of the centroids from quarter 3 pixel data , along with the monte - carlo - based @xmath74 uncertainties .
the centroids are converted into centroid offsets and offset distance with propagated uncertainties .
table [ tab : modeled_offsets ] shows the offset distance @xmath54 predicted by the modeling method described above for each target in the aperture .
we see that when the transit is on kepler-9 itself we expect a measurable centroid shift for kepler-9b and kepler-9c . in this case
the modeled centroid shift is about 3.7 times larger than that observed , though the signs of the offsets agree .
this exaggeration of the centroid offset has been traced to inaccuracies in the kic used to create the model images .
therefore , the centroid shifts in table [ tab : modeled_offsets ] should be scaled by a factor 1/3.7 .
if the transit were on one of the companion stars in the aperture , then the modeled centroid shift would be an order of magnitude larger than observed for kepler-9b and kepler-9c , ruling out the companion stars as the source for these signals .
companion stars are not as definitively ruled out for the koi-377.03 transit by this technique . after scaling the centroid offsets as above , modeled transits on companions 3 and 4 have offsets that are are about @xmath75 , while companion 2 is @xmath76 and companion 1 is less than @xmath74 .
the modeled transit on kepler-9 , however , is much smaller , consistent with the observed transit offset for koi-377.03 . as an initial test
, we modeled the light curves for each of these two signals assuming they are the result of an eclipsing binary physically associated with the target , i.e. , at the same distance ( hierarchical triple ) .
for this case the isochrone for the binary was taken to be the same as that of the primary , and corresponds to [ fe / h ] = + 0.12 and an age of 1 gyr .
the secondary and tertiary masses were allowed to vary freely between 0.10@xmath4 ( the lower limit in the models ; see footnote [ foot : models ] ) and 1.40@xmath4 , as mentioned earlier , seeking the best fit to the photometry .
the inclination angle was also free , and the orbits were assumed to be circular . in both kepler-9b and c , which have similar transit signals , we find that the best fitting hierarchical triple blend model corresponds to secondaries that are approximately 1.0 and 0.5 mag fainter than the primary , respectively , and tertiaries that are at the lower limit of the isochrone range ( late m dwarfs ) .
however these fits give a poor match to the photometry : is unable to simultaneously reproduce the total duration of the transit and the central depth , given the constraints on the brightness and size of the stars from the isochrones .
this type of blend scenario is therefore clearly ruled out .
we illustrate this for kepler-9b in figure [ fig : koi377b_ht ] , where the best - fit planet model is also shown for reference .
much better matches to the data can be found if additional light from a fourth star along the line of sight is incorporated into the model , providing extra dilution .
we find that this fourth star is required to be nearly as bright as the primary , and the optimal model changes in such a way that the secondary also becomes as bright as the primary ( so that its size enables the duration of the transits to be reproduced ) , while the tertiary remains a small star .
this rather contrived scenario requiring two bright stars that are nearly identical to the main star would be easily recognized in our high - resolution imaging for separations larger than about 01 ( see , e.g. , figure [ fig : aosensitivity ] ) , in our centroid analysis for separations larger than 006 , or would otherwise produce obvious spectroscopic signatures unless all three bright objects happened to have the same radial velocity .
we next considered blends involving eclipsing binaries in the background , by removing the constraint on the distance . in this case
a solar - metallicity isochrone was adopted for the binary , with a representative age for the field of 3gyr .
we explored a wide range of relative distances , and we first considered main - sequence stars only , again with circular binary orbits .
the results for kepler-9b and c are once again similar to each other , and we illustrate them for kepler-9c in figure [ fig : koi377c_backstar ] .
the axes correspond to the distance modulus difference @xmath5 as a function of the tertiary mass .
contours represent constant differences in the @xmath7 of the fit compared to the best - fit planet model , and are labeled in units of the statistical significance of the difference ( @xmath24 ) .
we draw two main results from this figure .
one is that the light curve fits strongly prefer the smallest available tertiary masses from the isochrones ( 0.10@xmath4 ) , and would in fact yield better fits for even smaller tertiaries ( i.e. , planets ) .
additionally , the best solutions cluster toward equal distances for the binary and the primary star , effectively converging toward the equivalent of the hierarchical triple scenario considered earlier .
no acceptable solutions exist with the binary at a significant distance behind the primary star .
the best fit to the light curve of kepler-9c is similar to the one shown in figure [ fig : koi377b_ht ] ( dashed curve ) , which is not particularly good .
the @xmath5 vs.tertiary mass diagram for kepler-9b is qualitatively the same . allowing the secondary to be a giant star gives a very poor fit to the photometry : the duration of the transit is very much longer than observed , there is out - of - eclipse modulation due to distortions in the giant , and all solutions place the binary at an implausibly large distance .
we conclude that blend configurations involving background eclipsing binaries in which the tertiary is a star are not a viable explanation for either of these two signals .
we then explored background eclipsing binaries in which the tertiaries are planets rather than stars .
this allows their radii to be smaller , possibly providing a better fit to the photometry .
the orbits were considered to be circular , as before .
figure [ fig : koi377c_backplanet ] shows the results for kepler-9c , this time in the plane of separation versus secondary mass .
once again the fits tends to favor an equal distance for the binary and the primary star , and background scenarios with the binary far behind provide unacceptably poor matches to the light curve .
a second noteworthy result is that these solutions have a strong preference for secondary stars that are quite similar to the primary .
all acceptable fits to the light curve correspond to relatively bright secondaries with @xmath77 mag ( see figure [ fig : koi377c_backplanet ] ) .
the best of these solutions is of about the same quality as a planet model , and has a secondary of mass 0.98@xmath4 that is only 100k cooler and 0.3 mag fainter than the primary in the band .
this somewhat artificial case of `` twin '' stars is a result we have seen often in simulations for other candidates .
the tertiary in this type of blend solution comes out about @xmath78 larger than in a planet model because the transit is diluted by another star of approximately equal size and brightness .
one may debate whether this situation should actually be referred to as a `` false positive '' for kepler-9 , since the signal would still correspond to a gas giant planet , only that this planet would be @xmath78 larger , and it would be orbiting a different star . alternatively , it could be thought of simply as an overlooked dilution factor in a true planetary system . in any event ,
the lack of evidence for this bright twin star in the spectroscopy or in our high - resolution imaging or centroid analysis for kepler-9 does not support this scenario . as a particular case of this family of configurations
, we examined blends in which the star - planet pair is constrained to be at the same distance as the primary , i.e. , effectively in a hierarchical system .
the secondary properties were therefore taken from the same isochrone as the primary , and the orbits were assumed to be circular .
an excellent fit to the light curve is possible for a tertiary that is about @xmath78 larger than in a true planet model , but not surprisingly , we find once more that the secondary must be as bright as the primary .
additional tests were run to examine the impact of changing the age adopted for the isochrone of the secondary in a background star - planet pair , or the addition of light from a fourth object in the aperture . in the first case ,
changing the age from 3gyr to 1gyr produced a small shift of the contours in figure [ fig : koi377c_backplanet ] downward and to the right that is simply due to the change in intrinsic brightness of the secondary star , and does not alter our conclusions . adding `` fourth light '' further attenuates the eclipses of the star - planet pair . to compensate , requires a slightly deeper eclipse , and in order to preserve the shape of the signal ( total duration , and slope of ingress / egress ) , this is achieved by bringing the secondary closer to the primary .
as a result , for relatively bright fourth light the contours are shifted downward by approximately the difference in magnitude between the primary and the fourth star , again without changing the conclusions .
one may also imagine blend scenarios in which the eclipsing binary is in the foreground , rather than the background .
we explored this possibility by extending the simulations to negative values of @xmath5 .
as before , we adopted circular orbits and a 3gyr isochrone for the foreground system .
binaries with stellar tertiaries are clearly ruled out as they yield fits to the light curve that do not match its shape , and additionally they predict a fairly obvious secondary eclipse that is not seen in the data .
we focus therefore on blends in which the tertiary is a planet , and we illustrate the results for kepler-9c . in this case
we find there are many acceptable solutions with @xmath7 values differing from the best planet fit at the level of 1@xmath24 or less .
these solutions span a range of secondary masses and a range of foreground separations , implying a wide range not only in apparent brightness for the secondary , but also in color .
models in which the secondaries are brighter than the primary and of significantly different spectral type would be inconsistent with the spectroscopic parameters derived for kepler-9 , and are excluded .
plausible solutions remain , in principle , for fainter foreground secondaries , which necessarily involve later - type stars .
we find that of these , the only ones that yield acceptably good fits to the photometry , with @xmath7 values differing from the planet fit by less than 3@xmath24 , correspond to secondaries that are within about 1.5 mag of the the primary in brightness , and are of course redder
. these would be valid blend configurations so long as the secondaries are close enough to the primary to be spatially unresolved ( angular separations @xmath79 ) , and at the same time faint enough to have gone undetected in the spectra .
stars that are within @xmath122.5 mag of the primary would generally have been seen spectroscopically , as indicated in sect .
[ sec : observations ] , and this would exclude these remaining foreground blend configurations . nevertheless , to be conservative , let us assume for the moment that a star 1.5 mag fainter than the primary has still managed to elude detection in our spectra .
this corresponds to the faintest secondary in a foreground blend scenario that still allows for a satisfactory fit to the light curve , and would be the most difficult case of this kind to disprove .
this fit is shown in the top panel of figure [ fig : colors ] , and is statistically indistinguishable from a planet model fit .
the secondary in this configuration is an m2 dwarf ( @xmath80@xmath4 ) 1.53 mag fainter than the primary , eclipsed by a 0.91@xmath8 planetary companion , and is located at a distance of 300 pc .
the primary in this scenario is at 750 pc .
other properties of this particular blend such as magnitudes and colors can be computed easily with , and compared with observations .
apparent magnitudes for kepler-9 are available from the kic for a variety of passbands including sloan @xmath81 , a special - purpose passband referred to as d51 ( centered on the b triplet at 518.7 nm ) , and @xmath82 from the 2mass catalog .
the lower panel of figure [ fig : colors ] shows various color indices ( @xmath83 ) predicted by both for the primary star alone and for the blend .
those of the primary are well reproduced by the model , and we find that a small amount of interstellar extinction leads to an even better match ( solid line in the figure ) .
the colors of the blend , on the other hand , disagree with the measured colors , and deviate by more than 0.4 mag for the reddest index , @xmath84 .
we are therefore able to exclude , solely on the basis of its color , this most difficult of the scenarios involving foreground star - planet pairs that could mimic the 19-day and 39-day signals in the light curve of kepler-9 .
larger - mass secondaries would not be as red and still allow for good fits to the photometry , but they are intrinsically brighter and would be recognized more easily .
the above simulations have all assumed circular orbits for the blended eclipsing binaries or star - planet pairs , which is not necessarily realistic given the relatively long periods of kepler-9b and c. eccentricity affects the speed of the secondary and tertiary in their relative orbit , and therefore can change the duration of the transit , making it shorter or longer than in a circular orbit , depending on the orientation ( longitude of periastron , @xmath85 ) .
it also changes the impact parameter , all else being equal . and finally , it shifts the location of the secondary eclipse . the magnitude of these effects is illustrated in figure [ fig : ecc ] for eccentricities between 0.1 and 0.7 .
the most important effect for our purposes is on the transit duration . given a fixed ( measured ) duration , eccentric orbits may allow blends with smaller or larger secondary stars than in the circular case to still provide satisfactory fits to the light curve , effectively increasing the pool of potential false positives .
the limiting cases correspond to @xmath86 and 270 , in which the line of apsides is aligned with the line of sight and the transit occurs at periastron ( accommodating larger secondaries ) or apastron ( smaller secondaries ) , respectively .
extensive simulations for these two extreme situations show that allowing for eccentric orbits does not change our conclusions regarding hierarchical triple systems , background eclipsing binaries , or background star - planet scenarios .
we show this for the latter blend category in figure [ fig : backplanet_ecc ] , illustrated for the case of orbits with eccentricities of 0.3 and 0.5 , and @xmath86 .
comparison with figure [ fig : koi377c_backplanet ] indicates that in both cases the blends are still bright enough that we would have seen signatures of them in the spectra of kepler-9 .
larger eccentricities of @xmath87 result in secondaries that are brighter still . for eccentric orbits oriented such that transits take place at apastron ( @xmath88 ) , we only find acceptable fits to the light curves for eclipsing star - planet pairs that are in the foreground ( and involve smaller stars ) .
however , as was the case for circular orbits , those blends are either too bright , too red , or both , and are thus also excluded . 0.1 in lcc[b ! ]
hierarchical triple with stellar tertiary , ms & & + @xmath89 circular and eccentric orbits & poor fits / sec.ecl .
& yes + @xmath89 added fourth light & twin star & yes ( imaging / spec./centr . ) + [ + 2.0ex ] hierarchical triple with planetary tertiary , ms & & + @xmath89 circular and eccentric orbits & twin star & yes ( imaging / spec./centr . ) + [ + 2.0ex ] background eb with stellar tertiary & & + @xmath89 circular and eccentric orbits , ms and giants & poor fits & yes + [ + 2.0ex ] background eb with planetary tertiary , ms & & + @xmath89 circular and eccentric orbits & twin star & yes ( imaging / spec./centr . ) + @xmath89 1 gyr isochrone for secondary & little change & yes ( imaging / spec./centr . ) + @xmath89 added fourth light & little change & yes ( imaging / spec./centr . ) + [ + 2.0ex ] foreground eb with stellar tertiary , ms & & + @xmath89 circular and eccentric orbits & poor fits / sec.ecl . & yes + [ + 2.0ex ] foreground eb with planetary tertiary , ms & & + @xmath89 circular and eccentric orbits & too bright / too red & yes ( imaging / spec./centr./color ) + [ -1.5ex ] the above , fairly exhaustive exploration of parameter space with allows us to conclude that no configuration involving an eclipsing binary ( or an eclipsing star - planet pair ) , either in the foreground or in the background , is able to provide a reasonable explanation for the signals of kepler-9b and c ( see table [ tab : sumlarge ] for a summary of the configurations tested , and the results ) .
many scenarios lead to light curves that match the detailed shape of the transit events , but none are simultaneously consistent with all of the other observational constraints .
this includes spectroscopy , high - resolution imaging , centroid measurements , and photometry ( colors ) .
therefore , even ignoring the evidence from ttvs , these results fully support the planetary nature of these objects and demonstrate the usefulness of for validating transiting planet candidates from .
we proceed next to examine false positive scenarios for the shallowest signal in kepler-9 , with @xmath90 days , which would correspond to a super - earth - size planet . because the period is so short in this case , and tidal forces in such binary systems
have likely circularized the orbit ( see , e.g. , * ? ? ?
* and references therein ) , we do not consider non - zero eccentricities .
additionally , blends in which the secondary star is a giant need not be considered , as those cases are obviously ruled out because of the short orbital period and small implied semimajor axis of the orbit . as for the larger signals considered above , hierarchical triple systems in which the tertiary is a star fail to provide good fits to koi-377.03 . a good match to the photometry
can be found when the tertiary is allowed to be a much smaller object ( i.e. , a planet ) , but as was the case earlier , it requires a secondary that is very similar to the primary in brightness .
the resulting size of the eclipsing object is @xmath78 larger than in a planet model , or slightly over 2@xmath10 .
this type of configuration was ruled out earlier based on the high - resolution imaging and the spectroscopy .
small tertiaries with appreciable mass , such as white dwarfs , induce tidal distortions on the primary due to the short orbital period that lead to significant out - of - eclipse variations in the light curve ( ellipsoidal variability ) .
these modulations are not seen in the photometry for kepler-9 , and such false positives are therefore also excluded .
blends with an eclipsing binary in the background ( the tertiary being a star ) are able to reproduce the light curve just as well as a planet model . in figure
[ fig : k3backstar ] we illustrate those results by showing the area of allowed parameter space in a diagram of distance modulus difference as a function of secondary mass .
acceptable fits with @xmath7 differing from the planet model by less than 3@xmath24 are possible over a wide range of relative separations ( @xmath91 ) , but the secondaries are restricted to a relatively narrow interval in mass centered on the mass of the primary .
these eclipsing binaries are all very distant and faint ( @xmath9222 ) , and have no effect on the colors of the blend .
the more distant scenarios place the binary at implausibly large distances of up to 42 kpc ( more than 10 kpc above the galactic plane ) . the nearest configuration ( @xmath93 ;
see figure [ fig : k3backstar ] ) has the binary at a distance of 5.3kpc , and the primary at @xmath1670 pc .
the secondary in this model is a late g star 5.5mag fainter than the primary in the passband , eclipsed by a late m dwarf that produces no detectable secondary eclipse .
the predicted brightness of this binary precisely matches that of the closest companion identified in the ao images ( comp 1 , table [ tab : companions ] ) , located 285 ne of the target .
however , this and all wider visual companions are already ruled out at more than the 3-@xmath24 confidence level by the lack of centroid motion , which would have revealed any blended eclipsing binaries at angular separations larger than about 074 ( sect.[sec : centroids ] ) . even without this constraint from astrometry ,
the predicted @xmath94 color of the secondary in this blend is considerably redder than measured for this close ao companion , which would also disqualify it .
eclipsing binaries that are between 5 and @xmath18.5 magnitudes fainter than the main star provide acceptably good fits to the light curve ( see figure [ fig : k3backstar ] ) , and if they were angularly closer than 074 from the target they may not be detectable in our ao or speckle observations , in our centroid motion analysis , nor in our spectra .
they remain viable blend configurations , and would necessarily be at distances greater than 5 or 6kpc .
an example is shown in figure [ fig : k3blend ] , to illustrate that the fit is indistinguishable from a planet fit . in the above calculations we have ignored interstellar extinction .
however , given the large distances for the binaries in some of these blend configurations , it is worth exploring the effect of dust more carefully , which we have done by repeating the simulations using a representative differential extinction coefficient of 0.5 mag kpc@xmath33 .
the results are shown in figure [ fig : k3backstarextin ] , and indicate that the blend scenarios providing good fits to the photometry of koi-377.03 are systematically shifted to smaller distances compared to the previous calculations .
their apparent brightness , however , changes relatively little , as can be seen by comparing the lines of equal @xmath95 with those in figure [ fig : k3backstar ] .
therefore , the overall impact of differential extinction on the permitted area of parameter space in terms of observable parameters is not as significant as might have appeared . allowing the tertiary to be a smaller object such as
a planet opens up a different area of parameter space for permissible background blends ( figure [ fig : k3backplan ] ) .
when including extinction as before , acceptable fits to the light curve are possible for @xmath5 values from zero up to about 4.5 .
this upper limit corresponds to distances for the star - planet pair of about 4.8kpc ( with @xmath96 , or apparent magnitudes of @xmath97 ) , and is set by the maximum size of 1.8@xmath8 we have allowed for a planet . as in the configurations described before , the solutions constrain the secondary masses to be near that of the primary in order to match the detailed shape and observed duration of the transits , with a range from about 0.9@xmath4 to 1.2@xmath4 .
therefore , the color of the blend is not as useful a discriminant in this case .
secondaries that are less than about 2 mag fainter than the primary would have been seen spectroscopically .
this excludes a good fraction of the space of parameters , as indicated by the lower dashed line in figure [ fig : k3backplan ] . of the remaining blends of this kind between @xmath98 and @xmath99 , only the ones with angular separations smaller than about 1
are allowed by the constraints from our ao imaging ( see figure [ fig : aosensitivity ] ) , but the centroid analysis is even more restrictive and rules out stars outside of 074 . at closer separations ,
the high - resolution images rule out all blends that are brighter than the sensitivity limit indicated in figure [ fig : aosensitivity ] ( i.e. , those that fall above the curves ) . as expected from the fixed duration and depth of the transit - like signal of koi-377.03 ,
the size of the tertiary in these configurations correlates with the secondary mass . due to this correlation
, small tertiaries with @xmath100 ( roughly neptune - size , and smaller ) are further excluded because the eclipses they produce are already very shallow , and further dilution by the primary would make them too shallow to fit the photometry . in order to avoid this
, the secondaries in those blends must be relatively small late - g type stars that are nearby , and would therefore be bright enough ( @xmath101 ) that they would have been detected in our spectra as a second set of lines .
thus , effectively places limits not only on the secondary , but also on the size of the tertiary ( see figure [ fig : k3backplanr3 ] ) . in particular , blends with a white dwarf eclipsing a background star are also ruled out for the same reason described above .
additionally , the predicted light curves for such cases with white dwarf tertiaries show ellipsoidal variability , which is not observed.30 , which should still allow that star to be seen spectroscopically if it were bright enough . in any case ,
white dwarfs are excluded as viable tertiaries for the reasons mentioned in the text . ]
many of the larger tertiaries correspond to gas giants ( saturn - size , or larger ) , which implies a qualitative difference in their nature compared to the alternate model of a true earth- or super - earth - size planet . in this sense these blends may properly be considered `` false positives '' , as opposed to the configurations discussed earlier requiring twin stars , which only change the tertiary radius by @xmath78 .
finally , we examine the possibility that the true period of the koi-377.03 signal is twice the nominal value .
alternating events would then correspond to the primary and secondary eclipses of a blended eclipsing binary ( the tertiary being a star in this case ) , which may in general be of different depth . in koi-377.03 there is no compelling evidence for a depth difference between odd- and even - numbered events , but this is difficult to establish in a faint star such as this for a signal that is only 0.2mmag deep .
the results of extensive simulations with for this scenario are illustrated in figure [ fig : k3backstar2p ] .
the best fits correspond to blended binaries far in the background , and do not indicate a significant difference in depth between the primary and secondary eclipses .
however , these fits provide only a poor representation of the light curve , and can therefore be confidently ruled out .
this is seen in figure [ fig:2p ] .
the top panel shows the closest fit to the full light curve together with the data , and in the bottom panel we have binned the measurements to facilitate the comparison .
this solution involves an eclipsing pair of mid - g dwarfs at 21kpc and 7.2 mag fainter than the primary , and the fit is visibly worse than that corresponding to a planet at half the period , which is shown with the dashed line .
in summary , the analysis of this section coupled with constraints from spectroscopy , high - resolution imaging , and centroid motion measurements rules out a large fraction of the false positives that could produce the 1.59-day signal , but not all ( see table [ tab : sumsmall ] ) .
the remaining configurations involve main - sequence background stars that are similar to the primary in spectral type ( late f to early k , or about 0.9@xmath4 to about 1.2@xmath4 ) , and are eclipsed either by another main - sequence smaller star or by a planet with @xmath102 ( neptune - size or larger ) .
these blends range in apparent brightness from @xmath103 to @xmath104 for stellar tertiaries , and from @xmath105 to @xmath97 if the tertiaries are planets , and must be closer than 074 from the target . at separations under 074
our imaging observations allow us to rule out the brighter of these blends , and only the ones with @xmath31 below the sensitivity curves in figure [ fig : aosensitivity ] would remain undetected .
lcc hierarchical triple with stellar tertiary , ms & poor fits & yes + @xmath89 white dwarf tertiaries & poor fits & yes + [ + 2.0ex ] hierarchical triple with planetary tertiary , ms & twin star & yes ( imaging / spec./centr . ) + [ + 2.0ex ] background eb with stellar tertiary , ms & & + @xmath89 with and without extinction & good fits & not all + @xmath89 giant secondaries & @xmath106 too short & yes + @xmath89 white dwarf tertiaries & poor fits & yes + @xmath89 twice the period & poor fits / sec.ecl . & yes + [ + 2.0ex ] background eb with planetary tertiary , ms & & + @xmath89 jupiters , neptunes , super - earths , with extinction & good fits & super - earths +
@xmath89 giant secondaries & @xmath106 too short & yes + @xmath89 white dwarfs , with extinction & good fits & yes ( imaging / spec./centr . ) + [ + 2.0ex ] foreground eb with stellar tertiary , ms & poor fits & yes + [ + 2.0ex ] foreground eb with planetary tertiary , ms & poor fits & yes + [ -1.5ex ] in the previous section we have considered a wide variety of possible blend scenarios for koi-377.03 involving secondaries of different spectral types ( main - sequence stars and giants ) , eclipsing objects of both planetary and stellar nature ( including white dwarfs ) , and configurations consisting of chance alignments with a foreground or background contaminant , as well as hierarchical triple systems . while these represent the most common and obvious configurations one can imagine , in principle there could also be more contrived scenarios that we have not thought of .
these should be intrinsically much less likely , a priori , but can nevertheless not be completely ruled out .
therefore , we proceed below on the assumption that any such situations we have not considered have a small rate of occurrence , at least compared to the ones we have discussed explicitly . in order to provide the basis for an estimate of the confidence level for the planetary status of koi-377.03 ,
we describe here the calculation of the likelihood that the signal is due to a background blend involving either a stellar tertiary or a planetary tertiary , taking into account the constraints on brightness and other properties indicated in the previous section . because this type of calculation is likely to be relevant for other candidates ,
we describe it here in some detail .
the frequency of stars in the mass range permitted by was estimated using the besanon galactic structure models of @xcite , specifically for the @xmath6 band , which is the closest available to the passband .
we used an aperture of 1 square degree centered on kepler-9 , and we performed the stellar density calculations in half - magnitude bins of apparent brightness , accounting for interstellar extinction as we did in the simulations , with a coefficient of 0.5 mag kpc@xmath33 in @xmath22 .
the range of allowed secondary masses for each magnitude bin was taken directly from figure [ fig : k3backstarextin ] for blends with stellar tertiaries , and from figure [ fig : k3backplan ] for blends with planetary tertiaries . using the density of stars in each magnitude bin , we calculated the fraction that would remain undetected after our high - resolution imaging ( speckle and ao observations ) , spectroscopy , and centroid motion analyses .
the results are listed in table [ tab : statistics ] .
the first two columns give the @xmath107 magnitude range of each bin and the magnitude difference @xmath95 compared to the target , calculated at the upper edge of the magnitude bins . for convenience the calculations for blends with stellar tertiaries and
planetary tertiaries are listed separately . for the stellar tertiary case ,
column 3 reports the density of stars obtained from the besanon models , restricted to the mass range allowed by as shown in figure [ fig : k3backstarextin ] .
column 4 lists the maximum angular separation @xmath108 at which stars in the corresponding magnitude bin would go undetected in our imaging observations , read off from figure [ fig : aosensitivity ] , and taken at the center of each magnitude bin .
centroid motion analysis rules out eclipsing binaries beyond 074 , so @xmath108 is constant at that value for the last few bins in which this provides a stronger constraint than the imaging limits . the total number of stars of the appropriate mass range in a circle of radius @xmath108 around kepler-9 is then given for each bin in column 5 , in units of @xmath109 .
we note that the secondary mass and angular separation constraints together reduce the number of background stars to be considered as potential contaminants by a factor of @xmath110 compared to the number that would otherwise be expected to fall within the photometric aperture of .
this is already indicative of a significantly reduced chance of having a false positive .
the intrinsic frequency of eclipsing binaries in the field is a key ingredient in the calculation , and for this we have relied on the results of @xcite , which are based on the observations themselves .
these authors found the average occurrence rate of eclipsing binaries among the targets down to @xmath111 to be approximately 1.2% across the entire field .
this may be a slight overestimate because it counts as eclipsing binaries targets that are actually blended with a background binary that is not a target , though the effect is likely small .
there is little information available on fainter eclipsing binaries , so we assume here that a similar frequency holds .
more importantly , many of these eclipsing binaries can not produce signals such as that of koi-377.03 because their light curves have the wrong shape .
examples include contact binaries and ellipsoidal variables , in which the brightness changes continuously throughout the cycle , rather than presenting sharp transit - like events such as we observe .
additionally , semi - detached systems would have eclipses that are too long and also of the wrong shape .
we therefore exclude these from the tally . with this adjustment ,
the frequency of eclipsing binaries capable of producing blends is 0.53% . multiplying the star counts in column 5 by this frequency
, we obtain the total number of blends expected in each magnitude bin , which is reported in column 6 .
columns 79 are similar to columns 35 , but for blends involving star - planet pairs .
note that the range of allowed magnitudes is different in this case , as is the range of secondary masses used to compute the densities in column 7 ( see figure [ fig : k3backplan ] ) . following the size ranges adopted by @xcite , we consider three categories of transiting planets as potential companions : super - earths ( 1.252@xmath10 ) , neptune - size planets ( 26 @xmath10 ) , and jupiter - size planets ( 622 @xmath10 , or equivalently @xmath10.52.0 @xmath8 ) .
planetary tertiaries smaller than 0.3@xmath112 are ruled out by , as false positives with such tertiaries can only reproduce the light curve if the secondaries are relatively bright , and those would have been detected spectroscopically .
this effectively eliminates all super - earths , and a fraction of the neptunes . for the intrinsic frequency of transiting planets we have relied on the results from the first 43 days of observations as reported by @xcite . that census is unlikely to have missed many neptune- or jupiter - size planets ( except ones with very long periods ) , although it may include false positives ,
so we consider the count to be conservative .
those authors presented a list of 306 targets with at least one transiting planet candidate , and described another 400 targets with one or more transit - like signals that have not yet been released .
of the 306 targets , 24% would correspond to jupiter - size planets , and 23% to neptune - size planets .
one may reasonably assume that the 400 sequestered targets contain a larger fraction of smaller earth- or super - earth - size planets ( which can not mimic the light curve of koi-377.03 ; see sect .
[ sec : koi377.03 ] ) , for at least two reasons .
smaller planets are the main focus of the mission , and they require more intensive follow - up efforts for validation , which is why those targets have not yet been made public . secondly , these 400 targets are brighter , which makes smaller planets around them easier to detect . consequently , the assumption of similar jupiter- and neptune - size planet frequencies as given above , but for the entire sample of @xmath113 targets showing transit - like features , is a conservative one . scaling to the full sample of @xmath114 targets @xcite , we find an upper limit to the frequency of transiting jupiter - size and neptune - size planets of 0.11% and 0.10% , respectively . with these adopted frequencies , the resulting numbers of blends involving these types of planets are listed in columns 10 and 11 of table [ tab : statistics ] .
the total blend frequencies in each of columns 6 , 10 , and 11 are calculated as the sum of the individual frequencies in each magnitude bin , and the three columns are then combined in the bottom section of the table to yield an overall blend frequency ( bf ) of @xmath115 .
this very small figure corresponds to the number of false positives we expect to find a priori for kepler-9 .
however , we point out that this _ does not _ translate directly into a false alarm probability , or equivalently into a confidence level that the candidate is orbited by a true super - earth - size planet , as that requires knowledge of the rate of occurrence of such planets . for a random candidate star in the field the rate of false positives relative to the rate of true planets ( false alarm rate , far )
can be written quite generally as @xmath116 , where @xmath117 is the number of false positives and @xmath118 is the number of planets in the sample .
thus , the larger the number of planets we expect , the smaller the far .
we consider kepler-9 to be fairly representative of a typical target in the field in terms of its spectral type ( solar ) , brightness , and background stellar density ( a function of galactic latitude ) . in that case
, the total number of blends can be taken to be approximately the product of bf and the size of the sample , or @xmath119 .
the number of small planets expected in the sample is of course not known , and determining it is precisely one of the goals of the mission .
if we accept a confidence level of 99.73% ( 3@xmath24 ) as being sufficient for validation of a transiting planet candidate ( corresponding to @xmath120 ) , then the minimum number of super - earth - size planets @xmath118 required in order to be able to claim this level of confidence is six .
even though @xmath118 is unknown , it is possible to make educated guesses as to what the _ minimum _ value would be in several ways , drawing on both theoretical and observational considerations .
ground - based doppler surveys continue to push toward the detection of smaller and smaller planetary signals .
@xcite have reported preliminary results from a sample of some 400 fgk stars observed with the harps instrument on the eso 3.6 m telescope ( see also * ? ? ?
* ) , suggesting that as many as @xmath130% of the targets may be orbited by close - in super - earth- and neptune - mass companions ( 530@xmath121 ) with periods up to 50 days .
the peak in the period distribution seems to be around 10 days . by making use of the corotlux tool @xcite we simulated a sample of @xmath114 stars in the field based on the besanon models employed earlier , and used the results from @xcite to assign planets at random to each star , with a log - normal distribution of periods centered at 10 days .
approximate planetary radii were inferred from the masses using the structure models of @xcite , by drawing masses at random and assigning radii over the full range of compositions allowed by these models .
we then retained only those in the super - earth category , with @xmath122 .
the number of these planets that undergo transits in the field is calculated to be @xmath123 .
if we were to accept this estimate , the corresponding false alarm rate would be @xmath124 .
a similar doppler survey of 166 g and k stars with the hires instrument on the 10-m keck i telescope @xcite has provided additional insights into the rate of occurrence of small - mass planets .
the results suggest that approximately 18% of solar - type stars harbor planets in the range of 3 to 30@xmath121 with periods under 50 days .
a calculation analogous to that carried out above for the transit probabilities and conversion from planetary masses to planetary radii leads to an estimate of @xmath125 for the field .
if we adopted this lower estimate , the corresponding false alarm rate would be @xmath126 .
population synthesis studies such as those of @xcite and @xcite based on the formation of planets by the core accretion process and subsequent migration have provided tentative predictions of the properties of planets , including distributions of their masses , periods , and other characteristics .
these theoretical models seem to point to a sizable population of super - earths that may be several times larger than the number of neptune- or jupiter - mass planets in those simulations . scaled to the size of the sample
, this would imply there could be several hundred small transiting planets .
however , the authors caution that those results should be considered with great care as some of the physical ingredients in these models are still very uncertain .
a further estimate may be obtained from the preliminary results as reported by @xcite . among the 306 targets listed there showing one or more periodic transit - like signals , 27 fall in the category of super - earths . while it is true that these candidates have not yet been followed up and validated , and may therefore include some fraction of false positives , additional super - earth - size planets are to be expected in the list of 400 unreleased candidates , which could make @xmath118 considerably larger .
this is particularly true since the targets in the latter list are all brighter than those in the publicly available set , and therefore the proportion of small planets is likely to be higher because the transit events are easier to detect . nevertheless ,
if we were to accept that @xmath118 is as small as 27 , then the corresponding false alarm rate would be @xmath127 .
there are caveats associated with each of the observational estimates mentioned above that should be kept in mind .
the results invoked in the previous paragraph are still preliminary , and although we regard our use of them to be conservative for the reasons described earlier , the true fraction of false positives in the sample remains unknown until all candidates have been followed up .
the doppler results are also preliminary to some degree , and are based on somewhat limited samples of stars . it is also possible that a fraction of those doppler candidates may turn out to be false positives , or given the @xmath128 ambiguity inherent in that technique , that some of them may have actual masses above the range considered for super - earths . additionally , there are uncertainties associated with the conversion we have applied between planetary masses and planetary radii , using theoretical models .
those uncertainties are difficult to quantify given our present state of knowledge . for these reasons , added to the fact that despite our best efforts to assess the blend frequency there could still be some exotic blend scenario that we have overlooked , it is not possible to present a more definitive value of the false alarm rate . nevertheless , the above estimates of the far based on consideration of all the blend scenarios that seem plausible to us
are all sufficiently small that they give us very high confidence that koi-377.03 is not a false positive , and they therefore validate it as a signal of planetary origin .
we designate this planet kepler-9d .
calculating the false alarm rate for targets with small signals such as kepler-9d is non - trivial because it depends crucially on the frequency of small transiting planets , which may only be fully known at the conclusion of the mission . of the arguments for the expected value of @xmath118 presented in the previous section ,
the one that relies on the preliminary results themselves is the most conservative , and already makes it highly unlikely that we are in the presence of a false positive such as those explored in this paper .
furthermore , that estimate is based on results from only the first @xmath143 days of operation of the spacecraft
. continued observations over the next two years will surely increase the number of candidates , which can only result in a larger confidence that the 1.6-day photometric signal is due to a true planet .
thus , we find the overall evidence for the planet interpretation very compelling .
it is also worth noting that the doppler surveys have found that a very large fraction of the smallest - mass planetary companions ( as many as 80% ) are in multi - planet systems ( e.g. , * ? ? ?
* ; * ? ? ?
because of the presence of kepler-9b and c , this makes it considerably more likely that kepler-9d is also a planet than if it were the only signal in the system .
furthermore , one may expect a priori that a planet interior to kepler-9b and c would have a high probability of presenting transits . indeed , with the reasonable assumption that the orbit of the inner planet is more or less coplanar with the outer two , the geometric probability of a transit at a period of 1.6 days would be close to 100% , instead of @xmath118% for random inclinations .
the light - curve parameters we obtain for kepler-9d by modeling the photometry using the formalism of @xcite are summarized in table [ tab : kepler9d ] , and supersede the preliminary estimates of @xcite .
the values and uncertainties were determined using a markov chain monte carlo technique , with four chains of length @xmath109 each .
our fits used the non - linear fourth - order limb - darkening law of @xcite , with coefficients for the band taken from the calculations by a. prsa referenced in footnote [ foot : prsa ] . for an adopted stellar radius for the primary star of @xmath15 , the estimated size of this planet
is @xmath129 , which is among the smallest yet reported.@xmath10 according to @xcite , and was revised to @xmath130 by @xcite .
the measured radius of the next smallest known planet , gj1214b , is @xmath131@xmath10 @xcite . ]
the uncertainly is currently dominated by the photometric errors , rather than the stellar parameters , and should improve as more measurements are gathered .
the impact parameter of kepler-9d is very poorly constrained by current data , but should also become better determined in the future . a mass determination for this object has not been made , and will be challenging given the small amplitude expected for the reflex motion of the star
. the radial - velocity semi - amplitude would be only about 2.3 m s@xmath33 assuming a similar mean density as the earth .
velocity measurements are further complicated by the presence of the other two planets in this system .
lc orbital period ( days ) & @xmath132 + [ + 0.5ex ] mid - transit epoch ( bjd ) & @xmath133 + [ + 0.5ex ] orbital semimajor axis ( au ) & @xmath134 + [ + 0.5ex ] transit duration ( hours ) & @xmath135 + [ + 0.5ex ] @xmath136 & @xmath137 + [ + 0.5ex ] @xmath138 ( @xmath10 ) & @xmath139 + [ + 0.5ex ] @xmath140 & @xmath141 + [ + 0.5ex ] equilibrium temperature ( k ) & @xmath142 + [ -1.5ex ] many of the most interesting candidates found by will correspond to earth - size planets , some of which are expected be in the habitable zone of their host star . for solar - type stars this implies reflex motions with radial - velocity amplitudes below current detection limits , making the spectroscopic measurement of the mass impossible .
other reasons may also hinder this type of validation , such as rapid rotation , chromospheric activity , or even the faintness of the star .
the information contained in the light curves on the shape of a transit - like event is a valuable asset for constraining the vast range of possible astrophysical false positives that might be masquerading as a planet .
here we have shown how modeling the light curve of a candidate directly as a false positive allows to rule out a significant fraction of blends involving background or foreground eclipsing binaries or star - planet pairs , as well as hierarchical triple systems , each of these possibly attenuated by the light of additional stars .
the combination of with follow - up observations consisting of high - resolution imaging , spectroscopy , and an analysis of the centroid motion of the target leaves no room for a false positive in the case of kepler-9b and c. these signals were previously known to correspond to bona - fide saturn - size planets because they display correlated ttvs @xcite .
nevertheless , the exercise serves to show that , lacking that evidence , it would still be possible to validate them using the same techniques , thus supporting the general approach and justifying the application to the more interesting kepler-9d signal . among the advantages of for is the ability to predict the brightness and overall color of a blend in many different passbands .
brightness information for virtually every target is available from the kic in the sloan @xmath81 and 2mass @xmath82 bands , as well as in the custom d51 passband ( 518.7 nm ) .
we have shown earlier how these can be used to rule out certain blend scenarios that might otherwise be viable .
the detailed fitting of the photometry with a false - positive model provides additional discriminating power .
rough estimates of the properties of a background eclipsing binary that can mimic a blend have sometimes been made in previous transit surveys based simply on the apparent brightness of the object and a representative depth for its undiluted eclipses ( such as 50% ) .
while such configurations may well reproduce the observed amplitude of a candidate light curve , not much can be said about the expected shape , which may be completely wrong .
an example of this is seen for kepler-9d in figure [ fig:2p ] , in which a back - of - the - envelope calculation of the depth one might predict from the brightness and @xmath150% deep eclipses of this binary may not be far off , but the detailed shape is not a good match to the observations , and easily rules out this scenario giving a poor @xmath7 for the fit . without these additional constraints on blend properties provided by the detailed light - curve fitting , the space of parameters open to false positives would be significantly larger , and more difficult to exclude by other means .
funding for this discovery mission is provided by nasa s science mission directorate .
we are grateful to leo girardi for computing isochrones for this work in the passband , to frederic pont for very helpful discussions on false alarm probabilities , and to david sing for advice on limb - darkening coefficients .
we also thank the anonymous referee for insightful comments on the original version of this paper .
cccccccccccc 13.814.3 & 0.5 & & & & & & & & & & + 14.314.8 & 1.0 & & & & & & & & & & + 14.815.3 & 1.5 & & & & & & & & & & + 15.315.8 & 2.0 & & & & & & & & & & + 15.816.3 & 2.5 & & & & & & 563 & 0.14 & 2.675 & 0.0029 & 0.0027 + 16.316.8 & 3.0 & & & & & & 595 & 0.17 & 4.168 & 0.0046 & 0.0042 + 16.817.3 & 3.5 & & & & & & 588 & 0.20 & 5.701 & 0.0063 & 0.0057 + 17.317.8 & 4.0 & & & & & & 430 & 0.23 & 5.514 & 0.0061 & 0.0055 + 17.818.3 & 4.5 & & & & & & 292 & 0.25 & 4.424 & 0.0049 & 0.0044 + 18.318.8 & 5.0 & 2 & 0.35 & 0.059 & 0.0003 & & 125 & 0.35 & 3.712 & 0.0041 & 0.0037 + 18.819.3 & 5.5 & 14 & 0.45 & 0.687 & 0.0036 & & 57 & 0.45 & 2.798 & 0.0031 & 0.0028 + 19.319.8 & 6.0 & 14 & 0.55 & 1.027 & 0.0054 & & 24 & 0.55 & 1.760 & 0.0019 & 0.0018 + 19.820.3 & 6.5 & 20 & 0.70 & 2.376 & 0.0126 & & & & & & + 20.320.8 & 7.0 & 13 & 0.74 & 1.726 & 0.0091 & & & & & & + 20.821.3 & 7.5 & 3 & 0.74 & 0.398 & 0.0021 & & & & & & + 21.321.8 & 8.0 & 4 & 0.74 & 0.531 & 0.0028 & & & & & & + 21.822.3 & 8.5 & 0 & 0.74 & 0.000 & 0.0000 & & & & & & + 22.322.8 & 9.0 & 0 & 0.74 & 0.000 & 0.0000 & & & & & & + & 70 & & 6.804 & * 0.0359 * & & 2674 & & 30.752 & * 0.0339 * & * 0.0308 * + + [ -1.0ex ] | light curves from the mission contain valuable information on the nature of the phenomena producing the transit - like signals . to assist in exploring the possibility that they are due to an astrophysical false positive
, we describe a procedure ( ) to model the photometry in terms of a `` blend '' rather than a planet orbiting a star .
a blend may consist of a background or foreground eclipsing binary ( or star - planet pair ) whose eclipses are attenuated by the light of the candidate and possibly other stars within the photometric aperture .
we apply to the case of kepler-9 , a target harboring two previously confirmed saturn - size planets ( kepler-9b and kepler-9c ) showing transit timing variations , and an additional shallower signal with a 1.59-day period suggesting the presence of a super - earth - size planet . using together with constraints from other follow - up observations we are able to rule out all blends for the two deeper signals , and provide independent validation of their planetary nature . for the shallower signal we rule out a large fraction of the false positives that might mimic the transits .
the false alarm rate for remaining blends depends in part ( and inversely ) on the unknown frequency of small - size planets .
based on several realistic estimates of this frequency we conclude with very high confidence that this small signal is due to a super - earth - size planet ( kepler-9d ) in a multiple system , rather than a false positive .
the radius is determined to be @xmath0 , and current spectroscopic observations are as yet insufficient to establish its mass . |
with the wealth of new and upgraded experimental facilities for the b - physics , the precision test of standard model is ever more promising .
the new facilities of babar at slac and belle at kek generate asymmetric @xmath0 collisions to trace the b - decays more accurately .
the symmetric @xmath0 collision facility of cleo - iii at cornell has also upgraded its luminosity by a factor of ten better .
furthermore , the hadon - hadron collision facilities such as lhcb at cern and btev at fermilab as well as the lepton - hadron collision facility such as hera - b at desy emerge as the powerful tools to investigate a lot of detailed b - decays @xcite .
certainly , an important motivation to study b - physics is to make a precision test of standard model especially associated with the unitarity of ckm mixing matrix .
one of the burning questions in physics is whether the complex phase is really the only source of the cp - violation or not . to make such a precision test
, the accurate analyses of exclusive semileptonic b - decays as well as rare b - decays are strongly demanded . as we will discuss in this talk ,
an effective use of light - cone ( lc ) degrees of freedom in those analyses seem crucial to make the calculations more accurate .
this also makes the model - building more scrutinized .
our talk is presented with the following outlines .
the theoretical overview of b - physics is given in the next section , section 2 .
especially , we will discuss the processes determining each ckm - matrix element and the profile of unitary triangle .
we ll also try to make a very brief survey of theoretical development in the last twenty years history of b - physics .
because of enormous works that people have done in b - physics , this survey would be in no way complete but very brief and limited . in any case
, the purpose of section 2 will be to motivate why the exclusive semileptonic b - decays are very important to constrain the ckm mixing matrix . in section 3 ,
then exclusive semileptonic decays are discussed .
some general remarks on the weak form factors will be made and the role of lc degrees of freedom in the exclusive semileptonic decays will be discussed . especially , the difficulties associated with the time - like processes such as the exclusive semileptonic decays in the lc formulation .
we will first identify the embedded states necessary to restore the covariance of the amplitudes and then present a way of handling the embedded states .
some preliminary numerical results are also presented in the exclusive semileptonic decays for @xmath1 with our new way of treating the embedded states .
conclusions follow in section 4 .
in the standard model , the only interaction relevant to the ckm mixing matrix @xmath2 is the weak charged - current interaction given by the lagrangian @xmath3 , where @xmath4 in the standard model , @xmath2 is unitary , i.e. @xmath5 , and all the matrix elements of @xmath2 can be written in terms of three real angles @xmath6 and one real phase - angle @xmath7 as explicitly shown in the particle data group @xcite . here
, @xmath8 is the usual cabbibo mixing angle .
wolfenstein @xcite realized the pattern of order of magnitude in each element and parametrized @xmath2 in the orders of @xmath9 .
up to the order of @xmath10 , @xmath2 is given by @xmath11 where four real wolfenstein s parameters are given by ( @xmath12 ) .
the unitarity condition leads to the definition of various unitarity triangles : @xmath13 where each off - diagonal element @xmath14 with the unequal quark - flavors @xmath15 and @xmath16 , i.e. @xmath17 , is given by the addition of three complex numbers due to the multiplication of two 3@xmath183 complex matrices , e.g. @xmath19 , while the diagonal elements are given by the sum of three real numbers .
since the off - diagonal element @xmath14 ( @xmath20 ) is zero ( i.e. the sum of three complex numbers is zero ) , it can be given as a triangle in the complex plane .
thus , each off - diagonal element @xmath14 corresponds to a unitarity triangle and there are six independent ones .
only four out of eighteen angles in the six triangles are independent and the area of all triangles is identical measure of cp - violation , i.e. , @xmath21 .
also , the magnitude of each mixing matrix element @xmath22 is independent of parametrization . now , let s focus on the current magnitude @xcite of each mixing matrix element and the corresponding experimental processes used for the determination .
the magnitude of @xmath23 has been determined by the superallowed @xmath24 nuclear @xmath25-decay , the nucleon @xmath25-decay ( @xmath26 ) and the pion @xmath25-decay ( @xmath27 ) .
the current average value is given by @xmath28 .
the magnitude of @xmath29 and @xmath30 is almost equal to the sine of the cabbibo angle , i.e. @xmath31 and @xmath32 , respectively .
@xmath33 is determined by the semileptonic kaon decay @xmath34 ( @xmath35 ) and the hyperon semileptonic decays such as @xmath36 , and @xmath37 , while @xmath38 is determined by the semileptonic @xmath39-meson decays of @xmath40 , @xmath41 as well as the leptonic decays @xmath42 .
the magnitude of @xmath43 is also determined by the @xmath39-meson semileptonic decays @xmath44 .
the analysis of heavy - to - heavy semileptonic @xmath45-decays such as @xmath46 and @xmath47 has been constrained by the heavy quark effective theory ( hqet ) @xcite and the small value of @xmath48 was determined as @xmath49 .
the hqet has also played the role of constraining the model building .
even smaller value of @xmath50 was determined by the heavy - to - light semileptonic @xmath45-decays of @xmath51 as well as the leptonic decays of @xmath52 . in this way
, the top two rows of @xmath2 matrix were determined mostly by the direct measurements of experimental processes
. however , the direct measurements are not feasible for the bottom row elements of @xmath2 matrix involving @xmath53-quark .
since the @xmath53-quark mass is so heavy ( @xmath54 gev ) and the lifetime of @xmath53-quark is much shorter than the strong interaction time scale , the @xmath53-quark does nt have any time to form a bound - state meson but quickly decays into @xmath55-quark and @xmath56 .
nevertheless , the rough magnitudes of @xmath57 are consistent with the semileptonic decays of @xmath53-quark @xmath58 measured at cdf and d@xmath59 in the fermilab where the @xmath53-quark evidence was confirmed . especially , the constraint given by @xmath60 is well satisfied by these magnitudes . while the direct measurements are not feasible for @xmath61 , the virtual transition through penguin " process will give more informations .
the first evidence of penguin " process was seen at cleo ii @xcite , where the branching ratio of radiative @xmath45-decay was determined as br(@xmath62 ) = ( @xmath63 .
the most interesting and promising unitarity triangle to be measured is ( db ) ; @xmath64 where @xmath65 .
dividing the l.h.s .
( [ eq : triangle ] ) , one gets @xmath66 where each term can be written in terms of wolfenstein s parameters @xmath67 and @xmath68 in the complex plane of ( @xmath69 ) .
the first term in eq.([eq : triangle - prime ] ) corresponds to the vector from ( 0,0 ) to ( @xmath70 ) and the last term does the vector from ( @xmath70 ) to ( 0,1 ) .
the second term @xmath71 corresponds to the vector from ( 0,1 ) to ( 0,0 ) .
the main interest in determining the profile of the unitarity triangle is then to find the exact location of the apex ( @xmath70 ) .
the allowed region of the apex is given by the overlapping region in fig .
[ fig:1](a ) .
the length between the two apices ( 0,0 ) and ( @xmath70 ) is mainly determined by the measurement of ratio @xmath72 .
the allowed region for this length is given by the rainbow around ( 0,0 ) in fig .
[ fig:1](a ) . also , the length of the adjacent side connecting ( @xmath70 ) to ( 1,0 ) is essentially determined by the measurement of @xmath73 mixing that can provide the value of @xmath74 .
since the value of @xmath75 is very close to that of @xmath76 , one can further constrain this length between ( @xmath70 ) to ( 1,0 ) by measuring the @xmath77 mixing . the rainbow around ( 1,0 )
shown in fig .
[ fig:1](a ) corresponds to the allowed region for this length .
the overlap of the two rainbows is further constrained by the measurement of cp - violation parameter @xmath78 in @xmath79-@xmath80 mixing .
since @xmath78 is a direct measure of the complex phase in @xmath2 , the constraint from @xmath79-@xmath80 mixing is drawn as a horizontal band in fig .
[ fig:1](a ) depending mostly on the value of @xmath68 . the final overlap yields then the allowed region for the apex ( @xmath70 ) of the triangle as shown in fig .
[ fig:1](a ) .
the three angles of the triangle denoted by @xmath81 are given by @xmath82 it is very interesting to note that these angles can also be determined from the cp - asymmetry measurement @xcite .
the cp - asymmetry is given by @xmath83 where @xmath84 , @xmath85 , @xmath86 , and @xmath87 . from the experimental measurements of @xmath88 ,
it may be possible to determine the values of angles @xmath89 .
for example , the value of @xmath25 may be determined by the measurement of @xmath90 .
more examples of process to determine each angle are shown in fig .
[ fig:1](b ) .
figure [ fig:1](b ) also shows other processes to fix the profile of the unitary triangle .
the theoretical efforts in the @xmath45 physics have also been very extensive in the last twenty years and we can not summarize all the developments in this talk . perhaps , here we are just content with a very brief survey . in early 80 s ,
the value of @xmath75 was investigated significantly with the qcd improved spectator model developed by altarelli et al .
@xcite . in the middle of 80 s , a model wavefunction due to wirbel
, stech and bauer called wsb model @xcite was introduced for the heavy quark system .
then , in the late 80 s , the heavy quark symmetry was extensively studied by many authors @xcite .
perhaps , the most quoted work has been done by isgur and wise @xcite introducing the isgur - wise function now frequently referred in the heavy quark effective theory ( hqet ) .
the basic idea of heavy quark symmetry is to realize the hidden symmetry of qcd that can be revealed only in the limit of infinitely heavy quark masses .
the analogous observation in the opposite extreme of zero quark mass limit is the chiral symmetry .
whether the quark masses are heavy or light may be categorized by the qcd scale @xmath91 .
the current quark masses of @xmath92 are much smaller than @xmath91 , while the masses of @xmath93 quarks are much larger than @xmath91 , _
i.e. _ @xmath94 .
thus , in the limit that @xmath95 and @xmath96 go to zero , the qcd reveals the chiral symmetry @xmath97 and the symmetry is spontaneously broken to @xmath98 due to the non - trivial qcd vacuum .
similarly , in the limit that @xmath99 and @xmath100 go to infinity , the heavy quark symmetry @xmath101(or @xmath102 ) is revealed in qcd . in the early 90 s
, the hqet was applied extensively in the heavy quark systems . at the same time , the constituent quark model(cqm ) was built by isgur , scora , grinstein and wise and named as isgw model @xcite .
this model emphasizes the importance of resonance contribution near the small invariant mass of final states in the inclusive semileptonic decays .
in the case of semileptonic @xmath45 decay such as @xmath103 , the invariant mass square @xmath104 of the final state , continuum @xmath105 and @xmath106 quarks , is bound by @xmath107 , where @xmath108 and @xmath109 are the four - momenta of electron , neutrino and the @xmath45 meson , respectively .
isgw criticized the early qcd improved spectator model by altarelli _ et .
@xcite because the region of @xmath110 is dominated by the resonances rather than the continuum . in the middle of 90 s , isgw model was extended to a relativistic version and called isgw2 model @xcite . throughout 90 s ,
the lattice qcd @xcite was also extensively used to analyze the heavy quark systems . in 90 s , the perturbative qcd(pqcd ) @xcite approach was used mainly to analyze the hadronic decays of heavy mesons .
then , in the later part of 90 s , dispersion relation was used for the analysis of timelike region using the inputs from cqm @xcite , lattice data @xcite , hqet and pqcd @xcite .
the light - cone quark model(lcqm ) was developed around this time .
perhaps , we may categorize the lcqm into four different versions .
first , the isgw2 model was extended to the lc formalism @xcite .
however , the same input parameters as the isgw2 model were used in this lcqm .
second , the lc version of hqet was developed @xcite . in this development ,
the wsb model developed in the middle of 80 s was ruled out because the wsb model does nt satisfy the constraint from the heavy quark symmetry .
third , the lc model wavefunction was also introduced @xcite . however
, modelling the lc wavefunction did nt give any information about the hadron spectra .
finally , we have implemented the variational principle to the qcd motivated effective lc hamiltonian to enable the analysis of meson spectra as well as many wavefunction - related observables such as the form factors , decay constants and electroweak decay rates , etc .
@xcite . in late 90 s ,
also other approaches such as the light - cone qcd sum - rule @xcite , the dyson - schwinger equation @xcite and the bethe - salpeter equation @xcite were used to analyze the heavy quark systems .
more recently , the quark - meson model @xcite utilizing both hqet and chrial perturbation ( @xmath111 ) theory was used to analyze the heavy - to - light meson decays such as @xmath112 . in the twenty years of @xmath45-physics history , we may note that one of the focal point in most analyses has been the accurate prediction of exclusive semileptonic decays such as @xmath113 and @xmath114 . in the next section
, we now present some more details of exclusive semileptonic decays .
the current matrix element of the semileptonic pseudoscalar to pseudoscalar ( ps@xmath115ps ) meson decays involve the two form factors : @xmath116 + f_0(q^2)\frac{m_1 ^ 2-m_2 ^ 2}{q^2}q^\mu\\ \nonumber & = & \sqrt{m_1 m_2}[h_+(q^2)(v_1+v_2)^\mu + h_-(q^2)(v_1-v_2)^\mu],\\ \label{eq : ps to ps}\end{aligned}\ ] ] where @xmath117 in the limit of @xmath118 , the only one form factor remains , _
i.e. _ @xmath119 and @xmath120 , where @xmath121 is known as the isgur - wise function . also , due to the ademello - gatto theorem @xcite , @xmath122 .
note here that @xmath123 corresponds to the charge form factor in the @xmath124 limit .
the similar decomposition of the matrix element of the semileptonic pseudoscalar ( ps ) to vector ( v ) meson decays can be made with the four transition form factors .
again in the heavy quark mass limit , only one form factor @xmath125 is needed .
also , the luke theorem @xcite applies in the zero recoil limit .
however , because of the limited space , we wo nt discuss the details of ps@xmath115v semileptonic decays in this presentation but here focus only on the ps@xmath115ps semleptonic decays .
the analysis of exclusive processes can be made efficiently in the lc formalism with the rational energy - momentum relation . in the lcqm calculations presented in ref .
@xcite , the @xmath126 frame has been used to calculate the weak decays in the timelike region @xmath127 , with @xmath128 and @xmath129 being the initial ( final ) meson mass and the lepton ( @xmath130 ) mass , respectively . however ,
when the @xmath126 frame is used , the inclusion of the nonvalence contributions arising from quark - antiquark pair creation ( z - graph " ) is inevitable and this inclusion may be very important for the heavy - to - light and light - to - light decays .
nevertheless , the previous analyses @xcite in the @xmath126 frame considered only valence contributions neglecting the nonvalence contributions . in this work ,
we treat the nonvalence state using the schwinger - dyson equation to connect the embedded - state shown as the black blob in fig .
[ fig : embed ] to the ordinary lc wave function ( white blob in fig .
[ fig : embed ] ) . to make the program successful , we need some relevant operator connecting one - body to three - body sector shown as the black box in fig . [
fig : embed ] .
the relevant operator is in general dependent on the involved momenta .
our main observation is that we can remove the four - body energy denomenator @xmath131 using the identity @xmath132 of the energy denominators and obtain the identical amplitude in terms of ordinary lc wave functions of photon and hadron ( white blob ) .
for the small momentum transfer , perhaps the relevant operator may not have too much dependence on the involved momenta and one may approximate it as a constant operator . in contact interaction case , we verified that our prescription of a constant operator in fig .
[ fig : embed](d ) is an exact solution of fig .
[ fig : embed](a ) . in our previous analysis
@xcite for the exclusive ps@xmath115ps semileptonic decays , the form factor @xmath133 obtained from @xmath134 in @xmath135 frame is not only immune to the zero - mode contribution but also in good agreement with the experimental data as well as other theoretical results . in order to obtain the form factor @xmath136 in @xmath137 frame , one has to use another component ( i.e. @xmath138 or @xmath139 ) of the current in addition to the @xmath134 .
however , as noted in ref .
@xcite , those @xmath138 and @xmath140 are not immune to the zero - mode contributions , which are not easy to be identified in lcqm .
thus , we use @xmath126 frame to determine the constant operator by equating the slope of @xmath141 at @xmath142 in @xmath126 frame to that in @xmath137 frame .
then , we apply the same operator to the calculation of @xmath136 .
we present here some preliminery results for @xmath143 using the approximation of constant operator to illustrate our method . [ cols="^,^,^,^,^ " , ] in table [ kpi ] , we summarize the experimental observables for the @xmath143 decays , where @xmath144 and @xmath145 .
we use our linear potential parameters given by refs .
@xcite in this analysis . as one can see in table [ kpi ] , our results for the slope @xmath146 of @xmath147 at @xmath142 and @xmath148=@xmath149 are now much improved and comparable with the data .
especially , our result of @xmath146= 0.025 obtained from our effective calculation is in excellent agreement with the data , @[email protected] .
more theoretical details on our effective method as well as heavy - to - heavy and heavy - to - light semileptonic processes will be presented in the future communication .
with the wealth of new or upgraded experimental facilities , precision test of standard model is ever more promising .
we presented a theoretical overview of b - physics and noted that the exclusive semileptonic b - decays and rare b - decays are very important to determine @xmath152 , @xmath153 , @xmath74 , @xmath76 , and @xmath154 stringently .
as we discussed , an effective use of lc degrees of freedom seems crucial to make predictions consistent with many other exclusive processes .
our initial attempt to accomodate the contributions from embedded states seems to give encouraging results .
we would like to thank prof .
chris pauli for organizing a stimulating meeting and his outstanding hospitality at the max plank institute .
we also thank chirag lakhani for drawing a figure .
this work was supported by a grant from the u.s .
doe under contracts de - fg02 - 96er40947 .
99 for the experimental overview of b - physics , see h. schrder in this proceedings .
particle data group , d. e. groom et al .
j. c 15 ( 2000 ) 1
. l. wolfenstein , phys .
51 ( 1983 ) 1945 .
n. isgur and m. b. wise , phys .
b 232 ( 1989 ) 113 ; phys .
b 237 ( 1990 ) 527 .
e. e. eichten and b. hill , phys .
b 234 ( 1990 ) 511 ; m. neubert , phys . lett .
b 338 ( 1994 ) 841 . cleo collab .
r. ammar et al .
71 ( 1993 ) 674 .
see e.g. y. nir and h. r. quinn , in b decays ( revised 2nd edition ) edited by s. stone , world scientific , singapore , 1994 .
g. altarelli , m. cabibbo , g. corbo , and l. maiani , nucl .
b 207 ( 1982 ) 35 ; n. cabibbo , g. corbo , and l. maiani , @xmath155 .
b 155 ( 1979 ) 93 .
m. wirbel , b. stech and m. bauer , z. phys .
c 29 ( 1985 ) 637 . n. isgur , d. scora , b. ginstein , m. b. wise , phys . rev .
d 39 ( 1989 ) 799 .
d. scora and n. isgur , phys .
d 52 ( 1995 ) 2783 . j. m. flynn and c. t. sachrajda , heavy quark physics from lattice qcd , to appear in heavy flavor , 2nd ed . ,
a. j. buras , m. lindner ( eds . ) , world scientific , singapore , hep - lat/9710057 ; c. w. bernard , a. x. el - khadra , and a. soni , phys . rev .
d 43 ( 1991 ) 2140 ; @xmath155 .
d 45 ( 1992 ) 869 ; ukqcd collaboration , k. c. bowler et al . ,
d 51 ( 1995 ) 4905 .
a. szczepaniak , e. m. henley , and s. j. brodsky , phys .
b 243 ( 1990 ) 287 ; c. e. carlson and j. milana , phys . rev .
d 49 ( 1994 ) 5908 .
d. melikhov , phys .
d 53 ( 1996 ) 2460 .
d. becirevic , phys .
d 54 ( 1996 ) 6842 . c. g. boyd , b. grinstein and r. f. lebed , phys .
d 56 ( 1997 ) 6895 .
i. l. grach , i. m. narodetskii , and s. simula , phys .
b 385 ( 1996 ) 317 ; f. cardarelli and s. simula , phys . lett .
b 421 ( 1998 ) 295 .
cheng , c .- y .
cheung , c .- w .
hwang , and w .- m .
zhang , phys .
d 57 ( 1998 ) 5598 . phys .
d 52 ( 1995 ) 2915
. w. jaus , phys .
d 53 ( 1996 ) 1349 . h. -m .
choi and c. -r .
ji , phys . rev .
d 59 ( 1999 ) 074015 .
choi and c. -r .
ji , phys .
d 59 ( 1999 ) 034001 ; phys .
b 460 ( 1999 ) 461 . p. ball and v. m. braun , phys .
d 58 ( 1998 ) 094016 ; t. m. aliev and m. savic , j. phys .
g 24 ( 1998 ) 2223 .
m. a. ivanaov , yu . l. kalinovsky and c. d. roberts , phys .
d 60 ( 1999 ) 034018 .
huang , phys .
d 56 ( 1997 ) 1579 .
a. deandrea , r. gatto , g. nardulli , and a. d. polosa , phys .
d 59 ( 1999 ) 074012 .
m. ademollo and r. gatto , phys .
( 1964 ) 264 .
m. e. luke , phys .
b 252 ( 1990 ) 447 .
w. jaus , phys .
d 60 ( 1999 ) 054026 . | theoretical overviews on the b - physics are presented with an emphasis on the light - cone degrees of freedom .
our new treatment of the embedded states seems to give an encouraging result . |
ten - meter class ground - based telescopes present the opportunity to obtain an unprecedented view of the galactic center in terms of both sensitivity and angular resolution .
initial studies of the galaxy s central cluster at high angular resolution relied upon speckle imaging techniques .
this first round of experiments was able to measure stellar motions on the plane of the sky , yielding estimates of the projected velocities ( eckart & genzel 1996 ; ghez et al .
1998 ) , projected accelerations ( ghez et al . 2000 ; eckart et al .
2002 ) , and three - dimensional orbital motions ( schdel et al . 2002 ; ghez et al .
2003 ) , which each provided a successively stronger case for a supermassive black hole at the center of the milky way and its association with the unusual radio source sgr a * ( lo et al . 1985 ) .
the advent of adaptive optics ( ao ) makes it practical to obtain spectroscopic observations of individual stars in the central @xmath9 of the galaxy .
such measurements in principle provide the critical third dimension of motion , as well as astrophysical information about these high velocity stars , such as spectral type and rotational velocity . while speckle spectroscopy was attempted by genzel et al .
( 1997 ) , it was limited in spectral resolution to r@xmath435 and by source confusion . the first ao - assisted spectroscopy of the galactic center was attempted with significantly higher spectral resolution , r@xmath42,000 ( gezari et al 2002 ) .
nonetheless , neither of these experiments yielded any line detections in the high velocity ( @xmath101000 km / s ) stars , indicating that the spectral lines are weak and would require either even higher spectral resolution or higher snr ( or possibly both ) to be detected . in this paper
we report the first measurement of spectral absorption lines in one of the high velocity stars , s0 - 2 .
section 2 describes the observations , section 3 details the results , and section 4 discusses how these observations improve the orbital solutions and raises questions regarding the origin of this apparently massive , young star .
on the nights of 2002 june 2 - 3 ( ut ) , near - infrared spectra of s0 - 2 were obtained with the w. m. keck 10 m telescope s adaptive optics system ( wizinowich et al .
2000 ) and nirc2 , the facility near - infrared instrument ( matthews et al .
, in prep ) .
ao corrections were made with a r=13.2 mag natural guide star ( usno 0600 - 28579500 ) located 30 " away from s0 - 2 . using nirc2 with its medium resolution grism , 2 pixel slit , and medium resolution pixel scale of @xmath4002/pixel , we achieved a spectral resolution of r@xmath44000 ( @xmath475 km / s ) and a spectral range of 0.3 @xmath11 across the 1024x1024 array . on the first night , the k - bandpass filter was used with the slit positioned such that the spectral coverage was 2.042.30@xmath1 m .
seven 20 min exposures were obtained with s0 - 2 at various locations on the slit .
hd 195500 ( spectral type a1v ) and hd 193193 ( spectral type g2v ) were observed as calibration sources at the same locations on the slit as s0 - 2 . on the second night , the setup was altered somewhat to help identify and minimize systematic effects , such as fringing , which might complicate any line detections .
the filter was changed to the k-bandpass , the slit was moved , giving a spectral coverage of 2.03 - 2.29@xmath1 m , and the calibration sources observed were hd 190285 ( spectral type a0v ) and hd 193193 .
four spectra of s0 - 2 were obtained with this set - up .
because of the high stellar densities at s0 - 2 s location , a dark patch of sky 1@xmath12 east of sgr a * was observed on both nights to measure a purely sky background .
the wavelength solution was obtained by identifying a set of 16 oh emission lines in the spectra of the sky and by fitting a low - order polynomial function to the location of those lines throughout the detector along the spatial direction .
the accuracy of this process is @xmath13km / s , as measured by the dispersion of the residual of the fit .
the sky - subtracted spectrum of s0 - 2 in each dataset was extracted over a 006 spatial interval along the slit and divided by the spectrum of the a - type star , extracted over the same window , to properly correct for the instrumental spectral response . prior to this step ,
the spectrum of the g - type star was used to remove the strong br@xmath0 absorption feature in the spectrum of the a - type star ( cf .
hanson et al . 1996 ) .
the spectrum was then divided by a blackbody of the temperature matching the spectral type of the calibrator to provide a spectrum corrected for all telluric absorption features .
the extracted spectra of s0 - 2 were still partially contaminated by background emission due to the gas present around the galactic center as well as to the presence of the nearby bright source irs 16c , whose wings extend along the spatial direction up to the location of s0 - 2 . to correct for this emission , we estimated a local background by averaging the flux in two 006 spatial intervals along the slit located 075 to the sw and 04 to the ne of s0 - 2 .
this background estimate , which amounts to 4050% of the level of s0 - 2 s continuum , was subtracted from the spectrum of the source to provide the final spectrum , which has snr per pixel of @xmath430 - 80 in the continuum ( see figure 1 ) .
there is significant br@xmath0 emission from the gas surrounding the galactic center and the intensity of this line varies spatially .
close to s0 - 2 , however , the wavelength of this feature does not vary , which helps to distinguish it from any potential photospheric feature . in general ,
as long as the magnitude of s0 - 2 s doppler shift is greater than @xmath4300 km / s , the line emission from the gas is unlikely to cause problems , either non - detections or biases , for r@xmath4 4000 measurements of stellar br @xmath0 absorption .
the zone of gas contamination is estimated based on the structure of the br@xmath0 emission line detected in the gas close to s0 - 2 .
the residual emission feature seen in the final spectrum of s0 - 2 corresponds to this local gas emission , and we believe that it is not physically associated with the star ( see 3 ) .
the spectrum of s0 - 2 , shown in figure 1 , has two identifiable spectral lines .
these are both seen in absorption and are identified as the h i ( 4 - 7 ) or br@xmath0 line at 2.1661 @xmath1 m and the he i triplet at 2.1126 @xmath1 m ( @xmath14 ) , which is a blend of three transitions at 2.11274 , 2.11267 , and 2.11258 @xmath1 m .
table 1 summarizes the properties of these two lines , which are obtained by fitting the background continuum over the whole spectrum with a low - order polynomial and fitting the lines with a gaussian profile .
also given in table 1 are limits on three lines that are not detected .
the reported values and uncertainties are based on the fits to the average of all 11 spectra and the standard deviation of the mean of the fits to 5 independent pairs of spectra .
the locations of the gaussian peaks provide estimates of s0 - 2 s doppler shift , after a correction of 18 km / s is applied to account for the earth s motion around the sun and the sun s motion towards the center of the galaxy ( binney & merrifeild 1998 ) .
the agreement between the doppler shifts inferred from the br@xmath0 and he i absportion line profiles supports the interpretation of the br@xmath0 emission in s0 - 2 s spectrum as residual gas emission as opposed to part of a p cygni line profile .
the weighted average of these two doppler shift measurements yields a radial velocity of -510@xmath240 km / s for s0 - 2 during 2002.4187 . combining this radial velocity with the tangential velocity measured by ghez et al .
( 2003 ) at the same epoch , results in a total space motion of 6660 @xmath2 730 km / s .
the gaussian half width at half maximum ( hwhm ) values provide estimates of the rotational velocities .
the intrinsic rotational profile of a star has a hwhm that corresponds to 90% of @xmath15 ( gray 1976 ) .
since the hwhm of both lines is significantly larger than the instrumental hwhm ( @xmath438 km / s ) , we simply scale the gaussian hwhm values by a factor of 1.1 to derive an estimated value of @xmath15 .
the values from the two lines are comparable and averaged together suggest a @xmath15 of 220@xmath240 km / s .
equivalent widths ( ew ) for the lines are also derived from the gaussian fits .
the ew(br@xmath0 ) is 2.8 @xmath2 0.3 @xmath3and the ew(hei[2.1126 @xmath1 m ] ) is 1.7 @xmath2 0.4 .
these values are checked against those obtained from integrating over 40 - windows centered on the lines , which provide similar results to within 37% . using the same window and the doppler shift measured for two detected lines ,
we derive the limits for he i ( 2.0581 @xmath1 m ) , n iii ( 2.1155 @xmath1 m ) , and he ii ( 2.1885 @xmath1 m ) , which are listed in table 1 . between our two consecutive nights of observations
there is an expected change of 60 km / s in the radial velocity ( 4.1 ) .
we therefore have also analyzed the spectra averaged over each night separately .
the results for the br@xmath0 line are reported in table 1 . while the change in radial velocities between the two nights is not significant ( @xmath16(2002.4205 )
- @xmath16(2002.4177 ) = -37 @xmath2 47 km / s ) , it is consistent with the predicted change . furthermore , the ew and @xmath15 measurements from the spectrum averaged over the two nights appears to be unaffected by this shift .
the individual nights @xmath17 are used in the dynamical analysis in 4.1 and the averaged spectrum values of ew and @xmath15 are applied in the stellar astrophysics discussion in 4.2 .
the measurements of a radial velocity for s0 - 2 provide a new and powerful constraint on its orbit .
s0 - 2 s motion on the plane of the sky provided the first estimates of the orbital parameters ( schdel et al .
2002 ; ghez et al .
2003 ) . by combining the radial velocity with the proper motions reported by ghez et al . and assuming a distance of 8.0 kpc ( reid 1993 ) , as was done in both proper motion analyses , we obtain the first orbital solution for s0 - 2 based on measured three dimensional motion ( table 2 ) .
we assume here that sgr a * has no significant velocity with respect to the galaxy along the line of sight , as is observed for its proper motion ( reid et al .
our fit produces a total mass estimate of @xmath18 , consistent at the @xmath42@xmath19 level with earlier estimates based on velocity dispersion measurements .
our results on s0 - 2 s orbit do not assume the location of the dynamical center , as opposed to the analysis of schdel et al .. despite these two additional free parameters , the uncertainties on the orbital parameters are reduced by a factor of 2 - 3 .
this first orbital estimate of the galaxy s dynamical center is not only consistent with the nominal infrared position of sgr a * to within the uncertainties on the latter ( reid et al .
2003 ) , but also a factor of 5 more precise ( @xmath2 2 milli - arcsec ) .
the addition of radial velocity measurements also breaks the ambiguity in the inclination angle , _
i_. with the proper motion data alone , only the absolute value of the inclination angle can be determined , leaving the questions of the direction of revolution and where along the line of sight the star is located behind the black hole unresolved .
our radial velocity measurements indicate a negative inclination angle and consequently that s0 - 2 is both counter - revolving against the galaxy and behind the black hole at the time of periapse . the improved location of the center of attraction from the orbital analysis results in a minimum offset of s0 - 2 from the blackhole in the plane of the sky of 11 @xmath2 2 milli - arsec , which is significantly larger than the expected einstein radius ( @xmath20 = 0.42 milli - arcsec for the s0 - 2 distance behind the black hole of @xmath4 100 au ) and therefore makes gravitational lensing an unlikely event ( wardle & yusef - zadeh 1992 ; alexander & loeb 2001 ) . in principle , the addition of radial velocities to the study of s0 - 2 s dynamics allows the distance to the galactic center , @xmath21 , to be a free parameter in the orbital fits ( salim & gould 1999 ) .
the measurements , however , were obtained just 30 days after the star s closest approach to the black hole when the radial velocity was changing very rapidly ( see figure 2 ) .
while the current radial velocity and proper motion data set constrains @xmath22 very effectively ( @xmath415% uncertainty ) , it does not yet produce a meaningful measurement of @xmath21 .
nonetheless , as figure 2 shows , the radial velocities from the currently allowed orbits quickly diverge , producing a spread of a few hundred km / s in one year . within the next few years
, the orbital fits based on both proper motions and additional radial velocity measurements should provide the most direct and precise estimate of the distance to the galactic center , making it a fundamental rung in the cosmic distance ladder .
the detection of absorption lines allows us to sort out the spectral classification ambiguities present when only photometric information is available and to determine if this star s photosphere has been altered as a result of its close proximity to the central black hole .
the average brightness at 2.2 @xmath1 m for s0 - 2 is k @xmath4 13.9 mag and there is no evidence of brightening after periapse passage ( ghez et al .
2003 ) . with a distance of 8.0 kpc and k - band extinction of 3.3 mag ( rieke , rieke , & paul 1989 ) , the 2.2 @xmath1 m brightness of s0 - 2 implies that , if it is an ordinary star unaltered by its environment , it could either be an o9 main - sequence star or a k5 giant star ; all supergiants are ruled out as they are too bright by at least 2 magnitudes in the k bandpass . kleinmann and hall ( 1986 ) provide a 2.0 - 2.5 @xmath1 m spectral atlas of late - type stars that demonstrates that if s0 - 2 is a k5 giant star , then it should have deep co absorption lines , which definitively were not detected in either this experiment or our earlier experiment reported by gezari et al .
in contrast , the spectral atlas of 180 o and b stars constructed by hanson , conti and rieke ( 1996 ) shows that an o9 main sequence star both lacks the co absorption and has br @xmath0 and he i ( 2.1126 @xmath1 m ) consistent with the observed values .
futhermore , stars in this comparison sample earlier than o8 show niii ( 2.115 @xmath1 m ) in emission and he ii ( 2.1885 @xmath1 m ) in absorption above our 3 @xmath19 thresholds , which are listed in table 1 ; the lack of photospheric he i ( 2.058 @xmath1 m ) absorption does not provide any additional constraints .
similarly , a dwarf b - type stars later than b0 have absorption equivalent widths that are too large .
together , the photometry and absorption line - equivalent widths permit dwarf spectral types ranging from o8 to b0 .
likewise , the rotational velocity of 224 km / s is reasonable for this range ( gatheier , lamers , & snow 1981 ) .
s0 - 2 , therefore , appears to have a spectral type , and hence temperature ( @xmath430,000 k ) , as well as luminosity ( @xmath410@xmath23 @xmath24 ) that are consistent with a main sequence star having a mass of @xmath415 @xmath5 and an age @xmath610 myr .
it is challenging to explain the presence of such a young star in close proximity to a supermassive black hole . assuming that the black hole has not significantly affected s0 - 2 s appearance or evolution , s0 - 2
must be younger than 10 myr and thus formed relatively recently . if it has not experienced significant orbital evolution , its apoapse distance of 1900 au implies that star formation is possible in spite of the tremendous tidal forces presented by the black hole , which is highly unlikely .
if the star formed at larger distances from the black hole and migrated inward , then the migration would have to be through a very efficient process .
current understanding of the distribution of stars , however , does not permit such efficient migration .
this problem is similar to that raised by the he i emission - line stars ( e.g. , sanders 1992,1998 ; morris 1993 , morris et al .
1999 ; gerhard 2001 ; kim & morris 2003 ) , which are also counter - revolving against the galaxy ( genzel et al .
1997 ) , but amplifies it with a distance from the black hole that is an order of magnitude smaller .
an alternative explanation for s0 - 2 s hot photosphere is that it may be significantly altered by its environment .
while its periapse passage is too large for it to be tidally heated by the black hole as explored by alexander & morris ( 2003 ) , it may be affected by the high stellar densities found in this region . on the one hand
, the high stellar densities might allow s0 - 2 to be an older giant star that has had its outer atmosphere stripped through collisions ; however , to generate the necessary luminosity , significant external heating is required ( alexander 1999 ) . on the other hand , high stellar densities might lead a cascade of merger events ( lee 1996 ) , which would allow s0 - 2 s formation process to have begun more than 10 myr ago .
however a large number of collisions would have had to occur to provide the necessary lifetime to bring it in from sufficiently large radii .
more exotically , it could be a `` reborn '' star , which occurs as the product of a merger of a stellar remnant with a normal star .
none of these possibilities is altogether satisfactory , leaving s0 - 2 as a paradox of apparent youth in the vicinity of a supermassive black hole .
we thank the staff of the keck obsevatory , especially randy campbell , grant hill , chuck sorensen , david lemignant , and director fred chaffee .
this work has been supported by the national science foundation through the individual grant ast99 - 88397 and the science and technology center for adaptive optics , managed by the university of california at santa cruz under cooperative agreement no .
ast - 9876783 .
the research of e.e.b .
, b.t.s . and
d.j.t . is supported by nasa .
keck observatory is operated as a scientific partnership among the california institute of technology , the university of california and the national aeronautics and space administration .
the observatory was made possible by the generous financial support of the w.m .
keck foundation .
ghez , a.m. , becklin , e.e .
, duchene , g. , hornstein , s. , m. morris , s. salim , a. tanner ( 2003 ) astron .
s1 , special supplement `` the central 300 parsecs of the milky way '' , eds .
a. cotera , h. falcke , t. r. geballe , s. markoff llll br@xmath0 ( 2.1661 @xmath1 m ) & & & @xmath25 avg .
spectrum & [email protected] & -510@xmath239 & 238@xmath262 @xmath25 june 2 spectrum & & -495@xmath236 & @xmath25 june 3 spectrum & & -532@xmath244 & he i ( 2.1125 @xmath1 m ) & [email protected] & -532@xmath296 & 216@xmath253 he ii ( 2.1891 @xmath1 m ) & @xmath60.5 & & n iii ( 2.1155 @xmath1 m ) & @xmath61.0 & & he i ( 2.0587 @xmath1 m ) & @xmath61.3 & & average & & -513@xmath236 & 224@xmath240 lll @xmath26 ( milli - arcsec ) & -2.7 @xmath2 1.9 & @xmath27 ( milli - arcsec ) & -5.4 @xmath2 1.4 & m ( @xmath28 ) & 4.07 @xmath2 0.62 & 3.7 @xmath2 1.5 a ( milli - arcsec ) & 125.6 @xmath2 5.5 & 119 @xmath2 15 p ( yrs ) & 15.78 @xmath2 0.82 & 15 @xmath2 1 e & 0.8736 @xmath2 0.0083 & 0.87 @xmath2 0.03t@xmath29 ( yrs ) & 2002.334 @xmath2 0.017 & 2002.30 @xmath2 0.05i ( @xmath30 ) & -47.3 @xmath2 2.5 & @xmath246 @xmath2 4 @xmath31 ( @xmath30 ) & 248.5 @xmath2 1.8 & 250 @xmath2 5 @xmath32 ( @xmath30 ) & 49.9 @xmath2 3.0 & 36 @xmath2 9 | we have obtained the first detection of spectral absorption lines in one of the high - velocity stars in the vicinity of the galaxy s central supermassive black hole . both br @xmath0 ( 2.1661 @xmath1 m ) and he i ( 2.1126 @xmath1 m )
are seen in absorption in s0 - 2 with equivalent widths ( 2.8 @xmath2 0.3 @xmath3 & 1.7 @xmath2 0.4 ) and an inferred stellar rotational velocity ( 220@xmath240 km / s ) that are consistent with that of an o8-b0 dwarf , which suggests that it is a massive ( @xmath415 @xmath5 ) , young ( @xmath610 myr ) main sequence star .
this presents a major challenge to star formation theories , given the strong tidal forces that prevail over all distances reached by s0 - 2 in its current orbit ( 130 - 1900 au ) and the difficulty in migrating this star inward during its lifetime from further out where tidal forces should no longer preclude star formation .
the radial velocity measurements ( @xmath7 = -510@xmath240 km / s ) and our reported proper motions for s0 - 2 strongly constrain its orbit , providing a direct measure of the black hole mass of @xmath8 .
the keplerian orbit parameters have uncertainities that are reduced by a factor of 2 - 3 compared to previously reported values and include , for the first time , an independent solution for the dynamical center ; this location , while consistent with the nominal infrared position of sgr a * , is localized to a factor of 5 more precisely ( @xmath22 milli - arcsec ) .
furthermore , the ambiguity in the inclination of the orbit is resolved with the addition of the radial velocity measurement , indicating that the star is behind the black hole at the time of closest approach and counter - revolving against the galaxy . with further radial velocity measurements in the next few years
, the orbit of s0 - 2 will provide the most robust estimate of the distance to the galactic center . |
elucidating the nature of chiral transition separating the high - temperature quark - gluon plasma phase from the low - temperature hadron phase has been one of the focal points of effort in recent numerical simulations of full lattice qcd including light dynamical quarks .
much of work in this direction has employed the kogut - susskind quark action since it retains a @xmath1 subgroup of chiral symmetry . on the other hand , studies with the wilson quark action
are less well developed in spite of the effort over the years@xcite .
a main problem for studies with wilson fermions originates from the presence of explicit chiral symmetry breaking in the action at non - zero lattice spacing , which is introduced to avoid the well - known species doubling problem .
clearly , before serious studies on the phase structure at finite temperature are initiated , more understanding on the effect of the explicit chiral symmetry breaking at zero temperature is necessary .
let me consider a phase space of @xmath2 and @xmath3 at zero temperature , where @xmath4 is the bare gauge coupling , @xmath5 the bare quark mass , and @xmath6 the lattice spacing .
there is much evidence from analytical considerations and numerical simulations of hadron masses that the pion mass vanishes along a line @xmath7 , the critical line , which runs from @xmath8 in the strong - coupling limit to @xmath9 in the weak - coupling limit .
however the massless pion appeared on @xmath10 can not be regarded as the nambu - goldstone boson of the spontaneous breaking of chiral symmetry since it is explicitly broken in the wilson fermion formulation at non - zero lattice spacing .
why does the massless pion appear then ? it has conventionally been explained as follows : the appearance of the massless pion on @xmath0 may be an evidence that the tuning @xmath11 effectively recovers the chiral symmetry , explicitly broken by the wilson fermion formulation .
off course , this in not an explanation , since it does not tell why the massless pion appear on @xmath0 .
it also can not tell what happens beyond @xmath0 . instead of the conventional explanation ,
a possible explanation was proposed some years ago , to explain the appearance of the massless pion@xcite , on the basis of new phase structure of lattice qcd with wilson fermions .
later some analytic@xcite and numerical@xcite evidences were given for the existence of the phase structure , but it is far from being completely established . in this report , after reviewing the previous accomplishment , more evidences are supplied to established the proposed phase structure .
in this section we review the possible explanation for the massless pion appeared on @xmath0 in the wilson fermion formulation@xcite .
let me consider lattice qcd with a single flavor first . in the theory there exists one `` @xmath12 meson '' , corresponding to a pseudo - scalar operator @xmath13 . at @xmath14 the correlation function of the operator for large
@xmath15 behaves such that @xmath16 where the correlation length @xmath17 is related to the mass of the @xmath12 meson as @xmath18 . since the correlation length diverges on @xmath0 the correlation function for large @xmath15 behaves at @xmath19 such that @xmath20 with some exponent @xmath21 .
if we go beyond @xmath0 ( @xmath22 ) , it is expected for large @xmath15 that @xmath23 with @xmath24 where @xmath25 is a space volume .
the above situation is summarized as follows . since the massless @xmath12 meson implies the divergent correlation length on @xmath0 , the operator corresponding to the @xmath12 meson has the long - range order at @xmath22 in such a way that @xmath24 .
this condensation spontaneously breaks the discrete parity at @xmath26 and the second order phase transition from the parity symmetric phase to the parity broken one occurs at @xmath27 .
therefore the massless @xmath12 meson can be interpretated as the divergent correlation length associated with this second order phase transition , not a nambu - goldstone boson . according to the above interpretation we expect near @xmath0 , @xmath28 with another critical exponent @xmath29 .
if we regards the effective theory for the @xmath12 meson as a four - dimensional scalar field theory , we can expect the phase transition to be mean - field like up to the logarithmic corrections and therefore @xmath30 , which reproduces well - known pcac relation that @xmath31 where a quark mass @xmath32 .
now we consider qcd with 2 flavors , where members of the pseudo - scalar mesons are @xmath33 ( non - singlet ) and @xmath34 ( singlet ) .
since the observed mass spectra of these mesons satisfy @xmath35 , it is expected at @xmath27 that @xmath36 . to satisfy this
, there exists a phase transition at @xmath19 , which implies the condensation such that @xmath37 at @xmath22 .
the above condensation spontaneously breaks both parity and flavor symmetry . in this case
the neutral @xmath38 meson corresponding to the @xmath39 operator is identified with the massless mode of this phase transition and becomes massless only at the transition point @xmath19 .
the charged @xmath40 mesons corresponding to the linear combination of @xmath41 and @xmath42 operators are nambu - goldstone bosons associated with the flavor symmetry breaking : @xmath43 in the phase with @xmath44 and @xmath45 in the phase with @xmath46 . on the other hand
the @xmath47 meson stays massive for all @xmath11 since @xmath48 is always zero .
this phase transition can naturally solve not only the problem of the massless pion but also the lattice u(1 ) problem@xcite .
one can rigorously prove that @xmath49 using _ a la _ vafa - witten theorem@xcite , while no such restriction can be applied to the operator @xmath39 , since the wilson fermion determinant is shown to be real - positive for @xmath50 even in the presence of the source term @xmath51 .
in the strong coupling ( @xmath52 ) and large @xmath53 limit we can analytically calculated the behavior of order parameter and meson masses@xcite : @xmath54 where the critical point is given by @xmath55(@xmath56 ) with @xmath57 . in this limit
there is no essential difference between @xmath58 and @xmath50 .
it was also shown that the @xmath12 meson becomes massless on @xmath59@xcite . in the strong coupling and 1/@xmath53 expansions
the behavior of the order parameter is essentially the same for @xmath58 case : @xmath60 at @xmath61 , and the pion becomes massless on @xmath62 , which was calculated in the expansion of @xmath63@xcite .
for @xmath64 we obtain @xmath65 at @xmath61 while @xmath66 for all @xmath67 , as expected@xcite .
meson spectrum also agrees with our expectation in the previous section . to calculate quark propagators on a given configuration iterative methods such as conjugate gradiant ( cg ) method are usually used to invert the lattice fermion matrix . for the case of the wilson fermion ,
if @xmath11 approaches @xmath0 , the inversion becomes more and more difficult : a number of cg iteration rapidly increases . to overcome this difficulty and to calculate non - zero order parameter
, we add an external field coupled to the order parameter and take @xmath68 limit after the average over configurations is taken .
let me consider the quenched qcd first .
if we add the external field @xmath69 to the wilson fermion action , it is easy to show that @xmath70 where @xmath71 means a value of an operator @xmath72 on a given gauge configuration @xmath73 and the external source @xmath74 .
if @xmath68 limit was non - singular , this would lead to @xmath75 before the average over gauge configurations .
therefore @xmath76 in the numerical simulation on a finite lattice , where the limit is always non - singular .
however this does not means @xmath77 .
since , in the parity broken phase , @xmath68 limit is singular as the volume @xmath25 goes to infinity one should take limits in the following order : @xmath78 .\ ] ] since actual numerical simulations are always performed on finite volumes , one should extrapolate to @xmath79 from non - zero @xmath74 , not put @xmath80 in the simulation , in order to obtain the correct value of @xmath81 at @xmath80 in the infinite volume limit .
this external source method has been applied to show that the parity breaking phase transition exists at @xmath82 , where @xmath83 is temperature , using the quenched qcd simulation at @xmath84 on a @xmath85 lattice@xcite .
such a phase transition disappear at @xmath86 , where @xmath87 is the critical temperature of the pure gauge theory , using the quenched qcd simulation at @xmath88 , 5.8 on a @xmath89 lattice@xcite . for the case of full qcd with 2 flavors , a term @xmath90 should be added . even in the presence of the term
the wilson fermion determinant is still real - positive .
therefore the hybrid monte carlo method is applicable to this case .
again we have to keep @xmath74 non - zero for simulations on a finite volume , in order to find the spontaneous breakdown of parity and flavor symmetry .
the hybrid monte carlo simulation has been performed for the full qcd with @xmath50 at @xmath84 on @xmath85 , @xmath91 , and @xmath92 lattices@xcite .
although some evidence for the parity and flavor symmetry breaking phase transition was observed , @xmath68 extrapolation was too difficult to prove the existence of the phase transition without doubt .
lattice sizes used may be too small for this @xmath63 .
since the numerical evidence of the phase transition is established for the quenched case but is ambiguous for the case of the full qcd , we carry out new full qcd simulation with @xmath64 at @xmath93 , where the analytic expressions for the order parameters are available for non - zero external field @xmath74 . in fig .
[ ordb0 ] , @xmath94 is plotted as a function of @xmath67 at @xmath52 on a @xmath85 lattice with @xmath95 , together with the analytic expressions with the corresponding @xmath74 in the strong coupling and large @xmath53 limits , since the agreement between the numerical data and the analytic predictions are very good , it can be established at least in the strong coupling limit that the parity and flavor symmetry breaking phase transition really occurs at @xmath96 ( @xmath97 ) and @xmath94 becomes non - zero in the @xmath68 extrapolation at @xmath98 ( @xmath26 ) .
we have checked that the finite size effect to the order parameter is very small by comparing the data on a @xmath85 lattice with the data on a @xmath92 lattice or with the analytic expressions corresponding to the values in the infinite volume limit . as a further check
we have calculated susceptibilities @xmath99 , @xmath100 , and @xmath101 , which are defined by @xmath102 ^ 2 \rangle \\
\chi_{00 } & = & \frac{1}{v}\langle [ \sum_x \bar\psi \psi ( x)]^2 \rangle \\ \chi_{50 } & = & \frac{1}{v}\langle [ \sum_x \bar\psi i\gamma_5\tau^3 \psi ( x ) ] \times [ \sum_y \bar\psi \psi ( y ) ] \rangle , \end{aligned}\ ] ] both analytically and numerically .
for example , @xmath99 is plotted in fig .
the agreement between them are reasonably good , except some deviations are observed deep in the broken phase
. therefore , the result for the susceptibilities also support our conclusion about the existence of the phase transition .
if we consider the above result from the opposite side it can be said that the success of our analytic calculations , except the deviation in the susceptibilities deep in the broken phase , establishes the validity of the strong coupling and large @xmath53 expansions near the limits .
furthermore , since the expansions predict the existence of the phase transition , we can conclude that the parity and flavor symmetry breaking is shown to exist in the strong coupling region .
finally let me comment on the deviations in the susceptibilities observed deep in the broken phase .
we think that the deviations are caused by the massless nambu - goldstone charged pions appeared in the broken phase , whose effect to the susceptibilities can not be included in the analytic calculation of the large @xmath53 limit .
we expect the agreement will become better if we include the effect in our analytic expressions .
in the previous section we have established the phase structure in the strong coupling region .
there are two critical lines @xmath62 and @xmath103 in @xmath63 - @xmath67 plane , which separate the broken phase from the symmetric phases : @xmath104 at @xmath105 and @xmath106 at @xmath107 or @xmath108 . in this section
we investigate how this phase structure will be modified in the weak coupling region . in fig .
[ phase ] we propose the expected phase structure of lattice qcd with 2 flavor wilson fermions in the @xmath63 - @xmath67 plane . at @xmath109
there are only two critical lines , while there are ten critical lines at @xmath110 and 5 points where two lines meat at @xmath111 .
each point corresponds to one continuum limit , whose low energy spectra are composed of a part of sixteen zero modes of the lattice fermion : one mode near @xmath112 appears at @xmath113 , four modes near @xmath114 and its permutation at @xmath115 , six modes near @xmath116 and its permutation at @xmath117 , four modes near @xmath118 its permutation at @xmath119 , and one mode near @xmath120 at @xmath121 .
note that the phase diagram is shown to be symmetric under @xmath122 transformation .
it is also noted that the width of the broken phase in the weak coupling should behave like @xmath123 $ ] where @xmath124 is some constant and @xmath125 is some regular function of @xmath126 , since the appearance of the width has not been detected by the usual perturbative expansion , and therefore should be non - perturbative .
there are several evidences for the existence of the multiple critical lines in the weak coupling region .
first of all , the phase structure similar to fig . [ phase ] has been found for the gross - nevue model in 2 dimensions in the large @xmath127 limit , regulated on the lattice using wilson fermions@xcite .
the existence of the multiple structure of the critical line has been also suggested from studies of eigen - value spectra of the wilson fermion matrix on quenched configurations@xcite and full qcd configurations@xcite . to establish the multiple structure of the critical lines we have performed new simulation using quenched configurations _ without _ adding the external field @xmath128 .
we first monitored the quantity so - called `` pion norm '' : @xmath129 ^ 2 \rangle $ ] .
this quantity becomes very large in the broken phase , due to the contribution from zero modes of the wilson fermion matrix .
although some signals for the multiple structure of the broken phases are observed in this quantity , more clear indicator for the multiple structure of the broken phases is found to be the number of the cg iteration necessary to invert the wilson fermion matrix .
this quantity becomes very large and is proportional to the volume in the presence of zero modes of the wilson fermion matrix , which are expected to exist in the broken phase .
we have plotted the number of cg iterations as a function of @xmath67 at @xmath84 , 5.5 , 5.6 , 5.7 on @xmath92 lattices in fig.[cg1 ] , and at @xmath130 5.8 , 5.9 , 6.0 on @xmath131 lattices in fig.[cg2 ] .
the multiple structure of the broken phase is clearly observed at @xmath132 .
finally we calculate again the order parameter @xmath133 with small non - zero external field , @xmath134 , in the weak coupling region ( @xmath135 ) . as seen from fig .
[ ordw ] , the multiple structure of the broken phases is again observed
the phase transition which causes or is caused by the massless pion appeared in lattice qcd with the wilson fermion formulation seems to exist at @xmath27 . in the strong coupling region at @xmath136 it is established that @xmath137 for the quenched qcd and @xmath138 for the 2 flavor full qcd . the evidence of the multiple structure of the critical lines are also given in the weak coupling region in the case of the quenched qcd .
as mentioned in the introduction , it is very important to study the phase structure of lattice qcd at finite temperature . in the case of the wilson fermion formulation one
should understand , based on the finding of this report , the relation between the critical line discussed in this report and the finite temperature phase transition line .
recently such an investigation has been carried out@xcite for the 2 flavor qcd .
the study concludes that the critical line of the massless pion at finite temperature forms a cusp at finite gauge coupling , and that the line of thermal transition runs past the tip of the cusp without touching the critical line .
the detail can be found in ref .
the part of this work has been completed in collaboration with a. ukawa and t. umemura .
numerical calculations for the present work have been carried out at center for computational physics , at university of tsukuba .
this work is supported in part by the grants - in - aid of the ministry of education(nos .
04np0701 , 06740199 ) .
m. fukugita , s. ohta and a. ukawa , phys .
* 57 * ( 1986 ) 1974 ; a. ukawa , nucl . phys . * b*(proc . suppl.)*17 * ( 1990 ) 463 .
see also , r. gupta _ et al .
rev . d*40 * ( 1989 ) 2072 ; k. m. bitar _
et al . , _ phys .
rev . d*43 * ( 1991 ) 2396 .
( 170,150 ) ( 10,10)(1,0)150 ( 85,10)(0,1)130 ( 10,140)(1,0)150 ( 87,10)(5,5)@xmath139 ( 165,5)(5,5)@xmath67 ( 15,10)(50,140)(50,140 ) ( 15,10)(35,70)(50,10 ) ( 50,10)(65,70)(85,10 ) ( 155,10)(120,140)(120,140 ) ( 155,10)(135,70)(120,10 ) ( 120,10)(100,70)(85,10 ) ( 12,1)(5,5)-4 ( 47,1)(5,5)-2 ( 83,1)(5,5)0 ( 118,1)(5,5)2 ( 153,1)(5,5)4 ( 83,140)(5,5)@xmath63 ( 85,135)(5,5)0 ( 83,40)(1,0)4 ( 77,38)(5,5)@xmath140 | the phase structure of lattice qcd with wilson fermions is discussed .
analytic and numerical evidences are given for the spontaneous breaking of parity and flavor symmetry , which naturally explains the existence of the massless pion at the critical hopping parameter @xmath0 without recourse to the chiral symmetry absent in the wilson fermion formulation .
new numerical evidences are also presented for the multiple structure of the critical lines in the weak coupling region .
a connection between the phase structure and the finite temperature phase transition is briefly mentioned .
= 23 cm = 15.1 cm = -0.9 cm = -1.5 cm |
the unusual transient sn2008s was discovered on 1.8 february 2008 ( arbour 2008 ) in the spiral galaxy ngc 6496 , the ninth supernova to be discovered in that galaxy in the last hundred years .
subsequently it was found by botticella et al .
( 2009 ) to be present on an image taken on 24 january 2008 , but not on an image taken on 16 january 2008 , allowing them to constrain the epoch of the explosion to ( 20@xmath74 ) january 2008 .
prieto et al . (
2008 ) reported that serendipitous pre - outburst deep optical observations of the supernova field did not reveal any source at its position down to 26th magnitude ( welch , clayton & sugerman 2008 provided further prediscovery deep optical photometric upper limits ) and that archival _
spitzer space telescope _
irac images revealed a bright mid - infrared source whose brightness had not varied significantly over the four years preceding the outburst .
the progenitor of sn2008s thus appeared to be enshrouded in a cloud of its own dust . from a 440 k blackbody fit to the ir fluxes , they estimated a pre - explosion luminosity of 3.5@xmath8 l@xmath0 and a modest mass of @xmath910 m@xmath0 for the progenitor star , close to the poorly defined boundary between agb and supergiant stars .
they suggested that sn 2008s was an electron capture sn event by such a star .
wesson et al . ( 2008 ) reported irac and mips _ spitzer _ observations of sn 2008s that were serendipitously obtained only 5 days after its discovery , fitting the photometry with a 500k modified blackbody with a luminosity of 2.1@xmath10 l@xmath0 .
although faint for a supernova , the amplitude and luminosity of the outburst of sn 2008s did not obviously correspond to any other class of object .
smith et al .
( 2009 ) presented optical photometry and spectroscopy of sn 2008s up to 280 days after outburst and argued that it was a ` supernova imposter ' originating from a super - eddington outburst of a @xmath1115 m@xmath4 star .
botticella et al .
( 2009 ) have presented a very large optical and infrared set of observations of sn 2008s up to day 304 , and a model of the post - outburst light echo .
they argued that the late - time decline rate observed for sn 2008s was consistent with the decay of a relatively small quantity of @xmath12co produced by an electron capture supernova event in a super - agb star of initial mass 6 - 8 m@xmath0 .
as noted by thompson et al .
( 2008 ) , smith et al .
( 2009 ) and botticella et al .
( 2009 ) , if a supernova event has not occurred , a relatively luminous star should be detectable once the 2008 outburst has faded sufficiently .
the discovery that sn2008s had a dust - enshrouded progenitor led thompson et al .
( 2008 ) to propose that it and two other recently discovered extragalactic transient sources were members of a new class of transient , formed when stars located near the upper mass limit for agb evolution go through a self - obscured phase shortly before exploding as electron capture sne .
depending on the amount of dust produced in this phase , and the proportion of stars for which the phase occurs , objects such as sn 2008s could potentially contribute to the substantial amounts of dust observed in young high - redshift galaxies ( e.g. bertoldi et al .
2003 ) , given that dust production by relatively massive stars ( @xmath135 m@xmath4 ) appears to be required in order to match the short timescales over which dust is inferred to have formed in the early universe ( morgan & edmunds 2003 ; dwek , galliano & jones 2009 ) the utility of light echoes from circumstellar dust as a probe into the prior evolution of supernova progenitors has been discussed and exploited by e.g. bode & evans ( 1980 ) , graham et al .
( 1983 ) , dwek ( 1983 ) , dwek ( 1985 ) and sugerman ( 2003 ) .
the post - outburst evolution of the sed of sn 2008s thus provides important information on the properties of its progenitor .
we present the mid - infrared observations of sn2008s that we obtained five days after its discovery , along with the further optical and infrared observations that we obtained in july 2008 .
we use these data to develop a self - consistent model for the evolution of the dust shell around sn2008s over three epochs , one pre - outburst and two post - outburst .
ngc 6946 was observed by _ spitzer _ for the sings legacy survey , and also by programs monitoring the recent type - ii supernovae sn2002hh and sn2004et .
the field containing the progenitor of sn2008s was observed eight times in the four years preceding its eruption .
prieto et al . ( 2008 ) reported the detection of the progenitor at 4.5 , 5.8 and 8.0 @xmath14 m . while monitoring sn2002hh , we serendipitously obtained observations of sn2008s with irac and mips on 6.80 and 7.53 february 2008 respectively ( wesson et al . 2008 ) .
we subsequently reobserved the field with _
spitzer _ in july 2008 .
the object was observed in all four irac bands on 18 july , and with mips at 24 @xmath14 m on 29 july .
we measured fluxes from these images by fitting a synthetic psf to the observed profile , using the _ daophot _ package in iraf .
for the mips 24-@xmath14 m data , we carried out psf fitting on difference images that were produced by subtracting pre - outburst images from each of the two post - outburst images .
we obtained an irs spectrum of sn 2008s on 8 july 2008 .
it shows a featureless spectrum , lacking obvious spectral lines or dust features . while of lower signal to noise , its continuum level is consistent with the mid - ir photometry obtained during the same month .
we also observed sn2008s with the 8.1 m gemini north telescope in hawaii during july 2008 .
we obtained 11.2-@xmath14 m photometry with michelle on 7 july and optical photometry and spectra with gmos on 14 july .
for the michelle observations , the total on - source exposure time was 2400s , and the data were flux - calibrated using observations of the standard stars hd192781 and hd198149 .
gmos images of sn 2008s were taken in the g@xmath15 , r@xmath15 and i@xmath15 filters , with an exposure time in each filter of 93s .
johnson magnitudes of v@xmath16 , r@xmath17 and i@xmath18 were derived from the gmos filter observations using the procedure described by welch et al .
( 2007 ) .
gmos spectra were taken using the b600 grating , giving a spectral resolution of 1.8 over a wavelength range of 4490 - 7356 .
we obtained three spectra , dithering the central wavelength by 5 nm between exposures to avoid gaps in the spectral coverage caused by gaps in the gmos detectors .
the total exposure time was 3600s , and the spectrum was flux - calibrated using observations of the standard star bd+28 4211 ( massey et al .
the shape of the spectrum was in good agreement with the optical photometry , but since all spectroscopic observations were taken with a narrow slit , only relative flux calibration was possible , so we scaled our spectrum up to match the optical photometry .
galactic foreground reddening in the direction of ngc 6946 is estimated to be e(b - v)=0.34 ( schlegel et al .
1998 ) , and so we dereddened our optical photometry and spectrophotometry using this value .
our july 2008 observations took place over a relatively narrow time interval , 169 days ( michelle ) , 176 days ( gmos ) , 180 days ( irac ) and 191 days ( mips ) after the outburst of sn2008s , taken to be on 20 january ( botticella et al .
we combined our optical and mid - ir observations from february and july 2008 with the day 16/17 bvr photometry of smith et al .
( 2009 ) and the day 174 jhk photometry of botticella et al .
( 2009 ) .
we present our measured fluxes from february 2008 and july 2008 in table [ day17fluxes ] .
figure [ spitzer_montage ] shows _ spitzer _ pre - outburst and post - outburst images of sn2008s at five wavelengths from 3.6 to 24 @xmath14 m . [ cols="^,^,^,^,^ " , ] + @xmath19july 2008 ; see section 2 for dates of observations .
we used this luminosity as a starting point from which to construct increasingly detailed models of the dust shell surrounding sn2008s , experimenting with both amorphous carbon and silicate grains , as well as mixtures of the two .
we constructed models of sn2008s and its circumstellar dust using the three - dimensional radiative transfer code mocassin ( ercolano et al .
2003 , 2005 , 2008 ) .
we began with the simplest case of a homogeneous constant density shell , obtaining a match to the observed mid - ir fluxes by varying the inner and outer radii , r@xmath20 and r@xmath21 , and the dust mass .
we assumed that the central source had a luminosity of 3@xmath2210@xmath23l@xmath0 and a temperature of 10000k .
the high visual optical depth of the dust shell being considered means that the resulting sed is insensitive to the choice of central star temperature . for the amorphous carbon models
, we used the grain optical constants of hanner ( 1988 ) , with particle sizes following a standard mrn distribution ( mathis , rumpl & nordsieck 1977 ) with a@xmath24=0.005@xmath14 m and a@xmath25=0.25@xmath14 m ; the best match to the observations was obtained with r@xmath20 = 75 au and r@xmath26 = 425 au , and a dust mass of 2@xmath2210@xmath27m@xmath0 .
these inner and outer radii are similar to those estimated by botticella et al .
( 2009 ) for the progenitor s dust shell .
we then constructed models in which the dust density in the shell was inversely proportional to the square of the radius .
this would be expected if the dust was forming continuously in an outflow from the star .
we varied the inner and outer radii , and the total dust mass in the shell , to obtain a good fit .
we found that a standard mrn dust size distribution with grain sizes between 0.005 and 0.25@xmath14 m could not reproduce the steepness of the observed sed between 3.6 and 4.5@xmath14 m .
prieto et al . ( 2008 ) suggested that the shape of the sed required the presence of large dust grains , and we found an improved fit by using an mrn distribution with grain sizes between 0.1 and 0.5@xmath14 m .
the match to our observations is quite sensitive to the adopted inner radius , which is found to be 85 au .
unfortunately , the constraints on the outer radius of the dust shell are weak , as this only significantly affects the sed at wavelengths of @xmath910@xmath14 m and longer .
we therefore considered two alternative physically plausible models . in the first case , we considered a dust shell with an outer radius 5 times the inner radius . for an expansion velocity of @xmath2870 km s@xmath5 ( consistent with the late - time [ ca ii ] line widths ; see section 3.4 )
, such an outer radius would correspond to an age of less than 30 years .
thompson et al . ( 2008 ) estimated that the dust - enshrouded phase for a star similar to the progenitor of sn 2008s could last up to ten thousand years .
we therefore constructed alternative models of the dust shell with the same inner radius but with outer radii of 20,000 au and 100,000 au , corresponding respectively to the distances from the star reached by a 70kms@xmath5 outflow after 1,500 years and 7,500 years .
all of these models gave a good fit to the observed sed , and lie below the upper limit to the 24-@xmath14 m flux .
the total dust mass is 1.2@xmath2210@xmath27m@xmath0 for r@xmath21 = 425 au , and 3.49@xmath2210@xmath29m@xmath0 for r@xmath21 = 100,000 au .
henceforth we refer to the model with r@xmath21 = 425 au as the ` small ' dust shell , and the r@xmath21 = 100,000 au model as the ` large ' dust shell .
figure [ progenitor_models ] compares the seds predicted by these two models with the prieto et al .
( 2008 ) optical and infrared observations of the progenitor .
we also investigated whether silicate dust shells could account for the observed sed of the progenitor of sn 2008s . using the ` astronomical silicate ' optical constants of draine & lee ( 1984 ) , the dust mass required to approximately match the sed was considerably higher than for amorphous carbon grains , due to the greater transparency in the visible and near - ir of silicate grains .
a full match to the observations was impossible because the dust masses required to match the steepness of the sed between 3.6 and 4.5 @xmath14 m give rise to a strong , broad silicate absorption feature at 10 @xmath14 m , which is inconsistent with the observed 8@xmath14 m flux .
silicate dust shells with larger outer radii of 20,000au or 100,000 au also predicted a 24-@xmath14 m flux much higher than the observed upper limit for the progenitor of sn2008s ( figure [ progenitor_models ] ) . with silicaceous dust
thus ruled out for the progenitor s sed , we modelled the dust shell at subsequent epochs using carbon dust having the same density distribution and the same minimum and maximum grain radii as the progenitor ` large ' dust shell model .
the progenitor model input parameters are summarised in table [ model_parameters ] for the large and small shell models , for both silicaceous and carbon dust .
although we constructed models using both amorphous carbon and graphite , using graphite optical constants from draine & lee ( 1984 ) , all dust masses quoted are for the case of amorphous carbon ; graphite grains required dust masses 7% higher in each case .
llllllllllll epoch & & + & & & @xmath30 & total mass ( m@xmath0 ) & visible fraction & composition & l(l@xmath0 ) & t(k ) & r(r@xmath0 ) + & ( au ) & ( light days ) & ( au ) & ( light days ) + progenitor & 85 & 0.49 & 425 & 2.45 & 11.2 & 1.2@xmath2210@xmath27 & 1.00 & ac@xmath19 & 2.6@xmath2210@xmath23 & 10000 & 53.7 + & 85 & 0.49 & 100000 & 577.6 & 13.5 & 3.49@xmath2210@xmath29 & 1.00 & ac@xmath19 & 3.3@xmath2210@xmath23 & 10000 & 60.5 + & 85 & 0.49 & 425 & 2.45 & 188.0 & 3.3@xmath2210@xmath31 & 1.00 & silicates@xmath32 & 3.9@xmath2210@xmath33 & 10000 & 208 + & 85 & 0.49 & 20000 & 115.5 & 225.4 & 1.9@xmath2210@xmath1 & 1.00 & silicates@xmath32 & 5.2@xmath2210@xmath33 & 10000 & 240 + + day 17 & 1250 & 7.22 & 100000 & 577.6 & 0.76 & 3.45@xmath2210@xmath29 & 0.07 & ac@xmath19 & 1.57@xmath2210@xmath34 & 8000 & 2060 + + day 180 & 1250 & 7.22 & 100000 & 577.6 & 0.76 & 3.45@xmath2210@xmath29 & 0.45 & ac@xmath19 & + + @xmath19amorphous carbon ; optical constants from hanner ( 1988 ) .
@xmath32astronomical silicates ; optical constants from draine & lee ( 1984 ) .
our _ spitzer _ observations of sn2008s took place on 6.8 february 2008 ( irac ) and 7.5 february ( mips ) , 17@xmath74 days after the explosion .
we combined our mid - ir flux measurements with the optical observations reported by smith et al .
( 2009 ) from 6.1 february ( b- and v - band ) and 6.7 february ( r - band ) .
we also followed botticella et al .
( 2009 ) in extrapolating their day 41 jhk photometry back to day 17 .
we dereddened the photometry using the estimated galactic foreground reddening of e(b - v)=0.34 ( section 2 ) .
the resulting sed can be well fitted by the sum of a 6000k blackbody plus a 550k blackbody modified by a @xmath35 emissivity law ( figure [ day17_bbs ] ) .
the luminosities of the two components are 12.9@xmath2210@xmath36l@xmath0 and 2.2@xmath2210@xmath36l@xmath0 respectively , giving a total luminosity of 15.1@xmath2210@xmath36l@xmath0 .
smith et al .
( 2009 ) estimated that the peak luminosity of sn 2008s was 3@xmath2210@xmath34l@xmath0 .
the flash would easily have vaporised dust at the inner edge of the shell .
the removal of the inner edge of the shell would then expose more dust to the intense radiation field , and dust destruction could proceed rapidly outwards . to investigate
how much of the shell could have been destroyed in this way , we ran a series of models with successively larger inner radii .
a sublimation temperature of 2200k is often assumed for amorphous carbon .
we found that for this value , the dust shell surrounding sn2008s could have been hollowed out to a radius of 215au .
this resulted in a new optical depth of 2.1 for the small shell and 4.5 for the large shell .
using this hollowed shell and varying only the source parameters , it was not possible to match the day 17 observations , as the optical depth was too high and gave too much reprocessed radiation .
we therefore investigated what new inner radius was required to match the day 17 observations , and what dust sublimation temperature would be required to allow the shell to be hollowed out to this radius .
we found that a radius of 1250 au was required to give dust cool enough at the inner edge , and a low enough optical depth to match the observations .
the mid - infrared observations on day 17 and in july 2008 require that there still be a significant amount of dust remaining around the supernova .
therefore , we can rule out the progenitor ` small ' shell model , as this would have been completely vaporised by the explosion , with no resulting mid - ir excess post - outburst . for the smith et al .
( 2009 ) peak luminosity of 3@xmath2210@xmath34l@xmath0 , the required dust sublimation temperature is 1050 k if dust inside 1250 au is to be vaporised .
experimentally , the sublimation temperature of solid carbon varies from 2870 k for a carbon partial pressure of 1 pa ( 10@xmath27 bar ) to 1850 k at 10@xmath37 pa ( grigoriev & meilikhov 1997 ) .
an extrapolation of their relation to 10@xmath38 pa , corresponding to the gas density of @xmath910@xmath33 cm@xmath29 at 1250 au ( section 4 ) combined with a carbon species abundance of 10@xmath29 , predicts a carbon sublimation temperature of 1440 k. for this value of the grain sublimation temperature a peak outburst luminosity of 1.5@xmath40 l@xmath0 would have been required in order to hollow out the dust shell to 1250 au . for a shell
in which the dust density is proportional to r@xmath1 , each annulus of constant thickness contains the same mass .
thus , hollowing out the shell to a new inner radius of 1250 au only removes a small fraction of the original mass .
however , the optical depth is determined by the column density of the dust , proportional to r@xmath41 - r@xmath42 for an r@xmath1 shell , and so a dust shell hollowed out by a radiation flash can have a much lower optical depth but a similar mass to the pre - existing shell .
the mass of the ` large ' shell , originally 3.49@xmath2210@xmath29m@xmath4 , is reduced to 3.45@xmath2210@xmath29m@xmath4 when hollowed out to 1250 au . meanwhile ,
its visual optical depth is reduced from 13.5 to 0.76 .
a consideration of light travel time is crucial to modelling the observed sed on day 17 . from the point of view of the observer , only dust lying within an ellipsoidal region defined such that the supernova - ellipsoid - observer distance is 17 light days greater than the sn - observer distance will have been heated by the flash from the supernova ( figure [ lighttraveltime ] ) .
emission reaching the observer from dust outside this ellipsoid left the dust before the supernova flash reached it .
the mass illuminated by the flash after 17 days is less than 7 per cent of the total mass of the shell , for an outer radius of 100,000 au . to account for this effect , we created a customised version of mocassin in which the sed was constructed by summing emission only from those cells lying within the ellipsoid described above , viewed from the direction in which the ellipse opens out .
we assumed a constant luminosity and temperature for the illuminating source over the first 17 days .
the bolometric luminosity estimates of botticella et al .
( 2009 ) indicate that the source luminosity changed by a factor of @xmath431.25 over this period . with an inner radius of 1250 au
, we obtained a fit to the observed day 17 sed using an 8000k source with a luminosity of 15.7@xmath2210@xmath36l@xmath0 ( table 2 ) , as shown in figure [ day17_model ] ; details of the model are given in table 3 .
the fit can be considered reasonable ; it lies somewhat below our measured 24-@xmath14 m flux level for day 17 , but this flux has a signal to noise ratio of only 3.4 . for their day 17 sed modelling , botticella et al .
used single - radius ( 0.5 @xmath14 m ) carbon grains having an r@xmath1 density distribution between r@xmath44 = 2000 au and r@xmath45 = 20000 au .
the dereddened sed of sn2008s in july 2008 , six months after outburst , can be fitted by the sum of a 4000k blackbody plus a 450k blackbody modified by a @xmath35 emissivity law , with luminosities of 8.2@xmath2210@xmath33l@xmath0 and 3.6@xmath2210@xmath33l@xmath0 , respectively .
the combined luminosity is then 1.2@xmath2210@xmath36l@xmath0 ( table 2 ) , corresponding to a fading of a factor of 12.5 between day 17 and day 180 .
we modelled the july 2008 fluxes using the dust distribution derived for the progenitor but again removing the dust in the inner regions , out to a radius of 1250 au , which had been vaporised by the outburst according to our day 17 modelling .
by july 2008 , 45 per cent of the 100,000au shell has been illuminated by the outburst of sn 2008s . to account for light travel time
, we constructed a model consisting of five nested ellipsoids , each corresponding to an interval of 36 days , with the luminosity and temperature of the illuminating source declining from the outer to inner ellipsoids .
the input luminosities were derived by averaging the bolometric light curve presented in figure 3 of botticella et al .
( 2009 ) over the corresponding 36 day period .
the central source temperatures were taken from table 8 of botticella et al .
( 2009 ) ; however , the model output was found to be insensitive to the adopted temperature .
the fraction of the shell seen to be illuminated by the flash after 180 days is 44.7 per cent .
the sed produced by this model gave a good fit to the observed optical and ir sed ( see table 3 and figure 6 ) .
botticella et al .
( 2009 ) invoked dust formation in order to account for the observed near - ir photometric flux levels after day 120 , but we find that our outburst - heated progenitor wind model can account for the observed near - ir flux levels in july 2008 ( @xmath9day 180 ) . our 14 july 2008 ( day 175 ) spectrum of sn 2008s is shown in figure [ spectrum ] , with identifications for some of the stronger emission lines indicated . a comprehensive list of emission line identifications for sn 2008s can be found in table a2 of botticella et al .
( 2009 ) . as noted previously by smith et al .
( 2009 ) and botticella et al .
( 2009 ) , the most striking aspect of the post - outburst spectra of sn 2008s is the narrowness of the h@xmath46 line : we measured its full width at half - maximum ( fwhm ) to be 690kms@xmath5 , much lower than the line widths of several thousand kms@xmath5 typically observed after the outbursts of ccsne .
the likely origin for the observed emission lines is the gaseous component of the dusty circumstellar outflow that predated the sn 2008s outburst .
botticella et al .
( 2009 ) observed smooth broad wings ( @xmath93000 km s@xmath5 fwhm ) in several lines , including h@xmath46 , in their early post - outburst spectra , which may have been caused by electron scattering , although an underlying high - velocity outflow is a possibility .
overall , they identified three line - width components , broad , intermediate and narrow , whose widths tended to decline with time .
we could obtain a good fit to our day 175 h@xmath46 profile with three such components : a broad component whose fwhm was 1460 km s@xmath5 , accounting for 25.1% of the total flux ; an intermediate component of fwhm 640 km s@xmath5 , accounting for 71.2% of the flux ; and a narrow component of fwhm 165 km s@xmath5 , accounting for the remaining 3.7% of the flux .
the [ ca ii ] 7291,7324 lines had a dominant ` narrow ' component of fhwm 140 km s@xmath5 ( all quoted fwhms were corrected in quadrature for the instrumental resolution of 1.8 fwhm ) .
if the intermediate and narrow emission line components originate in a spherically symmetric circumstellar outflow , then their half widths at half - maximum should be approximately equal to the expansion velocity in the region(s ) where they originate .
the day 175 h@xmath46 intermediate component then yields an expansion velocity of 320 km s@xmath5 , while the narrow h@xmath46 component and the forbidden [ ca ii ] 7291 , 7324 lines , which may originate from lower density material , further out in the wind , yield expansion velocities of 70 - 80 km s@xmath5 .
the intermediate - width velocity component may correspond to material in the innermost regions of the wind that has been accelerated by the radiation pressure pulse from the outburst .
our analysis of sn2008s shows that both the progenitor sed and the day 17 and day 180 seds can be self - consistently fitted by the same r@xmath1 density distribution of amorphous carbon grains , whose pre - outburst inner shell radius of r@xmath20 = 85 au increased to 1250 au after the outburst , due to grain evaporation by the light flash . as a result of this vaporisation of the innermost grains , the visual optical depth of the dust shell ( @xmath47 r@xmath41 - r@xmath42 ) decreased from 13.5 to 0.76 , but
the total mass of dust ( @xmath47 r@xmath21 - r@xmath20 ) decreased from its intial value by less than 2% .
for an expansion velocity of 70 km s@xmath5 , our derived parameters imply a dust mass - loss rate of 5.2@xmath6 m@xmath4 yr@xmath5 from the progenitor of sn 2008s .
if the gas to dust mass ratio in the outflow is 100 - 200 , then this corresponds to an overall mass - loss rate of 0.5 - 1.0@xmath3 m@xmath4 yr@xmath5 .
matsuura et al . ( 2009 ) find a high fraction of carbon - rich stars amongst the most luminous agb stars in the lmc , with over ten such stars having inferred mass - loss rates of 0.5 - 1.0@xmath3 m@xmath4 yr@xmath5 , similar to that inferred here for the progenitor of sn 2008s .
the total mass of the dust shell around sn 2008s could be larger than the 3.5@xmath2210@xmath29m@xmath0 in our dust model of outer radius 100,000 au , since the total mass is proportional to the outer radius , which is not constrained by the available data . for our derived dust density at 1250
au of 1.1@xmath48 g cm@xmath29 , the gas particle density n@xmath49 there is ( 0.5 - 1.0)@xmath50 cm@xmath29 , implying that n@xmath49 = ( 1 - 2)@xmath51 cm@xmath29 at 10 au , the effective emitting radius of the central source on both day 17 and day 180 , and ( 1 - 2)@xmath52 cm@xmath29 at 85 au , the presumed dust condensation radius of the progenitor wind .
the latter density is similar to that inferred by cherchneff et al .
( 2000 ) for the region of dust formation in late - type carbon - rich wolf - rayet outflows . while botticella et al .
( 2009 ) adopted a model in which sn 2008s had undergone episodic mass loss and a sharp change in dust chemistry , fitting the pre - outburst sed with silicate grains and the post - outburst sed with carbon grains , we find that a featureless grain type such as amorphous carbon is required to fit both the pre - outburst and post - outburst seds of sn 2008s .
this appears surprising at first sight , given that outflows from cool supergiants or super - agb stars might be expected to be oxygen - rich , thus favoring the formation of silicate grains .
however , we find silicate models to be completely incompatible with the observed optical and ir seds , producing a strong 10-@xmath14 m absorption feature for high-@xmath30 pre - outburst models and a strong 10-@xmath14 m emission feature for lower-@xmath30 post - outburst models .
comparison of our models for the july 2008 sed of sn 2008s with the _ spitzer _ and michelle mid - ir photometry and the _ spitzer _ irs spectrum constrain any silicate grain component to contribute significantly less than 10% of the overall dust mass .
since cool m supergiant stars always show silicate dust features in the mid - ir , the absence of silicates in the outflow around sn 2008s and its progenitor may count as evidence against a supergiant precursor and in favor of a super - agb star progenitor .
surface carbon enrichment of agb stars via the 3rd dredge - up can produce carbon stars .
however , at the higher mass end of the agb mass distribution , envelope burning is predicted to convert dredged - up carbon to nitrogen , lowering the c / o ratio below unity and allowing silicate grains to form . given the absence of silicate grains around sn 2008s , a mechanism to suppress or bypass envelope burning in super - agb stars may be necessary , perhaps related to the severe mass loss stripping of the envelope .
prieto et al . (
2009 ) have obtained a post - outburst _ spitzer _
irs spectrum of the 2008 optical transient in ngc 300 ( monard 2008 ) , whose evolution has shown many similarities to that of sn 2008s ( bond et al .
prieto et al .
find two strong emission features in its mid - ir spectrum , which they attribute to carbon - rich particles , from a comparison with the mid - ir spectra of a number of c - rich post - agb objects . as well as this similarity between the inferred dust compositions of sn 2008s and ngc 300-ot , prieto et al . also conclude that most of the dust around ngc 300-ot survived its outburst .
the progenitor of sn2008s is estimated to have had an initial mass in the range 610m@xmath0 ( prieto et al .
2008 ; botticella et al .
thompson et al . ( 2008 ) suggested that a dust - enshrouded phase may occur during the lives of a large fraction of all massive stars , which may indicate that this type of transient could be a significant additional source of dust in the early universe .
thompson et al . estimated that the dust - enshrouded phase of sn 2008s - type objects could last for up to 2@xmath8 years , in which case our derived dust mass - loss rate of 5.2@xmath6 m@xmath4 yr@xmath5 implies that up to 0.01 m@xmath4 of dust could be ejected , along with 1 - 2 m@xmath4 of gas , a significant fraction of the total stellar mass .
an ejected dust mass of 0.01 m@xmath4 is within a factor of two of the largest dust mass derived by any of the recent studies of young ccsne ejecta ( sn 2003gd ; sugerman et al .
up to 1.0 m@xmath4 of cold dust was claimed to be present in the young galactic supernova remnants cas a and kepler by morgan et al .
( 2003 ) and dunne et al .
( 2003 ) , although their analysis was disputed by krause et al .
recently dunne et al .
( 2009 ) and gomez et al .
( 2009 ) have returned to this issue with new observations , from which they still infer 0.1 - 1.0 m@xmath4 of cold dust in each remnant . since the dust around sn 2008s is definitely not silicaceous , and is consistent with amorphous carbon , the atoms in the dust will have been freshly synthesised , with the outflow enriching the ism in this dust component .
dwek et al .
( 2007 ) have estimated that the 2@xmath40 m@xmath4 of dust in the @xmath53 galaxy j114816.64 + 5251 must have been created within a timescale of 400 myr .
given that the lifetime of a 5 m@xmath4 star is only 100 myr ( maeder & meynet 1989 ) and that standard initial mass functions favour lower mass stars over high mass stars , then with some initial heavy element seeding from massive ccsne there could be scope for agb or super - agb stars with initial masses between 6 and 10 m@xmath4 , of which the progenitor of sn 2008s appears to be an example , to make a contribution to the dust enrichment of such galaxies .
we estimate the possible contribution of such stars to the dust content of j114816.64 + 5251 as follows : for a salpeter imf with limits of 0.1 and 120 m@xmath4 , 0.20 per cent of stars have masses between 6 and 10 m@xmath4 , while a further 0.29 per cent have masses between 4 and 6 m@xmath4 .
the star formation history of j114816.64 + 5251 is uncertain , with evidence for a current star formation rate of up to 3000 m@xmath0 yr@xmath5 ( dwek et al .
2007 ) . as they noted , such a high sfr almost certainly could not have been sustained for 400 myr ; @xmath9 200 m@xmath0 yr@xmath5 is likely to be a more representative average .
this would result in a total stellar mass of 8@xmath2210@xmath54 m@xmath0 , giving 6.6@xmath2210@xmath55 stars with masses between 4 and 6 m@xmath0 and 4.5@xmath2210@xmath55 stars with masses between 6 and 10 m@xmath0 .
if each star in both these mass ranges produce 0.01 m@xmath0 of dust , then their total contribution would be 6.6@xmath2210@xmath36 and 4.5@xmath2210@xmath36 m@xmath0 for the low and high mass ranges respectively , before dust destruction is accounted for .
a similar calculation can be made for ` normal ' supernovae ; assuming that all stars with initial masses greater than 8 m@xmath0 explode as supernovae , and that each produces 0.01 m@xmath0 of dust ( e.g. sugerman et al .
2006 ) , their total contribution before accounting for dust destruction would be 6.0@xmath2210@xmath36 m@xmath0 .
thus , the amount of dust produced by sn2008s - type events could be comparable to the contribution of ` normal ' supernovae .
for both types of object , these results are sensitive to the lower bound of the imf ; raising the imf lower mass limit from 0.1 m@xmath4 to 0.2 m@xmath0 would approximately double the quantity of dust produced .
another parameter that these estimates are sensitive to is the dust lifetime against destruction .
some fraction of the dust produced by both sources would be destroyed during the galaxy s 400 myr life .
the lifetime of dust in this environment , and hence the surviving fraction , is highly uncertain ; for a typical milky way dust lifetime of 5@xmath2210@xmath55 years ( dwek et al .
2007 ) , the surviving fraction after 400myr would be 70% . for star formation bursts as high as 3000 m@xmath0
yr@xmath5 , dust lifetimes could be as short as 25 myr ( dwek et al .
however , such high star formation rates are unlikely to be sustained for periods longer than this , and sn ejecta dust from the massive stars that resulted from the burst would seem more at risk from destruction than dust from longer - lived high - mass agb stars , which would be further from the sites of active star formation .
even for a dust lifetime of 500 myr , with 70% of grains surviving over 400 myr , 4 - 10 m@xmath0 agb stars and sne would respectively account for only 3.9 and 2.1 per cent of the 2@xmath2210@xmath55 m@xmath0 of dust observed in j114816.64 + 5251 .
it seems difficult for either agb stars or core collapse sne to account for the masses of dust inferred for such galaxies , if current estimates for the galaxy dust masses , and for the dust injection rates by stars , are correct .
if sn2008s was not a bona fide supernova event , then a star will remain at the centre of the dust shell , and could presumably explode later as a supernova .
we investigated whether a subsequent supernova could vaporise the dust shell , using grids formed from simple geometric expansion of our derived dust distribution . adopting a representative type ii supernova peak luminosity of 10@xmath56l@xmath0
, we find that if a supernova occurred in the very near future , then dust within about 7500au of the star would be vaporised , reducing @xmath30 from 0.76 to 0.15 .
the surviving dust shell would then have a mass of 3.2@xmath2210@xmath29 m@xmath0 , still 93% of the initial value .
if the supernova does not occur within the next 500 years , then a 70kms@xmath5 outflow velocity will have carried the inner edge of the shell beyond the vaporisation radius for this luminosity , and the entire dust shell would survive the supernova light flash .
we thank the telescope staff for their help in acquiring the july 2008 gmos and michelle observations ( program i d gn-2008b - q-44 ) which were obtained at the gemini observatory , operated by the association of universities for research in astronomy , inc .
, under a cooperative agreement with the nsf on behalf of the gemini partnership : the national science foundation ( united states ) , the science and technology facilities council ( united kingdom ) , the national research council ( canada ) , conicyt ( chile ) , the australian research council ( australia ) , ministerio da ciencia e tecnologia ( brazil ) and ministerio de ciencia , tecnologia e innovacion productiva ( argentina ) .
this work is based in part on observations made with the spitzer space telescope , which is operated by the jet propulsion laboratory , california institute of technology under a contract with nasa .
we thank an anonymous referee for constructive comments .
arbour r. , boles t. , 2008 , central bureau electronic telegrams 1234 , 1 barlow m. j. , et al .
, 2005 , apj , 627 , l113 bertoldi , f. , carilli , c. l. , cox , p. , fan , x. , strauss , m. a. , beelen , a. , omont , a. , zylka , r. , 2003 , a&a , 406 , l55 bode m. f. , evans a. , 1980 mnras , 193 , 21p bond h. e. , bedin l. r. , bonanos a. z. , humphreys r. m. , monard l. a. g. , prieto j. l. , walter f. m. , 2009 , apj , 695 , l154 botticella m. , et al . , 2009 , mnras , in press ( arxiv:0903.1286 ) draine b. t. , lee h. m. , 1984 , apj , 285 , 89 dunne l. , eales s. , ivison r. , morgan h. , edmunds m. , 2003 , nature , 424 , 285 dunne l. , maddox s. j. , ivison r. j. , rudnick l. , delaney t. a. , matthews b. c. , crowe c. m. , gomez h. l. , eales s. a. , dye s. , 2009 , mnras , 394 , 1307 dwek e. , 1983 , apj , 274 , 175 dwek e. , 1985 , apj , 297 , 719 dwek e. , galliano f. , jones a. p. , 2007 ,
apj , 662 , 927 ercolano b. , barlow m.j . , storey p.j .
, liu x - w . , 2003 , mnras , 340 , 1136 ercolano b. , barlow m.j . , storey p.j . , 2005 , mnras , 362 , 1038 ercolano b. , young p.r .
, drake j.j . , raymond j.c . , 2008 , apjs , 175 , 534 gomez h.l . , dunne l. , ivison r. , reynoso e.m . , thompson m.a . , sibthorpe b. , eales s.a . , delaney t.m
. , maddox s. , isaak k. , 2009 , mnras , in press ( arxiv:0905.2564 ) graham j. r. , meikle w. p. s. , selby m. j. , allen , d. a. , evans , a. , pearce , g. , bode , m. f. , longmore , a. j. , williams , p. m. , 1983 , nature , 304 , 709 grigoriev s. , meilikhov e. z. , 1997 , handbook of physical quantities , crc press , 324 hanner m. , 1988 , nasa conf . pub . 3004 , 22 karachentsev i. d. , sharina m. e. , huchtmeier w. k. , 2000 , a&a , 362 , 544 krause , o. , birkmann s. m. , rieke g. h. , lemke d. , klaas u. , hines d. c. , gordon k. d. , 2004 , nature , 432 , 596 maeder a. , meynet g. , 1989 , a&a , 210 , 155 massey p. , strobel k. , barnes j. v. , anderson e. , 1988 , apj , 328 , 315 mathis j. s. , rumpl w. , nordsieck k. h. , apj , 217 , 425 matsuura m. , barlow m. j. , zijlstra a. a. , whitelock p. a. , cioni m .-
l. , groenewegen m. a. t. , volk k. , kemper f. , kodama t. , lagadec e. , meixner m. , sloan g. c. , srinivasan s. , 2009 , mnras , 396 , 918 monard l. a. g. , 2008 , iauc 8946 , 1 morgan h. l. , edmunds m. g. , 2003 , mnras , 343 , 427 morgan h. l. , dunne l. , eales s. a. , ivison r. j. , edmunds m. g. , 2003 , apj , 597 , l33 prieto j. l. , kistler m. d. , thompson t. a. , ykse h. , kochanek c. s. , stanek k. z. , beacom j. f. , martini p. , pasquali a. , bechtold j. , 2008 , apj , 681 , l9 prieto j. l. , sellgren k. , thompson t. a. , kochanek c. s. , 2009 , apj , submitted ( arxiv:0907.0230 ) schlegel d. j. , finkelbeiner d. p. , davis m. , 1998 , apj , 500 , 525 smith n. , et al . , 2009 , apj , 697 , l49 sugerman b.e.k .
, 2003 , aj , 126 , 1939 sugerman b. e. k. et al . , 2006 , sci , 313 , 196 thompson t. a. , prieto j. l. , stanek k. z. , kistler m. d. , beacom j. f. , kochanek c. s. , 2008 , arxiv:0809.0510 welch d. l , clayton g. c. , campbell a. , barlow m. j , sugerman b. e. k. , meixner m. , bank s. h. r. , 2007 , apj , 669 , 525 welch d. l , clayton g. c. , sugerman b. e. k. , 2008 , central bureau electronic telegrams 1331 , 1 wesson r. , fabbri j. , barlow m. , meixner m. , 2008 , central bureau electronic telegrams 1381 , 1 | sn2008s erupted in early 2008 in the grand design spiral galaxy ngc 6946 .
the progenitor was detected by prieto et al . in _ spitzer space telescope _
images taken over the four years prior to the explosion , but was not detected in deep optical images , from which they inferred a self - obscured object with a mass of about 10m@xmath0 .
we obtained _ spitzer _ observations of sn 2008s five days after its discovery , as well as coordinated gemini and _ spitzer _ optical and infrared observations six months after its outburst .
we have constructed radiative transfer dust models for the object before and after the outburst , using the same r@xmath1 density distribution of pre - existing amorphous carbon grains for all epochs and taking light - travel time effects into account .
we rule out silicate grains as a significant component of the dust around sn 2008s .
the inner radius of the dust shell moved outwards from its pre - outburst value of 85 au to a post - outburst value of 1250 au , attributable to grain vaporisation by the light flash from sn 2008s .
although this caused the circumstellar extinction to decrease from a@xmath2 = 15 before the outburst to 0.8 after the outburst , we estimate that less than 2% of the overall circumstellar dust mass was destroyed .
the total mass - loss rate from the progenitor star is estimated to have been 0.5 - 1.0@xmath3 m@xmath4 yr@xmath5 .
the derived dust mass - loss rate of 5@xmath6 m@xmath0 yr@xmath5 implies a total dust injection into the ism of up to 0.01 m@xmath0 over the suggested duration of the self - obscured phase .
we consider the potential contribution of objects like sn 2008s to the dust enrichment of galaxies .
supernovae : individual : sn2008s ; circumstellar matter |
complex - energy `` eigenfunctions '' made their dbut in quantum mechanics through the hands of gamow , in the theory of alpha - decay @xcite .
gamow imposed an `` outgoing wave boundary condition '' on the solutions of the schrdinger equation for an alpha - particle trapped in the nucleus .
since there is only an _ outgoing _ flux of alpha - particles , the wavefunction @xmath0 must behave far from the nucleus as ( for simplicity , we consider an s - wave , and use units such that @xmath1 ) @xmath2 this boundary condition , together with the requirement of finiteness of the wave function at the origin , gives rise to a quantization condition on the values of @xmath3 ( and , therefore , on the values of @xmath4 ) .
it turns out that such values are complex : @xmath5 and so it follows that @xmath6 thus , if @xmath7 , the probability of finding the alpha - particle in the nucleus decays exponentially in time .
the lifetime of the nucleus is then given by @xmath8 , and the energy of the emitted alpha - particle by @xmath9 .
although very natural , this interpretation suffers from some difficulties .
how can the energy , which is an observable quantity , be complex ?
( in other words , how can the hamiltonian , which is a hermitean operator , have complex eigenvalues ? ) also , these `` eigenfunctions '' are not normalizable , since @xmath10 positive implies @xmath11 positive and so , according to ( [ |psi| ] ) , @xmath12 diverges exponentially with @xmath13 . in spite of such problems ( which , in fact , are closely related ) , it is a fact of life that alpha - decay , as well as other types of decay , does obey an exponential decay law and , in many cases , gamow s method provides a very good estimate for the decay rate .
why this method works is a question that has been addressed in the literature using a variety of techniques @xcite .
here we examine this question in a very elementary way , using techniques that can be found in any standard quantum mechanics textbook and some rudiments of complex analysis .
this paper is structured as follows . in section [ ii ] , we show gamow s method in action for a very simple potential . some of the results obtained there are used in section [ iii ] , where we study the time evolution of a wave packet initially confined in the potential well defined in section [ ii ] .
this is done with the help of the propagator , built with the _ true _ eigenfunctions ( i.e. , associated to _
real _ eigenenergies ) of the hamiltonian . as a bonus ,
we show that the exponential decay law is not valid either for very small or for very large times .
this is the content of section [ iv ] , where the region of validity of the exponential decay law is roughly delimited .
in order to exhibit gamow s method in action , we shall study the escape of a particle from the potential well given by : @xmath14 ( escape from this potential well was studied in detail in refs.@xcite . in this section
we follow closely the treatment of ref.@xcite . )
the positive dimensionless constant @xmath15 is a measure of the `` opacity '' of the barrier at @xmath16 ; in the limit @xmath17 , the barrier becomes impenetrable , and the energy levels inside the well are quantized . if @xmath15 is finite , but large , a particle is not confined to the well anymore , but it usually stays there for a long time before it escapes . if @xmath15 is small , the particle can easily tunnel through the barrier , and quickly escape from the potential well .
metastability , therefore , can only be achieved if the barrier is very opaque , i.e. , @xmath18 .
since this is the most interesting situation , we shall assume this to be the case in what follows . to find out how fast the particle escapes from the potential well
, we must solve the schrdinger equation @xmath19 @xmath20 is a particular solution of this equation , provided @xmath21 satisfies the time - independent schrdinger equation @xmath22 denoting the regions @xmath23 and @xmath24 by the indices 1 and 2 , respectively , the corresponding wave functions @xmath25 @xmath26 satisfy the free - particle schrdinger equation : @xmath27 since the wall at the origin is impenetrable , @xmath28 must be zero ; the solution of eq .
( [ free ] ) which obeys this boundary condition is @xmath29 to determine @xmath30 , we follow gamow s reasoning @xcite and require @xmath30 to be an outgoing wave . therefore , we select , from the admissible solutions of eq .
( [ free ] ) , @xmath31 the wave function must be continuous at @xmath16 , so that @xmath32 , or @xmath33 on the other hand , the derivative of the wave function has a discontinuity at @xmath16 , which can be determined by integrating both sides of ( [ sch ] ) from @xmath34 to @xmath35 , with @xmath36 : @xmath37 from which there follows another relation between @xmath38 and @xmath39 : @xmath40 combining ( [ cont1 ] ) and ( [ discont2 ] ) , we obtain a quantization condition for @xmath3 : @xmath41 the roots of eq .
( [ quant ] ) are complex and situated in the half - plane @xmath42 ; when @xmath18 , those which are closest to the origin are given by @xcite @xmath43 ( for each one of these roots , which are located in the fourth quadrant of the complex @xmath3-plane , there is a corresponding one in the third quadrant , given by @xmath44 .
the latter are associated to `` growing states '' @xcite and play no role in what follows . )
the corresponding eigenenergies are @xmath45 the imaginary part of @xmath46 gives rise to an exponential decay of @xmath47 , with lifetime equal to @xmath48 since the corresponding value of @xmath49 is very small ( @xmath50 ) , one may be tempted to say that the probability of finding the particle outside the well is negligible in comparison with the probability of finding the particle inside the well .
normalizing @xmath51 in such a way that the latter equals one when @xmath52 , the probability of finding the particle inside the well at time @xmath53 , if it were in the @xmath54-th decaying state at @xmath52 , would be @xmath55 the trouble with this interpretation is that @xmath56 , and so @xmath57 diverges exponentially as @xmath58 , since , according to ( [ phi2 ] ) , @xmath59 outside the well . because of this `` exponential catastrophe '' , the decaying states are nonnormalizible and , therefore , can not be accepted as legitimate solutions of the schrdinger equation ( although one can find in the literature @xcite the assertion that they are `` rigorous '' solutions of the time - dependent schrdinger equation ) .
we now return to eq .
( [ free ] ) and write , for the solution in region 2 , instead of ( [ phi2 ] ) , the sum of an outgoing plus an incoming wave : @xmath60 continuity of the wave function at @xmath16 implies @xmath61 as before , the derivative of the wave function has a discontinuity at @xmath16 , given by eq .
( [ discont1 ] ) , from which it follows , instead of ( [ discont2 ] ) , @xmath62 solving ( [ cont2 ] ) and ( [ discont3 ] ) for @xmath38 and @xmath39 , we find [ ab ] @xmath63 these expressions show a couple of interesting features : * @xmath64 for real values of @xmath3 , implying a zero net flux of probability through @xmath16 ; therefore , unlike the solution found in the previous section , there is no loss or accumulation of probability in the well region . *
@xmath65 if @xmath66 , except if @xmath3 is close to a pole of @xmath67 , in which case @xmath68 may become very large . to find the poles of @xmath38 we must solve the equation @xmath69 , which , after some algebraic manipulations , reads @xmath70 this is the same as eq .
( [ quant ] ) ! is this a coincidence ?
in fact , no . according to ( [ ab ] ) , @xmath38 and @xmath39 have the same poles ; therefore , near a pole both @xmath68 and @xmath71 are very large , and eqs .
( [ cont2 ] ) and ( [ discont3 ] ) become equivalent to eqs .
( [ cont1 ] ) and ( [ discont2 ] ) , respectively . in
what follows , we shall show that the poles of @xmath38 ( and @xmath39 ) play an important role in the decay process .
suppose that at @xmath52 the particle is known to be in the region @xmath72 with probability 1 ; in other words , its wave function @xmath73 is zero outside the well .
then , at a later time @xmath53 , the wave function is given by @xmath74 where the propagator , @xmath75 , can be written as @xmath76 the function @xmath77 is the solution of eq .
( [ sch ] ) corresponding to the energy @xmath4 : @xmath78 with this normalization , these functions satisfy the completeness relation @xcite @xmath79 eqs .
( [ psi1])([varphi ] ) give , for @xmath72 , @xmath80 where @xmath81 it is clear that the integral over @xmath3 is dominated by the resonances , i.e. , the neighborhood of the poles of @xmath67 .
since , for @xmath82 , @xmath83 when @xmath84 in the fourth quadrant , one can rotate must be an analytic function of @xmath3 , but this can be shown to be the case @xcite if @xmath85 is continuous in @xmath86 $ ] . ]
the integration contour by @xmath87 in the clockwise sense ( see fig
. 1 ) , thus obtaining @xmath88 where @xmath89 and @xmath90 the sum in ( [ psi3 ] ) takes into account the poles of @xmath67 which are situated in the region @xmath91 , and it corresponds to an expansion in gamow states ( for @xmath72 ) .
let us put aside , for a moment , the integral in ( [ psi3 ] ) ( it will be discussed in the next section ) .
then , the `` nonescape '' probability ( i.e. , the probability of finding the particle inside the well ) is given by @xmath92 where @xmath93 .
for @xmath18 , the interference terms are usually negligible , for @xmath94 and , therefore , the functions @xmath95 are approximately orthogonal . on the other hand ,
since the decay rate @xmath10 of the @xmath54-th decaying mode is a rapidly increasing function of @xmath54 ( @xmath96 ) , the decay becomes almost a pure exponential one when @xmath97 .
the system , therefore , `` loses memory '' of the initial state . finally , let us note that no exponential catastrophe occurs with @xmath98 .
in fact , one can easily show , using ( [ psi1 ] ) , ( [ g ] ) , ( [ completeness ] ) and the orthogonality of the eigenfunctions @xmath77 , that @xmath99 so that an exponential growth of @xmath100 is completely ruled out .
in order to derive expression ( [ p(t ) ] ) for the nonescape probability , we had to neglect the first term on the right hand side of ( [ psi3 ] ) . in this section we show that such approximation is not valid either for very small or for very large times . that it can not be valid for very small @xmath53 follows from the fact that initially the decay is slower than exponential @xcite .
this can be easily proved with the help of the continuity equation @xcite : @xmath101_{x = a}.\ ] ] since , by hypothesis , @xmath102 , it follows that @xmath103 when @xmath52 , whereas for the expression ( [ p(t ) ] ) one has @xmath104 at @xmath52 . on the other hand ,
the exponential decay does not last forever .
after a sufficiently long time , it obeys a power law @xcite . to see this , note that the integral in ( [ psi3 ] ) , which we shall denote here by @xmath105 , is dominated by small values of @xmath3 when @xmath106 , and so can it be approximated by @xmath107 therefore , the nonescape probability behaves asymptotically as decays as @xmath108 when @xmath106 , in contrast to eq .
( [ pas ] ) .
however , there is a flaw in their argument ; when properly corrected @xcite , it also leads to @xmath109 asymptotically . ]
@xmath110 comparing ( [ pas ] ) with ( [ p(t ) ] ) , and using ( [ tau ] ) , one finds that they become comparable in magnitude when @xmath111 or , since @xmath18 , when @xmath112 in practice , when the decay begins to obey a power law the nonescape probability is so small @xmath113 that it should be very difficult to observe deviations from exponential decay .
( on the other hand , experimental evidence for non - exponential decay at _ small _ times has been reported recently @xcite . ) in closing this section , we would like to remark that the breakdown of the exponential decay law for either small or large times is not a peculiar feature of the potential ( [ v ] ) .
it is possible to show that an exponential decay can not last forever if the hamiltonian is bounded below @xcite , and can not occur for very small times if , besides that , the energy expectation value of the initial state is finite @xcite conditions which certainly must be satisfied by any realistic hamiltonian or state .
in this paper we showed that decaying states , although plagued by the exponential catastrophe , give a fairly good description of the decay of a metastable state , provided some conditions are satisfied . in fact , the main objective of this paper was to show that one _ can _ compute the decay rate solving the time independent schrdinger equation subject to the `` outgoing wave boundary condition . ''
this is far from being a trivial result , since the corresponding eigenstates are unphysical .
the effectiveness of the decaying states in describing the decay may be understood by noticing @xcite that they are good approximate solutions to the time - dependent schrdinger equation , although nonuniform ones ( i.e. , they are not valid in the entire range of values of @xmath53 and @xmath114 ) .
we thank gernot muenster and pavel exner , for bringing some references to our attention , and gilberto hollauer , for useful discussions .
this work had financial support from cnpq , finep , capes and fujb / ufrj .
1 . complex @xmath3-plane .
the poles of @xmath67 are represented by the small circles .
those in the fourth quadrant give rise to the sum over decaying modes in eq .
( [ psi3 ] ) when one rotates the integration contour of eq .
( [ psi2 ] ) the positive real semi - axis by @xmath87 in the clockwise sense ( dashed line ) . | we examine gamow s method for calculating the decay rate of a wave function initially located within a potential well . using elementary techniques , we examine a very simple , exactly solvable model , in order to show why it is so reliable for calculating decay rates , in spite of its conceptual problems .
we also discuss the regime of validity of the exponential decay law . |
the atlas , cms and lhcb experiments feature detection capabilities at forward rapidities ( @xmath0 3 , see fig . [ fig:1 ] left ) which allows them to reconstruct jets in a kinematic range of interest for various higgs and qcd physics studies in @xmath1-@xmath1 collisions at tev energies .
+ on the one hand , having the possibility to reconstruct jets beyond @xmath2 3 in atlas and cms is crucial to signal higgs boson production in vector - boson - fusion ( vbf ) processes , @xmath3 ( with @xmath4 ) , where the valence quarks @xmath5 from each proton fragment into jets in the forward and backward hemispheres @xcite .
the presence of such low - angle jets is instrumental to significantly reduce the qcd backgrounds in various vbf higgs discovery channels at the lhc , particularly for low @xmath6 masses @xcite . in the case of lhcb , given the excellent secondary vertex capabilities of the detector , forward jet studies have focused on the reconstruction of @xmath7-jets aiming at the @xmath8 decay channel in higgs production associated with vector bosons ( about one third of the cross section falls within the lhcb acceptance ) @xcite .
+ on the other hand , forward jet production is in its own right an interesting perturbative qcd ( pqcd ) process whose study yields important information on the underlying parton structure and its dynamical evolution in the proton .
in particular , it provides valuable information on the gluon density @xmath9 in a regime of low momentum fraction , @xmath10 = @xmath11 10@xmath12 , where standard deep - inelastic @xmath13-@xmath1 data can only indirectly constrain its value @xcite , and where its evolution is expected to be affected by non - linear qcd dynamics @xcite .
indeed , in @xmath1-@xmath1 collisions , the _ minimum _ parton momentum fractions probed in each proton in a @xmath14 process with a jet of momentum @xmath15 produced at pseudo - rapidity @xmath16 are @xmath17 i.e. @xmath18 decreases by a factor of @xmath1910 every 2 units of rapidity .
the extra @xmath20 lever - arm motivates the interest of _ forward _ jet production measurements to study the pdfs at small values of @xmath10 . from eq .
( [ eq : x2_min ] ) , it follows that the measurement at the lhc of jets with transverse momentum @xmath21
= 20 gev / c at rapidities @xmath22 5 allows one to probe @xmath10 values as low as @xmath23 in partonic collisions with highly asymmetric longitudinal momenta in the initial - state ( fig .
[ fig:1 ] , right ) . + @xmath16 plane ( forward detectors are beyond the dashed vertical line ) @xcite .
right : log(@xmath24 ) distribution of two partons producing at least one jet above @xmath15 = 20 gev / c at forward rapidities in @xmath1-@xmath1 collisions at @xmath25 = 14 tev @xcite .
[ fig:1],title="fig:",scaledwidth=48.0% ] @xmath16 plane ( forward detectors are beyond the dashed vertical line ) @xcite .
right : log(@xmath24 ) distribution of two partons producing at least one jet above @xmath15 = 20 gev / c at forward rapidities in @xmath1-@xmath1 collisions at @xmath25 = 14 tev @xcite .
[ fig:1],title="fig:",scaledwidth=49.0%,height=230 ] in this contribution we summarise first the forward jet reconstruction capabilities of the three lhc experiments ( section [ sec : exp ] ) and , then , in sections [ sec : spec ] and [ sec : mn ] we present simulation studies of two cms forward - jet measurements @xcite : 1 . single inclusive jet cross section at moderate transverse momenta ( @xmath26 20 120 gev / c ) , 2 . azimuthal ( de)correlations of `` mueller - navelet '' @xcite dijet events , characterised by jets with similar @xmath15 separated by a large rapidity interval ( @xmath27 6 10 ) , which are sensitive , respectively , to the small-@xmath28 ( and high-@xmath29 ) proton pdfs , as well as to low-@xmath10 qcd evolution of the bfkl @xcite , ccfm @xcite and/or saturation @xcite types .
in atlas and cms , jets can be reconstructed calorimetrically at forward rapidities in the fcal @xcite and hf @xcite calorimeters 6.6 with the castor detector @xcite . ] ( 3@xmath305 ) , by means of standard jet algorithms of the cone or sequential - clustering types @xcite .
the jet radii are often chosen relatively small ( e.g. @xmath31 for the cone and @xmath32 for the @xmath33 algorithms ) so as to minimise the effects of hadronic activity inside the jet due to the underlying event and beam - remnants .
figure [ fig:2 ] ( left ) shows the energy resolution for forward jets reconstructed in cms with three different algorithms ( iterative cone , siscone and @xmath33 ) @xcite .
the obtained @xmath15 relative resolutions are of @xmath34 at 20 gev / c decreasing to 10% above 100 gev / c .
similar results are obtained for atlas @xcite .
we note that though the forward calorimeters have a coarser granularity than the barrel and endcap ones , the energy resolution is _
better _ in the forward direction than at central rapidities because ( i ) the _ total _ energy of the jet is boosted at forward rapidities , and ( ii ) the forward jets are more collimated and , thus , the ratio of jet - size / detector - granularity is more favourable . the position ( @xmath16 , @xmath35 ) resolutions ( not shown here ) for forward jets are also very good : @xmath36 0.045 at @xmath15 = 20 gev / c , improving to @xmath37 0.02 above 100 gev / c @xcite .
good @xmath35-@xmath16 resolutions are important when it comes to detailed studies of the azimuthal decorrelation as a function of the pseudorapidity separation in events with forward - backward dijets ( see section [ sec : mn ] )
. figure [ fig:2 ] ( right ) shows the efficiency and purity of forward jets reconstructed with the seeded cone finder ( @xmath38 ) in the atlas fcal calorimeter @xcite . above @xmath1935 gev / c ,
the efficiency saturates at around 95% with a purity below 4%
. resolutions for various jet algorithms in the cms hf calorimeter @xcite .
right : forward jet reconstruction efficiency as function of the true jet @xmath15 , and fake reconstruction rate versus the reconstructed jet @xmath15 in the atlas fcal calorimeter @xcite .
[ fig:2],title="fig:",scaledwidth=40.0%,height=219 ] resolutions for various jet algorithms in the cms hf calorimeter @xcite .
right : forward jet reconstruction efficiency as function of the true jet @xmath15 , and fake reconstruction rate versus the reconstructed jet @xmath15 in the atlas fcal calorimeter @xcite .
[ fig:2],title="fig:",scaledwidth=60.0%,height=210 ] in lhcb , jet reconstruction has focused on @xmath7-jets given the excellent vertexing capabilities of the detector .
the physics motivation is so far centered on the measurement of the @xmath8 channel for intermediate - mass higgs production associated with vector bosons ( 30% of such a signal falls within the lhcb acceptance ) @xcite .
both seeded - cone- and @xmath33-algorithms have been tested including information from the calorimeters and the tracking devices in a `` particle flow '' type of approach .
a neural - network is trained to identify @xmath7-jets and optimise the jet energy reconstruction .
the main current limitation of the measure is the saturation of the calorimeters ( designed originally mostly for single particle triggering / measurements at low and moderate @xmath15 s ) for jets with total energy beyond @xmath39 1.5 tev ( i.e. @xmath40 20 150 gev / c for @xmath22 3
5 ) given the very strong @xmath41 total - momentum boost at large rapidities .
figure [ fig:3 ] ( left ) shows the forward jet @xmath15 spectrum generated with pythia and reconstructed in the cms hf calorimeter with the siscone finder @xcite .
the spectrum is compared to fastnlo jet predictions @xcite with the mrst03 and cteq6.1 m pdfs .
the right plot shows the percent differences between the reconstructed spectrum and the two theoretical predictions .
the single jet spectra obtained for different pdfs are similar at high @xmath21 , while differences as large as @xmath42 appear below @xmath1960 gev / c .
the error bars include the statistical ( a total integrated luminosity of 1 pb@xmath43 is assumed ) and the energy - resolution smearing errors . the thin violet band around zero is the pdf uncertainty from the cteq6.1 m set alone .
the main source of systematic uncertainty is due to the calibration of the jet energy - scale ( jes ) .
assuming a conservative 5
10% jes error , one finds propagated uncertainties of the order 30
40% in the jet yields at @xmath44 35 60 gev / c ( yellow band ) which are similar to the theoretical uncertainty associated to the pdf choice .
if the jes can be improved at the 5% level or below , and the pdf uncertainties are indeed as large as the differences between mrst03 and cteq6 m , a forward jet measurement could help constrain the underlying pdf in global - fit analyses .
-@xmath1 at 14 tev compared to fastnlo predictions with mrst03 and cteq6.1 m pdfs .
right : percent differences between the forward jet @xmath15 spectrum and the two fastnlo predictions .
the yellow band shows the propagated yield uncertainty for a 5 10% jet - energy scale ( jes ) error.,title="fig:",scaledwidth=49.0%,height=230 ] -@xmath1 at 14 tev compared to fastnlo predictions with mrst03 and cteq6.1 m pdfs .
right : percent differences between the forward jet @xmath15 spectrum and the two fastnlo predictions .
the yellow band shows the propagated yield uncertainty for a 5
10% jet - energy scale ( jes ) error.,title="fig:",scaledwidth=50.0% ]
the interest in forward jet measurements goes beyond the _ single _ inclusive cross sections : the production of _ dijets _ with similar @xmath15 but separated by large rapidities , the so - called `` mueller - navelet jets '' @xcite , is a particularly sensitive measure of non - dglap qcd evolutions . the large rapidity interval between the jets ( e.g. up to @xmath45 12 in the extremes of cms forward calorimeters ) enhances large logarithms of the type @xmath46 which can be appropriately resummed within the bfkl @xcite , ccfm @xcite and/or saturation @xcite frameworks .
one of the phenomenological implications of this type of dynamics is an enhanced radiation between the two jets which results in a larger azimuthal decorrelation for increasing @xmath47 separations compared to collinear pqcd approaches .
cms @xcite has carried out an analysis with pythia @xcite and herwig @xcite selecting events with forward jets ( icone , @xmath48 0.5 ) which satisfy the following mueller - navelet ( mn ) type cuts : * @xmath49 gev / c ( good parton - jet matching and good jet trigger efficiencies in hf ) * @xmath50 gev / c ( similar @xmath21 to minimise dglap evolution ) * @xmath51 ( both jets in hf ) * @xmath52 ( each jet in a different hf , i.e. their separation is @xmath53 6 ) the data passing the mn - cuts are divided into 4 equidistant pseudorapidity bins with separations @xmath47=6.5 , 7.5 , 8.5 and 9.5 and the dijet cross section computed as @xmath54 , where @xmath55 and @xmath56 is the observed number of jets in the @xmath57 bin . for 1 pb@xmath43 , one expects a few 1000s ( 100s ) mn jets with separations @xmath58 6 ( 9 ) . figure [ fig : muller_navelet_jets ] , left , shows the expected pythia yields passing the mn cuts for @xmath27 7.5 .
the obtained dijet sample appears large enough to carry out detailed studies of the @xmath47 dependence of the yields , and look e.g. for a possible `` geometric scaling '' behaviour in the mueller - navelet yields @xcite .
+ ) for @xmath45 7.5 .
right : average of @xmath59 versus @xmath60 in pythia- , herwig - generated events compared to bfkl ( yellow band ) @xcite analytical estimates .
[ fig : muller_navelet_jets],title="fig:",scaledwidth=49.0%,height=215 ] ) for @xmath45 7.5 .
right : average of @xmath59 versus @xmath60 in pythia- , herwig - generated events compared to bfkl ( yellow band ) @xcite analytical estimates .
[ fig : muller_navelet_jets],title="fig:",scaledwidth=50.0%,height=215 ] an enhanced azimuthal decorrelation for increasing rapidity separation , measured e.g. by the average value ( over events ) of the cosine of the @xmath61 difference between the mn jets @xmath62 versus the @xmath47 between them , is the classical `` smoking - gun '' of bfkl radiation @xcite .
one expects @xmath63 = 1 ( 0 ) for perfect ( de)correlation between the two jets .
the results are shown in fig .
[ fig : muller_navelet_jets ] ( right ) for the two highest-@xmath15 jets in the event passing the mn cuts .
only the dominant ( statistical ) errors are presented . at the monte carlo truth level ( not shown here ) , the originating partons in pythia or herwig are almost exactly back - to - back for all @xmath47 in each such jet - pair events . yet , at the _ generator - level _ , the @xmath63 decorrelation increases to 15% ( 25% ) for pythia ( herwig ) , @xmath64 0.85 ( 0.75 ) , due to parton showering and hadronization effects . yet
, the forward dijet decorrelation observed in both mcs is smaller ( and less steep as a function of @xmath47 ) than found in bfkl approaches ( yellow band ) @xcite .
we have summarised the forward jet reconstruction capabilities of the atlas , cms and lhcb experiments .
the measurement of forward jets opens up the possibility to carry out interesting studies in the higgs ( tagging of vector - boson - fusion or higgs - to-@xmath65 in associated @xmath66 production ) and qcd ( low-@xmath10 parton densities and dynamics ) sectors of the standard model .
the author thanks salim cerci for his valuable collaboration with the cms forward jets analysis , as well as marco musy for discussions on lhcb jet reconstruction . support by the 7th eu framework programme ( contract fp7-erg-2008 - 235071 ) is acknowledged .
a. morsch [ alice collaboration ] , aip conf .
proc . * 1026 * , 72 ( 2008 ) .
a. djouadi , phys .
* 457 * ( 2008 ) 1 .
a. de roeck ( ed . ) _ et al .
_ [ cms collaboration ] , j. phys .
g * 34 * , 995 ( 2007).s .
asai _ et al . _ [ atlas collaboration ] , eur .
j. c * 32s2 * , 19 ( 2004).a . bay and c. potterat [ lhcb collaboration ] , lhcb note 2009 - 023 , lhcb note 2009 - 024 .
d. denterria , eur .
phys . j. a * 31 * , 816 ( 2007).see e.g. f. gelis , t. lappi and r. venugopalan , _ int .
j. mod .
e _ * 16 * , 2595 ( 2007 ) .
d. denterria , proceeds dis 2007 , pp .
1141 ; arxiv:0708.0551 [ hep - ex ] .
s. cerci and d. denterria [ cms collaboration ] , aip conf . proc .
* 1105 * ( 2009 ) 28 ; cms pas fwd-08 - 001 ( cms - an/2008 - 060 ) .
a. h. mueller and h. navelet , _ nucl .
b _ * 282 * , 727 ( 1987 ) . l.n .
lipatov , _ sov .
j. nucl .
* 23 * , 338 ( 1976 ) ; e.a .
kuraev , l.n .
lipatov and v.s .
fadin , _ zh . eksp .
. fiz _ * 72 * , 3 ( 1977 ) ; ya.ya .
balitsky , l.n .
lipatov , _ sov .
j. nucl .
_ * 28 * , 822 ( 1978).h .
jung and g. p. salam , eur .
j. c * 19 * ( 2001 ) 351 ; m. deak , f. hautmann , h. jung and k. kutak , jhep * 0909 * ( 2009 ) 121 .
a. artamonov _ et al .
_ , jinst * 3 * ( 2008 ) p02010 .
a. s. ayan _ et al .
_ , _ j. phys .
g _ * 30 * , n33 ( 2004 ) .
x. aslanoglou _ et al .
_ , _ eur .
j. c _ * 52 * , 495 ( 2007 ) .
c. buttar _
_ , arxiv:0803.0678 [ hep - ph ]
. g. aad _ et al .
_ [ the atlas collaboration ] , arxiv:0901.0512 .
t. kluge , k. rabbertz and m. wobisch , arxiv : hep - ph/0609285 ; and k. rabbertz , private comm .
a. sabio vera , f. schwennsen , _ nucl .
b _ * 776 * , 170 ( 2007 ) ; and private communication . c. marquet and c. royon , _ nucl
. phys .
b _ * 739 * , 131 ( 2006 ) ; and arxiv:0704.3409 [ hep - ph ] .
e. iancu , m. s. kugeratski and d. n. triantafyllopoulos , nucl .
a * 808 * , 95 ( 2008 ) . | the capabilities of the atlas , cms and lhcb detectors to reconstruct jets at forward rapidities ( @xmath0 3 ) in @xmath1-@xmath1 collisions at the cern large hadron collider are reviewed .
the qcd and higgs physics motivations for such measurements are summarised .
details are given on studies that provide information on the parton structure and evolution at small values of fractional momenta in the proton . |
matrix elements involving moving @xmath0 and @xmath3 mesons are important for studies of @xmath4 and @xmath5 decay .
it is necessary to understand the systematic errors in lattice qcd that come from the presence of non - zero momenta and the optimal way in which to handle such mesons on the lattice .
an easy place to study these effects is in the determination of the @xmath0 meson decay constant , @xmath1 @xcite . in the absence of discretisation errors
it should be true that @xmath6 we use both @xmath7 and @xmath8 with all current corrections and renormalisation through @xmath9 @xcite .
results shown are for an ensemble of 278 @xmath10 lattices at @xmath11 = 5.7 .
clover light quarks are used with @xmath12 and nrqcd heavy quarks are used with masses in lattice units of 2 and 8 ( @xmath13 = 4.0 ) .
all momenta are considered up to @xmath14 of 16 in units of @xmath15 , i.e. roughly 4 @xmath16 at this lattice spacing .
we generated heavy and light quark propagators with 3 different smearings : a delta function , a narrow gaussian ( width 1 ) and a broad gaussian ( width 3 ) .
we combined these smearings together to make 6 different smearings for the heavy - light meson and analysed the resulting smeared - smeared meson correlators using the constrained curve fitting methods described in @xcite .
the results are shown in figure 1 , which plots the amplitude of the ground state @xmath0 meson as a function of squared momentum for some of these smearings .
it is clear that the optimal smearing for a moving heavy - light meson is one in which the smearing on the heavy quark is a narrow one and the smearing on the light quark is a broad one .
this is not surprising if one considers that the heavy quark is carrying almost all of the meson momentum .
the overlap as a function of momentum is maximized if the heavy quark has a delta function smearing , but in that case the statistical noise is large because of the well - known problem that an unsmeared heavy quark has a poor signal / noise ratio @xcite . at zero momentum
this has led to the received wisdom that the heavy quark should have a broad smearing - figure 1 shows that this is not correct at non - zero momentum .
figure 2 amplifies this point by showing the rapid plateau and consequent small error on the kinetic energy possible with a good finite - momentum smearing , in contrast to that obtained with a smearing that might have been considered a good one at zero momentum .
we used smearing ` 5 ' from the above analysis and the constrained curve fitting methods to determine @xmath1 for moving @xmath0 mesons and the temporal axial current .
matrix elements for @xmath7 are obtained from a simultaneous fit to _ all _ @xmath17 ( i=0,1,2 ) and @xmath18 ( i=0,1,2,3,4 ) current correlators at a given momentum .
the construction of the continuum @xmath7 from @xmath19 is described in @xcite .
figure 3 shows results for the ratio : @xmath20 for @xmath21=8 using @xmath22 in the renormalisation .
the expected curve @xmath23 is also shown .
no disagreement is seen until the highest momentum . at @xmath21 = 2
, larger discrepancies appear @xcite .
the matrix elements for the spatial axial current behave rather differently to those for the temporal axial current since they must vanish when there is no component of @xmath24 along the direction of the current .
even the leading order current , @xmath25 = @xmath26 then has a matrix element @xmath27@xmath28 with respect to @xmath29 .
this is of the same order as the ` @xmath30 ' suppressed current contributions from @xmath31 = @xmath32 and @xmath33 = @xmath34 .
@xmath31 contributes at tree level and @xmath33 at one - loop to the final result for @xmath8 . to understand the size of different current contributions it is important to use a power - counting in both the external velocity and in @xmath35 @xcite .
figure 4 shows results for the ratio : @xmath36 as a function of @xmath37 for a range of values of @xmath38 , for @xmath21=8 .
the line shown is @xmath39 ( variation with @xmath38 of this line is not significant ) .
good agreement with the line is found for @xmath40 = 1,2 ( @xmath41 ) at all @xmath38 but @xmath40 = 3,4 show signs of deviation , presumably from missing current corrections that are higher order in the external velocity .
figure 4 shows agreement between @xmath1 from @xmath7 and @xmath8 which resolves a long - standing problem .
@xcite found a disagreement of @xmath27(10% ) when comparing the tree - level matrix elements for clover fermions .
this ignores the contribution of @xmath33 and does not allow for the clover equivalent of the different renormalisation of @xmath42 and @xmath43 .
both these effects are included here .
we have demonstrated the usefulness of optimal ( i.e. narrow ) smearing for heavy quarks in moving heavy - light mesons , and of constrained curve fitting in extracting results out to much higher momenta than have previously been attempted .
good agreement is found for @xmath1 at zero and non - zero momentum . for the temporal axial current at large @xmath21 this holds out to the highest momenta studied . to understand the behaviour of the currents at non - zero momentum it is important to use a power - counting appropriate to moving heavy - light mesons . * acknowledgements * sc , gc
, cd and jh are members of the ukqcd collaboration .
we are grateful to the following bodies for support of this work : pparc , doe , nsf and the eu under hprn-2000 - 00145 hadrons / lattice qcd .
9 j. simone , nucl .
. suppl . * 53 * ) 1997 386 .
c. morningstar and j. shigemitsu , prd*59 * ( 1999 ) 094504 .
g. p. lepage , these proceedings .
g. p. lepage , npb ( proc .
* 26 * ) 45 1992 .
s. collins _ et al _ , in preparation . | we show results for the @xmath0 meson decay constant calculated both for @xmath0 mesons at rest and those with non - zero momentum and using both the temporal and spatial components of the axial vector current .
it is an important check of lattice systematic errors that all these determinations of @xmath1 should agree .
we also describe how well different smearings for the @xmath0 meson work at non - zero momentum - the optimal smearing has a narrow smearing on the @xmath2 quark . |
the origin of non - fermi liquid ( nfl ) behavior in heavy - fermion ( hf ) systems has been studied intensively in the past decade but is still unclear up to now @xcite .
in particular two different scenarios are discussed for the quantum - critical point ( qcp ) , where long - range antiferromagnetic ( af ) order emerges from the hf state ; one in which nfl behavior arises from bragg diffraction of the electrons off a critical spin - density wave ( sdw ) @xcite , the other in which the bound - state structure of the composite heavy fermions breaks up at the qcp resulting in a collapse of the effective fermi temperature @xcite . in the sdw scenario , assuming threedimensional ( 3d ) spinfluctuations , singular scattering occurs only along certain `` hot lines '' connected by the vector * q * of the near af order while the remaining fermi surface still behaves as a fermi liquid .
therefore , the low - temperature ( @xmath11 ) specific heat coefficient @xmath20 , that measures the heavy quasiparticle ( qp ) mass , is expected to show an anomalous temperature dependence @xmath31 , but remains finite at the qcp @xcite .
a diverging qp mass , as evident from the @xmath32 behavior found e.g. in the prototypical system cecu@xmath33au@xmath34 for @xmath35 @xcite , would arise only if truly 2d critical spinfluctuations render the entire fermi surface `` hot '' @xcite .
on the other hand , measurements of the inelastic neutron scattering on cecu@xmath16au@xmath17 @xcite showed that the critical component of the spin fluctuations is almost momentum independent leading to the proposal of the locally critical scenario @xcite . since @xmath11-dependent measurements at the qcp alone
provide no information on how the heavy quasiparticles decay into the quantum critical state it is necessary to tune the system away from the magnetic instability in the landau fermi liquid ( lfl ) state and to follow the qp properties upon approaching the qcp . in this paper
we demonstrate that magnetic fields can be used for this purpose .
+ in the following , we consider the hf metal ybrh@xmath4si@xmath4 for which pronounced nfl phenomena , i.e. , a logarithmic divergence of @xmath20 and a quasi - linear @xmath11-dependence of the electrical resistivity below 10 k , have been observed above a low - lying af phase transition @xcite .
this system is very suitable for such an investigation , because the effect of disorder is negligible in crystals with a residual resistivity of less than 1 @xmath36 cm .
the application of pressure to ybrh@xmath4si@xmath4 increases @xmath37 @xcite as expected , because the ionic volume of the magnetic @xmath38 yb@xmath19-configuration is smaller than that of the nonmagnetic @xmath39 yb@xmath40 one .
expanding the crystal lattice by randomly substituting ge for the smaller isoelectric si atoms allows one to tune ybrh@xmath4(si@xmath5ge@xmath6)@xmath4 through the qcp without affecting its electronic properties and without introducing significant disorder to the lattice @xcite . in ybrh@xmath4(si@xmath5ge@xmath34)@xmath4 with the nominal ge concentration @xmath9 ( see below ) the nfl behavior extends to the lowest accessible temperatures , in particular @xmath41 is observed from above 10 k down to 10 mk @xcite . +
this article is organized as follows : after giving details concerning experimental techniques in the next section , we address in section 3 the pronounced nfl behavior at zero magnetic field in thermodynamic , magnetic and transport properties at the qcp in ybrh@xmath4(si@xmath42ge@xmath43)@xmath4 . in section 4
we concentrate on the undoped system and prove that magnetic fields can be used to tune the system continuously from the af ordered state through the qcp into a field - induced lfl state .
we present evidence for the divergence of the heavy quasiparticles at the qcp and end with the conclusions in section 5 .
single crystalline platelets of ybrh@xmath4(si@xmath5ge@xmath6)@xmath4 prepared with the nominal ge compositions @xmath7 and 0.05 were grown from in flux as described earlier @xcite .
a microprobe analysis ( ma ) of the @xmath9 sample revealed a ge concentration @xmath44 , only .
the solubility of ge atoms in in flux is much higher compared to that of si atoms .
this causes the actual ge concentration being smaller than 0.05 in these single crystals . to be consistent with previous publications @xcite ,
the ge - doped single crystals studied in this paper are labeled by their nominal ge composition @xmath9 . to demonstrate that the main effect of ge doping is primarily the expansion of the volume ,
measurements of the electrical resistivity under hydrostatic pressure have been performed on a @xmath9 single crystal @xcite . at a pressure of 0.63 gpa
the onset of af order has been observed at @xmath45 k. furthermore , the @xmath37 vs @xmath46 phase diagram of the ge - doped crystal matches perfectly with that found for pure ybrh@xmath4si@xmath4 @xcite if the pressure axis is shifted by @xmath47 gpa @xcite .
taking the value @xmath48 gpa for the bulk modulus as determined from recent moessbauer measurements under hydrostatic pressure @xcite , this pressure shift corresponds to a volume expansion of 0.1% for the ge - doped sample consistent with a ge content of @xmath49 .
+ the new generation of @xmath7 crystals show a residual resistivity @xmath50 cm , i.e. , twice as low as @xmath51 of the previous ones .
whereas for the latter the phase transition at @xmath37 discovered by ac - susceptibility measurements could not be resolved in the resistivity @xcite , the new crystals show a clear kink of @xmath52 at @xmath37 , see below .
+ for all low - temperature measurements , @xmath53he/@xmath54he dilution refrigerators were used .
the specific heat was determined with the aid of a quasi - adiabatic heat pulse technique .
the electrical resistivity , @xmath3 , and the magnetic ac - susceptibility , @xmath55 , were measured utilizing a linear research co. ( lr700 ) bridge at 16.67 hz .
amplitudes of 0.1 ma and 1 oe for the current and magnetic field , respectively , were chosen to determine @xmath3 and @xmath55 .
absolute values of @xmath2 have been determined from a comparison in the temperature range 2 k @xmath12 6 k with the results of the dc - susceptibility in 50 mt using a quantum design squid magnetometer .
low temperature magnetization measurements were performed utilizing a high - resolution capacitive faraday magnetometer as described in @xcite .
as @xmath56 vs @xmath57 ( a ) and the electrical resistivity @xmath3 ( b and c ) of high - quality ybrh@xmath4(si@xmath5ge@xmath6)@xmath4 single crystals with ge - contents @xmath7 and @xmath9 .
@xmath58 is the nuclear quadrupole contribution calculated from recent mssbauer results @xcite .
dotted line in ( a ) marks @xmath59 dependence with @xmath60 k. resistivity data shown in ( c ) were obtained for the current * j * flowing along ( @xmath61 ) and perpendicular ( @xmath62 ) to the @xmath22-axis .
the arrow indicates @xmath37 as obtained from the maximum in @xmath63.,scaledwidth=80.0% ] -direction , as @xmath64 vs @xmath11 for ybrh@xmath4(si@xmath42ge@xmath43)@xmath4 .
the dotted line indicates curie - weiss type dependence @xmath65 implying large fluctuating moments of 1.4 @xmath66-ion and a weiss temperature of @xmath67 k. the solid line represents @xmath68 with @xmath69 ( see text ) .
the inset shows @xmath55 vs @xmath11 on a logarithmic scale for ybrh@xmath4(si@xmath5ge@xmath6)@xmath4 for @xmath7 and 0.05 .
the arrow indicates @xmath37 for the undoped system.,scaledwidth=70.0% ] in undoped ybrh@xmath4si@xmath4 we observe a second order phase transition in specific heat ( fig .
1a ) which , according to ac - susceptibility measurements ( inset of fig .
2 ) , marks the onset of af order @xcite .
the resistivity follows a quasi - linear @xmath11-dependence down to about 80 mk , below which a sharp decrease , independent of the current direction , is observed ( inset of fig .
thus the resistivity does not show any signatures of the formation of a sdw for which an increase of @xmath52 along the direction of the sdw modulation , indicating the partial gapping of the fermi surface , should be expected .
the absence of this behavior favors the interpretation of local - moment type of magnetic order in ybrh@xmath4si@xmath4 .
the resistivity in the af ordered state ( at @xmath70 ) is best described by @xmath71 with a huge coefficient , @xmath72 cm / k@xmath73 , for 20 mk @xmath74 mk . extrapolating @xmath75 as @xmath76 to @xmath77 j / k@xmath73mol reveals an entropy gain at the af phase transition of only about @xmath78 .
this provides evidence for the weakness of the af order in ybrh@xmath4si@xmath4 .
the ratio of @xmath79 in the ordered state is close to that expexted for a lfl @xcite , i.e. , one with very heavy quasiparticle masses .
+ the af ordering is suppressed to below 10 mk by doping with a small amount of ge in ybrh@xmath4(si@xmath42ge@xmath43)@xmath4 .
this system must be located very close to the qcp , and pronounced nfl behavior is observed down to lowest @xmath11 : the electronic specific heat @xmath56 follows a @xmath59 ( @xmath60 k @xcite ) behavior between 0.3 and 10 k. in the same @xmath11-range a linear @xmath11-dependence of the elecritical resistivity is observed . thus above 0.3 k (
e.g. 0.0125 @xmath80 ) resembling @xmath11-dependences as reported for cecu@xmath16au@xmath17 are observed that could only be explained within the sdw scenario assuming truly 2d critical spinfluctuations @xcite .
below that temperature , which corresponds to 75 mk in cecu@xmath16au@xmath17 ( @xmath81 k ) , a cross - over into a different regime takes place .
whereas the resistivity continuous to follow a linear @xmath11-dependence down to below 15 mk , a pronounced upturn is observed in @xmath56 whose origin will be discussed in the conclusion . a cross - over around 0.3 k
is observed in the susceptibility , too : below 0.3 k , @xmath2 follows a curie - weiss type behavior ( fig .
2 ) implying fluctuating moments of the order of @xmath82 yb@xmath19-ion .
the weiss - temperature of @xmath83 k suggests strong af correlations in this regime . above 0.3 k
, the susceptibility can be described by @xmath84 with the exponent @xmath85 decreasing continuously from 1 ( below 0.3 k ) to 0.5 at 2 k. as displayed in fig .
2 , in the @xmath11 interval 0.2 k @xmath12 1.4 k the susceptibility roughly follows @xmath84 with @xmath69 and @xmath86 k , i.e. a `` non curie - weiss '' behavior reminiscent of cecu@xmath16au@xmath17 @xcite .
at varying temperatures in magnetic fields applied along and perpendicular to the @xmath22-axis .
arrows indicate critical fields @xmath27.,scaledwidth=65.0% ] as discussed in the previous section , the weak af order in ybrh@xmath4si@xmath4 can be suppressed by a tiny volume expansion upon replacing si by larger ge - atoms , and pronounced deviations from lfl behavior occur at the qcp in ybrh@xmath4(si@xmath42ge@xmath43)@xmath4 . to study
how the heavy quasiparticles decay into the quantum critical state it is required to investigate their properties upon crossing the qcp at zero temperature as a function of a control parameter that can be varied continuously . in the following we will show that the application of magnetic fields to ybrh@xmath4si@xmath4 can be used for this purpose .
this study was motivated by the observation of field - induced nfl behavior in the doped af systems cecu@xmath33ag@xmath6 @xcite and ybcu@xmath87al@xmath6 @xcite .
ybrh@xmath4si@xmath4 is highly suited to study the properties of a @xmath88-induced qcp , because the influence of disorder in this clean stoichiometric system should be negligible .
furthermore , the ordering temperature @xmath89 mk is the lowest among all undoped hf systems at ambient pressure and already a very small critical magnetic field @xmath90 is sufficient to push @xmath37 towards zero temperature .
+ we first discuss the low - temperature magnetization which proves that the af phase transition as a function of field is a continuous one .
ybrh@xmath4si@xmath4 exhibits a highly anisotropic magnetic response , indicating that yb@xmath19 moments are forming an `` easy - plane '' square lattice perpendicular to the crystallographic c - direction @xcite .
the isothermal magnetization ( fig .
3 ) shows a strongly nonlinear response for fields @xmath91 . for @xmath92 a clear reduction in slope
is observed above 0.06 t which indicates the suppression of af order resulting in a weakly polarized state .
a smooth extrapolation of @xmath93 for @xmath94 t towards zero field reveals a value of @xmath95 for the staggered magnetization in the af state , indicating that the size of the ordered moments is much smaller than that of the effective moments observed in the paramagnetic state above @xmath37 .
thus a large fraction of the local moments appears to remain ( quantum ) fluctuating within the easy plane in the af ordered state .
their continuous polarization for fields exceeding @xmath27 gives rise to a strong curvature in @xmath93 for @xmath91 .
for fields applied along the magnetic hard direction , @xmath96 , the magnetization shows an almost linear behavior ( fig .
1b ) which was found to extend at least up to 58 t @xcite . at @xmath92
a very tiny increase in the @xmath93 slope is observed below about 0.7 t which , according to the resistivity measurements discussed below , represents the critical field @xmath27 for @xmath96 .
si@xmath4 at varying magnetic fields applied perpendicular ( a ) and along the @xmath22-axis ( b ) . for clarity the different curves in @xmath97
were shifted subsequently by 0.1 @xmath36 cm .
up- and downraising arrows indicate @xmath37 - and upper limit of @xmath98 behavior , respectively .
dotted and solid lines represent @xmath99 with @xmath100 and @xmath101 , respectively.,scaledwidth=90.0% ] phase diagram for ybrh@xmath4si@xmath4 with @xmath37 as derived from @xmath63 vs @xmath11 and @xmath102 , the upper limit of the @xmath103 behavior , as a function of magnetic field , applied both along and perpendicular to the @xmath22-axis . for the latter ones the @xmath88-values
have been multiplied by a factor 11 .
lines separating the antiferromagnetic ( af ) , non - fermi liquid ( nfl ) and landau fermi liquid ( lfl ) phase are guides to the eye.,scaledwidth=70.0% ] in fig .
4 we show the evolution of the low - temperature resistivity upon applying magnetic fields along and perpendicular to the easy magnetic plane . at small magnetic fields
the nel temperature , determined from the maximum value of @xmath63 , shifts to lower temperatures and vanishes at a critical magnetic field @xmath27 of 0.06 t in the easy magnetic plane and of 0.66 t along the magnetic hard direction , i.e. the crystallographic @xmath22-axis . at @xmath104 ,
the resistivity follows a linear @xmath11-dependence down to the lowest accessible temperature of about 20 mk .
this observation provides striking evidence for field - induced nfl behavior at the critical magnetic fields applied along both crystallographic directions @xcite . at @xmath23
, we find @xmath105 for @xmath106 , with @xmath107 increasing and @xmath29 decreasing upon raising the applied magnetic field .
the evolution of @xmath37 and @xmath102 as a function of @xmath88 is shown in fig .
5 . the extremely low value of the critical field applied along the easy plane highlights the near degeneracy of two different heavy lfl states , one being weakly af ordered ( @xmath108 ) and the other one being weakly polarized ( @xmath23 ) .
vs @xmath11 ( on a logarithmic scale ) for ybrh@xmath4si@xmath4 at varying fields applied parallel to the @xmath22-axis .
@xmath109 is the nuclear quadrupole contribution calculated from recent moessbauer results@xcite.,scaledwidth=75.0% ] next we address specific - heat measurements ( fig .
6 ) which have been performed at magnetic fields along the @xmath22-axis . due to the strong magnetic anisotropy
, the sample plate used for the specific - heat measurement could not be aligned perfectly along the hard magnetic direction .
therefore , a critical field @xmath27 of only about 0.3 t was sufficient to suppress af order completely in these experiments .
as shown in fig .
6 , at @xmath104 the specific - heat coefficient @xmath110 increases down to the lowest @xmath11 , indicative of a field - induced nfl ground state . within 40
mk @xmath111 mk it follows a steep increase similar to the upturn observed for the @xmath9 single crystal below 0.3 k ( c.f .
while this anomalous contribution is strongly reduced upon increasing @xmath88 , at magnetic fields @xmath112 1 t , the nuclear contribution becomes visible at the lowest temperatures , above which a constant @xmath113 value is observed within a limited temperature window ( fig .
@xmath113 decreases in magnitude upon increasing the field . +
a very similar behavior has been found in the field dependence of the low temperature ac susceptibility , too @xcite . for magnetic fields exceeding @xmath27 constant values
@xmath114 have been observed which decrease in size upon increasing @xmath88 .
vs field @xmath88 .
data for @xmath88 perpendicular to the @xmath22-direction have been multiplied by 11 .
dashed line marks @xmath27 , solid line represents @xmath115 .
inset shows double-@xmath116 plot of @xmath117 vs @xmath118 and @xmath117 vs @xmath119 for different magnetic fields .
solid lines represent @xmath120cm(kmol / mj)@xmath73 and @xmath121cmk@xmath122/(m@xmath53/mol)@xmath73.,scaledwidth=90.0% ] in fig .
7 we present our analysis of the magnetic - field dependence of the coefficients @xmath117 , @xmath118 and @xmath119 observed for @xmath123 in the resistivity , @xmath124 , specific heat , @xmath125 @xcite , and magnetic ac - susceptibility , @xmath126 @xcite when approaching the qcp upon reducing @xmath88 towards @xmath27 .
as shown in the inset of fig .
7 , we observe @xmath127 and thus also @xmath128 , independent of the field orientation and for all @xmath88 values exceeding @xmath27 . like in the af ordered state ( at @xmath70 ) we find that the @xmath79 ratio roughly equals that observed for many hf systems @xcite
. a very large sommerfeld - wilson ratio @xmath129 of about 14 @xmath130 indicates a strongly enhanced susceptibility in the field - aligned state of ybrh@xmath4si@xmath4 pointing to the importance of low - lying ferromagnetic ( @xmath131 ) fluctuations , as also inferred from recent @xmath132si - nmr measurements @xcite .
+ since ybrh@xmath4si@xmath4 behaves as a true lfl for @xmath133 and @xmath134 , the observed temperature dependences should hold down to @xmath135 .
the field dependence @xmath29 shown in fig .
7 measures the qp - qp scattering cross section when , by field tuning , crossing the qcp at zero temperature . a @xmath28 divergence is observed indicating that the whole fermi surface undergoes singular scattering at the @xmath88-tuned qcp .
most importantly , the relation @xmath136 observed at elevated fields , suggests that also the qp mass diverges , i.e. , as @xmath137 , upon approaching @xmath138 .
the hf metal ybrh@xmath4si@xmath4 shows a weakly antiferromagnetically polarized ground state which is suppressed to @xmath139 either by a small volume expansion in ybrh@xmath4(si@xmath42ge@xmath43)@xmath4 or by the application of a critical magnetic field @xmath27 . at the qcp
, pronounced deviations from lfl behavior are observed below 10 k in the specific heat and the electrical resistivity , in particular @xmath140 ( above 0.3 k ) and @xmath141 ( down to the lowest @xmath11 ) . according to the sdw scenario
, such @xmath11-dependences would only arise assuming the scattering of truly 2d critical spinfluctuations .
+ in order to study the decay of the heavy quasiparticles in the quantum - critical state , we have used magnetic fields to suppress the af order in ybrh@xmath4si@xmath4 . by field
tuning the system through the qcp at a critical field @xmath27 one reaches a field - aligned state which at low temperatures can be described by the lfl model . the @xmath28 divergence of @xmath142 and @xmath30 indicates a divergence of the qp mass and qp - qp scattering cross section upon approaching the qcp . in the sdw model
the parameter @xmath143 , given by the square of the inverse magnetic correlation length in the 2d spin fluid , measures the distance from the qcp , i.e. , @xmath144 .
assuming that the spin fluid renders the entire fermi surface `` hot '' , the coefficient @xmath117 diverges as @xmath145 , whereas for the specific heat coefficient @xmath118 a much weaker divergence @xmath146 is expected @xcite .
thus , this model would predict the ratio @xmath79 to diverge upon decreasing @xmath88 instead of being constant .
the constancy of the kadowaki - woods ratio when approaching the qcp rather favors the locally critical scenario @xcite . +
a stronger than logarithmic divergence of the qp mass is also evident from our @xmath70 measurements of @xmath75 for ybrh@xmath4(si@xmath42ge@xmath43)@xmath4 below 0.3 k(fig .
the upturn can not be explained assuming a nuclear contribution to the low-@xmath11 specific heat .
furthermore a very similar behavior is observed for the volume thermal expansion coefficient @xmath147 too : as found for @xmath56 , above 0.3 k @xmath148 follows a logarithmic @xmath11 dependence @xcite .
below that temperature an even stronger divergence has been observed , providing additional evidence for the electronic origin of the upturn in the specific - heat coefficient . the curie - weiss behavior , observed in @xmath149 in the same @xmath11-range , where the upturn in @xmath75 occurs , hints to large unscreened fluctuating yb@xmath19 moments persisting down to the lowest @xmath11 at the qcp .
this strongly suggests a local nature of the critical fluctuations .
+ to conclude , the qp mass diverges faster than logarithmic upon approaching the qcp in ybrh@xmath4si@xmath4 in contradiction to the sdw picture and suggests a locally critical scenario for this system .
stewart , rev .
phys . * 73 * , 797 ( 2001 ) .
p. coleman et al . , j. phys
* 13 * , r723 ( 2001 ) .
hertz , phs . rev .
* b 14 * , 1165 ( 1976 ) .
a.j . millis , phys .
rev.*b 48 * , 7183 ( 1993 ) .
t. moriya and t. takimoto , j. phys .
. jpn . * 64 * , 960 ( 1995 ) .
q. si et al . ,
nature * 413 * , 804 ( 2001 ) .
lhneysen , j. phys .
* 8 * , 9689 ( 1996 ) .
a. rosch et al . , phys .
lett . * 79 * , 159 ( 1997 ) .
a. schrder et al . ,
nature * 407 * , 351 ( 2000 ) .
o. trovarelli et al . , phys .
lett . * 85 * , 626 ( 2000 ) .
o. trovarelli et al . , physica * b 312 - 313 * , 401 ( 2002 ) .
j. custers et al .
phys . pol .
* b 32 * , 3221 ( 2001 ) .
s. mederle at al . , in preparation .
m. abd - elmeguid et al .
, submitted .
t. sakakibara et al .
* 33 * , 5067 ( 1994 ) .
k. kadowaki and s.b .
woods , sol .
* 58 * , 507 ( 1986 ) .
k. heuser et al .
, phys . rev . *
b 57 * , r4198 ( 1998 ) .
k. heuser et al . , phys . rev . *
b 58 * , r15959 ( 1998 ) .
s. seuring et al .
, physica * b 281 & 282 * , 374 ( 2000 ) .
p. gegenwart et al . , phys .
lett . * 89 * , 056402 ( 2002 ) .
k. ishida et al . , to be published .
n. oeschler , unpublished results .
i. paul and g. kotliar , phys .
b 64 * , 184414 ( 2001 ) .
p. coleman , cond - mat/0206003 . | we report low temperature specific heat , @xmath0 , magnetization , @xmath1 , susceptibility , @xmath2 , and electrical resistivity , @xmath3 , measurements on high - quality single crystals of the heavy - fermion system ybrh@xmath4(si@xmath5ge@xmath6)@xmath4 ( @xmath7 and 0.05 ) .
the undoped compound shows weak antiferromagnetic ( af ) order at @xmath8 mk which is suppressed to below 10 mk by a tiny volume expansion in the @xmath9 system . in the latter pronounced deviations from landau fermi liquid ( lfl ) behavior
occur , e.g. @xmath10 over three decades in @xmath11 .
both thermodynamic and magnetic properties show a crossover at about 0.3 k : at 0.3 k @xmath12 10 k we observe @xmath13 and a `` non - curie '' behavior @xmath14 with @xmath15 similar to what was found for the prototypical system cecu@xmath16au@xmath17 . below 0.3 k
, @xmath2 turns into a curie - weiss dependence @xmath18 indicating large unscreened yb@xmath19 moments whereas in @xmath20 a pronounced upturn occurs .
+ in the undoped compound the af order is suppressed continuously by critical fields @xmath21 t and 0.7 t applied perpendicular and parallel to the @xmath22-axis , respectively . for @xmath23 a lfl state with @xmath24 and @xmath25 is induced , that fulfills the kadowaki - woods scaling @xmath26 . upon reducing the magnetic field to @xmath27 a @xmath28 dependence of @xmath29 and @xmath30 indicates singular scattering at the whole fermi surface and a divergence of the heavy quasiparticle mass . |
since the establishment of qcd the study of hadron states beyond the simple quark model , namely the qcd exotic states , has been a topical subject in hadron physics .
however , for a long time decisive evidence for qcd exotics has been absent from experimental measurements . until recently ,
more and more evidence in the heavy quark sector indicates possible candidates for the qcd exotics which are named as xyz " states .
very recently , the lhcb collaboration reported observation of two multiquark states , @xmath0 and @xmath1 , in the invariant mass spectrum of @xmath3 in the decay of @xmath4 @xcite .
the masses of these two states are @xmath5 mev and @xmath6 mev , respectively , and their widths are @xmath7 mev and @xmath8 mev , respectively .
their quantum numbers are determined by the partial wave analysis and the spins are @xmath9 and @xmath10 with opposite parities . since their decays into @xmath3 contain at least five quarks , they are ideal candidates for pentaquark states of @xmath11 .
this experimental result immediately initiates a lot of discussions and studies .
although there are immediate publications claiming their dynamic properties as a multiquark state @xcite , it was also investigated in refs .
@xcite that the anomalous triangle singularity ( ats ) may produce threshold enhancements to mimic the behavior of genuine states or to contribute on top of the genuine states .
it is interesting to note that since these two states were observed in their decays into @xmath3 , it is natural to expect that they can be produced in @xmath2 photoproduction via the @xmath12 and @xmath13-channel process .
this motivates us to investigate these two pentaquark candidates in @xmath2 photoproduction .
in particular , if they are genuine states , they should be created in @xmath2 photoproduction , while if they are the ats enhancement , they will not appear in @xmath14 since the ats condition can not be satisfied here .
therefore , their production in @xmath2 photoproduction can provide direct evidence for their nature as genuine states .
this process can be studied at jlab in the near future .
another interesting process is @xmath15 where the production of the pentaquark states in the invariant spectrum of @xmath3 is also free of the ats ambiguity .
such a possibility can be investigated by belle or future belle - ii .
there have been studies of exotic hadrons via meson photoproduction @xcite .
for instance , it was first proposed in ref .
@xcite to study the tetraquark candidate @xmath16 in meson photoproduction due to its strong coupling to @xmath17 .
this is an alternative way to probe the property of @xmath16 .
similar to @xmath16 , the pentaquark candidates @xmath0 and @xmath1 observed in their decays into @xmath3 as genuine states should have sizeable couplings to @xmath18 .
one can expect that their decays into @xmath18 should be through the @xmath19 annihilations , or it means that such hidden - charm baryons can be created in @xmath2 photoproduction via the @xmath12-channel excitations .
recent studies of hidden - charm pentaquark states can be found in the literature @xcite where , however , only low spin states were investigated .
the lhcb analysis favors that @xmath0 and @xmath1 to have higher spins of either @xmath9 or @xmath10 .
their production mechanism in @xmath18 scatterings should be investigated .
the available data for @xmath2 photoproduction show that the cross section is dominated by the diffractive process , i.e. the pomeron exchanges .
this is a @xmath20-channel process featured by the forward peaking in the differential cross section .
there have been tremendous studies of the diffractive process in the literature ( see e.g. ref .
@xcite ) . due to the dominance of the diffractive process
, it gives us an opportunity for searching for the hidden - charm pentaquark states in @xmath2 photoproduction which will be produced strongly via the @xmath12-channel transition and should dominate at off - forward kinematics . in this work
we study the photoproduction of the @xmath0 and @xmath1 off the proton , since this process can not satisfy the ats condition which makes it is a good place to distinguish a genuine state from the ats kinematic effect .
we include the @xmath20 channel diffractive pomeron exchanges and the @xmath12-channel pentaquark productions . by assuming that the pentaquark states dominantly decay into @xmath3
, we can extract their couplings to @xmath3 by different possible quantum numbers .
we then provide the different cross sections where the angular distributions for different quantum numbers will allow us to learn their nature from experiment .
as follows , we will present the formalism of our model in section ii .
the numerical results follows in section iii and a brief summary will be given in section iv .
in this section , we present detailed formulas for the dominant diffractive pomeron exchange in the @xmath20 channel and the pentaquark excitation in the @xmath12 channel in @xmath2 photoproduction . in @xmath14 , the dominant contribution is the diffractive process which is accounted for by the pomeron exchange model @xcite . in this scenario the pomeron mediates ( fig .
[ fig-2 ] ) the long range interaction between a confined quark and a nucleon .
although the nature of pomeron ( generally interpreted as multi - soft - gluon exchanges ) is still unclear , it is a useful tool for describing the diffractive production of neutral vector mesons in the high energy region .
, scaledwidth=60.0% ] with the same behavior of a @xmath21 isoscalar photon , the coupling between pomeron and quark and/or antiquark contributes a lorentz index @xmath22 .
so the pomeron - nucleon coupling is expressed as @xmath23 where @xmath24 gives the strength of the coupling of a single pomeron to a light constituent quark , and @xmath25 represents the structure function of the nucleon which is expressed as follows @xmath26 with @xmath20 the @xmath20-channel momentum transfer squared . taking the on - shell approximation for the quark loop , a gauge - fixing scheme @xcite
is adopted and the @xmath27 vertex has the following structure : @xmath28 the amplitude of pomeron exchange is @xmath29 where @xmath30 is regge trajectory of the pomeron and it is written as follows @xmath31 where @xmath32 . for any off - shell quark line a form factor , @xmath33 , is introduced , where @xmath34 is the four momentum transfer in the quark line and @xmath35 is the mass of the quark .
equally it will provide a form factor @xmath36 for the @xmath37 vertex .
then , @xmath38 is defined .
all the parameters used for pomeron are the same as those in ref .
@xcite .
the excitation of pentaquark candidates , @xmath0 and @xmath1 in @xmath2 photoproduction can occur via both @xmath12 and @xmath13 channel as illustrated by fig .
[ fig-3 ] ( a ) and ( b ) , respectively .
however , we should note that the @xmath13-channel contribution is negligible since the intermediate @xmath39 will be highly off - shell . and @xmath13-channel production of pentaquark states @xmath39 in @xmath2 photoptroduction . , scaledwidth=90.0% ] since the lhcb analysis suggests that @xmath0 and @xmath1 favor spin @xmath9 and @xmath10 , respectively , but with opposite signs , we then investigate the spin parity assignments of @xmath40 and @xmath41 for these two states in this study .
we adopt the effective lagrangians from ref .
@xcite for the @xmath42 couplings with @xmath43 or @xmath44 for the @xmath39 : @xmath45 where @xmath46 is the pentaquark fields with spin - parity @xmath43 or @xmath44 ; @xmath47 is the nucleon field .
the @xmath48 matrix is defined as follows : @xmath49 and superscript @xmath50 in the @xmath48 matrices means interchange the rows in eq .
( [ gamma - matrix ] ) .
the coupling constants in eq .
( [ photocoupling ] ) will be related to the couplings for @xmath51 interaction via vector meson dominance ( vmd ) .
we also adopt the effective lagrangians for @xmath51 couplings from ref .
@xcite as follows : @xmath52 one can see that for these higher partial wave states they can have different coupling structures which will make it difficult to determine their values .
however , notice that for @xmath0 and @xmath1 decays into @xmath3 , the momentum of the final states are fairly small compared with the nucleon mass .
thus , the higher partial wave terms proportional to @xmath53 and @xmath54 in eqs .
( [ photocoupling ] ) to ( [ strong - coupling ] ) can be neglected .
then , the transition matrix elements to the leading order can be written as @xmath55 where @xmath56 and @xmath57 are the polarization vectors for the initial state photon and final state @xmath2 , respectively ; @xmath58 and @xmath59 are the spin @xmath9 and @xmath10 rarita - schwinger spin projections for the corresponding baryon states , respectively .
they have the following expressions @xcite : @xmath60 \ , \\ \delta_{\rho\sigma ; \
\alpha\beta}(b , p ) & = & ( \slashed{p } + m_b ) \left[\frac 12 ( \bar g_{\rho\alpha } \bar g_{\sigma\beta } + \bar g_{\rho\beta } \bar g_{\sigma\alpha } ) -\frac 15 \bar g_{\rho\sigma } \bar g_{\alpha\beta } -\frac{1}{10 } ( \bar\gamma_\rho \bar\gamma_\alpha \bar g_{\sigma\beta } + \bar\gamma_\rho \bar\gamma_\beta \bar g_{\sigma\alpha } + \bar\gamma_\sigma \bar\gamma_\alpha \bar g_{\rho\beta } + \bar\gamma_\sigma \bar\gamma_\beta \bar g_{\rho\alpha } ) \right ] \ , \end{aligned}\ ] ] where @xmath61 and @xmath34 denote the mass and momentum of baryon @xmath62 and @xmath63 and @xmath64 are defined as the following : @xmath65 the decay of the @xmath39 states into @xmath18 should be via the @xmath19 annihilations . since these states couple to @xmath3 , we assume that their electromagnetic ( em ) couplings are via the @xmath2 pole .
we adopt the vmd leading order coupling between a vector meson and photon : @xmath66 where the coupling constant @xmath67 can be determined by the vector meson leptonic decay @xmath68 , @xmath69 is the em fine - structure constant . for the @xmath70 coupling
, we have @xmath71 . in the framework of vmd , the em coupling @xmath72 in eq .
( [ photocoupling ] ) can be related to the strong coupling in eq .
( [ strong - coupling ] ) : @xmath73 where the real photon condition @xmath74 is applied .
the strong coupling constant @xmath75 can then be estimated by the @xmath76 data .
if we assume @xmath3 channels accounts for the total widths of @xmath0 and @xmath1 , i.e. @xmath7 mev and @xmath8 mev .
we extract the coupling @xmath75 for different quantum numbers as @xmath77 which will give the upper limit of their production cross sections in @xmath2 photoproduction . for states which are off - shell
, we introduce a form factor as usually applied in hadronic models : @xmath78 where @xmath34 is the four - vector momentum carried by the @xmath46 . here
, we use the value of @xmath79 which is more suitable for heavier meson productions @xcite .
one notices that the cross section from these two @xmath46 states will be even larger if @xmath80 increases .
it should be mentioned that in this scheme , we do not consider the nucleon born terms or @xmath20-channel meson exchanges .
various studies have shown that these contributions are negligible in the energy region beyond the conventional @xmath81 region .
this actually provide advantages for photoproduction reaction to directly produce the hidden - charm pentaquark in the @xmath12 channel which can be distinguished from the @xmath20-channel processes . in this formalism
most of the parameters have been well constrained .
the only parameter @xmath75 will have the upper limit given by the experiment as given in eq .
( [ eq - coupling ] ) .
in terms of the c.m . energy . the red dashed , blue dotted , green dot - dashed and black solid curves are the contributions from pomeron , @xmath46 with spin @xmath9 , @xmath46 with spin @xmath82 and the coherent sum of all .
( a ) shows @xmath83 combination .
( b ) shows @xmath84 combination .
the coupling constants ( eq . ( [ eq - coupling ] ) ) between @xmath3 and the two @xmath46 states are extracted by assuming @xmath3 saturates all their total widths .
the experimental data are from refs .
@xcite , scaledwidth=90.0% ] except that the couplings are extracted by assuming @xmath3 channel only accounts for @xmath85 of their total widths.,scaledwidth=90.0% ] proceeding to the numerical results , we first show the total cross section from threshold to 8 gev of the c.m . energy .
the pomeron exchange model can describe well the experimental data at high energies .
the results are shown by the dashed line in figs .
[ fig-4 ] and [ fig-5 ] .
then we argue that the extrapolation of the pomeron exchange to the threshold region can still describe the behavior of the diffractive process in the @xmath2 photoproduction since the @xmath20-channel momentum transfer is still large at the forward angle .
although uncertainties of the pomeron exchange model should exist near threshold , we emphasize that they should not change the smooth behavior of the total cross section .
it means that interfering structures near threshold should indicate the contributions from non - diffractive processes such as the @xmath12 channel . near threshold
, the @xmath12-channel contributions from the pentaquark states can indeed bring obvious corrections to the total cross section as shown by the dotted and dot - dashed curves .
in particular , the contribution from the narrow @xmath1 is significant .
notice that the data for the total cross sections near threshold still have relative large errors .
the couplings between @xmath3 and the two @xmath46 states in fig .
[ fig-4 ] and fig .
[ fig-5 ] are extracted by assuming it accounts for their total widths and @xmath85 , respectively . the combined total cross section ( solid line ) in fig .
[ fig-5 ] appears to be consistent to some extent with the data which means @xmath3 channel is not the dominant decay channel for these two @xmath46 states . as shown in fig .
[ fig-5 ] ( b ) , the favored combination @xmath86 is also favored from the present photoproduction data .
meanwhile , a refined measurement of the cross section lineshape near the threshold region is required .
the calculation results are quite informative . on the one hand , we find that if the @xmath3 decay mode saturates the total widths of these two pentaquark candidates , their cross sections in @xmath2 photoproduction will significantly overestimate the experimental data . on the other hand , we find that a fraction of about @xmath85 for the @xmath3 decay mode can be accommodated by the present photoproduction data .
it means the these two pentaquark candidates , if exist , should not be a simple @xmath3 system .
they should have more stronger couplings to other channels , such as @xmath87 , @xmath88 , @xmath89 , or @xmath90 plus multi - pions .
the angular distributions corresponding to fig .
[ fig-5 ] at four energies , i.e. @xmath91 gev , 4.38 gev , 4.45 gev and 4.5 gev , are shown in figs .
[ fig-6 ] and [ fig-7 ] , respectively . from figs .
[ fig-6 ] ( a ) , [ fig-6 ] ( d ) , [ fig-7 ] ( a ) , [ fig-7 ] ( d ) , we can see that the @xmath12-channel contributions are restricted to a narrow kinematic region .
away from the @xmath46 states region it is still the pomeron exchange plays a key role .
the differential cross section at 4.38 gev and 4.45 gev are dominanted by @xmath0 and @xmath1 as expected .
the contributions from the two @xmath46 states make the differential cross section strongly deviated from the diffractive behavior as shown by figs .
[ fig-6 ] ( b ) , [ fig-6 ] ( c ) , [ fig-7 ] ( b ) , and [ fig-7 ] ( c ) .
these distributions can be measured by experiment at corresponding energies . in brief , either a detail scan of the total cross section in the low energy region or measurement of the angular distributions around 4.38 gev and 4.45 gev in @xmath2 photoproduction will help us pin down the nature of the two @xmath46 states .
it should be pointed out , if any other possible states which can also couple to @xmath3 strongly , they will also cause non - diffractive effects at off - forward angles which can be measured directly . in terms of @xmath92 with
@xmath93 the relative angle between photon and @xmath2 in the c.m .
the red dashed , blue dotted , green dot - dashed and black solid curves are the contributions from pomeron , @xmath46 with spin @xmath9 , @xmath46 with spin @xmath82 and the coherent sum of all with the combination @xmath83 .
( a ) , ( b ) , ( c ) and ( d ) are at the center energy @xmath91 gev , 4.38 gev , 4.45 gev and 4.5 gev respectively.,scaledwidth=90.0% ] similar to fig .
[ fig-6 ] except that quantum numbers of these two pentaquark states are @xmath84 .
, scaledwidth=90.0% ]
we propose to search for the newly observed hidden - charm pentaquark candidates @xmath0 and @xmath1 in @xmath2 photoproduction .
if they are genuine states , their production in photoproduction should be a natural expectation .
since the cross section of the @xmath2 photoproduction is dominated by the diffractive process which is described by the @xmath20-channel pomeron exchange , it should be relatively easy to isolate the direct production of these two states in the @xmath12-channel at off - forward angles in the differential cross section measurement .
we find that the total cross sections near threshold appear to have a non - trivial lineshape due to the excitations of the hidden - charm pentaquark states
. a detailed scan over the energy region from threshold to 5 gev should be able to tell whether @xmath0 and @xmath1 are genuine states or kinematic effects caused by the anomalous triangle singularity @xcite .
this work is supported , in part , by the sino - german crc 110 symmetries and the emergence of structure in qcd " ( nsfc grant no .
11261130311 ) , the national natural science foundation of china ( grant nos . 11035006 and 11425525 ) , and the japan society for the promotion of science under contract no .
p14324 and the jsps kakenhi ( grant no .
25247036 ) .
u. g. meiner and j. a. oller , arxiv:1507.07478 [ hep - ph ] .
v. v. anisovich , m. a. matveev , j. nyiri , a. v. sarantsev and a. n. semenova , arxiv:1507.07652 [ hep - ph ] .
r. f. lebed , phys .
b * 749 * , 454 ( 2015 ) [ arxiv:1507.05867 [ hep - ph ] ] .
f. k. guo , u. g. meiner , w. wang and z. yang , arxiv:1507.04950 [ hep - ph ] .
x. h. liu , q. wang and q. zhao , arxiv:1507.05359 [ hep - ph ] . m. mikhasenko , arxiv:1507.06552 [ hep - ph ] .
x. h. liu , q. zhao and f. e. close , phys .
d * 77 * , 094005 ( 2008 ) [ arxiv:0802.2648 [ hep - ph ] ] . | we propose to study the pentaquark candidates of @xmath0 and @xmath1 in @xmath2 photoproduction and look for further experimental evidence for their nature . since the photoproduction process does not satisfy the so - called anomalous triangle singularity " condition their presence in @xmath2 photoproduction would conclude that they should be genuine states and provide further evidence for their existence . |
the production of strange particles has always been a key observable in heavy - ion reactions and its enhancement was one of the first suggested signatures for quark - gluon plasma ( qgp ) formation @xcite .
the predicted enhancement of strangeness production in nucleus
nucleus collisions relative to proton
proton reactions was established experimentally some time ago @xcite and it was also found that this enhancement is increasing with the strangeness content of the particle type @xcite . however , a clear interpretation of these phenomena requires a systematic investigation of the energy and system size dependence of strangeness production . in the following we report on some aspects of such a study done by the na49 experiment
at mid - rapidity of @xmath4 ( a ) , @xmath5 ( b ) , @xmath6 ( c ) , and @xmath7 ( d ) divided by the pion rapidity densities ( @xmath8 ) in central pb+pb and au+au collisions as a function of @xmath9 @xcite .
the systematic errors are represented by the gray boxes .
also shown are na57 @xcite , ags @xcite , and rhic @xcite data , as well as calculations with hadronic transport models ( hsd , urqmd1.3 @xcite ) and a statistical hadron gas model ( shm(b ) @xcite ) . ] the total yield of @xmath10 divided by the total number of pions @xmath11 ( @xmath12 ) versus the center - of - mass energy @xcite .
the dashed curve shows the prediction from the hadronic transport model urqmd1.3 @xcite . a hadron gas model without strangeness suppression @xcite
is shown by the full curve .
the open squares represent the fits from @xcite including a strangeness under - saturation factor @xmath13 . ]
the rapidity densities @xmath14 at mid - rapidity of @xmath0 divided by the pion rapidity densities ( @xmath8 ) in central nucleus - nucleus collisions as a function of @xmath9 @xcite . also shown are na45/ceres @xcite , and rhic @xcite data , as well as calculations with hadronic transport models ( urqmd1.3 @xcite ) and a statistical hadron gas model ( hgm @xcite ) . ]
figure [ fig : dndypion ] shows a comparison of the energy dependence of mid - rapidity @xmath4 , @xmath5 , @xmath6 , and @xmath7 production to several models and results from other experiments . while the transport models urqmd1.3 and hsd provide a reasonable description of the @xmath4/@xmath15 amd @xmath5/@xmath15 ratios , they are clearly below the data points in case of the @xmath6 and @xmath7 .
this might indicate that an additional partonic contribution is necessary to reach the production rates observed for multi - strange particles .
statistical models on the other hand generally provide a better match to the data .
these models are based on the assumption that the particle yields correspond to their chemical equilibrium value and can thus be described by the parameters temperature @xmath16 , baryonic chemical potential @xmath17 , volume @xmath18 , and , in some implementations , by an additional strangeness under - saturation factor @xmath13 .
the curves shown in fig . [ fig : dndypion ] labeled shm(b ) are taken from @xcite and are based on parametrizations of the @xmath9 dependence of @xmath16 and @xmath17 .
the difference between the two model approaches discussed here is even more prominent for the @xmath2 , as demonstrated in fig .
[ fig : omegaedep ] . in this case
the deviation to the hadronic transport model is of the order of a factor of 10 , while both the statistical model approaches shown in fig .
[ fig : omegaedep ] are quite close to the data points .
while multi - strange hyperons generally seem to be close to the full equilibrium expectation at all energies , the @xmath0-meson exhibits significant discrepancies ( see fig .
[ fig : phiedep ] ) .
while at lower energies the @xmath0 production is close to both , the statistical model and the transport model urqmd1.3 , at top sps energies none of the models does match the measurements .
please note that the appearant discrepancy of urqmd1.3 with the @xmath0/@xmath15 ratios at lower energies , as visible in fig .
[ fig : phiedep ] , is rather due to an overestimate of the pion yields and not an underestimate of the @xmath0 yields @xcite . also shown in fig .
[ fig : phiedep ] is a measurement of the @xmath0 yields via the dielectron decay @xmath19 performed by the na45 collaboration at 158 @xcite .
this result agrees quite well with the na49 result , which has been measured using the hadronic decay branch @xmath20 .
divided by the average number of wounded nucleons @xmath21 of @xmath4 , @xmath5 , and @xmath6 at mid - rapidity for pb+pb collisions at 40@xmath22 and 158 , as well as for near - central c+c and si+si reactions at 158 , as a function of @xmath21 @xcite . also shown
are data of the na57 collaboration @xcite ( open stars ) and calculations with the hsd model @xcite ( dotted lines ) , the urqmd2.3 model @xcite ( dashed lines ) , and the core - corona approach ( solid lines ) @xcite . ]
values at mid - rapidity for pb+pb collisions at 40@xmath22 and 158 , as well as for near - central c+c and si+si reactions at 158 @xcite .
the ( anti-)proton data are taken from @xcite . also shown are the results from a fit for @xmath4 and @xmath5 with the core - corona approach ( solid lines ) . ] the system size dependence of @xmath4 , @xmath5 , and @xmath6 production close to mid - rapidity , as measured at sps energies , is summarized in fig .
[ fig : dndy_vs_nw ] . for @xmath4 and @xmath5 a relatively early saturation at @xmath23
is observed by na49 .
however , a clear discrepancy between the data of na49 and na57 is still present .
the transport models urqmd2.3 @xcite and hsd @xcite are close to the data points for @xmath4 , but are slightly below the @xmath5 measurements .
the @xmath6 production is clearly under - predicted at all system sizes .
the core - corona approach @xcite provides generally a much better description of the system size dependence of all strange particle species . here
the relevant quantity is the fraction of nucleons that scatter more than once @xmath24 which can be calculated in a glauber model .
this allows for an interpolation between the yields @xmath25 measured in elementary p+p ( @xmath26 ) and in central nucleus - nucleus collisions ( @xmath27 ) : @xmath28\end{aligned}\ ] ] please note that the curves shown in fig .
[ fig : dndy_vs_nw ] and fig .
[ fig : meanmt_vs_nw ] are based on a function @xmath24 that was calculated for pb+pb interactions
. therefore their comparison to the smaller systems c+c and si+si is not directly possible , since their surface to volume ratio is different .
it is interesting to observe that this approach not only works for yields , but also for dynamical quantities such as @xmath29 ( see fig .
[ fig : meanmt_vs_nw ] ) .
this suggests that the core - corona picture provides in general a reasonable way for understanding the evolution from elementary p+p to central pb+pb collisions .
( @xmath30 ) divided by the total yields of @xmath31 ( @xmath32 ) in p+p and nucleus - nucleus collisions at 158 as a function of the average number of wounded nucleons @xmath21 .
lower panel : the total yield of @xmath3 ( @xmath30 ) , @xmath33 , and @xmath0 in central pb+pb collisions at 158 divided by the expectation from a statistical model fit @xcite as a function of the resonance lifetime @xmath34 . ]
( @xmath30 ) divided by the total yields of @xmath31 ( @xmath32 ) in p+p and nucleus - nucleus collisions at 158 as a function of the average number of wounded nucleons @xmath21 .
lower panel : the total yield of @xmath3 ( @xmath30 ) , @xmath33 , and @xmath0 in central pb+pb collisions at 158 divided by the expectation from a statistical model fit @xcite as a function of the resonance lifetime @xmath34 . ] strange resonances are of particular interest due to their short lifetimes that are in the same order as the lifetime of the fireball . because of
this their yields can still be modified after chemical freeze - out via destruction and regeneration mechanisms .
for instance the particles resulting from the decay of such a resonance can rescatter in the fireball such that the resonance can not be reconstructed any more .
these effects can thus lead to deviations from the chemical equilibrium expectation .
new data on the @xmath3 ( @xmath30 ) production in central nucleus - nucleus collisions at 158 are summarized in fig .
[ fig : kstar ] . the @xmath3 ( @xmath30 )
are reconstructed via the decay @xmath35 ( @xmath36 ) . as shown in the upper panel of fig .
[ fig : kstar ] , the system size dependence of the total @xmath3 yield is clearly different than the one of charged kaons , the ratios @xmath3/@xmath31 ( @xmath30/@xmath32 ) decrease with increasing system size .
this could be indicative of a stronger reduction of the measurable @xmath3 yields in the larger fireball of central pb+pb reactions compared to the smaller one produced in c+c and si+si collsions , because here their decay products have a higher probability of rescattering with the medium .
the lower panel of fig .
[ fig : kstar ] compares the total yields of several resonances ( @xmath3 , @xmath33 , and @xmath0 ) to the expectations from a statistical model fit @xcite .
the fit did not include the resonances themselves .
the deviation is largest for the short lived @xmath3 , while it is slighly less pronounced for the @xmath33 and even less for the @xmath0 , which has a much longer lifetime than the other two resonances .
comparing the yields of resonances with different lifetimes can thus provide a means to study the time - like extension of the hot and dense fireball created in heavy ion reactions . | systematic studies on the production of strange hyperons and the @xmath0 meson as a function of beam energy and system size performed by the na49 collaboration are discussed .
hadronic transport models fail to describe the production of multi strange particles ( @xmath1 , @xmath2 ) , while statistical models are generally in good agreement to the measured particle yields at all energies .
the system size dependence is well reproduced by the core - corona approach .
new data on @xmath3 production are presented .
the yields of these short - lived resonances are significantly below the statistical model expectation .
this is in line with the interpretation that the measurable yields are reduced due to rescattering of their decay products inside the fireball . |
to possibly extend the ongoing census of exoplanet demographics from our solar neighborhood to the entire milky way we need to understand planet formation and its dependence on initial physical and chemical conditions .
a key step is to directly detect and characterize forming planets in their natal environment .
recently , based on sparse aperture masking observations , a few low - mass companion candidates have been revealed in the gap of their host star s transitional disks ( e.g. , lkca15 b , t cha b ; * ? ? ?
* ; * ? ? ?
* ) , but in some cases scattered light from the disk rim or other disk structures may be a valid explanation for the observed features ( e.g. , * ? ? ? * ) . until now no protoplanet has yet been found embedded in the optically thick gas and dust disk of its host star . here we present high contrast imaging data revealing a protoplanet candidate embedded in the disk around the herbig ae / be star hd100546 .
llc ra ( j2000 ) & 11@xmath833@[email protected] & ( 1 ) + dec ( j2000 ) & -70@xmath611@[email protected] & ( 1 ) + @xmath12 & 6.43@xmath13 mag & ( 2 ) + @xmath14 & 5.96@xmath15 mag & ( 2 ) + @xmath16 & 5.42@xmath13 mag & ( 2 ) + mass & [email protected] m@xmath17 & ( 3 ) + age & 5 ...
@xmath1810 myr & ( 3),(4 ) + distance & 97@xmath19 pc & ( 1 ) + sp .
type & b9vne & ( 5 ) + hd100546 ( see table [ parameters ] for stellar properties ) has a complex circumstellar environment consisting of an inner disk from @xmath30.24 au , a disk gap from @xmath3413 au , and a large outer disk from @xmath313 au out to a few hundred au ( e.g. , * ? ? ?
* ; * ? ? ?
a massive planet was suggested to be orbiting in the gap @xcite . based on asymmetries in the line profile of [ oi ] @xcite and oh @xcite dynamic evidence for such an object
was found . in the case of
oh the emission is thought to arise from an eccentric inner rim of the outer disk with the eccentricity being introduced by a planet .
the outer disk has been resolved at multiple wavelengths including scattered light ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) where it shows peculiar features such as large - scale spiral arms .
the remaining disk mass is estimated to be 10@xmath20
10@xmath21 m@xmath17 ( e.g. , * ? ? ?
recently , polarimetric differential imaging ( pdi ) in the near - infrared ( nir ) revealed distinct sub - structures in the innermost few tens of au of the disk @xcite .
lll date & 2011 - 05 - 30 & 2011 - 07 - 13 + ut start & 22h:48m:18s & 23h:00m:49s + ut end & 00h:14m:58s & 00h:26m:11s + ndit @xmath22 dit & 200 @xmath22 0.15 s & 200 @xmath22 0.15 s + nint & 130 & 126 + parallactic angle start & -17.33@xmath23 & 40.36@xmath23 + parallactic angle end & 10.33@xmath23 & 63.93@xmath23 + airmass range & 1.45 1.43 & 1.54 1.75 + mean dimm seeing [ @xmath24=500 nm ] & 0.6@xmath2 & 0.9@xmath2 + @xmath25 / @xmath26 / @xmath27 & 4.4 / 0.0 / 9.7 ms & 1.9 / 0.0 / 4.0 ms + pa start & -108.35@xmath23 & 129.34@xmath23 + pa end & -80.86@xmath23 & 153.04 @xmath23
hd100546 was observed with vlt / naco @xcite and its apodozing phase plate ( app ) coronagraph @xcite , which was already used in earlier exoplanet imaging projects @xcite .
we used the l27 camera ( @xmath3 27 mas pixel@xmath28 ) with the l@xmath11 filter ( @xmath29 m , @xmath30 m ) in angular differential imaging mode ( adi ; * ? ? ? * ) .
two datasets were taken on two different nights , referred to ashemisphere 1 " and hemisphere 2 " , with a 180@xmath6 offset in position angle of the app to allow its high - contrast half to cover most of the circumstellar environment . using the cube mode " option , all exposures , each 0.15 s long ,
were saved individually .
the core of the stellar psf was slightly saturated .
data cubes consisting of 200 exposures were taken using a 3point dither pattern along the detectors x - axis with roughly 7@xmath2 separation .
table [ observations ] summarizes the observations and the observing conditions .
for photometric calibration we observed hd100546 in the nb3.74 filter ( @xmath31 m , @xmath32 m ; centered on the pf@xmath33 line ) .
the observing strategy was identical to the deep science observations , but these exposures were in the linear detector regime ( 0.2 s exposure time ) and only 2 data cubes with 150 exposures each were recorded .
no photometric or astrometric standard star was observed .
the basic data reduction steps ( bad pixel cleaning , sky subtraction , image alignment ) were done in a similar way as described in @xcite . during the alignment process
the images were re - binned to twice their original resolution . from each image in the stack of aligned exposures we created a 2@xmath342@xmath2 sub - image centered on the star .
individual images showing bad ao correction or not covering the full size of the sub - images were disregarded . in the end
we had a stack of 16,117 images for hemisphere 1 and 18,916 images for hemisphere 2 .
the psf - subtraction was done using the principle component analysis based software package pynpoint @xcite , and the results were confirmed by using the loci algorithm @xcite . for the final pynpoint images we used 80 pca coefficients and kept only the best 12000 images in terms of total residuals over the whole image frame .
after psf subtraction , each image was de - rotated to the same field - orientation and we computed the mean image of the image stack clipping data points that were beyond 2.5@xmath35 of the mean .
the results shown below are robust against all of these numbers ( see next section ) . for loci
we median combined 20 consecutive exposures into a single image in those data cubes where this was possible , resulting in 734 stacked images .
the loci algorithm was then applied to this stack of images using the following loci parameters : fwhm=8 px , @xmath36=0.75 , @xmath37=8 , @xmath38=500 .
the choice of these values reflect the fact that the images have been re - binned to twice their original resolution ( see above ) .
all final images ( see , figure [ images ] ) were smoothed with a circular gaussian with a width of 3 pixels .
in figure [ images ] an emission source is revealed north of the central star in the hemisphere 1 dataset . to examine the robustness of this detection we did a series of tests :
* we varied the number of pca coefficients used in pynpoint ( between 20 and 200 ) . *
we split the dataset in half in different ways including random selections of @xmath350% of the images .
* we applied the loci algorithm as independent reduction approach . in all cases
we found a bright feature at the same location .
in addition , we analyzed the hemisphere 2 dataset in exactly the same way without finding a persistent source at any location in the final image . using the approach described in @xcite and a 8-pixels wide aperture ( @xmath39fwhm ) , the source has a signal - to - noise of @xmath315 in the final loci image .
the detected emission appears slightly elongated in the northern direction ( figure [ images ] ) . to estimate the photometry and astrometry of the source we did a detailed analysis inserting fake negative planets ( details see below ) .
it showed that the observed emission can be explained with a point source plus some extended component .
for the point source the projected separation amounts to @xmath30.48@xmath40 ( @xmath41 au ) .
the uncertainties in the exact location of the central star and the point source component in x and y on the detector were 0.5 and 1 pixel , respectively .
the position angle of the point source with respect to the star is @[email protected]@xmath6 .
this error excludes any systemic error in the orientation of the camera with respect to the true celestial north , which is estimated to be @xmath430.5@xmath6 based on calibration data from an ongoing large imaging program ( pi : j .- l .
beuzit ) . to estimate the brightness of the point source we inserted artificial negative planets in the individual exposures and re - ran pynpoint . for the fake objects we used an unsaturated psf of hd100546 from one of the photometric calibration datasets . to scale the flux of these objects , the difference in exposure time between the science and the calibration images had to be considered as well as the transmission curves of the two different filters . using published l - band spectra for hd100546 from iso and vlt / isaac ( * ? ? ?
* and references therein ) we derived a throughput fraction of @xmath3@xmath44 for the narrow band filter compared to the broadband l@xmath11 filter .
the error arises from changes in the pf@xmath33 line emission in hd100546 between the two datasets suggesting that the nb3.74 filter traces variable accretion activity .
also the whole nir and mir continuum varies with an offset of a factor of @xmath31.25 between the two datasets of @xcite , which impacts the error in the final photometry . for our contrast estimates we varied the brightness of the injected fake sources in steps of 0.1 mag and used two independent methods .
first , we searched for a negative source that , when subtracted , yielded a remaining flux at the object s location similar to the flux level in the surroundings , i.e. , in the extended flux component . secondly , we canceled out all the flux at the object s location .
our best estimate contrasts were @xmath45l@xmath46 mag and @xmath45l@xmath47 mag for both methods , respectively .
a contrast closer @xmath45l@xmath48 mag appears to be more likely , as strong residuals become present in the vicinity of the object not at the object s location itself if we cancel out its peak flux completely . for the rest of the analyses and discussion we use @xmath45l@xmath49 mag as default value .
the key points of the discussion and conclusions remain unchanged for a smaller value of the contrast . the observed flux densities for hd100546 based on the spectra mentioned above translate into an apparent brightness of l@xmath50 4.3 mag , which fits well to the l@xmath51 mag reported in @xcite .
hence we derive an apparent magnitude of l@xmath11=13.2 @xmath4 0.4 mag for the point source component . here
the error is the root of sum of squares of 0.3 mag for the uncertainty in the contrast and @xmath30.25 mag for the uncertainties in the mir continuum .
compared to these uncertainties the variability of the pf@xmath33 line flux and the intrinsic error in our photometric observations are negligible . assuming that the flux of the point source peaks in the l@xmath11 filter we can estimate its blackbody temperature using wien s law .
we can then derive a lower limit on the object s luminosity by taking into account its apparent l@xmath11 magnitude and its distance .
integrating over all frequencies this exercise yields a minimum luminosity of @xmath52 .
the vlt / naco pdi data presented in @xcite have sufficient spatial resolution and inner working angle to probe the disk surface on scales relevant for the app dataset .
those nir observations revealed sub - structures in the disk in the inner few tens of au .
in particular the existence of a disk hole " was suggested as both the final polarization intensity images as well as the polarization fraction images in h and k@xmath53 revealed a local flux deficit at the same location . in figure [ disk ]
we show the large scale disk environment revealed by hst / acs @xcite and then , zooming in the inner disk regions , the polarization fraction image of the pdi study .
we overplot in red the contours of the object detected here .
the disk is inclined by @xmath347@xmath423@xmath6 and the position angle of the disk major axis is @xmath3138@xmath424@xmath6 @xcite .
if the disk surface was smooth and azimuthally symmetric , the disk image shown in figure [ disk ] should be mirror symmetric with respect to the disk minor axis running with a position angle of @xmath348@xmath6 through the image center @xcite . however , there are clear asymmetries in form of a deficit in polarization fraction in northern direction , i.e. , along the position of the detected object . based on figure [ disk ] the disk hole "
extends to larger separations and appears more like a wedge " . as discussed in @xcite
the underlying physical reason for this feature is not clear at the moment ( e.g. , drop in surface density , disk surface geometry , changing dust properties ) .
however , finding an asymmetry at this specific location renders plausible a physical link between those structures and the source detected here .
based on the object s position angle , the disk inclination and the distance to hd100546 , the object s de - projected separation from the central star is @xmath368@xmath410 au , i.e. , within the large circumstellar disk .
different scenarios to explain both the l band emission and the observed disk structure can be assessed : _ background source : _ a background source would be observed through the hd100546 disk .
based on the disk model presented in @xcite background flux in the l band should be attenuated by a factor of @xmath36.7@xmath54 mag at a location of @xmath370 au . taking this factor into account
we used the besancon galactic model @xcite to estimate the number of objects in the apparent magnitude range 7 mag @xmath55 l @xmath55 9 mag .
this yielded @xmath3330 objects in a 2 square degree patch on the sky centered around hd100546 .
this number translates into a probability of having such a physically unrelated source in a 1@xmath2 @xmath22 1@xmath2 field of view around the star of @xmath56 .
furthermore , the fact that the l band emission appears to be extended argues against a background object . _
disk feature : _ the observed l@xmath11 brightness and minimum luminosity are difficult to explain with disk - internal processes alone as the expected temperature in the disk mid - plane at the location of the source is only @xmath350 k @xcite . furthermore , we are not aware of shock - processes that act only locally and might lead to the observed luminosity in a disk that appears to be not very massive .
if it was scattered light that we see , one would expect that also in the nir a maximum in scattered light would be seen . using the pdi images as tracer for scattered light we find a local _ minimum _ here as described above .
_ photospheric emission : _ if the observed point source flux arose solely from the photosphere of a young object the cond and dusty models suggest masses between @xmath315 20 m@xmath57 for an age of 5 10 myr @xcite .
models with lower specific entropy in the initial conditions for the formation process predict even higher masses ( cf . * ? ? ?
classical binary formation via core fragmentation or formation via disk instability when the disk was still massive would be the preferred formation mechanisms for an object of this mass . in this case the object formed roughly coeval with the star and would have had time to significantly alter the structure of the main disk , e.g. , dynamically clearing a large azimuthal gap , which has not been observed . _
ejected planet : _ another massive planet is thought to be orbiting in the inner disk gap ( e.g. , * ? ? ?
* ; * ? ? ?
* ) and we speculate that dynamical interactions between multiple planets and the disk could have led to an ejection event .
the emission we see in the l@xmath11 images would then be a combination of the planet s intrinsic emission plus extra luminosity from disk material being heated from the planet moving through the disk .
assuming that the planet was initially orbiting at 10 au its orbital period was @xmath320 years yielding an orbital velocity of @xmath33.2 au@xmath58yr@xmath28 ( @xmath315.2 km@xmath58s@xmath28 ) .
if the ejection velocity is a few times that value it would have taken the planet less then 20 years to reach its current location and within less than 100 years it would be beyond the extent of the observable disk .
given the age of the system , this timescale is extremely small and observing the object exactly at the right time is unlikely .
adding further complexity to this scenario , the ejection needs to occur roughly in the plane of the disk to make a link to the observed disk structures .
_ forming planet : _ in our view the best explanation for the observed morphology of both the disk and the emission source is the detection of a planet during its formation process .
the luminosity of the object is not coming from an isolated photosphere , but rather the planet is still accreting material from the disk .
young , forming gas giants with masses between one and a few jupiter masses are expected to have luminosities between 10@xmath5910@xmath20@xmath60 during the first few hundred thousand years after gas runaway accretion sets in @xcite , in agreement with our lower limit .
furthermore , an object in this mass range is expected to affect the disk structure much less and an azimuthal gap if it exists might be below our detection limits in the pdi data .
a narrow gap is hard to see in an inclined disk .
this scenario could also explain the extended component of the emission detected here with some disk material being heated in the accretion process similar to the case of lkca15 b ( see * ? ? ?
* for possible mechanisms to heat the surrounding material during the accretion process ) .
however , we acknowledge that from a theoretical perspective the formation of a gas giant planet at this location is not readily explainable using _ first principles_. for core accretion the timescales to assemble a massive rocky core seem to exceed the estimated age of the star , and given the observed disk parameters the disk does not seem to be gravitationally unstable . the different dynamics involved in the scenarios outlined above and multi - wavelength photometry and/or spectra will eventually help us to confirm which hypothesis is correct . the proper and parallactic motion of hd100546 will allow us to rule out a background source with second epoch observations obtained as early as mid 2013 . to distinguish between the ejection scenario and the forming planet " scenario ,
the baseline for follow - up observations needs to be a few years .
while the object is expected to orbit its star with a period of @xmath3360 years in the latter case , it should move away quickly in radial direction if it were ejected .
also , high spatial resolution alma observations will help to search for an azimuthal gap in the surface mass density ( gas and/or dust ) at the planet s location .
spatially resolved information about the existence ( or non - existence ) of an azimuthal disk gap may allow us to derive a dynamical mass estimate for the planet @xcite .
we have presented observational evidence that a gas giant planet could be forming in the circumstellar disk around the herbig ae / be star hd100546 at a separation of @xmath368 au .
this scenario , among others that we discussed , seems best capable of explaining most of the available data .
however , some aspects remain qualitative and follow - up observations are required to validate our proposed interpretation . together with lkca15 b @xcite ,
the object presented here is currently the best candidate for a forming young gas giant planet .
particularly interesting is that hd100546 b " would be the first protoplanet that is still embedded in a circumstellar disk and that it forms at large orbital separations . if confirmed , hd100546 would be a unique laboratory to study planet formation and the interaction between forming planets and the disk directly .
this research made use of the simbad database , operated at cds , strasbourg , france , and of nasa s astrophysics data system .
v. geers kindly provided us with the iso and isaac spectra .
we are indebted to f. meru , c. dominik , h. m. schmid and r. v. boekel for helpful discussions . , r. m. , skrutskie , m. f. , van dyk , s. , beichman , c. a. , carpenter , j. m. , chester , t. , cambresy , l. , evans , t. , fowler , j. , gizis , j. , howard , e. , huchra , j. , jarrett , t. , kopan , e. l. , kirkpatrick , j. d. , light , r. m. , marsh , k. a. , mccallon , h. , schneider , s. , stiening , r. , sykes , m. , weinberg , m. , wheaton , w. a. , wheelock , s. , & zacarias , n. 2003 , 2mass all sky catalog of point sources .
( the irsa 2mass all - sky point source catalog , nasa / ipac infrared science archive .
http://irsa.ipac.caltech.edu/applications/gator/ ) , c. a. , polomski , e. f. , henning , t. , stecklum , b. , woodgate , b. e. , telesco , c. m. , pia , r. k. , gull , t. r. , boggess , a. , bowers , c. w. , bruhweiler , f. c. , clampin , m. , danks , a. c. , green , r. f. , heap , s. r. , hutchings , j. b. , jenkins , e. b. , joseph , c. , kaiser , m. e. , kimble , r. a. , kraemer , s. , lindler , d. , linsky , j. l. , maran , s. p. , moos , h. w. , plait , p. , roesler , f. , timothy , j. g. , & weistrop , d. 2001 , , 122 , 3396 , m. a. , quanz , s. p. , meyer , m. r. , kasper , m. e. , lenzen , r. , codona , j. l. , girard , j. h. , & hinz , p. m. 2010 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol .
7735 , society of photo - optical instrumentation engineers ( spie ) conference series , r. , hartung , m. , brandner , w. , finger , g. , hubin , n. n. , lacombe , f. , lagrange , a. , lehnert , m. d. , moorwood , a. f. m. , & mouillet , d. 2003 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol .
4841 , society of photo - optical instrumentation engineers ( spie ) conference series , ed .
m. iye & a. f. m. moorwood , 944952 , g. , lacombe , f. , puget , p. , hubin , n. n. , gendron , e. , fusco , t. , arsenault , r. , charton , j. , feautrier , p. , gigan , p. , kern , p. y. , lagrange , a. , madec , p. , mouillet , d. , rabaud , d. , rabou , p. , stadler , e. , & zins , g. 2003 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol .
4839 , society of photo - optical instrumentation engineers ( spie ) conference series , ed .
p. l. wizinowich & d. bonaccini , 140149 | we present high - contrast observations of the circumstellar environment of the herbig ae / be star hd100546 .
the final 3.8 @xmath0 image reveals an emission source at a projected separation of [email protected]@xmath2 ( corresponding to @xmath347@xmath44 au ) at a position angle of [email protected]@xmath6 . the emission appears slightly extended with a point source component with an apparent magnitude of @xmath7 mag .
the position of the source coincides with a local deficit in polarization fraction in near - infrared polarimetric imaging data , which probes the surface of the well - studied circumstellar disk of hd100546 .
this suggests a possible physical link between the emission source and the disk . assuming a disk inclination of @xmath347@xmath6 the de - projected separation of the object is @xmath368 au . assessing the likelihood of various scenarios we favor an interpretation of the available high - contrast data with a planet in the process of forming .
follow - up observations in the coming years can easily distinguish between the different possible scenarios empirically .
if confirmed , hd100546
b " would be a unique laboratory to study the formation process of a new planetary system , with one giant planet currently forming in the disk and a second planet possibly orbiting in the disk gap at smaller separations . |
1 . bethe , h. , z. physik , 1931 , 71 , 205 ( english translation in the book ` the many - body problem : an encyclopedia of exactly solved models in one dimension ' ( ed .
mattis , d.c . ) , world scientific , 1993 2 .
hase , m. , tarasaki , i. and uchinokura , k. , phys .
lett . , 1993 , 70 , 3651 3 .
shiramura , w. , takatsu , k. , tanaka , h. , kamishima , k. , takahashi , m. , mitamura , h. and goto , t. , j. phys .
jpn . , 1997 , 66 , 1900 4 .
takatsu , k. , shimamura , w. and tanaka , h. , j. phys .
jpn . , 1997 , 66 , 1611 5 .
ramirez , a. p. , ann .
, 1994 , 24 , 453 6 .
chalker , j. t. , holdsworth , p. c. w. and shender , e. f. , phys .
lett . , 1992 , 68 , 855 7 .
huse , d. a. and elser , v. , phys .
lett . , 1988 , 60 , 2531 8 .
bernu , b. , lhuillier , c. and pierre , l. , phys .
lett . , 1992 , 69 , 2590 9 .
harrison , a. , annual reports on the progress of chemistry , 1992 , 87a , 211 10 .
villain , j. , bidaux , r. , carton , j. p. and conte , r. , journal de physique , 1980 , 41 , 1263 ; shender , e. f. , sov .
jetp , 1982 , 56 , 178 11 .
lecheminant , p. , bernu , b. , lhuillier , c. , pierre , l. and sindzingre , p. , phys .
b , 1997 , 56 , 2521 12 .
mila , f. , eur .
j. phys . , 2000 , 21 , 499 13 .
majumdar , c. k. and ghosh , d. k. , j. math .
phys . , 1969 , 10 , 1388 14 .
shastry , b. s. and sutherland , b. , phys .
lett . , 1981 , 47 , 964 15 .
bose , i. and mitra , p. , phys .
b , 1991 , 44 , 443 16 .
bhaumik , u. and bose , i. , phys .
b , 1995 , 52 , 12484 17 .
shastry , b. s. and sutherland , b. , physica b , 1981 , 108 , 1069 18 .
miyahara , s. and ueda , k. , phys .
lett . , 1999 , 82 , 3701 19 .
dagotto , e. and rice , t. m. , science , 1996 , 271 , 618 20 .
rice , t. m. , gopalan , s. and sigrist , m. , europhys .
lett . , 1993 , 23 445 21 .
bose , i. and gayen , s. , phys .
b , 1993 , 48 , 10653 22 .
xian , y. , phys .
b , 1995 , 52 , 12485 23 .
ornstein , j. and millis , a. j. , science , 2000 , 288 , 468 24 .
dagotto , e. , riera , j. and scalapino , d. j. , phys .
b , 1992 , 45 , 5744 25 .
uehara , m. , nagata , t. , akimitsu , j. , takahashi , h. , mori , n. and kinoshita , k. , j. phys .
jpn . , 1996 , 65 , 2764 26 .
dagotto , e. , rep .
phys . , 1999 , 62 , 1525 27 .
bose , i. and gayen , s. , j. phys . : condens .
matter , 1994 , 6 , l405 28 .
bose , i. and gayen , s. , j. phys .
: condens .
matter , 1999 , 11 , 6427 29 .
troyer , m. , tsunetsugu , h. and rice , t. m. , phys .
b , 1996 , 53 , 251 30 . rice , t. m. , z. phys .
b , 1997 , 103 , 165 31 .
lieb , e. , schultz , t. d. and mattis , d. c. , ann .
phys . , 1961 , 16 , 407 32 .
haldane , f. d. m. , phys .
lett . , 1983 , 50 , 1153 ;
a , 1983 , 93 , 464 33 .
affleck , i. , j. phys . : condens .
matter , 1989 , 1 , 3047 34 .
affleck , i. , kennedy , t. , lieb , e. h. and tasaki , h. , phys .
lett . , 1987 , 59 , 799 ; commun .
phys . , 1988 , 115 , 477 35 .
di tusa , j. f. , cheong , s - w . , park , j .- h . ,
aeppli , g. , broholm , c. and chen , c. t. , phys .
lett . , 1994 , 73 , 1857 36 .
granroth , g. e. , meisel , m. w. , chaparala , m. , jolicoeur , t. , ward , b. h. and talham , d. r. , phys .
lett . , 1996 , 77 , 1616 37 .
penc , k. and shiba , h. , phys .
b , 1995 , 52 r715 ; dagotto , e. , riera , j. , sandvik , a. and moreo , a. , phys . rev .
, 1996 , 76 , 1731 38 .
xu , g. et al . , science , 2000 , 289 , 419 39 .
malvezzi , a. l. and dagotto , e. , phys .
b , 2001 , 63 , 140409 40 .
bose , i. and chattopadhyay , e. , int .
b , 2001 , 15 , 2535 41 .
sondhi , s. l. , girvin , s. m. , carini , j. p. and shahar , d. , rev .
, 1997 , 69 , 315 ; sachdev , s. , science , 2000 , 288 , 475 ; sachdev , s. , quantum phase transitions , cambridge university press , cambridge , 1999 42 .
chaboussant , g. et al .
j. b , 1998 , 6 , 167 43 .
watson , b. c. et al . ,
, 2001 , 86 , 5168 44 .
landee , c. p. , turnbull , m. m. , galeriu , c. , giantsidis , j. and woodward , f. m. , phys .
b , 2001 , 63 , 100402 45 .
chaboussant , g. , crowell , p. a. , lvy , l. p. , piovesana , o. , madouri , a. and mailly , d. , phys .
b , 1997 , 55 , 3046 46 . rice , t. m. , science , 2002 , 298 , 760 ; nikuni , t. , oshikawa , m. , oosawa , a. and tanaka , h. , phys .
lett . , 2000 , 84 , 5868 ; normand , b. , matsumoto , m. , nohadani , o. , wessell , s. , haas , s. , rice , t. m. and sigrist , m. , j. phys . : condens .
matter , 2004 , 16 , s867 47 .
regg , ch .
nature , 2003 , 423 , 62 48 .
oshikawa , m. , yamanaka , m. and affleck , i. , phys . rev .
, 1997 , 78 , 1984 49 .
hida , k. , j. phys .
jpn . , 1994 , 63 , 2359 50 .
shiramura , w. et al .
, j. phys .
jpn . , 1998 , 67 , 1548 51 .
kageyama , h. et al . ,
lett . , 1999 , 82 , 3168 ; onizuka , k. , kageyama , h. , narumi , y. , kindo , k. , ueda , y. and goto , t. , j. phys .
jpn . , 2000 , 69 , 1016 52 .
momoi , t. and totsuka , k. , phys .
b , 2000 , 61 , 3231 53 .
kodama , k. et al . , science , 2002 , 298 , 395 54 .
katoh , n. and imada , m. , j. phys .
jpn . , 1995 , 64 , 4105
; troyer , m. , kontani , h. and ueda , k. , phys .
lett . , 1996 , 76 , 3822
; bose , i. and ghosh , a. , phys . rev .
b , 1997 , 56 , 3154 55 . johnsson , m. , trnroos , k. w. , mila , f. and millet , p. , chem .
2000 , 12 , 2853 ; brenig , w. and becker , k. w. , phys .
b , 2001 , 64 , 214413 ; totsuka , k. and mikeska , h - j . , phys . rev .
b , 2002 , 66 , 054435 56 .
kolezhuk , a. k. and mikeska , h - j . , int
b , 1998 , 12 , 2325 57 .
greiner , m. , mandel , o. , esslinger , t. , hnsch , t. w. and bloch , i. , nature , 2002 , 415 , 39 58 .
duan , l .- m . ,
demler , e. and lukin , m. d. , phys .
lett . , 2003 , 91 , 090402 ; garcia - ripoll , j. j. , martin - delgado , m. a. and cirac , j. i. , cond - mat/0404566 ; pachos , j. k. and plenio , m. b. , quant - ph/0401106 59 .
loss , d. and di vincenzo , d. p. , phys .
rev . a , 1998 , 57 , 120 ; lidar , d. a. and wu , l .- a .
2002 , 88 , 017905 ; bose , s. , phys . rev .
lett . , 2003 , 91 , 207901 ; subrahmanyam , v. , quant - ph/0307135 60 .
bose , i. and chattopadhyay , e. , phys .
a , 2002 , 66 , 062320 ; verstraete , f. , martin - delgado , m. a. and cirac , j.i . ,
, 2004 , 92 , 087201 ; fan , h. , korepin , v. and roychowdhury , v. , quant - ph/0406067 ; li , y. , shi , t. , song , z. and sun , c. p. , quant - ph/0406159 1 .
five types of interaction in the @xmath57 model .
the successive interactions are n.n . ,
diagonal , n.n.n .
, knights - move - distance - away and further - neighbour - diagonal .
2 . four columnar dimer states .
the dotted line represents a valence bond , i.e. , a singlet spin configuration .
3 . the shastry - sutherland model . the n.n . and diagonal exchange interaction strengths are @xmath40and @xmath41 respectively .
4 . a two - chain spin ladder . | there are several interacting spin systems which have a gap in their spin excitation spectrum
. the gap does not occur due to anisotropies present in the system but is quantum mechanical in origin .
we give a brief overview on the different types of spin gap ( sg ) antiferromagnets , the models proposed to describe their physical properties and experimental realizations of such systems .
our special focus is on exactly - solvable models and rigorous theories which provide the correct physical picture for the novel phenomena exhibited by sg systems .
keywords : antiferromagnet , spin gap , exchange interaction , frustration , spin ladder , quantum phase transition the last decade has witnessed an unprecedented research activity on undoped and doped quantum antiferromagnets ( afms ) in low dimensions .
several new materials exhibiting a variety of novel phenomena have been discovered .
the experimental effort is closely linked with theoretical ideas .
the field of low - dimensional quantum magnetism provides a fertile ground for rigorous theory .
powerful techniques like the bethe ansatz ( ba)@xmath0 and bosonization are available to study ground and excited state properties .
models of interacting spin systems are known for which the ground state and in some cases the low - lying excitation spectrum are known exactly .
theorems , furthermore , offer important insight on the nature of ground and excited states .
the knowledge gained provides the impetus to look for real materials so that experimental confirmation of theoretical predictions can be made . in this review
, we discuss a special class of low - dimensional afms , the so - called spin gap ( sg ) systems , to illustrate the rich interplay between theory and experiments .
the review is not meant to be exhaustive but focuses on some broad classes of phenomena exhibited by sg systems .
the sg afms are characterised by a gap in the spin excitation spectrum .
the magnitude of the gap , @xmath1 , is measured as the difference between the energies of the lowest excited state and the ground state .
the excitation spectrum of an afm is said to be gapless if @xmath2 in general , a gap appears in the spin excitation spectrum if some kind of anisotropy is present in the system . in sg afms , however , the gap is purely quantum mechanical in origin and can not be ascribed to any anisotropy .
the basic spin - spin interactions in afms are described by the heisenberg hamiltonian @xmath3 where @xmath4 is the spin operator located at the lattice site @xmath5 and @xmath6 denotes the strength of the exchange interaction .
real magnetic materials are three - dimensional ( 3d ) but behave effectively as low - dimensional systems if the dominant exchange interactions are intra - chain ( 1d ) or intra - planar ( 2d ) . for most materials ,
the exchange interaction is confined to only the nearest - neighbour ( n.n . ) spins on the lattice . also , @xmath6 s have the same magnitude @xmath7 for all the n.n . interactions .
there are , however , spin systems in which further - neighbour interactions can not be ignored and the exchange interaction strengths are inhomogeneous in character . in the next section , we discuss some possible origins of sg , the models proposed to describe such systems and experimental evidence for the different mechanisms of sg .
sg afms have spin - disordered ground states , i.e. , the spin - spin correlations in the ground state are short - ranged . the ground states , in the absence of long range magnetic order , can broadly be described as quantum spin liquids ( qsls ) .
the spin liquids are distinct from simple paramagnets , in certain cases one can define novel order parameters which have non - zero expectation values in the qsl states .
formation of such states is favoured by quantum fluctuations the effect of which is prominent in low dimensions and for low values of the spin . in experiments , the presence of the gap @xmath1
is confirmed through measurement of quantities like the susceptibility , @xmath8 which goes to zero exponentially at low t as @xmath9 .
some well - known examples of sg afms are : spin - peierls ( sp ) systems , afm compounds consisting of weakly - coupled spin dimers , frustrated spin systems , spin ladders and haldane gap ( hg ) afms .
the sp compounds are historically the first examples of magnetic systems exhibiting a sg .
the sp transition was originally observed in some organic compounds . in 1993 ,
hase et al.@xmath10 obtained the first experimental evidence of the sp transition in an inorganic compound , @xmath11 the sp transition generally occurs in quasi - one dimensional ( 1d ) afm spin systems with half - odd integer spins and is brought about by spin - phonon coupling .
below the sp transition temperature @xmath12 , a periodic deformation of the lattice sets in such that the distances between neighbouring spins are no longer uniform but alternate in magnitude .
this results in an alternation , @xmath13and @xmath14 in the strengths of the n.n . exchange interaction strengths .
the ground state is dimerized in which singlet spin pairs occupy the links with enhanced exchange couplings .
the sp ground state is spin disordered and a finite energy gap exists in the @xmath15 spin excitation spectrum .
some afm compounds can be described as crystalline networks of spin dimers .
well - known examples of such systems are the compounds @xmath16 ( @xmath17 ) in which a spin dimer arises from two antiferromagnetically coupled @xmath18ions@xmath19 .
the dimer ground state is a spin singlet with total spin @xmath20 the excited triplet state with total spin @xmath15 is separated from the ground state by an energy gap .
the excitation created on a particular dimer propagates through the network of dimers due to the weak inter - dimer exchange coupling .
* frustrated spin systems * quantum fluctuations in a spin system are enhanced due to the presence of frustration in the system .
frustration implies a conflict in minimising the exchange interaction energies associated with different spin pairs .
such conflicts arise mainly due to lattice topology and the presence of competing further - neighbour interactions@xmath21 .
afm systems with an odd number of bonds in the unit cell of the underlying lattice are frustrated spin systems .
examples include the triangular , kagom , pyrochlore and fcc lattices .
consider three ising spins located at the vertices of a triangular plaquette and interacting antiferromagnetically with each other .
the energy of an interacting spin pair is minimised when the spins are antiparallel .
the three interacting spin pairs can not , however , be simultaneously made antiparallel so that one pair is in the parallel spin configuraton in the lowest energy state .
such conflicts are absent when the elementary plaquette of the lattice contains an even number of bonds as in the case of the square lattice .
determination of the ground state of the afm heisenberg exchange interaction hamiltonian ( eq.(1 ) ) on various lattices poses a formidable theoretical problem . the task becomes simpler if the spins are treated as classical vectors , i.e. , the magnitude of spins @xmath22 the energy of a triangle of spins is given by @xmath23\label{2}\ ] ] the ground state , i.e. , the minimal energy spin configurations are obtained for @xmath24 in the case of the full triangular lattice , the classical ground state has a three - sublattice order with each elementary plaquette of spins satisfying the constraint @xmath25 in the case of the kagom lattice , which is constructed out of triangular plaquettes , the classical ground state is infinitely degenerate@xmath26 . in the first case , quantum fluctuations do not fully destroy the order in the ground state but reduce the sublattice magnetization significantly from its classical value .
recent calculations show that the @xmath27 heisenberg afm ( hafm ) hamiltonian on the triangular lattice has long range order ( lro ) in the ground state@xmath28 .
a rigorous proof of the existence of lro is , however , still missing .
experimental realizations of the triangular lattice hafm include @xmath29 etc@xmath30 . in the case
when the classical ground state is highly degenerate , i.e. , spin - disordered , thermal / quantum fluctuations may select a subset of states as ground states leading to new forms of spin order .
this is the phenomenon of ` order from disorder'@xmath31 .
another possibility is the opening up of a gap in the spin excitation spectrum .
the @xmath27 hafm on the kagom lattice appears to be a sg system with a singlet - triplet gap in the spin excitation spectrum .
an interesting feature of the sg system is the existence of a large number of singlet excitations in the gap , the number of which is proportional to @xmath32where n is the number of sites in the lattice@xmath33 .
the temperature dependence of the susceptibility is not affected by the presence of singlet excitations but the specific heat possibly has a power - law dependence , @xmath34 due to contributions from the singlets .
an experimental realization of a @xmath27 hafm on the kagom lattice is yet to be obtained .
the compound @xmath35@xmath21 is an example of a kagom - lattice afm with @xmath36 the dominant contribution to the low temperature specific heat of this compound appears to come from the singlet states .
the @xmath27 @xmath37 model in 1d , describing the majumdar - ghosh ( mg ) chain@xmath38 is the first example of a frustrated quantum spin model with further - neighbour interactions for which the ground state can be determined exactly .
the hamiltonian with periodic boundary conditions ( pbc ) is given by @xmath39 where @xmath40 and @xmath41 are the n.n . and next - nearest - neighbour ( n.n.n . )
exchange interaction strengths .
the exactly - solvable mg point corresponds to @xmath42 the exact ground state is doubly degenerate and the ground states are @xmath43\left[34\right]\left[56\right] ...... \left[n-1n\right],\phi_{2}\equiv\left[23\right]\left[45\right]\left[67\right] ...... \left[n1\right]\label{4}\ ] ] where @xmath44 $ ] denotes a singlet spin configuration , @xmath45 for spins located at the lattice sites @xmath46 and @xmath47 .
the up and down spin states are denoted as @xmath48 and @xmath49 .
a singlet state is also known as a valence bond ( vb ) .
the excitations in the model can be described in terms of scattering spin-@xmath50 defects acting as domain walls between the two exact ground states@xmath51 the scattering states form a continuum , the lowest branch of which is separated from the ground state by a gap .
the mg chain is thus a sg afm .
the sg phase survives for @xmath52 greater than a critical value @xmath53 for @xmath54 a gapless phase is obtained . if the n.n .
exchange interactions are alternating in strength , as in a sp system , the ground state is nondegenerate with the vbs forming along the stronger bonds .
the mg hamiltonian has been used to study the properties of the sp compound @xmath11 a large number of studies has been carried out on the frustrated @xmath37 and @xmath55 models in 2d where @xmath41 denotes the strength of the diagonal exchange couplings and @xmath56 that of n.n.n .
interactions in the horizontal and vertical directions .
the most general model studied so far is the @xmath57 model@xmath58 @xmath59 and @xmath60 are the strengths of the knights - move - distance - away and further - neighbour diagonal exchange interactions respectively ( figure 1 ) .
the four columnar dimer ( cd ) states ( figure 2 ) are the exact eigenstates of the @xmath57 hamiltonian for the ratio of interaction strengths @xmath61 each dotted line in figure 2 represents a vb , i.e. , a singlet spin configuration .
there is no rigorous proof as yet that the cd states are the exact ground states though approximate theories tend to support the conjecture@xmath62 one can , however , prove that any one of the cd states is the exact ground state when the dimer bonds are of strength @xmath63 and the rest of the exchange interactions are of strengths as specified in eq .
( 5 ) .
the excitation spectrum of the model is separated from the ground state by a gap .
the shastry - sutherland ( ss ) model@xmath64 is an example of a frustrated sg afm in 2d .
figure 3 shows the lattice on which the model is defined .
the n.n .
and diagonal exchange interactions are of strengths @xmath40 and @xmath41 respectively . for @xmath65 less than a critical value
@xmath66 0.7 , the exact ground state consists of singlets along the diagonals . at the critical point , the ground state changes from the gapped disordered state to an antiferromagnetically ordered gapless state .
the afm compound @xmath67 is an experimental realization of the ss model@xmath68 triplet excitations in the model are found to be almost localized .
a single triplet excitation can propagate in the ss lattice only at the sixth order perturbation in @xmath69 * spin ladders * spin ladders constitute one of the most well - known examples of sg agms .
the simplest ladder model consists of two chains coupled by rungs ( figure 4 ) . in general , the ladder may consist of @xmath70 chains . in the spin ladder model , each site of the ladder is occupied by a spin ( usually of magnitude @xmath50 ) and the spins
interact via the hafm exchange interaction hamiltonian ( eq .
( 1 ) ) .
the n.n .
intra - chain and rung exchange interactions are of strengths @xmath7 and @xmath71 respectively .
when @xmath72 one obtains two decoupled afm spin chains for which the excitation spectrum is known to be gapless . for all @xmath73
a gap opens up in the spin excitation spectrum@xmath74 the result is easy to understand in the simple limit in which the exchange coupling @xmath71 along the rungs is much stronger than the coupling along the chains .
the intra - chain coupling in this case may be treated as a perturbation . when @xmath75 the exact ground state consists of singlets along the rungs .
the ground state energy is @xmath76 , where @xmath77 is the number of rungs in the ladder .
the ground state has total spin @xmath20 in first order perturbation theory , the correction to the ground state energy is zero .
an @xmath15 excitation may be created in the ladder by promoting one of the rung singlets to the @xmath15 triplet state .
the weak coupling along the chains gives rise to a propagating @xmath15 magnon . in first order perturbation theory ,
the dispersion relation is @xmath78 where @xmath79 is the momentum wave vector .
the sg is given by @xmath80 the two - spin correlations decay exponentially along the chains showing that the ground state is a qsl . the family of compounds @xmath81 consists of planes of weakly - coupled ladders of @xmath82 chains@xmath83 for @xmath84 and @xmath85 one gets the two - chain and three - chain ladder compounds @xmath86 and @xmath87 respectively .
the first compound is a sg afm while the second compound has properties similar to those of the hafm hamiltonian in @xmath88 , which has a gapless excitation spectrum .
the experimental evidence is consistent with the theoretical prediction that in an @xmath70- chain ladder , the excitation spectrum is gapped ( gapless ) when @xmath70 is even ( odd)@xmath74 bose and gayen@xmath89 have studied a two - chain ladder model with frustrating diagonal couplings .
the intra - chain and diagonal couplings are of equal strength @xmath7 . for @xmath90 the exact ground state
consists of singlets along the rungs with the energy @xmath91 an excitation can be created by replacing one of the singlets by a triplet .
the triplet excitation is localised and separated by an energy gap from the ground state .
xian@xmath92 later pointed out that as long as @xmath93 the rung dimer state is the exact ground state .
ladders provide a bridge between @xmath88 and @xmath94 many body systems and are ideally suited to study how the electronic and magnetic properties change as one goes from a single chain to the square lattice limit .
the significant interest in @xmath94 many body systems is due to the unconventional properties of the @xmath95 panes in doped cuprate systems .
the latter exhibit high temperature superconductivity in appropriate ranges of dopant concentration .
many of the unusual properties of the cuprate systems arise due to strong correlation effects .
ladders are simpler systems in which some of the issues related to strong correlation can be addressed in a rigorous manner .
doped ladder models are toy models of strongly correlated systems . in these systems ,
strong coulomb correlations prohibit the double occupancy of a site by two electrons , one with spin up and the other with spin down . in a doped spin system ,
two processes are in competition : hole delocalization and exchange energy minimization .
the latter is minimised in the antiferromagnetically ordered nel - type of state .
a hole moving in such a background gives rise to parallel spin pairs which raise the exchange interaction energy of the system .
the questions of interest are : whether a coherent motion of the holes is possible , whether two holes can form a bound state and superconducting ( sc ) pairing correlations develop etc .
these issues are of significant relevance in the context of doped cuprate systems in which charge transport occurs through the motion of holes . in the sc phase ,
the holes form bound pairs with possibly @xmath96wave symmetry .
several proposals have been made so far on the origins of hole binding but the actual binding mechanism is still controversial@xmath97 the doped cuprate systems exist in a ` pseudogap ' phase before entering the sc phase .
in fact , some of the cuprate systems also exhibit a sg .
the two - chain afm ladder systems are sg systems and it is of interest to study how the gap evolves on doping . the possibility of binding of hole pairs in a two - chain ladder system was first pointed out by dagotto et al.@xmath98 in this case , the binding mechanism can be understood in a simple physical picture .
again , consider the case @xmath99 i.e. , a ladder with dominant exchange interactions along the rungs . in the ground state ,
the rungs are mostly in singlet spin configurations . on the introduction of a single hole ,
a singlet spin pair is broken and the corresponding exchange interaction energy is lost . when two holes are present , they prefer to be on the same rung to minimise the loss in the exchange interaction energy .
the holes thus form a bound pair . in the more general case ,
detailed energy considerations show that the two holes tend to be close to each other effectively forming a bound pair . for more than two holes ,
several calculations suggest that considerable sc pairing correlations develop in the system .
the superconducting state can be achieved only in the bulk limit .
theoretical predictions motivated the search for ladder compounds which can be doped wth holes . much excitement was created in 1996 when the doped ladder compound @xmath100 was found to become sc under pressure at @xmath101 the transition temperature @xmath102 is @xmath103 at a pressure of @xmath104 as in the case of cuprate systems , bound pairs of holes are responsible for charge transport in the sc phase .
experimental results on doped ladder compounds point out strong analogies between the doped ladder and cuprate systems@xmath105 bose and gayen@xmath106 have derived exact , analytical results for the ground state energy and the low - lying excitation spectrum of the frustrated @xmath107 ladder model doped with one and two holes .
the undoped frustrated ladder model has already been described .
the @xmath107 hamiltonian describing the ladder is given by @xmath108 the @xmath109 and @xmath110 are the electron creation and annihilation operators which act in the reduced hilbert space ( no double occupancy of sites ) , @xmath111 where @xmath112 is the spin index and @xmath113 are the occupation numbers of the @xmath5th and @xmath114th sites .
the first term in eq .
( 8) describes the motion of holes .
the hopping integral @xmath115 has the value @xmath116 for hole motion along the rungs and the value @xmath117 for both the intra - chain n.n . and diagonal hops .
the latter assumption is crucial for the exact solvability of the eigenvalue problem in the one and two hole sectors . though the model differs from the standard @xmath107 ladder model@xmath118 ( diagonal couplings missing in the latter model ) , the spin and charge excitation spectra exhibit similar features . for the frustrated @xmath107 ladder ,
the dispersion relation of the two - hole bound state branch can be obtained exactly and analytically .
the exact two - hole ground state is a bound state with centre of mass momentum wave vector @xmath119 and @xmath120-wave type symmetry .
the ladder exists in the luther - emery phase in which the spin excitation is gapped and the charge excitation gapless .
there is no spin - charge separation , a feature associated with luttinger liquids in which both the spin and charge excitations are gapless . in the exact hole eigenstates ,
the hole is always accompanied by a spin-@xmath121 the hole - hole correlation function can also be calculated exactly . when @xmath99 the holes of a bound pair are predominantly on the same rung .
for lower values of @xmath122 the holes prefer to be on n.n .
rungs so that energy gain through the delocalization of a hole along a rung is possible .
the novel phenomena exhibited by undoped and doped spin ladders have motivated a large number of theoretical studies .
several ladder compounds have been discovered / synthesized to date .
detailed information on the theoretical and experimental investigations may be obtained from two exhaustive reviews on ladders@xmath123 * haldane gap antiferromagnets * we have so far been discussing sg systems with half - odd integer spins . from the lieb - schultz - mattis ( lsm)@xmath124theorem one can show that the half - odd integer spin hafm chain has a gapless excitation spectrum in the infinite chain length limit . the theorem does not extend to integer spin chains .
haldane@xmath125 based on his analysis of the nonlinear @xmath112 model mapping of the large @xmath126 hafm hamiltonian in 1d , conjectured that the hafm spin chains with integer spins have a gap in the excitation spectrum , i.e. , are sg afms .
the ground state of an integer spin chain is disordered and the spin - spin correlation function has an exponential decay .
haldane s conjecture has now been verified both theoretically and experimentally@xmath127 the spin-@xmath128 hafm hamiltonian has the same ground state features as the spin-@xmath128 affleck , kennedy , lieb , tasaki ( aklt ) hamiltonian for which the so - called valence bond solid ( vbs ) state is the exact ground state .
consider a @xmath88 lattice each site of which is occupied by a spin-@xmath129 a spin-@xmath128 can be considered to be a symmetric combination of two spin-@xmath50 s . in the vbs state ,
each spin-@xmath50 component of a spin-@xmath128 forms a singlet ( vb ) with a spin-@xmath50 at a neighbouring site .
the aklt hamiltonian is a sum over projection operators onto spin @xmath130 for successive pairs of spins , i.e. , @xmath131 the presence of a vb between each neighbouring pair of sites implies that the total spin of each pair of spins can not be @xmath132 thus @xmath133 acting on @xmath134 , the wave function of the vbs state , gives zero .
the eigenvalues of the projection operator being positive , @xmath134 is the ground state of @xmath133 with eigenvalue zero . the vbs state is spin - disordered with an exponentially decaying spin - spin correlation function .
one can , however , define a non - local string order parameter which has a non - zero expectation value in the vbs state@xmath135 the explicit form of the aklt hamiltonian is given by @xmath136\label{11}\ ] ] the excitation spectrum of @xmath133 can not be determined exactly .
variational calculations show the existence of a gap in the excitation spectrum .
several haldane gap ( hg ) afms have been discovered so far .
these include the @xmath15 compound @xmath137@xmath138 @xmath139@xmath138 @xmath140@xmath141 and the @xmath142 compound ( @xmath143bipyridine ) trichloromanganese ( iii ) , @xmath144 ( bipyridine)@xmath145 experiments carried out on these compounds show that the vbs state provides the correct physical picture for the true ground state@xmath127 the doped spin-@xmath128 compound @xmath146 provides an example of a qsl in @xmath147 the parent compound @xmath140is a charge transfer insulator containing @xmath148 chains .
the ground state of the system is spin - disordered and the spin excitation spectrum is separated by the hg from the ground state .
the compound is doped with holes on replacing the off - chain @xmath149 ions by @xmath150ions .
the holes mostly appear in oxygen orbitals along the @xmath151 chains .
there is no evidence of metal - insulator transition but the dc - resistivity @xmath152 falls by several orders of magnitude .
this indicates that the holes are not fully mobile but delocalized over several lattice spacings .
inelastic neutron scattering ( ins ) experiments reveal the existence of new states within the hg .
several studies have been carried out so far to explain the origin of the sub - gap states@xmath153 a recent neutron scattering experiment@xmath154 provides evidence for an incommensurate double - peaked structure factor @xmath155 for the sub - gap states .
the ins intensity is proportional to the structure factor . for the pure compound
, the structure factor @xmath156 near the gap energy of @xmath157 mev , has a single peak at the wave vector @xmath158 indicative of afm correlations . for the doped compound
, @xmath155 has an incommensurate double - peaked structure factor , for energy transfer @xmath159 mev , with the peaks located at @xmath160 the shift @xmath161 is found to have a very weak dependence on the impurity concentration @xmath162 for @xmath162 in the range @xmath163.$ ] evidence of incommensurate peaks has also been obtained in the underdoped metallic cuprates .
the peaks are four in number and occur at @xmath164 and @xmath165 the crucial difference from the nickelate compound is that @xmath161 is proportional to the dopant concentration x. the incommensurability has been ascribed to inhomogeneous spin and charge ordering in the form of stripes@xmath97 malvezzi and dagotto@xmath166 have provided an explanation for the origin of spin incommensurability in the hole - doped @xmath15 nickelate compound .
they have shown that a mobile hole generates afm correlations between the spins located on both sides of the hole and this is responsible for the spin incommensurability seen in experiments .
xu et al.@xmath154 have given a different explanation for the origin of incommensurability .
the holes doped into the qsl ground state of the @xmath15 chain are located on the oxygen orbitals and carry spin .
they induce an effective ferromagnetic interaction between the @xmath167 spins on both sides .
the incommensurate peaks arise because of the spin density modulations developed around the holes with the size of the droplets controlled by the correlation length of the qsl .
bose and chattopadhyay@xmath168 have developed a microscopic theory of the origin of spin incommensurability in keeping with the suggestions of xu et al .
* quantum phase transitions * a quantum phase transition ( qpt ) occurs at @xmath169 and brings about a qualitative change in the ground state of an interacting many body system at a specific value @xmath170 of some tuning parameter @xmath171 examples of tuning variables include magnetic field @xmath172 pressure and dopant concentration . the origin of qpts lies in quantum fluctuations just as thermal fluctuations drive thermodynamic phase transitions . in the case of second order
thermodynamic phase transitions , the critical point is characterised by scale invariance and a divergent correlation length .
free energy and the different thermodynamic functions become non - analytic at the critical temperature @xmath173 the quantum critical point is also associated with scale invariance and a divergent correlation length with quantum fluctuations substituting for thermal fluctuations .
the ground state energy becomes non - analytic at the critical value @xmath170 of the tuning parameter .
if one of the phases is gapped , the gap goes to zero in a power - law fashion as @xmath174 quantum and thermal fluctuations are equally important in the so - called quantum critical regime extending into the finite-@xmath175 part of the @xmath175 versus @xmath176 phase diagram .
the macroscopic physical properties in this regime are in many cases independent of microscopic details .
a large number of theoretical and experimental studies has been carried out on qpts in condensed matter systems@xmath177 here we focus on a few specific examples of qpts in sg afms .
organic spin ladder compounds provide ideal testing grounds for theories of qpts .
consider the phase diagram of the afm two - chain spin ladder in the presence of a magnetic field @xmath178 at @xmath169 and for @xmath179 the ladder is in the sg phase . at @xmath180
there is a transition to the gapless luttinger liquid phase with @xmath181 where @xmath176 is the land splitting factor , @xmath182 the bohr magneton and @xmath183 the magnitude of the sg@xmath184 at an upper critical field @xmath185 there is another transition to the fully polarised ferromagnetic state .
both @xmath186and @xmath187 are quantum critical points .
the compound @xmath188 has been identified as a @xmath27 two - chain spin ladder in the strong coupling limit @xmath189 and @xmath190 with @xmath191 and @xmath192 the magnetization data exhibit universal scaling behaviour in the vicinity of @xmath186 and @xmath185 consistent with theoretical predictions . in the gapless regime
@xmath193 the ladder model can be mapped onto an xxz chain , the thermodynamic properties of which can be calculated exactly by the ba .
the theoretically computed magnetization @xmath194 versus magnetic field @xmath195 curve is in excellent agreement with the experimental data .
other organic ladder compounds exhibiting qpts are @xmath196@xmath197 and @xmath198 for inorganic spin ladder systems , the value of @xmath186 is too high to be experimentally accessible . the synthesis of organic ladder compounds has paved the way for experimental observation of qpts .
another notable example of experimentally observed qpts is that of field - induced @xmath199 magnetic ordering in low - dimensional sg afms , @xmath200@xmath201 and @xmath202@xmath203 which , as described earlier , consist of networks of dimers . the dimer ground state is a singlet for @xmath204
the triplet @xmath205 excitation is separated from the singlet ground state by an energy gap , @xmath206 application of an external magnetic field @xmath195 leads to zeeman splitting of the triplet excitation into three components : @xmath207 at a critical external magnetic field @xmath208 the lowest triplet component becomes energetically degenerate with the ground state .
this results in a qpt at @xmath209to a @xmath199 magnetically ordered state .
the critical point @xmath209 separates a gapped qsl state @xmath210 from a field - induced magnetically ordered state @xmath211 the triplet components can be regarded as diluted bosons and the qpt at @xmath209 can be treated as a bose - einstein condensation ( bec ) of low - lying magnons@xmath212 support for this idea comes from experimental findings on @xmath213 recently , the excitation spectrum in the magnetically ordered state of @xmath202 has been determined using ins@xmath214 the observed data are consistent with the theoretical prediction of a gapless goldstone mode characteristic of the bec . in the be condensed state ,
the state of each dimer is found to be a coherent superposition of the singlet and the @xmath215 triplet states .
the phase in the superposition specifies the orientation of the staggered magnetization in the plane transverse to the magnetic field direction .
the number of magnons in the condensed state is not , however , infinite as magnons can not occupy the same sites in a spin system due to a hard - core repulsion between them .
the interaction restricts the number of magnons to be large but finite .
recently , there is a resurgence of research interest in bec because of its experimental realization in ultracold gases of dilute atoms .
sg afms offer another testing ground for theories related to bec .
the definitive evidence of bec in the compound @xmath202 motivates the search for condensation phenomena in other sg systems .
* magnetization plateaus * the magnetization curve of a low-@xmath120 afm does not always show a smooth increase in magnetization , from zero value to saturation , as the magnetic field is increased in magnitude . in certain systems , the curve exhibits plateaus at certain rational values of the magnetization per site @xmath216 the phenomenon is analogous to the quantum hall effect in which electrical resistivity exhibits plateaus as a function of the external magnetic field .
oshikawa , yamanaka and affleck ( oya ) derived a condition for the occurrence of magnetization plateaus in quasi-@xmath88 afm systems by generalising the lsm theorem to include an external magnetic field@xmath217 for general spin systems , the quantization condition can be written as @xmath218 where @xmath219 @xmath70 being the number of spins of magnitude @xmath126 in unit period of the ground state and @xmath220 is the magnetization associated with the unit cell .
the quantization condition is necessary but not sufficient as not all plateaus predicted by the condition exist in general .
hida@xmath221 first predicted the existence of a magnetization plateau at @xmath222 in the magnetization curve of a @xmath27 afm chain with period @xmath223 exchange coupling .
one can readily check that the quantization condition @xmath224 is obeyed as in this case @xmath225 and @xmath226 @xmath227 a gapped phase is essential for the appearance of a plateau ( magnetization unchanging ) in the @xmath47 versus @xmath195 curve .
more than one plateau can occur if there is more than one gapped phase as @xmath195 is changed .
high - field measurements reveal the existence of magnetization plateaus in several afm compounds .
the @xmath27 material @xmath228 exhibits plateaus at @xmath229 and @xmath230 @xmath231 the @xmath27 sg afm @xmath67 is the first example of a @xmath94 spin system for which magnetization plateaus have been observed experimentally .
the plateaus are obtained for @xmath232 and @xmath233 momoi and totsuka@xmath234 have suggested that the appearance of plateaus in @xmath67 is due to a transition from a superfluid to a mott insulating state of magnetic excitations . as pointed out earlier ,
the triplet excitations in the ss model , which describes @xmath235 are almost localized . in the presence of a magnetic field and at special values of the magnetization , the triplet excitations localize into a superlattice structure to minimize energy so that the magnetization remains constant . as in the cases of @xmath200 and @xmath236
the compound @xmath67 is a network of coupled dimers .
a triplet excitation created on a dimer can propagate to a neighbouring dimer due to the inter - dimer exchange interaction .
the delocalization of triplets is similar to that of electrons in crystals . in the presence of a magnetic field ,
the triplet excitation is split into three components with the @xmath215 component being the lowest in energy .
the @xmath215 excitations can be regarded as bosons with a hard core repulsion .
the repulsive interaction arises from the @xmath237-component of the exchange interaction and disallows the occupation of a single dimer by more than one boson .
the @xmath238-part of the exchange interaction is responsible for the hopping of the triplet excitation to neighbouring dimers .
one thus has a system of interacting bosons in which itinerancy competes with localization .
the transition from itinerancy to localization is analogous to the mott metal - insulator transition in electronic systems .
if repulsive interactions dominate , the triplet excitations ( bosons ) localize to form a superlattice .
a direct measurement of the magnetic superlattice in @xmath67 has been made by kodama et al.@xmath234 using a high - field nmr facility .
the superlattice corresponds to @xmath239 which requires a high magnetic field strength of @xmath240 @xmath175 for its observation .
superlattice structures for higher @xmath194 values have not been detected as yet because of the requirement of very high magnetic fields .
* summary and future outlook * sg afms exhibit a variety of novel phenomena the observation and interpretation of which have been possible due to the intimate links between theory and experiments .
rigorous theories and exactly solvable models have predicted phenomena which were later verified experimentally .
quantum magnetism is one of the few reseach areas in which rigorous theories go hand in hand with experimental initiatives .
we have illustrated this interdependence through a few chosen examples .
the mg model illustrates the origin of sg due to the presence of further - neighbour interactions .
the ss and @xmath57 models extend the @xmath88 model to @xmath241 experimental realization of the ss model became a reality about twenty years after the original theoretical proposal was made .
the aklt model provides the correct physical picture of the grounds states of integer spin chains .
knowledge of the ground state has given rise to testable predictions which were later verified experimentally .
spin ladders were originally studied as toy models of strongly correlated systems .
theories of undoped and doped ladder models motivated the search for real materials which led to success in several instances .
the oya theorem provides the condition for the existence of magnetization plateaus which is in agreement with experimental results .
sg afms further exhibit novel quantum phenomena which illustrate how quantum effects influence ground and excited state properties .
we have discussed only a few of these in the present review .
two sg afm systems we have not described include the compounds @xmath242@xmath243 and the tellurate materials @xmath244@xmath245 or @xmath246the first compound is defined on a @xmath247-depleted square lattice which is a network of four - spin plaquettes connected by single links ( dimer bonds ) .
each square plaquette is in a resonating valence bond ( rvb ) spin configuration in the ground state . in this case , the rvb state is a linear superposition of two vb configurations . in one configuration ,
the vbs ( singlets ) occupy the horizontal links of the square plaquette . in the other configuration ,
the vbs occupy the vertical links .
more generally , a rvb state is a coherent linear superposition of vb states and is a well - known example of a qsl .
the tellurate materials can be described as networks of spin tetrahedra . in both @xmath248 and the tellurate materials ,
the existence of singlet excitations in the spin gap has been reported as in the case of the kagom lattice hafm .
the qpts considered in this review are brought about by tuning magnetic field strengths .
qpts can also occur by the tuning of exchange interaction strengths . in the case of the frustrated two - chain spin ladder model described earlier
, a first order qpt takes place at @xmath249 from the rung dimer state to the haldane phase of the @xmath15 chain@xmath250 kolezhuk and mikeska@xmath251 have constructed generalised @xmath27 two - chain ladder models with two - spin and four - spin exchange couplings for which the ground state can be determined exactly .
qpts to the various phases are obtained by varying the exchange interaction strengths .
a lattice of coupled two - chain ladders provides another example of a qpt which is brought about by tuning the inter - ladder exchange interaction strength @xmath252 for @xmath253 the spin lattice is in the sg phase . at @xmath254 a qpt to a long range
magnetically ordered state occurs .
most of the spin models considered in this review exhibit qpts at specific values of the exchange interaction strengths .
one might think that experimental observation of such qpts is not possible as exchange interaction strengths can not be changed at will .
there is now reason to believe that exchange interaction strengths can be controlled .
recent observations@xmath255 of the superfluid to mott insulator transition in a system of ultracold atoms defined in an optical lattice open up the exciting possibility of investigating phenomena associated with interacting many body systems in a controllable environment .
the optical lattice is generated as a light - wave interference pattern using several criss - crossing laser beams .
the lattice is equivalent to an energy landscape of mountains and valleys which can provide the confining potential to trap individual atoms in separate valleys .
proposals for the implementation of spin hamiltonians in optical lattices have been put forward with the aim to study the exotic quantum phases of interacting spin systems@xmath256 the exchange interaction between spins belonging to atoms in neighbouring valleys can be modified by controlling the intensity , frequency and polarization of the trapping light .
spin hamiltonians of interest can be engineered through the designing of appropriate optical lattice geometries .
practical implementation of some of these ideas may be possible in the not too distant future .
spin systems have recently been suggested as candidates for the realization of quantum computation and communication protocols@xmath257the spin systems considered so far include some sg afms like the mg and hg chains , the two - chain spin ladder etc.@xmath258 again , one anticipates intense research activity in the coming years on such problems . to sum up ,
the richness and vitality of the subject of sg afms are evident in the challenging problems of current research interest and also in the opening up of new avenues of research . |
the lowest - level quantum - kinematics of an extended many - electron system ( within the born - oppenheimer approximation ) is contained in 1- and 2-body densities and reduced density matrices ( rdms ) : @xmath18= electron density , @xmath19= 1-body rdm ( 1-matrix ) , @xmath20= pair density ( pd ) , @xmath21= 2-matrix . the well - tried and widely used density functional theory ( dft ) yields @xmath22 ( but not @xmath23 and @xmath24 ) .
if a density - matrix functional theory would exist , it would yield @xmath23 ( but not @xmath24 ) ; if a pair - density functional theory would exist , it would yield @xmath24 ( but not @xmath23 ) .
if an effective 2-body scheme would exist , which yields approximately the 2-matrix @xmath25 , then from it would follow both the pd @xmath24 ( by taking the diagonal elements ) and the 1-matrix @xmath23 ( by means of the contraction procedure @xmath26 and @xmath27 ) . for extended systems this requires the cumulant decomposition of @xmath25 . in this paper ( it continues from @xcite and
is related to @xcite ) , the * cumulant 2-matrix * @xmath0 ( dimensionless : @xmath28 ) is developed as a decisive key quantity and its properties are studied for a `` simple '' model system , namely the high - density spin - unpolarized homogeneous electron gas ( heg ) .
+ although not present in the periodic table , this heg is still an important and so far unsolved model system for the electronic structure theory , cf e.g. @xcite . its advantage : pure `` correlation '' and no `` multiple - scattering '' . in its spin - unpolarized version , the heg ground state ( gs )
is characterized by only one parameter @xmath29 , such that a sphere with this * wigner - seitz radius * @xmath29 contains _ on average _ one electron @xcite .
it determines the fermi wave number as @xmath30 in atomic units ( a.u . ) with @xmath31^{1/3}\approx 0.521062 $ ] and it measures simultaneously both the interaction strength and the homogeneous electron density @xmath4 , such that high density corresponds to weak interaction and hence weak correlation @xcite . for recent papers on this limit cf @xcite .
@xmath32 corresponds to the ideal fermi gas .
its gs energy per particle is @xmath33 , measured in units of @xmath34 .
the coulomb repulsion @xmath35 with @xmath36 causes deviations .
this is in lowest order the exchange energy @xmath37 .
but in next order the long range of the coulomb repulsion makes the correlation energy @xmath38 , defined by @xmath39 , to behave non - analytically at the high - density limit : @xmath40 $ ] @xcite .
this non - analytical behavior of @xmath41 carries over to the kinetic and potential components , @xmath42 respectively @xmath43 , through the virial theorem after march ( 1958)@xcite : @xmath44 the energy components @xmath42 and @xmath43 follow from the simplest quantum - kinematical quantities , namely @xmath1 and @xmath2 , the momentum distribution ( md ) and the ( static ) structure factor ( sf ) , respectively : @xmath45=c\ , \ ] ] @xmath46\frac{q_c^2}{q^2}\ , \quad s(0)=0\ , \quad \int\limits_0^\infty d(q^3)\ [ 1-s(q)]=1+c_1\ .
\ ] ] @xmath47 and @xmath3 are measured in units of @xmath48 and @xmath49 in units of @xmath50 .
the quantity @xmath51 is hereforth referred to as lwdin parameter , because lwdin was the first one who queried the meaning of the trace of the squared 1-matrix . because of @xmath52 , @xmath51 measures the @xmath29-dependent non - idempotency of @xmath1 , being zero for @xmath32 ( ideal fermi gas ) and increasing with @xmath29 .
thus @xmath51 measures simultaneously the strength of correlation .
note its particle - hole symmetry .
@xmath53 is another ( short - range ) correlation parameter , also vanishing for @xmath54 .
+ describing * electron pairs * , one has to distinguish sfs @xmath55 for pairs with antiparallel spin ( `` a '' ) and with parallel spins ( `` p '' ) and corresponding pair densities ( pds ) @xmath56 , where @xmath57 and @xmath58 describe the coulomb hole and the fermi hole , respectively .
the @xmath56 and @xmath55 are mutually related through fourier transform ( ft ) : @xmath59\ , \\
g_{\rm p}(r)-1&=&\int\limits_0^\infty d(q^3)\frac{\sin qr}{qr}\ [ s_{\rm p}(q)-1 ] \quad \leftrightarrow \quad s_{\rm p}(q)-1= \alpha^3\int\limits_0^\infty d(r^3)\frac{\sin qr}{qr}\ \frac{1}{2}[g_{\rm p}(r)-1]\ . \nonumber \end{aligned}\ ] ] from this follow the `` total '' pd @xmath60/2 $ ] and the `` total '' sf @xmath61 : @xmath62 \quad \leftrightarrow \quad s(q)-1= \alpha^3\int\limits_0^\infty d(r^3)\frac{\sin qr}{qr}\ [ \ g(r)-1]\ . \\\ ] ] @xmath63 fix the normalizations of @xmath64 , known as the perfect screening sr , whereas the coalescing ( or on - top ) values @xmath65 and @xmath66 fix the normalizations of @xmath67 and @xmath68 , respectively .
@xmath69 or @xmath70 shows the physical meaning of @xmath53 .
it determines the short - range correlation in terms of @xmath65 , the on - top value of the coulomb hole @xmath57 with @xmath71 . for further details ( normalizations ) cf ( [ b1 ] ) , ( [ b2 ] ) .
+ in view of eqs.([a2 ] ) and ( [ a3 ] ) , one may ask which peculiarities of @xmath1 and @xmath55 cause the non - analyticities of @xmath42 and @xmath43 , respectively . as shown in @xcite , the above - mentioned drastic changes ,
when switching on the long - range coulomb interaction , show up in the redistribution of the non - interacting md @xmath72 within thin layers ( thickness @xmath73 ) inside and outside the fermi surface @xmath17 and with a remaining finite jump discontinuity at @xmath17 : @xmath74 , @xmath75 .
they show up also in the behavior of @xmath2 within a small spherical region ( radius @xmath73 ) around the origin of the reciprocal space with the plasmon sr @xmath76 @xcite , which causes an inflexion point @xmath77 , where @xmath78 is the plasma frequency , measured in units of @xmath34 .
the mentioned non - analyticities and redistributions let ordinary perturbation theory fail , e.g. with @xmath79 and @xmath80 with @xmath72 and @xmath81 . that the gs energy @xmath82 diverges in 2nd order ,
has been shown by heisenberg ( 1947 ) @xcite .
macke ( 1950 ) @xcite has repaired this failure by means of an appropriate , physically plausible * partial summation of higher - order perturbation terms * for the gs energy per particle @xmath82 , a way which has been developed further by gell - mann and brueckner ( 1957 ) @xcite .
this so - called ring - diagram summation , cf fig.1 , is also known as the random phase approximation ( rpa ) .
this summation has been developed for @xmath1 by daniel / vosko ( 1960 ) @xcite and kulik ( 1961 ) @xcite , an analytical extrapolation for @xmath1 is given in @xcite , for the spin - polarized case see @xcite , recent quantum monte carlo calculations of @xmath1 for @xmath83 are in @xcite , @xmath84 for high - densities is in @xcite , @xmath84 calculated for @xmath85 is in @xcite .
the ring - diagram summation for @xmath2 has been done by glick / ferrell ( 1960 ) @xcite , geldart ( 1967 ) @xcite , and kimball ( 1976 ) @xcite . for attempts to go beyond rpa cf eg @xcite .
the on - top behavior of @xmath86 ( describing the short - range correlation ) influence the large-@xmath3 behavior of @xmath55 and @xmath1 .
this comes from theorems referred here as the coalescing cusp and curvature theorems , cf @xcite and @xcite and app.b .
+ with @xmath1 also @xmath42 is available ( but not @xmath43 ) and with @xmath2 it is available @xmath43 ( but not @xmath42 ) .
does a quantity exist which contains both @xmath1 and @xmath2 ? `` old '' and recent attempts in that direction ( with variational or direct calculations of the 2-matrix ) are in @xcite . in @xcite it has been shown that the self - energy @xmath87 as a _ functional _ of @xmath88 and @xmath89 is such a quantity @xcite . here
it is shown - using the concept of reduced density matrices ( rdms ) - that @xmath1 and @xmath2 have their common origin in a quantity called cumulant 2-matrix , @xmath90 with the symbolic variable @xmath91 .
this matrix appears if one tries to represent the 2-matrix of an interacting system in terms of the 1-matrix ( according to an independent - particle or hartree - fock model ) .
there remains an irreducible part , which can not be reduced due to correlation .
for the pd and for the sf , these * cumulant decompositions * are @xmath92 @xmath93 note the asymmetry with respect to `` a '' ( coulomb hole ) and `` p '' ( fermi hole ) . with the `` total '' cumulants @xmath94/2 $ ] and @xmath95 it follows
@xmath96 @xmath5 is the dimensionless 1-matrix [ following from @xmath1 , cf ( 2.4 ) ] and @xmath15 - referred to as hf - function - is the ft of @xmath16 , cf ( [ b15 ] ) .
@xmath97 are the cumulant pds ( cpds ) and their fts are the cumulant sfs ( csfs ) @xmath98 : @xmath99 @xmath100 within rdm theory one can show that the cumulant 2-matrix @xmath90 is the source for : ( i ) the csf @xmath9 [ or equivalently the cpd @xmath7 ] and ( ii ) also for the md @xmath1 .
whereas ( i ) results simply from @xmath102 , i.e. taking the diagonal elements @xmath103 , @xmath26 , the case ( ii ) follows from the slightly more complicated rdm contraction of @xmath90 with @xmath26 , @xmath104 and ft .
the eqs.([a9 ] ) and ( [ a10 ] ) contain the csf properties ( [ b1 ] ) and ( [ b2 ] ) needed below .
+ general remarks on * rdms and cumulants * : whereas finite many - body systems can be described by many - body wave functions @xmath105 as solutions of a schrdinger equation , extended systems have to be described by a hierarchy of @xmath106-representable rdms as solutions of the bbgky - like hierarchy of contracted schrdinger equations . and : whereas for finite systems the concept of cumulant ( 2-body , 3-body , ... ) matrices _ can _ be applied , for extended systems it is a _ must _ , because - unlike the rdms - all these cumulant matrices are _ size - extensive _ entities , what allows the thermodynamic limit with @xmath107 , @xmath108 , @xmath109 .
closely related with this property is that they are given within perturbation theory by non - vacuum _ linked _
feynman diagrams . in app.a , the systematic definition of cumulant matrices in terms of generating functionals is summarized .
recent ( quantum chemical ) papers on rdms , cumulants , contracted schrdinger equations , correlation strength , correlation entropy , quantum entanglement , berry phases etc . are @xcite-@xcite and refs . therein
@xcite and @xcite deal ( for finite systems ) with variational calculations of the 2-matrix .
recent heg papers are @xcite and @xcite and refs . within . + here
is roughly sketched , what will be presented in secs .
ii , iii , iv in more detail . in this paper
it is aimed to present @xmath28 in the rpa with all those terms of @xmath28 contributing to the correlation energy @xmath38 terms up to @xmath110 and @xmath111 .
the correctness of this @xmath28 is controlled / checked by deriving @xmath1 and @xmath2 .
a comparison is performed between these new rpa results with the old ones of daniel / vosko @xcite , kulik @xcite , and kimball @xcite , respectively . for this purpose
the diagrams of figs.1 - 5 are needed .
fig.1 shows the yukawa - like screening of the bare ( long - range ) coulomb repulsion @xmath35 , replacing it by the frequency - dependent interaction @xmath112 $ ] , where @xmath113 is the particle - hole propagator ( [ c3 ] ) .
the bare coulomb repulsion shows up in divergencies : for the correlation energy @xmath38 see @xcite , for @xmath1 and @xmath2 see @xcite .
these divergencies are eliminated through macke s partial summation of diagrams .
fig.2a shows the interaction of 2 particle - hole lines running from @xmath114 to 1 and from @xmath115 to 2 .
this is called `` d '' = direct diagram to distinguish it from `` x '' = exchange diagram of fig.3a with one line running from @xmath114 to 2 and another one from @xmath115 to 1 . these lowest - order rpa terms of figs .
2a and 3a are called @xmath116 and @xmath117 , respectively , `` r '' = ring - diagram summation of fig .
the cumulant matrices @xmath118 yield the cpds @xmath119 and the csfs @xmath120 of figs .
2b and 3b and the interaction energies of figs . 2c and 3c . but how to obtain @xmath121 ? using the self - energy @xmath122 ,
the lowest - order @xmath123 - diagrams are in figs.4c and 5c : @xmath124 .
of course , rdm - theory must end up with the same results .
indeed , @xmath125 of fig.4c comes from @xmath117 of fig.3a by the contraction @xmath26 and @xmath126 , indicated in fig.4b by a small circle with the final result @xmath125 of fig.4c .
similarly , @xmath127 of fig.5c comes from the ladder diagram ( in its exchange version ) of fig.5a , contraction makes fig.5b with the final result @xmath127 of fig.5c .
contraction in terms of diagrams mean to transform a 2-line diagram into a 1-line - diagram .
this study opens the door for variational 2-matrix calculations and for discussing heg in terms of cumulant geminals and their weights .
+ the outline of the paper is as follows . in sec.ii the basic definitions of rdms , their cumulant decompositions and
the contraction sr are presented .
it is derived in detail how the 2-body quantities @xmath2 and @xmath6 with its cumulant pedants @xmath9 and @xmath7 , together with the 1-body quantity @xmath1 , follow from the cumulant 2-body matrix @xmath28 .
the decisive contraction procedure is in eqs.([b25])-([b28 ] ) . in secs.iii and iv the lowest - order cumulant 2-matrices and what follows from them
are presented .
sec.v gives a summary and an outlook .
app.a explains the exponential - linked - diagram theorem as the base for a systematic cumulant decompositions .
app.b deals with sf properties in terms of the coalescing cusp and curvature theorems , of the plasmon sr , and of the inflexion - point trajectory .
table 1 and 2 contain the small- and large @xmath3-behavior of the csfs and sfs ( similar as in @xcite ) .
in app.c the particle - hole propagator , the macke function @xmath128 and contour integrations are described .
app.d lists characteristic correlation parameters , vanishing for @xmath54 .
* 2.1 the 1-matrix and the 2-matrix : definitions and properties * + the starting point is the definition of the 1-matrix @xmath23 and of the 2-matrix @xmath25 for the heg - gs in terms of creation and annihilation operators , @xmath129 and @xmath130 , respectively : @xmath131 the shorthands @xmath132 and @xmath133 are used .
the hermiticity is obvious and fermi symmetry means @xmath134 if 1 and 2 or 1 and 2 are interchanged .
@xmath135 is the total particle - number operator .
the bra vector @xmath136 as well as the ket vector `` @xmath137 '' are @xmath106-body states , therefore @xmath138 and @xmath139 are @xmath140-body states , thus @xmath141 holds for the normalization .
the contraction sr @xmath142 describes how the 1-matrix @xmath23 for finite @xmath106 results from the 2-matrix @xmath25 .
+ it follows the spin structure of @xmath23 , the dimensionless 1-matrix @xmath5 and its ft : @xmath143 @xmath144 , @xmath145 ( with @xmath146 and @xmath147 for @xmath32 ) .
its natural orbitals are plane waves , because the system is extended and homogeneous .
@xmath1 is the momentum distribution ( md ) , @xmath148 and @xmath149 are measured in units of @xmath48 and @xmath50 , respectively .
+ the symmetric and antisymmetric 2-body spin functions @xmath150 enter the spin structure of @xmath21 as @xmath151 @xcite .
it defines - together with the basic definition ( [ b1 ] ) - the singlet 2-matrix @xmath152 and the triplet 2-matrix @xmath153 , being ( in space ) symmetric and antisymmetric for the interchanges @xmath154 or @xmath155 , respectively . the normalizations @xmath156 follow from the contractions @xmath157 which are in their turn consequences of the total normalization ( [ b2 ] ) and the singlet - triplet representation ( [ b6 ] ) .
eqs.([b7 ] ) and ( [ b8 ] ) show the necessity to find an equivalent writing for @xmath25 , which allows the thermodynamic limit with @xmath158 , @xmath108 , @xmath109 .
this is just done by means of an additive decomposition of @xmath25 in a hartee - fock like term @xmath159 , which is defined in terms of the ( full , exact ) 1-matrix @xmath23 only , and a non - reducible remainder - the cumulant 2-matrix @xmath0 .
thus @xmath160 as the first step of a more general expansion , see app.a . +
* 2.2 the hf approximation * + the 2-matrix in a hartree - fock like approximation ( in terms of the exact 1-matrix only ) is @xmath161 this may by also addressed `` exact exchange '' . using eq.(2.4 ) , the same structure ( [ b6 ] ) as for the total 2-matrix @xmath21 results : @xmath162 with @xmath163\ .\end{aligned}\ ] ] their normalizations @xmath164 contain the above introduced lwdin parameter @xmath51 , which is originally defined by @xmath165 note @xmath166 and note that the homogeneous density @xmath4 , measured in units of @xmath167 , equals @xmath168 and that @xmath169 , @xmath170^{1/3}$ ] .
+ the contractions of @xmath171 are @xmath172 where @xmath173\ , \int d^3r\ c({\mbox{{\boldmath}${r}$}}|{\mbox{{\boldmath}${r}$}})=nc\ .\ ] ] @xmath174 and @xmath175 are referred to as `` lwdin function '' and `` lwdin matrix '' , respectively .
[ @xmath174 appears in @xcite as @xmath176 .
+ * 2.3 the cumulant decomposition and cumulant geminals * + with this prelude , the above mentioned decomposition @xmath177 defines the cumulant 2-matrix @xmath0 .
its spin - structure is similar as in ( [ b6 ] ) and ( [ b10 ] ) @xmath178 and it defines ( through @xmath179 ) the cumulant 2-matrices @xmath180 with the normalization @xmath181 following from the normalizations ( [ b7 ] ) and ( [ b12 ] ) .
this ensures the normalization of @xmath0 to be @xmath182 , as it should . with the help of eqs.([b8 ] ) and ( [ b14 ] ) the contractions of @xmath180 are @xmath183 eqs.([b18 ] ) and ( [ b19 ] ) show that the @xmath184 are size - extensively normalized and contracted . with @xmath185 , the dimensionless
cumulant 2-matrices @xmath186 are introduced .
their normalization follows from ( [ b18 ] ) as @xmath187 the introduction of the singlet / triplet matrices @xmath180 and @xmath188 , is only a necessary intermediate step to present its spin - parallel / antiparallel components @xmath189 in terms of linked diagrams . for diagonal spins
@xmath190 it is @xmath191 .
so for antiparallel spins ( @xmath192 , hence @xmath193 ) , respectively parallel spins ( @xmath194 , hence @xmath195 ) the spin structure ( [ b17 ] ) leads to @xmath196\ , \ \chi_{\rm p}=\chi_- \quad \leftrightarrow \quad \chi_+=2\chi_{\rm a}-\chi_{\rm p}\ , \
\chi_-=\chi_{\rm p}\ .\ ] ] the arguments of these matrices are @xmath197 . their diagonals @xmath198 determine the normalizations @xmath199 the singlet / triplet matrices @xmath188 define symmetric and antisymmetric cumulant geminals @xmath200 with corresponding weights @xmath201 and with @xmath202 as total weights .
which properties do these geminals have ?
are they discrete ( bound ) or / and continuous ( scattering ) states , the latter with phase shifts ?
is there a kind of an aufbau principle for the @xmath201 ? + * 2.4 linked diagrams * + as mentioned above , the cumulant 2-matrices @xmath188 are given by linked diagrams .
to each symmetrized direct diagram @xmath203 with 2 open particle - hole lines ( one running from @xmath204 to @xmath205 and another one from @xmath206 to @xmath207 ) belongs an exchange diagram @xmath208 , such that for @xmath209 and @xmath210 there is only one line running from @xmath204 to @xmath205 ( like the diagrams of the self - energy @xmath123 , cf @xcite ) . with these building elements @xmath211 and @xmath212 the singlet / triplet components are @xmath213 and from ( [ b21 ] ) follow the decisive relations @xmath214 again with the arguments @xmath197 .
the spin - averaged sum @xmath28 can be written as @xmath215 \quad { \rm or } \quad \chi=\frac{1}{4}\chi_++\frac{3}{4}\chi_- \quad { \rm or } \quad \chi=\chi_{\rm d}-\frac{1}{2}\chi_{\rm x}\ .\ ] ] the a- and p - components are equally weighted , whereas the singlet- and triplet - components have the weights 1/4 and 3/4 , and the d - and x - components have the `` weights '' 1 and -1/2 , respectively . from eqs.([b23 ] ) and ( [ b24 ] ) follow corresponding relations for the pds @xmath86 and their cumulant correspondings @xmath216
. note that for their fts @xmath217 and @xmath218 it holds @xmath219 and @xmath220 .
+ now the question is : how real physical properties like the pd s @xmath56 ( by taking the _ diagonal _ elements ) as well as the md @xmath1 ( by fourier transforming the contracted _ off - diagonal _ elements ) follow from the ( more formal ) key quantities @xmath221 , respectively @xmath222 . + * 2.5 pair densities and structure factors * + with the cumulant pds @xmath223 and @xmath224 and their fourier transforms @xmath98 following from ( [ a9 ] ) and ( [ a10 ] ) , the pds @xmath56 and sfs @xmath55 are given by ( 1.6)-([a8 ] ) .
this shows that for the spin - parallel quantities , the functions @xmath16 and its ft @xmath15 are needed , cf app.b .
so from the cumulant 2-matrices @xmath225 follow not only the spin - antiparallel pd @xmath57 , but also the spin - parallel pd @xmath58 ( supposed the md @xmath1 has been calculated first , cf next paragraph )
. rigorous theorems for @xmath226 , @xmath227 , and @xmath228 are summarized in app.b . + * 2.6 contraction sum rule and momentum distribution * + in terms of the lwdin function @xmath229 $ ] and of the contracted cumulant matrix ( the notation @xmath230 means @xmath209 and @xmath231 ) @xmath232 the contraction sr ( [ b19 ] ) can be written as @xmath233 . if the expressions ( [ b21 ] ) and ( [ b23 ] ) are contracted and fourier transformed , then @xmath234=0\ , \quad\quad { \rm \textroundcap{$\chi$}}_{\rm x}(k)=\frac{1}{2}[-c(k)-c(k)]=-c(k)\ .\end{aligned}\ ] ] this is the contraction sr , which allows to calculate @xmath1 from @xmath235 supposed @xmath236 is approximately known .
the direct diagrams @xmath237 with 2 lines do not contribute to @xmath1 , which alone derives from the 1-line diagram @xmath238 .
if @xmath239 means the md of the ideal fermi gas , then @xmath80 defines its interaction induced correction @xmath240 for @xmath241 being normalized as @xmath242 .
thus the contraction sr ( [ b26 ] ) reads finally @xmath243 ^ 2={\rm \textroundcap{$\chi$}}_{\rm x}(k)\ \quad { \rm for } \quad k\gtrless 1\ , \ ] ] what simplifies as @xmath244 for small correlation tails @xmath121 , as this is the case for the rpa with @xmath245 and @xmath246 .
the ellipsis mean beyond - rpa terms . from a diagrammatical point of view
one may expect , that already @xmath238 is the md , but this is not the case : as derived above , a factor @xmath247 has to be added , to transform @xmath238 to @xmath121 .
so contraction means @xmath209 , @xmath248 _ and _ adding a factor sign@xmath249 . - from @xmath121 follows the lwdin parameter @xmath51 , which fixes some csfs at long wavelengths : @xmath250 , @xmath251 , and @xmath252 . + * 2.7 energies * + from @xmath121 follows the kinetic correlation energy @xmath253 . the same @xmath121 makes @xmath254 with ( [ b18 ] )
this allows to calculate the interaction energy @xmath255 , which is built up from the hf part @xmath256 and the cumulant part @xmath257 : @xmath258 @xmath259 is the hf energy in lowest order .
@xmath220 is the ft of @xmath260/2 $ ] , cf ( [ a9 ] ) and ( [ a10 ] ) .
the cumulant interaction energy @xmath261 is part of the interaction correlation energy @xmath262 .
thus , the energy - components @xmath42 and @xmath43 are available , defining the total energy @xmath263 as a function of @xmath29 with @xmath264 and @xmath265 .
the correlation energy is @xmath266 . +
* 2.8 summary of sec .
ii * + if @xmath267 is given in terms of diagrams @xmath268 with 2 not - closed particle - hole lines ( e.g. fig.2a ) and corresponding exchange terms @xmath269 ( e.g. figs.3a , 5a ) , then : + ( i ) it follows @xmath1 from the contraction sr ( 2.30 ) and from @xmath1 it follows @xmath42 via ( [ a2 ] ) .
+ ( ii ) @xmath15 and @xmath16 , which are needed for @xmath270 and @xmath58 , respectively , follow also from @xmath1 , cf app.b .
+ ( iii ) @xmath97 and @xmath98 and the cumulant interaction energy @xmath257 follow from @xmath221 . + (
iv ) finally , @xmath271 and @xmath272 .
+ the resulting quantities are the gs - energies @xmath273 as functions of @xmath29 and the md @xmath1 as well as the pds @xmath56 for the many - body quantum kinematics on the lowest level . +
in the following , these basic definitions and relations are applied to the spin - unpolarized heg in the high - density limit @xmath54 , as an illustrative example .
all linked - diagram ( direct and corresponding exchange ) terms are taken into account , which contribute in the energy to terms @xmath274 $ ] .
these diagrams are shown in figs.2a , 3a , 4a , 5a . + * 3.1 the direct building block @xmath275 in rpa * + @xmath116 denotes the direct cumulant 2-matrix @xmath275 in its ring - diagram approximation of fig.2a : @xmath276 with an unkown ( hopefully small ) correction @xmath277 beyond rpa . in its position representation , this @xmath116 is given by @xmath278 with the short - hands @xmath279 and @xmath280 .
the integral operator @xmath281 generalizes the particle - hole propagator @xmath113 , cf ( [ c4 ] ) . the diagonal elements @xmath282 define a function @xmath283 which contributes with the rpa - beyond term @xmath284 to the cpd , @xmath285 .
when going from ( [ c1 ] ) to ( [ c2 ] ) , the contour integration @xmath286 , which replaces simultaneously the frequency integration by a velocity integration , leads to the real function @xmath287 for the particle - hole propagator , cf eg @xcite , eq.([b1 ] ) .
the effective interaction @xmath288=q_{\rm c}^2/[q^2+q_{\rm c}^2r(q , u)]$ ] is the rpa screened coulomb interaction with a yukawa - like cut - off @xmath289 .
the qualitative behavior of @xmath290 is in @xcite , fig.3 ; it starts at @xmath291 and decays , passes a zero - intercept and approaches 0 from below , its normalization is zero , see ( [ b1 ] ) .
the ft of @xmath292 , namely @xmath293 contributes to the csf @xmath294 @xcite with a beyond - rpa term @xmath295 .
how does @xmath296 behave for small and large momentum transfers , @xmath297 and @xmath298 , respectively , and in between ?
+ for small @xmath299 , the dynamic rpa - propagator @xmath300 can be approximated by its static truncation @xmath301 @xcite such that @xmath302 with @xmath303 and @xmath304 ( for the properties of @xmath305 see @xcite . ) thus @xmath306 in agreement with @xcite .
the ellipsis represents terms originating from the difference @xmath307 .
the function @xmath296 starts linearly in a small region ( @xmath299 ) , which shrinks and finally vanishes for @xmath54 .
the beyond - rpa factor @xmath308 is added by hand with the following argument . as shown in @xcite , the small-@xmath3 behavior of the hf - function @xmath15
is given in ( [ b18 ] ) .
it transfers via the sfs to the csfs and [ together with the assumption @xmath309 makes them behave correctly for small @xmath3 , according to the plasmon sr ( [ b3 ] ) and ( [ b13 ] ) , see also table 1 and 2 . + at the other end , the large-@xmath3 asymptotics does not require the ring diagram summation , so perturbation theory holds and the `` descreening '' replacement @xmath310 is possible . with the macke function @xmath311 ( explicitly given in @xcite ) it is @xmath312 `` in between '' @xmath313 and @xmath314 and according to ( [ a9 ] ) , the integral @xmath315 contributes to the correlation parameter @xmath316 , where the ellpipsis represents the missing terms beyond rpa .
the behavior of @xmath294 with the rpa - beyond term @xmath295 ( causing the correlation parameters @xmath317 ) is summarized in line 1 of table 1 with unknown coefficients @xmath318 . + * 3.2 the exchange building block @xmath236 in rpa * + as stressed by geldart et al .
@xcite , for a consistent small-@xmath29 description the leading exchange terms of @xmath127 , @xmath319 are needed .
they follow from @xmath236 with the ring - diagram part @xmath117 ( fig .
3a ) : @xmath320 with an unkown ( hopefully small ) correction @xmath321 beyond rpa . that exchange contribution @xmath117 , which corresponds to the above @xmath116 , arises from the exchange @xmath155 ,
i.e. from @xmath322 , thus @xmath323 the diagonal elements @xmath324 define a function @xmath325 which contributes to the cpd @xmath326 .
its ft @xmath327 contributes to the csf @xmath328 with the rpa - beyond term @xmath329 .
a careful study shows , that @xmath330 can be written as @xmath331 this is the exchange counterpart to the direct csf ( [ c3 ] ) , unfortunately not known as an explicit function of @xmath3 .
if one assumes for the small-@xmath3 behavior @xmath332 and @xmath333 , then the correct small-@xmath3 behavior of @xmath334 and @xmath335 results , namely @xmath336 , see ( [ b2 ] ) , and consequently @xmath337 , see ( [ a4 ] ) .
line 2 of table 1 contains so far unknown coefficients @xmath338 . at the other end
, the large-@xmath3 asymptotics of @xmath339 can be extracted from perturbation theory , replacing `` r '' by its lowest order `` 1 '' .
this allows the @xmath340-integration , by which appears the well - known energy denominator : @xmath341 thus , it results similar as @xmath342 , namely @xmath343 with an unknown coefficient @xmath344 . for @xmath345 in general and
possibly @xmath346 , see app.c . the behavior of @xmath328 is summarized in line 2 of table 1 .
an additional correlation parameter @xmath347 arises from the rpa - beyond term @xmath329 . +
* 3.3 how the cumulant structure factors @xmath218 follow from @xmath348 * + for the spin - antiparallel structure factors it holds @xmath349 . in rpa , @xmath350 is given by ( [ c3 ] ) and its large-@xmath3 asymptotics is given by ( 3.6 ) .
but beyond rpa the electron - electron coalescing cusp theorem @xmath351 $ ] ( app.b1 , @xcite ) makes the first term decorated with an additional factor @xmath352 .
one may expect the next term is modified in a similar way , thus @xmath353 with unknown correlation parameters @xmath317 vanishing for @xmath54 .
note that the spin - antiparallel csf @xmath354 starts with @xmath355 .
+ for the spin - parallel structure factors it holds @xmath334 , @xmath356 , and @xmath357 .
the large-@xmath3 asymptotics of @xmath358 is conjectured as ( with another correlation parameter @xmath347 ) @xmath359 such that the difference @xmath360 starts asymptotically according to @xmath361 with @xmath362 ( not with @xmath355 ) .
just this conclusion results also from the electron - electron coalescing curvature theorem with @xmath346 , cf ( [ b7 ] ) . at the other end
, the small-@xmath3 behavior @xmath363 follows from the above derived / assumed expressions for @xmath364 and @xmath365 : @xmath366 \omega_{\rm pl}\ .\ ] ] [ for the qualitative behavior of the corresponding cpd @xmath367 see @xcite , fig.4 ; it starts with @xmath368 , see ( [ b2 ] ) , ( [ b5 ] ) , ( [ b8 ] ) . ]
a consequence for the total csf @xmath369 is @xmath370\omega_{\rm pl } + \cdots\ .\ ] ] this , used in @xmath371 , cancels the first three terms .
thus emerges the quadratic term in agreement with the plasmon sr @xmath372 .
consequently an inflexion point appears near the origin .
its trajectory with @xmath54 is sketched in app.b .
table 2 gives the small-@xmath3 and large-@xmath3 behavior of @xmath217 .
for the small-@xmath3 behavior of @xmath219 the result ( [ b13 ] ) of the inflexion - point analysis with further correlation parameters @xmath373 is used , see table 2 .
it remains open to check the above assumption on @xmath365 .
+ * 3.4 how @xmath43 follows from the cumulant structure factor @xmath369 * + if ( [ c3 ] ) and ( [ c9 ] ) are used in ( [ b29 ] ) , then the cumulant interaction energy @xmath374 with ( `` r '' = rpa , ellipsis = beyond rpa ) @xmath375\ ] ] consists of a `` direct '' part @xmath376 and a corresponding `` exchange '' part @xmath377 . the first one , @xmath378 agrees ( of course ) with what follows from the total energy @xmath41 after macke ( 1950 ) @xcite and gell - mann / brueckner ( 1957 ) @xcite , @xmath379\right \}\ , \ ] ] by means of the virial theorem @xmath380 . the `` exchange '' part @xmath377 can be simplified as @xmath381 namely , whereas in @xmath339 the screened interaction @xmath382 is necessary to remove the long - range divergencies of the bare interaction @xmath383 , this is not the case for @xmath384 , which has been calculated by onsager / mittag / stephen ( 1966 ) @xcite with the result @xmath385(\alpha r_s)^2\ .\ ] ] note that @xmath386 .
+ the agreement of ( [ c18 ] ) with the results of macke and gell - mann / brueckner @xcite and of ( [ c20 ] ) with the calculations of onsager et al .
@xcite confirms also the above expressions for @xmath97 and @xmath98 in the rpa approximations `` dr '' and `` xr '' .
+ * 3.5 how @xmath387 follows from the contraction of @xmath117 * + from @xmath117 of ( [ c7 ] ) follows also a contribution to @xmath388 as explained in app.d , the md @xmath1 starts with a 2nd - order term , hence @xmath382 can be replaced by @xmath389 . therefore @xmath390\
\cdot \right . \\
\cdot\ \hat q_1(q,\eta)\hat q_2(q,-\eta ) { \rm e } ^{-{\rm i}({\mbox{{\boldmath}${k}$}}_1+{\mbox{{\boldmath}${k}$}}_2+{\mbox{{\boldmath}${q}$}}){\mbox{{\boldmath}${r}$}}_2}\ { \rm e}^{{\rm i}[({\mbox{{\boldmath}${k}$}}_1+{\mbox{{\boldmath}${q}$}}){\mbox{{\boldmath}${r}$}}_1+{\mbox{{\boldmath}${k}$}}_2{\mbox{{\boldmath}${r}$}}_1']}+{\rm h.c . } \right\}\ .
\nonumber \end{aligned}\ ] ] the @xmath207-integration and afterwards the @xmath391-integration yield ( including prefactors ) @xmath392 so that @xmath393 with @xmath394 ^ 2}\ , \ ] ] where @xmath395 for @xmath396 and @xmath397 otherwise
. the last step is derived in app.c4 .
this used in ( [ c24 ] ) gives for the md - correlation part the result @xmath398 with @xmath399 ^ 2}+{\rm c.c.}\right \}\ .\ ] ] with @xmath400 for @xmath401 and again with @xmath402 , it turns out @xmath403 ^ 2}\ , \ ] ] being @xmath404 for @xmath241 .
this is exactly the rpa - md of daniel / vosko ( 1960 ) and kulik ( 1961 ) , as it has been shown in detail in @xcite on the way therein from eqs.(3.34 ) to ( 3.38 ) . also therein at ( 3.10 )
it is shown , what @xmath125 contributes to the kinetic energy @xmath42 in agreement with the virial theorem ( [ a1 ] ) and the macke / gell - mann / brueckner energy ( [ c18 ] ) .
the asymptotic behavior is @xmath405 for rpa and @xmath406 beyond it .
this 2nd - order result is reduced by a factor of 1/2 through a still missing ladder diagram in its exchange version , as treated in the next section .
this ladder diagram has a particle - hole line running from @xmath114 to 2 and a 2nd line from @xmath115 to 1 , cf fig.5a .
its contraction @xmath209 and @xmath248 yields the diagram of fig.5b , which differs from fig.5c only by a factor sign@xmath249 according to ( [ b28 ] ) . in detail
the diagram @xmath407 of fig.5a is given by ( [ c12 ] ) .
if for simplicity the rpa screening is neglected ( corresponding to the replacement r@xmath4082 ) and a series of laborious contour integrations are carefully performed together with the contraction steps analog ( [ c21 ] ) , it finally yields @xmath409 the mentioned factor sign@xmath249 transforms @xmath410 to the corresponding md contribution @xmath127 , which thus reduces both the particle branch ( @xmath411 ) and the hole branch ( @xmath412 ) of the rpa - md @xmath125 , eq.([c27 ] ) . in particular the asymptotics for @xmath413 changes from 1/2 to 1/4 of @xmath414 as already mentioned at the end of sec.iii .
furthermore it is easy to show by means of the substitutions @xmath415 and @xmath416 , ( i ) the zero normalization @xmath417 and ( ii ) the kinetic energy @xmath418 where also the identity ( [ c20 ] ) is used .
the result ( [ d2 ] ) is just the exchange integral , which has been solved analytically by onsager et al . @xcite . whereas ( [ d2 ] ) is a non - divergent 2nd - order term , for @xmath419 an indeed complicated ring - diagram summation is needed , to ameliorate diverging terms .
this is a contribution to the theory of extended many - electron systems in terms of reduced - density matrices ( rdms ) , which replace @xmath420 , the many - electron wave functions of finite systems and form an infinite hierarchy , such that a lower - level rdm follows from the next - higher - level rdm by means of a certain contraction procedure , e.g. from the 2-body matrix ( 2-matrix ) @xmath25 follows the 1-matrix @xmath23 .
this paper is driven by the belief that the * cumulant 2-matrix * @xmath0 ( dimensionless : @xmath28 ) has a fundamental meaning .
this shows up for extended systems ( i ) in its size - extensive normalization and contraction and ( ii ) in its ( closely related ) _ linked_-diagram representation . besides it is a decisive key quantity , because from it follows both the pair density ( pd ) and the 1-matrix .
here it is aimed to unveil its secrets for the high - density spin - unpolarized homogeneous electron gas ( heg ) .
although this model is only a marginal corner in the complex field of electron correlation , one should get generally deeper insight into the many - body correlation by exploiting the concept of cumulant decomposition of rdms .
the ground state ( gs ) energy of an electronic system is a simple functional of @xmath421 = electron density , of @xmath422 = momentum distribution ( md ) , and of @xmath423 = pair density ( pd ) with @xmath91 .
these 1- and 2-body quantities have their common origin in the 2-matrix @xmath21 . whereas the pd simply follows from @xmath25 by taking the diagonal elements ,
the 1-body quantities need a procedure called contraction which means @xmath26 and @xmath126 .
but this procedure leads in the case of extended systems to difficulties with the thermodynamic limit ( tdl ) .
namely non - size - extensive expressions @xmath424 appear , resulting from the numbers of pairs @xmath425 .
this is avoided by the cumulant decomposition of the 2-matrix @xmath25 .
this means that @xmath25 is additively decomposed into a hf - like ( or exact exchange ) term @xmath159 , which reduces to a sum of products of the ( correlated ) 1-matrix , and a non - reducible remainder , called cumulant matrix @xmath0 .
its advantage : it is _ size - extensively _ normalizable and contractable . thus the problems with the tdl are eliminated .
this is worked out in detail for the heg with the homogeneous electron density @xmath4 , with the md @xmath1 and with @xmath2 , the structure factor ( sf ) [ or equivalently the 1-matrix @xmath426 and the pd @xmath427 , respectively ] .
the 2-body quantities @xmath2 and @xmath6 are decomposed into hf terms and remainders @xmath9 and @xmath7 , respectively called cumulant sf ( csf ) and cumulant pd ( cpd ) . in detail , this means @xmath8 and @xmath10 with @xmath15= fourier transform of @xmath16 . the contraction procedure is developed in eqs.([b25])-([b28 ] ) : a quadratic equation for @xmath428 yields the 2 branches for particles ( @xmath411 ) and holes ( @xmath412 ) .
+ closely related to the described _ size - extensivity _ of the cumulant 2-matrix is its form ( within many - body perturbation theory ) in terms of _ linked _ diagrams . under the impression that the triangle of `` cumulants @xmath429 size - extensivity @xmath429 linked diagrams '' is a fundamental relation ( with an exponential - linked - diagram theorem as a more general mathematical background @xcite )
, one may curiously ask for the * cumulant geminals * @xmath430 and their weights @xmath201 , which diagonalize @xmath0 @xcite , analogous to eq.([b16 ] ) in @xcite .
what are their properties ?
is there perhaps a 2-body scheme , which allows to calculate approximately these @xmath431 and @xmath201 ?
how they differ from the overhauser geminals @xcite , which parametrize the pd @xmath6 ?
one access to this subject are the contracted schrdinger equations @xcite . but instead of this , in this paper the cumulant 2-matrix of the high - density heg is derived .
the results for @xmath0 ( dimensionless @xmath28 ) prove to be correct , because they yield the known rpa expressions for @xmath2 , @xmath6 , and @xmath1 .
this is the 1st step for further curiousity studies along the line of overhauser @xcite , namely to find the cumulant geminals of the high - density heg , to be specific : the expressions ( [ c1 ] ) , ( [ c7 ] ) , and ( [ c12 ] ) have to be diagonalized .
+ furthermore , the coalescing cusp and curvature theorems , the plasmon sum rule ( and its consequence , the inflexion - point trajectory ) , as well as the small- and large-@xmath3 behavior of the csfs for electron pairs with parallel and antiparallel spins are systematically summarized ( analog to @xcite ) , see table 1 and 2 . within this , also the so far assumed asymptotic behavior of the structure factors for small and large @xmath3 is corrected .
the author is grateful to p. gori - giorgi , jian wang , k. morawetz , j. bodyfelt for discussions and hints and acknowledges p. fulde and the max planck institute for the physics of complex systems dresden for supporting this work and thanks th .
mller and r. schuppe for technical help .
the most compact quantity containing all the secrets of the heg gs @xmath137 is the generating functional @xmath432=\langle \hat p\ { \rm e}^{\int dx[\eta^\dagger(x)\psi(x)+\eta(x)\psi^\dagger(x)]}\rangle\ ] ] with @xmath433 and @xmath434 . @xmath435 and @xmath436 are grassmann variables and @xmath437 makes the normal ordering , i.e. orders the creation operators @xmath438 including sign changes to the left .
the volterra coefficients of @xmath439 $ ] are the rdms @xmath23 , @xmath25 , ... .
the ansatz @xmath439={\rm e}^{\chi[\eta^\dagger|\eta]}$ ] defines another generating functional , @xmath440 $ ] , the volterra coefficients of which are the so - called cumulant matrices @xmath441 , @xmath442 , ... , such that a hierarchy results with higher level rdms expressed by sums of products of lower level rdms and non - reducible remainders .
thus it holds e.g. @xmath443 , @xmath444 , ... with @xmath445 being an antisymmetrizer @xcite . within time - dependent perturbation theory ( @xmath446-matrix theory ) the cumulant matrices
@xmath441 , @xmath442 , ... are given by non - vacuum _ linked _ diagrams only and closely related to this they are size - extensively normalized i.e. @xmath447 , what allows for extended systems the thermodynamic limit with @xmath107 , @xmath448 , @xmath109 @xcite .
here are summarized the normalizations of @xmath98 , @xmath9 etc . as well as their behavior for small and large arguments .
+ * b1 : perfect screening sum rule .
cusp and curvature theorems * + @xmath2 starts with @xmath449 and approaches @xmath450 for @xmath298 .
@xmath6 starts with @xmath451 and oscillatory approaches @xmath450 for @xmath452 .
note , that @xmath453 , since they are probabilities .
the normalizations of @xmath454 and @xmath455 are contained in ( [ a5 ] ) with @xmath70 ( defining the correlation parameter @xmath53 ) and with @xmath449 ( the perfect screening sr ) , respectively : @xmath456=1+c_1\ , \quad \quad \alpha^3 \int\limits_0^\infty d(r^3)\ [ 1-g(r)]=1\ . \nonumber\ ] ] similarly , the normalizations of @xmath64 , @xmath67 , and @xmath457 are in ( 1.4 ) fixed with @xmath458 , @xmath66 , and @xmath63 . in terms of cumulant and `` a , p '' components , this reads as @xmath459 @xmath460 which follows from the fts ( [ a9 ] ) and ( [ a10 ] ) for @xmath461 , respectively for @xmath462 . with @xmath463 ,
the perfect screening sr @xmath449 is hidden in @xmath464 .
more general is the plasmon sr @xcite @xmath465 see table 2 , last line
. + according to the electron - electron coalescing cusp theorem @xmath466 for the spin - antiparallel cpd @xcite , it holds @xmath467 with @xmath468 @xcite .
exactly this high - density behavior of @xmath469 and @xmath470 results also from the ladder theory as the best method to treat short - range correlation @xcite . by means of the kimball trick
@xcite it follows the 1st term of eq.([c11 ] ) , whereas the 2nd term follows from @xmath471 and from ( [ c6 ] ) .
@xmath472 is another correlation parameter vanishing for @xmath54 .
@xmath53 appears also in the large-@xmath47 asymptotics @xmath473 .
thus short - range correlation determine the large - wave number asymptotics of @xmath1 @xcite and of @xmath9 @xcite .
+ for the spin - parallel quantities , the pauli principle makes @xmath474 or equivalently @xmath475 .
furthermore the electron - electron coalescing curvature theorem says @xmath476 or equivalently @xmath477 $ ] . with @xmath478 and with @xcite @xmath479 it holds @xmath480 .
this determines the large-@xmath3 asymptotics of @xmath481 as shown in the following by means of the kimball trick @xcite .
first a function @xmath482 is defined through @xmath483 .
then from the ft ( [ a10 ] ) it follows @xmath484 with the above curvature theorem it finally results @xmath485 comparing this with ( [ c13 ] ) , the sr @xmath486 seems to hold .
for @xmath54 it takes @xmath487 , hence @xmath346 [ for plausible arguments see text after ( [ c8 ] ) ] .
because @xmath488 decays so strong , from ( [ b6 ] ) it follows in addition to ( b.3 ) the relation @xmath489 since @xmath490 , cf @xcite , the integral has to be negative .
this is in agreement with ( [ b3 ] ) and ( [ b7 ] ) : a factor @xmath491 in front of @xmath481 makes the integral vanishing , cf ( [ b3 ] ) .
an additional factor @xmath491 enhances the negative asymptotic branch of @xmath488 , cf ( [ b7 ] ) .
besides , from @xmath492 it follows @xmath493 . for @xmath494 cf eg @xcite . + * b2 : plasmon sr and inflexion - point trajectory * + in ( [ c5 ] ) it has been shown how perturbation theory and rpa makes the plasmon - term @xmath495 arise from the linear term @xmath496 of the non - interacting system . here
it is asked on the contrary , if the interaction is switched off ( @xmath54 ) , how the quadratic behavior @xmath495 is transformed to the linear behavior @xmath496 . the answer : if @xmath2 starts for small @xmath3 according to the plasmon sr @xcite quadratically with @xmath491 , then there must be - already from a topological point of view - an inflexion point @xmath497 , @xmath498 , which moves with @xmath54 towards the origin .
so also the linear behavior at the inflexion point is transported to the origin , realizing thus the correct behavior of the unperturbed ideal fermi gas , as it should .
now this qualitative discussion is quantized by the ansatz @xmath499\ , \quad z=\frac{q}{\omega_{\rm pl}}\ ] ] and adjusting the @xmath29-dependent coefficients @xmath500 , @xmath501 appropriately . namely the power expansion of @xmath2 around @xmath502 @xmath503 is compared with its `` @xmath54 '' limit @xmath504 .
this gives 3 equations , which allows determining the 3 quantities @xmath497 , @xmath500 , @xmath501 : @xmath505 with the result ( for @xmath54 ) @xmath506 so ( [ b9 ] ) takes for @xmath54 the form ( with unknown correlation parameters @xmath373)@xcite @xmath507+\cdots\ .\ ] ] thus the inflexion point moves towards the origin according to @xmath508 and @xmath509 with a slope of @xmath510 being smaller than @xmath511 .
comparison of ( [ b13 ] ) with ( [ c15 ] ) shows @xmath512 and @xmath513 . - as quoted in @xcite , other consequences of ( [ b9 ] ) are @xmath514 and @xmath515=\frac{1}{\omega_{\rm pl}}\ , \quad \alpha^3\int\limits_0^\infty dr\ r^6[1-g(r)]=\frac{2 ^ 3 \cdot 5}{3 ^ 2}\frac{1}{\omega_{\rm pl}^3}\ .\ ] ] the equation with the weight @xmath516 is widely used in @xcite .
remind @xmath517 , what relates @xmath518 to @xmath519 for @xmath520 and @xmath521 . + * b3 : the hf - function @xmath15 * + for the spin - parallel quantities @xmath13 and @xmath14 the functions @xmath16 and its ft @xmath15 are needed , cf ( [ a7 ] ) , ( [ a8 ] ) and table 2 : @xmath522 @xmath15 transports the md @xmath1 , as seen from @xmath523 thus it is related to the probability of finding a pair of electrons with given relative momentum @xmath524 @xcite . for @xmath32 with @xmath239
the integral takes a simple geometric meaning : it is just the volume of two spherical calottes with the height @xmath525 , thus @xmath526 or @xmath527 \theta(2-q)\ , \quad
\int\limits_0^\infty d(q^3)\ \frac{1}{2}f_0(q)=1\ .\ ] ] [ by the way , ( i ) @xmath528 appears in the overhauser theory as the weight of its pd geminals @xcite and ( ii ) the singularities of @xmath15 at @xmath313 and @xmath529 are studied in @xcite . ]
the difference @xmath530 results from @xmath428 .
it holds @xmath531 and @xmath532 . +
* b4 : the jump discontinuity of @xmath533 and @xmath534 at @xmath535 * + for transition momenta @xmath536 , when passing this value from below , the topology changes from 2 overlapping to 2 non - overlapping fermi spheres . the consequence : @xmath537 , cf ( c.2 ) in @xcite @xmath538 @xcite .
the ellipsis represents exchange and beyond - rpa terms . for @xmath539
the relation @xmath540 , cf ( 5.2 ) in @xcite , has to be taken into account .
thus in lowest order @xmath541 , what causes ( in the direct space ) the friedel oscillations of @xmath6 .
it remains open to study the exchange term @xmath542 , which presumably reduces the direct term @xmath543 .
* c1 : * the building elements of the rpa feynman diagrams ( cf figs.2a , 3a , 4a , 5a ) are the coulomb repulsion @xmath544 with the coupling constant @xmath36 and the one - body green s function of free electrons with @xmath545 , @xmath546 * c2 : * from @xmath547 follows the particle - hole propagator @xmath113 in rpa according to @xmath548 with the result @xmath549\ , \\
\tau({\mbox{{\boldmath}${k}$}},{\mbox{{\boldmath}${q}$}})&=&t({\mbox{{\boldmath}${k}$}}+{\mbox{{\boldmath}${q}$}})-t({\mbox{{\boldmath}${k}$}})=q\left ( k\zeta+\frac{q}{2}\right ) , \quad \zeta=\cos\varangle ( { \mbox{{\boldmath}${k}$}},{\mbox{{\boldmath}${q}$}})\ . \nonumber\end{aligned}\ ] ] the denominators contain the excitation energy @xmath550 , to create a hole with @xmath148 inside the fermi sphere and a particle with @xmath551 outside the fermi sphere .
a complicated step function , defined by @xmath552 for @xmath553 and 0 otherwise , is a consequence of the pauli principle .
@xmath287 defines a real function .
a generalization of ( [ c3 ] ) is @xmath554f({\mbox{{\boldmath}${k}$}}_i)\ , \quad \tau_i=\tau({\mbox{{\boldmath}${k}$}}_i,{\mbox{{\boldmath}${q}$}})\ , \ ] ] defining the integral operator @xmath555 , @xmath556 . +
* c3 : * when calculating the csf @xmath557 , then the macke function @xmath128 appears via @xmath558 when dealing with the corresponding exchange term @xmath559 , then by means of contour integration it results @xmath560 where the integral operator @xmath561 and the exchange term @xmath345 are defined by @xmath562 @xmath128 and @xmath345 have the large-@xmath3 asymptotics @xmath563 whether the guessed value of @xmath346 [ text after ( [ b7 ] ) ] is correct , has to be studied .
the change of 2/5 to 8/5 is a heavy enhancement of the @xmath355-tail , which is plausible in view of the following qualitative arguments . in @xcite and @xcite , the (
explicitly not known ) function @xmath345 is compared with the explicitly known macke function @xmath128 .
whereas @xmath128 starts linearly with @xmath3 ( what causes the heisenberg divergence ) , its exchange pendant @xmath345 must start at least with @xmath491 .
otherwise the integral @xmath564 ( which is proportional to the congenially calculated onsager - et - al energy @xmath384 ) would not exist .
an estimate shows even @xmath565 .
this flattening in the small-@xmath3 region together with the `` area conservation '' @xmath566=0 $ ] enforces the mentioned tail enhancement . +
* c4 : * proof of ( [ c25 ] ) : in the following the notation @xmath567 is used .
( [ c4 ] ) inserted in ( [ c25 ] ) gives @xmath568_{{\mbox{{\boldmath}${k}$}}_1\to{\mbox{{\boldmath}${k}$}}-{\mbox{{\boldmath}${q}$ } } } \cdot \nonumber \\ \cdot\left[\frac{\theta(k_2 - 1)\theta(1-|{\mbox{{\boldmath}${k}$}}_2+{\mbox{{\boldmath}${q}$}}|)}{-\eta-\tau({\mbox{{\boldmath}${k}$}}_2,{\mbox{{\boldmath}${q}$}})-{\rm i}\delta}-\frac{\theta(1-k_2)\theta(|{\mbox{{\boldmath}${k}$}}_2+{\mbox{{\boldmath}${q}$}}|-1}{-\eta-\tau({\mbox{{\boldmath}${k}$}}_2,{\mbox{{\boldmath}${q}$}})+{\rm i}\delta}\right]_{{\mbox{{\boldmath}${k}$}}_2\to-{\mbox{{\boldmath}${k}$ } } } \nonumber \\
= \left[\frac{\theta(|{\mbox{{\boldmath}${k}$}}-{\mbox{{\boldmath}${q}$}}|-1)\theta(1-k)}{\eta-\tau({\mbox{{\boldmath}${|}$}}{\mbox{{\boldmath}${k}$}}-{\mbox{{\boldmath}${q}$}}|,{\mbox{{\boldmath}${q}$}})-{\rm i}\delta}-\frac{\theta(1-|{\mbox{{\boldmath}${k}$}}-{\mbox{{\boldmath}${q}$}}|)\theta ( k-1}{\eta-\tau({\mbox{{\boldmath}${|}$}}{\mbox{{\boldmath}${k}$}}-{\mbox{{\boldmath}${q}$}}|,{\mbox{{\boldmath}${q}$}})+{\rm i}\delta}\right ] \cdot \hspace{2cm}\nonumber \\
\cdot\left[\frac{\theta(k-1)\theta(1-|{\mbox{{\boldmath}${k}$}}-{\mbox{{\boldmath}${q}$}}|)}{-\eta-\tau({\mbox{{\boldmath}${k}$}},-{\mbox{{\boldmath}${q}$}})-{\rm i}\delta}-\frac{\theta(1-k)\theta(|{\mbox{{\boldmath}${k}$}}-{\mbox{{\boldmath}${q}$}}|-1}{-\eta-\tau({\mbox{{\boldmath}${k}$}},-{\mbox{{\boldmath}${q}$}})+{\rm i}\delta}\right]\ . \hspace{1.5
cm } \end{aligned}\ ] ] only the products `` over cross '' contribute . with the substitution @xmath569 it results @xmath570 ^ 2}+\frac{\theta(1-k)\theta(|{\mbox{{\boldmath}${k}$}}+{\mbox{{\boldmath}${q}$}}|-1)}{[\eta + \tau({\mbox{{\boldmath}${k}$}},{\mbox{{\boldmath}${q}$}})]^2}= \frac{\tilde \theta({\mbox{{\boldmath}${k}$}},{\mbox{{\boldmath}${q}$}})}{[\eta+\tau({\mbox{{\boldmath}${k}$ } } , { \mbox{{\boldmath}${q}$}})]^2}\ .\ ] ] to have finally exactly the same expression as in @xcite the substitution @xmath571 changes the sign in the dominator , proving thus ( [ c25 ] ) .
+ * c5 : * proof of ( [ d1 ] ) : the diagram of fig.5a means in detail @xmath572}\ \chi_{\rm lx2}({\mbox{{\boldmath}${k}$}}_{1,2},{\mbox{{\boldmath}${q}$}}_{1,2})+{\rm h.c.}\right\}\ , \end{aligned}\ ] ] where the following abbreviations are used : @xmath573 @xmath574 @xmath575 for the contribution of the ladder diagram @xmath576 to the md @xmath1 , the contraction steps analog ( [ c21 ] ) can be performed with the result @xmath577 } = \left(\frac{4\pi}{3}\right)^2 \delta({\mbox{{\boldmath}${k}$}}_1-{\mbox{{\boldmath}${k}$}})\delta({\mbox{{\boldmath}${k}$}}_2-{\mbox{{\boldmath}${k}$}}_1 ) \nonumber\ ] ] @xmath578 @xmath579 contains contour integrations in 4 complex frequency planes .
unfortunately , each of these 4 integrals runs over 3 frequency denominators in a difficult way .
but by means of a trick ( integration variables are substituted ) the @xmath580 integrations can be decoupled : @xmath581 .
this does not influence @xmath582 ( it is only written in reverse order ) , but it changes @xmath583 : @xmath584 @xmath585 now it is easy to perform the integration over @xmath586 , namely @xmath587 therein those cases do not contribute , where the 2 poles in the complex @xmath588-plane are both on the same side of the real axis .
only the combinations @xmath589 with @xmath590 contribute .
similarly the @xmath591-integration is handled : @xmath592 the next step would be the @xmath593-integration .
the integrand is a product of the rhs s of ( [ c17 ] ) and ( [ c18 ] ) with the 3rd factor of ( [ c16 ] ) : @xmath594\ { \mbox{{\boldmath}${q}$}}_1\cdot{\mbox{{\boldmath}${q}$}}_2}\ .
\nonumber\end{aligned}\ ] ] the integration region is defined by the functions @xmath595 .
the identity @xmath596 is used .
the remaining @xmath597-integration finally yields @xmath598[\omega_1-t({\mbox{{\boldmath}${k}$}})+{\mbox{{\boldmath}${q}$}}_1\cdot { \mbox{{\boldmath}${q}$}}_2\mp{\rm i}\delta]\ { \mbox{{\boldmath}${q}$}}_1\cdot{\mbox{{\boldmath}${q}$}}_2}= \frac{\theta({\mbox{{\boldmath}${k}$}},{\mbox{{\boldmath}${q}$}}_{1,2})}{({\mbox{{\boldmath}${q}$}}_1\cdot{\mbox{{\boldmath}${q}$}}_2)^2}\ ] ] where @xmath599 .
( [ c21 ] ) inserted in ( [ c14 ] ) gives ( [ d1 ] ) , qed .
here is a list of characteristic correlation parameters as functions of @xmath29 , vanishing for @xmath54 : + @xmath600 $ ] is the normalization of the md - correlation tails .
+ @xmath601=n-\int_0^\infty d(k^3)\ [ n(k)-n_0(k)]^2 $ ] is the lwdin parameter .
+ @xmath602 is the reduced discontinuity jump of @xmath1 at the fermi surface @xmath17 .
+ @xmath603 $ ] is the particle - hole symmetric correlation entropy , see eq.(22 ) in @xcite .
+ @xmath604 with @xmath605 being the on - top value of the coulomb hole , + @xmath53 simultaneously determines the large-@xmath3 asymptotics of @xmath606 and @xmath1 , + @xmath607 determine the higher - order large-@xmath3 asymptotics of @xmath606 , + @xmath608 and @xmath42 determine the on - top curvature of the fermi hole @xmath609 , + @xmath373 determine the small-@xmath3 behavior of @xmath2 beyond the plasmon - sr term @xmath610 , + @xmath611 influences the jump discontinuity @xmath612 .
+ all these parameters may serve as measures of the correlation strength .
they vanish for @xmath54 with terms @xmath613 , @xmath110 or @xmath614 . from a diagramatical point of view
all the quantities characterizing @xmath1 or derived from it , should start with @xmath111 .
the reason : the diagrams for @xmath1 start with @xmath615 , the 1st - order term @xmath616 ( fig .
4c ) is zero - in agreement with the vanishing 1st - order kinetic energy @xmath617 , what follows from the virial theorem ( [ a1 ] ) .
but there are starting dependencies as @xmath618 cf ( 3.46 ) in @xcite .
this peculiar behavior is an example for how higher - order partial summations may create lower - order terms .
it is a peculiarity of the coulomb repulsion .
@xmath619 results from that @xmath47-region @xmath620 , where the rpa reconstruction takes place .
a characteristic non - analyticity is also the behavior of @xmath1 at the fermi surface @xmath17 , namely @xmath621 .
tosi in : n.h .
march ( ed . ) , _ electron correlation in the solid state _ , imperial college press , 1999 , p. 1 .
- g. giuliani and g. vignale , _ quantum theory of the electron liquid _ , cambridge university press , cambridge , 2005 .
there is a non - vanishing probability of finding 0 , 2 , 3 , ...
electrons in such a wigner sphere . such particle number fluctuations in fragments have been studied by p. ziesche , j. tao , m. seidl , and j.p .
perdew , int .
j. quantum chem . * 77 * , 819 ( 2000 ) with the conclusion ` correlation suppresses fluctuations ' , cf also p. fulde , _ electron correlations in molecules and solids _ , 3rd ed .
, springer , berlin , 1995 , p. 157 .
the coulomb repulsion takes the form @xmath622 , if lengths and energies are measured in units of @xmath50 and @xmath34 , respectively .
this shows , that @xmath29 is not only a density parameter @xcite , but also measures the interaction strength or coupling constant .
the ft of @xmath622 is @xmath623 with @xmath3 measured in units of @xmath48 and @xmath36 .
- notice that @xmath624 gives the energy in a.u .
, e.g. the energy in zeroth order and the lowest - order exchange energy are @xmath625 and @xmath626 , respectively .
j. cioslowski , p. ziesche , and k. pernal , phys .
b * 63 * , 205105 ( 2001 ) ; j. chem . phys . *
115 * , 8725 ( 2001 ) .
p. ziesche , int .
j. quantum chem . * 90 * , 342 ( 2002 ) .
the last sentence of the abstract has to be deleted .
p. ziesche and j. cioslowski , physica a * 356 * , 598 ( 2005 ) . - here it is shown , which peculiarities of @xmath1 and @xmath2 caused by rpa lead to the logarithmic non - analyticity of the high - density correlation energy @xmath627 . - in eq . ( 27 ) the correct prefactor is @xmath628 .
r. diez muio , i. nagy , and p.m. echenique , phys .
b * 72 * , 075117 ( 2005 ) . j. cioslowski and p. ziesche , phys . rev .
b * 75 * , 085103 ( 2007 ) . here the self - energy @xmath629 is investigated .
p. ziesche , phys .
( b ) * 244 * , 2022 ( 2007 ) ; the relation @xmath630 has been studied for @xmath54 . - in appendix c the macke function @xmath128 is explicitly given .
the notation @xmath128 is related to the notation used in @xcite as @xmath631 and in @xcite as @xmath632 .
ziesche , ann .
physik ( leipzig ) * 16*(1 ) , 45 ( 2007 ) . here
it has been shown @xmath633 .
m.l . glasser and g. lamb ,
a * 40 * , 1215 ( 2007 ) .
p. ziesche , ann .
( berlin ) * 522*(10 ) , 739 ( 2010 ) . - in the 2nd line of eq .
( c.19 ) there is a prefactor @xmath491 missing .
read the last line on p. 762 as `` respectively .
one may guess , that @xmath345 starts with @xmath634 . '' .
w. heisenberg , z. naturf . *
2a * , 185 ( 1947 ) .
w. macke , z. naturf .
* 5a * , 192 ( 1950 ) . -
another access with the same result is due to d. bohm and d. pines , phys .
rev . * 92 * , 609 ( 1953 ) and d. pines , phys . rev . *
92 * , 626 ( 1953 ) .
m. gell - mann and k. brueckner , phys . rev . * 106 * , 364 ( 1957 ) .
march , phys . rev . * 110 * , 604 ( 1958 ) .
- note that the `` hellmann - feynman theorem '' had been formulated first by p. gttinger , z. phys .
* 73 * , 169 ( 1932 ) ; later it was to be found in h. hellmann , einfhrung in die quantenchemie ( deuticke , leipzig , 1937 ) , pp .
61 and 285 as well as in r.p .
feynman , ph.d .
thesis , mit , 1939 ; phys .
rev . * 56 * , 340 ( 1939 ) .
d. pines and p. nozi@xmath635res , the theory of electron liquids , vol .
( benjamin , new york , 1966 ) . from the @xmath636-sr or thomas - reich - kuhn sr of the dynamic sf
@xmath637 follows the plasmon sr @xmath76 of the static sf @xmath2 .
e. daniel and s.h .
vosko , phys . rev . * 120 * , 2041 ( 1960 ) .
kulik , zh . @xmath638ksp .
teor . fiz .
* 40 * , 1343 ( 1961 ) [ sov .
jetp * 13 * , 946 ( 1961 ) ] .
p. ziesche , phys.stat.sol .
( b ) * 232 * , 231 ( 2002 ) .
gori - giorgi and p. ziesche , phys .
b * 66 * , 235116 ( 2002 ) . - an _ analytical _ expression for @xmath1 in the range @xmath639 is carefully fitted to the well - known kinetic energy @xmath494 . - for the momentum distributions of the spin - polarized electron gas see @xcite .
p. ziesche and f. tasn@xmath640di , ann .
physik ( leipzig ) * 13 * , 124 ( 2004 ) , erratum * 13 * , 623 ( 2004 ) , sec.4 , fig.2 . -
besides in sec.5 it is also presented the spin - structure of the spin - polarized heg .
m. holzmann , b. bernu , c. pierleoni , j. mcminis , d.m .
ceperley , v. olevano , and l.d .
site , phys .
* 107 * , 110402 ( 2011 ) with monte carlo calculations for @xmath1 .
p. ziesche , thesis , th dresden , 1963 ; ann .
( leipzig ) * 21 * , 80 ( 1968 ) .
i.a . nechaev and e.v .
chulkov , phys .
b * 73 * , 165112 ( 2006 ) .
a.j . glick and r.a .
ferrell , ann .
11 * , 359 ( 1960 ) .
geldart , can .
* 45 * , 3139 ( 1967 ) .
kimball , phys .
b * 14 * , 2371 ( 1976 ) .
d.m . ceperley and b.j .
alder , phys .
* 45 * , 566 ( 1980 ) , with monte carlo calculations for @xmath6 .
other papers are based eg on the dielectric function @xmath641 and local field corrections .
kimball , phys .
rev . a * 7 * , 1648 ( 1973 ) .
rajagopal , j.c .
kimball , m. banerjee , phys .
b * 18 * , 2339 ( 1978 ) .
d.j.w . geldart , a. houghton , s.h .
vosko , can . j. phys . *
42 * , 1938 ( 1964 ) .
p. ziesche and f. tasn@xmath640di , ann .
( leipzig ) * 13 * , 232 ( 2004 ) ; erratum * 13 * , 624 ( 2004 ) . n. iwamoto , phys .
rev . a * 33 * , 1940 ( 1986 ) , eq .
gori - giorgi , f. sacchetti , g.b .
bachelet , physica a * 280 * , 199 ( 2000 ) , table on p. 201 ; phys .
b * 61 * , 7353 ( 2000 ) ; phys . rev .
b * 66 * , 159901(e)(2002 ) .
l. zecca , p. gori - giorgi , s. moroni , and g.b .
bachelet , phys .
b * 70 * , 205127 ( 2004 ) .
l. onsager , l. mittag , and m. j. stephen , ann .
physik ( leipzig ) * 18 * , 71 ( 1966 ) . here the exchange integral of the 3-dimensional heg has been evaluated _
analytically_. this rather herculean work was generalized and extended to the @xmath642-dimensional electron gas by m.l .
glasser @xcite .
glasser , j. comp .
appl . math . *
10 * , 293 ( 1984 ) .
p. ziesche , commun .
* 5 * , 301 ( 1967 ) .
p. ziesche , in : j. cioslowski ( ed . ) , _ many - electron densities and reduced density matrices _ , kluwer academic / plenum publishers , new york , 2000 , p. 33 .
p. ziesche , lecture series on computer and computational sciences * 7b * , 1051 , ( 2006 ) .
z. qian , phys .
b * 73 * , 035106 ( 2006 ) .
j. cioslowski and p. ziesche , phys .
b * 71 * , 125105 ( 2005 ) ; phys . rev .
b * 72 * , 239901(e)(2005 ) . j.c .
kimball , j. phys .
a * 8 * , 1513 ( 1975 ) .
h. yasuhara and y. kawazoe , physica * 85a * , 416 ( 1976 ) .
rassolov , j.a .
pople , and m.a .
ratner , phys .
b * 62 * , 2232 ( 2000 ) .
overhauser , can .
* 73 * , 683 ( 1995 ) .
succeeding papers are @xcite .
p. gori - giorgi and j.p .
perdew , phys . rev .
b * 64 * , 155102 ( 2001 ) : extended overhauser model . p. gori - giorgi and a. savin , phys .
a * 73 * , 032506 ( 2006 ) : electron - electron coalescence and short- and long - range correlation .
m. corona , p.
gori - giorgi , and j.p .
perdew , phys .
b * 69 * , 045108 ( 2004 ) .
r. asgari , m. cardenas , m. polini , b. davoudi , and m.p .
tosi , cond - mat/0408293 .
p. ziesche , phys .
b * 67 * , 233102 ( 2003 ) .
p. ziesche , phys .
( b ) * 241 * , 3544 ( 2004 ) : i ; ( b ) * 242 * , 2051 ( 2005 ) : ii . - it is shown , that the kimball - overhauser geminals do not satisfy the plasmon sr .
yasuhara , solid state commun . * 11 * , 1481 ( 1972 ) ; physica * 78 * , 420 ( 1974 ) ; j.m .
herbert , phys .
a * 66 * , 052502 ( 2002 ) ; r. asgari , m. polini , b. davoudi , and m.p .
tosi , solid state commun . * 125 * , 139 ( 2003 ) . l. kong and e.f .
valeev , j. chem . phys . * 134 * , 214109 ( 2011 ) .
alcoba , r.c .
bochicchio , l. lain , and a. torre , j. chem . phys . * 133 * , 144104 ( 2010 ) .
shamasundar , j. chem . phys .
* 131 * , 174109 ( 2009 ) .
s.d . hamieh and h. zaraket , eur .
j. d * 55 * , 229 ( 2009 ) .
m. nooijen , m. wladyslawski , and a. hazra , j. chem . phys .
* 118 * , 4832 ( 2003 ) .
d.a . mazziotti ( ed . ) , _ reduced density matrices : with applications to many - electron atoms and molecules _ , wiley - blackwell , hoboken , 2007 ; t. juh@xmath643sz and d.a .
mazziotti , j. chem .
phys . * 125 * , 174105 ( 2006 ) .
herbert , int .
j. quantum chem . * 107 * , 703 ( 2007 ) .
herbert , `` reconstructive approaches to one- and two - electron density matrix theory '' , dissertation , uni of wisconsin - madison , 2003 , p. 88 .
harriman , phys .
a * 65 * , 052507 ( 2002 ) .
w. kutzelnigg and d. mukherjee , j. chem phys .
* 110 * , 2800 ( 1999 ) ; chem .
. lett . * 317 * , 567 ( 2000 ) .
w.h . young and n.h .
march , proc .
a * 256 * , 62 ( 1960 ) .
m. nakata , b.j .
braams , k. fujisawa , m. fukuda , j.k .
percus , m. yamashita , and zh .
zhao , j. chem . phys . * 128 * , 164113 ( 2008 ) .
j.sun , j.p .
perdew , and m. seidl , phys .
b * 81 * , 085123 ( 2010 ) .
the correlation energy is interpolated between high- and low - density limits .
loos and p.w .
gill , phys .
b * 84 * , 033103 ( 2011 ) . here , the spin - polarized heg at high - density is treated .
the gs - energy @xmath41 is also a functional of @xmath644 and @xmath645 with @xmath646 @xcite and @xmath647 @xcite .
h. yasuhara , physica * 78 * , 420 ( 1974 ) . , @xmath3 = momentum transfer , @xmath588 = energy / frequency , dashed line = bare interaction @xmath648 , closed loop = particle - hole propagator @xmath113 of ( [ c3 ] ) , wavy line = effectively screened interaction @xmath649 $ ] . ] ( following from 2a = @xmath116 through the exchange replacement @xmath652 , 3b = @xmath653 or @xmath330 ( non - divergent ) , and 3c = @xmath377 .
as shown by onsager et al .
@xcite , already @xmath384 does not diverge ( needs no rpa renormalization ) . ]
transforms to the 1-body diagram 4b through the contraction [ with 2=2 and @xmath126 , see ( [ b25 ] ) ] , marked by a small circle and finally to the non - divergent rpa - diagram @xmath125 of fig.4c ( daniel / vosko and kulik ) @xcite .
the 1st - order term is zero ( analog to the 1st - order kinetic energy @xmath617 ) . ]
transforms to the 1-body diagram 5b through the contraction ( [ b25 ] ) and finally to the diagram @xmath127 of fig.5c . in sec.iv and app.c5
the rpa screening is neglected for simplicity corresponding to the replacement r@xmath4082 or @xmath654 . ]
= direct cumulant 2-matrix through its exchange pendant @xmath236 and the cumulant pd @xmath7 to the final results @xmath1 and @xmath6 , from which follow ( i ) @xmath5 = 1-matrix , @xmath42 = kinetic energy , @xmath51 = lwdin parameter and ( ii ) @xmath43 = interaction energy , respectively .
`` a''=spin - antiparallel , `` p''=spin - parallel . from ft arises a similar scheme for the structure factor @xmath2 , the cumulant structure factor @xmath9 , and the hf term @xmath15 [ @xmath655 , respectively ] .
note that @xmath656 $ ] and @xmath263 . ] | for the first time , the _ cumulant _ 2-body reduced density matrix (= 2-matrix ) of the spin - unpolarized homogeneous electron gas ( heg ) is considered .
this @xmath0 proves to be the common source for both the momentum distribution @xmath1 and the static structure factor @xmath2 . within many - body perturbation theory , this @xmath0
is given by only _ linked _
diagrams ( with 2 open particle - hole lines as well as with closed loops and interaction lines ) . here
it is worked out in detail , how the 1-body quantity @xmath1 follows from the 2-body quantity @xmath0 - through a certain contraction procedure , cf eqs.([b25])-([b28 ] ) . in particular , this @xmath0 is developed for the high - density heg .
its correctness is checked by deriving from it @xmath1 and @xmath2 , known from the random - phase approximation ( rpa ) .
this study opens the way to a more sophisticated heg description in terms of cumulant geminals or / and variational methods . besides , the cumulant structure factor ( csf ) of the exchange in lowest order
is explicitly given and sum rules for the csfs and their small- and large-@xmath3 behavior ( beyond rpa ) are systematically summarized within the plasmon sum rule , coalescing theorems , and the inflexion - point trajectory .
+ rdm reduced density matrix , sr sum rule , @xmath4 homogeneous electron density + md momentum distribution @xmath1 , @xmath5 1-body rdm + pd pair density @xmath6 , cpd cumulant pd @xmath7 : @xmath8 + sf structure factor @xmath2 , csf cumulant sf
@xmath9 : @xmath10 + `` a '' antiparallel spin : @xmath11 , @xmath12 + `` p '' parallel spin : @xmath13 , @xmath14 + ft fourier transform , @xmath15 = ft of @xmath16 , fs fermi surface @xmath17 * the two - body reduced density matrix of the high - density electron gas * p. ziesche + max - planck - institut fr physik komplexer systeme + nthnitzer str . 38 , d-01187 dresden , germany , [email protected] + pacs 71.10.ca , 05.30.fk |
deuterium is created during primordial nucleosynthesis ( e.g. weinberg 1977 ; boesgaard & steigman 1985 ; olive , steigman & walker 2000 ) and then destroyed in stars through d(p , @xmath0)@xmath1he reactions occurring at relatively low temperatures ( of a few 10@xmath2 k ) .
since it is hard to synthesize this loosely bound isotope in significant quantities in any astrophysical environment ( epstein , arnett & schramm 1976 ; epstein 1977 ; prodanovi & fields 2003 ) , its abundance in galaxies is expected to monotonically decrease with time owing to stellar cycling .
deuterium is the light nuclide produced during big bang nucleosynthesis ( bbn ) whose primordial abundance is most tightly tied to the baryon - to - photon ratio , @xmath3 ( or , equivalently , to the ratio of the baryon and critical densities at the present time , @xmath4 ) . in principle , once its primordial abundance , ( d / h)@xmath5 , is derived from measurements in unevolved systems , the value of @xmath4 is precisely known . however , there is still significant dispersion among the abundances derived for a handful of high - redshift , low - metallicity absorption - line systems towards quasi - stellar objects ( qsos ) with reasonably firm deuterium detections .
a more precise @xmath3 determination comes from the observations of the cosmic microwave background ( cmb ) power spectrum coupled to ly@xmath6 forest data ( spergel et al .
2003 ; see , however , pettini 2006 , his fig . 7 )
the suggested value , @xmath7 , where @xmath8 , when coupled to predictions from standard bbn theory , leads to a very narrow range for the primordial deuterium abundance , ( d / h)@xmath9 2.42.8 @xmath10 10@xmath11 ( e.g. coc et al .
2004 ) .
relating ( d / h)@xmath5 to the current local abundance of d meets with the non - trivial issue of estimating what the _ true _ local abundance of deuterium actually is .
in fact , a large scatter is found among nearly fifty d / h measurements towards as many galactic lines of sight .
possible explanations ( including deuterium depletion on to dust grains , or recent infall of unprocessed gas ; see e.g. pettini 2006 , and references therein ; prochaska et al .
2005 ) lead to different interpretations of the data . [ cols="<,^,^,^,^,^,^,<",options="header " , ] [ tab.ast ] notice that the dotted lines lay on the dashed ones , but they end up with a larger d / h at present , d / h = 1.9 @xmath10 10@xmath11 rather than d / h = 1.6 @xmath10 10@xmath11 . indeed , different d / h and o / h ratios are reached at late times according to the different imfs and stellar lifetimes adopted .
in particular , the higher the mass of massive stars ( responsible for a quick recycling of the gas ) and the lower the mass of very - low mass objects ( which just lock - up material , from the point of view of gce ) , the lower ( higher ) the d / h ( o / h ) ratio predicted at present . the fraction of stars with initial mass @xmath12 8 m@xmath13 ( @xmath14 1 m@xmath13 ) is larger ( smaller ) for the kroupa et al .
( 1993 ) imf than for the scalo ( 1986 ) one ( see romano et al .
2005b , their figure 1 ) .
this explains the different behaviours predicted by the models shown in fig . [ fig.mod ] . a primordial abundance of ( d / h)@xmath15 = 2.6 @xmath10 10@xmath11 is assumed for all the models , i.e. the weighted mean of the most reliable qso absorber data available in the literature ( see table [ tab.qso ] ) .
this value agrees well with that inferred from cmb anisotropy measurements and the standard bbn theory .
also shown in fig .
[ fig.mod ] are d / h measurements from qso absorber spectra ( circles ; see section [ sec.dat ] for references ) , the psc deuterium abundance after geiss & gloeckler ( 1998 ; sun symbol ) , and the representative ( d / h)@xmath16 value according to either linsky et al .
( 2005 ; ` high ' value ) or hbrard & moos ( 2003 ; ` low ' value ) .
two [ o / h ] values are displayed for the lism , namely that measured by oliveira et al .
( 2005 ) for the neutral medium inside the lb , log(o
/ h)+12 = 8.54 @xmath17 0.02 , and that given by esteban et al .
( 2004 ) for orion , log(o / h)+12 = 8.73 @xmath17 0.03 , which represents the oxygen content of the ionized matter and dust .
notice that part of the local oxygen is depleted on to dust grains ( @xmath1825 per cent ; see oliveira et al .
2005 ; meyer , jura & cardelli 1998 ; andr et al .
2003 ; cartledge et al .
2004 ) , which brings the two measurements into agreement .
all the ratios are normalized to the solar oxygen and silicon abundances from asplund et al .
an apparently striking feature of all the models is that only minor deuterium destruction is predicted as [ o / h ] varies on more than two orders of magnitude , increasing from [ o / h ] @xmath193 dex up to [ o / h ] @xmath18 @xmath200.5 dex .
the deuterium abundance is sensibly reduced afterwards .
this behaviour is easily explained by the large oxygen yield from the first type ii supernovae , which leads to a prompt oxygen enrichment of the ism only a few million years after the onset of the star formation .
indeed , we find [ o / h ] @xmath213.0 dex already 4 million years after the onset of the star formation in the galaxy , which raises to [ o / h ] = @xmath200.5 dex in the following @xmath18300 million years . in the meantime , continuing infall of gas of primordial chemical composition helps to keep the d abundance near its primordial value
infall of gas is particularly strong during the early galaxy evolution , so that the consumed d is being replenished fast by infall at those earlier times .
later on , when the contribution from low- and intermediate - mass stars ( which are not net oxygen producers ) to the chemical evolution of the galaxy becomes increasingly important and the infall rate is reduced , a steeper behaviour on the d / h versus o / h plot is observed . in table
[ tab.ast ] the astration factors by number for all the models displayed in fig .
[ fig.mod ] are listed and compared to each other .
all the models well reproduce the solar abundances of oxygen and silicon at @xmath22 gyr
. a model in which the scalo ( 1986 ) imf and the schaller et al .
( 1992 ) stellar lifetimes are adopted is able to destroy d at a level consistent with that suggested by linsky et al .
none of our models is able to predict astration factors larger than @xmath23 2 .
this leads us to favour the dust depletion scenario .
deuterium astration in galaxies is expected to be strongly dependent on the assumptions for the two competing processes of star formation and gas infall . in fig .
[ fig.com ] we compare the results obtained for d / h versus time by two gce codes , the one of chiappini et al .
( 2002 ; dashed line ) and that by tosi ( 1988a , b ; solid line ) , once adopting the same prescriptions on the stellar lifetimes ( i.e. schaller et al . 1992 ) and the stellar imf ( kroupa et al . 1993 ) .
the two models assume different prescriptions for the star formation and infall rates , yet the predicted deuterium evolution is almost the same .
this is because what really matters is _ the interplay _ between subtraction and replenishment of gas at each time , and this interconnection is severely constrained by a number of independent observables ( see , e.g. , matteucci 2004 ) . according to fig .
[ fig.com ] , only a mild deuterium destruction is predicted in the last 4.5 gyr .
this is consistent with the mild galactic evolution at late times suggested by the small increase of the overall metallicity of the gas from the time of sun formation up to now ( esteban et al .
2004 , and references therein ) .
also the observed g- and k - dwarf metallicity distributions clearly show that the majority of the stars in the solar neighbourhood formed at low metallicities , [ fe / h ] @xmath24 0 , thus pointing again to a low star formation activity ( and , hence , less evolution ) at late times .
also shown in fig .
[ fig.com ] is the chiappini et al .
( 2002 ) model with scalo s ( 1986 ) imf and schaller et al.s ( 1992 ) stellar lifetimes , which best reproduces both the solar and local ` high ' d / h values ( dotted line ) .
it is worth emphasizing that with these prescriptions the model also well reproduces all the other observational constraints ( romano et al .
2005a ) .
in general , our models predict ( d / h)@xmath16 values in the range 1.42.0 @xmath10 10@xmath11 .
we need to enhance the ratio of the star formation to gas infall efficiencies if a larger fraction of the initial deuterium has to be destroyed .
but can a more efficient star formation and/or less effective infall of unprocessed gas be accommodated within successful gce models ?
it is immediately seen that the almost constant metal abundance of the ism from the time of sun s formation up to now ( esteban et al .
2004 ) sharply contrasts with rapid star formation and/or absence of external sources of unenriched gas . in fig .
[ fig.bad ] , we show the outputs of a model especially designed to produce a significant drop in deuterium abundance from its psc value to the local one ( for further modeling and discussions see tosi et al .
this model , assuming @xmath25 = 1.5 gyr rather than 7 gyr for the thin - disc formation time scale , fits the solar and local ` low ' d / h values ( fig .
[ fig.bad ] , upper panel , solid line ) , as well as other observational constraints ( such as for instance the current gaseous and stellar mass surface densities at the solar position ) .
however , it is barely consistent with the metal abundances of the sun , it predits a sharp increase of the metal abundance of the ism in the last 4.5 gyr ( at variance with the observations ) , and it can not account for the metallicity distribution of solar neighbourhood g- and k - dwarf stars ( fig .
[ fig.bad ] , lower panel , solid line ) .
the metallicity distribution of long - lived stars in the solar neighbourhood is a fundamental constraint for gce models , which mainly constrains the infall time scale and strength ( see , e.g. , chiappini et al . 1997
; matteucci 2004 ) . in fig .
[ fig.bad ] , lower panel , we show as a dashed , dotted , and solid histogram the g - dwarf metallicity distributions observed by wyse & gilmore ( 1995 ) , rocha - pinto & maciel ( 1996 ) and jrgensen ( 2000 ) , respectively . also shown is the k - dwarf metallicity distribution by kotoneva et al .
( 2002 ; big dots ) .
k dwarfs have lifetimes older than the present age of the galactic disc , and are thus ideal stars for investigating the chemical evolution of the disc .
g dwarf stars instead are sufficiently massive that some of them have begun to evolve away from the main sequence , so that corrections must be taken into account when determining their space densities and metallicity ( kotoneva et al . 2002 ) .
these corrections are at least partly responsible for the non - negligible differences among different observational g - dwarf metallicity distributions ( in particular , the low - metallicity tail of the distribution and the exact position of the peak may vary significantly according to different authors ) . in conclusion ,
in the framework of our models we can not explain at the same time both the lowest observed ( d / h)@xmath16 values and the observed g- and k - dwarf metallicity distributions . in fig .
[ fig.grad ] we display the galactic gradients of d / h , d / o , and o / h at the present time predicted by adopting the prescriptions of model a of chiappini et al .
( 2001 ) as far as the formation and evolution of the galactic disc are concerned , and by varying the prescriptions for the imf and the stellar lifetimes ( solid line : scalo s 1986 imf and maeder & meynet s 1989 stellar lifetimes ; dashed line : kroupa et al.s 1993 imf and maeder & meynet s 1989 stellar lifetimes ; dotted line : scalo s 1986 imf and schaller et al.s 1992 stellar lifetimes see romano et al .
2005a ) . a value of ( d / h)@xmath15 = 2.6 @xmath10 10@xmath11 is assumed for all the models .
also shown are the relevant observations .
the filled squares in the upper panel of fig .
[ fig.grad ] stand for the representative d / h value in the lism according to either linsky et al .
( 2005 ; ` high ' value ) or hbrard & moos ( 2003 ; ` low ' value ) .
the filled triangle in the same panel refers to the measurement by rogers et al .
( 2005 ) for the outer galactic disk .
the filled squares in the middle panel of fig .
[ fig.grad ] stand for the representative d / o ratio according to either hbrard & moos ( 2003 ; lower value ) or linsky et al .
( 2005 ; higher value ) . to obtain ( d / o)@xmath16 in this latter case , we divided ( d / h)@xmath16 from linsky et al .
( 2005 ) by the oxygen abundance of orion ( esteban et al .
2004 ) . at first glance
, it seems that gce models explain the lowest d / o value better than the highest one , while the opposite is true for d / h ( upper panel ) .
however , one should recall that the current oxygen abundance displays a non - negligible spread . indeed , several indicators hii regions , b - type stars , the neutral medium show that the local oxygen abundance varies in the range 8.3 @xmath24 log(o / h)+12 @xmath24 9.0 ( fig .
[ fig.grad ] , lower panel , and oliveira et al .
2006 , their table 12 ) .
so , in fig .
[ fig.grad ] , middle panel , we show also the d / o ratio recently measured by friedman et al .
( 2006 ) for the line of sight towards lse44 . along this line of sight , ( d / h)@xmath26 = 2.24@xmath27 @xmath10 10@xmath11 , consistent with the lower limit suggested by linsky et al .
( 2005 ) , and ( o / h)@xmath26 = 11.3@xmath28 @xmath10 10@xmath29 ( that is the highest o / h value measured locally by _ fuse _ ) .
now , within the errors , the model adopting the scalo ( 1986 ) imf and the schaller et al .
( 1992 ) stellar lifetimes is in satisfactory agreement with both the observed ` high ' ( d / h)@xmath16 and ` low ' ( d / o)@xmath16 , provided that ( d / o)@xmath16 is ` low ' as a consequence of both a high deuterium and a high oxygen abundance in the lism .
it is seen that the shape of the gradients slightly changes when changing the prescriptions about the imf and the stellar lifetimes . in the galactocentric distance range @xmath30 = 412 kpc the theoretical gradients of log(d / h ) , log(d / o ) , and log(o / h)+12 are 0.03 @xmath17 0.01 dex kpc@xmath31 , 0.08 dex kpc@xmath31 , and @xmath200.055 @xmath17 0.005 dex kpc@xmath31 , respectively . in the outermost regions of the disc the gradient of d / h flattens , while those of d / o and o / h steepen , owing to the lower and lower amount of gas processed by stars
outward in the disc . before concluding that , according to gce model predictions , one shall observe d / h ratios close to the primordial value at the largest radii in spiral galaxies , it is worth reminding that the model predictions depend on a number of uncertain free parameters . model a of chiappini et al .
( 2001 ) produces a fairly steep o / h gradient at the largest radii ( fig .
[ fig.grad ] , bottom panel ) , while the majority of the data suggest a flatter behaviour .
a flatter o / h gradient in the outer disc is easily obtained by suppressing the gas density threshold which regulates the star formation process during the preceding halo phase .
the solid lines in fig . [ fig.grad2 ]
stand for chiappini et al.s ( 2001 ) model c results .
this model is analogous to model a ( fig .
[ fig.grad ] , solid lines ) , except that it allows star formation in the halo to go on even when the gas density drops below the threshold value ( @xmath18 4 m@xmath13 pc@xmath32 ) .
the expected d / h ratio is now @xmath18 2.5 @xmath10 10@xmath11 at the outermost radii , i.e. below the adopted primordial value [ ( d / h)@xmath15 = 2.9 @xmath10 10@xmath11 for this model ] . also shown in fig .
[ fig.grad2 ] are the predictions from the model of tosi et al .
( in preparation ) , which is an updated version of the model by tosi ( 1988a , b ) adopting the kroupa et al .
( 1993 ) imf and the schaller et al .
( 1992 ) stellar lifetimes , and extending to the whole disc [ dashed lines : ( d / h)@xmath15 = 2.9 @xmath10 10@xmath11 ; dotted lines : ( d / h)@xmath15 = 2.3 @xmath10 10@xmath11 ] .
this model tends to predict higher oxygen abundances , hence lower d / o ratios , across the disc .
we conclude that , in principle , accurate measurements of both d / h and o / h in the outermost regions of spiral galaxies can allow us to discriminate among existing scenarios of galaxy formation and evolution .
in fact , only minor uncertainties related to the adopted stellar yields affect the interpretation of the data in this case , owing to the well - known nucleosynthetic origin of both species .
hence , observational efforts in this direction are highly desirable . in summary ,
notice that all the gce models discussed above favour minor deuterium variations [ @xmath34log(d / h)/@xmath35 0.02 dex kpc@xmath31 ] in the galactocentric distance range sampled by current observations ( @xmath37 811 kpc ) , hence favouring the highest ( d / h)@xmath38 determinations as representative of the local deuterium abundance .
in general , observations of deuterium abundances provide _ lower bounds _ on its primordial abundance
. however , at sufficiently high redshifts and low metallicities , one expects the primordial abundance of deuterium to reveal itself as a plateau . instead
, statistically significant scatter of the d / h measurements has been found in several high - redshift qso absorbers . before claiming that a dispersion around the mean value is actually present , one has to carefully check possible systematics affecting the abundance derivation .
it has been suggested that in the absorbers with the lowest hi column densities , interlopers might indeed contribute to the inferred di column densities , while in those with the highest hi column densities , interlopers might affect the wings of the hi lines ( steigman 2004 , 2006 , and references therein ) .
yet , it is worth noticing that the placement of the continuum and the fitting of the damping wings are closely related for dlas : if one gets the continuum wrong , it is unlikely that he could fit the line profile satisfactorily ( m. pettini , private communication ) . in this paper , we assume ( d / h)@xmath15 = ( 2.6 @xmath17 0.3 ) @xmath10 10@xmath11 , i.e. the weighted mean of d / h measurements towards qsos ( indirect determinations give a consistent value ; see discussions in sections [ sec.int ] and [ sec.dat ] )
. of more concern is the dispersion of local d / h data . in the lism , d / h is found to vary from @xmath18 0.5@xmath10 10@xmath11 to @xmath18 2.2@xmath10 10@xmath11 ( hbrad et al .
2005 ; linsky et al . 2005 ; williger et al . 2005 ; friedman et al . 2006 ; oliveira et al . 2006
; to quote only the most recent works )
. basically , two mechanisms are suggested in the current literature which might explain the observed variation ( pettini 2006 , and references therein ; see also lemoine et al .
1999 ; prochaska et al .
2005 ) : 1 . differing ism conditions along different sight lines , which determine different degrees of deuterium depletion on to dust grains ; 2 .
localized infall of unprocessed gas , which modifies the deuterium abundance in the surveyed region while leaving unchanged those of more abundant species , such as oxygen . in the first case
, one should recover the true local d / h in interstellar clouds which have been accelerated by supernova shocks , since the depletion of refractory elements is much reduced relative to that of the normal quiescent ism there ; in this case , the true local abundance of deuterium would be the highest observed one ( linsky et al .
2005 , and references therein ) . in the second case
, the true local abundance of deuterium would be the lowest observed one . in any case ,
the mean d / h value measured within the local bubble should not be considered any more as representative of the actual degree of astration suffered by deuterium in the solar neighbourhood . by assuming that the primordial abundance of deuterium is reasonably well known
, one can usefully bind the deuterium astration factor of the solar vicinity .
an astration factor ( by number ) either as low as @xmath39 or a factor of 23 higher is permitted by the observations . in this paper ,
we have computed a number of gce models for the solar vicinity as well as for the whole galactic disc and conclude that : 1 . only low astration factors ( not in excess of @xmath40 ) are compatible with gce requirements : our models predict ( d / h)@xmath16 = 1.42.0 @xmath10 10@xmath11 starting from a primordial value of ( d / h)@xmath15 = 2.6 @xmath10 10@xmath11 .
this is not surprising : in the absence of supernova - driven winds which efficiently remove the metal - rich ejecta of dying stars from the galaxy and/or in the absence of a peculiar imf , the relatively high present - day gas content and low metallicity are indicative of modest astration ( and , hence , modest d destruction ; see also tosi et al .
we emphasize here once again that the low astration factors are due to the combination of moderate star formation and continuous infall of gas which are needed in order to reproduce the available milky way data .
small variations in the predicted d astration factor are produced by changing the prescriptions on the imf and the stellar lifetimes .
in particular , by adopting the scalo ( 1986 ) imf and the schaller et al .
( 1992 ) stellar lifetimes we predict a d astration factor which clearly favours the low one suggested by linsky et al .
( 2005 ) . with this choice for the imf and the stellar lifetimes
, the model also gives a good fit to many other observational constraints for the galaxy ( romano et al .
3 . when the model is forced to reproduce the lowest d / h values observed locally , the agreement between model predictions and relevant observations is lost , for one or more observables ( the most relevant observational constraints being in this context the g - dwarf metallicity distribution , the mild increase of the overall metallicity of the ism from the time of sun s formation up to now , the present - day mass and gas surface densities , the present - day infall rate ; see also matteucci 2004 ) .
4 . in order to attain the lowest d / h values observed in the lism ,
the smooth , gentle decline which nicely accounts for the psc d abundance , must turn into a steeper behaviour during the latest phases of galaxy s evolution .
however , a large gas cycling through stars in the last @xmath185 gyr is unlikely , in the light of the small increase of the global metal abundance from the time of sun s formation up to now suggested by several independent indicators ( esteban et al .
2004 ; their table 15 and references therein ) .
5 . in principle
, joint observations of deuterium and oxygen abundances in the outermost regions of galactic discs can shed light on the mechanisms of spiral galaxy formation and evolution . in conclusion
, we favour a scenario in which d / h in the solar neighbourhood declines mildly during galaxy s evolution , due to the combined effect of moderate star formation and continuous infall of external gas . in this framework , the dispersion in the current local d / h data might be at least partly explained by different degrees of dust depletion along different lines of sight .
dr and mt wish to thank j. geiss , g. gloeckler , g. hbrard , j. linsky , and t. bania of the lola - ge team for useful conversations at the international space science institute in bern . the warm hospitality and the financial support at issi are gratefully acknowledged .
we are also grateful to g. steigman for stimulating discussions and to w. moos and m. pettini for interesting conversations about many observational aspects .
we thank j. linsky and m. pettini for providing us with a copy of their papers in advance of publication .
andr m. k. , oliveira c. m. , howk j. c. , et al . , 2003 ,
apj , 591 , 1000 asplund m. , grevesse n. , sauval a. j. , 2005 , in barnes t.g .
iii , bash f. n. , eds , asp conf .
, cosmic abundances as records of stellar evolution and nucleosynthesis .
, san francisco , vol . 336 , p. 25 audouze j. , lequeux j. , reeves h. , vigroux l. , 1976 , apj , 208 , l51 boesgaard a. m. , steigman g. , 1985 , ara&a , 23 , 319 burles s. , tytler d. , 1998a , apj , 499 , 699 burles s. , tytler d. , 1998b , apj , 507 , 732 cartledge s. i. b. , lauroesch j. t. , meyer d. m. , sofia u. j. , 2004 , apj , 613 , 1037 charbonnel c. , primas f. , 2005 , a&a , 442 , 961 chiappini c. , matteucci f. , gratton r. , 1997 , apj , 477 , 765 chiappini c. , matteucci f. , romano d. , 2001 , apj , 554 , 1044 chiappini c. , renda a. , matteucci f. , 2002 , a&a , 395 , 789 coc a. , vangioni - flam e. , descouvemont p. , adahchour a. , angulo c. , 2004 , apj , 600 , 544 crighton n. h. m. , webb j. k. , carswell r. f. , lanzetta k. m. , 2003 , mnras , 345 , 243 crighton n. h. m. , webb j. k. , ortiz - gil a. , fernndez - soto a. , 2004 , mnras , 355 , 1042 epstein r. l. , 1977 , apj , 212 , 595 epstein r. l. , arnett w. d. , schramm d. n. , 1976 , apjs , 31 , 111 esteban c. , peimbert m. , garca - rojas j. , ruiz m. t. , peimbert a. , rodrguez m. , 2004 , mnras , 355 , 229 fields b. d. , 1996 , apj , 456 , 478 fields b. d. , olive k. a. , silk j. , cass m. , vangioni - flam e. , 2001 , apj , 563 , 653 friedman s. d. , hbrard g. , tripp t. m. , chayer p. , sembach k. r. , 2006 , apj , 638 , 847 galli d. , palla f. , ferrini f. , penco u. , 1995 , apj , 443 , 536 geiss j. , gloeckler g. , 1998 , space sci .
, 84 , 239 geiss j. , reeves h. , 1972 , a&a , 18 , 126 hbrard g. , moos h. w. , 2003 , apj , 599 , 297 hbrard g. , tripp t. m. , chayer p. , et al . , 2005 , apj , 635 , 1136 jenkins e. b. , tripp t. m. , woniak p. r. , sofia u. j. , sonneborn g. , 1999 , apj , 520 , 182 jrgensen b. r. , 2000 , a&a , 363 , 947 jura m. , 1982 , in kondo y. , ed .
, advances in ultraviolet astronomy .
washington , nasa , p. 54
kirkman d. , tytler d. , suzuki n. , omeara j. m. , lubin d. , 2003 , apjs , 149 , 1 kotoneva e. , flynn c. , chiappini c. , matteucci f. , 2002 , mnras , 336 , 879 kroupa p. , tout c. a. , gilmore g. , 1993 , mnras , 262 , 545 lemoine m. , audouze j. , ben jaffel l. , et al . , 1999 ,
new astr . , 4 , 231 levshakov s. a. , dessauges - zavadsky m. , dodorico s. , molaro p. , 2002
, apj , 565 , 696 linsky j. l. , 1998 , space sci . rev . ,
84 , 285 linsky j. l. , draine b. t. , moos h. w. , et al . , 2005 ,
baas , 37 , 1444 lubowich d. a. , pasachoff j. m. , balonek t. j. , millar t. j. , tremonti c. , roberts h. , galloway r. p. , 2000 ,
nat , 405 , 1025 maeder a. , meynet g. , 1989 , a&a , 210 , 155 mahaffy p. r. , donahue t. m. , atreya s. k. , owen t. c. , niemann h. b. , 1998 , space sci .
, 84 , 251 matteucci f. , 2004 , in mcwilliam a. , rauch m. , eds , carnegie obs .
ser . , origin and evolution of the elements .
cambridge univ . press , cambridge , p. 87
matteucci f. , romano d. , molaro p. , 1999
, a&a , 341 , 458 meyer d. m. , jura m. , cardelli j. a. , 1998 , apj , 493 , 222 moos h. w. , sembach k. r. , vidal - madjar a. , et al .
, 2002 , apjs , 140 , 3 olive k. a. , steigman g. , walker t. p. , 2000
, 333 , 389 olive k. a. , skillman e. d. , 2004 , apj , 617 , 29 oliveira c. m. , dupuis j. , chayer p. , moos h. w. , 2005 , apj , 625 , 232 oliveira c. m. , moos h. w. , chayer p. , kruk j. w. , 2006 , apj in press ( astro - ph/0601114 ) omeara j. m. , tytler d. , kirkman d. , suzuki n. , prochaska j. x. , lubin d. , wolfe a. m. , 2001 , apj , 552 , 718 pettini m. , 2006 , in sonneborn g. , moos h. w. , andersson b .-
, eds , asp conf . ser .
, astrophysics in the far ultraviolet .
, san francisco , vol . 348 , in press ( astro - ph/0601428 ) pettini m. , bowen d. v. , 2001 , apj , 560 , 41 prantzos n. , 1996 , a&a , 310 , 106 prantzos n. , ishimaru y. , 2001 , a&a , 376 , 751 prochaska j. x. , tripp t. m. , howk j. c. , 2005 , apj , 620 , l39 prodanovi t. , fields b. d. , 2003 , apj , 597 , 48 rocha - pinto h. j. , maciel w. j. , 1996 , mnras , 279 , 447 rogers a. e. e. , dudevoir k. a. , carter j. c. , fanous b. j. , kratzenberg e. , bania t. m. , 2005 , apj , 630 , l41 rogerson j. b. , york d. g. , 1973 , apj , 186 , l95 romano d. , chiappini c. , matteucci f. , tosi m. , 2005a , a&a , 430 , 491 romano d. , chiappini c. , matteucci f. , tosi m. , 2005b , in corbelli e. , palla f. , zinnecker h. , eds , assl , the initial mass function 50 years later .
springer , dordrecht , vol .
327 , p. 231
( astro - ph/0409474 ) romano d. , tosi m. , matteucci f. , chiappini c. , 2003 , mnras , 346 , 295 scalo j. m. , 1986 , fund .
cosmic phys . , 11 , 1 schaller g. , schaerer d. , meynet g. , maeder a. , 1992 , a&as , 96 , 269 scully s. , cass m. , olive k. a. , vangioni - flam e. , 1997 , apj , 476 , 521 sembach k. r. , wakker b. p. , tripp t. m. , et al . , 2004 , apjs , 150 , 387 sonneborn g. , tripp t. m. , ferlet r. , jenkins e. b. , sofia u. j. , vidal - madjar a. , woniak p. r. , 2000 , apj , 545 , 277 spergel d. n. , verde l. , peiris h. v. , et al . , 2003 ,
apjs , 148 , 175 steigman g. , 1994 , mnras , 269 , l53 steigman g. , 2004 , in freedman w. l. , ed . , carnegie obs .
ser . , measuring and modeling the universe .
cambridge , cambridge univ . press ,
p. 169 steigman g. , 2006 , int .
journal of modern phys .
e , 15 , 1 steigman g. , tosi m. , 1992 , apj , 401 , 150 steigman g. , tosi m. , 1995 , apj , 453 , 173 tosi m. , 1988a , a&a , 197 , 33 tosi m. , 1988b , a&a , 197 , 47 tosi m. , 1996 , in leitherer c. , fritze - von - alvensleben u. , huchra j. , eds , asp conf .
ser . , from stars to galaxies : the impact of stellar physics on galaxy evolution .
soc . pac .
, san francisco , vol . 98 , p. 299
tosi m. , steigman g. , matteucci f. , chiappini c. 1998 , apj , 498 , 226 vangioni - flam e. , olive k. a. , prantzos n. 1994 , apj , 427 , 618 weinberg s. , 1977 , the first three minutes . a modern view of the origin of the universe .
london : andre deutsch williger g. m. , oliveira c. , hbrard g. , dupuis j. , dreizler s. , moos h. w. , 2005 , apj , 625 , 210 wood b. e. , linsky j. l. , hbrard g. , williger g. m. , moos h. w. , blair w. p. , 2004 ,
apj , 609 , 838 wyse r. f. g. , gilmore g. , 1995 , aj , 110 , 2771 | estimates of the interstellar deuterium abundance span a wide range of values . until recently , it was customary to adopt the abundance of deuterium measured in the local bubble as representative of the local one .
now , it is becoming unclear whether the true local deuterium abundance is significantly higher or lower than this value , depending on the interpretation given to current data .
it is important to deal with the issue of the deuterium variation and see whether it challenges our current understanding of the galaxy evolution . to this aim
, we study the evolution of deuterium in the framework of successful models for the chemical evolution of the milky way able to reproduce the majority of the observational constraints for the solar neighbourhood and for the galactic disc .
we show that , in the framework of our models , the lowest d / h values observed locally can not be explained in terms of simple astration processes occurring during the galaxy evolution .
indeed , the combination of a mild star formation and a continuous infall of unprocessed gas required to fit all the available observational data allows only a modest variation of the deuterium abundance from its primordial value .
therefore , we suggest that depletion of deuterium on to dust grains is the most likely physical mechanism proposed so far to explain the observed dispersion in the local data .
= = = = = = = = # 1 # 1 # 1 # 1 @mathgroup@group @mathgroup@normal@groupeurmn @mathgroup@bold@groupeurbn @mathgroup@group @mathgroup@normal@groupmsamn @mathgroup@bold@groupmsamn = `` 019 = ' ' 016 = `` 040 = ' ' 336 = " 33e = = = = = = = = # 1 # 1 # 1 # 1 = = = = = = = = [ firstpage ] cosmology : observations galaxy : abundances galaxy : evolution ism : abundances |
it is quite plausible that the structure of space - time in the vicinity of planck scale is described by a fuzzy quantum space - time " . as shown by doplicher _
@xcite , this quantum structure of space - time , where the space - time coordinates are operator - valued and satisfy a non - commuting coordinate algebra , can be one of the most plausible ways to prevent the gravitational collapse arising from the attempt to localise a space - time event within a planck length scale . since making a guess regarding the non - commutative structure of the coordinate algebra
is difficult , one tries to postulate a simple structure of this form and study the geometry of the resulting non - commutative spaces .
the simplest examples are the 2d moyal plane ( @xmath4 ) : @xmath5 = i \theta , \label{mp}\ ] ] and fuzzy @xmath6 : @xmath7 = i \lambda \epsilon_{ijk } \hat{x}_k .
\label{fs}\ ] ] fuzzy sphere @xmath8 corresponds to a 2d subspace of @xmath6 with radius quantized as @xmath9 @xcite . because of the inherent uncertainty relations satisfied by these coordinates the usual concepts like points , lines etc loose their meaning in these kinds of spaces .
it thus becomes essential to use the mathematical formalism of non - commutative geometry(ncg ) as developed by connes @xcite to study the geometry of such spaces .
+ we would like to mention in this context that connes himself , along with his other collaborators , have invested a lot of effort in formulating a completely new mathematical framework to describe the standard model of particle physics , by invoking the so - called almost commutative spaces " - built out of the usual commutative ( euclideanised ) 4d curved space - time .
this is expected to describe physics upto gut scale ( @xmath10 @xmath11 ) . however , as we mentioned in the beginning , one perhaps has to take quantum space - times , i.e. spaces where the coordinates become operators satisfying a non - commutative algebra and for which ( [ mp ] ) and ( [ fs ] ) provide prototype examples , more seriously in the higher energy scales - like in the vicinity of planck scales .
( see for example @xcite for a review ) .
it is only recently that the above mentioned mathematical framework of ncg , _ a la _ connes , has been used to compute the spectral distances on quantized spaces " like the moyal plane , fuzzy sphere etc . @xcite,@xcite-@xcite . on the other hand ,
an algorithm was devised in @xcite to compute this distance using the hilbert - schmidt operatorial formulation of quantum mechanics @xcite , @xcite .
this hilbert - schmidt operatorial formulation has the advantage that it bypasses the use of any star product , like moyal or voros , and is therefore free from any ambiguities that can arise from there @xcite .
furthermore , it has an additional advantage that the above - mentioned algorithm / formula is adaptable to this hilbert - schmidt operatorial formulation and infinitesimally it has essentially the same structure as that of the induced metric from the hilbert space inner product , obtained by provost and vallee @xcite ( see also @xcite ) , when expressed in terms of the density matrix @xcite and yields for the moyal plane the correct distance in the harmonic oscillator basis " and the flat metric in the coherent state , upto an overall numerical factor .
however , the corresponding finite distance can not simply be obtained just by integrating " along the geodesics , as the very concept of geodesics in the conventional sense ( i.e. like on a commutative differentiable manifold ) may not exist at all . this motivates us to undertake the task of extending our algorithm of @xcite so that one is able to compute finite distances as well .
this new formula is shown to involve the ` transverse ' component ( @xmath12 ) , in addition to the longitudinal " component ( @xmath13 ) where @xmath14 is the difference between normal states represented by density matrices , as in @xcite .
this in turn shows that the formula of @xcite actually corresponds only to the lower bound of the distance and not the actual distance .
indeed , on the way to our derivation of the generalised formula in section [ sec5 ] we point out a flaw , commented on in @xcite through a counter example , in the analysis in @xcite where the same expression was shown to correspond to the upper bound as well .
it should , however , be mentioned that the counter example in @xcite does not satisfy the boundedness condition imposed in @xcite .
this error was not serious in examples studied previously to compute infinitesimal distances as these distances coincided with the exact distance for discrete states and differed from the exact distance by a numerical factor for the coherent states for both @xmath4 and @xmath8 @xcite .
however , as one can easily see , a straightforward calculation to compute _ finite _ distances , using the same formula does not yield any sensible result indicating a non - trivial role for the transverse component @xmath12 . in the generic case
there can be many choices of @xmath12 to a given @xmath14 ( in fact an infinite number of them in @xmath4 ) .
consequently the computation of the infimum involving @xmath12 occurring in the revised formula turns out to be quite non - trivial .
one therefore has to try with different forms of @xmath12 and improve the estimate of the distance as best as one can . + on the other hand , we can emulate and adapt the approach of @xcite to our hilbert - schmidt operatorial formulation to obtain an upper bound to the distance and then look for an optimal element in the algebra saturating this upper bound . in case
at least one such optimal element can be identified , this upper bound itself can obviously be recognised as the true distance .
otherwise one has to be content with the above - mentioned best possible estimate only .
in fact this paper deals with an interplay of both the approaches , as they seem to complement each other in some sense .
this brings out some stark differences between @xmath4 and @xmath8 . particularly for
@xmath8 corresponding to any finite @xmath0-representation of @xmath1 one can not define a geodesic in the conventional sense ; it reduces to commutative @xmath15 only in the limit @xmath16 .
the distance turns out to be much smaller than the geodesic distance of @xmath15 , the latter coinciding with the above - mentioned upper bound .
further in the case of maximal non - commutativity i.e. for @xmath2 , even the finite distance is shown to coincide exactly with the lower bound with @xmath12 playing no role and any pair of pure states is shown to be interpolated by a one - parameter family of mixed states , lying in the interior of @xmath8 .
the distance between any mixed state to the nearest pure state can be taken as a measure of the ` mixedness ' .
+ the whole analysis , particularly the `` ball condition '' is carried out in the eigen - spinor basis furnished by the respective dirac operators .
this is clearly a natural choice of basis , as the ball condition involves the dirac operator .
this definitely simplifies the computations considerably and even allows us to study the geometry of @xmath8 for @xmath17 , apart from reproducing many of the existing results in the literature , albeit with the help of _
mathematica_. however the corresponding analysis for @xmath18 remains quite intractable , even with _
mathematica_. + the paper is organised as follows . in section [ sec2 ]
, we provide a brief review of the hilbert - schmidt operatorial formulation of non - commutative quantum mechanics on the 2d - moyal plane and fuzzy sphere and the associated spectral triples , introduced in @xcite - required to study the geometrical aspects of them .
we also introduce the corresponding dirac operators and their eigen - spinors for both non - commutative spaces .
to begin , we revisit the derivation of the formula given in @xcite in section [ sec5 ] and derive the corrected form .
we then provide a computation of ( finite ) distances on the moyal plane in the coherent and the harmonic oscillator " basis in sections [ sec3 ] and [ sec4 ] , respectively .
we then proceed onto the case of the fuzzy sphere in quite the same way adopted as in the moyal plane in section [ sec6 ] .
we note some fundamental differences from the case of the moyal plane and adopt a different algorithm using the dirac eigen - spinors and study the @xmath2 and @xmath17 representations of the fuzzy sphere algebra .
we finally conclude in section [ sec_con ] .
the hilbert - schmidt operatorial formulation of non - commutative quantum mechanics ( ncqm ) on the 2d moyal plane @xcite , described by the non - commutative heisenberg algebra i.e. the co - ordinate algebra augmented by the following commutation relations involving linear momenta operators @xmath19 satisfying ( in units of @xmath20 ) @xmath21 = i \delta_{ij } ~~;~~ [ \hat{p}_i , \hat{p}_j ] = 0 \label{ha}\ ] ] begins by introducing an auxiliary hilbert space @xmath22 furnishing a representation of just the coordinate algebra ( [ mp ] ) . in this particular situation , since the algebra ( [ mp ] ) is isomorphic to the algebra @xmath23 = i$ ] of a 1d harmonic oscillator , @xmath22 can be constructed exactly in the same manner i.e. @xmath24 where @xmath25 and @xmath26 are the respective lowering and raising operators satisfying @xmath27 = 1 $ ] and the `` ground state '' @xmath28 satisfy @xmath29 .
this @xmath22 , however , can not furnish a representation of the linear momentum operators @xmath30 . as shown in @xcite ,
we need to introduce the space @xmath31 comprised of hilbert - schmidt operators acting on @xmath22 . loosely speaking ,
these are essentially the trace - class bounded set of operators and forms a hilbert space on its own referred to as the quantum hilbert space .
physical states @xmath32 ( denoted by round kets , rather than angular kets @xmath33 ( [ hc ] ) ) , having generic forms as @xmath34 is defined as @xmath35 where the subscript @xmath22 in indicates that the trace has to be computed over @xmath22 .
+ we reserve @xmath36 to denote the hermitian conjugation on @xmath22 ( [ hc ] ) , while @xmath37 denotes the hermitian conjugation on @xmath31 .
note that @xmath31 has a natural tensor product structure as @xmath38 ( @xmath39 being the dual of @xmath22 ) , enabling one to express the elements of @xmath31 in the form @xmath40 or their linear spans .
one can refer to @xmath41 ( respectively @xmath42 ) as the left ( respectively right ) hand sector . +
a unitary representation of the non - commutative heisenberg algebra ( [ mp ] ) and ( [ ha ] ) is obtained by the following actions : @xmath43 ) \label{urep}.\ ] ] note that we are using capital letters ( without hats ) to distinguish them from the operators acting on @xmath22 .
apart from the harmonic oscillator " basis @xmath44 , introduced in , satisfying @xmath45 , one can also introduce normalized coherent states in terms of a dimensionless complex number @xmath46 as @xmath47 where @xmath48 is a unitary operator furnishing a projective representation of the translation group .
these states provide an over - complete basis in @xmath22 .
the corresponding non - orthogonal projection operator @xmath49 is an operator acting on @xmath22 and is an eigenstate of @xmath50 ( the representation of @xmath51 in @xmath31 ) : @xmath52 .
this represents a quantum state with maximal localization , where the position measure must now be interpreted in the context of a weak measurement ( positive operator - valued measurement , povm ) rather than a strong measurement ( projective valued measurement , pvm ) .
as was shown in @xcite this quantum hilbert space @xmath31 has a built - in structure of an algebra in the sense that under the multiplication map @xmath53 the usual operator product of any arbitrary pair of elements @xmath54 yields another element of @xmath31 : @xmath55 in the coherent state i.e. @xmath56 is obtained by composing the respective representations of individual states i.e. @xmath57 and @xmath58 by using the voros star product @xmath59 : @xmath60 furthermore , it was shown in @xcite that @xmath61 also provides an over - complete basis on @xmath31 - the counterpart of @xmath62 , provided that they are composed using the above mentioned voros star product : @xmath63 correspondingly , one can introduce the unnormalized projection operators @xmath64 which are positive ( i.e. @xmath65 ) but unnormalized ( @xmath66 ) and non - orthogonal .
they , however , form a complete basis and therefore provide a povm ( positive operator valued measure ) that one can use to provide a consistent probability interpretation by assigning the probability density @xmath67 of finding the outcome of a position measurement to be @xmath68 if the system is in a state described by the density matrix @xmath69 .
in particular , if @xmath70 is a pure state density matrix , then @xmath71 which clearly goes into the corresponding commutative result in the limit @xmath72 .
+ we briefly mention in this context that , just as the basis @xmath61 has a natural association with the voros star product , it was shown in @xcite that one can like - wise construct an appropriate basis , which is naturally associated with the moyal star product @xmath73 ( in the cartesian @xmath74 basis ) .
however , this basis is somewhat unphysical in the sense that it is the common eigenstate of mathematically constructed unphysical commuting position - like observables @xmath75 , obtained by taking the average of left and right actions of @xmath76 as @xmath77 satisfying @xmath78 = 0 $ ] .
furthermore , they do not conform to the requirement of a povm , unlike the voros case discussed above . besides , the representation of the quantum states in this moyal basis lead to a different class of functions than the corresponding voros ones .
this implies that the equivalence between different choices of star products can not be guaranteed , especially in a path integral formulations , without taking due care of this point .
hence the classical configuration space @xmath90 of a fuzzy space of type ( [ fs ] ) is given by @xmath91 each @xmath0 corresponds to the fixed sphere of radius @xmath92 such that for a fixed @xmath0 the corresponding hilbert sub - space is a @xmath93-dimensional sub - space @xmath94 however , since the quantum hilbert space , in which the physical states are represented , consists of those operators generated by coordinate operators only and since these commute with the casimir the elements of the quantum hilbert space must in addition commute with the casimir , i.e. , must be diagonal in @xmath0 .
therefore , the quantum hilbert space @xmath31 of the fuzzy space of type ( [ fs ] ) splits into the direct sum of sub - spaces as follows : @xmath95 where @xmath96 represents the quantum hilbert space of a fuzzy sphere with fixed radius @xmath92 .
the coherent states are the quantum states which saturate the heisenberg s uncertainty relation .
each coherent state @xmath97 is obtained by the action of a unitary operator @xmath98 on the highest weight state @xmath99 , satisfying @xmath100 , and is specified by a complex number @xmath101 .
it is well - known that @xmath102 as @xmath103 and hence implies that coherent states have the properties of classical states . here
we review the construction of generalized coherent states of the @xmath1 group @xcite such that the non - commutative analog of the homogeneous space of the fuzzy sphere can be constructed by using the perelomov s @xmath1 coherent states @xcite .
note that the heisenberg uncertainty relations for the fuzzy sphere is given by @xmath104 for both bases build on @xmath99 and @xmath105 .
thus , we can choose the highest weight state as @xmath106 and the perelomov @xmath1 coherent states can be obtained by the action of a representation @xmath107 of @xmath108 on @xmath99 .
note that any action of @xmath84 on the state @xmath99 does not change it so that the group element generated by @xmath84 is the stability subgroup of @xmath1 .
this implies that the set of generalized coherent states of @xmath1 is topologically isomorphic to the coset space @xmath109 .
however , geometrically it reduces to @xmath15 only in the limit @xmath110 . we know that the operator @xmath111 can be expressed in terms of euler angles as @xmath112 @xcite . for @xmath15 , @xmath113 such that a generic coherent state specified by a point on @xmath15 is given by @xmath114 here
, @xmath101 is the stereographic variable of @xmath15 projected from the south pole to the complex plane @xmath115 and @xmath116 .
the basic ingredients of studying the geometrical content , particularly connes distance on @xmath22 is facilitated by specifying the spectral triple @xmath117 where the algebra @xmath118 is identified with @xmath31 as mentioned above ( [ multi - map ] ) and the hilbert space @xmath119 is the module and identified with @xmath120 so that it represents the space of appropriate spinors " where a typical element @xmath121 acts on @xmath122 through the diagonal representation @xmath123 as @xmath124 the dirac operator , as explained in detail in @xcite ( see also the appendix a ) , is identified as @xmath125 having a well - defined left action on @xmath120 as @xmath126 the eigen - spinors of the dirac operator for the moyal plane are given by the following , @xmath127 with the eigenvalues @xmath128 for any state @xmath129 given by @xmath130 they furnish a complete and orthonormal basis for @xmath120 , so that the resolution of the identity takes the form @xmath131 for the fuzzy sphere the spectral triple was already constructed in @xcite
. we will be working with the same one here .
the spectral triple consists of the algebra @xmath132 , the hilbert space @xmath133 , and the dirac operator @xmath134 where @xmath135 is related to @xmath136 by @xmath137 and @xmath138 are the usual ladder operators .
as mentioned earlier we will make use of the dirac eigen - spinors @xmath139 , which for the fuzzy sphere are given by @xcite @xmath140 with @xmath141 and @xmath142 .
the respective eigenvalues are given by @xmath143 for any @xmath144 or @xmath145 .
thus , the eigenvalues for a particular @xmath0 representation is independent of @xmath144 or @xmath145 and is responsible for a @xmath146-fold degeneracy in the positive eigenvalue sector and a @xmath147-fold degeneracy in the negative eigenvalue sector .
this can be understood from the tensor product structure of dirac operator and the clebsch - gordon decomposition of a tensor product of a pair of @xmath1 representations .
for example , if @xmath148 in corresponds to the @xmath2 representation i.e. @xmath149 , then it will split into the direct sum of @xmath17 ( triplet ) and @xmath150 ( singlet ) representations of three and one dimension , respectively .
states @xmath151 are positive linear functionals of norm 1 over @xmath118 .
pure states play a rather fundamental role and are defined as those functionals that can not be written as a convex linear combination of two other functionals .
the connes spectral distance between two states is then defined by @xmath152 \|_{op } \leq 1 \right\rbrace,\\ \| \mathcal{a } \|_{op } = \sup_{\phi \in \mathcal{h } } \frac{\| \mathcal{a\phi } \|}{\| \phi \|}.\\ \end{split}\end{aligned}\ ] ] an algorithm , adaptable to this hilbert - schmidt operatorial formulation and subject to the following conditions : * the states @xmath153 are normal states ( see @xcite for the definition ) , * the states @xmath151 and @xmath154 are separately bounded on @xmath50 , i.e. @xmath155 and @xmath156 , @xmath157 , * let @xmath158\|_{op } = 0 \right\rbrace$ ] , then the states @xmath151 , @xmath154 are such that @xmath159 , @xmath160 , was devised to compute connes spectral distance @xcite .
the present analysis will also assume the above conditions .
the first two conditions are actually quite mild and just imply that a generic state @xmath151 can be represented by a density matrix @xmath161 .
furthermore , if the state @xmath151 is pure , then the corresponding density matrix @xmath161 too will be pure . more specifically , to illustrate through an example , let us consider the case of the moyal plane @xmath162 - parametrized by the complex number @xmath101 .
it has a one - to - one correspondence with the coherent state @xmath163 or more precisely the density matrix @xmath164 .
like - wise , the harmonic oscillator " state @xmath165 is associated with the density matrix @xmath166 . or @xmath167 are density matrices from the perspective of @xmath22 and belong to @xmath31 . they should not be confused with real quantum density matrices , which should be constructed by taking outer products of states @xmath168 as @xmath169 .
the fact that @xmath170 allows us to treat them as vectors , facilitating the analysis of the present paper .
this is precisely the advantage of this hilbert - schmidt operatorial formulation . ]
note that both @xmath171 and @xmath167 can be regarded as pure states in the @xmath172-algebraic framework and @xmath173 can indeed be identified with an involutive algebra , which is a dense sub - algebra of a @xmath172-algebra , where the hermitian conjugation ( @xmath36 ) plays the role of involution operator .
since both @xmath161 and @xmath174 are elements of @xmath175 , the connes distance between a pair of states , now represented by density matrices @xmath161 and @xmath176 , can be recast in terms of the inner product as @xmath177 on the other hand the third condition implies a certain irreducibility condition , as explained in @xcite , @xcite . in this context , it is worthwhile to recall another important role of the third condition in the present analysis .
if this condition were to be violated then we can find an @xmath178 such that @xmath179 .
however , since the ball condition places no constraint on @xmath180 , it is clear that no upper bound exists for this distance function and the resulting distance diverges .
when this condition holds , the spectral distance is always finite .
this point was illustrated through the example of the @xmath181 model in @xcite .
this example , which was first considered in @xcite , is described by the following spectral triple : @xmath182 note that the dirac operator has been written here in diagonal form with @xmath183 , @xmath184 being its eigenvalues .
it was shown there that the space of pure states corresponds to @xmath185 . now with @xmath186 , the dirac operator will become proportional to the identity operator , so that @xmath187 \|_{\text{op } } = 0 ~\forall~a \in m_2(\mathds{c})$ ] holds trivially and no constraint is imposed by the ball condition on any @xmath188 , giving rise to infinite distance between any pair of pure states . even setting @xmath189 subsequently , one finds that only parts of the algebra and more specifically the off - diagonal elements @xmath190 of @xmath191 gets constrained as @xmath192 , whereas no constraint is imposed on the real diagonal elements @xmath193 and @xmath194 .
consequently , the distance between any pair of states , belonging to different latitudes diverges and distance between those belonging to the same latitude can only be finite and calculated .
furthermore , note that one may also construct states where the second condition is violated i.e. @xmath195 and @xmath196 are not separately bounded @xmath197 . in this case
the spectral distance will also diverge .
indeed , such states were constructed explicitly in @xcite .
it is a simple matter to verify that the pure state defined by the unit vector @xmath198 in eq(3.18 ) of @xcite is not bounded on @xmath50 , but that elements of the algebra with components of the form given in eq(3.16 ) of @xcite lead to divergent results .
we are not interested in this kind of pathological states and our entire analysis will be restricted to the cases where the above mentioned conditions hold .
+ in the next section we take up the issue of this same computation , but for finite distances , by devising an appropriate and a more general algorithm that also corrects the error in @xcite .
as mentioned earlier , knowledge of the infinitesimal distance is not enough to compute the finite distance between a pair of finitely separated points by integrating along the geodesic connecting these points , if the concerned non - commutative space does not allow one to define a geodesic in the conventional sense .
in fact , as we shall see later , although the moyal plane @xmath4 allows one to identify the straight line as the geodesic just like the commutative @xmath199 , the great circle in the case of fuzzy sphere @xmath8 can not be identified in this manner .
as mentioned earlier , we consider the particular case where the states in are normal and bounded so that they are representable by density matrices . with this the connes
distance function becomes @xmath200 where , @xmath201 is the density matrix associated with the state @xmath151 so that @xmath202 and @xmath203 we also introduce the following subsets : @xmath204\|_{\text{op } } \ne 0\ } , \label{3.1.6}\ ] ] @xmath205\|_{\text{op } } \le 1\ } , \label{3.1.7}\ ] ] @xmath206 following the line of reasoning as in @xcite .
+ first we find the lower bound . for this ,
let us make an assumption that the optimal element @xmath207 is aligned in a direction which is either parallel or anti - parallel to @xmath14 .
equivalently @xmath208 can be taken to be proportional to @xmath14 .
we thus consider the one - parameter family of algebra elements @xmath209||_{op } } \right\rbrace \subset b^\prime.\ ] ] taking the extremal element @xmath210\|_{\text{op}}}$ ] yields the lower bound as @xmath211\|_{\text{op } } } .
\label{3.1.9}\ ] ] the same lower bound can be obtained alternatively in a more rigorous manner by noting that the trace - norm of any element within the ball @xmath212 is bounded above . to see this ,
consider an element @xmath213 s.t .
@xmath214\|_{\text{op } } \le 1 $ ] .
now writing @xmath215 in terms of the unit vector " @xmath216 satisfying @xmath217 , allows us to extract @xmath218 out of this inequality to write @xmath219\|_{\text{op } } \le 1 $ ] yielding @xmath220\|_{\text{op}}}.\ ] ] therefore @xmath221 is bounded above by @xmath222\|_{\text{op}}. \label{3.1.10}\ ] ] we can now decompose @xmath216 in a longitudinal " ( @xmath223 ) and transverse " ( @xmath224 ) component as @xmath225 , where @xmath226 and @xmath227 , is taken to be orthogonal to @xmath223 i.e. @xmath228 and corresponds to a unit vector in the plane formed by @xmath216 and @xmath223 .
it is clear from , that we can choose @xmath229 to be an acute angle i.e @xmath230 ensuring that both @xmath231 , @xmath232 are positive . note that here @xmath233
since we want @xmath234 otherwise @xmath235 in would collapse to zero .
we can then re - write @xmath236\|_{\text{op}}$ ] , as in @xcite and invoke the triangular inequality to get @xmath237\|_{\text{op } } = \|\left [ \mathcal{d},\cos\theta \ , \pi(\widehat{\delta\rho } ) + \sin\theta \ , \pi(\widehat{\delta\rho}_{\perp } ) \right]\|_{\text{op } } \le |cos\theta| \|
\left [ \mathcal{d},\pi(\widehat{\delta\rho } ) \right ] \|_{\text{op } } + |\sin\theta| \| \left [ \mathcal{d},\pi(\widehat{\delta\rho}_{\perp } ) \right ] \|_{\text{op}}. \label{3.1.10.1}\ ] ] this , however , just means that the infimum of the l.h.s is bounded from above as @xmath238\|_{\text{op } } \le \inf_{\theta \in [ 0 , \pi / 2 ) } |cos\theta|\|[\mathcal{d},\pi(\widehat{\delta\rho})]\|_{\text{op } } + |\sin\theta| \|[\mathcal{d},\pi(\widehat{\delta\rho}_{\perp})]\|_{\text{op}}= \min\{\|[\mathcal{d},\pi(\widehat{\delta\rho})]\|_{\text{op } } , \|[\mathcal{d},\pi(\widehat{\delta\rho}_{\perp})]\|_{\text{op}}\}. \label{3.1.11}\ ] ] here , of course , the mod over @xmath231 and @xmath239 are quite redundant , as @xmath229 is an acute angle as we have mentioned above . since @xmath233 , we are forced to identify @xmath240 \|_{\text{op } } + |\sin\theta| \| [ \mathcal{d},\pi(\widehat{\delta\rho}_{\perp } ) ] \|_{\text{op } } = \|[\mathcal{d},\pi(\widehat{\delta\rho } ) ] \|_{\text{op}}.\ ] ] thus , using , , we observe that @xmath241 is just bounded from above by @xmath242\|_{\text{op}}$ ] @xmath243\|_{\text{op } } , \label{3.1.11.1}\ ] ] but can not be identified , as was done in @xcite , with it _ in general_. thus the rhs of was erroneously identified as an upper bound in @xcite .
recall in this context that by definition of the infimum , it should rather correspond to the highest lower bound say @xmath244 , satisfying @xmath245 + \sin\theta [ \mathcal{d } , \pi(\widehat{\delta\rho}_{\perp } ) ] \|_{\text{op } } ~\forall~~ \delta\rho_{\perp } \in w , ~\theta \in [ 0 , \pi / 2).\ ] ] since the determination of a general formula to obtain @xmath244 is difficult , we have to be content with just writing @xmath246
\|_{\text{op } } = \inf_{\substack{\delta\rho}_{\perp } \in w \\ \theta
\in [ 0 , \pi / 2 ) } \| \cos\theta [ \mathcal{d } , \pi(\widehat{\delta\rho } ) ] + \sin\theta [ \mathcal{d } , \pi(\widehat{\delta\rho}_{\perp } ) ] \|_{\text{op } } \label{3.1.12}\ ] ] and find @xmath241 case by case .
returning to our derivation , we begin by rewriting the infinitesimal connes distance function , as @xmath247 where @xmath248 is given by @xmath249 from it is clear that the connes distance function can only depend on @xmath14 , i.e. @xmath265 it is also elementary to see that for any unitary transformation @xmath266 @xmath267 this also implies , from , that @xmath268 in general we note from and that @xmath269 and @xmath270 depend on the direction of @xmath14 , in the sense that even if @xmath271 , @xmath272 .
however , if @xmath271 implies @xmath273 , this dependence disappears using , @xmath270 is a constant as @xmath274 and @xmath223 both have norm one and @xmath275 .
this is the case for the coherent state basis in the moyal plane , where equality of the trace norms implies that @xmath276 and @xmath14 differ by a rotation of the form @xmath277 .
this explains why the connes distance on the moyal plane is proportional to the trace norm , which is simply the euclidean distance , infinitesimally and for finite distances .
we corroborate this result in the next section through a more explicit calculation . more generally , it is not difficult to verify that @xmath271 implies @xmath273 if @xmath14 and @xmath276 are the difference of two orthogonal pure states , in which case @xmath270 is again a constant .
this , and more general scenarios under which this holds will be explored elsewhere .
however , this clearly can not hold in general , but when this is the case , it readily follows that @xmath248 , as determined from ( [ 3.1.13 ] ) indeed corresponds to a numerical constant and upto this constant the metric @xmath278 that can be read off from the infinitesimal distance ( [ newb1 ] ) , is indeed given by the provost - vallee form i.e. @xmath279 @xcite in the coherent state basis in particular . in the case of the moyal plane , this readily yields a flat metric , as in @xcite , so that the straight lines are expected to play the role of geodesics .
indeed this fact was used implicitly in the parametrization below . on the other hand , for the case of the fuzzy sphere ,
although the metric is that of commutative sphere - upto an overall numerical factor , it turns out that the finite distance is quite different from the geodesic ( great circle ) .
indeed , for the @xmath2 representation of @xmath1 i.e. for the case of maximal non - commutativity , the distance turns out to be half of the chordal distance @xcite and there does not exist any geodesic in the conventional sense , preventing one to integrate the infinitesimal distance to compute finite distance and the commutative result is obtained only in the @xmath16 limit . as we shall see in the sequel essentially similar results are obtained by applying our formula .
we begin with the computation of the finite distance between coherent states on the moyal plane in the next section .
the purpose of this section is to determine the spectral distance , _ a la _ connes , between an arbitrary pair of finitely separated coherent states @xmath280 , or more precisely between @xmath281 and @xmath282 .
although a formal algorithm ( [ rev_formula ] ) and ( [ n_finite ] ) was devised for this purpose in the preceding section , this is not very user - friendly , as the identification of the right @xmath12 , for which the infimum is reached in ( [ n_finite ] ) is an extremely difficult job . on the other hand , the lower bound ( [ 3.1.9 ] ) can be easily computed as was done in @xcite ( at most up to a numerical constant ) .
the strategy we therefore adopt is to emulate @xcite to obtain the corresponding upper bound and then look for an optimal element @xmath283 for which the saturation condition holds .
if we can identify at least one @xmath283 ( note that this may not be unique ! ) then we can identify the upper bound to be the true distance .
it may also happen in some situations that both upper and lower bounds coincide . in this case ,
their common value can be identified as the distance . otherwise , one has to play with different choices of @xmath12 in ( [ n_finite ] ) to find the best possible estimate , as the upper bound can not be identified as the true distance
. we shall encounter a variety of such situations in the rest of the paper , which will help us to study and contrast various non - commutative spaces through the examples of @xmath4 and @xmath8 .
the present section deals with @xmath4 .
we begin by considering the action of the state @xmath284 , ( associated to the fuzzy point @xmath101 in the moyal plane , in the spirit of gelfand and naimark ) , on a generic algebra element @xmath285 as , @xmath286 note that we have made use of here .
this means that the algebra element @xmath287 gets translated by the adjoint action of @xmath48 thereby furnishing a proper representation of the translation group . without loss of generality
we therefore only have to compute the distance between the pair of states @xmath284 and @xmath288 ( taken to correspond to the origin " @xmath289 ) as can easily be seen by invoking the transformational property of the dirac operator @xmath290 under translation , ( as explained in appendix a ) and can be written , by using , as @xmath291 intuitively , @xmath292 is the maximum change in the expectation values of the @xmath293 and the translated algebra element @xmath294 in the same state @xmath295 .
this is somewhat reminiscent of the transition from the schrdinger to heisenberg picture , where the operators are subjected to the unitary evolution in time through an adjoint action of the unitary operator , while the states are held frozen in time .
+ to proceed further , let us introduce a one - parameter family of density matrices @xmath296 with @xmath297 $ ] taken to be , for convenience , a real affine parameter along the straight line connecting the origin @xmath289 and the point @xmath101 in the complex plane .
we can then introduce @xmath298 consequently , we have the following inequality , @xmath299 since @xmath300 + we can make use of the hadamard identity @xmath301 + \frac{1}{2 ! } [ g , [ g , a ] ] + \frac{1}{3 ! } [ g , [ g , [ g , a ] ] ] + \cdots , \ \label{1.7 } \end{split}\end{aligned}\ ] ] where @xmath302 , to get @xmath303 |0\rangle + t \langle 0|[g , [ g , a]]|0\rangle+ \frac{t^2}{2 ! } \langle 0|[g , [ g , [ g , a]]]|0\rangle + \cdots \\ & = \langle 0|(\exp(tg ) \ ; [ g , a ] \ ; \exp(-tg ) ) |0\rangle.\\ \end{split}\ ] ] on further simplification , this can be recast as @xmath304 ) + z \omega_{zt}([b , a]^{\dagger})| \label{1.8}\ ] ] since only a hermitian element @xmath305 can give the supremum in the connes distance function @xcite .
this yields an upper bound for @xmath306 by making use of the cauchy - schwarz inequality : @xmath307 ) + z \omega_{zt}([b , a]^{\dagger})| \label{1.9}\\ & \leq \sqrt{2}|z| \sqrt{|\omega_{zt}([b , a ] ) ^{2 } + |\omega_{zt}([b , a]^{\dagger } ) ^{2}| } \label{1.10 } \\ & \leq \sqrt{2}|z| \sqrt { \lvert[b , a]\rvert_{op}^{2 } + \lvert[b , a]^{\dagger}\rvert_{op}^{2 } } \label{1.11}\end{aligned}\ ] ] note that in the last step , we have made use of the fact that states @xmath151 s are linear functionals of unit norm .
+ now with the dirac operator @xmath290 , one can prove [ see appendix a.1 ] the following identity @xmath308\|_{op } = \sqrt{\frac{2}{\theta } } \| [ b , a ] \|_{op } = \sqrt{\frac{2}{\theta } } \| [ b^\dagger , a ] \|_{op}.\ ] ] using this , the ball " condition reduces for @xmath293 to @xmath309 \|_{op } = \|[b^{\dagger},a]\|_{op } \leq \sqrt{\frac{\theta}{2}}. \label{1.12}\ ] ] from and , one can therefore write @xmath310 hence from eq ( [ condis ] ) , ( [ 1.6 ] and ( [ 1.13 ] ) we have the following upper bound for the connes distance : @xmath311 clearly , the rhs can be identified as the connes distance , provided there exists an optimal element @xmath312 for which the inequality of is saturated .
we therefore look for an optimal element @xmath313 , satisfying @xmath314 s.t .
@xmath315 a simple inspection of shows that @xmath208 should satisfy @xmath316 = \sqrt{2 \theta}|z| ~~\mathrm{and}~~ [ g , [ g , a_{s } ] ] = 0,\ ] ] where @xmath302 , ensuring that all higher order nested commutators vanish .
observe that since @xmath317 act irreducibly on @xmath318 , we must have , using schur s lemma , @xmath319 $ ] to be proportional to the identity operator , as happens here .
this yields , @xmath320 where @xmath321 .
+ one can check at this stage that although @xmath322 @xmath323\|_{op } = 1,\ ] ] it fails to be a trace - class operator : @xmath324 consequently @xmath325 , but can be thought of as belonging to the multiplier algebra where @xmath326 and @xmath327 .
a resolution of this problem was provided in @xcite .
we briefly describe their approach here , as we propose an alternative approach in the next sub - section . + here
one weakens the strong requirement that @xmath328 and rather looks for a sequence @xmath329 , @xmath330 and also @xmath331 , by inserting a suitable operator - valued `` gaussian '' factor to ensure convergence of the trace - norm and thereby rendering it a trace - class operator : @xmath332 then the following proposition was proved in @xcite ( proposition 3.5 ) : [ [ proposition ] ] proposition : + + + + + + + + + + + + let @xmath333 be a fixed translation and @xmath334 .
define @xmath335 , where @xmath336 . then there exists a @xmath337 s.t .
@xmath338 ( lipschitz ball ) for any @xmath339 .
+ using this proposition , any generic element of the sequence can be written in terms of above @xmath340 as @xmath341 now , it can be easily shown that @xmath342 therefore , @xmath343 it is clear , on the other hand , from the translational symmetry of the dirac operator @xmath290 ( see appendix a ) that it is quite adequate to look for the optimal element @xmath283 at the level of the infinitesimal distance itself .
the anticipated advantage is that @xmath283 when projected to a finite dimensional sub - space will be automatically trace - class .
besides , it will help us to put the computation presented in @xcite in the context of the present analysis and to zoom in on the source of the mismatch between the result of @xcite and @xcite , by a factor of @xmath344 .
we therefore turn our attention towards the computation of the connes distance between infinitesimally separated coherent states @xmath345 and @xmath346 .
clearly , this can just be read off from to yield , by invoking translational symmetry ( see appendix a.1 ) , @xmath347 and the apparent optimal element for the infinitesimal case is easily obtained by projecting @xmath283 ( [ 1.15.1 ] ) into the 2d subspace @xmath348 to get @xmath349 where @xmath350 upto @xmath351 . note that the projection @xmath352 was employed to construct @xmath353 , as @xmath354 lives in the above - mentioned 2d subspace and consequently only the projected component of @xmath283 can contribute to the distance , given in terms of the inner product .
we therefore observe that in this case we have to deal with only finite dimensional subspaces , the optimal element is trace - class , @xmath355 by default and we need not invoke any sequence here .
however , in this case it does not belong to the ball @xmath50 any more : @xmath356 \|_{op } = \sqrt{3 } > 1 \implies a_s \notin b$ ] . in other words , can just be written as @xmath357 the difference with being the _ absence _ of the ball condition : @xmath322 ; all that we have is @xmath358 .
this situation is therefore quite opposite to the finite case where @xmath322 but @xmath359 .
this @xmath353 therefore can not be identified with the _ true _ optimal element .
this @xmath353 can , at best , be identified as an _ effective optimal _ element .
one way to try to remedy this situation would be to replace @xmath353 by @xmath360 as this will now satisfy the ball condition .
this results in the distance expression @xmath361 , thus reproducing the result of @xcite .
one can , however , recognise easily that the occurrence of the @xmath344 factor violating the ball condition is just an artefact of this inappropriate projection procedure .
to see it more transparently , let us project the infinite dimensional matrix @xmath283 to a more general , but finite ( say @xmath362)-dimensional subspace of @xmath31 by @xmath363 , and then compute @xmath364^\dagger [ \mathcal{d } , \pi(p_{n+1}a_sp_{n+1})]$ ] .
we find the matrix to be living on a higher @xmath365-dimensional subspace , having a block - diagonal form : @xmath366^\dagger [ b , p_{n+1}a_sp_{n+1 } ] = \frac{\theta}{2 } \left ( \begin{array}{c|c } \mathds{1}_{(n-1 ) \times ( n-1 ) } & 0_{(n-1 ) \times 3}\\ \hline 0_{3 \times ( n-1 ) } & a\ \end{array } \right),\ ] ] but with @xmath367 being a @xmath368 non - diagonal block matrix : @xmath369 the corresponding operator norm thus turns out to be a linearly divergent @xmath269-dependent function @xmath370 . for @xmath371 this is just @xmath344 as mentioned above .
thus , although @xmath354 ( [ a1 ] ) and @xmath353 ( [ 1.16 ] ) are proportional , the in - appropriate projector @xmath352 ( [ 1.16 ] ) generates the undesirable @xmath344 factor , otherwise the lower bound itself would have yielded the desired result ( [ eqa7 ] ) . +
we now present an alternative to this gaussian " sequence approach of ( [ eqa6])-([1.15.2 ] ) @xcite by constructing a sequence of projected @xmath372 in @xmath373 , rather than projected @xmath283 in @xmath31 as in , using a projector which is appropriate for the eigen - spinor basis of the dirac operator .
this will allow us to evade the problem associated with the violation of the ball condition projector @xmath374 for @xmath31 only .
+ we begin with the diagonal representation of @xmath375 , i.e. @xmath376 . in particular for @xmath377 , we have @xmath378 where the columns and rows are labelled by @xmath379 of ( [ dir_bas_m ] ) , respectively . note that it has vanishing entries in the remaining rows / columns , indexed by @xmath380 , with @xmath381 .
+ proceeding with the same proposed optimal element , we now project it on the representation space spanned by the eigen - spinors . to begin with , we first project it on the same above @xmath382d subspace spanned by @xmath383 .
@xmath384 on computation this yields @xmath385 consequently , @xmath386 \equiv \begin{pmatrix } 0 & -\frac{1}{\sqrt{2 } } & \frac{1}{\sqrt{2 } } & 0 & 0\\ \frac{1}{\sqrt{2 } } & 0 & 0 & -\frac{1}{2 } & \frac{1}{2}\\ -\frac{1}{\sqrt{2 } } & 0 & 0 & -\frac{1}{2 } & \frac{1}{2}\\ 0 & \frac{1}{2 } & \frac{1}{2 } & 0 & 0\\ 0 & -\frac{1}{2 } & -\frac{1}{2 } & 0 & 0\ \end{pmatrix}\ ] ] and finally , in contrast to , we have @xmath386^\dagger [ \mathcal{d } , \mathds{p}_2 \pi(a_s ) \mathds{p}_2 ] \equiv \left ( \begin{tabular}{c|c } $ \mathds{1}_{3 \times 3}$ & $ 0_{3 \times 2}$\\ \hline $ 0_{2
\times 3}$ & $ b_{2 \times 2}$\ \end{tabular } \right),\ ] ] where @xmath387 is a @xmath388 square matrix having eigenvalues @xmath389 and
this is in contrast to the matrix @xmath367 .
further , @xmath391 refers to a rectangular null matrix with @xmath392 rows and @xmath0 columns .
thus clearly we have in this case , @xmath393\|_{op } = 1\ ] ] and , @xmath394 where @xmath395 denotes the inner product between a pair of elements @xmath373 given by @xmath396 and is the counter part of .
again , the subscript @xmath120 indicates that the trace has to computed over @xmath120 .
further note that a factor of @xmath397 has been inserted in ( [ innpro_moy ] ) in anticipation to relate the inner products and , in case both @xmath398 and @xmath399 in are representations of @xmath400 such that @xmath401 and @xmath402 . in that case
, they can be related as , @xmath403 .
of course here one can easily see that @xmath404 any @xmath285 s.t .
@xmath405 and one can not simply relate with any inner products @xmath406 of @xmath31 . indeed , if it were to exist , we could have identified this ` @xmath174 ' , using ( [ mix_ball ] ) and ( [ innpro_moy ] ) , to be the optimal element itself , which by definition has to belong to @xmath407 , or at best to the multiplier algebra .
in fact , this will be a persistent feature with any finite @xmath408-dimensional projection @xmath409 with @xmath410 one can note at this stage , however , that one can keep on increasing the rank of the projection operator @xmath411 indefinitely without affecting ( [ mix_ball ] ) and ( [ innpro_moy ] ) in the sense that the counter part of these equations still has the same form @xmath412 \|_{op } = 1\ ] ] and @xmath413 are independent of @xmath269 if @xmath414 .
+ these equations again follow from the fact that @xmath415^\dagger [ \mathcal{d } , \mathds{p}_n \pi(a_s ) \mathds{p}_n ] = \left ( \begin{array}{c|c } \mathds{1}_{(2n-1 ) \times ( 2n-1 ) } & o_{(2n-1 ) \times 2}\\ \hline o_{2 \times ( 2n-1 ) } & b\ \end{array } \right)\ ] ] with @xmath50 again appearing as the lower block and @xmath416 has `` support '' only on the first @xmath417 block .
+ finally , since in the limit @xmath418 , @xmath419 by , we have @xmath420 .
one can thus interpret @xmath421 \|_{op } = 1\ ] ] and @xmath422 thus , instead of inserting a gaussian factor , as in , we have a sequence @xmath423 of trace - class operators living in @xmath373 ( note that @xmath373 can be regarded as hilbert - schmidt operators acting on @xmath120 ) and each of them satisfy the ball condition .
this is accomplished by projecting in the finite dimensional subspaces spanned by dirac eigen - spinors rather than projecting just to @xmath31 by @xmath374 in , where the ball condition gets violated and the operator norm diverges linearly .
this latter projector could be associated naturally to a different orthonormal and complete basis @xmath424 for @xmath120 as @xmath425 and @xmath426 has the block - diagonal form @xmath427 .
note that the eigen - spinor basis is easily obtained from by first leaving out @xmath428 separately and then pairing @xmath429 and @xmath430 as @xmath431 the projector @xmath411 ( [ p - n ] ) is then clearly the natural choice for the ball condition due to its natural association with the dirac operator .
furthermore , note that we have to make use of the entire @xmath283 ( [ 1.15.1 ] ) as the optimal element .
we would also like to point out in this context that @xmath416 ( [ pi_d_rho ] ) and @xmath432 ( [ proj - a_s ] ) are not proportional anymore , unlike their counterparts ( [ 1.16 ] ) .
however , here too we can easily split @xmath432 or for that matter @xmath372 itself in the limit @xmath433 , into the longitudinal and transverse components , but now in @xmath434 and not in @xmath31 .
this requires a slight generalization of the analysis presented in section [ sec5 ] .
this , however , is not very useful in this context and we do not pursue it here anymore .
+ it is finally clear from the above analysis that the upper bound is saturated in the infinitesimal case through the sequence @xmath423 in the limit @xmath433 , allowing one to identify @xmath435 invoking translational symmetry with @xmath48 as in and the transformational property of the dirac operator , it is clear that @xmath436 \right\rbrace\ ] ] and one concludes that for finitely separated states , one can write @xmath437 reproducing the result and identify the straight line joining @xmath289 to @xmath101 to be geodesic of the moyal plane enabling one to integrate the infinitesimal distance ( [ inf - dis ] ) along this geodesic to compute finite distance .
in fact , one can easily see at this stage that the distance can be written as the sum of distances @xmath438 and @xmath438 as , @xmath439 where @xmath440 is an arbitrary intermediate pure state from the one - parameter family of pure states , introduced in , so that the respective triangle inequality becomes an equality . as we shall subsequently see
this feature will not persist for other generic non - commutative spaces and we will demonstrate this through the example of the fuzzy sphere later . before that we , however , complete our study of the moyal plane by computing the distance between the discrete harmonic oscillator " states in the next section .
for the discrete case , we consider a pair of states , which are separated by an infinitesimal " distance . by this
we mean the nearest states , which are eigenstates of @xmath443 . to compute the distance between the states @xmath444 and @xmath445 we take a similar approach i.e. start with @xmath446 and re - express this as the difference in the expectation value of the transformed algebra element and that of itself in the same state @xmath441 as @xmath447 + ab)b^\dagger - ( n+1)a| n \rangle|.\\\end{aligned}\ ] ] on simplification this yields @xmath448| n+1 \rangle|\\ & = \sup_{a \in b } \frac{1}{\sqrt{n+1 } } |\langle n+1 |[b^{\dagger},a]| n \rangle| .
\end{split}\ ] ] we can now invoke bessel s inequality @xmath449 ( written in terms of the matrix elements @xmath450 of an operator @xmath451 in some orthonormal bases ) , to write ( using ) @xmath452| n+1 \rangle|\\ & = \sup_{a \in b } \frac{1}{\sqrt{n+1 } } |\langle n+1 | [ b^{\dagger},a]| n \rangle|\\ & \leq \frac{1}{\sqrt{n+1 } } \|[b , a]\|_{op } = \frac{1}{\sqrt{n+1 } } \|[b^\dagger , a]\|_{op}.\ \end{split}\ ] ] this finally yields @xmath453 for the finite case , to compute the distance between @xmath445 and @xmath465 with the difference between the two integer labels @xmath392 and @xmath0 being @xmath466 .
we start by writing , @xmath467 as shown in the infinitesimal case eq , @xmath468 | n+i \rangle.\ ] ] therefore , proceeding as in the infinitesimal case , @xmath469 | n+i \rangle \big| \label{2.2.1}\\ & \leq \sup_{a \in b } \sum_{i=1}^{k } \frac{1}{\sqrt{n+i } } \big|\langle n + ( i-1 ) | [ b , a]| n+i \rangle\big| \label{2.2.2}\\ & \leq \sqrt{\frac{\theta}{2}}\sum_{i=1}^{k } \frac{1}{\sqrt{n+i } } , \label{2.2.3 } \end{aligned}\ ] ] by using eq and . + to find an optimal element @xmath312 for which the above inequality is saturated , we demand @xmath470| n+i \rangle\big| = \sqrt{\frac{\theta}{2 } } \sum_{i=1}^{k } \frac{1}{\sqrt{n+i}}.\ ] ] equivalently , @xmath471 now if we let @xmath472 it implies @xmath473 .
constructing such an @xmath312 we get , @xmath474 where @xmath475 .
this gives @xmath476 from the above equation it is also seen that for the harmonic oscillator " basis @xmath477 reproducing the result of @xcite . + finally note that @xmath208 is no longer proportional to @xmath478 unlike ( [ 2.1.4.1 ] ) .
consequently , the distance computed here will exceed the lower bound and @xmath12 contributes non - trivially in ( [ n_finite ] ) .
we approach the problem of the fuzzy sphere in quite the same way as the moyal plane .
there are , however , some fundamental differences between these cases and we will comment on these as we proceed . to begin with , we shall first try to find the distance in the discrete basis and later we will look at the continuous coherent state basis . as will be seen in the subsequent discussion , it is convenient to adopt different techniques for these two cases , namely we need to use the dirac operator eigen - spinors in the latter case .
we begin with a particular fuzzy sphere , corresponding to a particular @xmath0 .
the discrete set of basis are indexed by @xmath144 as @xmath479 or just @xmath480 in an abbreviated from , where the index ` @xmath0 ' is suppressed .
we shall rather use a subscript ` @xmath0 ' to denote the distance function @xmath481 between a pair of states @xmath151 and @xmath482 .
we first compute the distance between the states @xmath483 and @xmath484 ( this being the `` infinitesimal separation '' as far as discrete basis is concerned ) .
similar to the moyal case , we start with @xmath485 then , @xmath486| n_{3}+1 \rangle|.\\\end{aligned}\ ] ] again invoking bessel s inequality , we can write @xmath487\|_{op}}{\sqrt{n(n+1)-n_{3}(n_{3}+1 ) } } = \frac{\|[j_{+},a]\|_{op}}{\sqrt{n(n+1)-n_{3}(n_{3}+1 ) } } \leq \frac{r_n}{\sqrt{n(n+1)-n_{3}(n_{3}+1 ) } } , \label{4.46}\ ] ] where we have made use of the inequality , as proved in appendix [ ap_fuz ] , where for @xmath293 , we have @xmath488 \|_{op } = \|[j_{+},a ] \|_{op } \leq r_n ; \ \
r_n = \lambda \sqrt{n(n+1)}. \label{4.47}\ ] ] like in the case of the moyal plane , we look for an optimal element i.e. an algebra element saturating the upper bound in , so that it can be identified as the distance .
however , as mentioned before , the optimal element may not be unique . here ,
we provide two such optimal elements which saturate the above inequality .
we try with the form corresponding to the lower bound ( [ 3.1.9 ] ) : @xmath489||_{op } } ~~\mathrm{where}~~ d\rho = |n_3 + 1\rangle\langle n_3 + 1| - |n_3\rangle\langle n_3|.\ ] ] the operator norm @xmath456 \|_{op}$ ] can now be computed using the dirac operator eigen - spinor basis as @xmath458=\frac{1}{r_n}\left ( \begin{tabular}{c|c } 0 & $ a$\\ \hline $ -a^\dagger$ & 0\ \end{tabular } \right),\ ] ] where @xmath490 with the rows and columns labeled from top to bottom and left to right respectively by @xmath491 and @xmath492 . from this
we get the operator norm as @xmath461 \|_{op } = \frac{2}{r_n}\sqrt{n(n+1)-n_3(n_3 + 1)}.\ ] ] again as tr@xmath462 , we can readily compute this infinitesimal distance to get , @xmath493 \|_{op}}=\frac{r_n}{\sqrt{n(n+1)-n_3(n_3 + 1)}}. \label{2.1.5.1}\ ] ] there is yet another optimal element given by @xmath494 for both these elements , the `` infinitesimal distance '' is the upper bound itself .
furthermore it also coincides with the lower bound .
thus , @xmath495 this reproduces the result of @xcite . for the finite case
, we try to compute the distance between the states @xmath496 and @xmath497 with @xmath498 and @xmath499 . again , adopting the same technique as in the moyal plane , we write @xmath500| n_{3}+i \rangle|}{\sqrt{n(n+1)- ( n_{3 } + i)(n_{3 } + i-1)}}\\ & \leq \sum_{i=1}^{k } \frac{r_n}{\sqrt{n(n+1 ) - ( n_{3}+ i)(n_{3}+
i -1 ) } } .\end{aligned}\ ] ] it can be easily verified that the element @xmath312 , for which the above inequality is saturated is given by @xmath501 which gives @xmath502 from the above equation it is evident that for the discrete basis , @xmath503 i.e. the distance between finitely separated states are additive and they saturate the triangle inequality for the fuzzy sphere , much like in the case of the moyal plane . in particular the distance between the north pole ( n ) , denoted by @xmath504 and south pole ( s ) , denoted by @xmath505 is given by @xmath506 since both the discrete states @xmath507 and @xmath508 , representing north and south poles are also coherent states , we should be able to reproduce the same results in our coherent state calculations , to be taken up in the next subsection . furthermore , since the @xmath2 case corresponds to maximal non - commutativity let us just write down the distance between the north ( n ) and south poles ( s ) for the lowest three cases @xmath2 , @xmath3 and @xmath509 respectively - as special cases .
they are given by @xmath510 the coefficients in front of @xmath511 , @xmath512 and @xmath513 here indicates that they are highly deformed spaces and the corresponding distances are way below the corresponding commutative spheres @xmath514 , @xmath515 and @xmath516 respectively .
this is just indicative of the fact that none of the fuzzy spheres with fixed @xmath0 allows one to define a geodesic in the conventional sense .
we have more to say on this in the following subsections .
however , one can expect the results to match with the commutative ones in the large @xmath0 limit so that the distance as given in , in the limit @xmath16 , should match asymptotically with @xmath517 , which is the geodesic distance between north and south pole for a commutative sphere . + to implement the commutative limit , we take @xmath518 and @xmath519 as the increment such that @xmath520 as @xmath16 . then we can rewrite @xmath521 where we have set @xmath522 to obtain the last integral .
let us proceed along the same line for the moyal plane .
the perelomov coherent states for the fuzzy sphere is constructed by @xcite @xmath523 and @xmath524 is the abbreviation for @xmath525 .
note that we have taken for convenience the azimuthal angle @xmath526 .
this can be done without loss of generality .
now for the state @xmath527 we have , @xmath528 to calculate the upper bound , we write the distance as @xmath529 we again construct the one - parameter family @xmath530 , with @xmath297 $ ] being a real parameter and @xmath531 .
we then have for pair of states @xmath151 and @xmath482 , @xmath532 using the cauchy - schwartz inequality we get @xmath538}{|\alpha}|\right)\big| & = |\omega_{zt}([j_{+},a ] ) - \omega_{zt}([j_{-},a])|\\ & \le \sqrt{2 } \sqrt { |\omega_{zt}([j_{+ } , a ] ) |^{2 } + |\omega_{zt}([j_{- } , a]^ ) |^{2}}\\ & \le \sqrt{2 } \sqrt { ||[j_{+ } , a]||_{op}^{2 } + ||[j_{- } , a]||_{op}^{2}}\\ & \le 2 r_n . \end{split}\ ] ] thus we get from and , @xmath539 therefore the upper bound of the connes distance on the fuzzy sphere is actually the geodesic distance on the commutative sphere @xmath540 the stark difference with the moyal plane is that we can not find any algebra element saturating this inequality , not even through a sequence or through projections .
this limit is actually the distance on a commutative sphere and is reached only for the commutative limit , @xmath16 as we have shown earlier .
this implies that for any finite @xmath0 representation , the distance between two points on the fuzzy sphere is less than the geodesic distance for a commutative sphere ( see , for example ) and the distance does not follow the conventional ` geodesic ' path as we know it .
indeed , we will show that for @xmath2 , the distance actually corresponds to half of the chordal distance between a pair of points on the surface of the sphere or more precisely between the associated pure states and interpolated by a one parameter family of mixed states . only in the limit @xmath110
this slowly deforms to become the great circle path on the surface . for the calculation of distances for coherent states or more precisely the operator - norm occurring in the ball condition
, we now make use of the eigen - spinors of the dirac operator .
we will first sketch the outline of the algorithm .
one can calculate in a straightforward manner ( at least in principle ) the commutator @xmath541 $ ] in the above - mentioned eigen - spinor basis to obtain @xmath542 = \frac{1}{r_n } \left ( \begin{array}{c|c } 0_{(2n+2 ) \times ( 2n+2 ) } & a_{(2n+2 ) \times 2n } \\
\hline -~a^\dagger_{2n \times ( 2n+2)}&0_{(2n ) \times ( 2n)}\ \end{array } \right),\ ] ] where we denote the non - vanishing rectangular matrices by @xmath367 and @xmath543 as @xmath544 | n , n_3^\prime \rangle\rangle_- & \equiv \frac{1}{r_n}a_{(2n+2 ) \times 2n}\\ a^\dagger_{2n \times ( 2n+2)}=(2n+1)_{-}\langle\langle n , n'_3|\pi(a)| n , n_3 \rangle\rangle_+~~\rightarrow ~~_-\langle\langle n , n_3^\prime | [ \mathcal{d},\pi(a ) ] | n , n_3 \rangle\rangle_+ & \equiv -\frac{1}{r_n } a^\dagger_{2n \times ( 2n+2 ) } \end{split}\ ] ] with the respective ranges for @xmath144 and @xmath145 given by : @xmath545 and @xmath546 . the occurrences of rectangular null matrices in the diagonal blocks of stems from the degeneracy of the spectrum of the dirac operator ( see comments below ) .
this yields a block diagonal form for @xmath547^\dagger [ \mathcal{d},\pi(a ) ] = \frac{1}{r_n^2 } \left ( \begin{tabular}{c|c } $ ( a a^\dagger)_{(2n+2 ) \times ( 2n+2)}$ & $ 0_{(2n+2 ) \times 2n}$\\ \hline $ 0_{2n \times ( 2n+2)}$ & $ ( a^\dagger a)_{2n \times 2n}$\ \end{tabular } \right)\ ] ] and , from the c@xmath548-algebra , property we have @xmath549\|_\text{op}^2=\| [ \mathcal{d},\pi(a)]^\dagger [ \mathcal{d},\pi(a ) ] \|_\text{op } = \frac{1}{r_n^2 } \| aa^\dagger \|_\text{op } = \frac{1}{r_n^2 } \| a^\dagger a \|_\text{op}.\ ] ] in what follows we shall make use use of to first obtain the lower bound of the infinitesimal distance for a general @xmath0 representation .
this will help us improve the estimate by introducing a suitable transverse part @xmath12 in .
we consider the @xmath2 and @xmath17 representations separately in the subsequent subsections and try to calculate the finite connes distance between the ` north pole ' and some other points represented by appropriate coherent states .
we will see that although the revised formula indicates that @xmath12 might play a role in the distance calculation for @xmath2 the lower bound will itself correspond to the exact distance , leaving no room for @xmath12 to contribute .
however , for @xmath17 it will indeed play an important role in improving our estimation of the lower bound , except for @xmath550 , where again @xmath12 will make no contribution . following the algorithm as explained in the previous subsection [ subsec_dirop_fuz ]
, we easily find that @xmath354 is in this case given by @xmath551 using the infinitesimal version of , this yields @xmath552.\ ] ] making use of , we get the matrix form in the dirac eigenbasis for @xmath553 $ ] as @xmath554 \equiv \frac{1}{r_n } \left ( \begin{tabular}{c|c } 0 & $ a$\\ \hline $ -a^\dagger$ & 0\ \end{tabular } \right),\ ] ] where @xmath555 here rows / columns are labeled from up to down / left to right respectively by @xmath556 and @xmath557 .
subsequently , we get @xmath554^\dagger [ \mathcal{d } , \pi(d\rho ) ] = \frac{1}{r_n^2 } \left ( \begin{tabular}{c|c } $ a a^\dagger$ & 0\\ \hline 0 & $ a^\dagger a$\ \end{tabular } \right)\ ] ] with @xmath558 since @xmath456^\dagger [ \mathcal{d } , \pi(d\rho)]\|_\text{op } = \frac{1}{r_n^2 } \|a^\dagger a\|_\text{op}$ ] , as follows from , we have @xmath461\|_\text{op } = \frac{1}{r_n } \sqrt{\|a^\dagger a\|_\text{op } } = \frac{d\theta}{r_n } \sqrt{n ( 3n + \frac{1}{2})}.\ ] ] using this , the lower bound comes out to be @xmath559 \|_\text{op } } = r_n d\theta \sqrt{\frac{2n}{6n + 1}}.\ ] ] clearly , for @xmath2 this comes out to be @xmath560 .
we show later that it coincides with the correct infinitesimal version of the finite distance for the @xmath2 case , to be computed in the next subsection . clearly here
the algebra element can be taken to be a traceless hermitian @xmath388 matrix without loss of generality .
the traceless condition stems from the fact that the @xmath388 identity matrix @xmath561 commutes with dirac operator and therefore does not contribute to the ball condition .
one can thus parametrise any generic algebra element ` @xmath174 ' in terms of a @xmath562-vector @xmath563 , by writing @xmath564 where @xmath565 stands for the three pauli matrices .
+ this enables us to express all the matrix elements of @xmath566 $ ] in terms of the coefficients @xmath567 . for any generic pair of states @xmath568 and @xmath569 we have , by making use of the general framework sketched in section , the matrix element , @xmath570 | m^\prime \rangle\rangle_- = \frac{2}{r_{1/2 } } ~_+\langle\langle m | \pi(a ) | m^\prime \rangle\rangle_-,\ ] ] where @xmath392 takes values @xmath571 and @xmath572 . on explicit computation ,
one finds the following matrix elements for the commutator : @xmath573 so that the final form of the commutator matrix ( [ 6.3.2 ] ) in this case can be written as @xmath574 & \equiv \frac{1}{r_{1/2 } } \left ( \begin{tabular}{c|c } $ 0_{3 \times 3}$ & $ \mathrm{a}_{3 \times 1}$\\ \hline $ - \mathrm{a}^\dagger_{1 \times 3}$ & 0\ \end{tabular } \right ) \end{split } , ~\text{where}~\mathrm{a}_{3 \times 1 } = 2~ _ + \langle\langle m | [ \mathcal{d } , \pi(a ) ] | m^\prime \rangle\rangle_-= \begin{pmatrix } \sqrt{2 } ( a_1 + i a_2)\\ 2a_3\\ - \sqrt{2 } ( a_1 - i a_2 ) \end{pmatrix}.\ ] ] finally , @xmath575^\dagger [ \mathcal{d } , \pi(a ) ] = \frac{1}{r^2_{1/2 } } \left ( \begin{tabular}{c|c } $ \mathrm{a}\mathrm{a}^\dagger$ & 0 \\ \hline 0 & $ \mathrm{a}^\dagger
\mathrm{a}$\ \end{tabular } \right).\ ] ] since @xmath576^\dagger [ \mathcal{d } , \pi(a ) ] \|_\text{op}$ ] and @xmath577 is just a number , we have @xmath578^\dagger [ \mathcal{d } , \pi(a ) ] \|_\text{op } = \frac{1}{r^2_{1/2 } } \|\mathrm{a}^\dagger \mathrm{a}\|_\text{op } = \frac{4}{r^2_{1/2 } } \left ( a_1 ^ 2 + a_2 ^ 2 + a_3 ^ 2 \right ) = \frac{4}{r^2_{1/2 } } |\vec{a}|^2.\ ] ] thus , the ball condition reduces to @xmath578 \|_\text{op } = \frac{2}{r_{1/2 } } |\vec{a}| \leq 1.\ ] ] equivalently , and interestingly , this yields a solid ball in @xmath579 : @xmath580 , all pure states are coherent states such that we can parametrise a pair of generalized points ( pure states ) as 1 .
@xmath581 2 .
@xmath582 , @xmath583 , where in general @xmath584 .
we define @xmath585 .
being a traceless hermitian matrix , this too can be expanded like ( [ al_element ] ) : @xmath586 with this , we have @xmath587 since both @xmath588 and @xmath589 , the supremum of @xmath590 will be attained when @xmath588 and @xmath591 are parallel or anti - parallel to each other .
thus , with the ball condition ( [ n2 ] ) and , we get the spectral distance between a pair of pure states @xmath592 and @xmath593 as @xmath594 this is just half of the chordal distance connecting @xmath595 , reproducing the result of @xcite .
since the supremum is attained here where @xmath588 and @xmath596 are parallel , the distance function just corresponds to the lower bound ( [ 3.1.9 ] ) .
the corresponding infinitesimal distance @xmath597 can easily be seen to match exactly with ( [ low_bound_fuz ] ) , by setting @xmath598 .
coherent states for @xmath598.,width=264 ] let us now consider a family of mixed states ( as shown in the figure [ fig:1 ] )
. a generic mixed state @xmath599 , represented by the point @xmath600 on the chord @xmath595 inside the sphere , is obtained from the following convex sum : @xmath601 clearly , @xmath602 and @xmath603 . introducing @xmath604 and @xmath605 such that @xmath606 we get the spectral distances between the mixed state @xmath607 and @xmath608 and @xmath609 respectively as @xmath610 the fact that the distance of this mixed state @xmath607 from the extremal , pure states @xmath608 and @xmath611 are proportional to the parameters @xmath612 and
@xmath613 respectively indicates that we can identify a unique pure state @xmath614 ( see fig 1 ) nearest to a given mixed state @xmath600 just by extending the straight line @xmath615 from the center @xmath616 to the surface of the sphere .
this distance can therefore be used alternatively to characterize the ` mixedness ' of a spin-@xmath397 system .
further , we have @xmath617 this is just analogous to , except that the intermediate state @xmath607 is not pure .
this chord therefore be identified as a conventional geodesic .
this same family can be parametrised alternatively @xcite as @xmath618 where @xmath619 is the magnitude the vector @xmath620 parametrising each of the mixed states between the two extremal pure states and is given by @xmath621 clearly , @xmath619 is strictly less than @xmath389 : @xmath622 except for the extremal pure states at @xmath261 and @xmath623 .
further the mixed state @xmath624 for the open interval @xmath625 represents a point @xmath600 in the chord connecting the north pole @xmath626 and point @xmath627 , and therefore lies in the interior of the sphere .
indeed , these two different parameters @xmath612 and @xmath229 for the same state can be related by setting @xmath628 , to get @xmath629 thus , we can recast the spectral distance between a mixed state represented by @xmath624 and the pure states @xmath630 and @xmath631 ( [ dis - in - t ] ) respectively as @xmath632 for the case @xmath633 , we get @xmath634 and @xmath635 . finally , we can also obtain the distance between pure states represented by @xmath636 ( south pole s ) and @xmath593 ( p ) as @xmath637 this implies that @xmath638 ^ 2+[d(\rho_{\pi},\rho_{\theta_0})]^2=r^2_{1/2}. \label{pytha}\ ] ] that is , the pythagoras identity @xmath639 is obeyed .
all these features , however , will not persist for higher ` @xmath0 ' , as we shall see . for the spectral triple ,
the space of vector states is @xmath641 @xcite .
we can parametrize a pair of @xmath641-doublets , associated to the pair of points lying in the same latitude ( i.e. the same polar angle @xmath229 ) @xcite as @xmath642 here , the spectral distance between these two points is obtained as @xmath643 which corresponds to the distance measured along the chord connecting the pair of points ( @xmath644 ) and ( @xmath645 ) , at the same latitude @xmath229 .
now , let us define a family of mixed states out of this pair of pure states @xmath161 and @xmath646 in an analogous way as for the @xmath647 representation : @xmath648 similarly , we obtain the distances between the mixed state @xmath649 and the corresponding pure states @xmath161 representing the point ( @xmath644 ) and @xmath646 representing ( @xmath645 ) as @xmath650 clearly , we have @xmath651 the computation for @xmath3 , presented in this section , is expected to be much more complicated than the @xmath598 case , simply because we expect deviations from the straight line chord and the associated pythagoras relation ( [ pytha ] ) to show here , while in the extreme @xmath652 limit this should merge with the great circle of the commutative sphere .
it is obvious that one needs to consider the entire @xmath653 algebra , which in the @xmath654 matrix representation are spanned by 8 traceless gell - mann matrices , as in .
the identity matrix is not considered , since it commutes with the dirac operator and hence will make no contribution to the operator norm @xmath655\rvert_{op}$ ] .
we therefore consider the algebra element @xmath174 to be traceless throughout this section .
it is quite tempting to start directly by identifying the algebra element @xmath174 as a linear combination of the gell - mann matrices .
this algebra element with some extra restrictions provide us with a simple expression of the distance using , which we then corroborate with a more rigorous calculation using .
the role of @xmath12 turns out to be very important in ( [ rev_formula ] ) and ( [ n_finite ] ) for the @xmath3 fuzzy sphere and we employ the most general form of @xmath12 possible to improve the estimate of the spectral distance as best as we can from the lower bound i.e. @xmath656 case in .
the determination of an exact value , even with the help of @xmath657 , remains a daunting task . to begin with the @xmath17 case we first proceed in the same way as in section [ section2 ] to obtain , using , @xmath658|n_3'\rangle\rangle_- = \frac{3}{r_1}~_+\langle\langle n_3|\pi(a)|n_3'\rangle\rangle_-~~\text{and}~~_-\langle\langle n_3'|[\mathcal{d},\pi(a)]|n_3\rangle\rangle_+ = -\frac{3}{r_1}~_-\langle\langle n_3'|\pi(a)|n_3\rangle\rangle_+,\ ] ] where the ranges of @xmath659 are respectively given by @xmath660 and @xmath661 so that with the diagonal representation the commutator @xmath541 $ ] takes the following off - block diagonal form : @xmath662 = \frac{1}{r_1}\left ( \begin{tabular}{c|c } $ 0_{4\times 4}$ & $ a_{4\times 2}$\\ \hline $ -a^{\dagger}_{2\times 4}$ & $ 0_{2\times 2}$\ \end{tabular } \right),\ ] ] where the rectangular matrices @xmath663 and @xmath664 are given by @xmath665 @xmath666 note that here @xmath667 are the usual matrix elements @xmath668 , with @xmath669 . using the above result for @xmath541 $ ]
we readily obtain : @xmath554^\dagger [ \mathcal{d } , \pi(d\rho ) ] = \frac{1}{r_1 ^ 2 } \left ( \begin{tabular}{c|c } $ ( a a^\dagger)_{4\times 4}$ & $ 0_{4\times 2}$\\ \hline $ 0_{2\times 4}$ & $ ( a^\dagger a)_{2\times 2}$\
\end{tabular } \right).\ ] ] by exploiting the properties of the operator norm one has the freedom to choose between the two block - diagonal square matrices as @xmath670^{\dagger}[\mathcal{d},\pi(a)]\rvert_{op } = \lvert[\mathcal{d},\pi(a)]\rvert_{op}^2 $ ] .
we choose to work with the more convenient one i.e. the @xmath671 matrix @xmath672 which turns out to be @xmath673 where @xmath674 this matrix has two eigenvalues @xmath675 : @xmath676 here both @xmath677 and @xmath600 can be written as a sum of several whole square terms and thus they are both positive definite for any algebra elements @xmath174 . clearly , @xmath678 yielding , for a particular @xmath679 , @xmath680 \rvert_{\text{op } } = \frac{1}{r_1 } \sqrt{e_+}.\ ] ] the corresponding infimum @xmath681 \rvert_{\text{op}}$ ] is computed by varying the entries in the algebra elements , within the admissible ranges and obtaining the global minimum of @xmath682 .
this gives @xmath683 \rvert_{\text{op}}=\frac{1}{r_1}\text{min}\big(\sqrt{e_+}\big)= \frac{1}{r_1}\sqrt{\text{min}(e_+)}.\ ] ] the eigenvalue @xmath682 will always have a concave - up " structure in the parametric space as it can be written as the sum of square terms only ( [ mat1 ] ) and ( [ eig_gen ] ) .
there can be points in the parametric space where @xmath682 and @xmath684 are equal , namely points where @xmath600 becomes @xmath390 , but since @xmath682 can not become less than @xmath684 , determining the minimum of @xmath682 will alone suffice in calculating the infimum of the operator norm as is clear from .
so we work with @xmath682 alone and use _ mathematica _ to get the desired result .
+ the pure states corresponding to points on the fuzzy sphere can be obtained by the action of the su(2 ) group element @xmath685 on the pure state @xmath686 corresponding to the north pole ( @xmath269 ) of @xmath687 ( with @xmath17 ) i.e. @xmath688 . note that we have taken for convenience the azimuthal angle @xmath689 .
this can be done without loss of generality .
+ correspondingly , @xmath690 like the @xmath2 case , all entries are real here ; indeed by writing @xmath691 ( @xmath692 s are the gell - mann matrices ) the coefficients of @xmath693 , @xmath694 and @xmath695 vanishes .
this @xmath14 however only provides us with a lower bound of the distance in connes formula and the actual distance is reached by some optimal algebra element ( @xmath696 ) of the form @xmath697 for which the infimum is reached ( say in ) .
this should be contrasted with the optimal element , for which the supremum is reached in . in any case , let us first try to have an improved estimate of the upper bound of the distance .
this will be followed by the computation involving @xmath12 .
the upper bound for the spectral distance , obtained previously in corresponded to that of a commutative sphere @xmath15 , but that lies much above the realistic distance for any fuzzy sphere @xmath8 associated to the @xmath0-representation of @xmath1 , as discussed in section @xmath698 .
it is therefore quite imperative that we try to have a more realistic estimate of this where this upper bound will be lowered considerably . at this stage
, we can recall the simple example of @xmath699-atom , where the energy gap between the ground state @xmath700 and first excited state @xmath701 is the largest one and the corresponding gaps in the successive energy levels go on decreasing and virtually become continuous for very large @xmath0 ( @xmath702 ) .
one can therefore expect a similar situation here too .
indeed , a preliminary look into the distance between north and south poles already support this in the sense that @xmath703 .
one therefore expects the distance function @xmath704 to be essentially of the same form as that of @xmath705 , except to be scaled up by a @xmath706-factor and a miniscule deformation in the functional form . for large - values of @xmath0 ,
the corresponding ratios @xmath707 , and the functional deformations are expected to be pronounced .
however , the exact determination of this form is extremely difficult and we will have to be content with a somewhat heuristic analysis in this subsection and a more careful analysis , using , in the next subsection . to that end
, we start here with the most general form of an algebra element @xmath174 , as a linear combination of all the @xmath708 gell - mann matrices ( @xmath692 ) and look for an optimal element from @xmath293 giving @xmath709 , with some additional restrictions which are to be discussed later .
we write @xmath710 in analogy with for @xmath2 @xmath8 . again
the rows / columns are labeled from top to bottom / left to right by @xmath711 .
now we calculate tr@xmath712 using the @xmath14 matrix and the above algebra element to get @xmath713 this clearly demonstrates the expected independence of imaginary components viz . @xmath714 and @xmath715 .
we therefore set @xmath716 to begin with .
this simplifies the matrix elements of @xmath717 , using , as @xmath718 like - wise the above expression simplifies as @xmath719\bigg| .
\label{guess}\end{aligned}\ ] ] our aim is to obtain a simple form of the ball " condition and eventually of connes spectral distance so that we might obtain an improved estimate for the spectral distance over the lower bound , obtained by making use of @xmath14 , which is more realistic than .
we shall see shortly that with few more additional restrictions , apart from the previous ones ( like vanishing of @xmath720 and @xmath715 imposed already ) it is possible to simplify the analysis to a great extent , which in turn yields a distance estimate which has the same mathematical structure as that of the exact distance for the @xmath2 case upto an overall factor . to that end , we impose the following new constraints : @xmath721 as a simple observation of suggests that it simplifies even further to the following form : @xmath722\big| = 2\sqrt{2}\sqrt{x_1 ^ 2 + 2x_3 ^ 2}\big|\sin\big(\frac{\theta}{2}\big ) \cos\big(\zeta-\frac{\theta}{2}\big)\big|,\ ] ] where @xmath723 and @xmath724 .
moreover , putting the above constraints in the equation - , we get @xmath725 with this , the eigenvalue @xmath682 and the corresponding ball condition can be obtained as @xmath726 \rvert_{\text{op } } = \frac{1}{r_1}2\sqrt{x_1 ^ 2 + 2x_3 ^ 2}\le 1.\ ] ] using this ball condition in , we get @xmath727 hence , a suggestive form of the spectral distance between a pair of pure states @xmath728 and @xmath729 for the @xmath3 representation can be easily obtained by identifying the optimal value of the last free parameter @xmath730 to be given by @xmath731 .
this yields @xmath732 the corresponding form of the optimal algebra element @xmath283 is obtained after a straightforward computation to get @xmath733 when @xmath734 , the distance is exactly the same between the two pure coherent states @xmath735 and @xmath736 and the above distance gives @xmath737 which exactly matches with the one computed in using the discrete formula .
note that we have made use of all the restrictions @xmath738 and @xmath731 , imposed at various stages .
finally , we would like to mention that this simple form was obtained by imposing the above ad - hoc constraints resulting in @xmath739 .
consequently , one can not a priori expect this to reflect the realistic distance either . at best
, this can be expected to be closer to the realistic one , compared to .
the only merit in is that it has essentially the same structure as that of for the @xmath598 case . nevertheless , as we shall show below , the computation involving @xmath12 , using - matches with to a great degree of accuracy .
note that we are denoting the analytical distances as @xmath740 to distinguish them from other distances to be calculated in the next section . also note that since the analytical distance has the same form as , it corresponds to @xmath706-times the half of the chordal distance .
furthermore , it satisfies the pythagoras equality just like the @xmath741 case .
before we conclude this subsection , we would like to point out that we could have perhaps reversed our derivation by simply requiring the matrix @xmath717 to be diagonal : @xmath742 .
but in that case @xmath743 for a particular choice of algebra element , satisfying the aforementioned conditions viz @xmath744 .
now given the structures of @xmath745 and @xmath746 they will have shapes which are concave upwards , when the hyper - surfaces are plotted against the set of independent parameters occurring in @xmath747 @xmath748 , where @xmath749 represents the subregion in the parameter space , defined by these conditions .
now it may happen that @xmath750 , in which case one of them , say @xmath751 , exceeds the other : @xmath752
. then clearly @xmath753\rvert_{\text{op}}=\frac{1}{r_1}\sqrt{\text{min}_{a\in\mathcal{r}}(m_{11})}. \label{n-1}\ ] ] otherwise , the hyper - surfaces given by @xmath751 and @xmath754 will definitely intersect and will reduce to @xmath755\rvert_{\text{op}}=\frac{1}{r_1}\sqrt{\text{min}_{a\in\bar{\mathcal{r}}}(m_{11})}=\frac{1}{r_1}\sqrt{\text{min}_{a\in\bar{\mathcal{r}}}(m_{22 } ) } , \label{n-2}\ ] ] where @xmath756 represents the subregion where @xmath757 .
in fact , this is a scenario which is more likely in this context , as suggested by our analysis of infinitesimal distance presented in the next subsection ( see also fig .
2 ) . we therefore also set @xmath758 . with the additional condition like @xmath759
, we can easily see that one gets , apart from , another set of solutions like , @xmath760 this , however , yields the following ball condition @xmath761 the counterpart of , and in contrast to , can not in anyway be related to its counterpart here , given by @xmath762 we therefore reject from our consideration , as it will not serve our purpose . in this section
we employ the modified distance formula by constructing the most general form of @xmath12 for both finite as well as infinitesimal distances .
we show that in both of the cases the distance calculated using this more general ( numerical ) method matches with the corresponding result given by @xmath763 to a very high degree of accuracy suggesting that @xmath763 should be the almost correct distance for arbitrary @xmath229 .
[ [ infinitesimal - distance-1 ] ] infinitesimal distance + + + + + + + + + + + + + + + + + + + + + + in this case @xmath354 takes a simpler form by expanding ( [ mat2 ] ) and keeping only the leading order terms in @xmath764 : @xmath765 the most general structure of the transverse part @xmath766 here is obtained by taking all possible linear combinations of the generic states @xmath767 i.e. @xmath768 where @xmath769 because of the hermiticity of @xmath770 .
clearly , the complex parameters @xmath771 are exact analogues of suitable combinations of @xmath772 s in .
the orthogonality condition ( [ new1 ] ) here requires the coefficient @xmath773 to be purely imaginary .
moreover we can demand that the matrix representation of @xmath12 should be traceless as discussed in section [ sec1 ] .
we thus impose @xmath774 . to better understand the significance of each term we write ( [ new2 ] ) in matrix form as follows :
@xmath775 with this our optimal algebra element @xmath696 becomes : @xmath776 where we have absorbed @xmath777 inside the coeffiecients of @xmath770 . with @xmath778 independent parameters , it is extremely difficult to vary all the parameters simultaneously to compute the infimum analytically .
however , as far as infinitesimal distances are concerned , it may be quite adequate to take each parameter to be non - vanishing one at a time .
thus , by keeping one of these diagonal / complex conjugate pairs like @xmath779 and @xmath780 to be non - zero one at a time and computing the eigenvalues of the @xmath671 matrix ( [ mat1 ] ) using ( [ new1 ] ) , ( [ new4 ] ) and ( [ new3 ] ) , it is found that only the real part of @xmath779 contributes non - trivially to the infimum of the operator norm @xmath655\rvert_\text{op}$ ] in the sense that the operator norm of this object with @xmath781 in is a monotonically increasing function of @xmath782 but yields a non - trivial value for the infimum which is less than the one with vanishing @xmath779
i.e. @xmath354 itself . like - wise , both real and imaginary parts of @xmath780 and @xmath783 in
does not contribute to the infimum . and
@xmath746,title="fig:",width=302][fig:2 ] on computation , we get @xmath784\rvert_{\text{op}}=\frac{1}{r_1}\sqrt{\text{max}\{m_{11},m_{22}\ } } , ~~\text{where}~~m_{11 } = \beta_1 ^ 2 + 2\sqrt{2}\beta_1d\theta + \frac{7}{2}d\theta^2 ~~\text{and}~~ m_{22 } = 7\beta_1 ^ 2 + 2\sqrt{2}\beta_1d\theta + \frac{1}{2}d\theta^2;\ ] ] where @xmath785 and the @xmath671 matrix takes a diagonal form @xmath786 thus trivially yielding eigenvalues . from the plot of these
eigenvalues(see fig.2 ) it is clear that the infimum of the operator norm over the full range of @xmath787 is given by the minimum value of the two intersections at @xmath367 ( @xmath788 ) and @xmath50 ( @xmath789 ) which comes out to be @xmath790 i.e. @xmath655\rvert_\text{op}=\frac{\sqrt{2 } d\theta}{r_1}$ ] .
note in this context that @xmath791 also , since @xmath792 , as follows from , the infinitesimal spectral distance is given by @xmath793 we now corroborate the same result by varying all the @xmath778 parameters simultaneously . of course we shall have to employ _ mathematica _ now . to that end ,
first note that @xmath794 ( [ mat1 ] ) and ( [ eig_gen ] ) are now given as : @xmath795 where @xmath796 and @xmath797 .
interestingly enough we get the same infimum i.e. @xmath798 by finding the minimum of the eigenvalue @xmath682 as discussed in section [ sec5 ] with the @xmath778 parameter eigenvalue where p and q are given by ( [ p1 ] ) and ( [ q1 ] ) .
we therefore recover the distance which also matches with for @xmath799 .
[ [ finite - distance-1 ] ] finite distance + + + + + + + + + + + + + + + for any finite angle @xmath229 , the @xmath14 matrix ( [ mat2 ] ) can be directly used to compute the square of the trace norm @xmath800 .
moreover , this @xmath14 can be used as the algebra element @xmath174 to compute the eigenvalues of and then the operator norm @xmath655\rvert_{op}$ ] , yielding the lower bound for the spectral distance using .
more specifically for @xmath801 ( i.e. the distance between the north pole @xmath269 and any point @xmath802 on the equator ) the lower bound is found to be @xmath803 contrasting with the corresponding value of @xmath804 , we see that the transverse component @xmath805 must play a role here . on the other hand for @xmath550 ( i.e. the distance between north and south pole )
we find @xmath806 which matches exactly with the result of for the distance between discrete states @xmath735 and @xmath736 .
this means that there is no contribution of @xmath12 to this distance for @xmath550 .
there are , however , non - trivial contributions from @xmath807 to the distance for any general value of the angle @xmath808 as we have illustrated above through the example of @xmath801 . now the most general @xmath807 in this case can be constructed as follows : @xmath809 , \end{split}\ ] ] where @xmath810 .
again we can take @xmath696 to be traceless as before , which implies @xmath811 and we can eliminate one of them , say @xmath812 .
furthermore , imposing the orthogonality condition we have a relation between all the remaining 8 parameters and one of them can be eliminated from that . in this case , using the @xmath14 and @xmath12 in while absorbing @xmath777 inside the coefficients of @xmath12 , we have @xmath813 where @xmath814 and @xmath815 are the real components and @xmath816 and @xmath817 are the imaginary components of @xmath818 and @xmath780 respectively . with these substitutions the eigenvalues of the matrix become a function of @xmath778 parameters .
we now calculate p , q for the @xmath801 case to get @xmath819.\end{aligned}\ ] ] minimizing @xmath682 as before , we get @xmath820 which , remarkably , exactly matches the result of . as for @xmath550 ,
the terms p , q of the eigenvalues @xmath675 come out to be @xmath821.\end{aligned}\ ] ] again we need to calculate the spectral distance by minimizing @xmath682 which gives @xmath822 .
this is precisely the lower bound result @xmath823 of as discussed previously and hence support our claim that @xmath12 will not contribute here at all . for arbitrary angle @xmath229 , we present in table 1 both the distances i.e. the one calculated using the formula and the other calculated using the global minima of the eigenvalue @xmath682 ( say @xmath824 ) for various angles between @xmath390 and @xmath123 . [ cols="^,^,^",options="header " , ]
it is very striking that the distance @xmath825 matches almost exactly with @xmath824 for all these angles as one sees from table 1 .
this strongly suggests that is indeed very very close to the exact distance !
in fact , for larger angles like @xmath826 and above the results agree upto @xmath827 decimal places , whereas for smaller angles ( @xmath828 ) they agree upto @xmath382 decimal places and show some miniscule deviations from @xmath829 decimals onwards .
one can expect to see more pronounced deformations away from in the functional form when the overall scale of magnification starts reducing monotonically with @xmath110 and eventually merge with the commutative results .
we have provided here a general algorithm to compute the finite spectral distance on non - commutative spaces , in our hilbert - schmidt operator formulation , and we find here that the formula quoted in @xcite actually corresponds to the lower bound .
however , as far as the computation of infinitesimal distances is concerned , this formula is quite adequate .
this is because it can reproduce the local infinitesimal distance up to an overall numerical factor as shown in @xcite .
we should , however , keep in mind that the knowledge of infinitesimal distances may not be adequate to capture the geometry of a generic non - commutative space as such spaces , unlike a commutative differentiable manifold , may not allow a geodesic to be defined in the conventional sense . by the word conventional " ,
we mean that the geodesic passes through a one parameter family of pure states .
in this situation , one can not simply integrate the infinitesimal distance to compute the finite distance . here
, we have studied extensively the geometry of the moyal plane ( @xmath4 ) and that of the fuzzy sphere ( @xmath8 ) . for this
, we made use of both the above mentioned revised algorithm and also emulated the method of @xcite to compute the upper bound and then look for an optimal element saturating this upper bound . in the case of the moyal plane , we succeed in identifying such optimal element ` @xmath283 ' belonging to the multiplier algebra .
we then constructed a sequence of projection operators @xmath830 in the finite dimensional subspace spanned by eigen - spinors of the dirac operator that converge to @xmath372 and saturates the upper bound , allowing us to identify the upper bound itself with the distance .
eventually , this enables us to relate the one parameter family of pure states to the geodesic of the moyal plane which is nothing but the straight line .
in contrast , on the fuzzy sphere , although an analogous upper bound can be constructed for any finite @xmath0-representation of @xmath79 , there simply does not exist an optimal element @xmath283 saturating the inequality . indeed , for the case of extremal non - commutativity @xmath598 , the finite distance turns out to be half the chordial distance . here ,
except for the extremal points , the interpolating
points " correspond to mixed states .
this in turn helps us to find the distance between a given mixed state and a uniquely defined nearest pure state lying on the surface " of @xmath8 .
the corresponding distance can then be taken as an alternative characterization of the mixedness " of a state .
this exercise shows that in connes framework no discrimination is made between pure and mixed states ; it scans through the entire set of pure and mixed states to compute the supremum in .
all these calculations are enormously simplified by working in the eigen - spinor basis of the respective dirac operator , so much so that we are able to compute the distance in the ` @xmath3 ' fuzzy sphere , using this revised algorithm .
since this algorithm involves also the transverse @xmath12 components in addition to the longitudinal @xmath14 component , this becomes somewhat less user - friendly .
for the ` @xmath3 ' case , for example , it involves a minimization in seven parameters .
needless to say that we have to make use of _ mathematica _ after solving the quadratic characteristic equation . for higher @xmath0
s , the corresponding characteristic equations will not only involve higher degree polynomials , it will also involve a large number of independent parameters to be varied .
consequently , the computation for the @xmath831 fuzzy sphere , even with the help of _ mathematica _ , remains virtually intractable and for the moyal plane the number of parameters is simply infinite ! to put our findings in a nutshell
, we observe that the finite distance for @xmath3 and that of @xmath598 have almost the same functional form except for an overall scaling by a factor of @xmath706 and a miniscule deformation at small ` @xmath229 ' and that too only from the sixth decimal onwards .
we thank prof . a.p .
balachandran for his valuable suggestions .
ap thanks the inspire and kk thanks snbncbs for the financial support to carry out their project works .
both of them thank snbncbs for the hospitality during their stay when parts of this work were completed .
in order to construct a dirac operator for the moyal plane , we need to consider the momentum operator satisfying the non - commutative heisenberg algebra ( [ mp ] , [ ha ] ) .
now since only @xmath31 can furnish a complete representation of the entire heisenberg algebra we need to construct our dirac operator on this space .
thus we consider the operators @xmath833 , which acts adjointly only on @xmath31 .
the dirac operator is then constructed in terms of pauli matrices @xmath834 and @xmath835 as @xmath836 note that in this construction the dirac operator has a natural action on @xmath837 through the adjoint action of @xmath838 , i.e. on a generic element @xmath839 it acts as @xmath840\\ [ -i \hat{b},| \phi_1 ) ] \end{pmatrix } .\ ] ] this , on turn implies that the action of the commutator @xmath541 $ ] on @xmath841 , on using , gives @xmath842\phi = \sqrt{\frac{2}{\theta}}\begin{pmatrix } 0 & [ i \hat{b}^\dagger , a]\\ [ -i \hat{b},a ] \end{pmatrix}\phi.\ ] ] now regarding @xmath841 as a test function , we can identify the dirac operator @xmath290 as @xmath843 this is further simplified by considering the transformation @xmath844 and @xmath845 , which just corresponds to a @xmath846 rotation by an angle @xmath847 in @xmath848 space . with this transformation
the dirac operator takes the following hermitian form : @xmath849 which precisely has the same form as , used throughout the paper . the most important point
to note is that this very structure of the dirac operator allows us to make it also act directly from the left on @xmath850 . in this context
, we would like to mention that this dirac operator @xmath290 responds to a @xmath846 rotation in the @xmath851 plane by an arbitrary angle @xmath852 as @xmath853 since @xmath854=[\hat{b},\hat{b}^\dagger]=1 $ ] , one can build the tower of states , parametrized by @xmath852 , analogous to ( [ hc ] ) and is related to the ones in ( [ hc ] ) by a phase factor : @xmath855 ) , the dirac operator transforms as @xmath856 since the operators @xmath51 and @xmath857 are shifted by a c - numbers only , it has no impact on the commutator @xmath541 $ ] occurring in the ball condition ( [ condis ] ) . finally , note that the dirac operator here is hermitian and differs from that of @xcite by a factor ( @xmath858 ) , which however , is quite inconsequential as it does not affect the operator norm @xmath859 \|_{op}$ ] , that appears in the ball condition .
+ we are now going to prove an identity that will be used in our calculations throughout the paper . for our dirac operator
we have for @xmath121 @xmath575 = \sqrt{\frac{2}{\theta } } \begin{pmatrix } 0 & \left[\hat{b } , \pi(a)\right]\\ \left[\hat{b}^\dagger , \pi(a)\right ] & 0 \ \end{pmatrix},\ ] ] yielding @xmath860^\dagger[\mathcal{d } , \pi(a)]\|_\text{op } = \frac{2}{\theta}~ \text{max } \left ( \| \hat{q}^\dagger \hat{q } \|_\text{op } , \| \hat{q } \hat{q}^\dagger \|_\text{op } \right),\ ] ] where we have introduced the operator @xmath861 as @xmath862 = - [ \hat{b } , \pi(a)]^\dagger ~;~~ \hat{q}^\dagger = [ \hat{b}^\dagger , \pi(a)]^\dagger = - [ \hat{b } , \pi(a)].\ ] ] note that here the algebra element @xmath121 has been taken to be hermitian ( @xmath863 ) since , as shown in @xcite , the optimal element for which the supremum in the connes distance formula is reached belongs to the subset of hermitian elements of the algebra @xmath118 .
+ we now state a theorem in standard functional analysis @xcite : if @xmath864 , we have @xmath865 , where the superscript @xmath866 stands for the involution operator .
+ from this theorem it then immediately follows that @xmath867 . since in our case
the involution operation is given by the hermitian conjugation ( @xmath36 ) , as mentioned earlier .
besides for a legitimate spectral triple @xmath868 $ ] and @xmath869 $ ] both are bounded operators and thus the theorem holds .
+ we combine this result with to obtain , @xmath578\|_\text{op } = \sqrt{\frac{2}{\theta}}~ \| [ \hat{b}^\dagger , \pi(a)]\|_\text{op } = \sqrt{\frac{2}{\theta}}~ \| [ \hat{b } , \pi(a)]\|_\text{op}.\ ] ] this simplifies the problem of finding the operator norm @xmath187 \|_\text{op}$ ] for any hermitian element @xmath870 . then for a hermitian algebra element @xmath872 , @xmath873\|^{2}_\text{op } & = \|[\mathcal{d},\pi(a)]^\dagger[\mathcal{d},\pi(a)]\|_\text{op}\\ & = \frac{1}{r^{2}}\left\|\left[\begin{array}{cc } [ j_{+},a]^{\dagger}[j_{+},a ] + [ j_{3},a]^{\dagger}[j_{3},a ] & [ [ j_{-},a],[j_{3},a]]\\ - [ [ j_{+},a],[j_{3},a ] ] & [ j_{-},a]^{\dagger}[j_{-},a ] + [ j_{3},a]^{\dagger}[j_{3},a ] \end{array}\right ] \right\|\\ & \ge \frac{1}{r^{2}}\sup_{{\stackrel{\psi \in \mathcal{h}}{\|\psi\| = 1 } } } \langle\psi_{1}| [ j_{+},a]^{\dagger}[j_{+},a ] + [ j_{3},a]^{\dagger}[j_{3},a ] |\psi_{1}\rangle \ \ \bigg(\psi = \begin{pmatrix } 0 \end{pmatrix } \bigg)\\ & \ge \frac{1}{r^{2 } } \sup_{{\stackrel{|\psi_{1}\rangle \in \mathcal{h}_{c } } { | \psi_{1}\rangle\langle\psi_{1}|= 1}}}\langle\psi_{1}| [ j_{+},a]^{\dagger}[j_{+},a]|\psi_{1}\rangle+ \frac{1}{r^{2}}\sup_{{\stackrel{|\psi_{1}\rangle \in \mathcal{h}_{c}}{| \psi_{1}\rangle\langle\psi_{1}| = 1 } } } \langle\psi_{1}| [ j_{3},a]^{\dagger}[j_{3},a ] |\psi_{1}\rangle\\ & \ge \|[j_{+},a]\|^{2}_\text{op } + \|[j_{3},a]\|^{2}_\text{op}. \end{split } \label{a.4}\ ] ] therefore we get @xmath874\|_\text{op } \le \|[\mathcal{d},\pi(a)]\|_\text{op } \ \ \ \frac{1}{r } \|[j_{3},a]\|_\text{op }
\le \|[\mathcal{d},\pi(a)]\|_\text{op}. \label{a.5}\ ] ] for @xmath875^{\dagger } = - [ j_{-},a]$ ] and using @xmath876 as shown above , we get @xmath877\|_\text{op } \le \|[\mathcal{d},\pi(a)]\|_\text{op } . \label{a.6}\ ] ] for @xmath878 i.e @xmath879\|_\text{op } \le 1 $ ] it thus follows that @xmath537\|_\text{op } \le r ~~;~~ \|[j_{-},a]\|_\text{op } \le r. \label{a.7}\ ] ] galluccio s , lizzi f and vitale p 2008 _ phys .
d _ * 78 * 085007 ; + balachandran a p and martone m 2009 _ mod .
a _ * 24 * 1721 ; + balachandran a p , ibort a , marmo g and martone m 2010 _ phys .
* 81 * 085017 . | we revise and extend the algorithm provided in @xcite to compute the finite connes distance between normal states .
the original formula in @xcite contains an error and actually only provides a lower bound .
the correct expression , which we provide here , involves the computation of the infimum of an expression which involves the transverse " component of the algebra element in addition to the longitudinal " component of @xcite .
this renders the formula less user - friendly , as the determination of the exact transverse component for which the infimum is reached remains a non - trivial task , but under rather generic conditions it turns out that the connes distance is proportional to the trace norm of the difference in the density matrices , leading to considerable simplification .
in addition , we can determine an upper bound of the distance by emulating and adapting the approach of @xcite in our hilbert - schmidt operatorial formulation .
we then look for an optimal element for which the upper bound is reached .
we are able to find one for the moyal plane through the limit of a sequence obtained by finite dimensional projections of the representative of an element belonging to a multiplier algebra , onto the subspaces of the total hilbert space , occurring in the spectral triple and spanned by the eigen - spinors of the respective dirac operator .
this is in contrast with the fuzzy sphere , where the upper bound , which is given by the geodesic of a commutative sphere is never reached for any finite @xmath0-representation of @xmath1 .
indeed , for the case of maximal non - commutativity ( @xmath2 ) , the finite distance is shown to coincide exactly with the above mentioned lower bound , with the transverse component playing no role .
this , however starts changing from @xmath3 onwards and we try to improve the estimate of the finite distance and provide an almost exact result , using our new and modified algorithm . |
* setup . * let us assume that we know the probability distribution @xmath25 of a random variable @xmath129 .
let @xmath2 denote the mean and @xmath130 the variance of the @xmath25 distribution .
a value of @xmath125 can not be observed directly , but only via experiment results .
we would like to estimate teh value of @xmath125 from an experiment result @xmath43 , the probablity of which @xmath131 depends on @xmath125 .
we look for an estimator @xmath132 with the smallest average error w.r.t . the @xmath25 distribution . for the average error @xmath133 the optimal estimator
is known to be : @xmath134 in such a choice of estimator we have : @xmath135 where @xmath136 is average probability of obtaining the result @xmath43 . +
* bayesian cramer - rao bound*. the bayesian cramer - rao inequality bounds from below the average error @xmath137 of any estimator @xmath138 @xcite . for a gaussian prior distribution @xmath25 with a variance @xmath130 it is as follows : @xmath139 where @xmath140 is the fisher information for the @xmath141 probability defined as @xmath142 , for a quantum setup see also eq .
( [ eq : fisher ] ) .
we are interested in estimating a random phase @xmath143 , where the random phases @xmath63 have a gaussian distribution @xmath87 with a covariance matrix @xmath29 and the same means equal @xmath2 , i.e. @xmath144 , and @xmath145 , @xmath79 .
the distribution @xmath90 of @xmath86 is gaussian with the variance equal @xmath146 and the mean equal @xmath2 .
we can not observe a @xmath86 value directly , but only via an experiment result @xmath43 .
the probability of obtaining a result @xmath43 @xmath147 depends on all values @xmath63 .
the probability of obtaining @xmath43 when @xmath148 is @xmath149 , where @xmath150 .
the best estimator of @xmath86 according to eq .
( [ eq : bayesest ] ) is : @xmath151 where @xmath152 is the best bayesian estimator of a random phase @xmath153 w.r.t .
@xmath87 . from eqs . ( [ eq : bayeserror ] ) and ( [ eq : bcr ] ) we arrive at : @xmath154 where @xmath155 and @xmath156 is the fisher information for the @xmath157 probability , i.e. @xmath158 .
we perform a povm measurement @xmath9 on a state @xmath159 .
the probability of obtaining a result @xmath43 equals @xmath160 .
when @xmath63 are gaussian random variables with the same means equal @xmath2 , we have @xmath161 , where @xmath162 corresponds to the dephased state @xmath39 , see eq .
( [ eq : state ] ) .
+ to use the bayesian cramer - rao bound in eq .
( [ eq : bcr ] ) we need the fisher information @xmath163 for the probability distribution @xmath164 , where @xmath165 is a state obtained by integrating @xmath166 over a set @xmath167 with a conditional probility @xmath168 , where @xmath90 is the @xmath86 probability distribution .
below we prove that @xmath169 $ ] , where @xmath170 .
therefore , for @xmath171 we have @xmath172 . as the quantum fisher information is convex w.r.t .
density matrices , we also have @xmath173 .
thus , we arrive at a quantum version of the bayesian cramer - rao inequality above in eq .
( [ eq : bayesc ] ) : @xmath174 we now prove that @xmath169 $ ] . as @xmath73
, we obtain : @xmath175 therefore : @xmath176 and @xmath177\nonumber\\ & + & g_c(\varphi)^{-1}\sum_{j=1}^n\gamma_j\int_{\mathbb{r}^{n-1 } } \mathrm{d}\varphi_1 ...
\mathrm{d}\varphi_{j-1 } \mathrm{d}\varphi_{j+1} ... \mathrm{d}\varphi_n\ , g_\phi\left(\varphi_1, ... ,\varphi_{j-1},\gamma_j^{-1}\varphi-\gamma_j^{-1}\sum_{k=1,k\neq j}^n\gamma_k\varphi_k,\varphi_{j+1}, ... ,\varphi_n\right)\nonumber\\ & & \qquad\qquad\times\ , \,-\gamma_j^{-1}\left ( \sum_{k=1 , k\neq j}^n ( c^{-1})_{jk}(\varphi_k-\phi)\,+\ , ( c^{-1})_{jj } \left(\gamma_j^{-1}\varphi-\gamma_j^{-1}\sum_{k=1,k\neq
j}^n\gamma_k\varphi_k-\phi\right ) \right ) \nonumber\\ & & \qquad\qquad\times\,\rho_{\varphi_1, ... ,\varphi_{j-1},\gamma_j^{-1}\varphi-\gamma_j^{-1}\sum_{k=1,k\neq j}^n\gamma_k\varphi_k,\varphi_{j+1}, ... ,\varphi_n}\nonumber\\ & -&\rho'_{\varphi } g_c(\varphi)^{-1 } \frac{\mathrm{d}}{\mathrm{d}\varphi } g_c(\varphi ) \nonumber\\ & = & -i\left[h,\rho'_{\varphi}\right]- \rho'_{\varphi } \delta_{c}^{-2 } ( \varphi-\phi ) + \rho'_{\varphi } \delta_{c}^{-2 } ( \varphi-\phi)\,=\ , -i\left[h,\rho'_{\varphi}\right]\quad\blacksquare.\end{aligned}\ ] ]
let us prove that the choice of @xmath178 in eq .
( [ eq : bestest ] ) , where @xmath83 and @xmath179 is the best bayesian estimator of the random phase @xmath153 w.r.t .
the @xmath180 distribution , @xmath79 , is locally unbiased . given that @xmath181 saturates the cramer - rao inequality in eq .
( [ eq : fisher ] ) , the following will prove its optimality .
let @xmath182 denote average w.r.t .
we have @xmath184 from eq .
( [ eq : bayesc ] ) and therefore also @xmath185 .
we also have : @xmath186 where @xmath187 and @xmath188 , @xmath27 .
thus , since @xmath83 is also the best bayesian estimator of @xmath189 w.r.t .
@xmath180 we obtain : @xmath190 which implies @xmath191 , i.e. @xmath181 is locally unbiased . | we derive upper bounds on the quantum fisher information in interferometry with @xmath0 subsystems , e.g. two - level atoms or gaussian modes , in the presence of arbitrarily correlated gaussian dephasing including independent and collective dephasing .
the derived upper bound enables us to analyse the fisher information asymptotic behaviour when @xmath1 .
dephasing introduces random phases to subsystems dynamics , which lowers the precision of estimating the phase difference @xmath2 in an interferometer .
the method presented uses bayesian estimation of the random phases and eliminates dephasing noise by calculating their weighted arithmetic mean , which correponds to the phase @xmath2 estimated in interferometry . *
introduction . * in numerous areas of modern physics , e.g. spectroscopy in atomic clocks @xcite or gravitational interferometers @xcite , it is necessary to estimate an unknown value of a parameter of quantum system dynamics .
when using a quantum system , the precision of parameter estimation is bounded , from below , by the inverse of the fisher information which depends on parameter encoding details , available resources , such as number @xmath0 of atoms / gaussian modes in the system and initial system state preparation , and the measurement performed on the system state with an encoded parameter . when a parameter is encoded via the unitary dynamics of the system
, the estimation error scaling can be improved from the classical shot - noise scaling @xmath3 to the heisenberg scaling @xmath4 by an entangled initial state @xcite .
this quantum enhancement in precision , however , may be significantly limited in the presence of decoherence , i.e. when the system interacts with an uncontrolled enviroment @xcite , @xcite .
the aim of interferometry is to estimate a phase @xmath2 encoded in an evolved system state , see fig .
[ fig : scheme ] . here
we discuss interferometry with @xmath0 subsystems , including two - level atoms or gaussian modes , in the presence of dephasing .
dephasing introduces additional random phases to subsystems dynamics , thus lowering the precision of the estimation of the phase @xmath2 . only independent and collective dephasing
have been successfully considered so far .
independent random phases have been discussed e.g. in @xcite and the derived upper bound on fisher information shows linear scaling with @xmath0 . in @xcite the second case of identical random phases
was considered and an upper bound , which converges to a constant when @xmath1 , was derived .
we , however , provide a new unified approach to interferometry with @xmath0 subsystems , which delivers an upper bound on the fisher information in the presence of arbitrarily correlated dephasing .
this bound depends on both correlationsof the random phase and initial system state preparation .
it is tight for weak decoherence , because the heisenberg scaling of the estimation error is recovered as dephasing disappears .
furthermore , the bound enables us to analyse the asymptotic behaviour of the fisher information , thus obtaining already familiar constant asymptotics for collective dephasing and the linear scaling in the independent case . + * interferometry with dephasing . *
the interferometry setup is sketched out in fig .
[ fig : scheme ] .
the system is first prepared in an initial state @xmath5 @xmath6 .
then it undergoes the dynamics described by a channel @xmath7 leading to an evolved state @xmath8 which has an enconded value of the @xmath2 phase .
finally , a povm measurement @xmath9 ( @xmath10 , @xmath11 , @xmath12 , @xmath13 ) is performed on @xmath14 in order to obtain information about @xmath2 . is encoded in an initial state @xmath5 via a quantum channel @xmath15 . a dephasing channel @xmath16 which commutes with unitary encoding of @xmath2
is considered .
the interferometer is described by a hamiltonian @xmath17 which is the generator of unitary encoding .
a povm measurement @xmath9 is performed on @xmath14 to retrive the information about the value of @xmath2 .
quality of the setup is quantified by the fisher information .
[ fig : scheme],scaledwidth=50.0% ] in the absence of dephasing , @xmath2 is encoded via unitary dynamics by a hamiltonian @xmath17 , i.e. the evolved state is @xmath18 . when interaction with an enviroment leads to dephasing , we have @xmath19 , where @xmath16 represents the dephasing channel , which commutes with unitary dynamics generated by @xmath17 , see eq . ( [ eq : state ] ) .
thus , dephasing can also be viewed as an imperfect preparation of the initial state @xmath5 such that in fact the dephased state @xmath20 is prepared and used in the unitary interferometry setup .
the estimation precision is lowered by dephasing , as all possible initial states are effectivily the mere outputs of the dephasing channel @xmath16 .
let @xmath21 , where @xmath22 is a self - adjoint operator on @xmath23-th subsystem . dephasing introduces random phases into the dynamics of the subsystems : @xmath24 where @xmath25 is the distribution of the random phases @xmath26
, @xmath27 .
one works with the averaged state @xmath28 as it is not possible to access enviroment degrees of freedom in order to choose values of the random phases .
the @xmath28 state is influenced by correlations of the random phases .
usually @xmath25 is assumed to be gaussian when it is fully determined by random phase means and a covariance matrix @xmath29 . without loss of generality , we assume the means to equal 0
. independent identically distributed phases decribe * independent dephasing * when fully correlated @xmath30 , @xmath31 correspond to * collective dephasing*. for a state @xmath5 of @xmath32 two - level atom we have @xmath33 , where @xmath34 is the pauli matrix along the @xmath35-axis . for dephasing given by a gaussian random phase with a variance @xmath36 and a mean 0 ,
we obtain the following dephased state @xmath37 ( in the eigenbasis of @xmath34 ) : @xmath38 + * fisher information . * the quality of the interferometry setup can by quantified using the fisher information . here
, it is enough to discuss the case when @xmath2 is unitarily encoded in an initial state @xmath5 , i.e. @xmath18 , and a povm measurement is performed on @xmath39 . in the presence of dephasing we
simply replace @xmath5 by @xmath37 .
we have to estimate @xmath40 only from a result @xmath41 of the povm measurement @xmath9 performed on a state from the family @xmath42 .
a result @xmath43 is obtained with probability @xmath44 . in order estimate @xmath2
we use an estimator - @xmath45 .
let us consider the case of @xmath46 , where @xmath47 is known and @xmath48 is a small fluctuation that we want to estimate .
we compare estimators by using the local error defined as @xmath49 . for any locally unbiased estimator at @xmath50 ( @xmath51 and @xmath52 ) this error
is bounded from below in the cramer - rao inequality : @xmath53 where @xmath54 is the fisher information .
this fisher information quantifies the quality of the interferometry setup as it bounds from below the phase estimation precision , thus simplifying the optimisation of the setup since we no loger need to refer to an estimator .
the fisher information depends on the choice of measurement @xmath55 . whatever the measurement is @xcite : @xmath56 , \label{eq : qfisher}\ ] ] where @xmath57 is the quantum fisher information and @xmath58 is the symmetric logarythmic derivative .
the @xmath58 eigenbasis corresponds to the optimal projective measurement for which @xmath59 . as the quantum fisher information is the same for all @xmath2 , let us drop the index @xmath2 .
we have @xmath60 and @xmath61 $ ] .
optimisation of the interferometry setup is reduced to choosing the initial state @xmath5 .
+ it is not easy to find the maximum of the quantum fisher information w.r.t .
the initial state in the presence of dephasing , even numerically . in order to discuss the asymptotic behaviour of the quantum fisher information in the presence of general dephasing
, we need to derive a new upper bound .
our method uses knowledge about the dephasing channel @xmath16 , i.e. the random phase probability distribution .
first , we estimate random phase values using the bayesian approach .
then , we eliminate dephasing noise by calculating a weighted arithmetic mean of the random phase estimators , which corresponds to @xmath2 .
this mean can be related to the optimal estimator of @xmath2 which saturates the cramer - rao bound in eq .
( [ eq : fisher ] ) . our approach provides a clear and simple insight into the quantum fisher information behaviour in the presence of general dephasing . + * the bayesian approach .
* let us consider one run of the intererometry experiment .
first , unknown random phase values are chosen according to a gaussian distribution in with all means equal @xmath62 and a covariance matrix @xmath29 . next , these phases are shifted by a common phase @xmath46 , where @xmath47 is known and @xmath48 is an uncontrolled fluctuation to be estimated .
let @xmath63 denote the shifted phase values and @xmath64 their probability distribution with all the means now equal @xmath2 , and the unchanged covariance matrix @xmath29 : @xmath65 .
these values are then encoded in the initial state @xmath66 and we go on to perform the povm measurement @xmath55 . in order to estimate @xmath2 from a result @xmath43
, we first estimate the shifted random phases @xmath67 and then their common mean which equals exactly @xmath2 .
as we can not choose the random phase value , in many experiments we obtain a result @xmath41 with an average probability @xmath68 , where @xmath69 is the dephased state in eq .
( [ eq : state ] ) and @xmath70 .
we would expect the fisher information to appear , since @xmath71 is small .
we know the @xmath64 distribution except for the mean @xmath2 , which we need to estimate .
let us consider the following _
gedankenexperiment_. we assume that we can observe phases @xmath67 directly . in order to estimate @xmath2 we eliminate random dephasing noise by calculating a weighted arithmetic
mean @xmath72 with @xmath73 , thus guaranteeing that @xmath74 is an unbiased estimator of @xmath2 .
using e.g. lagrange multipliers one can show that @xmath75 leads to minimum local estimation error @xmath76 .
let us note that it is sufficient to measure just one phase @xmath77 in order to estimate @xmath2 .
for independent dephasing we have @xmath78 , @xmath79 , and @xmath80 , where @xmath81 is the variance of every random phase .
we can not , however , observe phases directly , but only via a measurement result @xmath43 . in order to estimate the shifted phases we use knowledge about their gaussian distribution @xmath64 and the bayesian estimation .
the bayesian approach provides the estimators @xmath82 , @xmath79 , which have minimum error w.r.t .
@xmath64 ( see appendix [ ap : bayes ] ) . as we do not know the exact value of @xmath46 , we make an informed guess assuming @xmath50 in order to obtain the random phase estimators .
inspired by our results for direct phase observation , we decided to take this a step further .
we chose the estimator ( abusing the notation ) @xmath83 in order to find the @xmath2 value .
this choice proved optimal up to a linear transformation which guarantees local unbiasedness at @xmath50 ( see appendix [ ap : optimal ] for proof ) : @xmath84 as the cramer - rao inequality in eq .
( [ eq : fisher ] ) is saturated , which we prove as follows .
we have @xmath85 bacuse of the definition of the @xmath86 phase and the fact that @xmath87 is gaussian .
we also have @xmath88\,\pi_x\}$ ] . therefore : @xmath89\,\pi_x\ } } { \mathrm{tr}\{\bar{\rho}_{\phi_0}\pi_x\}}= \frac{\partial } { \partial \phi}|_{\phi=\phi_0}\log(\bar p_{\phi}(x))\quad \mathrm{and}\\ \delta_{c}^{-4}\delta^2_{\phi_0}\hat\phi&=&\delta_{c}^{-4}\int_x
\mathrm{d}x\,\bar{p}_{\phi_0}(x ) \,(\hat\phi(x)-\phi_0)^2\,=\ , f_{\phi_0,\bar{\rho},\pi}. \label{eq : bayesfigen}\end{aligned}\ ] ] the relation in eq .
( [ eq : bayesfigen ] ) was presented in a different context in @xcite in the case of one - dimensional gaussian distribution , which can be related to collective dephasing . in order to obtain the upper bound on the quantum fisher information ,
let us look at the bayesian estimation of the random phase @xmath86 , which has a gaussian distribution @xmath90 with a mean @xmath2 and a variance @xmath91 .
the above mentioned estimator @xmath74 is also the best bayesian estimator for phase @xmath86 when @xmath50 .
therefore , the average error of phase @xmath86 estimation equals ( see appendix [ ap : bayesmulti ] ) : @xmath92 according to eq .
( [ eq : bayesfigen ] ) , the optimal measurements in the bayesian estimation of @xmath86 and in the fisher information approach to @xmath2 estimation are exactly the same .
the average error @xmath93 is bounded from below by the bayesian cramer - rao inequality @xcite .
for a gaussian distribution @xmath90 and the interferometry setup , we have ( see appendix [ ap : bayesquantum ] ) : @xmath94 where @xmath95 is the quantum fisher information for the initial state @xmath5 .
combining eqs .
( [ eq : bayesfigen ] ) , ( [ eq : bayeserrormain ] ) we obtain : @xmath96 .
we thus arrive at the * main result * of this paper , maximising @xmath97 w.r.t . to the measurement : @xmath98 which , in turn , leads to the following bound on phase @xmath2 estimation precision for any locally unbiased estimator @xmath74 : @xmath99 this bound , which takes into account both dephasing strength via @xmath91 and the available resources via @xmath100 , can be interpreted as follows .
if a perfect random phase @xmath86 observation were possible , the local error would be @xmath101 .
as this is not possible , the error is greater by @xmath102 , taking into account the noise of observing phases only via the results of the interferometry experiment .
we recover precision @xmath103 , which characterises the unitary dynamics , when the random phases variances converge to 0 , since that implies @xmath104 .
this guarantees that the bound will be tight in the presence of weak dephasing . +
* examples .
* let us consider the two following examples of correlated dephasing with the covariance matrices : @xmath105 we obtain : @xmath106 we see that , for any value @xmath107 , for constant correlations ( discrete topology ) the bound in eq .
( [ eq : boundgenerror ] ) converges to a constant @xmath108 , whereas for exponentially decaying correlations in one dimension , we obtain a better asymptotic scaling @xmath109 , see the lhs in fig .
[ fig : bounds ] . [ cols="^,^ " , ] the case @xmath110 corresponds to collective dephasing , both for @xmath111 and @xmath112 : @xmath113 which , for a single - mode gaussian state @xmath5 of photons in a two - arm interferometer with an average number @xmath114 , due to @xmath115 the form @xmath116 . in @xcite this
was proved using a variational approach to the quantum fisher information .
the bound in eq .
( [ eq : boundfc ] ) has a * constant asymptotic behaviour*. as the case @xmath117 corresponds to independent dephasing , we obtain : @xmath118 given that for a state @xmath5 of @xmath0 two - level atoms we have @xmath119 , where @xmath34 is the pauli matrix along the @xmath35-axis , and thus @xmath120 , we arrive at @xmath121 . in @xcite
a different upper bound @xmath122 was proved .
the bounds are compared on the rhs in fig .
[ fig : bounds ] .
the bound in @xcite works better for strong dephasing @xmath123 , but does not provide heisenberg scaling when @xmath124 .
both bounds show the scaling @xmath3 when @xmath1 .
the bound in eq .
( [ eq : boundgenerror ] ) provides an insight into interferometry in the presence of dephasing , the asymptotic precision of which is determined by the noise correlations . in the case of collective dephasing ,
if we were able to estimate the phases of atoms perfectly , it would be only one phase @xmath125 being a gaussian variable with the variance @xmath36 .
this is the exact bound in eq .
( [ eq : boundfc ] ) when @xmath1 .
if the phases are strongly correlated , as in the @xmath112 example , we effectively have a finite number of independent noise realisations and can not completely eliminate the dephasing noise , even if @xmath1 .
thus we observe that the bound in eq .
( [ eq : boundgenerror ] ) converges to the constant @xmath126 .
if the phases are weakly correlated , as in the @xmath111 example , we can eliminate the noise , but the best possible scaling will be reduced from the heisenberg scaling @xmath127 to the shot - noise scaling @xmath128
. + * summary and comments . * in this paper we present a new upper bound on the quantum fisher information in the presence of arbitrarily correlated gaussian dephasing that we have derived .
this bound , as shown in eq .
( [ eq : boundgen ] ) , takes into account both dephasing correlations and initial system state preparation .
moreover , it enables one to analyse the asymptotic scaling of phase estimation precision when the number of subsystems @xmath1 .
we also show that weak ( exponentially decaying ) correlations of dephasing noise preserve the scaling @xmath3 which is characteristic in independent dephasing .
arbitrarily small , but strong ( non - decaying ) correlations limit the precision scaling when @xmath1 to a constant error .
the bound derived can be fruitfully modified to frequency estimation @xcite
. + + * acknowledgements . *
the author is grateful to mdlin gu and sammy ragy for inspiring discussions .
we would also like to thank michael hush for helpful comments on the draft of this work .
this research was supported by the school of mathematical sciences and the school of physics & astronomy at the university of nottingham .
10 d. j. wineland _
et al . _ ,
spin squeezing and reduced quantum noise in spectroscopy , _ phys .
rev . a _ * 46 * , r6797 ( 1992 ) , d. leibfried _ et al .
_ toward heisenberg - limited spectroscopy with multiparticle entangled states , _ science _ * 304 * , 1476 ( 2004 ) .
c.f .
roos _
et al . _ , designer atoms for quantum metrology , _ nature _ * 443 * , 316 - 319 ( 2006 ) .
ligo scientific collaboration , a gravitational wave observatory operating beyond the quantum shot - noise limit : squeezed light in application , _ nature phys . _ * 7 * , 962 ( 2011 ) .
v. giovannetti , s. lloyd , l. maccone , advances in quantum metrology , _ nature photonics _ * 5 * , 222 ( 2011 )
. b. m. escher , r. l. de matos filho , l. davidovich , general framework for estimating the ultimate precision limit in noisy quantum - enhanced metrology , _ nature physics _ * 7 * , 406411 ( 2011 ) .
r. demkowicz - dobrzaski , j.koodyski and m. guta , the elusive heisenberg limit in quantum - enhanced metrology , _ nature comm .
_ , * 3 * , 1063 ( 2012 ) .
c. w. helstrom , _ quantum detection and estimation theory _
( academic press , new york , 1976 ) , s. l. braunstein and c. m. caves , _ phys . rev .
lett . _ * 72 * , 3439 ( 1994 ) .
b. m. escher , l. davidovich , n. zagury , and r. l. de matos filho , quantum metrological limits via a variational approach , _ phys .
rev .
lett . _ * 109 * , 190404 ( 2012 ) .
k. macieszczak , r. demkowicz - dobrzaski , m. fraas , optimal quantum frequency estimation , _ arxiv e - prints _ ( 2013 ) , arxiv:1311.5576 [ quant - ph ] .
h. l. van trees , _ detection , estimation and modulation theory _
, part i ( wiley , new york , 1968 ) . k. macieszczak [ in preparation ] . |
chain polymers carrying electro - negative atoms ( e.g. oxygen or nitrogen ) in their repeat unit can act as solvents for certain salts .
well - known examples are li - salts dissolved in polyethylene - oxide ( peo ) . at temperatures sufficiently above the glass - transition temperature these polymer - salt solutions
show significant dc ionic @xmath0 conductivities .
such `` polymer electrolytes '' offer widespread applications in batteries , sensors and fuel cells . from the scientific point of view
, an important goal is to improve our understanding of the electrical conduction mechanism in polymer electrolytes , and in particular to elucidate the interplay between ion diffusion and the polymer network dynamics .
dynamic monte carlo ( mc ) simulation of the diffusion coefficient of few particles ( ions ) in a rearranging environment of polymer chains is hampered by the need to move every monomer ( polymer bead or ion ) with the same probability .
therefore most of the computational time is spent moving the polymeric host without affecting the ionic configurations . on the other hand ,
earlier studies indicate that important features of the ion diffusion within a dynamical matrix of chain molecules can be described by a more coarse - grained model@xcite .
the idea is to map the diffusion process onto dbp - theory @xcite and to determine the central quantity entering this theory , the renewal time distribution @xmath1 , from the time dependence of the local occupational correlation function due to the polymer chain dynamics . in this communication
we perform tests of such a procedure for a hard - core lattice gas and for tracer diffusion in systems of athermal lattice polymers .
it turns out that the dbp - concept compares favourably with simulations of the complete system dynamics and thereby saves about one order of magnitude in computer time .
in order to explain our procedure we treat the simple case of a hard - core lattice gas , i.e. point particles moving on a lattice with no interaction besides hard core repulsion . the tracer diffusion constant @xmath2 as a function of concentration @xmath3 is known to a high degree of accuracy via dynamic pair approximations @xcite which yield a tracer correlation factor @xmath4{\langle}\cos\theta { \rangle}},\ ] ] with @xmath5 for a simple cubic lattice .
now we consider a lattice with static disorder in which randomly chosen sites are blocked and thus not accessible for the tracer particle .
for this standard percolation problem we denote the mean square displacement of the tracer by @xmath6 .
if on the other hand the blocked sites are globally and instantaneously rearranged according to an arbitrary waiting - time distribution @xmath1 , generalized dbp - theory yields the following diffusivity at zero frequency in @xmath7 dimensions @xcite @xmath8 our aim is to map the complete system of coupled ions and lattice chains onto this coarse - grained dbp - model . while @xmath6 can be obtained in a straightforward manner from simulations of ion diffusion in the frozen network , the determination of @xmath1 requires more explanation . within the spirit of previous work on the hard - core lattice gas we propose to determine @xmath1 from the local occupational correlation function @xmath9 of a site i adjacent to a fixed tracer position
let us introduce the probability @xmath10 that no renewal takes place within the time interval @xmath11 $ ] after a previous renewal at an arbitrary @xmath12 .
the joint probability @xmath13 that the lattice point @xmath14 is occupied at @xmath15 and at @xmath16 ( not necessarily by the same particle ) consists of two distinct contributions .
the first one is the probability that @xmath14 is occupied at @xmath15 and that no renewal occurs until @xmath16 , which is given by @xmath17 .
the second contribution describes the situation where one or more renewals have taken place until @xmath16 .
the corresponding probability is given by @xmath18 .
hence @xmath10 is related to @xmath13 via @xmath19 according to @xcite , @xmath20 where @xmath21 denotes the mean renewal time .
equation ( [ eq:2 ] ) thus can be rewritten as @xmath22 where we obtain @xmath10 from ( [ eq:3 ] ) by means of mc simulations for @xmath23 .
the resulting @xmath10 can be fitted with sufficient accuracy by a combination of @xmath24 exponential functions , @xmath25 on the other hand , @xmath26 can by expressed in terms of the simulated mean square displacement @xmath27 of a tracer particle making @xmath28 steps in a frozen network via @xmath29 where @xmath30 is the attempt frequency of the tracer . inserting ( [ eq:5 ] ) and ( [ eq:6 ] ) into ( [ eq:4 ] ) and carrying out the integration one gets @xmath31 independent calculations for @xmath32 showed that in order to keep the error caused by the finiteness of @xmath33 and @xmath34 below the statistical error , values of @xmath35 and @xmath36 were sufficient .
note that the main decay of @xmath10 is roughly represented by a single exponential with decay rate @xmath37 . in figure ( [ fig:1 ] )
we compare the results of this theory with mc simulations ( see also @xcite ) and the dynamic pair approximation equation ( [ eq:1 ] ) and also with an existing effective medium approximation ( ema ) for a percolating lattice renewing itself with only one rate constant @xmath38 @xcite .
as one can see , using more than one rate constant in the representation of @xmath10 according to ( [ eq:5 ] ) improves the results significantly .
this observation suggests to apply the generalized dbp - theory to the situation described in the next section where the non - exponential character of the renewal process is even more pronounced .
in the case of a hard - core lattice gas the procedure described bears no computational advantage over existing methods but was merely considered as a test case .
the situation changes , however , when we apply our approach to ion diffusion in a polymer network . no analytic approximations for @xmath2 of a quality similar to ( [ eq:1 ] )
are available then , and full simulations of @xmath39 become much more demanding because of the internal degrees of freedom of the host molecules and the larger statistical errors connected with the small concentration of tracer particles .
let us consider , for example , a system where the chains are made of ten beads , assigned to lattice sites , and linearly connected via nearest - neighbour bonds .
apart from site exclusion , which mimics a hard core repulsion , no explicit interactions between beads are assumed .
the size of the simulation box is @xmath40 and periodic boundary conditions in all directions are employed . for the chain dynamics we use the standard algorithm as described in ref .
@xcite , while point - particles individually perform nearest - neighbour hops . to apply our theory we again determine @xmath10 in analogy to the hard - core lattice gas , see ( [ eq:5 ] ) , which is now highly non - exponential , indicating the importance of temporal correlations in the associated renewal processes . a superposition of three exponential functions in ( [ eq:5 ] )
, however , still gives a good description of @xmath10 .
mean - square displacements @xmath27 of ions within a frozen network of chains are determined using @xmath41 steps , and the diffusion constant is calculated as before via equation ( [ eq:7 ] ) . in fig.([fig:2 ] ) we compare our results for the tracer correlation factor as a function of the total concentration of occupied sites , c , with full mc simulations .
due to the connectivity of chains blocking effects at a given @xmath3 on average are reduced and the correlation factor @xmath42 is larger than in the case of point particles , at least as long as @xmath43 .
more important from the computational point of view is the fact that dbp - theory very well agrees with the full simulations , but saves about one order of magnitude in computer time .
mapping that problem onto a coarse - grained dbp - model thus appears to constitute an accurate and efficient method for investigating diffusion through a dynamic host of chain molecules . to support this conclusion further work
is being carried out , in particular under varying chain lengths .
k. nakazato and k. kitahara , progr .
64 * , 2261 ( 1980 ) r. a. tahir kheli and r. j. elliott , phys .
b * 27 * , 844 ( 1983 ) w. dieterich , j. stat .
* 39 * , 583 ( 1985 ) h. j. de bruin and g. e. murch , philos . mag .
* 27 * , 1475 ( 1973 ) r. granek and a. nitzan , j. chem
. phys . * 92 * , 1329 ( 1990 ) p.
h. verdier and w. h. stockmayer , j. chem . phys . *
36 * , 227 ( 1962 ) h. j. hilhorst and j. m. deutch , j. chem .
63 * , 5153 ( 1975 ) k. kremer and k. binder , computer reports * 7 * , 259 ( 1988 ) | diffusion of ions through a fluctuating polymeric host is studied both by monte carlo simulation of the complete system dynamics and by dynamic bond percolation ( dbp ) theory .
comparison of both methods suggests a multiscale - like approach for calculating the diffusion coefficients of the ions .
charge transport in polymer ion conductors : a monte carlo study + |
the potential to realize interesting quantum spin liquid states with fractionalized excitations and topological order has driven research on frustrated magnets in the last decades.@xcite one of the most promising candidate models is the spin-1/2 heisenberg antiferromagnet on the two - dimensional kagome lattice .
many theoretical attempts have been made to unravel its groundstate properties , which are still not fully understood .
while early approaches supported a symmetry broken valence bond solid state,@xcite various different groundstates have been proposed since .
recent numerical works based on the density matrix renormailzation group ( dmrg ) method provide strong evidence for a gapped @xmath2 spin liquid state,@xcite whereas projected wavefunction studies favor a gapless @xmath3-dirac spin liquid groundstate,@xcite but this issue is not settled yet.@xcite both of these states do not break lattice symmetries and lack conventional long - range magnetic order due to strong quantum fluctuations associated with the frustrated spin - exchange interactions .
the interest in chiral spin liquids , which break time - reversal and parity symmetries , was triggered by kalmeyer and laughlin , who proposed that bosonic analogues of fractional quantum hall states could be realized in frustrated magnets.@xcite within a slave - fermion approach these chiral states are stable phases of matter , because gauge fluctuations are gapped by a chern - simons term.@xcite more recently , various theoretical works showed that such chiral spin liquids can be stabilized on the kagome lattice either by including further - neighbor interactions or additional terms that explicitly break time - reversal symmetry.@xcite as far as experiments are concerned , the mineral herbertsmithite as well as organic charge transfer salts are the most promising candidate materials to host a spin liquid groundstate.@xcite while measurements on the triangular lattice organic salts are consistent with a gapless spin liquid , the kagome lattice compound herbertsmithite likely has a gapped spin liquid groundstate .
inelastic neutron scattering experiments are compatible with a continuum of fractionalized spinon excitations,@xcite and recent nmr measurements indicate that the groundstate is gapped.@xcite the fact that no sharp onset of the two - spinon continuum was observed in neutron scattering has been attributed to the presence of a flat band of topological vison excitations in gapped @xmath2 spin liquids,@xcite as well as to the contribution from impurities at low energies.@xcite various different spin liquid states have been proposed as potential groundstates of kagome heisenberg antiferromagnets . in order to relate theoretical results to inelastic neutron scattering experiments , a better characterization of dynamical structure factors in kagome systems is clearly beneficial . in this work
we take a step in this direction by computing dynamical spin structure factors of simple chiral spin liquids using schwinger - boson mean - field theory.@xcite our approach is based on an earlier projective symmetry group classification of time - reversal symmetry breaking mean - field anstze by messio , lhuillier , and misguich.@xcite we show that the dynamical spin structure factor @xmath4 shows clear signatures of time - reversal symmetry breaking , in contrast to static ( equal - time ) structure factors . in particular , momentum inversion symmetry @xmath5 is lost and consequently the six - fold rotation symmetry of @xmath4 around the @xmath1-point is reduced to three - fold rotations .
moreover , we show that the onset of the two - spinon continuum is rather flat for the _ cuboc1 _ state , which has been argued to minimize the groundstate energy of the kagome heisenberg antiferromagnet within the schwinger - boson approach.@xcite this particular chiral spin liquid state is a quantum disordered version of the magnetically ordered _ cuboc1 _ state , which is a possible non - coplanar state of the classical afkm model .
@xcite it is important to note that this schwinger - boson construction does not lead to chiral spin liquids of the kalmeyer - laughlin type .
this is due to the fact that the condensation of boson bilinears reduces the gauge symmetry from @xmath3 to @xmath2 .
consequently , the effective low energy theory is a chern - simons - higgs theory with a condensed charge-2 higgs field , the topological properties of which are typically equivalent to @xmath2 gauge theory.@xcite the remainder of our paper is structured as follows : in section [ sec : diag ] we review the schwinger - boson mean - field theory ( sbmft ) for time - reversal symmetry breaking anstze and calculate the spinon dispersions . in section [ sec : self ]
we determine the mean - field parameters self - consistently for all sbmft anstze considered in this work . in section [ sec : ssf ] the spin structure factor for a general chiral sbmft ansatz is derived .
lastly , in section [ sec : results ] we present and discuss the numerically computed structure factors . we conclude with section [ sec : conclusion ] .
the hamiltonian of the antiferromagnetic heisenberg model is given by @xmath6 where @xmath7 , @xmath8 is the spin operator on lattice site @xmath9 and the sum runs over nearest neighbor sites . in the following we use the schwinger - boson representation of spin operators @xmath10 where here , and throughout this paper
, we employ a summation convention over repeated greek indices , and @xmath11 , @xmath12 are bosonic annihilation and creation operators , respectively , of spin @xmath13 on site @xmath9 .
consequently , the hamiltonian can be written as @xmath14 where the langrange multiplier @xmath15 constrains the number of bosons per site to @xmath16 , with @xmath17 the length of the spin .
note that this length constraint is only imposed on average here and in the following mean - field approximation .
it can be enforced exactly by allowing for a space- and imaginary time dependent lagrange multiplier , which leads to a theory of bosonic spinons coupled to an emergent u(1 ) gauge field.@xcite however , for the spin liquid states considered in this work , the condensation of bosonic bilinears gaps out gauge fluctuations and the mean - field approximation is justified .
we now introduce the @xmath18-invariant bond operators @xmath19 where @xmath20 is the fully antisymmetric tensor of @xmath18 .
one can show that @xmath21 for @xmath22 and the hamiltonian can be rewritten as @xmath23 next , we apply a mean - field decoupling of the bond operators resulting in the mean - field hamiltonian @xmath24 @xmath25 and @xmath26 are free complex mean - field parameters that will be computed self - consistently by extremizing the free energy .
even though most sbmft studies use one or the other , including both @xmath25 and @xmath26 has been proven to lead to a better description of the spectrum of excitations in frustrated magnets,@xcite where @xmath25 describes singlet amplitudes and @xmath26 describes boson hopping amplitudes .
a set @xmath27,@xmath28 specifies a mean - field ansatz .
for the symmetric , time - reversal breaking spin - liquids considered in refs .
[ , ] the mean - field parameters take the form @xmath29 where the moduli @xmath30 and @xmath31 are the same on each bond , but the phases @xmath32 and @xmath33 are bond - dependent .
the detailed form of these anstze is shown in fig .
[ fig : unitcell ] . taking the fourier transform of the schwinger - boson operator as @xmath34 where @xmath35 is a band index , @xmath36 is the position of site @xmath9 , and @xmath37 is the number of @xmath38-points summed over in the fourier transform , and adopting a general chiral ansatz following the notation of ref . [ ]
, the hamiltonian in reciprocal space reads @xmath39 where we have introduced the spinor @xmath40 with the superscript letters denoting one of the six bands @xmath41 of the six - site unit - cell shown in fig .
[ fig : unitcell ] , and @xmath42 is the hermitian block matrix @xmath43 with @xmath44 the @xmath45 identity matrix , and 1@xmath46 2@xmath47 3@xmath48 @xmath49 @xmath50 here we denote the real - space vectors @xmath51 , @xmath52 , and @xmath53 , and @xmath54 , as shown in fig .
[ fig : unitcell ] , with @xmath55 and @xmath56 the inter - site spacing , which we set to unity .
finally we perform a bogoliubov transformation by defining the bosonic operators @xmath57 which are related to the schwinger - boson ladder operators by the linear transformation @xmath58 whereby @xmath59 will satisfy the canonical commutation relations and diagonallze @xmath60 if @xmath61 where the bogoliubov rotation matrix takes the block form @xmath62 furthermore , @xmath63 and @xmath64 is a @xmath65 diagonal matrix representing the bosonic eigenenergies for up - spins at momentum @xmath66 and down - spins at momentum @xmath67 , where the @xmath68 diagonal matrix @xmath69 carries the bosonic eigenenergies @xmath70 along its diagonal with @xmath35 the band index whose values range in @xmath41 , the six bands comprising our unit cell as illustrated in fig .
[ fig : unitcell ] , while @xmath71 and @xmath13 are the momentum and spin polarization , respectively , and @xmath72 is the @xmath68 zero matrix .
it is to be noted here that @xmath73 has this form due to @xmath18 symmetry .
note , however , that for chiral anstze one has @xmath74 , because the @xmath75 symmetry is broken
. nevertheless , we still have @xmath76 due to @xmath18 symmetry .
' '' '' height 1pt @a xhline & & & & & + ' '' '' height 1pt @a xhline @xmath77 & & & & & + ( r)1 - 1 @xmath78 & & & & & + ( r)1 - 1 @xmath79 & & & & & + ( r)1 - 1 @xmath80 & & & & & + ( r)1 - 1 @xmath81 & & & & & + ( r)1 - 1 @xmath82 & & & & & + ( r)1 - 1 @xmath15 & & & & & + ( r)1 - 1 ' '' '' height 1pt @a xhline @xmath83 & & & & & + ( l)6 - 6 ' '' '' height 1pt @a xhline [ table ] before proceeding with the numerical computation of the spin structure factors we first determine the self - consistent mean - field parameters for each ansatz that we consider by extremizing the free energy @xmath84 where @xmath85 are the free mean - field parameters ( denoted by an asterisk in table [ table ] ) of the given ansatz , and @xmath86 is the mean - field free energy derived from eq . to be @xmath87 where b.z .
stands for the 1st brillouin zone . in this work
we consider two non - chiral anstze ( _ q=0 _ and @xmath88 ) as well as three further anstze ( _ cuboc1 _ , _ cuboc2 _ , and _ octahedral _ ) that can break time - reversal.@xcite in the following we set @xmath89 and @xmath90 . with this artificially small value of the spin
we ensure that all anstze describe a state deep in the spin - liquid phase , which is what we re interested in .
we find the stationary point of @xmath86 by an adaptive - grid method that seeks to minimize @xmath91 , stopping only when this sum is of the order of @xmath92 or better .
the results of this extremization procedure for all anstze that we consider are shown in table [ table ] .
the self - consistent mean - field parameters for the different anstze turn out to be quite close to one another . in particular , all five anstze exhibit very similar values for @xmath78 , @xmath80 , and @xmath15 , the only parameters that are free in all anstze .
the main difference is in the phases , some of which are fixed by the specific form of an ansatz , while others are free .
it is important to note that the phase @xmath93 for the _ cuboc2 _ as well as the _ octahedral _ ansatz turns out to be equal to @xmath94 within numerical accuracy .
consequently , the saddle points of these two anstze describe non - chiral spin - liquid phases , where time - reversal and parity symmetries are restored . in our computation
only the _ cuboc1 _ ansatz turns out to be chiral .
note , however , that interactions beyond nearest neighbors can stabilize chiral saddle points of the _ cuboc2 _ form , where the phase @xmath93 takes a non - trivial value.@xcite
the dynamic spin structure factor is defined as @xmath95 which , using the bogoliubov operators , can be expressed at @xmath96 and in the absence of a spinon condensate as @xmath97\delta(\omega-\epsilon^n_{\bm{q}\downarrow}-\epsilon^m_{\bm{k}-\bm{q}\uparrow})+\\ \nonumber & 2x^*_{sn}(-\bm{q})y^*_{sm}(-\bm{k}+\bm{q})\bigg[y_{rm}(-\bm{k}+\bm{q})x_{rn}(-\bm{q})+y_{rn}(-\bm{q})x_{rm}(-\bm{k}+\bm{q})\bigg]\delta(\omega-\epsilon^n_{\bm{q}\downarrow}-\epsilon^m_{\bm{k}-\bm{q}\downarrow})+\\ \nonumber & 2v_{sn}(\bm{q})u_{sm}(\bm{k}-\bm{q})\bigg[u^*_{rm}(\bm{k}-\bm{q})v^*_{rn}(\bm{q})+u^*_{rn}(\bm{q})v^*_{rm}(\bm{k}-\bm{q})\bigg]\delta(\omega-\epsilon^n_{\bm{q}\uparrow}-\epsilon^m_{\bm{k}-\bm{q}\uparrow})+\\ & v_{sn}(\bm{q})y^*_{sm}(-\bm{k}+\bm{q})\bigg[y_{rm}(-\bm{k}+\bm{q})v^*_{rn}(\bm{q})-u^*_{rn}(\bm{q})x_{rm}(-\bm{k}+\bm{q})\bigg]\delta(\omega-\epsilon^n_{\bm{q}\uparrow}-\epsilon^m_{\bm{k}-\bm{q}\downarrow } ) \bigg\},\end{aligned}\ ] ] where @xmath35 , @xmath98 , @xmath99 , and @xmath100 are band indices taking values in @xmath41 .
the static ( equal - time ) structure factor is obtained by integrating over frequencies @xmath101 .
we use the vegas@xcite monte carlo integration routine to numerically evaluate the dynamic structure factors of the anstze shown in table [ table ] , while approximating the dirac @xmath102-functions in eq . as lorentzian functions with a width @xmath103 for numerical reasons . the non - chiral anstze _
q=0 _ and @xmath88 were first discussed in ref .
[ ] , and their dynamic structure factors were calculated in ref . [ ] , although using an ansatz with @xmath104 .
similar dynamical structure factors for fermionic mean - field spin liquids have been computed in ref .
[ ] . in fig .
[ fig : nonchiral ] , we show their dynamic structure factors along the @xmath1-m - k-@xmath1 high - symmetry lines in the gapped spin liquid phase with @xmath90 , using the self - consistent mean - field parameters shown in table [ table ] .
note that we ve adopted a normalization where the maximum of the structure factors is set to unity for convenience .
our results are qualitatively very similar to those in ref .
the small differences come from the fact that the spinon dispersions are slightly altered when taking non - zero @xmath105 s into account
. figs .
[ fig : cuboc1 ] , [ fig : cuboc2 ] , and [ fig : octahedral ] show the dynamic structure factors in the @xmath66-plane for fixed frequencies @xmath106 , as well as along the @xmath1-m - k-@xmath1 high - symmetry lines for the _ cuboc1 _ , _ cuboc2 _ , and _ octahedral _ anstze , respectively , in the gapped spin liquid phase with spin @xmath90 . the dynamic structure factor at @xmath107 for _ cuboc1 _ in fig .
[ fig : cuboc1 ] shows that inversion symmetry is lost with respect to the @xmath1-point due to time - reversal symmetry breaking .
consequently , the usual six - fold rotational symmetry is reduced to a three - fold one .
one would expect to see the same for the _ cuboc2 _
( at @xmath107 ) and _ octahedral _ ( at @xmath108 ) anstze in figs .
[ fig : cuboc2 ] and [ fig : octahedral ] , respectively , as they allow for nontrivial aharonov - bohm phases when a spinon is taken around a plaquette .
however , as shown in table [ table ] , the saddle point value of @xmath93 at @xmath90 is equal to @xmath94 within numerical accuracy , which , along with the fixed phase @xmath109 ( _ cuboc2 _ ) or @xmath94 ( _ octahedral _ ) , leads to a time - reversal invariant ansatz . the @xmath75 symmetry as well as the six - fold rotational symmetry is thus retained in the dynamical structure factor . on the other hand , for the _ cuboc1 _
ansatz , the free phase @xmath79 takes on a value other than @xmath110 ( @xmath111 ) , leading to a chiral spin liquid with broken time - reversal and parity symmetry . an interesting feature in the _ cuboc1 _ dynamical structure factor along the @xmath1-m - k-@xmath1 high - symmetry lines in fig .
[ fig : cuboc1 ] is the relatively flat onset of the two - spinon continuum compared to the _
q=0 _ and @xmath112 case . the static structure factors for the _ cuboc1 _ , _
cuboc2 _ , and _ octahedral _ anstze , shown in fig .
[ fig : ssf ] , are qualitatively very similar .
note that the static structure factor for the chiral _
cuboc1 _ state does nt show signs of time - reversal or parity symmetry breaking .
this can be understood by recognizing from eq .
that one can write @xmath113 by a simple relabeling of the site indices .
it is clear from eqs . and that @xmath114 only if @xmath115 is time reversal invariant
. on the other hand , the static ( equal - time ) structure factor is always invariant under @xmath0 .
lastly , the variational groundstate energies of the five different anstze considered here are listed in the last line of table [ table ] .
we find that the non - chiral @xmath112 ansatz has the lowest energy at @xmath90 .
note that this is in contrast to ref .
[ ] , who find that the _ cuboc1 _ state has the lowest energy .
it is worth mentioning here that the sbmft approach is not quantitatively reliable to find the true groundstate of the kagome heisenberg antiferromagnet , nor does it give variational upper bounds to the groundstate energy .
this is due to the fact that the spin - length constraint is only imposed on average , and thus unphysical states are only approximately projected out .
consequently , the main purpose of our work is not to determine the true groundstate of the kagome heisenberg antiferromagnet , but to highlight features in dynamical correlation functions of different chiral spin liquid states .
we computed static and dynamic spin structure factors of several chiral and non - chiral sbmft anstze deep in the gapped spin liquid phase at spin @xmath90 . even though the _ cuboc1 _ , _ cuboc2 _ , and _ octahedral _ anstze allow for time - reversal symmetry breaking , only the saddle - point of the _ cuboc1 _
ansatz is chiral , which can be seen directly in the dynamic structure factor at fixed frequency , as seen in fig .
[ fig : cuboc1 ] .
time - reversal symmetry breaking manifests itself by breaking the inversion symmetry with respect to the @xmath1-point , as well as reducing the usual six - fold rotational symmetry present for non - chiral anstze to a three - fold rotational symmetry . | we calculate dynamical spin structure factors for gapped chiral spin liquid states in the spin-1/2 heisenberg antiferromagnet on the kagome lattice using schwinger - boson mean - field theory .
in contrast to static ( equal - time ) structure factors , the dynamical structure factor shows clear signatures of time - reversal symmetry breaking for chiral spin liquid states . in particular , momentum inversion @xmath0 symmetry as well as the six - fold rotation symmetry around the @xmath1-point are lost .
we highlight other interesting features , such as a relatively flat onset of the two - spinon continuum for the _ cuboc1 _ state .
our work is based on the projective symmetry group classification of time - reversal symmetry breaking schwinger - boson mean - field states by messio , lhuillier , and misguich . |
quasars form a class of objects that frequently warrant superlatives in their descriptions : 1745 + 624 ( 4c+62.29 ) is no exception .
it was identified as one of the highest redshift x - ray quasars at the time of its discovery ( @xmath0=3.87 ; * ? ? ?
* ) from spectroscopic followup of statistically significant sources from the _ einstein _ x - ray observatory .
further optical spectroscopy by @xcite refined the redshift to @xmath0=3.89 , affirmed by @xcite ; the latter value is adopted here .
after the _ einstein _ detection , it was established as a bona fide x - ray source by _ rosat _
@xcite , _ asca _
@xcite , and _ bepposax _ @xcite . by virtue of its high - redshift ,
it is one of the most radio - luminous quasars known ( cf .
figure 1 of * ? ? ?
* ) its observed 1.4 ghz luminositykm s@xmath7 mpc@xmath7 , @xmath8 and @xmath9 and have converted quoted literature values of luminosities and proper motions to this cosmology .
if we assume the cosmology adopted in @xcite , @xmath10 w hz@xmath7 ( 1 watt = 10@xmath11 erg s@xmath7 ) . ] , @xmath12 w hz@xmath7 , is over an order of magnitude greater than that of the archetype luminous quasar 3c 273 @xcite .
it is probably no coincidence that its 2.5 long jet ( 18 kpc projected ; * ? ? ?
* ) is also the most luminous radio jet currently known ( that we are aware of ; cf . * ? ? ?
* ; * ? ? ?
* ) , accounting for @xmath131/4th of the source s total 1.4 ghz luminosity .
this quasar is of interest to us because of this prominent radio jet , visible on both milli - arcsecond ( * ? ? ?
* ; * ? ? ?
* figure [ figure-1 ] ) and arcsecond - scales ( figures [ figure-2 ] & [ figure-3 ] ) .
we were prompted to identify such high - redshift systems with large - scale jets after the _ chandra _ x - ray observatory detection of a @xmath132.5 long x - ray extension @xcite in the @xmath0=4.3 quasar gb 1508 + 5714 @xcite .
subsequent vlba imaging of gb 1508 + 5714 has revealed a parsec - scale jet that can be traced out to @xmath13100 milli - arcseconds ( 0.7
kpc projected ; * ? ? ?
* ) , aimed in the general direction of the kpc - scale structure , supporting the jet interpretation .
these are the two most distant quasars with a _ kiloparsec - scale _ jet=5.47 blazar q0906 + 6930 by @xcite , but on vlbi scales only . ] detected at any wavelength @xcite .
the identification of such systems in the early universe is extremely interesting for a number of reasons .
jets are signposts for `` active '' black hole / accretion disk systems ( e.g. , * ? ? ?
* ) , requiring central nuclei of high-@xmath0 galaxies to be sufficiently well - developed to sustain gravitational collapse . at such early epochs , it is remarkable that the jet production process is apparently efficient and persists long enough ( @xmath14myr ) to have produced such large ( 10s100 s kpc - scale ) structures . in this context
, jets may even serve as tracers of the directionality of the bh / disk axis in these early accreting systems , as chronicled in the `` bumps and wiggles '' in our jet images . also , these jets are an efficient means to shock heat ambient gas thus triggering early star formation ( e.g. , * ? ? ?
* ) . among other important issues ,
such high - redshift systems are potentially very important to our understanding of the emission processes responsible for the production of the broad - band jet emission , and the x - rays in particular .
although the number of quasars with kpc - scale jets detected in the x - rays is growing @xcite , the origin of this radiation is still in active debate .
@xcite noted early on that a comparison of cosmologically distant jets with local examples could in principle provide crucial arguments to the debate .
this is because the two main contending models for the x - ray emission of large - scale quasar jets synchrotron radiation and inverse compton scattering of the cosmic microwave background ( ic / cmb ) have , in their simplest versions , markedly different predictions of the x - ray jet properties with redshift ( see [ section - z ] ) .
a high - resolution _ chandra _ image of the @xmath0=3.89 quasar 1745 + 624 was obtained in an early observing cycle ( table [ table-1 ] ) and x - ray emission from the arcsecond - scale radio jet is in fact detected and is quite prominent also .
apart from the gb 1508 + 5714 case mentioned above , evidence has been presented recently for at least one more high-@xmath0 quasar with extended x - ray emission ( j22192719 at @xmath0=3.63 ; * ? ? ?
* ) with a radio counterpart ( * ? ? ?
* and manuscript in preparation ) .
this makes 1745 + 624 one of only three very high - redshift ( @xmath13 ) systems with extended x - rays associated with a radio jet . with the scarcity of such high - redshift systems and their importance in current discussions of x - ray jet emission models
, we have independently analyzed the archival _ chandra _ data .
it was also realized that 1745 + 624 hosts an exceptional jet making it conducive to study .
this is because while both the gb 1508 + 5714 and j22192719 cases each show a single faint ( @xmath131 mjy at 1.4 ghz ) arcsecond - scale feature separated from their bright nuclei , the 1745 + 624 jet is 2 orders of magnitude brighter than the other two cases and shows _ resolved structure_. in addition , it is the most luminous x - ray jet observed thus far ( cf .
* ) . to fully exploit the potential of the _ chandra _ observation , we have also analyzed archival multi - frequency imaging data from the very large array ( vla ) , and a hubble space telescope ( hst ) image to map the spectral energy distributions of the extended components in the quasar jet .
these data , along with a new deep very long baseline interferometry ( vlbi ) map of the extended subarcsecond - scale jet are described in
[ section - vlbi][section - chandra ] , with a critical assessment of the derived spectra of the kpc - scale components in [ section - spectra ] .
the physical properties of these extended components are then discussed ( [ section - discuss ] ) , with comparisons made to the other two quasars with x - ray detected knots at comparably high - redshift ( @xmath153 ) .
these distant examples are compared to other currently known x - ray jets in quasars at lower redshift ( @xmath162 ) in light of expectations from synchrotron and inverse compton emission models .
our results are summarized in [ section - summary ] .
vlbi maps of 1745 + 624 show a prominent parsec - scale jet and the emission fading with distance out to @xmath1340 mas @xcite . the lower frequency , 2.3 ghz usno radio reference frame image database maps ( rrfid , * ? ? ?
* ) suggest that the jet is even more extended .
we therefore uniformly reprocessed the six best ( @xmath17 ) datasets at this frequency from the rrfid ( obtained between 1997 march to 1998 august ) and produced images on the same grid , reconvolved with a common 4 mas beam .
we then averaged the 6 images to suppress the noise in the background resulting in improved definition of the outer jet ( figure [ figure-1 ] ) .
the parsec - scale jet in this
average " image is now clear out to @xmath13110 mas ( 0.8 kpc projected ) allowing us to trace ( [ section - describe ] ) the vlbi - scale emission to the outer structure sampled by the vla data ( figure [ figure-2 ] and [ section - radio ] ) . at 4 mas resolution
, the initial bright segment of the vlbi jet is oriented at position angle ( pa ) 212 and contains two prominent peaks at roughly 14 mas ( 22.3 mjy / bm ) and 23 mas ( 10.4 mjy / bm ) . at @xmath1350 mas from the core , the jet position angle increases ( by @xmath134 ) and transitions to the outer fainter portion .
curiously , the eastern ridgeline of the outer portion lines up with the central pa of the inner jet .
the integrated flux of the visible jet in the 2.3 ghz map is @xmath13110 mjy with the outer portion contributing about 20 mjy of this .
we measure a jet to counter - jet ratio of @xmath1450 ( defined as the ratio of the peak to 3 times the rms noise on the counter - jet side ) of the first visible peak in the 2.3 ghz map . in light of the high signal - to - noise ( s / n ) ratio _ chandra _ detection of the arcsecond - scale jet in 1745 + 624 ( [ section - chandra ] ) , we desired comparable information about its structure and spectra in the radio band .
@xcite first discovered the extended radio emission in short vla observations at 1.5 , 5 and 15 ghz . the two main x - ray emitting knots which we dub k1.4 and k1.8 ( indicating their distance from the core in arcseconds ) , and the terminal ( brightest ) jet feature at 2.5 are readily identifiable in their 5 ghz map .
for the sake of discussion , we classify the terminal knot , k2.5 , as the radio `` hot spot '' in the context of similar x - ray detected features in other powerful jets ( see below and [ section - hotspot ] ) .
@xcite noted also an 8.5 ghz vla detection of this terminal feature by @xcite from a 100 sec .
observation , though no image is presented .
these four datasets form the basis of our radio analysis of the jet . at the two higher frequencies ,
these data are supplemented by the only other available vla datasets
sparse snapshots from the more compact configurations to help fill in the shorter ( @xmath17 ) spacings ( see table [ table-1 ] for a summary of the data ) .
the data were obtained from the nrao archive and calibrated in aips @xcite utilizing scans of the calibrators 3c 48 and 3c 286 to set the amplitude scale . in the high - resolution a - configuration 15 ghz experiment ( program ab414 ; table [ table-1 ] ) , 3c 48 is heavily resolved so could not be relied on for this purpose .
fortunately , the phase calibrator 1803 + 784 was monitored frequently at this frequency by the umrao @xcite and was found to be fairly stable over the two month period bracketing the vla epoch ( within @xmath137@xmath4 of the average of 3 jy over 7 observations ) , so the amplitude scale was set with this data .
the calibrated ( @xmath17 ) data were exported into the caltech difmap package @xcite for self - calibration and imaging .
the multi - frequency images were registered on the image peaks , i.e. , the arcsecond - scale radio core .
this registration must be good to within a small fraction of the vla beams because the majority of the vlbi flux is concentrated within a few 10 s mas ( [ section - vlbi ] ) .
improving on the previous studies , we were able to separate and detect emission from the radio hot spot _ and _ the jet at all four frequencies ( figure [ figure-2 ] ) .
the 1.5 ghz image is a `` super - resolved '' image , created by reconvolving the data with a beam 1/3rd of the uniformly weighted one .
an excess due to the jet is apparent ( figure [ figure-2 ] ) .
there is also an apparent excess adjacent to the core , on the side opposite of the visible jet ; a similar excess is seen in the 5 ghz image .
this could be suggestive of real emission on the counterjet side , although it is equally likely that they are spurious due to the gaps in ( @xmath17 ) coverage of the datasets .
we considered the individual knot spectra ( [ section - spectra ] ) but were concerned with the differing ( @xmath17 ) coverages affecting the spectral measurements of the jet . to mitigate against this
, we tapered and convolved the three higher - frequency images with a common 0.35 beam ( see e.g. , figures [ figure-3 ] & [ figure-5 ] ) and integrate the radio jet emission over several beamwidths when we measured the `` total '' jet spectral index between the 3 frequencies ( [ section - spectra ] ) .
additional polarization leakage - term calibration of the 5 ghz dataset was performed using observations of the calibrator , 1748253 , from an adjacent program am337 . the electric vector position angle ( evpa ) of the resultant polarization image
was determined by setting the evpa of 3c 48 to 106@xmath18 . the integrated rotation measure ( rm ) of [email protected] radians
m@xmath20 @xcite toward 1745 + 624 amounts to a small correction of 6 at 5 ghz , and this correction has been applied to the polarization map in figure [ figure-3 ] . at 5 ghz , the initial segment of the radio jet is highly polarized at @xmath1320@xmath4 out to knot k1.4 with stronger interknot polarization of 30@xmath4 .
the rm corrected map shows the evpa perpendicular to the direction of the jet along most of this length , i.e. , the implied magnetic field is parallel to the jet quite typical in powerful one - sided jets @xcite . also typical
is the low polarization in the core ( @xmath134@xmath4 ) .
the parallel - to - the - jet magnetic field configuration roughly continues into k1.8 and the terminal radio feature , which we classify as a hot spot based on its terminal position and compactness .
the hot spot is unresolved with an average deconvolved major axis of 0.15 and aspect ratio of 0.5 elongated at position angle of 254 ( fits consistent between the 5 , 8.5 , and 15 ghz datasets ) .
this outer portion of the jet exhibits lower radio polarization levels of @xmath1356@xmath4 .
the polarization pattern in k1.8 is not uniform and field cancellation may account for its low polarization .
the dimensions of the jet knots , k1.4 and k1.8 , are not obviously resolved in our highest resolution ( 8.5 and 15 ghz ) images though these are the frequencies at which the jet is faintest .
we set an upper limit to their angular sizes ( diameters ) of 0.2 based on these data .
lastly , a new radio knot is found in the highest resolution ( 15 ghz ) image 0.4 from the nucleus ( k0.4 ) .
though not well - separated from the nucleus in the lower - resolution images , some excess emission presumably from k0.4 is seen connecting the core to the outer jet in all three lower - frequency maps ( figure [ figure-2 ] ) .
its identification in the 15 ghz data helps delineate the overall ( projected ) path of the radio jet from parsec - scales ( out to @xmath210.1 ;
[ section - vlbi ] ) to kiloparsec scales ( [ section - describe ] ) . a deep hubble space telescope ( hst ) image of 1745 + 624 was obtained from the stsci archive to search for optical emission from the jet .
the data analyzed was taken with stis using the ccd in clear ( unfiltered ) imaging mode .
this image is sensitive over a broad range in the optical band ( @xmath132,00010,000 ) with an effective wavelength of 5,740 .
four roughly equal exposures totaling just under 3,000 sec .
( table [ table-1 ] ) were aligned , then cosmic - ray rejected and combined using the crreject task in the stsdas package in iraf .
the quasar appears prominently in the image and has slightly saturated the inner few ccd pixels ( the following description follows figure [ figure-4 ] closely ) . in the vicinity of the radio jet
, a faint resolved optical source is apparent just north of the jet ( between k1.8 and the hot spot ) , but it is clearly unassociated with the extended radio source .
we modeled the central source with a series of elliptical isophotes using the ellipse task in iraf , successfully out to @xmath131.2 .
other than a clear source @xmath130.95 roughly south of the quasar , this subtraction revealed the alluded to optical emission apparently coincident with k0.4 .
as this is adjacent to the central mildly saturated source , and similar features at two different position angles about the nucleus are also seen in the subtracted image ( one is lined up with the direction of a diffraction spike ) , it is difficult to gauge its reality .
it is formally a 5@xmath22 detection in a 0.15 radius circular aperture ( contains 79@xmath4 of encircled energy ) in the central source subtracted image ( figure [ figure-4 ] ) , where @xmath22 here is the standard deviation in the background in apertures adjacent to the jet .
some of the optical emission may indeed originate from the radio source ; its radio - to - optical spectra index of @xmath131 ( table [ table-2 ] ) is quite typical in quasar jets @xcite , so we tentatively identify it as an optical counterpart to the radio knot k0.4 .
this knot is unresolved from the nucleus in the _ chandra _ image so is not discussed in the subsequent sections .
we analyzed the outer knots and hot spot similarly with 0.2 radius apertures ( 87@xmath4 encircled energy ) centered on the radio positions in the unprocessed image and found no net positive signal .
measurements of the poisson noise in these apertures result in a formal 3@xmath22 point source detection limit of 0.01 @xmath23jy .
however , experience with similarly configured stis clear imaging data @xcite allow us to more practically gauge detection limits for such jet knots we favor the more conservative 3@xmath22 limit of 0.06 @xmath23jy derived from the previously published work .
an extinction correction of 10@xmath4 was estimated using ned s quoted values from the @xcite maps and though basically negligible , has been applied to the optical measurements ( table [ table-2 ] ) .
the quasar 1745 + 624 was observed with the acis camera aboard the _ chandra _ x - ray observatory @xcite and we retrieved this data from the _ chandra _ archive . in this observation ,
the source was placed 30 from the default aim - point position on the acis - s backside illuminated ccd , chip s3 ( proposers observatory guide , pog ) . the data were collected in a full readout mode of 6 ccds resulting in 3.241 seconds readout time .
the quasar s count rate of 0.056 cts s@xmath7 gives 6 - 7@xmath4 pileup in the core for this choice of the readout mode .
after standard filtering , the effective exposure time for this observation was 18,312 sec ( table [ table-1 ] ) .
the x - ray data analysis was performed with the ciao 3.2 software using calibration files from the caldb3 database .
we ran acis_process_events to remove pixel randomization and to obtain the highest resolution image data .
the x - ray position of the quasar ( r.a.= 17 46 14.057 , decl.= + 62 26 54.79 ; j2000.0 equinox ) agrees with the radio position to better than 0.2@xmath24 , which is smaller than _ chandra _
s 90@xmath4 pointing accuracy of 0.6 @xcite .
the x - ray images were rebinned to 1/4 the native acis pixel - size of 0.492 in the final image ; the extended x - ray jet is already quite obvious in this image ( figure [ figure-5 ] ) . to determine the extent of the psf at the location of the quasar , we ran a ray - trace using the _ chandra _ ray tracer ( chart ) and then marx to create a high s / n simulation of a point source .
we modeled the quasar core as a point source with the energy spectrum given by the spectral fitting described below .
for the simulation , we added 7@xmath4 errors to account for uncertainty in the ray - trace model @xcite .
we then extracted a radial profile from both the _ chandra _ data and the simulated point source image assuming annuli separated by 0.45 and centered on the quasar .
the psf was normalized to match the peak surface brightness of the core . a comparison between the quasar profile and the simulated psf shows an excess of counts above the psf beginning at @xmath25 from the quasar centroid .
therefore , in our analysis of the quasar s core x - ray emission ( [ section - core ] ) , we define a 1.2 radius circular region centered on the x - ray centroid which excludes any overlap with the region defined for the extended jet . for our analysis of the extended emission ( [ section - xjet ] ) , we defined two pie - shaped regions based on the radio images
one for the jet , and another for the terminal hot spot .
spectral analysis was performed in _
sherpa _ @xcite using counts in the energy range 0.37 kev . based on the psf analysis described above , a 1.2 radius circular region encircles about 91@xmath4 of the total quasar counts .
a total of 1,132 counts are detected in this region corresponding to 1,107.6@xmath26 net counts .
we fit the spectrum between 0.3 and 7 kev with an absorbed power law model assuming the photoelectric absorption model with abundance tables from @xcite .
our analysis included the pileup model specified by @xcite , which gives the fraction of piled photons within 69@xmath4 .
the best - fit power law model has a photon index , @xmath27 and a total absorbing hydrogen column of @xmath28 cm@xmath20 for an absorber at @xmath0=0 ( @xmath30 dof , and we give 1@xmath22 errors for one significant parameter ) .
note that the fitted column is greater than the equivalent column within the galaxy ( 3.3@xmath31 cm@xmath20 ) in the direction of the quasar at high significance .
this model gives a 0.52 kev flux of 1.5@xmath32 erg cm@xmath20 s@xmath7 and 210 kev flux of 3.0@xmath33 erg cm@xmath20 s@xmath7 for the core .
if we now assume that the quasar spectrum is absorbed by two absorption components , a known ( and fixed ) galactic column at @xmath0=0 and an intrinsic absorber located at the quasar redshift , @xmath0=3.89 , we obtain an intrinsic @xmath34 cm@xmath20 and the best - fit photon index of @xmath35 ( @xmath36=78.5/82 dof ) ; see figure [ figure-6 ] .
this large intrinsic absorption column confirms the previous _ asca _ result from analysis of the total emission @xcite and is in agreement with the average column observed in samples of high - redshift radio - loud quasars @xcite .
these are comparable to the absorptions observed in broad absorption line ( bal ) quasars @xcite , where it is usually associated with an outflow .
the photon index is also typical of that observed in lower redshift quasars with x - ray jet detections ( e.g. , * ? ? ?
* ) . in this model ,
the 0.52 kev and 210 kev fluxes are equal to 1.6@xmath37 erg cm@xmath20 s@xmath7 and 3.4@xmath38 erg cm@xmath20 s@xmath7 , respectively ; these correspond to quasar x - ray luminosities , @xmath39(0.52 kev ) = 1.6@xmath40 erg s@xmath7 and @xmath39(210 kev)=5.7@xmath40 erg s@xmath7 ( unabsorbed and k - corrected ) .
we determined that there is no fe - line present in the quasar spectrum with a 3@xmath22 upper limit on the equivalent width of 210 ev , consistent with previous _
results @xcite .
the _ chandra _
data are also consistent with little or no variability in both the x - ray flux and spectrum since previous observations of ( the total emission from ) the quasar @xcite .
we extracted the spectrum of the total extended x - ray emission ( from the jet and hot spot ) assuming a pie - shaped region between 1 to 2.75 centered on the quasar core .
the assumed background region was equal to a pie region spanning the same radii , but excluding the jet+hs region .
a total of 241 counts corresponding to 224.5@xmath41 net counts was detected in this region .
the binned spectrum was fit with an absorbed power law model , and the best fit photon index for a fixed galactic column is equal to @xmath42 ( @xmath36=17.9/19 dof ) .
this gives a 0.52 kev flux of 3.1@xmath43 erg cm@xmath20 s@xmath7 , and 210 kev flux of 6.1@xmath44 erg cm@xmath20 s@xmath7 ( figure [ figure-6 ] ) .
since we found that the radio hot spot spectrum is significantly steeper than that of the jet ( [ section - spectra ] ) , we attempted to analyze the jet and hot spot emissions separately , assuming pie - shaped regions between 1 to 2.05 and 2.05 to 2.75 , respectively , centered on the quasar core .
the background region was equal to the annulus with the same radii excluding the jet and hot spot regions .
a total of 181 counts were detected in the jet region which result in 173.2@xmath45 net counts .
we binned the spectrum to have a minimum of s / n=3 and fit an absorbed power law model to the jet spectrum within the 0.37 kev energy range assuming the galactic equivalent hydrogen column of 3.3@xmath31 cm@xmath20 .
we obtain the best fit photon index @xmath46 ( @xmath36= 10.1/14 dof ) corresponding to a 0.5 - 2 kev flux of 2.2@xmath47 erg cm@xmath20 s@xmath7 and 2 - 10 kev flux of 5.1@xmath48 erg cm@xmath20 s@xmath7 .
similarly , for the hot spot ( net [email protected] counts only ) , we found @xmath49 ( 1@xmath22 ) and fluxes 6.9@xmath50 erg cm@xmath20 s@xmath7 ( 0.52 kev ) and 7.9@xmath50 erg cm@xmath20 s@xmath7 ( 210 kev ) .
these results suggest that , as in the radio band , the hot spot s x - ray spectrum is steeper than the jet , although the formal difference is not well determined because of the large uncertainty in the hot spot spectrum ( @[email protected] ) .
the model results from this paragraph are the ones used in the subsequent discussion .
we also allowed the total absorption column of the jet emission to vary and obtained a slightly steeper photon index @xmath52 and a 3@xmath22 upper limit to the absorption column of @xmath53 cm@xmath20 .
this fitted photon index is consistent within 1@xmath22 of the value found with the column fixed .
we thus conclude that there is little gained by allowing the column to vary . in the _ chandra _ image , the x - ray intensity ratio of the knots k1.4
to k1.8 is @xmath130.6:0.4 . assuming that the x - ray spectrum does not change drastically between these two regions ( there is no significant difference in radio spectra ; [ section - spectra ] ) , we estimated their fluxes using their respective ratios of the total flux ( table [ table-2 ] ) .
taking our flux measurements of the hot spot at the four radio frequencies , we fit a spectral index of @[email protected] ( @xmath55 ) , compared to @xmath54=1.45 measured by @xcite from mostly the same data ( tables [ table-1 ] & [ table-2 ] ; figure [ figure-7 ] ) . in comparison to the individual jet knot measurements ( below )
, we believe this result is more robust because the hot spot is bright , compact , and well - detected at the observed frequencies .
the fitted spectral index is insensitive to the presented low resolution 1.5 ghz measurement ( @[email protected] omitting this measurement in the fit ) .
this 1.5 ghz measurement though may be contaminated by emission from the adjacent k1.8 ( figure [ figure-2 ] ) .
since the flux ratio of k1.8 to hs at the other bands is about 1:3 , reducing the 1.5 ghz flux of the hot spot by 25@xmath4 will change the spectrum to [email protected] which is still within the uncertainty .
the jet knots k1.4 and k1.8 are well - separated from the bright nucleus in the 5 , 8.5 and 15 ghz images .
these are the first reported detections of the _ jet _ at frequencies other than 5 ghz @xcite .
we measured a `` total '' jet spectral index of [email protected] between these 3 frequencies on the maps reconvolved to a common 0.35 beam ( [ section - radio ] ) .
spectral indices measured from the full - resolution images give larger uncertainties ( table [ table-2 ] ) because the signal - to - noise ratio of the individual jet knot detections are low . including the `` super - resolved '' 1.5 ghz image , the resultant 4-frequency radio spectral index is [email protected] , which we adopt as the best determined spectral index of the jet ( table [ table-2 ] ; figure [ figure-7 ] ) .
the overall impression is that the jet knot radio spectra are fairly steep on the steep end of what is observed in more nearby examples @xcite . as indicated by previous work @xcite ,
the hot spot radio spectrum is significantly steeper than the jet ( @xmath56-@xmath57=0.35 ) and this is also suggested to be the case in the x - ray data ( [ section - xjet ] ) .
the slope of the hot spot spectrum is as steep as the kpc - scale feature detected in gb 1508 + 5714 @xcite and is similar to those measured in comparably high - redshift radio galaxies ( e.g. , * ? ? ? * and references therein ) .
these steep radio spectra are probably due to a combination of the increased importance of ic / cmb losses on the higher energy electron population and higher @xmath58 rest - frame frequencies sampled by the observations this effect may be important when comparing low and high - redshift jets ( [ section - z ] ) .
in addition to the radio spectra , the optical upper limits of all the extended features preclude a simple power - law extrapolation of the radio data into the x - ray band ( i.e. , @xmath59 and @xmath60 ; figure [ figure-7 ] ) .
this behavior is common among the x - ray detected jets in quasars ( e.g. , * ? ? ?
one should bear in mind that this assumes that the radio - to - x - ray radiation all originate in the same emitting region , and also , the optical limits assumed a point source . though the radio knots are unresolved at @xmath130.150.2resolution ( [ section - radio ] ) , it does not preclude the optical emission being more extended or that there are multiple emitting components .
although there are complicating factors in the measurements of the radio and x - ray spectra as discussed above , our analysis suggests that the power - law slopes measured in the two bands are different for the jet .
formally , this difference , @[email protected] , is significant .
no similar evidence is evident in the hot spot spectra , as we are limited by the low statistics of its x - ray detection .
at @xmath0=3.89 , the 1745 + 624 system displays a fully formed jet as observed by us when the universe was only about 12@xmath4 of its present age . for its projected length @xmath62 kpc , and assuming a constant hot spot advance velocity @xmath63 ( see in this context * ? ? ? * but also comments at the end of [ section - hotspot ] ) , the lifetime of the radio source is roughly @xmath64 myr ( the jet viewing angle @xmath65 is taken here for illustration ; see below ) .
this structure signals long - term active accretion at a very early cosmic epoch , along with a roughly stable black hole / accretion disk axis ( e.g. , * ? ? ?
such large - scale jets ( deprojected @xmath13100 kpc in this case ) at high - redshifts are not necessarily rare , though they have not been systematically searched for and studied .
* and manuscript in preparation ) has only recently identified @xmath1310 other jets on a comparable scale in a vla study of a sample of @xmath13.4 ( out to @xmath0=4.7 ) flat - spectrum - core radio sources ; 1745 + 624 remains the most radio - luminous example observed ( [ section - intro ] ) . despite the extreme luminosity of the jet in 1745 + 624 ,
its extended radio and x - ray emissions appear quite similar to other more local examples in regard to its morphology , spectral energy distribution , and radio polarization .
this is quite surprising as higher-@xmath0 jets have propagated through the ( supposedly different ) intergalactic medium of the early universe ( e.g. , * ? ? ?
in the following , we first consider available evidence for relativistic beaming in the 1745 + 624 jet on both parsec and kpc - scales ( [ section - describe ] ) .
then , we discuss the emission properties of the kpc - scale jet ( [ section - discussjet ] ) and the hot spot ( [ section - hotspot ] ) , confronting the multiwavelength data with expectations from different emission models . along the way , we make comparisons to other x - ray jets at low and high redshift .
finally , the broad - band data allow us to estimate the energetics of the jet , which we compare to the accretion parameters of the quasar nucleus ( [ section - energetic ] ) .
several lines of evidence suggest that relativistic beaming in the 1745 + 624 jet is significant on the observed scales .
both the radio ( vlbi and vla - scale ) and x - ray structures appear one - sided to our detection limits .
the best constraints on the sidedness comes from the presented 2.3 ghz usno map ( jet to counter - jet ratio @xmath1450 ; [ section - vlbi ] ) and proper motion studies with vlbi .
twelve epochs of higher resolution 8 ghz data from the same usno program , obtained over @xmath132.5 years @xcite , show a basically stationary inner ( @xmath1311.5 mas distant ) component , although velocities of a few @xmath66 are allowed by the formal fit ( b.g .
piner , 2005 , private communication ) .
maps from the caltech - jodrell bank survey obtained over a longer time baseline at 5 ghz reveal three superluminal components presumably further down the jet ( * ? ? ?
* ; * ? ? ?
* this is the highest redshift object in the sample see figure 1 in the latter ) . converting to our adopted cosmology , their proper motions correspond to apparent speeds ranging from @xmath136@xmath66 up to @xmath1316@xmath66 .
this restricts the maximum angle of the small - scale jet to our line of sight to be @xmath67719 .
the superluminal vlbi jet is oriented at pa@xmath13205208 in the sky @xcite , increasing gradually to @xmath13216 out to @xmath13110 mas ( figure [ figure-1 ] ; [ section - vlbi ] ) .
this further increases out to the visible arcsecond - scale structure ( the hot spot pa=227 ) .
the maximal ( projected ) changes in position angle between the different features is 6 ( cf .
table [ table-2 ] ) .
the required intrinsic bend in the jet to account for these changing pas is @xmath136 @xmath68@xmath132 , or even less ( supposing the jet inclination @xmath6720 as indicated by the proper motions ) .
since we can trace the vlbi - scale emission out to the observed _ chandra _ scales , the vlbi proper motions ( and only small intrinsic bends ) constrain the kiloparsec - scale jet to be aligned within @xmath6925 to our line - of - sight , with a likely range of @xmath131020 , or less .
radio asymmetry studies of large scale quasar jets require jets like this to be relativistic ( @xmath70@xmath141 , with strict upper limits of @xmath70@xmath6723 ; * ? ? ?
* ) ; these constraints have bearing on the following discussion of the origin of the x - ray emission in the 1745 + 624 jet .
the absence of optical emission and the steep radio spectra of the jet knots in 1745 + 624 preclude a straightforward power - law extrapolation of the low energy ( synchrotron ) emission into the x - ray band ( [ section - spectra ] ) .
the resultant convex broad - band spectral energy distribution ( sed ; figure [ figure-7 ] ) is a common characteristic of x - ray emitting knots in the jets of powerful quasars ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
as widely discussed , ( unbeamed ) inverse compton emission also can not account for the observed excess of x - rays , and it is evident that this applies also to the present example at very high - redshift ( [ section - basic ] ) .
we explore the hypothesis that the bright x - ray emission is produced by _ relativistically beamed _
inverse - compton scattering of the cmb photons @xcite as advocated by many workers .
the main feature of this model is that in order to fulfill energy equipartition , significant beaming of the large - scale quasar jet emission is required ( [ section - beamed ] ) .
`` non - standard '' synchrotron models have also been proposed to reproduce the overall convex radio - to - x - ray seds and we discuss this possibility in [ section - synchrotron ] .
the main difference between these models is the expected diverging redshift - dependence of the x - ray emission , and we discuss issues and prospects for distinguishing such a dependence from the observations ( [ section - z ] ) . generally , one can calculate the magnetic field in a synchrotron emitting radio source if we assume magnetic field energy equipartition with the radiating particles ( corresponding to the minimum total energy of the system ) . in the 1745 + 624 radio jet , we use the observed radio properties of k1.4 and k1.8 ( assuming negligible contribution from relativistic protons and no beaming , @xmath71=1 ) to calculate fairly high values of @xmath72 @xmath23 g and @xmath73 @xmath23 g , respectively .
these values are more than an order of magnitude higher ( due to the exceptionally luminous nature of this radio jet ; [ section - intro ] ) than those computed in lower - redshift quasar jets @xcite , but are comparable to values inferred for lower power fri radio galaxies @xcite . the corresponding magnetic field energy densities of the knots , @xmath74 erg@xmath75 , are @xmath1367 times greater than the cmb energy density at the quasar s redshift , @xmath76 erg@xmath75 @xmath77 erg@xmath75 .
the latter quantity is , however , comparable to the energy densities of the knots synchrotron photons ( in the case of sub - relativistic jet velocity ) , @xmath78 namely , @xmath79 2.8 and 3.8 @xmath80 erg@xmath75 for k1.4 and k1.8 , respectively . in this estimate
, we approximated the knots as spheres with radii , @xmath81 kpc ( [ section - radio ] ) , and took the synchrotron luminosities , @xmath82 ergs@xmath7 , as appropriate for the radio continuum extending over a few decades in frequency with spectral index @xmath83 .
for the spectral index @xmath83 , the expected x - ray flux densities ( @xmath84 ) from the synchrotron self - compton ( ssc ) plus ic / cmb processes are simply : @xmath85 using the 5 ghz radio flux measurements ( @xmath86 ) , this severely underpredicts ( @xmath130.054 njy ( k1.4 ) and @xmath130.074 njy ( k1.8 ) at 1 kev ) the observed x - ray jet emission by factors of 87 and 42 in knots k1.4 and k1.8 , respectively ( table [ table-2 ] ) .
the bulk velocity of kpc - scale quasar jets are likely to be at least mildly relativistic ( @xmath87 , * ? ? ?
* ; * ? ? ?
this can significantly influence the expected inverse compton x - ray fluxes calculated above . specifically , relativistic beaming ( @xmath88 ) increases the ic / cmb emission , and decreases the ssc one . in 1745 + 624 ,
the superluminal vlbi jet can be traced out to kpc - scales ( [ section - describe ] ) , so the beamed ic / cmb x - ray emission will dominate over that produced by ssc in the visible jet . assuming energy equipartition as an additional constraint , the expected ic / cmb flux is ( for @xmath891 ) : @xmath90 since in the jet rest frame ( denoted by primes ) @xmath91 , and the equipartition magnetic field @xmath92 . here
, @xmath70 is the jet bulk lorentz factor , and @xmath93 .
this gives the required @xmath94 and 3.9 for the knots k1.4 and k1.8 , respectively , to explain the observed x - rays via inverse compton emission .
in this interpretation , the ( small ) decrease of the required doppler factor along the jet is a direct consequence of the fading x - rays with the corresponding brightening of the radio emission ( figure [ figure-5 ] ; cf .
similar multi - wavelength morphologies in other well - studied jets like , 3c 273 @xcite , 0827 + 243 @xcite , pks 1127145 @xcite , and pks 1136135 @xcite ) .
although there is evidence that the 1745 + 624 jet is highly relativistic on parsec scales ( [ section - describe ] ) , there are few independent constraints ( the one - sidedness of the radio and x - ray jets ) to directly support the fairly high doppler factors on these large scales , as implied in this interpretation .
the inferred doppler factors require minimum lorentz factors , @xmath95 = 2.4 ( k1.4 ) and 2.1 ( k1.8 ) .
these approach the strict upper limits on the speeds of large - scale quasar jets deduced from radio asymmetry studies ( @xmath70@xmath6723 ; * ? ? ?
if we assume the 1745 + 624 jet flow is moving near these maximal values , the derived doppler factors require that the jet is inclined within @xmath691215 to our line of sight , in agreement with the values inferred in [ section - describe ] .
given the discussed lorentz factors of @xmath1323 , the doppler factor is fairly insensitive to jet angles considered ; for instance , @xmath71=5.53.6 ( @xmath70=3 ) and @xmath71=4.03.2 ( @xmath70=2.1 ) for jet angles , @xmath96=015 .
thus we can _ just _ plausibly explain the level of x - ray jet emission in this jet via the ic / cmb mechanism if we push the speeds to be consistent with upper limits inferred from previous studies , together with the small viewing angles .
it should be stressed that these estimates assumed a single - zone emitting region , and there are generally more free parameters in both compton and synchrotron models invoking substructure in the jet ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
related to the beaming issue , we should remark that the _ observed _ radio power of the 1745 + 624 jet ( excluding the hot spot region ) is very large , @xmath97 ergs@xmath7 , suggesting its synchrotron luminosity of order @xmath98 ergs@xmath7 or greater .
such enormous radio powers from _ large - scale jets _ ( exclusive from extended radio lobes ) are quite extreme , so invoking even moderate beaming factors would reduce the intrinsic jet frame luminosity ( by 12 orders of magnitude ) to more comfortable levels .
this relativistically beamed ic / cmb model also requires that there is a sufficient number density of low energy electrons to upscatter cmb photons into the observed _ chandra _ band .
radio emission would be radiated by these electrons at very low - frequencies ( 10 s mhz in this case ) and subarcsecond - resolution radio imaging capabilities at these energies are not widely available .
curiously , @xcite noticed that the _ integrated _ radio spectrum of 1745 + 624 contains a steep - spectrum excess at low - frequencies ; this quasar would be catalogued as a steep - spectrum radio quasar in low - frequency surveys and/or if it were local ( the @xmath99 shift in observed frequencies ) . in our own compilation of radio data from the literature ( figure [ figure-7 ] ) ,
we find @[email protected] from 38 mhz to 1.4 ghz and @xmath54@xmath130 at the higher frequencies where our imaging observations were taken .
if we extrapolate the observed individual spectra of the ( dominant @xmath54@xmath130 ) core and steep - spectrum extended component to lower frequencies , it implies that the latter dominates the total low - frequency emission .
the well - studied @xmath0=0.158 quasar 3c 273 shows a similar low - frequency excess apparent down to @xmath1310 s mhz @xcite and resolved low - frequency imaging shows its extended x - ray emitting ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) radio jet does indeed dominate the total flux at low frequencies ( cf .
figure a1 in * ? ? ?
further , although this low - frequency emission in 1745 + 624 is probably dominated by the ( brighter and steeper spectrum at cm - wavelengths ; [ section - hotspot ] ) hot spot , this does not preclude a low - frequency contribution from the jet .
interestingly , the x - ray spectral index of the jet ( table [ table-2 ] ) matches the low - frequency integrated radio one , as would be expected in the ic / cmb scenario .
future low - frequency imaging observations can clarify this by showing the relative contributions of the jet and hot spot in this and similar systems ( see * ? ? ?
* for future prospects ) . in lieu of the beamed ic / cmb interpretation ,
synchrotron models with non - standard and/or multiple electron components are able to account for the `` excess '' x - ray emission @xcite .
these models are attractive because in nearby low - power fri jets , it is widely believed that synchrotron x - ray emission is produced ( e.g. , * ? ? ?
* ) , although they do not show the same convex seds , and emit at much smaller ( few kpc ) scales than the discussed quasar jets ( 10s100 s kpc scales ) .
the challenge remains to explain the strong upturn in the x - ray band ( i.e. , the small values of @xmath100 ) observed in powerful jets such as in 1745 + 624 . while it is true that an increased energy density of the cmb photons at @xmath101 will result in a stronger radiative cooling of the ultra - relativistic electrons , the maximum electron energy available assuming efficient ( although realistic ) continuous acceleration process is still high enough to allow for production of an appreciable level of kev synchrotron photons . in the case of the 1745 + 624 jet , the comoving energy density of the magnetic field ( by assumption close to equipartition with the radiating electrons , as given above )
is lower than the comoving energy density of the cmb photons , @xmath102 , if only @xmath103 .
thus , assuming even very moderate bulk velocity and beaming , the jet electrons cool mainly by inverse - comptonization of the cmb photon field . in
the equipartition derived magnetic field of @xmath104=200 @xmath105 , the electrons emitting the highest energy photons detected ( @xmath136 kev , observed ) have electron energies , @xmath106 . since their cooling are dominated by inverse compton losses ,
the appropriate timescale can be evaluated roughly as @xmath107 yrs .
this evaluation is rough , because in fact the considered high energy electrons are expected to radiate in a transition between thomson and klein - nishina cooling regimes , since @xmath108 with the energy of the cmb photon @xmath109 milli - ev , and the jet bulk lorentz factor @xmath70 of the order of a few ( see in this context * ? ? ?
* ) . for such parameters and the considered electron energies , the ` optimistic ' electron acceleration timescale @xmath110 yr
, where @xmath111 is the electron gyroradius and @xmath112 is the efficiency factor depending on the jet plasma conditions ( see * ? ? ?
this is still much shorter than the timescale for radiative losses , while the latter is order of magnitudes longer than the timescale for the electron escape from the emission region .
the postulated stochastic acceleration mechanism can produce a flat - spectrum high - energy electron component , with the total energy density not exceeding the magnetic field energy density ( see * ? ? ?
* for a discussion ) .
the _ observed _ synchrotron spectrum of such particles ( from the unresolved emitting region ) is expected to show a spectral index modified ( increased ) by the radiative losses . with the condition of continuous injection of freshly accelerated electrons ( characterized by a power - law energy spectrum @xmath113 ) ,
such a spectral index would be @xmath114 for any electron index @xmath115 ( see * ? ? ?
* ) , which is very close to the observed x - ray spectral index ( @xmath116 ) of the 1745 + 624 jet . moreover , the total luminosity of such a high - energy synchrotron component , if indeed limited only by the energy equipartition requirement , should be comparable to or less than the luminosity of the low - energy ( radio ) synchrotron component .
that is in fact the case , since the _ total _ @xmath117 kev luminosity of the 1745 + 624 jet ( excluding the terminal feature discussed in [ section - hotspot ] ) is @xmath118 erg s@xmath7 while , as we argue in [ section - basic ] above , the radio jet luminosity is @xmath119 erg s@xmath7 .
note , that in the framework of this synchrotron interpretation , the x - ray - to - radio jet luminosity ratio is not expected to change systematically with redshift in contrast to expectations in an ic / cmb interpretation , and this issue is discussed in the next subsection .
the simplest versions of synchrotron and ic / cmb models give significantly different predictions for the x - ray jet emission at high redshift because of the strong increase of the cmb energy density @xmath120 ( e.g. , * ? ? ?
with all else equal , they predict divergences in the observed x - ray to radio monochromatic luminosity ( @xmath121 ) ratio , @xmath122/@xmath123 exceeding a factor of 100 or more at @xmath0@xmath142 ( cf . equation [ equation - iccmb ] ) .
this difference should be readily apparent in observations of @xmath1244 jets like in 1745 + 624 if the x - rays were dominated by ic / cmb emission .
x - ray jets detected so far have observed @xmath122/@xmath123 values which range over 4 orders of magnitude and there is no obvious sign of a @xmath58-dependence in the latest large compilation of @xcite . however , there are complicating factors which will skew such correlations , in both models , such as source - to - source ( and knot - to - knot within a single jet ) variations in jet magnetic fields , electron energy densities , and speeds . particularly in current samples biased toward `` blazars '' , the x - ray jet emission is particularly sensitive to the ( very small ) jet angles to our line of sight if there is significant relativistic beaming on these large scales . for an ic / cmb origin , (
figure 4 therein ) outlined such a scenario to account for the large @xmath125 ratio of the @xmath0=4.3 gb 1508 + 5714 quasar jet ( @xmath14 100 ; among the largest value found thus far ) in comparison to other known , lower redshift ( @xmath162 ) jets due mostly to its extreme redshift can be equivalently expressed as the radio - to - x - ray power - law slope ; @xmath126=0.73 for gb 1508 + 5714 @xcite while @xmath127 , translating to @xmath128 at lower redshifts ( see figures 5 and 9 in * ? ?
* ; * ? ? ?
* respectively).[footnote - x ] ] .
this scenario can be extended to the @xmath0=3.63 core - dominated quasar j22192719 where 6 total extended x - ray counts were recently detected in an @xmath138 ksec exposure @xcite , and can be attributed to a @xmath131 mjy radio feature 2 south of the quasar ( thus @xmath129 also ; * ? ? ? * and manuscript in preparation ) .
two quasars showing particularly low @xmath130 ratios , despite their fairly high - redshifts ( @xmath0=1.4 and 2 ) also stood out
these are especially lobe - dominated quasars in comparison to the other examples , and a connection to beaming was suggested ( see * ? ? ?
* for a discussion ) .
one caveat to note is that the doppler factors required for the knots in the 1745 + 624 jet ( in the minimum - energy + ic / cmb framework ) are rather moderate , @xmath131 , and similar to that found in the gb 1508 + 5714 case ( * ? ? ?
* [ section - intro ] ) .
the required @xmath71 in these two @xmath1244 quasar jets are at the low end of the corresponding values calculated for lower - redshift objects ( @xmath132425 ) .
in fact , these few critical data points at high-@xmath0 help to define the trend discussed by ( * ? ? ?
* therein ) that the maximal values of @xmath71 required to explain the x - ray jet emission as ic / cmb _ decreases with redshift_. also , in contrast to the gb 1508 + 5714 and j22192719 cases , the @xmath122/@xmath123 values of the 1745 + 624 jet are closer to those observed at lower - redshifts ( table [ table-2 ] and footnote [ footnote - x ] ) .
although second - order effects from the intrinsic jet properties should be factored in , the low implied jet doppler factors and lack of a very obvious redshift - dependence of @xmath133/@xmath123 , can be taken as evidence against an ic / cmb interpretation .
in particular , if there were more extreme doppler factors than in the @xmath0@xmath134 quasars observed so far , the large - scale jets would be orders of magnitude brighter and will outshine the active cores in x - rays via the ic / cmb process @xcite .
currently however , no such sources have been found in a growing number of high-@xmath0 radio - loud quasars imaged with subarcsecond - resolution by _
@xcite , so where are the very highly beamed high-@xmath0 x - ray jets ?
there is of course a natural malmquist bias inherent in studying very high - redshift objects , so this and other selection biases must be addressed in future samples . on the other hand ,
if these trends persist with further observations , it may instead reflect intrinsic differences between large - scale quasar jets located at different redshifts and/or their different environments , which are then probed with observations .
for example , in the ic / cmb interpretation , the trend may imply that high-@xmath0 jets of this kind are intrinsically slower than their lower - redshift counterparts .
this may be due to a more disturbed environment ( evidenced by the `` alignment effect '' of high-@xmath0 radio galaxies ; e.g. , rees 1989 ) through which high-@xmath0 jets are propagating .
such an scenario , in fact , was suggested early on to explain the distorted morphologies observed in a large samples of high-@xmath0 ( @xmath131.53 ) quasars and radio galaxies ( e.g. , * ? ? ?
* ) . as current x - ray studies have tended toward known well - studied radio jets ( e.g. , *
* ; * ? ? ?
* ) that are relatively local , it is important to test if this and other trends persist and can be clarified with larger , more homogenous samples of distant ( @xmath124 ) jets . in analogy
to vlbi proper motion studies , the earliest superluminal motions detected tended to be the highest @xcite and larger ensembles of source measurements now tend toward lower values @xcite .
this may mean that beaming factors of kpc - scale x - ray jets as inferred from ic / cmb calculations are more typically smaller than the highest ones found thus far .
recent _ chandra _ studies have aimed to distinguish the x - ray detected terminal jet features ( i.e. , the hot spots ) from the jets @xcite .
many of these knots show much the same problem as in the jet , i.e. the inability to extrapolate the radio - to - optical power - law slopes smoothly in the x - ray band , and the underprediction of not - beamed inverse - compton ( ssc and ic / cmb ) emission to the observed . in this context
, it is useful to discuss the terminal feature k2.5 of the 1745 + 624 jet as a terminal hot spot .
knot k2.5 follows the conventional empirical definition of a hot spot only loosely @xcite , as applying these strict criteria to such high - redshift , small angular - size jets is quite restrictive .
hot spots are strong terminal shocks formed at the tips of powerful jets where the strong compression of the jet magnetic field leads to a transverse configuration and relatively high radio polarization .
this expectation is in good agreement with observations of sources located at low redshifts .
k2.5 is situated at the edge of the radio source and it is indeed compact with a spatial extent of @xmath134 kpc ( [ section - radio ] ) , which is typical for the known x - ray detected hot spots in quasars and fr ii radio galaxies .
however , the parallel configuration of the magnetic field with respect to the jet axis , as well its low degree of linear polarization relative to the rest of the jet ( figure [ figure-3 ] ) , are not expected , nor typical of the hot spots resolved in more nearby radio sources @xcite .
_ this may betray the dominance of an underlying jet flow in this feature .
_ the 5 ghz flux of the hot spot in 1745 + 624 implies a monochromatic luminosity , @xmath135 ergs@xmath7 .
this is higher , although still comparable to the 5 ghz luminosities of hot spots at lower redshifts ( e.g. , * ? ? ?
assuming energy equipartition as in the jet , and a cylindrical geometry for the emitting region with radius @xmath136 kpc ( [ section - radio ] ) and similar length , we calculate the hot spot magnetic field , @xmath137 @xmath23 g .
this also is in rough agreement ( though again slightly higher ) with the minimum energy magnetic fields derived for the other knots @xcite . the equipartition magnetic energy density of the hot spot , @xmath138 erg@xmath75 , is more than one order of magnitude greater than the energy density of the cmb radiation at the quasar s redshift
however , the energy density of the synchrotron photons thereby , @xmath139 erg@xmath75 , is roughly comparable to @xmath140 . in this evaluation
, we took the hot spot radio spectral index , @xmath141 , its synchrotron luminosity @xmath142 , the geometry as described above , and neglected relativistic corrections .
the implied 1 kev flux of the hot spot due to the ssc plus ic / cmb processes is then @xmath130.18 njy ( equation [ equation - ssc ] ) . as in the jet emission ( [ section - basic ] ) ,
this is inconsistent with the observed x - ray flux , though here the deviation is less drastic ( factor of @xmath1314 ; table [ table-2 ] ) .
however , one should keep in mind many of the underlying assumptions in such calculations before claiming a true disagreement between model predictions and the observations .
in particular , the hot spot structure is in fact unresolved , hence a slightly smaller volume than considered , together with , e.g. small deviations from energy equipartition and spectral curvature ( see below ) , could foreseeably bring the measurement into better agreement with the model prediction
. however , if we take the observed excess in x - rays over the estimated ssc flux at face value , this would not be the first such case for a hot spot .
@xcite has argued for a synchrotron origin of this x - ray excess in many local hot spot sources , though in the case of 1745 + 624 , the optical upper limit precludes a straightforward extrapolation of the ( very steep ) radio spectrum up to the x - ray band . on the other hand , @xcite has argued that there is an ic / cmb contribution from an underlying relativistic portion of the jet terminus ( unresolved by our observations ) to the few x - ray detected hot spots studied in several @xmath143 quasars .
the latter possibility is quite attractive in the case of 1745 + 624 : as we noted earlier , the terminal jet feature in this source does not have typical hot spot - like radio polarization properties , and there could very well be an unresolved portion of the jet mixed in .
if the multi - band radiative output of k2.5 is dominated by a relativistic jet flow , this would imply @xmath71=5 for an ic / cmb origin of the x - ray flux ( using equation [ equation - iccmb ] with all the parameters discussed above ) .
this beaming factor is quite high , though comparable to values derived in lower - redshift quasars @xcite .
it is also slightly higher than the values obtained for the proper jet ( knots k1.4 and k1.8 ; see [ section - beamed ] ) , though again , there are the usual uncertainties in these calculations to keep in mind .
regardless , the problem of terminal x - ray / radio knots in this and many quasar jets can be viewed as a simple extension of that in the jet knots . one relevant issue in this discussion is the apparent lack of a hot spot on the side opposite of the nucleus to the prominent jet .
if the advance speed of the jet is sub - relativistic , then some emission should have been visible on the counter - jet side .
this can be reconciled by the fact that the usually inferred ( sub - relativistic ) advance speeds of hot spots in powerful radio sources assume that the jet thrust is balanced with the ram - pressure of the ambient material within a typical galaxy group / poor cluster environment @xcite . however
, if the jets propagate into the radio cocoon formed in a previous epoch of radio activity , i.e. in an environment with a substantially decreased number density and pressure , the advance velocity of the jet may be close to @xmath66 .
in such a case , combined doppler and light - travel effects may result in an apparent lack of a visible counter hot spot .
such a scenario was proposed by @xcite for the 3c 273 jet .
since the observed properties of 1745 + 624 and 3c 273 radio quasars are so similar , one could also apply this argument to explain the lack of emission visible on the counter - jet side in 1745 + 624 .
let us finally discuss yet another issue regarding this dominant radio feature , namely the extrapolation of its radio spectrum to lower frequencies . as discussed in
[ section - beamed ] , an extrapolation of its observed @xmath13@xmath144 ghz continuum @xmath145 to lower frequencies exceeds the extrapolation of the flat - spectrum core component @xmath146 at about @xmath147 mhz , and joins smoothly with the observed integrated low frequency emission at @xmath148 mhz ( see figure [ figure-7 ] [ left ] ) . in other words
, one can attribute the @xmath1330300 mhz integrated emission to the extended radio structure . since the extended emission
is probably dominated by the brightest feature , the hot spot ( we can not exclude a contribution from the fainter inner jet ; [ section - beamed ] ) , this would imply a broken power - law character of its synchrotron spectrum , @xmath149 for @xmath150 mhz and @xmath145 for @xmath151 mhz . in the standard continuous - injection model of terminal features in powerful jets
@xcite , the spectral break would correspond to the energies , @xmath152 .
this low - frequency break is expected in face of the large computed hot spot equipartition magnetic field , @xmath137 @xmath23 g ( see a discussion in * ? ? ?
* ; * ? ? ?
the multiwavelength emissions of the 1745 + 624 jet allow us to estimate the kinetic power of the outflowing plasma in this source .
the kinetic power of the ultrarelativistic electrons and magnetic field characterizing the proper jet , under the minimum - power hypothesis , is roughly @xmath153 , where @xmath154 kpc is the jet radius and @xmath155 erg s@xmath7 is the jet comoving magnetic field energy density , as discussed in [ section - discussjet ] .
assuming for illustration , a jet viewing angle @xmath156 ( see [ section - describe ] ) and jet bulk lorentz factor at the kpc - scale @xmath157 ( leading to the ` comfortable ' value of the jet doppler factor @xmath158 ) , one obtains a kinematic factor @xmath159 , and therefore a jet kinetic luminosity ( regarding solely ultrarelativistic jet electrons and the jet magnetic field ) @xmath160 erg s@xmath7 .
if we regard the terminal feature as a `` true '' hot spot , we can estimate the kinetic energy of the jet .
we approximate the broad - band synchrotron spectrum of the hot spot by a broken power - law : @xmath149 between the assumed minimum synchrotron frequency @xmath161 mhz and the break frequency @xmath162 mhz , followed by @xmath163 between @xmath164 and the assumed maximum synchrotron frequency @xmath165 hz ( see [ section - hotspot ] ) . the radiative efficiency factor ,
i.e. the ratio of the power emitted by the hot spot electrons via synchrotron radiation , and the total power injected to these electrons at the terminal shock , is : @xmath166 the reconstructed radio continuum of the hot spot implies also its total synchrotron luminosity @xmath167 erg s@xmath7 .
thus , the total kinetic power of the jet transported from the active center to the jet terminal point is roughly : @xmath168 where @xmath169 is the fraction of the jet kinetic energy transformed in the terminal shock to the internal energy of ultrarelativistic ( i.e. synchrotron emitting ) electrons . in the case of energy equipartition between the electrons and the magnetic field
reached at the hot spot , one should expect very roughly @xmath170 ( see * ? ? ? * for a wider discussion ) .
this gives @xmath171 erg s@xmath7 .
in fact , @xmath172 should be regarded as an upper limit , keeping in mind a possible proton contribution and work done by the outflow in pushing out the ambient medium , making @xmath3 erg s@xmath7 a safe lower limit for the kinetic power of the 1745 + 624 jet .
the jet power estimated at the beginning of this subsection is an order of magnitude lower than its total kinetic power constrained from the hot spot emission .
this may indicate a dynamical role of non - radiating jet particles ( either cold electrons or protons ; see in this context , * ? ? ?
let us compare the kinetic power of the 1745 + 624 jet estimated above with the accretion parameters of the quasar core .
due to the large distance of the quasar , and hence related observational difficulties , these accretion parameters can not be evaluated precisely thus should be treated with some caution .
nevertheless , we note @xcite and @xcite found a broad emission civ line in the spectrum of the 1745 + 624 quasar at the observed wavelength @xmath173 , and an observed @xmath174 .
this gives the intrinsic velocity dispersion : @xmath175 @xcite reported also a steep - spectrum ( @xmath176 ) , non - variable and low - polarized optical continuum of the 1745 + 624 quasar , which we consider below as being produced by the circumnuclear gas solely .
in other words , we assume that the host galaxy and the nuclear portion of the jet do not contribute significantly to the observed optical emission of the discussed object . at the observed wavelength @xmath177 ,
corresponding to the source rest frame wavelength @xmath178 , the detected continuum flux is @xmath179 mjy .
thus , the intrinsic continuum luminosity at the emitted frequency @xmath178 , is simply @xmath180_{6600 } \approx 5.4 \times 10^{46}$ ] erg s@xmath7 .
these estimates allow us to find the bolometric luminosity of the quasar core ( assumed to be about ten times the @xmath181-band core luminosity ) , @xmath182 erg s@xmath7 , as well as the mass of the central supermassive black hole via the virial relation @xcite : @xmath183 this can be taken as a lower limit in this case because the relation was calibrated for the broad component of the civ line , which was not measured independently from the narrow one @xcite ; including the latter underestimates the measured broad - line width @xcite which could be @xmath143,000 km s@xmath7 .
the corresponding eddington luminosity is @xmath184 erg s@xmath7 and the eddington ratio , @xmath185 .
the latter is quite large and may indicate contribution of the emission due to an unresolved portion of the jet to the nuclear optical continuum produced by the accreting matter .
finally , the total observed @xmath186 ghz flux of the 1745 + 624 radio source ( 780.6 mjy ; * ? ? ? * ) , gives the total monochromatic radio luminosity @xmath187 erg s@xmath7 thus a radio loudness parameter , @xmath188 .
note that the discussed object , which possesses the most powerful observed radio jet known , is extremely ` radio - loud ' , as the standard division between radio - quiet and radio - loud sources @xcite is @xmath189 . with the obtained values of the eddington ratio and radio loudness ,
1745 + 624 fits into the ` radio - loud ' trend formed on the @xmath190 plane by local ( @xmath191 ) fr i radio galaxies and radio selected broad - line agns , as discussed in detail by @xcite .
we have analyzed multi - frequency radio , optical and x - ray imaging data for the kpc - scale jet in the @xmath0=3.89 quasar 1745 + 624 .
this quasar hosts the most powerful large - scale radio and x - ray jet yet observed , with monochromatic radio and x - ray luminosities , @xmath192 erg s@xmath7 and @xmath193 erg s@xmath7 , respectively ( excluding emission from the powerful terminal `` hot spot '' ) . aside from its large radiative output , and related large inferred equipartition magnetic field ( in comparison to local powerful jet sources ) , its properties ( multi - wavelength morphology , radio polarization , and broad - band spectral energy distribution ) are broadly similar to other lower - redshift quasar jets .
this is unexpected in light of the dramatic increase in the cmb energy density and strong cosmic evolution of the intergalactic medium at such high - redshifts ; this should have manifested in a dramatically different appearance of such large - scale outflows .
the spectral energy distributions of the resolved linear structures are discussed . in this jet ,
inverse - compton scattered emission on the cmb photons can just account for the x - ray emission if the jet is inclined close to our line of sight ( @xmath6910 ) and if it is also moving relativistically , @xmath70=23 .
several indirect arguments from the multi - scale radio observations of 1745 + 624 support such values of the large - scale jet kinematic parameters .
the data are also consistent with a synchrotron interpretation provided the electron acceleration timescale is much shorter than that of the radiative losses , which we deem likely
. the main distinguishing feature of the two models should manifest in drastically different x - ray / radio emission properties of such high - redshift jets though with only three current @xmath13 examples , no trends are yet apparent . via the virial relation
, we estimate that a @xmath194 supermassive black hole resides at the center of this galaxy .
the broad - band emission of the extended components allow us to additionally estimate a jet kinetic power of @xmath195 ergs@xmath7 , which is a small fraction of the eddington luminosity ( @xmath196 ) corresponding to this black hole mass .
as the bolometric luminosity of the quasar is @xmath197 , the inferred jet kinetic power seems to be much less than the estimated accretion power . despite this , the 1745 + 624 quasar , hosting the most powerful radio jet known , is extremely ` radio - loud ' in comparison to more local radio selected quasars with a radio - loudness parameter @xmath198 , and is surprisingly similar to local radio selected broad line agns regarding the accretion parameters .
this work began while c. c. c. spent the initial year of his fellowship at the mit kavli institute .
he is grateful to the high energy transmission grating group there for their hospitality , and the kipac at stanford for currently hosting his fellowship .
s. acknowledges support by mein through the research project 1-p03d-003 - 29 in years 2005 - 2008 , and by the enigma network through the grant hprn - ct-2002 - 00321 .
this research is funded in part by nasa ( a. s. ) through contract nas8 - 39073 and through chandra award number go5 - 6113x issued by the chandra x - ray observatory center , which is operated by the smithsonian astrophysical observatory for and on behalf of nasa under contract nas8 - 39073 .
the vla is a facility of the national radio astronomy observatory is operated by associated universities , inc . under a cooperative agreement with the national science foundation . based in part on observations made with the nasa / esa hubble space telescope , obtained from the data archive at the stsci .
stsci is operated by the association of universities for research in astronomy , inc . under nasa contract
nas 5 - 26555 .
this research has made use of the united states naval observatory ( usno ) radio reference frame image database ( rrfid ) .
we thank alan fey for providing us these usno datasets and glenn piner for sharing results prior to publication .
we thank dan harris for his practical comments on the discussion .
cohen , a. s. , lane , w. m. , kassim , n. e. , lazio , t. j. w. , cotton , w. d. , condon , j. j. , perley , r. a. , & erickson , w. c. 2006 , in `` from clark lake to the long wavelength array , '' asp conference series , 345 , 299 notes. + knot names indicate their distances relative to the core in arcseconds .
the position angle ( pa ) is positive east of north defined to be zero degrees .
+ the flux densities , @xmath199 , are at the radio frequencies 1.5 , 4.9 , 8.5 , and 14.9 ghz , optical ( opt ; 5.2@xmath20010@xmath201 ghz ; extinction corrected values are indicated in parentheses ) , and 1 kev x - rays . for the hot spot and total jet
, we converted the 210 kev fluxes from the spectral fits into flux density using the fitted spectral indices ( @xmath202 ) indicated .
the flux densities of the individual features come from the ratios of the ( background subtracted ) counts to the total jet , k1.4:k1.8 = 0.6:0.4 .
+ the radio ( @xmath59 ) , radio - to - optical ( @xmath203 ) , radio - to - x - ray ( @xmath126 ) , and x - ray ( @xmath202 ; @xmath204 ) spectral indices where @xmath205 .
the broad band spectral indices use 5 ghz values ( for k0.4 , the 15 ghz measurement was converted to 5 ghz assuming @xmath59=0.9 ) . + * jet fluxes in the radio were measured directly from the vla maps . at 4.9 , 8.5 , and 14.9
ghz , the maps were reconvolved with a common 0.35beam with ( @xmath17 ) tapering applied at the two higher frequencies . at 1.5 ghz
, we used the super - resolved map ( figure [ figure-2 ] ) . | a newly identified kiloparsec - scale x - ray jet in the high - redshift @xmath0=3.89 quasar 1745 + 624 is studied with multi - frequency very large array , hubble space telescope , and _
chandra _ x - ray imaging data .
this is only the third large - scale x - ray jet beyond @xmath1 3 known and is further distinguished as being the most luminous relativistic jet observed at any redshift , exceeding @xmath2 erg / s in both the radio and x - ray bands .
apart from the jet s extreme redshift , luminosity , and high inferred equipartition magnetic field ( in comparison to local analogues ) , its basic properties such as x - ray / radio morphology and radio polarization are similar to lower - redshift examples .
its resolved linear structure and the convex broad - band spectral energy distributions of three distinct knots are also a common feature among known powerful x - ray jets at lower - redshift .
relativistically beamed inverse compton and ` non - standard ' synchrotron models have been considered to account for such excess x - ray emission in other jets ; both models are applicable to this high - redshift example but with differing requirements for the underlying jet physical properties , such as velocity , energetics , and electron acceleration processes .
one potentially very important distinguishing characteristic between the two models is their strongly diverging predictions for the x - ray / radio emission with increasing redshift .
this is considered , though with the limited sample of three @xmath1 3 jets it is apparent that future studies targeted at very high - redshift jets are required for further elucidation of this issue .
finally , from the broad - band jet emission we estimate the jet kinetic power to be no less than @xmath3 erg / s , which is about 10@xmath4 of the eddington luminosity corresponding to this galaxy s central supermassive black hole mass @xmath5 @xmath6 estimated here via the virial relation .
the optical luminosity of the quasar core is about ten times over eddington , hence the inferred jet power seems to be much less than that available from mass accretion .
the apparent super - eddington accretion rate may however suggest contribution of the unresolved jet emission to the observed optical flux of the nucleus . |
stellar orbits in the center of a galaxy are not static , and sometimes stars walk into trouble . if an unlucky star passes within @xmath2 of the galaxy s central black hole ( bh ) , the bh s tidal gravity exceeds the star s self - gravity , and the star is disrupted .
for bhs with @xmath3 , the disruption of a solar - type star occurs outside the horizon and is likely accompanied by a week- to year - long electromagnetic flare ( e.g. , @xcite ) .
gravitational interactions between stars ensure that all supermassive bhs tidally disrupt nearby stars ( e.g. , @xcite ) .
the scattering process might be accelerated by resonant relaxation very close to the bh @xcite , or interactions with `` massive perturbers '' like a massive accretion disk @xcite or giant molecular clouds @xcite .
in addition , the galactic potential may be triaxial so stars need not be scattered at all : they may simply follow their chaotic orbits down to @xmath4 @xcite . given these uncertainties , predictions for the timescale between tidal disruptions in a given galaxy range from @xmath5 to @xmath6 years .
the rate remains uncertain , but tidal disruption must occur .
indeed , a handful of candidate events have been detected to date .
the accreting stellar debris is expected to emit blackbody radiation from very close to the bh , so x - ray and uv observations probe the bulk of the emission .
several candidate tidal disruption events were discovered in the rosat all - sky survey ( see * ? ? ? * ) and the xmm - newton slew survey @xcite ; the galex deep imaging survey has so far yielded three candidates @xcite . for rosat ,
these detections are consistent with a rate @xmath7 per galaxy @xcite , but the data are sparse .
however , we are entering a new era of transient surveys : in the optical , surveys like pan - starrs ( ps1 , then all four telescopes ) ( e.g. , * ? ? ? * ) , the palomar transient factory @xcite , and later the large synoptic survey telescope will have fast cadence , wide fields of view , and unprecedented sensitivity .
wide - field transient surveys with rapid cadence are also planned at other wavelengths , including the radio ( e.g. , lofar and the ata ) , near - infrared ( e.g. , sasir ) , and hard x - rays ( e.g. , exist ) .
how many tidal flares these surveys find depends on their luminosity and spectra as a function of time . in this paper
, we predict the light curves and spectra of tidal disruption events as a function of time .
since the early work on tidal disruption ( e.g. , * ? ? ?
* ) , it has been well - appreciated that the bulk of the emission occurs in the uv and soft x - rays , with a possible extension to harder x - rays .
taking into account only this emission , optical wavelengths are not the most promising for detecting tidal flares , because the blackbody temperature of the inner accretion disk is @xmath8 .
we show , however , that there are two additional sources of optical emission that likely dominate the optical flux in many , though not all , cases : ( 1 ) emission produced by a super - eddington outflow at early times and ( 2 ) emission produced by the irradiation and photoionization of unbound stellar debris ( see also the earlier work of * ? ? ?
* ) . in the latter case ,
much of the optical emission is in the form of very broad emission lines , while in the former it is primarily continuum ( although some lines may also be present ) . throughout this paper
, we typically discuss these two sources of emission separately , largely because the physics of the photoionized stellar debris is more secure than that of the super - eddington outflows .
the remainder of this paper is organized as follows . in [ outflows ] , [ diskmodel ] and [ unbound ] , we describe our models for the polar super - eddington outflow , accretion disk and the equatorial unbound material , respectively ; then in [ emissionprops ] we calculate the luminosity and spectral signatures of tidal disruption events .
we predict detection rates in [ rates ] , and summarize and discuss our results in [ discussion ] .
[ data ] and [ discussion ] include a discussion of our models in the context of rosat and galex observations of tidal flare candidates .
we consider a star approaching the bh on a parabolic orbit with pericenter distance @xmath9 .
once the star reaches the vicinity of the tidal radius ( @xmath10 ) , the tidal gravity stretches it radially and compresses it vertically and azimuthally .
the acceleration is @xmath11 and acts for a dynamical time @xmath12 near pericenter , resulting in velocity perturbations @xmath13 , where @xmath14 is the star s radius and @xmath15 is the star s escape velocity .
the change in velocity @xmath16 is smaller than the star s orbital velocity at pericenter , @xmath17 , by a factor of @xmath18 .
because @xmath16 is at least as large as the sound speed inside the star , the stellar gas may shock vertically and azimuthally ( e.g. , * ? ? ?
* ; * ? ? ?
. once the shredded star passes through pericenter , the compression subsides and the star re - expands , cooling adiabatically ; thermal pressure becomes negligible and the particles travel away from the bh ballistically .
we assume that the particle trajectories become ballistic when the star passes through pericenter . at that time , the particles have perturbed azimuthal , vertical , and radial velocities @xmath19 .
the particles have a range in specific energy @xmath20 ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , due to their relative locations in the bh s potential well and differences in their azimuthal speeds .
initially , approximately half of the stellar mass is bound and half is unbound @xcite . after a time @xmath21 the most bound material returns to pericenter .
( here we have defined @xmath22 , @xmath23 , and @xmath24 . )
less bound gas follows , at a rate @xmath25 @xcite .
there will be some deviations from this canonical @xmath26 scaling at early times , depending on the precise structure of the star @xcite , but we use equation ( [ mdotfallback ] ) for simplicity . as matter returns to pericenter , it shocks on itself , converting most of its bulk orbital energy to thermal energy ( see * ? ? ? * ) .
the viscous time is typically shorter than the fallback time , so at least some of the matter begins to accrete . for @xmath27
, the mass fallback rate predicted by equations ( [ tfallback ] ) & ( [ mdotfallback ] ) can be much greater than the eddington rate @xmath28 for a period of weeks to years ; here @xmath29 , @xmath30 is the eddington luminosity , and 0.1 is the assumed efficiency of converting accretion power to luminosity .
the fallback rate only falls below the eddington rate at a time @xmath31 where @xmath32 .
while the fallback rate is super - eddington , the stellar gas returning to pericenter is so dense that it can not radiate and cool . in particular ,
the time for photons to diffuse out of the gas is longer than both the inflow time in the disk and the dynamical time characteristic of an outflow .
the gas is likely to form an advective accretion disk accompanied by powerful outflows ( e.g. , * ? ? ?
* ) , although the relative importance of accretion and outflows in this phase is somewhat uncertain ( see [ discussion ] ) .
later , when @xmath33 ( @xmath34 ) , the outflows subside , and the accretion disk can radiatively cool and becomes thin . in [ outflows ] , we describe our model for the super - eddington outflows , and in [ diskmodel ] , we describe our model for the accretion disk .
we discuss uncertainties in these models in [ discussion ] .
when the fallback rate to pericenter is super - eddington , radiation produced by the shock and by viscous stresses in the rotating disk is trapped by electron scattering . by energy conservation , this material is initially all bound to the bh , but it is only weakly bound because the radiation can not escape and because the material originated on highly eccentric orbits .
some fraction of the returning gas is thus likely unbound ( see , e.g. , the simulations of * ? ? ?
* ) , with energy being conserved as other gas accretes inward @xcite .
if the outflow s covering fraction is high , most of the radiated power will be emitted from the outflow s photosphere , which can be far outside @xmath35 @xcite .
we now estimate the properties of this outflowing gas ( see @xcite for related estimates in the context of short - duration gamma - ray bursts ) . in our simplified scenario ,
stellar debris falls back at close to the escape velocity and shocks at the launch radius @xmath36 , converting bulk kinetic energy to radiation : @xmath37 where @xmath38 is the temperature at @xmath39 , @xmath40 is the density of gas at @xmath39 , and @xmath41 . outflowing gas
is launched from @xmath39 at a rate @xmath42 and with terminal velocity @xmath43 we approximate the outflow s geometry as spherical , with a density profile @xmath44 inside the outflow where @xmath45 ; the density falls quickly to zero at @xmath46 .
we define the trapping radius @xmath47 via @xmath48 ( where @xmath49 is the opacity due to electron scattering ) : inside @xmath47 , the gas is too optically thick for photons to escape and so the outflowing gas expands adiabatically . because the outflow remains supported by radiation pressure , @xmath50 .
the photosphere of the outflow @xmath51 is where @xmath52 . because @xmath53 is likely not much smaller than @xmath54 , @xmath55
; we thus neglect any deviations from adiabaticity between @xmath47 and @xmath51 . at the earliest times for small @xmath56 and @xmath57
, the fallback rate can be so large and the density so high that the edge of the outflow limits the location of the photosphere to be just inside @xmath58 . in that case
, the density of the photosphere is still given by @xmath59 ; lacking a detailed model , we assume that the photospheric gas near the edge is on the same adiabat as the rest of the gas , so that @xmath60^{1/3 } \\ & \sim & 3\times10 ^ 4 f_{v}^{-1/3 } \mbhrat^{1/36}\rprat^{-1/8}\mstarrat^{-1/12}\rstarrat^{1/12}\left(\frac{t}{\rm day}\right)^{-7/36 } \k \ , .
\nonumber \label{tphedge}\end{aligned}\ ] ] note that the photospheric temperature during the edge - dominated phase is essentially independent of all parameters of the disruption ( e.g. , @xmath61 , @xmath57 , etc . ) , and is only a weak function of time .
the total luminosity during this phase grows as @xmath62 while the luminosity on the rayleigh - jeans tail increases even more rapidly , @xmath63 .
after a time @xmath64 the density falls sufficiently that the photosphere lies well inside @xmath58 ; the photosphere s radius is then @xmath65 and its temperature is @xmath66 the adiabatically expanding outflow preserves the photon distribution function generated in the shock and accretion disk close to the bh .
estimates indicate that this gas is likely to be close to thermal equilibrium and thus we assume that the escaping photons have a blackbody spectrum @xmath67 when the photosphere lies inside the edge of the outflow ( i.e. , @xmath68 so @xmath69 ) , equations ( [ rph ] ) and ( [ tphnoedge ] ) imply that the total luminosity of the outflow is @xmath70 the total luminosity of the outflow is thus of order the eddington luminosity : see figure [ lpeakfig ] , discussed in [ outflow_emission ] .
note that the total luminosity decreases for lower outflow rates , @xmath71 , because the photosphere s surface area is smaller .
the luminosity on the rayleigh - jeans tail ( generally appropriate for optical and near - infrared wavelengths ) declines even faster for lower @xmath72 , scaling as @xmath73 these relations only apply if @xmath74 because otherwise the outflow is optically thin ; we impose this lower limit to @xmath51 in our numerical solutions described later .
we now consider the bound stellar debris that accretes onto the bh . after shocking at pericenter ,
this gas circularizes and viscously drifts inward , forming an accretion disk .
the disk extends from @xmath75 down to the last stable orbit , @xmath76 .
we expect the viscous time in the disk to be substantially shorter than the fallback time for at least a few years ( see @xcite for the case of @xmath77 , assuming a thick disk ) , and check this expectation at the end of [ diskmodel ] ; we thus assume that accretion during this period proceeds at @xmath78 . during the super - eddington phase ,
the time for photons to diffuse out of the disk is longer than the viscous time , and so the disk is thick and advective . in contrast , at later times when @xmath79 , the disk is thin and can cool by radiative diffusion .
we derive an analytic `` slim disk '' model ( similar to the numerical work of * ? ? ?
* ) to describe the structure of the disk in both regimes .
to calculate the disk s properties , we solve the equations of conservation of mass , momentum , and energy : @xmath80 where we have approximated the radial entropy gradient as @xmath81 . here
@xmath82 is the accretion rate , @xmath83 is the cylindrical distance from the bh , @xmath84 is the disk scale height , @xmath85 is the density , @xmath86 is the radial velocity , and @xmath87 is the midplane temperature .
the no - torque boundary condition at the inner edge of the disk implies @xmath88 .
we neglect gas pressure , since radiation pressure is dominant throughout the disk for at least a few years ; we further assume that the viscous stress is proportional to the radiation pressure , so that @xcite @xmath89 with sound speed @xmath90 and @xmath91 , where @xmath92 .
simulations indicate that this assumption is reasonable and that such disks are thermally stable @xcite .
the vertically integrated heating and cooling rates are given by @xmath93 and @xmath94 , where the half - height optical depth is @xmath95 and @xmath49 is the electron scattering opacity .
these relations form a quadratic equation for the dimensionless quantity @xmath96 , @xmath97 solving equation ( [ quadratic ] ) yields the effective temperature of the disk , @xmath98 @xmath99^{-1}. \label{tdisk}\ ] ] combining this relation with equation ( [ mdotfallback ] ) , we calculate the luminosity and spectrum of the disk as a function of time , modeling it as a multicolor blackbody .
the solution to ( [ quadratic ] ) also yields the disk scale height ratio , @xmath100 @xmath99^{-1 } \ , .
\label{hoverr}\ ] ] the scale height @xmath101 while @xmath102 , and decreases as @xmath26 at fixed @xmath83 at later times . the viscous time at a radius @xmath83 in the disk is @xmath103 which is @xmath104 times the local dynamical time during the super - eddington phase , and later increases as @xmath105 at fixed @xmath83 . for @xmath106 ,
the viscous timescale evaluated at the disk s outer edge is shorter than the time @xmath107 since disruption for @xmath108 ; our assumption of steady - state accretion during this period is thus consistent .
while half of the initial star becomes bound to the bh during the disruption , the other half gains energy and escapes from the bh on hyperbolic trajectories . from the viewpoint of the bh , this unbound material subtends a solid angle @xmath109 , with a dispersion @xmath110 in azimuth and a dispersion @xmath111 in orbital inclination .
this material absorbs and re - radiates a fraction of the luminosity from the accretion disk .
we now estimate the dimensions of the unbound wedge . in the orbital plane at a fixed time @xmath112 ,
the unbound stellar debris lies along an arc , as the spread in specific energy produces a spread in radius and azimuthal angle ( see fig .
[ ubdiagram ] ) .
the most energetic particles escape on a hyperbolic orbit with eccentricity @xmath113 .
these particles race away from the bh at a substantial fraction of the speed of light , @xmath114 ( ignoring relativistic effects ) and lie furthest from the bh at a distance @xmath115 they also have the smallest angle away from stellar pericenter , @xmath116 , where @xmath117 obeys @xmath118 so that @xmath119 ( see also * ? ? ?
particles with lower energies and thus smaller eccentricities are closer to the bh and make a larger angle relative to pericenter , up to @xmath120 .
this produces an azimuthal dispersion @xmath121 .
particles having the maximum vertical velocity @xmath122 have vertical specific angular momentum @xmath123 , and total specific angular momentum @xmath124 \sim \rp\vp\left[1+\frac{1}{2}\left(\frac{\rstar}{\rp}\right)^{2}\right]\ ] ] to lowest order in @xmath125 . the orbital inclination @xmath126 is given by @xmath127 , so @xmath128 .
the resulting inclination dispersion is @xmath129 ( our result is consistent with evans & kochanek 1989 but we disagree with khokhlov & melia 1996 ) .
the finite inclination dispersion produces a vertical wall of debris whose inside face scatters , absorbs , and re - radiates a fraction of the disk s emission .
this face subtends a solid angle @xmath130 the number density of particles in the unbound wedge is @xmath131 , where @xmath132 is the radial dispersion of the material at fixed @xmath133 .
this dispersion is due to differences in the particles radial velocities and azimuthal positions when the star passes through pericenter .
particles at @xmath134 travel on orbits whose pericenter is shifted from the star s
pericenter by an angle @xmath135 , which produces a spread in radial position @xmath136 .
the number density is then @xmath137 and the radial column density seen by the black hole is @xmath138 as the unbound material expands , it cools very quickly ; after at most a few weeks , the gas would all be neutral if not for the disk s ionizing radiation .
this radiation ionizes the surface layer of the unbound material .
the ionized gas in turn emits via bremsstrahlung , radiative recombination , and lines .
the physical conditions and processes here are similar to those in the broad line region of an active galactic nucleus ( agn ) .
the ionized gas can reach photoionization equilibrium provided conditions change more slowly than the hydrogen recombination rate @xmath139 .
the recombination coefficient for hydrogen is @xmath140 , and @xmath141 is the electron number density . in the ionized region , @xmath142 , as we show below .
the material can remain in equilibrium for at least a few years , until @xmath143 : @xmath144 the column depth of the ionization front is @xmath145 , where @xmath146 is the ionization parameter , @xmath147 @xmath148 the electron density in the ionized layer is @xmath149 and the fractional depth of the ionization front is @xmath150 @xmath151 so the ionized layer is typically thin and highly ionized .
we use the results of [ bound ] and [ unbound ] to calculate the emission due to the tidal disruption of a solar - type star as a function of time and wavelength .
we consider a solar - type star because the stellar mass - radius relation and typical stellar mass functions imply that these stars should dominate the event rate .
the two key parameters we vary are the star s pericenter distance @xmath57 and the bh mass @xmath56 .
we consider the mass range @xmath152 .
early on , when @xmath102 ( @xmath153 ) , outflowing gas likely dominates the emission .
we calculate its properties using results from [ outflows ] .
in figure [ outflow_spect ] we plot the spectral energy distribution at various times during the outflow phase , for three fiducial models : @xmath154 and @xmath155 ; @xmath156 and @xmath157 ; and @xmath158 and @xmath157 .
we take nominal values of @xmath159 and @xmath160 ; we discuss the uncertainties in these parameters in [ outflow_rates_section ] and [ discussion ] .
the photosphere lies well inside the edge of the outflow at all times shown in figure [ outflow_spect ] .
the emission from the outflow has a blackbody spectrum , initially peaking at optical / uv wavelengths . as time passes and the density of the outflow subsides , the photosphere recedes and the emission becomes hotter but less luminous . in figure [ outflow_lc ]
, we plot @xmath161- ( 4770 ) and @xmath126- ( 7625 ) band light curves for the three fiducial models . in the leftmost panel at @xmath162 ,
the edge of the outflow limits the size of the photosphere , so the photosphere initially expands , following the edge of the outflow . after a time @xmath163 ( eq . [ [ tedge ] ] ) , however , the photosphere begins to recede inside the edge of the outflow and the luminosity declines . in the middle and rightmost panels ,
the photosphere lies well inside @xmath58 for virtually the entire outflow phase .
the optical emission decreases as the photosphere s emitting area decreases and the temperature rises only slowly . as figures [ outflow_spect ] & [ outflow_lc ]
demonstrate , the peak optical luminosity of the outflow is substantial , @xmath164 , comparable to the optical luminosity of a supernova .
the color of the emission is quite blue ( @xmath165 ) . to illustrate how the peak luminosity depends on the parameters of the disruption event , figure [ lpeakfig ] shows the peak bolometric and @xmath161-band luminosities of the outflow as a function of @xmath56 , for @xmath166 and @xmath157 . for sources at cosmological distances ( which _ are _ detectable ;
[ rates ] ) , the negative k - correction associated with the rayleigh - jeans tail implies that the rest - frame @xmath161-band luminosity in figure [ lpeakfig ] underestimates the peak optical luminosity visible at earth . when @xmath102 , a fraction of the falling - back gas is blown away while the remainder likely accretes via an advective disk ( [ diskmodel ] ) . as the fallback rate declines below eddington , the photons are able to diffuse out of the region close to the bh and the disk cools efficiently , but also becomes less luminous .
the vertical dotted line in figure [ outflow_lc ] delineates the super - eddington fallback ( and outflow ) phase from the sub - eddington fallback phase .
the accretion disk irradiates the surface of the equatorial unbound stellar material ( [ unbound ] ) . in this section
we calculate the combined emission produced by the accretion disk and the irradiated stellar debris . in order to isolate the more theoretically secure emission by the disk and photoionized material ,
we do _ not _ consider the emission from super - eddington outflows in this section .
we show results for the disk and photoionized material at both @xmath167 and @xmath34 ; depending on the geometry of the outflow , and the viewing angle of the observer to the source , it is possible that all three emission components could be visible at early times .
because the mass driven away by outflows during the super - eddington phase can also be photoionized by the central source at times @xmath34 , our emission line predictions are likely a lower limit to the total emission line fluxes ( [ discussion ] ) .
we calculate the photoionization properties of the unbound material using version 07.02.02 of the publicly available code cloudy , last described by @xcite .
we simplify the geometry : the unbound spray traces out a widening spiral shape with most of the area coming from close to @xmath168 , so we approximate it as a cloud of area @xmath169 located a distance @xmath170 from the ionizing source .
our model cloud has constant density @xmath171 ( eq . [ [ density ] ] ) , column depth @xmath172 ( eq . [ [ coldensity ] ] ) , and is irradiated by the accretion disk having the luminosity and spectrum described in [ diskmodel ] . the total emission calculated here is the sum of the emission from this photoionized layer and the emission from the central accretion disk .
we focus on non - rotating bhs ( @xmath173 ) , although we quote results for rapidly rotating holes ( @xmath174 ) as well .
in figure [ sed_fiduc ] we plot the spectral energy distribution 30 days , 100 days , 300 days , and 1000 days after disruption , for our three fiducial models : @xmath154 and @xmath155 ; @xmath175 and @xmath157 ; and @xmath158 and @xmath157 .
the early - time short - wavelength peaks at @xmath176 with luminosity @xmath177 are emission from the disk . after a time @xmath178 , the mass fallback rate declines below the eddington rate , and the disk begins to cool and fade . for @xmath179 and @xmath180 ,
the optical light is dominated by lines and continuum from the photoionized material . for larger @xmath56 ( and larger @xmath181 ) ,
the equatorial debris subtends a smaller solid angle ( see eq . [ [ domega ] ] ) and the disk s luminosity is larger , so the disk dominates the optical emission . [ cols="^,^,^,^,^,^",options="header " , ] @xmath182 rates for the emission from the disk and unbound equatorial debris ; the range corresponds to @xmath183 ( low ) , @xmath184 ( high ) .
this emission is relatively faint in the optical / infrared and may be difficult to detect relative to the host bulge ( see [ discussion ] ) .
+ @xmath185 rates for the emission from super - eddington outflows , restricted to @xmath186 , for @xmath160 and @xmath159 ( see fig . [ outflowratesfig ] ) .
+ @xmath187 the survey strategy for sasir has not yet been finalized .
+ @xmath188 these parameters are for comparison with the all - sky rate calculation by @xcite .
+ @xmath189 rosat and galex do not have regular cadences .
+ @xmath190 rosat and galex have insufficient cadence and/or sky coverage to detect flares from super - eddington outflows .
figure [ optir_fiduc ] zooms in on the uv / optical / near - infrared spectra for our three fiducial models .
the emission lines characteristic of the broad line region of an agn are typically the strongest lines here as well : e.g. , ly@xmath191 , civ 1548 + 1551 , h@xmath192 , and h@xmath191 . in most cases , these lines are optically thick for more than a year .
the lines are extremely broad , since the marginally bound gas has a speed close to zero while the most energetic gas leaves the bh at @xmath193 , @xmath194 , and @xmath195 for @xmath196 , @xmath166 ; @xmath197 , @xmath157 ; and @xmath198 , @xmath157 , respectively ( see eq . [ [ vmax ] ] ) .
in addition , the mean velocity along our line of sight will usually be substantial , so the lines should have a large redshift or blueshift on top of the galaxy s redshift . for clarity , we plot the spectra with a mean redshift of zero .
figure [ lines_fiduc ] focuses on the evolution of five strong lines , plotting the ratio @xmath199 at line center for each .
the quantity @xmath200 is the line intensity at line center accounting for the significant broadening .
as the surface area of the equatorial wedge grows in time , line luminosities grow until @xmath201 and irradiation by the disk subsides .
the quantity @xmath202 is the continuum intensity , which includes the contributions of both the disk ( blackbody ) and the photoionized unbound material ( bremsstrahlung and radiative recombination)again , the emission from the super - eddington outflows is not included in @xmath202 .
the lines remain prominent for a few years , and are strongest and broadest for small @xmath56 and small @xmath181 .
the uv lines are the strongest lines when the unbound material dominates the continuum ( left panel ) , while the near - infrared lines are the strongest when the disk dominates the continuum ( middle and right panels ) .
we next describe the broadband optical evolution of a tidal disruption event .
figure [ outflow_lc ] , also discussed in [ outflow_emission ] , plots the optical light curve for each fiducial model , showing the total emission at both @xmath161- and @xmath126-bands .
for @xmath203 ( left of the dotted lines ) , the emission is dominated by the super - eddington outflows , while for @xmath204 the emission is dominated by the accretion disk and photoionized equatorial debris .
once @xmath34 , i.e. , @xmath205 , the disk s optical luminosity falls off gently , approximately as @xmath206 : although the bolometric luminosity is declining as @xmath26 , the optical emission lies on the rayleigh - jeans tail . increasing @xmath181 and/or @xmath56 increases the disk s luminosity by up to two orders of magnitude because of the disk s larger emitting area and/or because @xmath30 rises . at all times
, the disk emission is quite blue ( @xmath207 ) . for large @xmath56 and/or large @xmath181 ( middle and right panels in fig .
[ outflow_lc ] ) , the disk outshines the photoionized material at optical wavelengths , and the light curves and color evolution are determined by the disk emission alone . by contrast , for @xmath208 and @xmath166 ( left panel ) , the photoionized material s optical line emission is initially an order of magnitude brighter than the disk . as the illuminating power of the disk declines but the unbound debris becomes less dense ,
different lines wax and wane .
the significant redshift or blueshift of the unbound material further complicates the photometry by altering which lines contribute in which wavebands ( again , our figures assume a mean redshift of zero ) .
these effects can produce a non - monotonic light curve and a complicated color evolution , depending on the exact redshift of the source and the velocity of the equatorial debris .
we use our calculated spectra and light curves to predict the number of tidal disruption events detectable by observational surveys .
we focus on an ( almost ) all - sky optical survey like the pan - starrs ps1 @xmath209 survey , but we also predict results for surveys with more rapid cadence ( e.g. , ptf and lsst ) and discuss the results of our models compared to rosat and galex observations .
our assumed survey parameters are listed in table [ surveytable ] , along with some of our results . ]
our results can readily be scaled to other surveys using equation ( [ ratescaling ] ) discussed below .
to predict rates , we use the redshift - dependent bh mass function given by @xcite . at @xmath210 ,
the bh density is @xmath211 for @xmath212 and gently falling at higher masses ; as @xmath213 rises to 3 , the bh number density falls by @xmath214 .
we assume that the bh number density for @xmath215 is the same as for @xmath212 , although it is poorly constrained observationally .
we do not consider tidal disruption events beyond @xmath216 .
( our results can easily be scaled to other assumed bh mass densities ; see eq . [ [ ratescaling ] ] ) .
the rate of tidal disruptions within a single galaxy is @xmath217 . to predict detection rates ,
we assume that @xmath218 is independent of bh mass .
we adopt @xmath219 as found by @xcite using the rosat all - sky survey , which is also in line with conservative theoretical estimates .
we further assume that this rate is distributed equally among logarithmic bins of stellar pericenter distance @xmath57 , so that @xmath220 . in the limit of @xmath221 , the equation for the predicted rate is @xmath222 where @xmath223 is the fraction of sky surveyed
; when necessary we use the generalization of equation ( [ rateeqn ] ) that includes cosmological effects . when the duration of a flare @xmath224 is shorter than the cadence of the survey @xmath225 , we approximate the probability of detection as @xmath226 .
we start by considering emission from only the accretion disk and photoionized equatorial debris .
then in [ outflow_rates_section ] , we include the emission from super - eddington outflows , where the physics is somewhat less certain , but the observational prospects are particularly promising . for all but the largest @xmath56 ,
the duration of peak optical emission for the accretion disk and photoionized material is @xmath227 ( eq . [ [ tedd ] ] ) .
this timescale depends on the bh s mass and the star s pericenter distance , as shown in figure [ disk_tevent ] . for @xmath228 ,
the flare lasts for @xmath229 and then decays only gently since the disk dominates the optical emission .
however , for @xmath230 and @xmath231 , the optical flare is shorter , @xmath232 , and then the emission decays more quickly since irradiation of the unbound material the dominant source of optical emission subsides .
for @xmath233 , @xmath234 ( eq . [ [ tfallback ] ] ) because the fallback rate in these systems is never super - eddington .
figure [ ratesfig ] shows our calculated rates for optically - detected tidal flares ( for a survey like the pan - starrs 3@xmath235 survey ) , for both non - rotating and rapidly rotating bhs . for @xmath236 ,
the disk contributes most of the emitted power , so the rates increase with @xmath56 as @xmath237 increases .
the rates are dominated by @xmath238 and @xmath239 ( non - rotating bhs ) and @xmath240 ( rapidly rotating bhs ) .
since most of the flares that dominate the rates have relatively long durations ( fig . [ disk_tevent ] ) , imperfect survey cadence only modifies the detection rates by @xmath241 . at @xmath242 and small @xmath57 ,
the photoionized material re - emits a relatively large fraction of the disk s power in the optical and boosts the detection rates significantly .
integrated over @xmath243 , our estimated rates for the pan - starrs 3@xmath235 survey , assuming non - rotating bhs , are @xmath244 and @xmath245 in @xmath161-band and @xmath126-band , respectively .
the mass range @xmath242 contributes another @xmath246 ( both @xmath161-band and @xmath126-band ) , assuming ( probably optimistically ) that @xmath247 and @xmath218 are the same at @xmath248 as at @xmath249 .
if the bh is rotating faster , @xmath181 can be smaller .
this allows an accretion disk to form for even @xmath250 , widens the disk for all @xmath56 , and increases the solid angle of the unbound material .
indeed for @xmath251 and @xmath252 , the unbound material covers a quarter of the sky !
( at this point @xmath253
so our approximations begin to break down . )
these effects raise the total predicted rates for rapidly rotating bhs to @xmath254 .
figure [ rpbbrates ] plots the detection rates as a function of @xmath181 for @xmath255 , @xmath256 , and @xmath257 , for the disk alone ( light lines ) and for the disk plus photoionized material ( heavy lines ) .
the photoionized material enhances detection rates significantly for most @xmath181 at @xmath258 and @xmath256 , but has little effect for @xmath257 .
the rates decrease substantially for @xmath259 because then the outer radius of the disk @xmath260 is at the last stable orbit ; our disk model assumes a no - torque boundary condition at @xmath76 , implying that there is essentially no emission from the disk when @xmath261 .
although the rates quoted above and in figures [ ratesfig ] and [ rpbbrates ] are for a survey covering 3/4 of the sky , assuming constant @xmath262 and constant bh mass density below @xmath263 , our predicted rates can be scaled to other assumed parameters : @xmath264 assuming that the sources are at @xmath265 , where @xmath266 .
the pan - starrs mds , ptf , and lsst will have cadences of only a few days , giving them sensitivity to all events within their survey volumes .
however , for detecting this relatively long - lived emission , the advantage of fast cadence is only minor ; more significant is the spatial volume probed by the survey ( see table [ surveytable ] ) .
the mds will deeply ( @xmath267 ) image a relatively small ( 84-deg@xmath268 ) region , and so will detect @xmath269 as many events as the @xmath209 survey , @xmath270 per year .
ptf will image a larger region ( @xmath271 ) but less deeply ( @xmath272 ) , and so will have detection rates similar to the mds .
the emission from early - time super - eddington outflows may significantly increase these rates , as we discuss in
[ outflow_rates_section ] . in the next decade
, lsst will image a large region ( 20,000 @xmath273 ) deeply ( @xmath274 ) , and so will discover hundreds of tidal flares .
mds and lsst will detect events at cosmological distances ( @xmath275 ) , where the negative k - correction of the disk s blackbody peak enhances rates by a factor of a few . by co - adding images up to @xmath276
, these short - cadence surveys will also be able to raise their detection rates of events having @xmath277 by a factor of a few . also in the next decade , sasir plans to deeply image 140 @xmath273 each night in the near infrared ( to @xmath278 at @xmath279-band ; @xcite ) .
their observing strategy is not yet finalized ; assuming a 10-day cadence covering 1400 deg@xmath268 ( a good strategy for detecting flares during the super - eddington outflow phase ; see [ outflow_rates_section ] ) , this survey should detect a flare every few years from the accretion disk plus photoionized debris . as a check on our model , we have calculated rates for the rosat all - sky survey for comparison with @xcite s result of 42 events per year over the whole sky for @xmath280 ( @xmath281 kev ) . year but irregular
; we assume it is perfect for simplicity .
the rates we predict thus may be slightly high . ] considering blackbody emission alone , we predict @xmath245 ( @xmath282 ) or @xmath283 ( @xmath284 ) . this strong sensitivity to @xmath285 arises because the rosat band is on the wien tail of the disk emission .
if we assume that 10% of the emission is in an x - ray power - law tail with a photon index @xmath286 ( not unreasonable assumptions for x - ray emission from accreting bhs ; e.g. , * ? ? ?
* ) , our predicted rates are @xmath287 ( @xmath288 ) and @xmath289 ( @xmath284 ) .
given the large uncertainties in the x - ray emission , our predictions are consistent with the observational results .
we also compare with detection rates in the galex deep imaging survey @xcite .
they search an area of 2.882 square degrees , observed at fuv ( 1539 ) and nuv ( 2316 ) down to @xmath290 ab magnitudes .
gezari et al .
detect 2 events in this region over 3 years .
this is somewhat higher than our predicted rates in the nuv , @xmath291 if @xmath282 and @xmath292 if @xmath284 ( assuming perfect cadence for simplicity ) , and may suggest that the disruption rate per galaxy @xmath218 is a factor of a few higher than we have assumed here .
as the above estimates demonstrate , consistency with galex and rosat constraints prefers a rate per galaxy of @xmath293 .
significantly larger disruption rates , as some calculations predict ( e.g. , @xcite ) , are inconsistent with current observational limits unless dust obscuration has significantly biased the rosat and galex results or the large disruption rates are confined to brief epochs in a galaxy s life ( e.g. , a merger ) .
note also that this constraint only applies to massive bhs with @xmath294 , because uv and x - ray surveys select for these systems ( [ discussion ] ) .
we now calculate optical detection rates accounting for emission from the luminous but shorter - lived super - eddington outflows .
we begin by plotting the duration of peak emission for the outflow phase , @xmath295 , in figure [ outflow_tevent ] as a function of @xmath56 for @xmath155 and @xmath157 . for @xmath166 and @xmath296 ,
the duration of the outflow is set by the time at which the photosphere recedes inside the edge of the outflow ( @xmath297 ; eq . [ [ tedge ] ] ) while for larger @xmath56 , the duration is set by the timescale for the most bound material to return to pericenter ( @xmath298 ; eq . [ [ tfallback ] ] ) ; for @xmath157 , the transition from @xmath299 to @xmath300 occurs at a somewhat lower bh mass of @xmath301 .
figure [ outflow_tevent ] shows that most flares last longer than the few - day cadences of surveys like ptf , pan - starrs mds , and lsst , but are much shorter than the months - long cadence of the pan - starrs @xmath209 survey .
rosat and galex are unlikely to have detected events during this phase due to insufficient cadence and sky coverage
. in figure [ outflowratesrpfig ] , we plot detection rates as a function of @xmath181 for @xmath302 , @xmath256 , and @xmath257 , all for the pan - starrs @xmath209 survey at @xmath161-band . at small
@xmath181 , the edge of the outflow limits the size of the photosphere . as @xmath181 increases
, the radius of the photosphere at @xmath163 increases and so the rates increase . for @xmath303 and @xmath257 at the largest @xmath181 , the photosphere is no longer limited by the edge of the outflow , and the maximum luminosity occurs at @xmath300 . for large @xmath181 , the rate declines as the photosphere recedes inward . in figure [ outflowratesfig ] ,
we plot overall detection rates at @xmath161-band for pan - starrs @xmath209 and mds , ptf , and lsst . in the leftmost panel , we assume that @xmath304 of the falling back material flows out in the wind ( @xmath305 ) , as we have previously . here
we restrict detections to redshifts @xmath186 , where our assumed tidal disruption rate per galaxy is appropriate .
the pan - starrs @xmath209 survey should detect @xmath306 , while the mds should detect @xmath307 because of its smaller spatial volume .
ptf should detect @xmath308 as well , since its fast cadence makes up for its smaller spatial volume relative to the pan - starrs @xmath209 survey .
lsst s large spatial volume and rapid cadence should allow it to detect @xmath309 ! assuming the survey parameters and strategy described in
[ diskub_rates_section ] and table [ surveytable ] , sasir should detect @xmath310 as well .
in the central panel of figure [ outflowratesfig ] , we plot detection rates for redshifts @xmath311 to highlight the possibility of studying tidal disruption events at cosmological distances .
deep surveys like mds will be sensitive to events far beyond @xmath312 .
even the shallower @xmath209 survey could detect @xmath214 events per year at @xmath311 .
our predicted rates at @xmath311 are particularly uncertain : higher redshift sources are more likely to be obscured , and the lower mass bhs of interest for tidal disruption are still growing significantly in mass at these redshifts @xcite so the tidal disruption rate and mass function become less certain . nonetheless , the fact that tidal disruption flares will be detectable at @xmath311 with forthcoming surveys highlights that these sources may become a strong probe of the evolution of @xmath313 bhs . in the rightmost panel of figure [ outflowratesfig ] , we show the sensitivity of our predictions to uncertainty in the outflow model by plotting the rates for the pan - starrs @xmath209 survey for different assumptions about the fraction of the material that is blown away in the outflow , @xmath314 . for @xmath315 ,
the detection rate is @xmath316 , while for @xmath317 it falls to @xmath318 .
the rate falls rapidly at the highest @xmath56 because the outflow is optically thin .
figure [ outflowratesfig ] shows that even if we very conservatively assume that only @xmath319 of the material is blown away when the fallback rate is super - eddington , upcoming optical surveys like pan - starrs should still be able to detect a significant number of tidal disruptions during the super - eddington phase .
we have calculated the spectra and light curves produced by the tidal disruption of a solar - type star by a massive black hole .
upcoming optical transient surveys should detect many such events ( [ rates ] ) .
our results demonstrate that there are at least three different emission components that are important during tidal flares : ( 1 ) outflows at early times when the fallback rate is super - eddington , ( 2 ) a compact ( @xmath320 ) accretion disk around the bh , and ( 3 ) stellar debris that is unbound during the disruption and forms an outflowing `` wedge '' in the equatorial plane ( see fig . [ ubdiagram ] ) .
it is also possible that the super - eddington fallback powers a lower - density magnetically - dominated jet , but the properties of such a jet are difficult to predict so we do not consider this potential source of emission .
each of these three components contributes to the total emission from tidal disruption flares . at early times
, the super - eddington outflows likely dominate , producing a few to 10day optical - infrared flare with a luminosity comparable to that of a supernova ( figs .
[ outflow_spect]-[lpeakfig ] & [ outflow_tevent ] ) . as the fallback rate decreases below the eddington rate
, these outflows will diminish , revealing the underlying accretion disk that emits primarily in the uv to soft x - rays ( fig .
[ sed_fiduc ] ) ; at this time , the optical emission is likely to be much less than that of a typical agn and well below that of the super - eddington phase ( fig .
[ outflow_lc])because the accretion disk is not very spatially extended .
the central uv and soft x - ray source photoionizes the inner edge of equatorial stellar debris , producing a spectum of broad emission lines ( figs .
[ sed_fiduc ] , [ optir_fiduc ] , & [ lines_fiduc ] ) whose `` rest '' wavelength should be either blueshifted or redshifted with respect to the host galaxy depending on the line of sight of the observer relative to the escaping material .
we find that this spectroscopic signature of tidal flares is the strongest for low mass bhs because the equatorial stellar debris occupies the largest solid angle in these systems ( eq . [ [ domega ] ] ) .
although the above stages are the focus of this paper , for completeness we briefly discuss the rest of the evolution of a tidal disruption event . as the fallback rate continues to decrease below the eddington rate , the viscous time in the thin disk ( eq . [ [ tvisc ] ] ) increases and becomes comparable to the time @xmath107 since disruption ; at this point , matter begins to build up rather than rapidly accreting onto the bh .
for disruption at @xmath238 , which is likely to produce a significant fraction of the events ( figs .
[ rpbbrates ] & [ outflowratesrpfig ] ) , we estimate that this occurs @xmath321 years after disruption , when the fallback rate has decreased to @xmath322 ( where @xmath323 is the dimensionless viscosity ) . separately
, we expect a significant change in the thermodynamics of the disk when @xmath324 . below this accretion rate ,
the material will no longer cool efficiently when it circularizes and shocks upon returning to pericenter .
instead of cooling to form a thin disk , the material will be shock heated to form a geometrically thick , radiatively inefficient accretion flow . in general
, both geometrically thin / optically thick and geometrically thick / radiatively inefficient disks appear to be stable accretion solutions for a given @xmath325 .
however , in this case , the boundary condition that matter shocks up to the virial temperature upon returning to pericenter picks out the radiatively inefficient solution once @xmath326 .
because the viscous time in a thick disk is @xmath104 times the dynamical time , once the accretion becomes radiatively inefficient , the viscous time is always much shorter than the orbital period of matter returning to pericenter .
moreover , the transition to radiatively inefficient accretion happens at a time comparable to when matter would otherwise begin accumulating in a thin disk , particularly for more massive bhs .
this suggests that there is typically only a limited range of accretion rates ( and time ) during which the `` spreading disk '' solution of @xcite applies .
instead , at late times matter will rapidly accrete via a thick disk and the accretion rate will decay as @xmath327 . as in x - ray binaries
@xcite , we expect that the thermodynamic transition at @xmath328 will be accompanied by a significant change in the luminosity and spectrum of the disk , and perhaps also by the production of relativistic jets .
this should be explored in more detail in future work .
having summarized our key results and the timeline of a tidal disruption event , we now discuss some uncertainties in our models , observational challenges to detecting tidal flares , and the astrophysical applications of studies of tidal disruption events . in
[ outflows ] , we described our simple model for outflows driven when the fallback rate is super - eddington .
energy conservation implies that the falling back material initially remains bound after returning to pericenter and circularizing , but even small amounts of accretion can release additional energy and drive a powerful outflow .
there is , however , a significant uncertainty in precisely how much of the falling back material is blown away , and in the kinematics of the outflow .
we assume that the gas expands roughly spherically from the bh , but the flow is probably somewhat collimated along the pole , due to the original angular momentum of the stellar debris . some photons
can then leak out through the sides of the outflow rather than continuing to drive the expansion ; in this case , the overall emission would be somewhat hotter and fainter , with a dependence on viewing angle .
we have parameterized the terminal velocity of the outflow using @xmath329 , and the mass outflow rate using @xmath330 ; in all of our calculations , we have assumed that @xmath331 .
if the gas actually expands more slowly ( @xmath332 ) , its density will be larger , so the photosphere will be larger , increasing the optical fluxes and detection rates . in the extreme case in which there is no unbound outflow , but super - eddington fallback leads to a radiation - pressure - supported atmosphere around the bh that slowly expands until the photons can diffuse out , we also expect significant optical luminosities during the super - eddington phase ( e.g. , @xcite ) .
it is worth noting that our predictions for the radiation from super - eddington outflows are particularly uncertain for low @xmath56 and small @xmath181 , when the edge of the outflow limits the radius of the photosphere and determines both the peak luminosity ( fig .
[ outflow_lc ] ) and duration ( fig .
[ outflow_tevent ] ) of the flare ; this does not , however , significantly influence our total predicted rates ( fig . [ outflowratesrpfig ] ) .
figure [ outflowratesfig ] shows that even if the outflow rate is just a few percent of the fallback rate ( @xmath333 ) , the outflowing gas is sufficiently bright in the optical that forthcoming surveys should detect a significant number of tidal disruption flares .
this makes early - time optical flares from the tidal disruption of stars an extremely promising candidate for current and future optical / infrared transient surveys .
( we discuss some practical issues associated with detecting these sources in
[ disc_obs ] . ) in this context , we note that in the radiation hydrodynamic simulations of accretion at @xmath334 carried out by @xcite , @xmath335 of the gas becomes unbound , @xmath335 accretes , and the remaining @xmath336 is marginally bound and may ( or may not ) eventually accrete as well .
these precise values will depend on the pericenter of the star , with smaller @xmath181 likely leading to smaller @xmath314 , i.e. , a smaller fraction of the gas being blown away .
future observational constraints on the luminosity , spectrum , and timescale of the super - eddington outflow phase should be able to strongly constrain the value of @xmath314 in individual events .
these results will have important implications for how massive bhs grow .
in particular , if @xmath314 is typically modest , this would imply that black holes can accrete at rates far above the eddington rate , perhaps helping to explain how supermassive bhs ( @xmath337 ) can be observable as luminous quasars as early as @xmath338 .
figure [ outflowratesfig ] also demonstrates that deep optical surveys such as the pan - starrs mds and lsst will be sensitive to tidal flares at high redshift .
these surveys may thus provide a powerful probe of the bh mass function and stellar dynamics in galactic nuclei as a function of redshift .
for example , at @xmath339 , bhs having @xmath340 are still growing significantly in mass @xcite and thus their disruption rates may evolve significantly with redshift .
in addition , galaxy mergers , which are more common at @xmath341 , could substantially increase the tidal disruption rate : @xcite find rates of up to @xmath342 for @xmath343 after the merger due to three - body interactions between stars and a binary bh .
we now consider several aspects of our model for the accretion disk , broad emission lines , and the material unbound during the disruption ( [ diskmodel ] and [ unbound ] ) .
our accretion disk model is designed to describe the emission from the time when the disk first forms through the following few years . during much of this period , the fallback rate is super - eddington , and we expect the disk to be optically and geometrically thick , with radiation pressure dominating gas pressure . once the fallback rate becomes sub - eddington , the disk becomes geometrically thin and may be subject to viscous instabilities ( although it is thermally stable ; @xcite ) . these instabilities may lead to additional time dependence not captured in our models , particularly at late times when the viscous time in the thin disk becomes comparable to the orbital period of the material falling back to pericenter . as described above ,
once @xmath344 , the density is sufficiently low that the flow becomes radiatively inefficient and our model is no longer appropriate .
the disk will then heat up and its spectrum will become significantly harder
. this phase may be detectable by hard x - ray transient surveys like exist .
it is also unclear how much tidal forces will spin up the rotation of the star as it approaches pericenter .
we have assumed that the star is maximally spun up , so that stellar debris is accelerated to relative velocities @xmath345 in the azimuthal direction .
if in fact the spin - up is less effective , the onset of the flare , which occurs at the fallback time ( eq . [ [ tfallback ] ] ) , will be later by a factor of few and the solid angle subtended by the unbound equatorial debris will be somewhat smaller .
this will not change our qualitative conclusions , only some of our quantitative results .
it is important to stress that the line emission we predict may well be an underestimate in all cases : the accretion disk will also photoionize the _ back edge _ of the material that was blown away during the super - eddington phase , which is far from the bh once the outflows subside .
simple estimates indicate that the density and velocity of this outflowing gas are similar to that of the gas unbound at the time of disruption ; as a result , irradiation of this gas will produce additional broad hydrogen lines .
the equivalent width of these lines depends on the solid angle subtended by the super - eddington winds , which , although uncertain , is likely to be significant .
these lines are unlikely to depend as sensitively on @xmath56 and @xmath181 as the emission lines from the equatorial debris ( see fig .
[ optir_fiduc ] for the latter ) . as a result
, observations of the line emission will help constrain the geometry of the outflowing gas created during the tidal disruption event .
in addition to emission lines from the back edge of the outflowing gas at late times , the photosphere of the super - eddington outflow may at early times show strong _ absorption lines _ , particularly in the ultraviolet ( much like the photosphere of a star ) ; these lines would likely be highly blueshifted relative to the lines of the host galaxy .
finally , we reiterate that we expect very little narrow forbidden line emission from tidal disruption events : the outflowing stellar debris is too dense to produce forbidden lines , and there is insufficient time to photoionize ambient lower density gas far from the bh .
it is possible that there is ambient gas in the galactic nucleus sufficiently close to the bh to produce forbidden lines on a @xmath346 timescale , but the prevalence of such gas is currently poorly understood .
candidate detections of tidal disruption flares have thus far been selected by their uv and soft x - ray emission ( predominantly via galex , rosat and xmm - newton ) .
the emission at these wavelengths is primarily produced by the accretion disk , which is brighter for larger @xmath56 . as a result ,
these surveys are most sensitive to bhs having @xmath347 .
uv and x - ray selected events are likely to be discovered somewhat after the initial period of super - eddington fallback , because the outflow during that phase probably precludes direct observation of the underlying accretion disk from many viewing angles .
this outflowing material could also be a significant source of obscuration even at late times when the fallback rate is sub - eddington .
the accretion disk emission we predict for uv - selected events is broadly consistent with the galex candidates : uv luminosities of @xmath348 , optical luminosities of @xmath349 , blackbody temperatures of @xmath350 , and bolometric luminosities of @xmath351 .
the events selected from soft x - rays have less data , but typically have soft x - ray luminosities of @xmath352 .
this emission may be from the accretion disk , at energies just above the blackbody peak , or may be from an x - ray power - law tail with @xmath353 of the bolometric luminosity .
in addition , as discussed in [ data ] , our model is consistent with rosat and galex rate estimates , provided that the tidal disruption rate per galaxy for bhs with @xmath354 is @xmath355 .
this constraint already suggests that galactic nuclei in the nearby universe are relatively spherical , rather than triaxial , because the expected disruption rate is significantly higher in the latter case @xcite . as described in [ emissionprops ] , the spectral signature of the equatorial stellar debris is a transient spectrum of broad emission lines shifted in wavelength relative to the host galaxy .
we do not expect forbidden lines ( e.g. [ nii ] , [ sii ] , [ oi ] , [ oiii ] ) to be present , because the density in the unbound material is too high .
line of [ oiii ] does appear at late times for the lowest @xmath56 and @xmath57 , where the densities fall below this ion s critical density after about a year .
] for several reasons , it is not surprising that this spectral signature has yet to be seen .
first , the tidal flare candidates are likely from relatively high - mass bhs ; in those events , the unbound stellar debris subtends a small solid angle ( eq . [ [ domega ] ] ) and
so the emission lines should be @xmath356 of the bulge luminosity .
future optically - selected tidal flares are more likely to show detectable lines .
it is also important to note that most standard searches for agn in optical / infrared surveys use forbidden lines to identify nuclear activity , and have not specifically looked for faint , broad lines offset from the host galaxy s lines .
tidal disruption events may yet be hiding in archival spectroscopic data .
we predict that outflows during the super - eddington fallback phase have peak optical luminosities of @xmath357 and characteristic decay timescales of @xmath0 days ( figs .
[ lpeakfig ] & [ outflow_tevent ] ) .
these events are sufficiently bright that a natural concern is whether our predictions can already be ruled out by optical supernova searches such as the supernova legacy survey and stripe 82 in the sloan digital sky survey .
although a careful search of archival data is clearly warranted , we do not believe that current observations are necessarily that constraining , for two reasons .
first , the outflow phase can be relatively brief and many survey cadences may be insufficient to find these events .
most importantly , however , tidal flares could be readily mistaken for agn and thus discarded .
indeed , most supernova searches discard galactic nuclei in order to avoid confusion with agn and optimize their probability of detecting supernovae . optically detecting a tidal flare
may require disentangling the flare emission from that of the bh s host galaxy .
for example , at a distance of 300 mpc , a ground - based optical survey with a resolution @xmath358 " should just be able to resolve a kiloparsec - sized bulge .
bulges are found to be @xmath359 times more massive than their central bhs @xcite .
super - eddington outflows typically shine at @xmath360 ( see fig . [ lpeakfig ] ) , so this phase should be at least as bright as the host galaxy and readily detectable given sufficient attention to sources in galactic nuclei and careful screening to rule out an unsteady agn .
by contrast , the typical optical luminosity of the accretion disk itself is @xmath361 ( see fig .
[ outflow_lc ] ) .
the accretion disk would thus brighten the host bulge by only a few percent .
photometric detections of late - time flares will require very careful bulge subtraction . as a result ,
shallow , wide - area surveys such as ptf are more likely to find the late - time disk emission than narrow deep surveys such as the pan - starrs mds . as an additional complication to finding tidal flares , optical extinction in galactic nuclei can be significant ( although less than at uv or soft x - rays ) ; as a result , some fraction of optical tidal flares may not be detectable due to obscuration .
infrared surveys such as sasir , which are also very sensitive to tidal disruption events ( table [ surveytable ] ) , will be particularly immune to the effects of obscuration .
type ii supernovae in the nuclear regions of galaxies may be confused with tidal disruption events , as both have quite blue colors . for sources at @xmath362 mpc
, we estimate that such supernovae will occur within @xmath363 of the galactic nucleus at a rate of @xmath364 , perhaps an order of magnitude more often than tidal disruption events ; at higher redshift the contamination from supernovae will be more significant , but follow - up imaging at high spatial resolution and/or spectroscopic follow - up should help classify these events and distinguish tidal flares from nuclear supernovae .
theoretical calculations of the tidal disruption rate per galaxy , @xmath218 , vary substantially and can have complicated dependences on bh mass and pericenter distance ( e.g. , * ? ? ?
the rate model we implement is consistent with theoretical estimates and is sufficiently simple that the reader can easily scale our results to different model parameters ( eq . [ [ ratescaling ] ] ) .
we have assumed that the rate at which stars enter the disruption zone ( @xmath365 ) is independent of bh mass , and constant with @xmath366 .
a star may venture deep into the disruption zone ( @xmath367 ) on its last orbit if its change in angular momentum over one dynamical time is large enough at least of order the maximum angular momentum for disruption .
this condition is satisfied in the full loss cone regime , and marginally satisfied in the outskirts of the diffusive regime . for realistic stellar density profiles ,
the disruption rate is dominated by the boundary between these two regimes ( e.g. , * ? ? ?
* ) , so many stars probably do take large enough angular momentum steps to arrive at @xmath368 . in the diffusive regime ,
the disruption rate per @xmath369 varies weakly with @xmath369 , consistent with our assumption in [ rates ] .
given the sensitivity of the optical - infrared emission from tidal flares to @xmath57 , upcoming surveys should significantly improve our knowledge of the stellar dynamics in galactic nuclei .
our results demonstrate that optical transient surveys will be quite sensitive to the lowest mass bhs in galactic nuclei , both because of the outflows produced when the fallback rate is super - eddington and because of the large angle subtended by the equatorial stellar debris .
such bhs are otherwise difficult to detect because their host galaxies are faint , it is difficult to resolve their small spheres of influence , and even when they are active , their eddington luminosities are low .
the space density of @xmath242 bhs and the stellar density profiles in the galaxies they inhabit are only moderately well - constrained at present ( e.g. , * ? ? ?
* ) , as is the role of tidal disruption in growing these bhs .
optical searches for tidal flares should thus prove to be a powerful probe of low - mass bhs and their host galaxies .
we thank josh bloom , phil chang , eugene chiang , luis ho , and enrico ramirez - ruiz for helpful discussions ; we also thank phil hopkins for providing the black hole mass function in tabular form ; and we thank gary ferland , ryan porter and peter van hoof at the cloudy discussion board for rapid and helpful replies to our questions . les and eq are supported in part by the david & lucile packard foundation and nasa grant nng06gi68 g . les dedicates this work to the memory of two shining stars , sandra strubbe and doug baker .
abramowicz m.a . , czerny b. , lasota j.p . , szuszkiewicz e. 1988 , apj , 332 , 646 alexander t. 2005 , physics reports , 419 , 65 ayal s. , livio m. , piran , t. 2000 , apj , 545 , 772 blandford , r.d . , begelman m.c .
1999 , mnras , 303 , l1 bloom j.s .
, prochaska j.x . , lee w. et al . 2009 , arxiv:0905.1965v1 [ astro-ph.im ] bogdanovi t. , eracleous m. , mahadevan s. , sigurdsson s. , laguna p. 2004
, apj , 610 , 707 brassart m. , luminet j .- p .
2008 , , 481 , 259 cannizzo j.k .
, lee h.m .
, goodman j. 1990 , apj , 351 , 38 chen x. , madau p. , sesana a. , liu , f.k .
2009 , arxiv:0904.4481v1 [ astro-ph.ga ] donley j.l .
, brandt w.n . , eracleous m. , boller t. 2002 , apj , 124 , 1308 esquej p. , saxton r.d . , freyberg m.j . ,
read a.m. , altieri b. , sanchez - portal m. , hasinger g. 2007 , a&a , 462 , l49 evans c. , kochanek c. 1989 , apjl , 346 , l13 ferland g.j .
, korista k.t . ,
verner d.a .
, ferguson j.w . , kingdon j.b . , verner e.m .
1998 , pasp , 110 , 761 gezari s. , martin d.c .
, milliard b. et al .
2006 , apj , 653 , l25 gezari s. , basa s. , martin d.c .
2008 , apj , 676 , 944 gezari s. , heckman t. , cenko s.b .
2009 , arxiv:0904.1596v1 [ astro-ph.co ] greene j.e .
2007 , apj , 670 , 92 grindlay j.e .
2004 , aip conf .
proc . , 714 , 413 guillochon j. , ramirez - ruiz e. , rosswog s. , kasen d. 2008 , arxiv:0811.1370v2 [ astro - ph ] haring n. , rix h.w .
2004 , apj , 604 , l89 heckman t.m . , kauffmann g. , brinchmann j. , charlot s. , tremonti c. , white s.d.m .
2004 , apj , 613 , 109 hirose s. , krolik j. , blaes o. 2009 , apj , 691 , 16 hopkins p.f . , richards , g.t . , hernquist , l. 2007 , apj , 654 , 731 khokhlov a. , melia f. 1996 , apjl , 457 , l61 kochanek c.s .
1994 , apj , 422 , 508 komossa s. 2002 , rev .
15 , 27 koratkar a. , blaes o. 1999 , pasp , 111 , 1 lacy j.h . , townes c.h . , hollenbach d.j .
1982 , apj , 262 , 120 li l. , narayan r. , menou k. 2002 , apj , 576 , 753 lodato g. , king a.r . ,
pringle j.e .
2009 , mnras , 392 , 332 loeb a. , ulmer a. 1997 , apj , 489 , 573 magnier e. 2007 , aspc , 364 , 153 , ed .
c. sterken ( san francisco : asp ) magorrian j. , tremaine s. 1999 , mnras , 309 , 447 merritt d. , poon m.y .
2004 , apj , 606 , 788 ohsuga k. , mori m. , nakamoto t. , mineshige s. 2005 , apj , 628 , 368 perets h.b . , hopman c. , alexander t. 2007 , apj , 656 , 709 phinney e.s .
1989 , in iau symp .
136 , the galactic center , ed .
m. morris ( dordrecht : kluwer ) ramirez - ruiz e. , rosswog s. 2009 , apjl , 697 , l77 rau a. , kulkarni s.r . , law n.m .
2009 , arxiv:0906.5355v1 [ astro-ph.co ] rauch k.p . ,
tremaine s. 1996 , newa , 1 , 149 rees , m. 1988 , nature , 333 , 523 remillard r. , mcclintock j. 2006 , ara&a , 44 rossi e.m . , begelman m.c .
2009 , mnras , 392 , 1451 rybicki g.b .
, lightman a.p .
1979 , radiative processes in astrophysics ( new york : john wiley & sons , inc . ) shakura n.i . ,
sunyaev r.a .
1973 , a&a , 24 , 337 ulmer a. 1999 , apj , 514 , 180 zhao h. , haehnelt m.g . , rees , m.j .
2002 , newa , 7 , 385 | a star that wanders too close to a massive black hole ( bh ) is shredded by the bh s tidal gravity .
stellar gas falls back to the bh at a rate initially exceeding the eddington rate , releasing a flare of energy . in anticipation of upcoming transient surveys ,
we predict the light curves and spectra of tidal flares as a function of time , highlighting the unique signatures of tidal flares at optical and near - infrared wavelengths .
a reasonable fraction of the gas initially bound to the bh is likely blown away when the fallback rate is super - eddington at early times .
this outflow produces an optical luminosity comparable to that of a supernova ; such events have durations of @xmath0 days and may have been missed in supernova searches that exclude the nuclear regions of galaxies . when the fallback rate subsides below eddington , the gas accretes onto the bh via a thin disk whose emission peaks in the uv to soft x - rays
. some of this emission is reprocessed by the unbound stellar debris , producing a spectrum of very broad emission lines ( with no corresponding narrow forbidden lines ) .
these lines are the strongest for bhs with @xmath1 and thus optical surveys are particularly sensitive to the lowest mass bhs in galactic nuclei . calibrating our models to rosat and galex observations
, we predict detection rates for pan - starrs , ptf , and lsst and highlight some of the observational challenges associated with studying tidal disruption events in the optical .
upcoming surveys such as pan - starrs should detect at least several events per year , and may detect many more if current models of outflows during super - eddington accretion are reasonably accurate .
these surveys will significantly improve our knowledge of stellar dynamics in galactic nuclei , the physics of super - eddington accretion , the demography of intermediate mass bhs , and the role of tidal disruption in the growth of massive bhs .
= -0.7 cm [ firstpage ] galaxies : nuclei black hole physics optical : galaxies |
the study on exotic hadron spectroscopy is experiencing a renaissance in the last decade .
more and more charmonium - like and bottomonium - like states ( called @xmath2 ) have been announced by experiments in various processes ( see refs .
@xcite for a review ) .
several charged structures with a hidden @xmath3 or @xmath4 , such as the@xmath5 @xcite , @xmath6 @xcite , @xmath7 @xcite , and @xmath8 @xcite were observed by experiments , which would be exotic state candidates .
very recently , the d0 collaboration observed a narrow structure @xmath1 in the @xmath9 invariant mass spectrum with @xmath10 significance @xcite . the mass and width are measured to be @xmath11 mev and @xmath12 mev , respectively .
the quark components of the decaying final state @xmath13 are @xmath14 ( or @xmath15 ) , which requires @xmath1 should be a structure with four different valence quarks . after the discovery of @xmath1 , several theoretical investigations are carried out in order to understand its underlying structure . in ref .
@xcite , the mass and decay constant of @xmath1 were computed within the two - point sum rule method using the diquark - antidiquark interpolating current .
the mass and pole residue are studied with the qcd sum rules in @xcite .
its mass spectrum was also calculated in ref .
@xcite using diquark - antidiquark type interpolating current . in ref .
@xcite , the authors estimated the partial decay width @xmath16 with @xmath1 being an @xmath17-wave @xmath18 molecular state .
it is necessary to study some other possibilities , before we claim that @xmath1 is a genuine particle , such as tetraquark or molecular state .
some non - resonance explanations have ever been proposed to connect resonance - like peaks with kinematic singularities induced by the rescattering effects @xcite .
it is shown that sometimes it is not necessary to introduce a genuine resonance to describe a resonance - like structure , because some kinematic singularities of the rescattering amplitudes will behave themselves as bumps in the invariant mass distributions . in this work
, we are trying to use the so - called triangle singularity ( ts ) mechanism to describe the observation of @xmath1 .
the work is organized as follows : in section ii , the ts mechanism is briefly introduced ; in section iii , we discuss several rescattering processes where the ts can be present and simulate the resonance .
a brief summary is given in section iv .
is @xmath19 ( @xmath20@xmath211 , 2 , 3 ) .
the momentum symbols also represent the corresponding particles . ] the possible manifestation of the ts was first noticed in 1960s .
it is found that the ts s of rescattering amplitudes can mimic resonance structures in the corresponding invariant mass distributions @xcite .
this offers a non - resonance explanation about the resonance - like peaks observed in experiments .
unfortunately , most of the proposed cases in 1960s were lack of experimental support .
the ts mechanism was rediscovered in recent years and used to interpret some exotic phenomena , such as the largely isospin - violation processes , the production of some exotic states and so on @xcite . for the triangle diagram in fig .
[ tsmechanism ] , all of the three internal lines can be on - shell simultaneously under some special kinematic configurations .
this case corresponds to the leading landau singularity of the triangle diagram , and this leading landau singularity is usually called the ts .
the physical picture concerning the ts mechanism can be understood like this : the initial particles @xmath22 and @xmath23 firstly scatter into particles @xmath24 and @xmath25 , then the particle @xmath26 emitted from @xmath24 catches up with @xmath25 , and finally @xmath24 and @xmath25 will rescatter into particles @xmath27 and @xmath28 .
this implies that the rescattering diagram can be interpreted as a classical process in space - time with the presence of ts , and the ts will be located on the physical boundary of the rescattering amplitude @xcite .
the ts mechanism is very sensitive to the kinematic configurations of rescattering diagrams .
it is therefore necessary to determine in which kinematic region the ts can be present . in fig .
[ tsmechanism ] , we define the invariants @xmath29 , @xmath30 and @xmath31 . the locations of ts can be determined by solving the so - called landau equations @xcite . for the diagram in fig .
[ tsmechanism ] , if we fix the internal masses @xmath19 , the external invariants @xmath32 and @xmath33 , we can obtain the solutions of ts in @xmath34 , i.e. , @xmath35},\end{aligned}\ ] ] with @xmath36 .
likewise , by fixing @xmath19 , @xmath34 and @xmath33 we can obtain the similar solutions of ts in @xmath32 , i.e. , @xmath37}.\end{aligned}\ ] ] by means of the single dispersion representation of the 3-point function , we learn that within the physical boundary only the solution @xmath38 or @xmath39 will correspond to the ts of rescattering amplitude , and @xmath40 and @xmath41 are usually defined as the anomalous thresholds @xcite . for convenience
, we further define the normal threshold @xmath42 ( @xmath43 ) and the critical value @xmath44 ( @xmath45 ) for @xmath34 ( @xmath32 ) as follows @xcite , @xmath46 , \nonumber \\ & & s_{2n}=(m_1+m_3)^2,\ s_{2c}=(m_1+m_3)^2 + \frac{m_3}{m_2}[(m_2-m_1)^2-s_3].\end{aligned}\ ] ] if we fix @xmath33 and the internal masses @xmath47 , when @xmath34 increases from @xmath48 to @xmath49 , @xmath39 will move from @xmath50 to @xmath51 .
likewise , when @xmath32 increases from @xmath51 to @xmath50 , @xmath38 will move from @xmath49 to @xmath48 .
this is the kinematic region where the ts can be present .
the discrepancies between normal and anomalous thresholds can also be used to represent the ts kinematic region .
the maximum values of these discrepancies take the form @xmath52 , \nonumber \\
\delta_{s_2}^{\max}&=&\sqrt{s_{2c } } - \sqrt{s_{2n}}\approx \frac{m_3}{2m_2(m_1+m_3)}[(m_2-m_1)^2-s_3].\end{aligned}\ ] ] in ref .
@xcite , it was argued that for the single channel rescattering process , when the corresponding resonance - production tree diagram is added coherently to the triangle rescattering diagram , the effect of the triangle diagram is nothing more than a multiplication of the singularity from the tree diagram by a phase factor . therefore the singularities of triangle diagram can not produce obvious peaks in the dalitz plot projections .
this is the so - called schmid theorem .
but for the coupled - channel cases , the situation will be quite different from the single channel case discussed in ref .
@xcite . for the rescattering diagrams which will be studied in this paper , the intermediate and final states are different , therefore the singularities induced by the rescattering processes
are still expected to be visible in the dalitz plot projections .
the reader is referred to refs .
@xcite for some comments about the schmid theorem , and refs .
@xcite for some discussions about the coupled - channel case .
we will focus on the coupled - channel cases in this work .
via the triangle rescattering diagrams ( a ) @xmath53$]-loop and ( b ) @xmath54$]-loop .
@xmath55 represents the incident state . ]
the mass of @xmath1 is very close to the @xmath56-threshold , which is about 5555 mev @xcite .
one may wonder whether there are some connections between @xmath1 and the coupled - channel scattering @xmath57 near threshold . in the high energy collisions ,
the production of @xmath0 may receive contributions from rescattering diagrams illustrated in fig .
[ feymantri ] .
the intriguing characteristic of this kind of diagrams is ts s are expected to be present in the rescattering amplitudes , which may result in some resonance - like bumps in @xmath0 distributions around the @xmath58-threshold accordingly .
.ts kinematic region corresponding to the rescattering diagrams in fig .
[ feymantri ] , in unit of gev . [ cols="^,^,^,^,^,^,^",options="header " , ] [ kmregion ] the momentum and invariants conventions of fig .
[ feymantri ] are the same with those of fig .
[ tsmechanism ] . according to eqs .
( [ s1ns1c ] ) and ( [ deltas1s2 ] ) , the kinematic region where the ts can be present is displayed in table [ kmregion ] .
it can be seen that the kinematic region of ts in @xmath34 is very large for both of the diagrams in fig .
[ feymantri ] .
@xmath59 is nearly 1 gev for each of the diagrams .
firstly , this is because the quantity @xmath60 $ ] in eq .
( [ deltas1s2 ] ) is large .
physically , this quantity corresponds to the phase - space factor for @xmath61 ( @xmath62 ) , which is sizable .
secondly , the ratio @xmath63 is equal to @xmath64 , which is also quite large .
this means that the kinematic conditions of the presence of ts can be fulfilled in a very broad energy region of incident states .
the kinematic requirement on the incident state would be largely relaxed , which is an advantage to observe the effects resulted by the ts mechanism .
on the other hand , the kinematic region of ts in @xmath32 is relatively smaller .
@xmath65 is about 0.2 gev for each of the diagrams , which implies that the ts peaks in @xmath0 distributions may not stay far away from the @xmath58-threshold ( normal threshold @xmath43 ) .
we will naively construct some effective lagrangians to estimate the behaviour of the rescattering amplitudes .
taking into account the conservation of angular momentum and parity , the quantum numbers of the incident state @xmath55 are set to be @xmath66=@xmath67 .
some of the lagrangians read @xmath68,\end{aligned}\ ] ] where @xmath69 and @xmath70 represent the light vector and pesudoscalar mesons respectively .
the process @xmath57 can be a @xmath71-wave scattering , and the corresponding lagrangian takes the form @xmath72 the @xmath71-wave scattering implies that the quantum numbers of @xmath58 and @xmath0 systems would be @xmath66=@xmath73 .
it should be mentioned that the processes @xmath74 in fig .
[ feymantri ] can also happen through weak interactions in the high energy collisions , therefore the parity does not have to be conserved for this vertex .
-dependence of the rescattering amplitude squared @xmath75 .
plots ( a ) and ( b ) correspond to figs . [
feymantri](a ) and ( b ) , respectively .
@xmath76 is the invariant mass @xmath77 of incident state @xmath55 .
the vertical dashed line indicates the position of @xmath1.,title="fig : " ] -dependence of the rescattering amplitude squared @xmath75 .
plots ( a ) and ( b ) correspond to figs . [ feymantri](a ) and ( b ) , respectively .
@xmath76 is the invariant mass @xmath77 of incident state @xmath55 .
the vertical dashed line indicates the position of @xmath1.,title="fig : " ] when the kinematic conditions of ts are fulfilled , it will imply that the particle @xmath24 in fig .
[ tsmechanism ] will be unstable .
we then introduce a breit - wigner type propagator @xmath78^{-1}$ ] to account for the width effect when calculating the triangle loop integrals .
the complex mass of the intermediate state will remove the ts from physical boundary by a distance @xcite .
if the width @xmath79 is not very large , the ts will lie close to the physical boundary , and the scattering amplitude can still feel the influence of the singularity . the numerical results for @xmath0 invariant mass distributions corresponding to the rescattering processes in fig .
[ feymantri ] are displayed in fig .
[ distributiontri ] .
we ignore the explicit couplings but just focus on the line - shape behavior here .
the distributions are calculated at several incident energy points . from fig .
[ distributiontri ] , it can be seen that some bumps arise around the position of @xmath1 .
since the bumps around @xmath1 can be present for a very broad incident energy region , it is possible that the observation of @xmath1 is due to some kind of accumulative effects of the rescattering amplitudes at different incident energies .
the bumps in fig .
[ distributiontri](a ) are broader compared with those in fig .
[ distributiontri](b ) .
this is because the decay width of @xmath80-meson ( 149 mev ) is larger than that of @xmath81-meson ( 50 mev ) @xcite .
the larger decay width will remove the corresponding ts in the complex - plane further away from the physical boundary .
the @xmath71-wave scattering @xmath57 will also smooth the ts peaks to some extent .
-channel contributions for @xmath57 .
( a ) : photon - exchange ; ( b ) : @xmath82-exchange . ] corresponding to the @xmath83-channel scatterings in fig . [ emphi ] .
@xmath84 represents the scattering energy .
] the scattering process @xmath57 would be ozi suppressed . in eq .
( [ contact ] ) , we assume a contact interaction , which may account for the short range part of this interaction . if we take into account the @xmath83-channel contributions , because the momentum transfer in this process will be very small , some long range interactions , such as the electromagnetic ( em ) interaction , may become important . to judge whether the em interaction may play a role in @xmath57
, we will compare the contributions of @xmath83-channel processes illustrated in fig .
[ emphi ] . in figs .
[ emphi](a ) and ( b ) , we use the photon- and @xmath82-exchange diagrams to partly account for the em and strong interactions , respectively .
some effective lagrangians are constructed as follows @xmath85 we firstly compare the coupling constants of these two diagrams .
if adopting the vector meson dominance model @xcite , the ratio @xmath86 would be equal to @xmath87 , where the couplings @xmath88 and @xmath89 are estimated to be @xmath90 and @xmath91 according to the decay widths of @xmath92 and @xmath93 , respectively .
the ratio @xmath94 is then obtained to be about @xmath95 . according to this naive estimation
, we can see that the em couplings may not be very smaller compared with the ozi suppressed strong couplings . without introducing any form factors to account for the off - shell effects
, we further integrate over the moment transfer @xmath83 and obtain the cross section ratios @xmath96 for different scattering energies , which is displayed in fig .
[ emratio ] .
we can see that the cross section corresponding to fig .
[ emphi](a ) is larger than that corresponding to fig .
[ emphi](b ) by about three orders of magnitude .
this is mainly because the quantity @xmath97 is much larger than @xmath98 .
we can then make a quantitative judgment that the contribution of em interaction may be comparable with that of strong interaction .
taking into account the above arguments , the triangle diagram in fig .
[ feymantri ] can be changed into the box diagram in fig .
[ feynmanbox ] accordingly , and the numerical results for the @xmath0 invariant mass distributions are displayed in fig .
[ distributionbox ] .
because the masses of @xmath99 and @xmath100 are different , it can be judged that there will no infrared divergence in these box diagrams @xcite .
compared with the resonance - like bumps in fig .
[ distributiontri ] , the bumps in fig .
[ distributionbox ] are much narrower and more like resonance peaks .
this implies that the interaction details of the scattering @xmath57 may affect the ts mechanism to some extent .
if the long range em interactions play a dominant role in @xmath57 , one can not expect there will be ts peaks arising in the @xmath101 invariant mass distributions , because there is no @xmath102 vertex .
we can further conclude that if the observation of @xmath1 is due to the rescattering effects and the long range em interaction dominates the scattering @xmath57 , there will no charge neutral partner of @xmath1 . via the box rescattering diagrams ( a ) @xmath103$]-loop and ( b ) @xmath104$]-loop . ]
-dependence of the rescattering amplitude squared @xmath75 .
plots ( a ) and ( b ) correspond to figs . [
feynmanbox](a ) and ( b ) , respectively .
@xmath76 is the invariant mass @xmath77 of incident state @xmath55 .
the vertical dashed line indicates the position of @xmath1.,title="fig : " ] -dependence of the rescattering amplitude squared @xmath75 .
plots ( a ) and ( b ) correspond to figs . [
feynmanbox](a ) and ( b ) , respectively .
@xmath76 is the invariant mass @xmath77 of incident state @xmath55 .
the vertical dashed line indicates the position of @xmath1.,title="fig : " ] as stated before , the process @xmath74 can also happen via the weak interactions , such as @xmath105 .
interestingly , according to the quark model calculation in ref .
@xcite , it can be noticed that there are many charm - beauty mesons @xmath106 , of which the masses just fall into the region @xmath107 gev .
this energy region has a large overlap with the ts kinematic region displayed in table [ kmregion ] .
this is another support that the ts mechanism may play a role in the observation of @xmath1 .
in this paper , we investigated the invariant mass distributions of @xmath0 via different rescattering processes .
because the ts s of rescattering amplitudes could be present for a very broad incident energy region , one can expect that the ts peaks may mimic the resonance - like structure @xmath1 . the ts mechanism is highly process - dependent , which is different from other dynamic mechanisms .
if the kinematic conditions of the ts are not fulfilled , there will be no peaks arising in the amplitudes .
however , one would expect that the genuine particles should also appear in the processes where kinematic conditions of the ts are not fulfilled .
if the observation of @xmath1 is due to the rescattering effects , because of the @xmath71-wave scattering characteristic of the process @xmath57 , the quantum numbers of `` @xmath1 '' would be @xmath108 .
this is different from the assignment of diquark - antidiquark picture @xcite or @xmath109 molecular state model @xcite , where the quantum numbers are set to be @xmath110 or @xmath67 .
further experiments with angular distribution analysis may help us to clarify the ambiguities and check different mechanisms .
helpful discussions with makoto oka , wei wang , qiang zhao and shi - lin zhu are gratefully acknowledged .
this work is supported in part by the japan society for the promotion of science under contract no .
p14324 , the jsps kakenhi ( grant no .
25247036 ) , the national natural science foundation of china ( grant no .
11275113 ) and the natural science foundation of shandong province ( grant no .
zr2015jl001 ) . | we investigate the invariant mass distributions of @xmath0 via different rescattering processes .
because the triangle singularity ( ts ) could be present for a very broad incident energy region , it can be expected that the ts peaks may simulate the resonance - like bump @xmath1 observed by the d0 collaboration . the highly process - dependent characteristic of ts mechanism offers a criterion to distinguish it from other dynamic mechanisms . |
the roots of the mass renormalization program of quantum electrodynamics ( qed ) trace back to the late @xmath2 and early @xmath1 centuries notion that the electron s inertia has a purely electromagnetic origin .
this notion was central to the pre - quantum attempts of the time to construct a consistent electromagnetic theory of matter and radiation that would unify electrodynamics and mechanics , now known as classical electron theory .
when lorentz proposed that abraham s @xcite purely electromagnetic , strictly rigid spherical electron ( see appendix a.2 ) had to be replaced by an electron that undergoes the lorentz - fitzgerald contraction , the laws of special relativity began to take shape in the papers of lorentz @xcite , and poincar @xcite ; see @xcite for an excellent account of the historical developments .
ironically , special relativity and renormalization later contributed to the demise of relativistic lorentz electrodynamics ( led ) more than quantum mechanics . of course , quantum mechanics removed led from the list of contenders for a truly fundamental electromagnetic theory of matter and radiation , but it did not jeopardize the viability of led as a physical theory in the classical limit . after all , bohr s correspondence principle was undisputed in the quest for the fundamental quantum theory , and as such one would not have been , nor would one be unhappy with a consistent classical microscopic relativistic theory of electromagnetic matter and radiation that could serve as the empirically correct classical limit of a consistent quantum theory .
special relativity on the other hand had emancipated itself in the works of einstein @xcite and minkowski @xcite as a theory of spacetime structure independent of led or any other particular matter model @xcite .
thus , when lorentz objected to the uhlenbeck
goudsmit @xcite proposal of an electron spin magnetic moment by pointing out that led would forbid the required magnitude of about one bohr magneton because the purely electromagnetic electron would have to reach an impossible equatorial rotation speed of about 10 times the speed of light ( see @xcite , p. 35 ) , he had furnished an argument that eventually weighed heavily against the physical sensibility of led itself , and which has persisted ever since .
judged from our modern perspective on mass renormalization @xcite , the purely electromagnetic calculations by lorentz @xcite and his contemporaries do , however , not qualify as proper .
a proper renormalization analysis would investigate the ` flow ' toward vanishing bare rest mass for a consistent relativistic led _ with _ positive bare rest mass ( ` massive ' led ) in which the bare particle parameters of the model can be chosen such that the particle charge , the particle spin magnetic moment , and the renormalized particle mass match the empirical electron data without involving superluminal gyration speeds of the model electron .
if the limit of vanishing bare rest mass exists , this limit would define a renormalized led which in a sense completes lorentz program of a purely electromagnetic electron theory , and whose physical viability is no longer objectable to on basis of the superluminal arguments that grounded the original lorentz theory .
in this paper we will show that such a properly renormalized approach to the lorentz program of electrodynamics does indeed lead to a satisfactory lorentz electrodynamics . for simplicity
we consider only a single lorentz electron coupled to the maxwell lorentz fields in the whole spacetime .
the extension to a many particle theory does not pose any difficulties .
we first present a novel , manifestly lorentz - covariant massive led which displays several crucial features expected of a realistic , consistent classical electrodynamics .
in particular : * the dynamical equations constitute a cauchy problem for the evolution of the physical state of massive led ; * scattering does not change the nature of the renormalized particle , which moves as a _ soliton _ ; * the parameters of the bare particle can be chosen such that charge , magnetic moment , and renormalized mass of the ` free ' particle can be matched to the physical electron data without involving superluminal gyration speeds .
we then study the renormalization flow to vanishing bare rest mass of our empirically matched massive led .
the ` renormalized led ' which emerges in the limit is characterized by the following additional features : * the equatorial gyration speed reaches the speed of light @xmath3 precisely in the limit of vanishing bare rest mass so that the bare gyrational mass converges to a nonvanishing contribution .
it is fitting to call this a ` photonic ' mass ; * the renormalized spin magnitude is a _
derived _ property in our model .
it turns out to be @xmath4 rather than @xmath5 , with corrections of order @xmath6 ( sommerfeld s fine structure constant ) , and we have @xmath7 rather than @xmath8 or @xmath9 . * our renormalized led is , strictly speaking , a classical field theory with ultraviolet cutoff corresponding to the physical electron s compton length !
it is not unexpected that in our classical theory the relation between spin and magnetic moment holds with a land factor @xmath10 . on the other hand ,
since we operate here with a minimalist model , it is not _ a priori _ clear whether a slightly more intelligently constructed classical model would not account also for the correct @xmath11 . in any event ,
the issue of the @xmath12 factor should not be confused with the superluminal speed problem of lorentz . among the many physically correct features of the model
, the truly rewarding one is its soliton dynamics . without it the rest mass , magnetic moment and spin of the renormalized particle
would typically be altered by each scattering process , and the theory would then simply fail to account for the classical motion of electrons and their electromagnetic fields . another gratifying feature is the size of the renormalized , purely electromagnetic model electron , whose radius comes out to @xmath13 physical electron compton lengths , about @xmath14 times larger than the somewhat misnamed `` classical electron radius '' @xmath15 .
our renormalized led is thus intrinsically equipped with the ultraviolet cutoff at which heuristic discussions @xcite have traditionally set the spatial limit of applicability of classical electrodynamics as a physical theory .
remarkable is furthermore that the cauchy problem for the state evolution bears a certain resemblance to the problem of motion of a black hole in general relativity .
unexpectedly , massive led thus provides a playground for this much harder problem of general relativity .
our model owes much to the monumental work of nodvik @xcite .
although in nodvik s lorentz - covariant massive led one can match the particle parameters to the empirical electron data for charge , mass , and magnetic moment without encountering superluminal gyration speeds , one can do this only for sufficiently large bare rest mass .
more specifically , the renormalization flow toward smaller bare rest mass terminates at about @xmath16 of the empirical electron mass , i.e. before reaching the purely electromagnetic limit of vanishing bare rest mass , and its spin is then merely about @xmath17 .
in addition to this deficiency , nodvik s massive led suffers from a serious dynamical defect which it shares with the purely electromagnetic models of abraham and lorentz ; namely , the initial value problem is singular .
interestingly , both deficiencies of nodvik s massive led are caused by his choice of bare rest mass distribution , which concentrates all the bare inertia at the center of the particle so that its bare moment of inertia vanishes . by endowing the particle with a bare rest mass distribution that gives it a strictly positive bare moment of inertia
our massive led overcomes both deficiencies and , in this sense , improves decisively on nodvik s massive led .
the presentation of our results is organized as follows . to begin with , in section 2
we stipulate the relativistic notation used in this paper . in section 3
we discuss the relativistic kinematical prerequisites of the model .
massive led is defined by a lorentz invariant action functional that ` minimally couples ' the dynamics of the bare particle and the electromagnetic fields .
we first formulate the relativistic mechanics of a gyrating bare particle for given minkowski force and torque of unspecified origin ( section 4 ) and then , in section 5 , present nodvik s @xcite covariant maxwell
lorentz equations which determine the electromagnetic fields also in our model , given the kinematical history of the charged particle .
our action functional for massive led is introduced in section 6 , where we also discuss its symmetries and associated conservation laws .
the dynamical equations of our massive led are presented in section 7 . in section 8
we address the cauchy problem for the state evolution .
the scattering problem is discussed in section 9 , where we will see that the renormalized particle behaves like a soliton . in section 10 the renormalization group flow toward vanishing bare rest mass is studied , in detail for the stationary particle , and in an outline for the dynamical particle .
we conclude the main body of our paper in section 11 with a summary and an outlook on follow - up work and open problems .
in particular , we also speculate about possible implications of our results in the quest for a consistent qed . in order not to interrupt the flow of the presentation , some technical material adapted from @xcite ,
e.g. the derivations of the euler - lagrange equations , has been relegated to appendix a.1 . to facilitate the comparison with nodvik s model @xcite
, appendix a.2 contains a brief discussion of the point mass limit in which our massive led formally reduces to the massive led of nodvik @xcite . in particular , we explain that this limit is singular and severely overdetermines the initial value problem .
incidentally , a similar assessment holds for abraham s @xcite purely electromagnetic versions of semi - relativistic led with and without spin .
since the ill - posedness of the initial value problem for the purely electromagnetic abraham model is itself a new observation , appendix a.3 gives a concise summary of the semi - relativistic massive led and its purely electromagnetic , singular limit .
there we start with the massive model with spin and show next that the more familiar semi - relativistic massive ` led without spin ' actually is a theory with spin in ` hidden ' form , obtained from the massive model with spin in a non - singular limit .
we then address the singular purely electromagnetic abraham limit .
a fair list of references has been selected from the vast pertinent literature .
various equivalent conventions for the minkowski space formalism exist in the literature . by and large ,
we here follow the conventions of the book by misner , thorne and wheeler @xcite
. however , since the model is essentially a flat space model , we will not introduce forms . instead , throughout this work we use four - vectors and tensors constructed from them .
thus , abstract minkowski space @xmath18 is a flat four - dimensional pseudo - riemannian manifold with lorentzian metric of signature @xmath19 .
the invariance group acting on minkowski space is the inhomogeneous _ poincar group _ ; its homogeneous subgroup is called the _ lorentz group _ ; for good discussions , see @xcite .
we can choose any point @xmath20 and identify @xmath18 with the tangent space @xmath21 at @xmath22 .
a tetrad @xmath23 of fixed unit four - vectors in @xmath21 that satisfy the elementary inner product rules @xmath24 is a basis for @xmath21 , called a _
lorentz frame_. with respect to a lorentz frame , any real - valued vector @xmath25 has the representation @xmath26 with @xmath27 given by @xmath28 and by @xmath29 for @xmath30 .
therefore , @xmath21 can be identified with @xmath31 , the set of ordered real @xmath32-tuples @xmath33 , which is a global coordinate chart for @xmath18 . to emphasize _ spacetime _ among the various physical realizations of abstract minkowski space , spacetime points
will be denoted by @xmath34 and called events , and their four - vector representation w.r.t .
@xmath35 will be denoted as @xmath36 or @xmath37 etc .
the coordinates of @xmath36 in a lorentz frame @xmath35 are sometimes grouped as @xmath38 where @xmath39 is called a ` point in space , ' and @xmath40 an ` instant of time , ' with respect to the given lorentz frame . here , @xmath3 is the speed of light _ in vacuo_. event
four - vectors @xmath41 are classified into spacelike , lightlike , and timelike according as @xmath42 , @xmath43 , or @xmath44 , respectively , where the inner product of any two event four - vectors @xmath41 and @xmath45 is evaluated using ( [ eq : qvte ] ) , distributivity and the elementary inner product rules ( [ eq : ipr ] ) . a spacelike four - vector
is connected through an orbit of the lorentz group with a four - vector of the form @xmath46 , a timelike one with a four - vector of the form @xmath47 .
the lightlike vectors form the ` light cone , ' a double cone in @xmath31 which separates the spacelike from the timelike four - vectors .
this classification is then carried over to four - vectors @xmath48 in abstract minkowski space . for convenience
, @xmath49 will sometimes be abbreviated thus , @xmath50 where @xmath51 is the minkowski ` norm ' of @xmath48 , defined as the principal value of @xmath52 .
notice that @xmath53 is negative for timelike vectors .
the tensor product between any two four - vectors @xmath54 and @xmath55 is a tensor of rank two , denoted by @xmath56 , and defined by its inner - product action on four - vectors , thus @xmath57 ( notice that our convention here differs from the one used , for instance , in @xcite . )
given a frame @xmath58 , any tensor of rank two , @xmath59 , can be uniquely written as @xmath60 a rank - two tensor @xmath59 for which @xmath61 in ( [ eq : gent ] ) is called _ symmetric _ ( @xmath62 sign ) , respectively _ anti - symmetric _
( @xmath63 sign ) . for symmetric ( @xmath62 ) and anti - symmetric ( @xmath63 ) tensors we have @xmath64 but for a general tensor @xmath65 a special example for an anti - symmetric tensor of rank two is the exterior product between two four - vectors , denoted by a wedge - up product @xmath66 , and defined by @xmath67 examples for symmetric tensors of rank two are , ( i ) the symmetrized tensor product between two four - vectors , denoted by a wedge - down product @xmath68 , and defined by @xmath69 and ( ii ) the _ metric tensor _ @xmath70 with @xmath71 given in ( [ eq : ipr ] ) .
notice that @xmath72 has the same components @xmath73 in all lorentz frames because the @xmath74 are lorentz scalars .
notice in particular also that @xmath72 acts as identity on four - vectors , i.e. @xmath75 . the _ ( anti-)commutator _ of any two tensors of rank two , @xmath76 and @xmath77 is defined as usual , @xmath78_\pm{\stackrel{\textrm{\tiny def}}{=}}{{\textbf{\textsf{a}}}}\cdot{{\textbf{\textsf{b } } } } \pm { { \textbf{\textsf{b}}}}\cdot{{\textbf{\textsf{a } } } } \ , .\ ] ] the action of tensors on a vector is associative , i.e. @xmath79 so that @xmath78_\pm\cdot{{{\textbf{\textsl{v}}}}}= { { \textbf{\textsf{a}}}}\cdot({{\textbf{\textsf{b}}}}\cdot{{{\textbf{\textsl{v } } } } } ) \pm { { \textbf{\textsf{b}}}}\cdot({{\textbf{\textsf{a}}}}\cdot{{{\textbf{\textsl{v } } } } } ) \ , .\ ] ] _ warning : _ in general , @xmath80 the four - trace , or contraction , of a tensor of rank two is given by @xmath81 in particular , @xmath82 , while @xmath83 for any anti - symmetric tensor @xmath84 .
for a differentiable function @xmath85 we denote its four - gradient with respect to the metric @xmath86 by @xmath87 . in local lorentz frame
coordinates it is given by @xmath88 where @xmath89 is the usual three - gradient . in analogy with the conventional curl ,
we also define the four - curl of a differentiable four - vector function as the anti - symmetric four tensor function @xmath90 where the @xmath91 are the entries of the conventional rank - four levi - civita tensor .
the four - laplacian with respect to @xmath86 is just the ( negative of the ) dalembertian , or wave operator , given by @xmath92
a spherically extended , ` rigidly ' gyrating particle in spacetime is kinematically described by a pair of curves in minkowski space .
we shall assume that both curves are of class @xmath93 .
one curve is the so - called _ world - line _ of the particle , a timelike future - oriented curve in spacetime itself .
the concept of the world - line is familiar from the relativistic mechanics of point particles .
for an extended ` rigid ' particle an event on its world - line marks the location of the particle s center in spacetime . the world - line has associated with it an invariant ` arc - length ' element @xmath94 . in units
such that the speed of light @xmath95 , which henceforth we shall use , this invariant arc - length element becomes the _ proper - time _ element .
choosing any particular event @xmath96 on the world - line as reference world point and assigning to it the proper - time @xmath97 , and stipulating the ` future ' direction as the one along which @xmath94 increases , then by integrating @xmath94 along the world line automatically determines a unique invariant proper - time @xmath98 for any other event @xmath34 on the world - line .
since the world - line is timelike future - oriented , its proper - time assignment is one - to - one onto and order - preserving , whence serves as a natural parameterization for the world - line . in any particular lorentz frame
@xmath35 , the particle world - line is then given by a mapping @xmath99 .
the proper - time element accordingly is given by @xmath100 , where @xmath101 is taken along the world - line . with this parameterization of the world - line
, the four - velocity of the particle at the event @xmath102 is obtained as @xmath103 .
the map @xmath104 , called ` four - hodograph , ' traces out a complicated curve in _ four - velocity space _
, located on the future unit hyperboloid which is the genus - zero three - dimensional sub - manifold defined by @xmath105 and @xmath106 . when @xmath107 is constant , the four - hodograph degenerates into a minkowski point .
we will sometimes group the four - velocity into its time and space components w.r.t .
the chosen lorentz frame , as @xmath108 where @xmath109 is the usual three - velocity in the space part of the chosen lorentz frame , and where @xmath110 finally , by taking the second derivative w.r.t .
@xmath111 we obtain the four - acceleration of the particle at the event represented by @xmath102 , thus @xmath112 .
the map @xmath113 is a section in the tangent bundle to the future unit hyperboloid of four - velocity space .
the world - line does not contain any kinematical information about the gyration of the extended rigid particle .
that information is contained in what may be called the _
gyrograph _ of the particle , a spacelike - oriented curve in four angular velocity space , or _
gyrospace _ for short .
the introduction of gyrospace and the gyrograph map requires the consideration of two non - inertial reference frames : * the body frame @xmath114 ; * the fermi
walker frame @xmath115 .
the body frame @xmath114 is a co - moving and co - rotating frame , fixed in the particle .
it is given by a tetrad of orthonormal basis vectors @xmath116 , with @xmath117 , that satisfy the equation @xcite @xmath118 where @xmath119 is the anti - symmetric tensor of the instantaneous rate of four - gyration of the particle s body frame @xmath114 as seen in the lorentz frame @xmath35 .
the components of four - vectors with respect to @xmath114 will be marked by primes : @xmath120 .
the fermi
walker frame @xmath115 is the natural co - moving tetrad of orthonormal basis vectors @xmath121 , with @xmath122 , satisfying the law of _ fermi walker transport _
@xcite , @xmath123 where @xmath124 is the anti - symmetric fermi walker tensor .
the components of vectors with respect to @xmath115 will be marked by overbars , i.e. @xmath125 . fermi walker transport is a kinematical effect which does not discriminate between objects of different physical characteristics like inertia or charge . the most prominent effect associated with fermi
walker transport is the celebrated _ thomas precession _
@xcite of the space axes of @xmath115 relative to the lorentz frame @xmath35 , described by the angular velocity three - vector @xcite , @xmath126 here @xmath127 is the usual three - acceleration of the origin of @xmath115 as seen from @xmath35 , and @xmath109 its three - velocity .
if the four - gyration of the body frame w.r.t .
the lorentz frame differs from the thomas precession , one ascribes this difference to the particle s _ intrinsic four - gyration _ , which is captured by the anti - symmetric tensor @xmath128 the subscript ` @xmath129 ' here is meant as reference to euler s monumental analysis of newtonian gyration of an inert rigid body . beside being anti - symmetric , the tensor @xmath130 satisfies @xmath131 so that @xmath130 has only three independent components .
it follows that @xmath130 is dual to a spacelike four - vector @xmath132 which satisfies @xmath133 and @xmath134 . in @xmath115
, @xmath132 is given in the form @xmath135 , where @xmath136 is the usual angular velocity three - vector which is directed along the instantaneous axis of body gyration in the spacelike three - slice of the frame @xmath115 . at each instant of proper - time @xmath111
, the instantaneous state of intrinsic four - gyration of the particle is captured by the four - vector @xmath137 .
as @xmath138 varies along the particle s world - line , the map @xmath139 traces out the gyrograph of the particle in gyrospace , which is the minkowski space consisting of all possible four - vectors with the meaning of an @xmath132 .
the gyrograph is spacelike , since for @xmath140 , we have @xmath141
formally stripping the particle from its electromagnetic features leaves one with the _
bare particle_. our bare particle is characterized by a strictly positive inertia against changes of its translational and gyrational momenta , respectively . in our model , all forms of bare inertia
are computed with a compactly supported , spherical ` density ' @xmath142 of inert mass of unspecified non - electromagnetic origin . in general , @xmath143 is a measure on @xmath144 $ ] .
we demand that @xmath145 generates a strictly positive moment of inertia ( see below ) .
for instance , two admissible mass densities are : * _ ( volume inertia ) _ @xmath146 , * _ _ ( surface inertia)__@xmath147 . here , @xmath148 is the dirac distribution , and @xmath149 the characteristic function of the interval @xmath144 $ ] with @xmath150 .
a fraction of the mass in @xmath145 is allowed to be a point mass , but not all of it .
notice that the complete point mass density @xmath151 violates our condition that @xmath145 must have a strictly positive moment of inertia .
the singular limiting case of a point mass density is addressed in appendix a.2 .
the most basic inertia is the _ bare rest mass _ , given by @xmath152 where @xmath153 . since in general the bare particle
is in a state of gyration , its instantaneous effective inert mass in the fermi
walker frame is the sum of its bare rest mass and the relativistic kinetic energy associated with its euler gyration , which we call the _ bare gyrational mass _ , a lorentz scalar given by @xmath154 the ` integration over a simultaneity space slice ' in ( [ eq : gyromass ] ) follows the insight of fermi @xcite , who used this setup in his solution to the infamous @xmath155 problem ( see @xcite for discussions of the @xmath155 problem ) . in the limit of small inertial angular velocity , @xmath156 , a maclaurin expansion of ( [ eq : gyromasseval ] ) in powers of @xmath157 gives the expected result @xmath158 where @xmath159 is the bare particle s non - relativistic principal moment of inertia . by the spherical symmetry of the particle in any of its rest frames the principal inertias are of course degenerate .
the corresponding non - relativistic moment - of - inertia three - tensor of the bare particle generalizes to the tensor of the _ minkowski inertia ( about @xmath160 ) _ , given by @xmath161 which captures the inertia associated with changes of the state of intrinsic four - gyration of the bare particle .
the various inertias acquire their meaning from their role as ` coefficients of resistance ' against changes of the bare particle s various minkowski momenta .
the bare particle s four - vector of _ minkowski momentum _ is given by @xmath162 and the four - vector of its _ minkowski spin _ by @xmath163 for a given minkowski force @xmath164 and minkowski torque @xmath165 , here of unspecified origin , but satisfying @xmath166 and @xmath167 , the evolution of these momenta is governed by the coupled dynamical equations @xmath168 and @xmath169 due to the occurrence of @xmath102 in the gyrational mass , these dynamical equations are of second order for @xmath160 and of first order for @xmath132 , so they need to be supplemented with initial data for @xmath160 , @xmath170 , and @xmath132 .
a discussion of these equations is interestingly in itself , but will not be pursued here .
if the charged particle s world - line @xmath102 and gyrograph @xmath171 are given , the anti - symmetric electromagnetic tensor field of rank two @xmath172 can be computed by solving the linear , relativistic maxwell lorentz equations @xmath173 @xmath174 where @xmath175 is the ( left ) hodge dual of @xmath172 , and where the covariant charge - current density four - vector field @xmath176 is of the general form ` charge density times four - velocity . '
the specific form of the charge - current density of course depends on the particle model under consideration .
our massive led inherits the charge - current density from nodvik s @xcite model of a gyrating lorentz electron , which generalizes the non - relativistic expression for a gyrating , strictly rigid abraham electron ( appendix a.3 ) to the manifestly covariant expression @xmath177 here , @xmath178 is the four - velocity of a particle element located at @xmath41 in @xmath35 , with @xmath179 , and @xmath180 is the spherically symmetric electrical charge ` density ' ( in general a finite measure ) of the particle at rest . to avoid infinite electromagnetic energy densities and forces , we request that @xmath181 does not feature any singular dirac point charge distribution anywhere on its support ; more general dirac distributions are however allowed . for instance , the following two possibilities are admissible : * _ _ ( volume charge)__@xmath182 , * _ _ ( surface charge)__@xmath183 . here , @xmath184 is the elementary charge as appropriate for a model of an electron
. problems with infinite self energies and self forces aside , however , it should be noticed that in the formal limit @xmath185 , together with the assumption that @xmath186 , the nodvik charge - current density ( [ eq : jnodvik ] ) reduces to the familiar expression for the spinless point charge @xcite , @xmath187
in order to obtain the dynamical equations of massive led , we have to couple the relativistic maxwell lorentz equations for the electromagnetic fields and the relativistic equations for the particle s world - line and gyrograph in a consistent manner . since both the dynamics of the bare particle and the dynamics of the maxwell lorentz field derive separately from some _ principle of ` least ' action _ , we can follow nodvik @xcite and apply the standard procedure of minimal coupling to the action functionals and then postulate the principle of least action for this new , joint action . as a general reference for relativistic action principles we recommend @xcite .
the action is defined through an integral over any four - domain @xmath188 which is sandwiched between two disjoint spacelike three - subspaces that contain the intersections of the supports of @xmath189 ( and of @xmath190 ) with the spacelike slices @xmath191 of events which are simultaneous to the events at @xmath102 for @xmath192 and @xmath193 , respectively .
such a four - domain is called ` admissible . '
it contains _ ordered particle histories _ , defined by particle world - lines which connect the supports of @xmath189 in the slices @xmath191 at @xmath192 and @xmath193 in such a way that successive space slices of simultaneity do not intersect in the support of @xmath189 ; see figure [ fig : slice1 ] .
not every world - line corresponds to an ordered history .
the space slices of simultaneity for two different values of @xmath111 may intersect in the support @xmath194 if the particle acceleration is too strong ; see figure [ fig : slice2 ] . to formulate our minimally coupled action functional
, we also need to introduce the four - vector of the electromagnetic potential .
we recall that ( [ eq : hommleq ] ) implies that @xmath172 is the exterior derivative of an electromagnetic potential four - vector @xmath195 , i.e. @xmath196 .
we will work within the class of @xmath197 that satisfy the so - called lorentz gauge condition @xmath198 , which still allows the gauge transformation @xmath199 for any @xmath200 satisfying the wave equation @xmath201 .
given an admissible @xmath188 , the action functional @xmath202 is given as integral @xmath203 where @xmath204 is the lagrangian four - density for the minimally coupled field - particle system , @xmath205 the first term is schwarzschild s lagrangian density for the free electromagnetic field of maxwell , i.e. @xmath206 and the second term is the standard lagrangian density for the minimal electromagnetic coupling of particle and field degrees of freedom , i.e. @xmath207 where @xmath176 is the nodvik charge - current density given in ( [ eq : jnodvik ] ) .
the last term is the lagrangian density for our bare particle , @xmath208 our action functional is invariant under several symmetry transformations of base and target spaces , associated with which are invariants that give the physical conservation laws .
these invariants are listed below .
we consider only ordered histories , so that the domain of integration in the conservation laws can be written @xmath209 instead of @xmath188 .
invariance of the action functional under translations of spacetime implies conservation of the _ total minkowski momentum _ four - vector @xmath210 of the field - particle system , the constant time component of which expresses energy conservation , the space components the conservation of linear three - momentum .
the minkowski momentum of the field - particle system can be written in the form @xmath211 where @xmath212 is the symmetric field - particle tensor mass , itself composed of the following tensor masses : the first term on its r.h.s . is just the bare gyrational mass of the particle written as diagonal tensor with the help of the metric tensor @xmath72 ; the second term on its r.h.s is the symmetric _ nodvik tensor mass _ , @xmath213_+ \right]_+ { f_{\text{e}}}\big({\left\| { { { \textbf{\textsl{x}}}}}\right\| } \big)\ , \delta({{{\textbf{\textsl{u}}}}}(\tau ) \cdot { { { \textbf{\textsl{x}}}}})\,{\mathrm{d}}^4{{{\textbf{\textsl{x}}}}}\ , , \label{eq : spinorbitmass}\ ] ] extracted ( see appendix a.1 ) from the minkowski momentum four - vector of electromagnetic spin - orbit coupling given in @xcite ; the last term on its r.h.s . finally is the electromagnetic field mass tensor , with @xmath214 the conventional traceless tensor of the electromagnetic energy - momentum - stress density , a symmetric tensor field of rank two .
it s four - divergence , given by @xmath215 does in general not vanish on the support of the particle densities @xmath145 and @xmath181 ; in a nutshell , this is the origin for the need of the fermi prescription of integration over space slices of simultaneity ( cf .
@xcite ) .
invariance of the action functional under spacetime rotations implies conservation of the anti - symmetric tensor of _ total minkowski angular momentum _
@xmath216 of the field - particle system , @xmath217 where @xmath210 and @xmath59 are given above , and where @xmath218 is the anti - symmetric tensor of _ bare minkowski spin ( about @xmath160 ) _ of the particle .
the tensor @xmath219 is dual to the bare spin four - vector ( [ eq : barespinvect ] ) , hence it is of space - space type w.r.t .
@xmath170 , viz .
@xmath220 since @xmath216 is anti - symmetric , its constancy expresses six individual physical conservation laws .
the constancy of the three space - space components expresses the conservation of the total angular momentum , and the constancy of the three time - space components correspond to the statement that the center of mass of the field - particle system moves with constant three - velocity along a straight line in the three - space slice of the lorentz frame .
invariance of the action functional under gauge transformations @xmath199 implies charge conservation locally , i.e. ( [ eq : jnodvik ] ) satisfies the continuity equation @xmath221 which is demanded by the maxwell
lorentz equations and readily verified to hold also by direct computation with ( [ eq : jnodvik ] ) , cf . @xcite . of course
, ( [ eq : continuityeq ] ) implies the conservation of the total _ charge _
@xmath222 i.e. @xmath223 in our model of the electron , for all @xmath224 .
in addition , there is a symmetry associated with the spherical rest symmetry of the particle , giving invariance of the action functional under shift of the euler angle around the instantaneous axis of gyration .
( for the definition of the euler angles , see appendix a.1 . )
this corresponds to the conservation of @xmath225 where @xmath226_+\ ] ] is the space - space part w.r.t .
@xmath170 of the anti - symmetric tensor of _ total minkowski spin _ , @xmath227 with @xmath228 the tensor of the particle s _ electromagnetic minkowski spin _ about @xmath102 .
we remark that despite its appearance , @xmath229 is invariant under a gauge transformation @xmath199 .
the law for @xmath230 means that the dual four - vector to @xmath231 evolves on the spacelike genus - one hyperboloid .
our action principle is formulated for a fixed four - domain @xmath188 , with the constraints @xmath232 for @xmath233 and @xmath193 .
likewise , the variations of the world - line @xmath234 and of the euler angles ( see appendix a.1 ) vanish at @xmath233 and @xmath193 . following standard physics terminology
we refer to this hamiltonian principle of stationary action as ` principle of least action , ' although we do not investigate the second variation here .
the particle world - line satisfies the variational principle @xmath235 under an independent deformation of the world - line @xmath236 ( with @xmath237 $ ] ) into a world - line @xmath238 ( with @xmath111 in the same range ) , with euler variables @xmath239 and four - potential @xmath197 fixed .
the gyrograph satisfies the variational principle @xmath240 under an independent variation of the gyrograph of the particle , with fixed world - line @xmath241 and four - potential @xmath197 .
the fields satisfy the standard schwarzschild principle @xcite , @xmath242 for prescribed world - line @xmath241 and gyrograph @xmath243 .
the dynamical equations are the euler - lagrange equations for the variational principle @xmath244 with @xmath245 and the euler angles @xmath239 treated as independent variables .
we will state the relativistic dynamical equations in a format closely reminiscent of the semi - relativistic equations of the massive abraham theory described in appendix a.3 . recalling that @xmath246 by definition , which already implies the homogeneous maxwell lorentz equations ( [ eq : hommleq ] ) , the principle of least action together with @xmath247 now yields the equivalent potential form of the inhomogeneous maxwell
lorentz equations , i.e. the inhomogeneous wave equation @xmath248 with @xmath176 nodvik s four - vector of charge - current density given in ( [ eq : jnodvik ] ) .
the world - line equations read @xmath249 where @xmath250 is the _ minkowski momentum _ of the particle , with @xmath251 ( see ( [ eq : smmt ] ) ) , and where @xmath252 is nodvik s
@xcite abraham lorentz type _
minkowski force_. the equations for the gyrograph read @xmath253_- = { { { \textbf{\textsf{t}}}}}\ , , \label{eq : ggeq}\ ] ] where @xmath254 is the anti - symmetric tensor of the abraham lorentz type _
minkowski torque _
, with @xmath255 the projection of the four - vector @xmath256 onto the space slice w.r.t .
the space projector @xmath257 under the integral guarantees that ( [ eq : ggeq ] ) preserves the intrinsic space - space character of @xmath219 . the commutator with the fermi
walker tensor accounts for the four - precession of the fermi walker - transported , natural co - moving frame with respect to which the inertial gyrations are defined .
in the concise component - free minkowski space notation the similarity of the world - line equation ( [ eq : wleq ] ) , supplemented with ( [ eq : minkenmom ] ) and ( [ eq : minkalforce ] ) , to the newtonian type evolution equation ( [ eq : aleqtmb ] ) of the semi - relativistic massive led with spin is clearly recognizable , and so is the similarity of the gyrograph equation ( [ eq : ggeq ] ) , supplemented with ( [ eq : minkaltorque ] ) , to the eulerian evolution equation ( [ eq : aleqrmb ] ) .
the almost familiar appearance of the dynamical equations of our relativistic massive led is , however , deceptive .
as we will now see , the physically natural interpretation of these dynamical equations as a set of evolution equations for the _ dynamical state _ of the theory unveils some radically new elements of physical reality ( as defined by the theory ) .
the characteristic of a physical state , loosely speaking , is that its specification at any particular instant completely determines , via first - order evolution equations , the time derivative of the state at that instant and , hence , allows to continue uniquely to the state at the next instant . to find out what precisely constitutes a ` state at any particular instant ' in our massive led
, we have to analyze for which dynamical variables the equations pose a first - order initial value problem ( cauchy problem ) in an orderly manner .
this is done in the following two subsections .
we then address the problem of nearly free global evolutions .
since the evolution equations of our massive led are implicit in a rather unfamiliar manner , we first rewrite them into a quasi - explicit format .
we then propose a direct generalization of the familiar picard iteration to solve the equations in principle .
we finally remark on the question of global well - posedness of the cauchy problem for arbitrary admissible states .
this issue is only partially settled .
various possible ` mechanisms ' that may cause some sort of singularity to occur at a finite time have to be controlled .
however , for the special case of purely gyrational particle dynamics coupled to the maxwell lorentz fields we succeeded in proving , by using our generalized picard iteration , a global existence and uniqueness result .
this proof is going to appear in @xcite .
we first need to clarify the phrase _ at an instant of time_. in special relativity , this is a quite flexible notion .
generally it refers to the collection of all the events on a three - dimensional , simply connected , spacelike hypersurface .
we assume without loss of generality that such a hypersurface is defined as the level set @xmath258 of a differentiable function @xmath259 on spacetime . here , a three - dimensional hypersurface @xmath260 is called spacelike if @xmath261 is timelike at each @xmath41 in @xmath260 . if the intersection of any two such spacelike hypersurfaces @xmath262 and @xmath263 , with @xmath264 , is empty , and if furthermore the four - dimensional set @xmath265 covers spacetime , @xmath265 is called a _ foliation _ of spacetime ( generated by @xmath259 ) , and each three - dimensional hypersurface @xmath260 is called a _ leaf _ of that foliation .
for example , the standard foliation of a lorentz frame @xmath266 into affine space slices of the form @xmath267 is generated by the function @xmath268 , which has a constant timelike four - gradient , @xmath269 .
there are uncountably many foliations of spacetime among which one has to find any one that is _ compatible _ with the structure of the dynamical equations , preferably a convenient one .
unfortunately , as we shall see in a moment , the standard foliation is not compatible with the dynamical equations of massive led . in any event ,
a ` state at an instant @xmath270 ' thus means the specification of the complete set of dynamical variables on the leaf with index @xmath270 of a compatible foliation of spacetime . to find out what constitutes such a complete set of dynamical variables , we notice that our dynamical equations are of first order in the proper - time derivative of each of the kinematical particle variables @xmath160 , @xmath170 , @xmath130 , and of first order in the lorentz time derivative of the fields @xmath172 . a dynamical state in
our massive led will therefore be a set of the form @xmath271 , specified on a leaf @xmath260 of a compatible foliation of spacetime indexed by @xmath270 , satisfying various consistency conditions that constrain the values of @xmath272 at @xmath111 and the field @xmath273 on the leaf .
in particular , the four - velocity has to satisfy @xmath274 , and the anti - symmetric rank - two euler tensor has to satisfy @xmath275 .
furthermore , the maxwell
lorentz equations impose the constraint that @xmath172 satisfy the three - divergence part of the maxwell
lorentz equations , which fixes the longitudinal part of @xmath172 as a passive attribute of the charged particle .
these are all _ local constraints _ imposed by relativity , which occur also in the traditional limiting theories , i.e. the relativistic point charge mechanics in given external force fields , the mechanics of a point spin in given external torque fields , and the maxwell electrodynamics with given source terms , in all of which cases the constraints are known to propagate in time if they are satisfied initially
. this will be true also for our massive lorentz electrodynamics .
however , our covariant first - order dynamical equations are also mildly _ non - local _ , and this is going to be reflected in the compatibility conditions to be satisfied by the foliation of spacetime .
namely , since the first - order proper - time derivatives of @xmath170 and @xmath130 at proper - time @xmath111 are given ( implicitly ) in terms of integrals of the fields over the particle s support in its instantaneous space slice of simultaneity , the transversal part ( i.e. the dynamical degrees of freedom ) of @xmath276 must be known for all @xmath277 satisfying @xmath278 ( recall that the passive longitudinal part of the field @xmath276 is not a dynamical degree of freedom ) .
hence , we say that a foliation of spacetime is _ compatible _ with a world - line @xmath279 of an extended lorentz electron ( i.e. , a particle history ) if each leaf of the foliation contains precisely one of the sets @xmath280 ( for simplicity , we assume that @xmath281 )
. a compatible foliation sets the spacetime parameterization w.r.t . which the maxwell
lorentz equations must be solved in order to enable the evaluation of the minkowski force and torque .
a compatible foliation is of course not known explicitly until the particle world - line is known ; it has to be determined along with the particle world - line and the fields by solving the second variation equations for the extrinsic curvature of the leaf @xmath260 coupled with the constraint equations of gauss codacci and gauss in the complement of the sets @xmath277 satisfying @xmath278 , which is a non - linear free boundary problem .
interestingly , these circumstances inject technical elements from general relativity into this special relativistic theory . _
remark : _ no foliation is compatible with a particle history that is not ordered in the sense described earlier ( see fig.1 and 2 ) .
the condition that the particle history be ordered imposes size restrictions on the accelerating fields .
indeed , we easily find locally consistent data for which the particle s three - acceleration in its instantaneous lorentz rest frame exceeds the value @xmath282 , where @xmath283 is the radius of the particle .
for such accelerations , part of the support of the particle in its instantaneous space slice of simultaneity will be launched _
backward in lorentz time _ rather than forward , and this motion does not generate an ordered history .
this implies the following bound on the particle s initial three - acceleration in the lorentz frame in which the particle is instantaneously at rest , @xmath284 which has to be translated into bounds on the accelerating field strengths in the instantaneous particle rest frame .
we will soon see that these bounds on the field strengths are about @xmath285 times the particle s self - fields at it s surface which , happily , is way outside the physically reasonable range of data for a classical electrodynamics .
last not least , the condition that the gyrational speed for points on the particle s equator must be subluminal imposes a size restriction on @xmath130 , viz .
( recall that @xmath95 ) @xmath286 we now formulate the initial value problem for the states .
we begin by specifying the generic initial data on a convenient initial leaf .
the compatible foliation itself is characterized subsequently by dynamical equations which together with the equations of massive led ( rewritten w.r.t .
the compatible foliation ) pose a nonlinear , implicit cauchy problem for the joint evolution of the state and the leaf of the foliation on which the state is defined . since by a suitable spacetime translation we can always transform to a lorentz frame
in which the event of the particle s initial spacetime position is at the origin of that lorentz frame , and since by a suitable lorentz boost we can furthermore transform to another lorentz frame with the same origin but in which the particle s initial three - velocity vanishes , and since finally we can choose the proper - time of the particle equal to zero at that initial event , the cauchy data for the world - line @xmath236 can be chosen conveniently as @xmath287 notice that @xmath288 satisfies the constraint @xmath289 required of a four - velocity .
moreover , by at most a space rotation of this lorentz frame we can achieve that @xmath290 , the dual vector to @xmath291 , has only a single non - vanishing component , say the @xmath292 component , @xmath293 , with @xmath294 satisfying ( [ eq : gyrosubc ] ) but otherwise at our disposal .
the cauchy data for the particle s gyrograph @xmath295 are then of the form @xmath296 clearly , @xmath291 is anti - symmetric , and since @xmath297 , also of space - space type w.r.t . @xmath298 .
the lorentz frame in which ( [ eq : eventinitcond ] ) , ( [ eq : hodoinitcond ] ) , and ( [ eq : gyroinitcond ] ) holds will be referred to as the _ initial rest frame _ @xmath299 .
the initial leaf @xmath300 of all of our foliations is now chosen to be the particle s affine space slice of simultaneity in the initial rest frame @xmath299 , i.e. the set @xmath301 on which the lorentz time @xmath302 .
accordingly , the cauchy data for @xmath276 are specified on @xmath300 , @xmath303 where @xmath304 satisfies the three - divergence equations in the maxwell
lorentz equations restricted to @xmath300 . to make this somewhat implicit characterization of @xmath305 more explicit , we decompose @xmath305 into its electric and magnetic components w.r.t .
@xmath299 and group them together into a complex electromagnetic three - vector field on @xmath300 , @xmath306 whose real and imaginary part are , respectively , the electric ( i.e. time - space ) and magnetic ( i.e. space - space ) components of the field tensor @xmath305 in @xmath300 .
the three - divergence equations in @xmath300 now read @xmath307 where @xmath308 is the time component of @xmath309 , uniquely determined in terms of the initial values of the other state variables @xmath310 and @xmath311 .
_ remark : _ through the presence of @xmath312 in the right - hand side of ( [ eq : jnodvik ] ) the ( initial ) constraint appears to be implicit .
however , by direct computation it is readily verified that the @xmath312 contribution to @xmath176 drops out initially as a consequence of the sphericity of the charge distribution in the particle s instantaneous rest frame , which initially coincides with the lorentz frame @xmath299 .
the initial fields are chosen in accordance with the asymptotic condition that @xmath313 as @xmath314 , the real part as @xmath315 , the imaginary part sufficiently fast for surface integrals at infinity involving @xmath316 to vanish .
to each differentiable particle world - line starting with ( [ eq : eventinitcond ] ) and ( [ eq : hodoinitcond ] ) and generating an ordered particle history ( referred to by an index @xmath317 ) , we have to associate a convenient , compatible foliation of spacetime which contains @xmath300 as initial leaf .
unfortunately , the standard foliation is generally not a compatible foliation , for notice that any world - line is the soul of the tubular set of events @xmath318 , called the future _ world - tube _ of the particle , inside of which the foliation is necessarily given by the @xmath111-indexed affine segments @xmath319 which intersect with the standard leaf of lorentz time @xmath270 but are generally strict subsets of that leaf .
however , one expects that it is possible to select a compatible foliation whose leafs , sufficiently far away from the world - line , will arbitrarily closely approximate the leafs of the standard foliation of the initial lorentz frame , i.e. @xmath300 parallelly transported in lorentz time . by further choosing the @xmath270-parameterization of the compatible foliation to be lorentz time at spatial infinity , by selecting the corresponding value of @xmath111 through the identification @xmath320 , and by parameterizing points in @xmath260 by the points @xmath321 through following the integral curves of @xmath322 , we obtain a foliation which nearly coincides with the standard foliation in most of spacetime , except for a small `` wiggle '' accompanying the world - line of the particle . notice that the world - line is one such integral curve , so that in this parameterization the segment of the world - tube that belongs to @xmath260 is always centered at the origin of @xmath300 .
equipped with such a global parameterization of a compatible foliation , spacetime is diffeomorphic to the product manifold @xmath323 , i.e. @xmath324 , with spacetime metric @xmath325 where @xmath326 is the _ lapse function _ , and the @xmath327 are the components of the _ first fundamental form _ @xmath328 of the foliation induced on @xmath260 from the minkowski metric .
such an extension of the family of affine segments of the future world - tube to the outside of the world - tube is found by solving the structure equations for a foliation of minkowski spacetime , supplemented with standard foliation asymptotics at spatial infinity , and with appropriate boundary conditions along the perimeter of each segment in the tube which guarantee a differentiable transition at the segment s boundary .
the boundary is determined by @xmath102 and @xmath329 ( given @xmath181 , which as a model parameter does not vary during the dynamics ) .
since we have to determine the world - line along with the foliation outside the world - tube , our problem belongs to the category of the dynamical free - boundary problems . notice that even if the world - line is analytic , the foliation is surely not analytic unless the particle is stationary , in which case one can choose the standard foliation of spacetime . by following @xcite
, we conclude that outside the world - tube the foliation is a solution of the system of evolution equations @xmath330 @xmath331 where @xmath332 is the _ second fundamental form _ of the leaf of the foliation , i.e. the extrinsic curvature ( actually defined by ( [ eq : gevol ] ) if @xmath333 were known ) , @xmath334 is the ricci tensor of @xmath260 , and @xmath335 denotes covariant gradient w.r.t .
@xmath328 . equation ( [ eq : kevol ] ) is known as the second variation formula .
in addition , @xmath328 , @xmath200 and @xmath332 have to satisfy the constraints given by the gauss codacci equation @xmath336 and the gauss equation @xmath337 which propagate in time if satisfied on the initial leaf .
this set of equations has to be supplemented by boundary conditions along the perimeter of the world - tube segment for @xmath200 and for @xmath332 ; and by asymptotic conditions as @xmath338 .
we demand that the foliation be _ normal _ , i.e. that @xmath339 as @xmath338 , which means that at spatial infinity the foliation tends to the standard foliation of spacetime .
furthermore , a `` gauge '' condition is at our disposal , for instance the maximality condition @xmath340 the maxwell
lorentz equations for the field tensor @xmath172 have to be rewritten w.r.t .
the foliation in terms of the associated decomposition of @xmath172 into electric and magnetic components ( cf .
@xcite ) , here conveniently grouped together as a complex electromagnetic three - vector field , @xmath341 the initial decomposition being defined already earlier .
the maxwell lorentz evolution equations for @xmath342 then read @xcite @xmath343 where @xmath344 means first - order foliation time derivative ( which equals lorentz time derivative far away from the particle ) , @xmath345 is the curl w.r.t .
@xmath346 , and the electrical current density three - vector @xmath347 is the @xmath260 projection of @xmath176 . clearly , ( [ eq : mlevolveqfoli ] ) is an explicit equation for the foliation time derivative of @xmath342 ; however , note that @xmath348 is a ( vector ) functional not only of all the other state variables @xmath349 but in addition also of @xmath312 . beside the evolution equation ( [ eq : mlevolveqfoli ] )
, @xmath342 has to satisfy the three - divergence equations @xmath350 where @xmath351 , the time component of @xmath176 , is likewise a functional of all the other state variables @xmath349 and in addition also of @xmath312 .
however , ( [ eq : mldiveqfoli ] ) is already included in our set of conditions to be satisfied by a state , for notice that by taking the ( three-)divergence of ( [ eq : mlevolveqfoli ] ) it follows that a solution @xmath352 of ( [ eq : mlevolveqfoli ] ) for @xmath353 given in terms of consistent @xmath102 and @xmath171 , automatically satisfies ( [ eq : mldiveqfoli ] ) for all @xmath354 if it satisfies ( [ eq : mldiveqfoli ] ) initially . the maxwell
lorentz equations are supplemented by the asymptotic condition that @xmath355 as @xmath314 , the real part as @xmath356 , the imaginary part sufficiently fast for surface integrals at infinity involving @xmath357 to vanish .
finally , following @xcite one converts @xmath352 into @xmath172 on a given leaf , which inserted into the world - line and gyrograph equations ( [ eq : wleq ] ) and ( [ eq : ggeq ] ) closes the system of evolution equations for the state in massive led . in summary ,
the cauchy problem for the state evolution turns out to bear a surprising resemblance to the general relativistic evolution problem of a black hole interacting with electromagnetic radiation as described in @xcite , and to the evolution problem of the vacuum einstein equations @xcite .
this gives the equations of massive led an unexpected additional appeal as a somewhat simpler `` toy problem '' for these more formidable type of problems of general relativity . at the same time , the enormous progress recently made in the rigorous treatment of the einstein equations @xcite gives hope that the complete control of the cauchy problem for the state in massive led can be achieved along similar lines of analysis .
we shall have recourse to the notion of the state in massive led in our section about scattering .
in particular , the invariants stated earlier have to be converted into conservation laws for the evolution w.r.t
. a compatible foliation . in the remainder of this section
now we will discuss an alternate , global way of looking at massive led , including an algorithm for the actual computation of solutions to any degree of precision .
if , instead of evolving the state locally in time , we look at the putatively harder problem of global particle histories and field decorations of spacetime , one gains however the advantage that one can work with the standard foliation . the price to pay , of having to solve the global problem , will be somewhat lessened by noting that for all practical purposes of computation , one will need only an approximate solution .
we restrict ourselves to nearly free global future histories and decorations , with data prescribed at @xmath358 on @xmath300 , as before .
nearly free evolutions start from cauchy data which are sufficiently close to a stationary state . for such evolutions ,
the system of fully implicit massive led evolution equations can be converted into an equivalent , regular _ quasi - explicit _ first - order system .
we then explain that this quasi - explicit system of equations can be treated via a generalized picard iteration .. ( a ) a pair of equations @xmath359 , @xmath360 is called an explicit first - order system .
( b ) a pair @xmath361 , @xmath362 which is not also of the form ( a ) is called a regular quasi - explicit first - order system .
thus , in a regular quasi - explicit system each dynamical variable satisfies an explicit equation _ conditioned on the histories of the other variables_. ( c ) a pair @xmath363 , @xmath364 which is not also of the form ( b ) or ( a ) is called a fully implicit first - order system .
the class of fully implicit systems contains the singular quasi - explicit systems , which are of the form @xmath365 , @xmath366 with coefficient functions @xmath367 and @xmath368 that vanish somewhere in the range of their arguments . ]
we now work with the standard foliation , so that ( [ eq : ebfoli ] ) is the decomposition of @xmath172 into electric and magnetic components w.r.t .
this standard foliation .
the maxwell lorentz evolution equations for @xmath342 now read simply as @xmath369 where @xmath344 means first - order lorentz time derivative , @xmath370 is the conventional curl , and the electrical current density three - vector @xmath347 is the space component of @xmath176 .
clearly , ( [ eq : mlevolveqqexpl ] ) is an explicit equation for the lorentz time derivative of @xmath342 .
once again , @xmath348 is a ( vector ) functional not only of all the other state variables @xmath349 but in addition also of @xmath312 . beside the evolution equation ( [ eq : mlevolveqqexpl ] ) , @xmath342 has to satisfy the three - divergence equations @xmath371 where @xmath351 , the time component of @xmath176 , is likewise a functional of all the other state variables @xmath349 and in addition also of @xmath312 . as explained earlier , ( [ eq : mldiveqqexpl ] )
will be satisfied if it is satisfied by a initial state .
the maxwell
lorentz equations are supplemented by the asymptotic condition that @xmath355 as @xmath314 , the real part as @xmath356 , the imaginary part sufficiently fast for surface integrals at infinity involving @xmath357 to vanish .
the gyrograph equation ( [ eq : ggeq ] ) , when viewed as an equation for @xmath130 , does not have the desired format in the first - order proper - time derivative of @xmath130 because @xmath372 involves @xmath373 in a rather complicated manner . however , it is possible and also convenient to view ( [ eq : ggeq ] ) as an equation for @xmath219 . to see that ( [ eq : ggeq ] )
can indeed be viewed as an equation for @xmath219 , we recall that we required @xmath145 to generate a strictly positive moment of inertia @xmath374 . as a consequence ,
the map @xmath375 is strictly positive , increasing , and strictly convex for @xmath376 .
this implies that we can invert @xmath377 to get the euler angular velocity tensor @xmath130 uniquely in terms of the bare spin tensor @xmath219 , viz .
@xmath378 , where @xmath379 . with @xmath130 understood in this way as unique ( vector ) functional of @xmath219 , the minkowski torque @xmath380 given in ( [ eq : minkaltorque ] ) is now a ( tensor ) functional of @xmath381 , so that ( [ eq : ggeq ] ) , as equation for @xmath219 , @xmath382_-\ , + { { { \textbf{\textsf{t}}}}}\ , , \label{eq : gyroeqqexpl}\ ] ] is already in the desired format regarding @xmath372 .
notice that not only @xmath380 but also @xmath383 still depends on @xmath312 .
we now come to the world - line equation ( [ eq : wleq ] ) , which is of second order in the proper - time derivatives of @xmath160 .
since we can treat @xmath160 and @xmath170 as independent degrees of freedom , this second - order equation is equivalent to two first - order equations ; namely , ( [ eq : wleq ] ) , now viewed as equation for @xmath384 , together with @xmath385 for @xmath160 . while ( [ eq : qeqqexpl ] ) already is in the desired format , ( [ eq : wleq ] ) is not .
in fact , @xmath312 enters through @xmath386 in the l.h.s . of ( [ eq : wleq ] ) and through the tensor @xmath387 in the r.h.s . of ( [ eq : wleq ] ) , so that ( [ eq : wleq ] ) is fully implicit first - order no matter whether viewed as equation for @xmath170 or for @xmath388 . switching to @xmath388 as dynamical variable
therefore does not seem to offer an advantage in the conversion of ( [ eq : wleq ] ) into the desired format .
continuing thus to handle ( [ eq : wleq ] ) as an equation for @xmath384 , we first sort out all terms involving @xmath312 . beginning with the r.h.s .
of the world - line equation ( [ eq : wleq ] ) , we notice that @xmath389 can be decomposed into a sum of two terms , one of which is independent of @xmath312 and the other one linear in @xmath312 , @xmath390 the l.h.s . of ( [ eq : wleq ] )
can be decomposed in a similar fashion .
for @xmath391 we have @xmath392 with @xmath393 given in ( [ eq : gyromass ] ) , and with @xmath394_+ \big]_+ { f_{\text{e}}}({\left\| { { { \textbf{\textsl{x}}}}}\right\| } ) \delta({{{\textbf{\textsl{u}}}}}\cdot{{{\textbf{\textsl{x}}}}}){\mathrm{d}}^4{{{\textbf{\textsl{x}}}}}\notag \\ & - \int_{{\mathbb r}^{1,3 } } \big[{{{\textbf{\textsl{x}}}}}{\otimes}{{{\textbf{\textsl{x}}}}},\big[{{{\textbf{\textsl{u}}}}}\cdot{{\boldsymbol { { \nabla } } } _ { \text{g}}}{{{\textbf{\textsf{f}}}}}({{{\textbf{\textsl{q}}}}}+{{{\textbf{\textsl{x } } } } } ) , { \mbox{${\boldsymbol { { \omega } } } _ { \text{\textsc{e}}}$}}\big]_+\big]_+ { f_{\text{e}}}({\left\| { { { \textbf{\textsl{x}}}}}\right\| } ) \delta({{{\textbf{\textsl{u}}}}}\cdot{{{\textbf{\textsl{x}}}}}){\mathrm{d}}^4{{{\textbf{\textsl{x}}}}}\notag \\ & + \int_{{\mathbb r}^{1,3 } } \big[{{{\textbf{\textsl{x}}}}}{\otimes}{{{\textbf{\textsl{x } } } } } , \big[{{{\textbf{\textsf{f}}}}}({{{\textbf{\textsl{q}}}}}+{{{\textbf{\textsl{x}}}}}),{{\mbox{${\boldsymbol { { \omega } } } _ { \text{\textsc{e}}}$}}}\big]_+ \big]_+ { f_{\text{e}}}({\left\| { { { \textbf{\textsl{x}}}}}\right\| } ) ( { { \stackrel{\,\circ}{{{{{\textbf{\textsl{u}}}}}}}}}\cdot{{{\textbf{\textsl{x}}}}}){{{\textbf{\textsl{u}}}}}\cdot{{\boldsymbol { { \nabla } } } _ { \text{g}}}\delta({{{\textbf{\textsl{u}}}}}\cdot{{{\textbf{\textsl{x } } } } } ) { \mathrm{d}}^4{{{\textbf{\textsl{x}}}}}\ , .\end{aligned}\ ] ] the tensor @xmath393 is independent of @xmath312 , and so are the first three terms of @xmath395 , while the last term in @xmath395 is linear in @xmath312 . in conclusion ,
the world - line equation can therefore be written as @xmath396 where the four - force @xmath397 depends only on @xmath160 , @xmath170 , @xmath130 , @xmath373 , @xmath172 , and @xmath398 , thus @xmath399_+\cdot{{{\textbf{\textsl{u}}}}}\big ) { f_{\text{e}}}({\left\| { { { \textbf{\textsl{x}}}}}\right\| } ) \delta({{{\textbf{\textsl{u}}}}}\cdot{{{\textbf{\textsl{x}}}}}){\mathrm{d}}^4{{{\textbf{\textsl{x}}}}}\ , , \end{aligned}\ ] ] and the pseudo - inertia tensor @xmath400 only on @xmath160 , @xmath170 , @xmath130 , @xmath172 , and @xmath398 , thus @xmath401_+ \big]_+ { f_{\text{e}}}\big({\left\| { { { \textbf{\textsl{x}}}}}\right\| }
\big ) \delta({{{\textbf{\textsl{u}}}}}\cdot{{{\textbf{\textsl{x}}}}})\,{\mathrm{d}}^4{{{\textbf{\textsl{x}}}}}\notag \\ & - \int_{{\mathbb r}^{1,3 } } { { { \textbf{\textsl{x}}}}}{\otimes}{{{\textbf{\textsl{x}}}}}\big({{{\textbf{\textsl{x}}}}}\cdot{\mbox{${\boldsymbol { { \omega } } } _ { \text{\textsc{e}}}$}}\cdot { { { \textbf{\textsf{f}}}}}({{{\textbf{\textsl{q}}}}}+{{{\textbf{\textsl{x}}}}})\cdot{{{\textbf{\textsl{u}}}}}\big ) { f_{\text{e}}}\big({\left\| { { { \textbf{\textsl{x}}}}}\right\| } \big){{{\textbf{\textsl{u}}}}}\cdot{{\boldsymbol { { \nabla } } } _ { \text{g}}}\delta({{{\textbf{\textsl{u}}}}}\cdot{{{\textbf{\textsl{x}}}}})\,{\mathrm{d}}^4{{{\textbf{\textsl{x}}}}}\notag \\ & - \int_{{\mathbb r}^{1,3 } } { { { \textbf{\textsf{f}}}}}({{{\textbf{\textsl{q}}}}}+{{{\textbf{\textsl{x}}}}})\cdot{{{\textbf{\textsl{u}}}}}{\otimes}{{{\textbf{\textsl{x}}}}}{f_{\text{e}}}\big({\left\| { { { \textbf{\textsl{x}}}}}\right\| } \big ) \delta({{{\textbf{\textsl{u}}}}}\cdot{{{\textbf{\textsl{x}}}}})\,{\mathrm{d}}^4{{{\textbf{\textsl{x}}}}}\ , .\label{eq : tildemtensor}\end{aligned}\ ] ] _ remark : _ the first three terms in ( [ eq : tildemtensor ] ) are manifestly symmetric , but the fourth one is not and , as far as we can see , also can not be symmetrized by adding a term whose action on @xmath312 yields zero .
although this brings some technical inconveniences for concrete computations , it does not interfere with the remaining steps in our analysis . at this point ,
the system of evolution equations ( [ eq : mlevolveqqexpl ] ) , ( [ eq : gyroeqqexpl ] ) , ( [ eq : qeqqexpl ] ) , and ( [ eq : wleqqexpl ] ) is already a quasi - explicit first - order system ; however the tensor @xmath400 could be singular ( as an operator acting by matrix multiplication ( w.r.t . the minkowski metric ) on four - vectors ) .
our next step is to show that in typical situations @xmath400 is not singular .
we will establish the invertibility of @xmath400 for admissible initial data in a large neighborhood of those data that correspond to the stationary solutions with empirical electron characteristics that covers essentially all physically sensible situations .
we begin with the observation that , since we have chosen a lorentz frame for the initial value problem which coincides with the particle s initial rest frame , we can in particular choose initial data such that the particle and the fields remain stationary for all time .
the stationary electric coulomb field and the stationary gyro - magnetic field field , denoted here by an index @xmath402 , are defined by the equations @xmath403 with @xmath404 constant , and @xmath405 together with the condition that @xmath406 as @xmath407 . for a stationary situation in which the charge , gyro - magnetic moment , and renormalized mass are matched to those of the empirical electron , with @xmath408 but small ( details are given in section 10.1 ) , the first and third field integrals in ( [ eq : tildemtensor ] ) , evaluated with the corresponding @xmath409 , are of magnitude @xmath6 ( sommerfeld s fine structure constant ) relative to the gyrational bare term , while the second field integral in ( [ eq : tildemtensor ] ) , evaluated with the corresponding @xmath409 , vanishes .
the field integrals in ( [ eq : tildemtensor ] ) are therefore only small perturbations ( in the mathematical sense of operator theory ) to the gyrational bare term @xmath410 in ( [ eq : tildemtensor ] ) , which in turn is invertible because it acts just as multiplication by the strictly positive coefficient @xmath411 ( @xmath412 , the empirical electron mass ) recall that the metric tensor @xmath72 acts as identity on four - vectors . hence , we conclude that for a stationary , empirical data - honoring situation , @xmath400 is invertible .
having established the invertibility of such a stationary @xmath400 , we now turn to the discussion of initial data which do not launch a stationary solution .
since @xmath413 , in such a nonstationary situation the tensor @xmath400 will remain invertible roughly up to the point where the field integrals in ( [ eq : tildemtensor ] ) become comparable to the gyrational bare term .
inspection of the field integrals reveals that this limit is reached ( a ) for applied initial field strengths about @xmath285 times the surface field strength of the stationary electron , i.e. about @xmath414 , or ( b ) for values of @xmath415 about @xmath416 times the surface field strength of the stationary electron per radial size of the electron .
interestingly , field configurations that strong are precisely those that saturate the admissibility condition ( [ eq : initaccelbound ] ) . while therefore it is to be expected that a proof , if at all possible , of the general invertibility of @xmath400 for all _ a priori_admissible initial data will be a technically hard problem , from a physical perspective it is idle to worry about field strengths that strong within the lorentz model .
these critical field strengths are way outside the reasonable range of applied field strengths for which one can hope that the lorentz model makes physically sensible predictions ; whence , we will not address the issue of the critical field strengths any further and simply conclude with the observation that for essentially all physically acceptable initial data the tensor @xmath400 is a small perturbation of @xmath410 and therefore invertible
. we may now multiply ( [ eq : wleqqexpl ] ) by the inverse of @xmath400 to obtain the first - order evolution equation @xmath417 valid surely initially and , if a continuously differentiable solution @xmath170 of ( [ eq : wleqqexpl ] ) exists , valid also for at least a short ensuing time interval during which @xmath400 will continue to be invertible , then , literally by continuity .
conversely , if ( [ eq : hgeq ] ) has a continuously differentiable solution @xmath170 , then as long as @xmath400 remains invertible this solution will also solve ( [ eq : wleqqexpl ] ) .
equation ( [ eq : hgeq ] ) is called the _ hodograph _ equation for @xmath170 .
it is of the desired format in @xmath312 . before we address the question of how to solve our quasi - explicit system of evolution equations ,
a remark is in order regarding the following subtle point .
since @xmath418 , by taking the inner product of ( [ eq : hgeq ] ) we see that we must have @xmath419 .
unfortunately , this last equation is not an automatic identity ( as is , for instance , @xmath420 , which follows just by the anti - symmetry of @xmath172 , irrespective of whether @xmath170 and @xmath172 are dynamically related or not ) , and this may cause some concern .
however , it is only necessary that @xmath421 vanishes identically for solutions to the dynamical equations , and this can be shown to be the case .
in fact , it holds for any solution of ( [ eq : hgeq ] ) with given fields @xmath172 and gyrograph @xmath130 _ conditioned _ on @xmath295 being a solution of ( [ eq : ggeq ] ) . in brief
, the apparent difficulty traces back to the fact that nodvik s minkowski force @xmath389 given in ( [ eq : minkalforce ] ) is generally not four - orthogonal to @xmath170 @xcite because it couples the spin - generated electromagnetic ( dipole and higher multipole ) moments to the inhomogeneities of the electromagnetic fields over the size of the particle .
we have @xmath422 which vanishes identically ( irrespective of @xmath172 ) only for a so - called ` particle without spin ' for which @xmath423 ( more on particle models without spin in appendices a.2 and a.3 ) .
this non - orthogonality of four - velocity and minkowski force is countered on the left side of ( [ eq : wleq ] ) by the spin - orbit coupling term @xmath424 in @xmath393 , which creates a generally anisotropic translational inertia of the gyrating charged particle so that the minkowski momentum @xmath388 and the four - velocity @xmath170 are not anymore parallel to each other but related by the genuine tensor proportionality ( [ eq : minkenmom ] ) . by adapting the corresponding computations of nodvik @xcite to the present model
, it can now be shown that the equation obtained by taking the inner product of ( [ eq : wleq ] ) with @xmath170 is identically satisfied conditioned on @xmath295 being a solution of the gyrograph equations ( [ eq : ggeq ] ) , ( [ eq : minkaltorque ] ) .
the correspondingly conditioned identity @xmath419 now follows from the fact that ( [ eq : hgeq ] ) is equivalent to ( [ eq : wleq ] ) . having succeeded in writing the whole evolution problem as a regular quasi - explicit first - order system ( [ eq : mlevolveqqexpl ] ) , ( [ eq : gyroeqqexpl ] ) , ( [ eq : qeqqexpl ] ) , and ( [ eq : hgeq ] ) , in the final step in our local analysis of the initial value problem we now explain that this regular quasi - explicit first - order system can be treated by a direct generalization of the familiar picard iteration scheme .
this then verifies that a unique regular solution exists and , above all , shows how it can actually be computed .
thus , we consider the coupled sequences of approximate world - lines @xmath425 , @xmath426 ; approximate hodographs @xmath427 , @xmath426 ; approximate gyrographs @xmath428 , @xmath426 ; and approximate electromagnetic fields @xmath429 , with @xmath430 and @xmath431 , jointly defined by @xmath432 @xmath433 @xmath434_-\ , + { { { \textbf{\textsf{t}}}}}^{(n ) } \ , , \label{eq : gyroeqitera}\ ] ] and @xmath435 together with the initial - time constraint @xmath436 starting the iteration with functions that are constant in time , @xmath437 and with correspondingly vanishing time derivatives of the starting data , @xmath438 this set of equations determines the first - order time derivatives of the maps @xmath439 , @xmath440 , @xmath441 , and @xmath442 , in terms of the order @xmath443 maps @xmath444 , @xmath445 , @xmath446 , and @xmath447 , _ and _ the first - order time derivatives of the order @xmath443 maps .
the maps @xmath439 , @xmath440 , @xmath441 , and @xmath442 themselves are now obtained by simply integrating the whole system with respect to the respective time variables .
this is done for the maxwell lorentz equations ( [ eq : mlevolveqitera ] ) , ( [ eq : mldiveqitera ] ) by the standard laplace transform technique , thus picking up the consistent initial data @xmath448 ( i.e. @xmath305 ) posed at @xmath449 through the familiar integration by parts on @xmath450 . equations ( [ eq : zeqitera ] ) ,
( [ eq : hgeqitera ] ) , and ( [ eq : gyroeqitera ] ) are simply integrated w.r.t .
the proper - time from @xmath451 to @xmath111 , thereby picking up the initial data @xmath452 , @xmath298 , and @xmath453 ( viz .
@xmath291 ) from the integration of the explicit proper - time derivatives in the respective left - hand sides .
this completes the @xmath454 iteration cycle . starting with the prescribed order zero maps ( [ eq : initdataitera ] ) and their derivatives ( [ eq : dotinitdataitera ] ) , one can
thus systematically proceed to higher order iterates .
the upshot is an algorithm for the practical computation of the solution , provided the iteration converges . _
remark : _ it is important to notice that `` practical computation '' means to be able to obtain an approximate solution to any degree of precision in some finite region of spacetime . to reach this goal it is _ not _ necessary to actually compute the global iterands in spacetime ( which would be impossible practically ) . a finite number of iterations on a nested , decreasing sequence of spacetime regions , starting with a sufficiently big initial region , will suffice .
the convergence of our algorithm is equivalent to a contraction principle for the fixed - point map in the global future particle histories starting with the specified cauchy data , their time derivatives ( treated as independent functions in the iteration and in the fixed point map ) , and the field decorations of spacetime in the future of @xmath300 , starting with the initial field configuration on @xmath300 . proving the contraction principle establishes the convergence of the iteration to a unique global solution in the future ( and under similar conditions also the past ) of the initial time for those admissible cauchy data for which the tensor @xmath400 remains a small perturbation of @xmath410 and therefore invertible .
so far , for the model in which the bare rest mass and the charge are both concentrated on the same sphere , and provided the ratio of electrostatic to bare rest mass is smaller than @xmath455 , we proved global existence and uniqueness for the iteration algorithm in the special case that only the particle s spin degrees of freedom participate in the particle + field dynamics . in the further specialization that the axis of rotation remains fixed
, we were able to prove exponentially fast approach to a stationary gyration state for the field - particle system , convergence understood on families of nested compact sets , superposed on which is an outgoing field of electromagnetic radiation .
for the proofs of these results , see @xcite . for more general evolutions
, convergence has not yet been established .
the dynamical equations of our massive led may spoil the global existence of the dynamics in several ways .
first , the equatorial rotation speed has to remain subluminal for a particle with finite bare rest mass @xmath456 .
( as we will see later on , the situation is different in the limit @xmath457 where the rotation speed can reach the speed of light , or actually is locked in at that speed ) .
second , the acceleration of the particle has to remain below a critical threshold ( mentioned above ) to guarantee an ordered particle history .
third , but only relevant to the multi - particle systems not considered here , any two particles have to have a minimum distance between them to avoid overlapping cutoffs .
we now comment on the first two problems .
the first problem can be controlled by an appropriate _ a priori _ choice of the bare rest mass distribution .
it suffices to discuss two representative choices , volume and surface inertia .
they differ in their limiting behavior when @xmath458 .
the volume choice for @xmath145 is representative for the larger class of continuous functions @xmath459)$ ] , in all of which cases @xmath460 .
in such a case it takes only a finite amount of gyrational energy to bring an equatorial point to the speed of light .
this signals that the choice of a continuous @xmath145 will create a mathematically delicate problem when it comes to the question whether the gyrational dynamics remains subluminal , and so one may expect that this choice may easily lead to a singularity in finite time .
in contrast , for the surface inertia choice for @xmath145 , the function @xmath411 is not only increasing and convex , but satisfies @xmath461 as @xmath462 . in this case
it would require an infinite amount of gyrational energy to bring an equatorial point to the speed of light , so that it is to be expected that the dynamics will not lead to a singularity in finite time .
these considerations show that surface inertia is preferable , although we do not claim that this requirement is necessary for global or even local existence .
the second problem is much harder to control _ a priori_. for a moving particle one now has to rule out that constructive interference of electromagnetic radiation fields leads to a violation of the acceleration limit ( in our lorentz frame ) @xmath463 where @xmath283 is the radius of the particle and @xmath464 . at the initial instant @xmath465 ,
the particle is at rest and the estimate ( [ eq : condonaccel ] ) reduces to ( [ eq : initaccelbound ] ) .
notice finally that the smallness condition that we need in our proof in @xcite is surely going to be violated if one lets the bare rest mass flow to zero .
the control of this situation apparently requires completely different techniques . due to all these technical difficulties , the global existence and uniqueness of regular solutions , at least for the physically relevant parameter range of the model ,
still awaits its verification . in the following sections we proceed under the assumption that the dynamics of the field - particle system is globally well - defined . put differently ,
whatever is concluded in those sections applies only to those situations for which a unique solution exists for all time .
having established that the dynamical state in our single - particle massive led in spacetime @xmath31 is given by the transversal part of @xmath342 ( viz .
, @xmath172 ) on a spacelike hypersurface of spacetime containing the particle support , together with the particle variables @xmath160 , @xmath170 , and @xmath130 ( or @xmath219 ) coincidental with the particle support in the spacelike hypersurface , we now turn the important question which of these states are scattering states , and in which sense . loosely speaking
, a _ scattering state _ is a state whose backward evolution tends asymptotically in the infinite past to a superposition of free evolutions of propagating ( in space ) field and particle degrees of freedom , and similarly its forward evolution tends asymptotically in the infinite future to a(nother ) superposition of free evolutions of propagating field and particle degrees of freedom .
the _ scattering problem _ is concerned with establishing the existence of the scattering states , with their unique identification , and further with their classification according as to how their asymptotically free evolution in the infinite past is connected with their asymptotically free evolution in the infinite future .
we will limit our discussion to situations in which the conserved quantities @xmath466 , @xmath216 , and @xmath230 are finite . moreover , we take a foliation in which @xmath467 , i.e. a center of mass frame for the field - particle system .
we assume the center of mass is at the origin of the initial leaf .
we expect that in our massive led there are basically three categories of states , namely the non - propagating ( a.k.a .
bound ) states , the scattering states , and the ` catastrophic states ' . for the catastrophic states
the evolution either terminates at a finite time , or blows up in infinity time , or suffers some other pathological behavior like loss of uniqueness .
as explained in the previous section , the global existence and uniqueness problem is largely unsettled so that we will not be able to rule out catastrophic states nor characterize them as physically irrelevant ( although we expect the latter to be true ) .
however , the special global existence and uniqueness result in @xcite shows that for a certain class of initial conditions the evolution does exist globally , and these states are in fact scattering states .
we shall show in this section that the scattering states have the remarkable , and in fact physically indispensable character of soliton dynamics for the renormalized particle degrees of freedom .
as for the bound states , we readily characterize the stationary states , but we have no control yet on the question of whether periodic ( or perhaps quasi - periodic ) non - propagating states exist as well . this section on scattering
is mainly included for the purpose of demonstrating the important solitonic character of the renormalized particle dynamics in scattering situations .
as such , we will necessarily be brief on the other scattering issues . as a general technical reference to scattering theory we recommend @xcite , though
most of the emphasis there is on quantum mechanical potential scattering ; in this context , see also @xcite for subsequent technical breakthroughs , and @xcite for a very clear exposition of the conceptual issues involved and further references .
various classical scattering results in electrodynamics are discussed in @xcite .
massive led describes the interacting dynamics of two _ a priori _ subsystems , the charged bare particle and the electromagnetic field , each of which is characterized by its own dynamical degrees of freedom . with only a single
charged bare particle interacting with the electromagnetic field in otherwise empty space , it could seem that there are no bound states .
however , some of the particle s spin degrees of freedom couple strongly to some of the dynamical degrees of freedom of the electromagnetic field , as a result of which bound states exist .
the most elementary ones give rise to the notion of the _
renormalized particle_. in addition , more subtle scenarios may possibly lead to dynamical bound states , on which we comment only briefly .
the most obvious bound states are stationary solutions with nonvanishing particle spin .
clearly we can choose the standard foliation of spacetime for these states , i.e. a lorentz frame @xmath35 in which the particle is at rest at the origin of the space slice @xmath468 ; i.e. @xmath469 for all @xmath111 so that @xmath470 .
the stationary states compatible with our conventions made in the cauchy problem have been defined earlier ; all other stationary bound states that satisfy our convention about the center of mass are now obtained by at most a rotation from these stationary states . for later convenience
, we here list the solutions explicitly .
the stationary particle in the center of mass frame is spinning with constant angular velocity tensor @xmath130 .
the stationary charge - current density four - vector thus reads @xmath471 we seek a stationary solution of the vector wave equation @xmath472 satisfying the lorentz gauge condition and vanishing at spatial infinity . recalling that @xmath473 , and introducing the time - space decompositions for the current density @xmath474 and electromagnetic potential four - vector @xmath475 , for the stationary coulomb potential @xmath476 we find @xmath477 and for the stationary vector potential @xmath478 , @xmath479 the electric and magnetic fields are obtained in the usual manner as @xmath480 and @xmath481 , so that @xmath482 since we chose a spherical distribution @xmath483 with compact support in @xmath484 , the coulomb and the vector potential take a universal form outside the support of the particle , given by @xmath485 where @xmath486 is the magnetic moment for the stationary @xmath487 , @xmath488 clearly , away from its support the stationary charged particle s electromagnetic signature is that of an electric point charge and a magnetic point dipole , as was to be anticipated . finally , we remark that the static limit @xmath489 for the stationary bound state is no longer a bound state of the field - particle system , in the sense that only the passive coulomb field remains attached to the charged bare particle . in various approximate , linearized versions of ( massive ) led , time - periodic non - propagating field - particle solutions have been found .
time - periodic solutions do not exist if the _ wiener _ condition is satisfied , viz .
the fourier transform of @xmath181 be strictly positive .
however , notice that neither the surface density nor the volume density discussed in this paper satisfy the wiener condition .
in principle one should therefore be prepared for the possibility of periodic solutions in our massive led .
on the other hand , rigorous studies of simpler semi - relativistic models of a particle interacting with a scalar wave field show that time - periodic bound states of their linearized version are structurally unstable to the nonlinear perturbation of the linear dynamics @xcite ; see also @xcite .
no such study has yet been carried out for our massive led . in our center of mass initial leaf
, we can divide the scattering states into two categories : ( a ) scattering of electromagnetic radiation off a non - moving particle , and ( b ) scattering of radiation and moving particle in an encounter .
scattering with a non - moving particle is considerably simpler because the family of space slices of simultaneity for the particle are just the standard foliation of spacetime , as for a stationary particle .
the only dynamically active degree of freedom of the particle is its euler tensor , or in the stationary lorentz frame of our foliation , the angular velocity vector @xmath490 .
the dynamical degrees of freedom of the electromagnetic field are restricted to be compatible with the condition that the particle sit still at the origin of space .
for instance , large classes of axisymmetric field decorations of spacetime will be fine .
this case has been studied in by us in rigor ; details will be given in @xcite .
here we summarize the main results , obtained under the condition that the ratio of the particle s electrostatic coulomb energy to the bare rest mass is smaller than @xmath455 .
the best understood situation is when the particle s axis of rotation does not change during the evolution , say @xmath491 , so that @xmath492 is the only remaining dynamical particle degree of freedom . for all admissible initial conditions corresponding to these restrictions , we know that on families of nested compact sets the field - particle system converges exponentially fast to a stationary state which is a bound state if the conserved @xmath493 initially and the static state if @xmath494 , superposed on which in either case is an outgoing field of electromagnetic radiation .
the evolution of the electromagnetic field is thus given by the scattering formula @xmath495 where @xmath496 is a divergence - free electromagnetic field , orthogonal to all bound - state fields , uniquely determined by the initial state , characterizing the outgoing radiation .
similarly @xmath497 where @xmath498 denotes complex conjugate .
we remark that @xmath499 is the same stationary bound state in both formulas , which is a simple illustration of the soliton dynamics of the renormalized particle in this special situation . in the general non - moving particle case where the axis of @xmath487 is not fixed during the evolution , we know that every admissible initial state compatible with the constraints evolves uniquely into the future ( respectively , past ) , but we do not yet know that on families of nested compact sets the field - particle system converges ( exponentially fast ) to a stationary state .
however , we do conjecture that this will be the case . assuming this to be true
we were able to show that for large times the evolution of the electromagnetic field satisfies the scattering formulas @xmath500 and @xmath501 where now @xmath502 and @xmath503 are generally _ not the same _ stationary bound state ; however , they differ by at most a space rotation , and this once again illustrates the soliton dynamics of the renormalized particle . in both situations ,
the explicit characterization of the scattering operator from the `` in '' states to the `` out '' states has not yet been worked out .
scattering with a moving particle can mean any scattering process in which the particle moves for at least a finite interval of time . in that case we need to work with a nonstandard foliation , as explained in the section on the cauchy problem for the state .
needless to say , so far this is the least explored case . the explicit characterization of the scattering states , and of the scattering operator from the `` in '' states to the `` out '' states still have to be worked out .
however , although it is rigorously unknown which initial states are scattering states for which the particle moves , _ a priori _ speaking there are several types of scattering processes to be distinguished : ( a ) the particle moves only a finite distance , associated with which are bound states in the distant past and future this is a perturbation of the previous `` scattering of radiation off a bound state ; '' ( b ) the particle is at rest only in the past ( respectively , only in the future ) but moves freely in the asymptotic future ( respectively , past ) ; ( c ) the particle moves freely in the asymptotic past _ and _ future .
furthermore , one can give the following partial characterization of the `` in '' and `` out '' states .
in the situations ( a ) and the parts of ( b ) associated with a stationary bound state , the scattering asymptotics differs from the one discussed in the previous subsection by at most a translation of the stationary bound state . in the situations ( c ) and the parts of ( b )
not associated with a stationary bound state , the scattering asymptotics in the specified foliation is given as follows .
as for the asymptotic particle motion , all the `` in '' states @xmath504 , as well as all the `` out '' states @xmath505 , are obtained from the stationary particle states considered above by the action of an element of the poincar group , more explicitly a combination of a translation , a rotation , and a lorentz boost .
the free evolution of a particle is obtained by applying the lorentz - time propagation operator to these states , or expressed in terms of proper - time @xmath111 ( recall that @xmath506 ) , @xmath507 @xmath508 @xmath509 and correspondingly for @xmath510 using the `` in '' states .
associated with these free asymptotic evolution of the particle are the co - moving electromagnetic fields .
sufficiently far away from the particle these are just co - moving in the standard foliation to which the compatible foliation is asymptotically close except in a neighborhood of the world - tube .
their electromagnetic `` in '' and `` out '' states are parameterized by the space part @xmath511 of @xmath512 , the constant velocity @xmath513 in @xmath514 , and the angular velocity vector @xmath515 corresponding to @xmath516 . to have a handy abbreviation for this parameter set
, we choose the index @xmath517 for ` co - moving . '
we set @xmath518 and where @xmath519 and @xmath520 .
after these preparations , we can give the asymptotic future evolution of the electromagnetic field for such a scattering state as @xmath521 similar formulas hold for the asymptotic past evolution in terms of the `` in '' potentials . while the rigorous treatment of the scattering problem of massive led is still in its infancy , in all instances of scattering a remarkable conclusion about the gyrational particle degrees of freedom can be made .
navely one would expect that generally the orientation _ and _ the norm of the angular velocity of the particle would change during the scattering process and with it the rotational and renormalized mass of the particle , its spin and magnetic moment .
however , precisely this does not happen .
any scattering process connects two freely evolving particle states with identical renormalized mass and identical magnitudes of spin and magnetic moment .
the relativistic lorentz electron , equipped with non - vanishing positive bare inertia , thus has the dynamical characteristics of an elementary particle in the best sense one could have hoped for in a classical theory . by adapting the terminology of spohn @xcite and collaborators ,
who considered charged particles without spin , we call the renormalized particle a _ spinning charge soliton . _ as remarkable as this soliton dynamics itself , is the fact that it is a simple joint consequence of the conservation of @xmath522 together with the fact that @xmath523 for a stationary bound state , and the invertibility of the map @xmath524 for such a state . indeed , we need only consider a stationary bound state , for we know that the asymptotic future ( past ) evolution of the scattered particle dressed with its co - moving fields ( the renormalized particle ) is just a translated and boosted such state . explicitly , for a stationary bound state we have , in the lorentz frame and in dual vector notation , @xmath525 the first term being the bare spin , the second the electromagnetic spin of the particle .
for the bare spin @xmath526 we already saw , in our section on the cauchy problem , that it is parallel to @xmath487 and the map @xmath527 invertible . using our formula ( [ eq : stata ] ) for the stationary vector potential of a bound state ,
we see by explicit computation that also the total spin @xmath528 is parallel to @xmath487 , and the map @xmath529 is invertible .
we have @xmath530 , with @xmath531 and map @xmath532 increasing , vanishing for @xmath533 .
now , @xmath534 , so that from the invertibility of @xmath532 we conclude that @xmath535 determines @xmath157 uniquely in a stationary bound state .
thus , from the constancy of @xmath230 we conclude that @xmath536 in the asymptotic future state has the same value as in the asymptotic past state .
the solitonic character of the renormalized particle is proved .
we carry out the renormalization group analysis for a simple example where the various dynamical functions can be computed explicitly .
specifically , we here consider ` densities ' @xmath145 and @xmath181 which are dirac measures concentrated on the surface of a two - sphere of radius @xmath283 , viz .
@xmath537 the elementary charge @xmath538 will not be tempered with and therefore not displayed as variable in the arguments of the functions below . beside having set the charge @xmath539 of the model particle equal to that of the real electron , we also demand that its total model mass @xmath540 matches the empirical electron mass , @xmath541 , and that its magnetic moment @xmath542 matches the empirical electron magnetic moment , @xmath543 . _
remark : _ in this and the next section we resort to units in which the speed of light @xmath3 is displayed explicitly , to facilitate the comparison with the formulas used in physical data books .
the renormalized particle spin will be _ derived _ in our model in the limit of vanishing bare rest mass , whence is _ not _ to be matched to a ` corresponding empirical value , ' but rather compared .
navely speaking it is to be expected that in our classical theory with ultraviolet cutoff the spin to magnetic - moment relation @xmath544 with land factor @xmath545 should hold .
interestingly , we will find @xmath546 .
we remark that , without further elaboration , there remains a certain ambiguity in matching our stationary model data to empirical electron data inferred with the help of quantum mechanics from data of _ dynamical _ experiments , viz .
@xmath547 in scattering experiments , and @xmath548 in resonance frequency experiments in a penning trap .
the ultimate justification for these choices has to wait for some future dynamical analysis of our model . by performing at most a space translation
, we can assume that the stationary particle , and its self - fields , are centered the at the origin of the space slice @xmath550 of the lorentz frame in which the system is stationary .
since the time - component of the spin four - vector and the angular velocity four - vector vanish in our stationary frame , after at most a space rotation we can furthermore assume that the spin four - vector of the stationary bare particle , @xmath551 , and of its self - fields , @xmath552 , points along the @xmath553 direction .
since @xmath554 , we can canonically identify these spin four - vectors with the familiar bare and field spin three - vectors @xmath526 , @xmath555 , respectively the angular velocity four - vector with the angular velocity three - vector @xmath487 , which moreover can be written as @xmath556 , @xmath557 , and @xmath558 , with @xmath559 . with these conventions , and with ( [ eq : shell ] ) , the relevant formulae for the renormalization analysis are the following . beside the electric monopole moment ( i.e. the charge @xmath539 ) , the magnetic dipole moment of the empirical electron is the second electromagnetic characteristic measurable in a classical manner .
the magnetic moment three - vector @xmath560 given in ( [ eq : magmom ] ) ( and re - installing a factor @xmath561 ) evaluates to @xmath562 [ eq : xxv ] + since it points in the @xmath563 direction as well , we will write @xmath564 , with @xmath565 read off from ( [ eq : xxxii ] ) . the mass @xmath540 of the renormalized particle is defined by @xmath566 here , @xmath567 is the gyrational energy of the bare particle , defined earlier in ( [ eq : gyromass ] ) and evaluated for general spherical @xmath145 in ( [ eq : gyromasseval ] ) .
re - installing a factor @xmath568 and inserting now ( [ eq : shell ] ) for @xmath145 in ( [ eq : gyromasseval ] ) , we obtain @xmath569 furthermore , @xmath570 is the electromagnetic field energy of the gyrating charged particle , @xmath571 our system of equations for the stationary renormalization flow is given by @xmath572 with the function @xmath573 given through ( [ eq : xxix ] ) , ( [ eq : xxvi ] ) , ( [ eq : xxviii ] ) , and with the function @xmath574 given through ( [ eq : xxxii ] ) .
explicitly , we have @xmath575 these two coupled equations determine a curve in @xmath576-space for which we will study the limit @xmath457 . equation ( [ eq : renormmuexpl ] ) can immediately be solved for @xmath492 in terms of the radius @xmath283 , @xmath577 inserted into ( [ eq : renormmexpl ] ) , this gives @xmath578 which rewrites into a formula for the bare rest mass as function of @xmath283 , @xmath579 we are interested in the inverse function @xmath580 .
in particular , we want to know whether @xmath581 can be taken without encountering superluminal equatorial rotation speeds .
we readily read off from ( [ eq : lxv ] ) that the right side is well - defined only on the @xmath283-interval @xmath582 . from ( [ eq : renormmuexpl ] ) and ( [ eq : comptonl ] ) it furthermore follows that @xmath583 for @xmath584 , viz .
, no superluminal rotation speeds occur for @xmath283 in the domain of definition for the r.h.s . of ( [ eq : lxv ] ) .
furthermore , @xmath585 is a monotonically increasing function , so that the limit @xmath586 for @xmath456 will answer the question whether one can take the limit @xmath457 for @xmath283 .
a simple inspection of ( [ eq : lxv ] ) reveals that @xmath587 as @xmath588 and that @xmath457 as @xmath589 . hence , by monotonicity there is a unique limit of vanishing bare rest mass for the radius of the renormalized particle , given by @xmath590 it also follows from our discussion that the equatorial rotation speed in this limit is precisely equal to the speed of light @xmath3 .
the renormalization flow analysis reveals that the lorentz program of a `` purely electromagnetic electron '' can be completed successfully , in the stationary setting at least .
the resolution of the puzzlement is that beside the static electromagnetic field energy contributions considered by lorentz , the mass of the properly renormalized purely electromagnetic particle consists of an additional contribution which we call _ photonic _
, for it is the mass of an object without rest mass which rotates with equatorial speed equal to the speed of light .
this contribution was missed by lorentz who calculated without bare rest mass from the outset .
it is instructive now to make use of the empirical relation for the electron magnetic moment , @xmath591 , where @xmath592 is the _ bohr magneton _ and @xmath593 the anomaly factor ( see , e.g. , @xcite ) . up to corrections of @xmath594 , equation ( [ eq : rlim ] )
then reveals that the radius of the renormalized purely electromagnetic electron is @xmath595 where @xmath596 is the electron s _
compton length_. this result is quite remarkable in many respects .
it shows that our lorentz - covariant led in fact contains the correct ultraviolet cutoff length where heuristic discussions traditionally place the limit of applicability of a classical field theory , viz . at the order of the electron s compton length @xmath597 , rather than at the (
lamentably so - called ) _ classical electron radius _ , given by @xmath598 which gives the size of a ` purely electrostatic classical electron . ' recalling _ sommerfeld s fine structure constant _ , given by @xmath599 we have the relation @xmath600 so that the true lorentz electron is actually about @xmath601 times larger in diameter than predicted by a purely electrostatic calculation .
we conclude this subsection with a discussion of the total spin , given by @xmath602 here , @xmath526 is the spin three - vector of the bare particle , which for @xmath603 evaluates to @xmath604 which , by inserting ( [ eq : renormmuexpl ] ) into ( [ eq : xxx ] ) , with ( [ eq : bohrmagneton ] ) for @xmath605 , becomes @xmath606 using now ( [ eq : lxv ] ) , we get @xmath607 the second spin term , @xmath555 , is the stationary field spin three - vector associated with the rotating charged particle , which evaluates to @xmath608 we write @xmath609 , with @xmath610 and @xmath611 given in ( [ eq : spinbofr ] ) and ( [ eq : xxxi ] ) .
we consider the spin coefficients @xmath610 and @xmath611 as function of @xmath283 .
letting @xmath612 , i.e. taking @xmath457 , for the bare spin we get , @xmath613 while for the field spin , recalling that @xmath614 as @xmath457 , we immediately obtain from ( [ eq : xxxi ] ) that @xmath615 finally , adding @xmath611 and @xmath610 and taking the limit equals the sum of the individual limits , whence the renormalized limit spin magnitude is given by @xmath616 put differently , from ( [ eq : spinlim ] ) we read off that the spin magnitude @xmath617 of the renormalized electron satisfies @xmath618 to high precision , with relative corrections of order @xmath6 , compared to our nave expectation @xmath619 , and compared to @xmath620 in a quantum mechanical framework . curiously , furthermore , a closer look at the corrections reveals that in our led an expansion of the @xmath12-factor in the land relation ( [ eq : mugspin ] ) in terms of powers of @xmath6 holds .
we set @xmath621 , then @xmath622 which is surprisingly similar to the familiar expansion in qed @xcite , @xmath623 we finally come the dynamical renormalization problem . in contrast to the stationary problem , no rigorous results concerning the existence of this limit are available yet .
however , it is of interest to inquire into the features expected for a renormalized dynamical limit of vanishing bare rest mass of massive led .
first of all , the matching of empirical electron data in the dynamical renormalization problem with scattering states has to be carried out in the asymptotic past ( or future ) , for we want to `` send in '' the particle with the correct physical characteristic particle data .
the incoming - data matching in the asymptotic past will fix @xmath283 and @xmath456 for the rest of the motion , since as parameters of the model these do not vary during the motion .
this matching in the asymptotic past will also fix the value of @xmath157 in the asymptotic future , which is due to the soliton dynamics of the renormalized particle . during the motion , of course
, @xmath157 will in general vary .
however , our results on the scattering with a non - moving particle indicate that @xmath157 approaches its asymptotic value extremely fast . extrapolating from this observation
, we expect that the soliton dynamics of the particle in massive led in fact will try to keep the angular frequency @xmath157 of the particle as close as possible to the asymptotic value in the free motions . by the same token
the gyrational mass ( hence , also the renormalized mass , which is a small perturbation of the former in the stationary state ) of the particle should remain ( in the instantaneous rest frame ) very close to the empirical mass during the motion .
an important conclusion can then be drawn if one can show that the gyrational bare mass of the particle with @xmath408 stays _ uniformly close _ to the empirical mass as @xmath457 for an @xmath456-indexed sequence of motions .
namely , in that case the equatorial gyration speed of the renormalized purely electromagnetic particle must be locked at the speed of light during the whole motion ! for assuming to the contrary that the gyrational speed would remain bounded away from the speed of light for at least a finite fraction of the evolution ( of course still close to it in the uniform sense stipulated ) , it would then follow that the bare gyrational mass converges to zero , while the electromagnetic mass would contribute at best an amount of order @xmath6 of the empirical mass .
this is in contradiction to its hypothesized uniform proximity to the empirical mass .
we infer that in the limit @xmath457 the bare gyrational mass @xmath411 is independent of @xmath487 , hence a _ constant of the motion _ identical to what we have called the photonic mass of the stationary particle ; explicitly , @xmath624 by the same reasoning , in the limit of vanishing bare rest mass the bare spin vector becomes the _
photonic spin _ , a four - vector with only two independent variables .
it s dual tensor is given by @xmath625 where @xmath626 is anti - symmetric , of space - space type , has norm 1 , satisfying @xmath627 .
the inertial gyro - frequency @xmath130 and the spin tensor @xmath626 are related by a simple proportionality , @xmath628 emphasizing the characteristic euler angular frequency @xmath629 explicitly .
similarly , @xmath630 , with @xmath631 after these preparations , we can write down the putative dynamical equations of renormalized led .
the world - line equation of renormalized led reads @xmath632 where @xmath633 is the energy - momentum four - vector of the renormalized purely electromagnetic electron , with @xmath634 given by @xmath635_+ \right]_+ { f_{\text{e}}}\big({\left\| { { { \textbf{\textsl{x}}}}}\right\| } \big)\ , \delta({{{\textbf{\textsl{u}}}}}\cdot { { { \textbf{\textsl{x}}}}})\,{\mathrm{d}}^4{{{\textbf{\textsl{x}}}}}\label{eq : smmtnull}\ ] ] now the renormalization limit of the tensor mass @xmath393 .
the gyrograph equation of renormalized led reads @xmath636_- = \nonumber \\ & \int_{{\mathbb r}^{1,3 } } { { { \textbf{\textsl{x}}}}}\wedge\bigl({{{\textbf{\textsf{g}}}}}+ { { { \textbf{\textsl{u}}}}}\otimes{{{\textbf{\textsl{u}}}}}\bigr)\cdot { { { \textbf{\textsf{f}}}}}({{{\textbf{\textsl{q}}}}}+{{{\textbf{\textsl{x}}}}})\cdot \big({{{\textbf{\textsl{u}}}}}- ( { \mbox{${\boldsymbol { { \omega } } } _ { \text{\textsc{fw}}}{}$}}+ \kappa { { { \textbf{\textsf{s}}}}}_{\text{ph}})\cdot{{{\textbf{\textsl{x}}}}}\big ) { f_{\text{e}}}({\left\| { { { \textbf{\textsl{x}}}}}\right\| } ) \,\delta({{{\textbf{\textsl{u}}}}}\cdot{{{\textbf{\textsl{x}}}}})\,{\mathrm{d}}^4{{{\textbf{\textsl{x}}}}}\ , .
\label{eq : spineqnull}\end{aligned}\ ] ] the maxwell
lorentz equations for renormalized led have their familiar structure , but the charge - current density is now given by @xmath637 the state in renormalized led differs from that of massive led in that only the orientation but not the norm of @xmath638 is a dynamical degree of freedom .
this imposes some consistency conditions also on the initial conditions .
we remark that ( [ eq : spineqnull ] ) does _ not _ automatically preserve @xmath639 for an arbitrary electromagnetic tensor field @xmath172 and world - line @xmath241 .
however , it is only mandatory that the norm of @xmath638 be conserved for consistent joint solutions of the world - line , gyrograph , and field equations ; the situation is similar to the one related to the conservation of @xmath640 by the world - line equation .
in the previous sections we have constructed the first viable relativistic lorentz electrodynamics with positive bare rest mass ( massive led ) .
we studied its renormalization flow limit to vanishing bare rest mass ( renormalized led ) .
this modern completion of lorentz program of a purely electromagnetic theory avoids the superluminal pitfalls of lorentz purely electromagnetic calculations . in the limit
we obtain a particle with correct empirical mass , charge and magnetic moment , though with spin @xmath4 .
although the bare rest mass vanishes , the particle retains a `` photonic mass '' resulting from luminal equatorial motions , which actually accounts for most of the particle s mass while the traditional electromagnetic self - energies contribute merely a correction of order @xmath6 .
the size of the limit particle is of the order of the electron s compton length , about @xmath14 classical electron radii .
the finite size of this ultraviolet cutoff in the ` purely electromagnetic ' limit endows our led with a mildly _ a - causal _ appearance . however , the ultraviolet cutoff is itself an individual but _ non - local event _ which can not be subdivided into a collection of individual point events
. the cutoff moves as a non - local unit in a lorentz - covariant way , in fact accelerated in a _ causal _ manner in the sense that fields outside the cutoff event do not influence its motion by the field integrated over the cutoff at any given instant of its proper - time .
it is perhaps surprising that such an instantaneous , non - local action is compatible with lorentz covariance and causality , yet it is .
the existence of this physically viable properly renormalized led reinforces some hopes that led may ultimately have some bearing on the quest for a consistent qed , with led or some improved version thereof ( led@xmath641 say ) as its ( semi-)classical limit . recall that the perturbative series for qed @xcite is most likely divergent @xcite and
hence does not give a mathematically acceptable definition of qed .
many experts nowadays are inclined to believe that the constructive quantum field theory approach to qed @xcite will not lead to a mathematically consistent theory either @xcite , though the failure of this approach to qed has not yet been conclusively demonstrated @xcite .
there is also a strong anticipation that a physically correct and mathematically well - defined relativistic quantum theory must involve also the strong and weak nuclear interactions , as well as gravity , and all this supersymmetrically @xcite .
it should however not be necessary to unify physics at the planck scale in order to obtain a consistent working theory for electromagnetic phenomena above the electron s compton scale .
inspired by kramer s program of ` non - relativistic qed ' @xcite , which aims at the quantization of led s semi - relativistic predecessor ( see appendix a.3 ) , and which is expected to describe the low energy limit of qed @xcite , one may want to investigate the possibility of deforming our led into a ` quantized led ' , perhaps along the ideas of kontsevich @xcite , and it would be interesting to see whether that would teach us something about qed proper . such fundamental hopes aside , our led is interesting in itself as a nonlinear relativistic microscopic theory which couples particle and field degrees of freedom in a self - consistent manner .
while the rigorous control of its semi - relativistic predecessor has made significant advances in recent years due to the efforts of spohn and collaborators and others ( discussed in the appendix ) , the extension of this rigorous control to our relativistic led has just started @xcite .
the most surprising new dynamical challenge is the need to construct a compatible foliation of spacetime along the evolution of the state in led by also solving the vacuum einstein equations as a free - boundary problem coupled to the evolution of the state .
unexpectedly , massive and renormalized led may therefore serve as playground for the harder problem of black hole dynamics in general relativity . among the future projects ,
one is certainly the rigorous extraction of effective equations of motion for the particle in the adiabatic regime , where to lowest order one expects equations of the type discussed in @xcite , with radiation reaction - corrections showing in the next order of approximation .
another interesting project is the many particle problem , and in particular the rigorous microscopic foundation of relativistic vlasov
maxwell theory @xcite , which so far could not even be attempted for lack of a consistent microscopic theory of electromagnetic particles and fields .
appendix a.1 contains the more technical derivations of the euler - lagrange equations for our massive led , appendix a.2 discusses the singular point mass limit in which our massive led reduces to nodvik s massive led , and appendix a.3 provides supplementary material about the semi - relativistic led . in this subsection , we first collect various formulas needed to carry out the variations of the action functional .
next we apply these formulas to derive the euler - lagrange equations for the action principle . to a given world - line @xmath236
we may add a small perturbation @xmath642 to get a deformed particle world - line @xmath643 .
here , @xmath111 is the proper - time of the unperturbed world - line .
the proper - time along the perturbed world - line at the point @xmath644 will in general _ not _ coincide with the value @xmath111 , i.e. the proper - time of the unperturbed world - line . to first order ,
the difference in the proper - times along the deformed and original world - line reads @xmath645 accordingly , the derivative of @xmath646 with respect to @xmath111 , viz .
@xmath647 , is generally _ not _ the first - order variation in the four - velocity , @xmath648 , which is defined by taking the derivative of @xmath649 w.r.t .
its proper - time , then computing the difference @xmath650 to first order . as a function of @xmath111 , the first - order variation in the four - velocity
is given by the space projection ( w.r.t .
@xmath170 ) of @xmath647 , @xmath651 where @xmath72 is the metric tensor of minkowski space , defined in ( [ eq : gt ] ) . taking the inner product of ( [ eq : deltau ] ) with @xmath170 , and recalling that @xmath72 acts as identity on four - vectors , we find that @xmath652 , as required by @xmath653 in first order .
the variation of @xmath383 under an independent change @xmath654 , the gyrograph being kept fixed , is given by @xcite : @xmath655 defining quasi - coordinates @xmath656 equation ( [ eq : varthom1 ] ) can be rewritten into @xmath657_- + { \mbox{${\boldsymbol { { \omega } } } _ { \text{\textsc{fw}}}{}$}}{{{\textbf{\textsl{u}}}}}\,\cdot{{\stackrel{\,\circ}{{\delta { { { \textbf{\textsl{q}}}}}}}}}\ , , \ ] ] which proves convenient later on . moreover , although the gyrograph is fixed during an independent variation of the world - line , the fermi
walker frame changes , and with it the tensor of the euler gyration , for the gyrograph is defined w.r.t .
the fermi walker frame . the variation in @xmath130 under an independent change @xmath654
is given by @xcite @xmath658 whence @xmath659 the orientation of the set of three spacelike vectors @xmath660 relative to the spacelike vectors @xmath661 can be specified by euler angles @xmath662 , @xmath663 , @xmath200 .
independent changes @xmath664 , @xmath665 , @xmath666 of the particle s history of euler angles , i.e. keeping its world - line fixed , give rise to the variation @xcite @xmath667_- \ , , \label{eq : vareul2}\ ] ] where the anti - symmetric tensor @xmath668 is given by @xmath669 with @xmath670 the tetrad of the frame @xmath115 and @xmath671 given by @xmath672 cf .
@xcite for details .
we first turn to the action of the bare particle , which we rewrite as @xmath673 where @xmath674 is the lagrangian for the bare particle .
we first define the volume @xmath188 bounded by two space slices defined by @xmath675 at @xmath233 and @xmath193 .
its translation by @xmath160 is denoted by @xmath676 . under an independent variation @xmath677
the resulting first variation of @xmath678 reads @xmath679 where the last term under the integral results from the change in proper time along the perturbed world - line , see equation ( [ eq : deltadtau ] ) . as for the variation of @xmath680 w.r.t .
@xmath170 , we get @xmath681 where we used @xmath682 then integrated by parts , carried out the differentiations , and noticed that terms proportional to @xmath683 in the integrand vanish upon integration against @xmath684 .
but with the help of the four - orthogonality of @xmath170 and @xmath648 we see that @xmath685 as for the variation of @xmath680 w.r.t .
@xmath130 , we find that @xmath686 where @xmath219 is the spin tensor of the bare particle , defined earlier in ( [ eq : barespintensor ] ) .
thus , we have @xmath687 inserting now ( [ eq : vareuler1 ] ) for @xmath688 , with ( [ eq : deltau ] ) in place for @xmath648 , then using @xmath689 once , next integrating by parts terms containing @xmath690 , recalling that @xmath691 at the boundaries of integration , then using again @xmath689 and also noticing once again that terms proportional to @xmath683 in the integrand vanish upon integration against @xmath684 , we obtain @xmath692 here , @xmath693 is the minkowski momentum four - vector of the bare particle , which for later convenience we write in the form @xmath694 where @xmath695 is the diagonal bare mass tensor , with @xmath696 the gyrational bare mass of the particle . turning next to the action of field - particle coupling ,
we rewrite it as @xmath697 where @xmath698 is the lagrangian for the coupling of the particle to the electromagnetic fields .
since @xmath130 and @xmath312 occur in @xmath699 only in the combination @xmath700 , we can treat @xmath699 as a function of @xmath701 .
thus , abusing notation a little bit , we write @xmath702 and obtain @xmath703 under independent variation of the world - line . as for the variation of @xmath699 w.r.t .
@xmath170 , we have @xmath704 using once again ( [ eq : dudelta ] ) , then integrating by parts , carrying out the differentiations , and noticing once again that terms proportional to @xmath683 in the integrand vanish upon integration against @xmath684 , we obtain @xmath705 using the four - orthogonality of @xmath170 and @xmath648 , we thus have @xmath706 as for the variation of @xmath699 w.r.t .
@xmath160 , we have @xmath707 while for the variation of @xmath699 w.r.t .
@xmath119 we simply have @xmath708 we use ( [ eq : varom1 ] ) for the variation of the total gyration tensor @xmath119 along the independently perturbed particle world - line , and find @xmath709 { f_{\text{e}}}\big({\left\| { { { \textbf{\textsl{x}}}}}\right\| } \big ) \,\delta\big({{{\textbf{\textsl{u}}}}}\cdot{{{\textbf{\textsl{x}}}}}\big ) \,{\mathrm{d}}^4{{{\textbf{\textsl{x}}}}}\\ \qquad\qquad\qquad + \int_{{{\xi_{{{{\textbf{\textsl{q } } } } } } } } } \big [ { { { \textbf{\textsl{x}}}}}+ ( { { { \textbf{\textsl{u}}}}}\cdot{{{\textbf{\textsl{x}}}}}){{{\textbf{\textsl{u}}}}}\big ] \big({{{\textbf{\textsl{u}}}}}-{{\boldsymbol { { \omega } } } } \cdot{{{\textbf{\textsl{x}}}}}\big)\cdot{{{\textbf{\textsl{a}}}}}({{{\textbf{\textsl{q}}}}}+{{{\textbf{\textsl{x } } } } } ) { f_{\text{e}}}\big({\left\| { { { \textbf{\textsl{x}}}}}\right\| } \big)\,\delta'\big({{{\textbf{\textsl{u}}}}}\cdot{{{\textbf{\textsl{x}}}}}\big ) \,{\mathrm{d}}^4{{{\textbf{\textsl{x}}}}}\bigg\}\delta{{{\textbf{\textsl{q}}}}}\,{\mathrm{d}}\tau \\ - \int_{\tau_1}^{\tau_2 } \int_{{{\xi_{{{{\textbf{\textsl{q } } } } } } } } } { { \boldsymbol { { \nabla } } } _ { \text{g}}}{\otimes}{{{\textbf{\textsl{a}}}}}({{{\textbf{\textsl{q}}}}}+{{{\textbf{\textsl{x}}}}})\cdot\big[{{{\textbf{\textsl{u}}}}}-{{\boldsymbol { { \omega } } } } \cdot{{{\textbf{\textsl{x}}}}}\big ] { f_{\text{e}}}\big({\left\| { { { \textbf{\textsl{x}}}}}\right\| } \big ) \,\delta\big({{{\textbf{\textsl{u}}}}}\cdot{{{\textbf{\textsl{x}}}}}\big ) \,{\mathrm{d}}^4{{{\textbf{\textsl{x}}}}}\,\delta{{{\textbf{\textsl{q}}}}}\,{\mathrm{d}}\tau\ , .
\label{eq : deltasmf}\end{gathered}\ ] ] we next simplify the previous equation with the help of the following useful identity . performing an integration by parts
, one can easily check that @xmath710 \cdot { { \boldsymbol { { \nabla } } } _ { \text{g}}}{\otimes}{{{\textbf{\textsl{a}}}}}({{{\textbf{\textsl{q}}}}}+{{{\textbf{\textsl{x } } } } } ) { f_{\text{e}}}\big({\left\| { { { \textbf{\textsl{x}}}}}\right\| } \big ) \,\delta\big({{{\textbf{\textsl{u}}}}}\cdot{{{\textbf{\textsl{x}}}}}\big ) \,{\mathrm{d}}^4{{{\textbf{\textsl{x}}}}}\\ = \int_{{\mathbb r}^{1,3 } } { { { \textbf{\textsl{a}}}}}({{{\textbf{\textsl{q}}}}}+{{{\textbf{\textsl{x } } } } } ) ( \operatorname{tr}{{\boldsymbol { { \omega } } } } ) { f_{\text{e}}}\big({\left\| { { { \textbf{\textsl{x}}}}}\right\| } \big ) \,\delta\big({{{\textbf{\textsl{u}}}}}\cdot{{{\textbf{\textsl{x}}}}}\big ) \,{\mathrm{d}}^4{{{\textbf{\textsl{x}}}}}\\ \qquad - \int_{{{\xi_{{{{\textbf{\textsl{q } } } } } } } } } { { { \textbf{\textsl{a}}}}}({{{\textbf{\textsl{q}}}}}+{{{\textbf{\textsl{x}}}}})\big [ { { { \textbf{\textsl{u}}}}}\cdot{{{\textbf{\textsl{x}}}}}-{{{\textbf{\textsl{x}}}}}\cdot{{\boldsymbol { { \omega } } } } \cdot{{{\textbf{\textsl{x}}}}}\big ] { f_{\text{e}}}'\big({\left\| { { { \textbf{\textsl{x}}}}}\right\| } \big)/{\left\| { { { \textbf{\textsl{x}}}}}\right\| } \,\delta\big({{{\textbf{\textsl{u}}}}}\cdot{{{\textbf{\textsl{x}}}}}\big ) \,{\mathrm{d}}^4{{{\textbf{\textsl{x}}}}}\\ - \int_{{{\xi_{{{{\textbf{\textsl{q } } } } } } } } } { { { \textbf{\textsl{a}}}}}({{{\textbf{\textsl{q}}}}}+{{{\textbf{\textsl{x}}}}})\,{{{\textbf{\textsl{u}}}}}\cdot\big [ { { { \textbf{\textsl{u}}}}}-{{\boldsymbol { { \omega } } } } \cdot{{{\textbf{\textsl{x}}}}}\big ] { f_{\text{e}}}\big({\left\| { { { \textbf{\textsl{x}}}}}\right\| } \big ) \,\delta'\big({{{\textbf{\textsl{u}}}}}\cdot{{{\textbf{\textsl{x}}}}}\big ) \,{\mathrm{d}}^4{{{\textbf{\textsl{x}}}}}. \label{eq : intermedstepa}\end{gathered}\ ] ] by the anti - symmetry of @xmath119 we have @xmath711 , so that the first integral of the r.h.s . of ( [ eq : intermedstepa ] ) vanishes . again by the anti - symmetry of @xmath119 , we also have that @xmath712 , and furthermore we have @xmath713 .
therefore , the second integral of the r.h.s . of ( [ eq : intermedstepa ] )
vanishes , too . as for the last integral of the r.h.s .
of ( [ eq : intermedstepa ] ) , we use that @xmath714 with @xmath715 , and moreover that @xmath716 , to rewrite it into @xmath717 { f_{\text{e}}}\big({\left\| { { { \textbf{\textsl{x}}}}}\right\| } \big ) \,\delta'\big({{{\textbf{\textsl{u}}}}}\cdot{{{\textbf{\textsl{x}}}}}\big ) \,{\mathrm{d}}^4{{{\textbf{\textsl{x}}}}}\notag \\
= & \qquad\qquad\qquad\int_{{{\xi_{{{{\textbf{\textsl{q } } } } } } } } } { { { \textbf{\textsl{a}}}}}({{{\textbf{\textsl{q}}}}}+{{{\textbf{\textsl{x } } } } } ) \big[1+{{\stackrel{\,\circ}{{{{{\textbf{\textsl{u}}}}}}}}}\cdot{{{\textbf{\textsl{x}}}}}\big ] { f_{\text{e}}}\big({\left\| { { { \textbf{\textsl{x}}}}}\right\| } \big ) \,\delta'\big({{{\textbf{\textsl{u}}}}}\cdot{{{\textbf{\textsl{x}}}}}\big ) \,{\mathrm{d}}^4{{{\textbf{\textsl{x}}}}}\ , .
\label{eq : intermedstepb}\end{aligned}\ ] ] finally , the r.h.s . of ( [ eq : intermedstepb ] )
is integrated by parts once , and we arrive at the final identity @xmath718 \cdot { { \boldsymbol { { \nabla } } } _ { \text{g}}}{\otimes}{{{\textbf{\textsl{a}}}}}({{{\textbf{\textsl{q}}}}}+{{{\textbf{\textsl{x } } } } } ) { f_{\text{e}}}\big({\left\| { { { \textbf{\textsl{x}}}}}\right\| } \big ) \,\delta\big({{{\textbf{\textsl{u}}}}}\cdot{{{\textbf{\textsl{x}}}}}\big ) \,{\mathrm{d}}^4{{{\textbf{\textsl{x}}}}}\notag \\ & \qquad\qquad\qquad = { \frac{{\mathrm{d}}\phantom{s}}{{\mathrm{d}}\tau}}\int_{{{\xi_{{{{\textbf{\textsl{q}}}}}}}}}{{{\textbf{\textsl{a}}}}}({{{\textbf{\textsl{q}}}}}+{{{\textbf{\textsl{x } } } } } ) { f_{\text{e}}}\big({\left\| { { { \textbf{\textsl{x}}}}}\right\| } \big ) \,\delta\big({{{\textbf{\textsl{u}}}}}\cdot{{{\textbf{\textsl{x}}}}}\big ) \,{\mathrm{d}}^4{{{\textbf{\textsl{x}}}}}. \end{aligned}\ ] ] this identity is now used to anti - symmetrize the last integral in ( [ eq : deltasmf ] ) .
noticing next that @xmath719 for an arbitrary four - vector @xmath720 , and recalling that @xmath721 , the variation ( [ eq : deltasmf ] ) of @xmath722 takes the form @xmath723\ , \delta{{{\textbf{\textsl{q}}}}}\,{\mathrm{d}}\tau , \label{eq : deltasmfnew}\ ] ] where @xmath724 is the abraham lorentz type _
minkowski force _ , and @xmath725 { f_{\text{e}}}\big({\left\| { { { \textbf{\textsl{x}}}}}\right\| } \big)\ , \delta({{{\textbf{\textsl{u}}}}}\cdot { { { \textbf{\textsl{x}}}}})\,{\mathrm{d}}^4{{{\textbf{\textsl{x}}}}}\label{eq : nodvikpspinorbit}\ ] ] is nodvik s minkowski momentum for the electromagnetic spin - orbit coupling . for later convenience
, we notice that the charged particle s co - moving electromagnetic field , viz .
its coulomb field and its instantaneous gyro - magnetic field ( both defined w.r.t .
the fermi
walker frame ) do not contribute to @xmath726 , so that @xmath172 in ( [ eq : nodvikpspinorbit ] ) can be replaced by the _
reduced field tensor _ @xmath727 , which obtains from @xmath172 by subtracting the particle s co - moving electromagnetic field . furthermore , making use of the identity @xmath728 , and using the fact that terms proportional to @xmath729 in the integrand vanish upon integration against @xmath730 , we can rewrite the four - vector @xmath726 in the more appealing format @xmath731 where we introduced the symmetric nodvik tensor mass of spin - orbit coupling , @xmath732_+\big]_+ { f_{\text{e}}}\big({\left\| { { { \textbf{\textsl{x}}}}}\right\| } \big)\ , \delta({{{\textbf{\textsl{u}}}}}\cdot { { { \textbf{\textsl{x}}}}})\,{\mathrm{d}}^4{{{\textbf{\textsl{x}}}}}\ , .
\label{eq : mspinorbit}\ ] ] the euler - lagrange equation for @xmath733 under the variation of @xmath102 for given @xmath276 and @xmath171 can now be read off from ( [ eq : deltasmfnew ] ) and ( [ eq : barevar ] ) as @xmath734 where @xmath735 is the four - vector of the _ minkowski momentum _
, with @xmath736 .
equation ( [ eq : alberteq ] ) is the _ minkowski equation of motion _ with the expected abraham lorentz type minkowski force , but with a somewhat unusual anisotropic tensor mass @xmath393 . using ( [ eq : vareul2 ] ) and integrating by parts , observing that the variations of the euler variables vanish at the boundaries , for the first variation of the action @xmath678 of the bare particle we find @xmath737_- \bigr ) \cdot\delta{{\boldsymbol { { \theta } } } } \bigr ) \,{\mathrm{d}}\tau , \ ] ] and for the first variation of the coupling action we get @xmath738_- \bigr ) \cdot\delta{{\boldsymbol { { \theta } } } } \bigr )
\,{\mathrm{d}}\tau\ ] ] where we introduced the tensor of _ electromagnetic minkowski spin ( about @xmath160 ) _ of the charged particle , @xmath739 we remark that @xmath229 is invariant under a gauge transformation @xmath740 .
defining now the _ total spin _ of the particle as @xmath741 and introducing the abbreviation @xmath742_- , \label{eq : ndefine}\ ] ] the vanishing of the first variation of the action functional with respect to all anti - symmetric @xmath668 that are of space - space type with respect to @xmath170 now simply reads @xmath743 at this point we recall that , given any tetrad , any anti - symmetric tensor @xmath76 can be uniquely decomposed into a sum of its space - space and time - space parts . with @xmath170 as the timelike unit vector of the tetrad , we have @xmath744 where @xmath745_+ \text { is the space - space part of $ { { \textbf{\textsf{a}}}}$ , and } \\ { { \textbf{\textsf{a}}}}^\| & { \stackrel{\textrm{\tiny def}}{=}}{{{\textbf{\textsl{u}}}}}\ , \wedge\,({{\textbf{\textsf{a } } } } \cdot { { { \textbf{\textsl{u } } } } } ) \qquad\qquad\qquad \text { its time - space part . }
\end{array } \right.\ ] ] armed with the above decomposition , we conclude that equation ( [ eq : varian ] ) can only be satisfied if the space - space part w.r.t .
@xmath170 of the anti - symmetric tensor @xmath746 vanishes , explicitly @xmath747 which is the euler - lagrange equation for the rotational variations . to extract the evolution equation for @xmath130 from ( [ eq : elrot ] ) , we also decompose the anti - symmetric spin tensor @xmath748 into its space - space and time - space parts with respect to @xmath170 , @xmath749 inserting the above decomposition of @xmath748 into the definition ( [ eq : ndefine ] ) of @xmath746 simply gives @xmath750_- + { \frac{{\mathrm{d}}\phantom{s}}{{\mathrm{d}}\tau}}{{{\textbf{\textsf{s}}}}}^\| + \big[{{\boldsymbol { { \omega } } } } , { { { \textbf{\textsf{s}}}}}^\| \big]_- \ , .\ ] ] with the help of the _ helicity _ four - vector @xmath751 we can write @xmath752 by straightforward calculation we furthermore find that @xmath753_- = { { { \textbf{\textsl{u}}}}}\wedge \left ( { { \stackrel{\,\circ}{{{{\textbf{\textsl{h } } } } } } } } + \ , { \mbox{${\boldsymbol { { \omega } } } _ { \text{\textsc{e}}}$}}\cdot{{\textbf{\textsl{h } } } } \right)\ ] ] and that @xmath754_-\right)\cdot{{{\textbf{\textsl{u}}}}}= { { \textbf{\textsl{0 } } } } \ , .\ ] ] recalling next that @xmath755 ( i.e. @xmath756 , so that @xmath757 ) we find that @xmath758 , viz . @xmath759 from where we read off that @xmath760 , i.e. @xmath761 is itself spacelike for all @xmath111 .
this implies that @xmath762 is spacelike , and we conclude that @xmath763_- & = { { { \textbf{\textsf{n}}}}}^\| \ , , \\ { \frac{{\mathrm{d}}\phantom{s}}{{\mathrm{d}}\tau}}{{{\textbf{\textsf{s}}}}}^\perp + \big[{{\boldsymbol { { \omega } } } } , { { { \textbf{\textsf{s}}}}}^\perp \big]_- & = { { { \textbf{\textsf{n}}}}}^\perp \ , .\end{aligned}\ ] ] hence , the euler - lagrange equation ( [ eq : elrot ] ) reads @xmath764_- = { { \textbf{\textsf{0 } } } } \ , .
\label{eq : elspineq}\ ] ] it remains to show that ( [ eq : elspineq ] ) is identical to ( [ eq : ggeq ] ) .
we recall that @xmath766 , and notice that decomposition into space - space and time - space parts commutes with addition , so that @xmath767 .
we also recall that @xmath768 .
we furthermore notice that @xmath769 , which implies that @xmath770_- = { { \textbf{\textsl{0}}}}$ ] , so that @xmath771_- = \big[{\mbox{${\boldsymbol { { \omega } } } _ { \text{\textsc{fw}}}{}$}},{{{{\textbf{\textsf{s}}}}_{\text{b}}}}\big]_- $ ] . using all these pieces of information
we see that ( [ eq : elspineq ] ) states that @xmath772_- = - { \frac{{\mathrm{d}}\phantom{s}}{{\mathrm{d}}\tau}}{{{{{\textbf{\textsf{s}}}}_{\text{f}}}}^\perp } -\ , \big[{{\boldsymbol { { \omega } } } } , { { { { \textbf{\textsf{s}}}}_{\text{f}}}}^\perp \big]_- \ , .
\label{eq : elspineqrewrite}\ ] ] the l.h.s . of ( [ eq : elspineqrewrite ] ) is manifestly identical to the l.h.s . of ( [ eq : ggeq ] ) . to see that the r.h.s . of ( [ eq : elspineqrewrite ] )
is identical to the r.h.s . of ( [ eq : ggeq ] ) , one can rewrite the calculation in appendix c of @xcite into our notation , using the nodvik maxwell lorentz equations ( [ eq : hommleq ] ) , ( [ eq : inhmleq ] ) , ( [ eq : jnodvik ] ) , and a few integrations by parts . the details are not repeated here . in the renormalization flow study of
our massive led with the choice ( [ eq : shell ] ) for the mass and charge densities , the renormalized particle reaches a finite smallest size @xmath773 in the ` renormalized purely electromagnetic ' limit of vanishing bare rest mass @xmath549 .
this renormalization program for massive led aside , it is of interest to investigate which types of theories emerge in the point limits for mass and charge , either taken separately or jointly . in the point
mass limit @xmath774 but with @xmath181 fixed and ` regular ' ( not a point charge density ) , and with the condition that @xmath775 , our massive led formally reduces to the massive led of nodvik @xcite . in this limit ,
the square root terms in ( [ eq : gyromass ] ) and ( [ eq : barelagrangiandens ] ) become just factors of unity on the support of @xmath776 , so that the bare particle s gyrational mass ( [ eq : gyromass ] ) degenerates identically into a constant , @xmath777 and our bare action reduces simply to the action of a free mechanical point particle with bare rest mass @xmath456 , @xmath778 employed by nodvik @xcite .
the field and coupling actions remain the same , however .
accordingly , the electromagnetic part of the two models is identical , i.e. the maxwell
lorentz equations in nodvik s massive led are given by ( [ eq : hommleq ] ) , ( [ eq : inhmleq ] ) , and ( [ eq : jnodvik ] ) .
moreover , the world - line equation in nodvik s massive led can still be written as @xmath779 but now one has @xmath780 with @xmath781_+\big]_+ { f_{\text{e}}}\big({\left\| { { { \textbf{\textsl{x}}}}}\right\| } \big)\ , \delta({{{\textbf{\textsl{u}}}}}\cdot { { { \textbf{\textsl{x}}}}})\,{\mathrm{d}}^4{{{\textbf{\textsl{x}}}}}\ , , \label{eq : minknodvik}\ ] ] which is a slightly simpler mass tensor than @xmath393 .
on the other hand , the point mass limit eliminates the moment of inertia of the bare particle while the norm of @xmath130 remains bounded ( by hypothesis ) so that @xmath782 identically ; whence , the gyrograph equation ( [ eq : ggeq ] ) degenerates into the statement @xmath783 , i.e. , upon using ( [ eq : minkaltorque ] ) , @xmath784 these are the equations of the massive led of nodvik @xcite .
we will now explain that this is a qualitatively quite different dynamical theory from our massive led .
first of all , since one loses a proper - time derivative for @xmath130 by passing from ( [ eq : ggeq ] ) to ( [ eq : ggeqnodvik ] ) , the point mass limit is a _ singular limit _ from the dynamical perspective of the cauchy problem for @xmath130 .
this fact in itself already means that it will no longer be possible to arbitrarily choose initial data for @xmath130 from the set of admissible cauchy data for ( [ eq : ggeq ] ) . indeed , choose ( which we always may ) a lorentz frame in which initially ( i.e. at @xmath785 ) we have @xmath786 and @xmath787 , and pick the fields @xmath305 on @xmath788 at random from our set of admissible cauchy data for @xmath305 .
then ( [ eq : ggeqnodvik ] ) at @xmath789 has only a nontrivial space part , which in three - space notation reads @xmath790 where @xmath791 is the space part of the dual four - vector for @xmath311 , and where @xmath792 is the initial electric torque on the particle and @xmath793 its initial ` magnetic spin . ' clearly , any attempt to also prescribe cauchy data for the angular velocity tensor @xmath311 , chosen at random from our set of admissible data , will almost surely violate ( [ eq : initconstraint ] ) .
more specifically , ( [ eq : initconstraint ] ) , with ( [ eq : etorque ] ) and ( [ eq : bspin ] ) , imposes not just on the set of admissible initial data for @xmath130 but also on the fields @xmath305 , in the sense that for most choices of @xmath305 _ no _ choice for @xmath311 will satisfy ( [ eq : initconstraint ] ) .
thus , one extracts directly from ( [ eq : initconstraint ] ) , ( [ eq : etorque ] ) , ( [ eq : bspin ] ) that the originally admissible field data are now further restricted as follows : either @xmath316 satisfies @xmath794 , and in this case also @xmath795 has to satisfy @xmath796 ; or @xmath316 satisfies @xmath797 , in which case @xmath795 has to satisfy @xmath798 . only if these alternate constraints on the initial fields are satisfied is it possible to satisfy ( [ eq : initconstraint ] ) by a compatible class of initial data @xmath311 ( viz . , @xmath791 ) .
namely , if @xmath794 and @xmath799 , then the admissible data for @xmath791 are the same as for our model .
if @xmath797 and @xmath800 , however , then @xmath791 has to be of the form @xmath801 which leaves only one degree of freedom to choose from instead of three , i.e. the coefficient @xmath6 in front of @xmath802 is at our disposal .
this clarifies to which extent initial data for nodvik s model have to be compatible .
compatibility of a set of initial data in the sense just explained does not , however , imply that all is well then , for not only has one lost a proper - time derivative on @xmath130 by passing from ( [ eq : ggeq ] ) to ( [ eq : ggeqnodvik ] ) , a proper - time derivative of @xmath130 does not show any more at all in ( [ eq : ggeqnodvik ] ) .
nodvik s ( [ eq : ggeqnodvik ] ) is therefore not an orderly evolution equation for @xmath130 , but rather a _ constraint _
equation for the dynamical field variables @xmath172 and the dynamical particle variables @xmath160 and @xmath170 , which in case of actually being satisfied by these variables for some time interval after the initial instant , still leaves a certain amount of freedom for choosing the value of the now purely kinematical variable @xmath130 .
therefore , to the extent that compatible initial data for @xmath172 , @xmath160 , and @xmath170 , with a compatible initial choice of @xmath130 , can be continued at all according to nodvik s equations , such a continuation is presumably not unique .
ever since dirac s surmise @xcite that , as long as we are willing to let the bare mass take the value @xmath803 to compensate for the @xmath804 electromagnetic self - energies , we may view the abraham lorentz dirac ( ald ) equation @xmath805 as the _ exact _ equation for the radiation reaction - corrected motion of a spinless point particle with empirical electron mass @xmath540 and empirical electron charge @xmath539 in external electromagnetic fields @xmath806 , there has been a strong interest in making sense out of the peculiar third - order proper - time derivative of @xmath160 that figures in ( [ eq : aldeq ] ) , e.g. @xcite . a rigorous clarification of dirac s surmise is , however , not so easy to come by .
notable rigorous attempts to construct a consistent renormalized electrodynamics with spinless point charges are @xcite and @xcite ; see also @xcite and @xcite for work on the classical pauli fierz model , a semi - relativistic dipole approximation to the problem , which also serves as point of departure for setting up a so - called non - relativistic qed @xcite . naturally one may wonder whether ( [ eq : aldeq ] ) emerges in some spinless point particle limit of our massive led , or to begin simpler , in the point particle limit of a spinless version of our massive led . this massive led without spin obtains from our model by simply discarding ( [ eq : ggeq ] ) and setting @xmath423 in the remaining equations .
thus , the relativistic maxwell lorentz equations are once again @xmath807 for the hodge dual of @xmath172 , and @xmath808 for @xmath172 itself , but now with @xmath809 the world - line equation still reads @xmath810 but now with @xmath811 and with @xmath812 given in ( [ eq : ueqsl ] ) . notice that the structure of @xmath145 does not figure in this spinless led , which is therefore identical to the spinless version of nodvik s led , due to spohn @xcite , in which @xmath145 is a point mass distribution from the outset .
the question then is : what happens if we let @xmath181 tend to a delta function concentrated at the same point ?
the first step to the answer is the rigorous control of this spinless massive led which , however , is not very developed yet . in this appendix
we give a brief summary of semi - relativistic led .
this material can also be used as a primer for the relativistic theory .
we begin with the massive model with spin .
its limiting cases of the more familiar abraham model and the popular ` model without spin ' are discussed in subsequent subsections . as in the relativistic theory , the electric field @xmath813 and the magnetic field @xmath814 at the point @xmath815 at time @xmath816 satisfy the classical maxwell
lorentz equations , @xmath817 although not manifestly lorentz co - variant , the maxwell lorentz equations ( [ eq : mldivb ] ) , ( [ eq : mlrote ] ) , ( [ eq : mldive ] ) , ( [ eq : mlrotb ] ) appear to have a relativistic format already , but whether in fact they do satisfy relativistic lorentz transformation laws or not depends on the model for the charge distribution . in the classical works of abraham @xcite before the dawn of special relativity @xcite , the electrical charge and current densities formally combine as @xmath818 but do _ not _ transform like a relativistic four - current density vector . in the semi - relativistic formulation of abraham , the charge of a particle ( @xmath539 for a model of the electron ) is distributed around the instantaneous location @xmath819 of the particle by a charge density @xmath181 with @xmath820 symmetry and compact support , satisfying @xmath821 .
this charge density is rigidly carried along by the particle with linear velocity @xmath822 and rotating rigidly with angular velocity @xmath823 .
thus , with a single charged particle source in ( [ eq : mldive ] ) , ( [ eq : mlrotb ] ) , we have @xmath824 and @xmath825 the evolution equations for the dynamical variables of the particle , i.e. position @xmath826 , linear velocity @xmath822 , and angular velocity @xmath823 , are likewise not lorentz covariant and given by newton s and euler s evolution equations equipped with the abraham lorentz expressions for the volume - averaged lorentz force and torque @xcite , respectively , that act on the charged particle .
newton s equation of motion here reads @xmath827 { f_{\text{e}}}(|{{\boldsymbol { { x } } } } -{{{\boldsymbol { { q } } } } } ( t)| ) \ , { \mathrm{d}}^3x\ , , \label{eq : aleqtmb}\ ] ] where @xmath828 is the newtonian linear particle momentum , with @xmath456 the bare mass ( ` material mass ' in @xcite ) .
euler s equation of motion reads @xmath829 { f_{\text{e}}}(|{\boldsymbol { { x } } } -{\boldsymbol { { q } } } ( t)| ) \ , { \mathrm{d}}^3x\ , , \label{eq : aleqrmb}\ ] ] where @xmath830 is the classical particle spin associated with the bare moment of inertia @xmath831 . with the bare inertias @xmath832 and @xmath833 ,
the semi - relativistic equations listed above pose a cauchy problem for the following initial data at time @xmath834 : for the mechanical variables of the particles , the data are @xmath835 , @xmath836 ( equivalently : @xmath837 ) , and @xmath838 ( equivalently : @xmath839 ) ; for the fields , the data are @xmath840 satisfying ( [ eq : mldivb ] ) at @xmath841 , and @xmath842 satisfying ( [ eq : mldive ] ) at @xmath841 . notice that ( [ eq : mldivb ] ) and ( [ eq : mldive ] ) are merely initial constraints , which then remain satisfied by the fields @xmath843 and @xmath844 in the ensuing evolution . for ( [ eq : mldivb ] ) this is seen by taking the divergence of ( [ eq : mlrote ] ) . for ( [ eq : mldive ] ) this is seen by taking the divergence of ( [ eq : mlrotb ] ) and the time - derivative of ( [ eq : mldive ] ) , then using the continuity equation for the charge , @xmath845 which is demanded by the maxwell
lorentz equations and indeed satisfied for @xmath846 given by ( [ eq : alcharge ] ) and @xmath847 by ( [ eq : alvfield ] ) ; cf .
@xcite .
the semi - relativistic massive led with spin has the following conservation laws @xcite .
charge , @xmath848 energy , @xmath849 linear momentum , @xmath850 and angular momentum , @xmath851 are all conserved during the evolution . a mathematically regular limit which , unfortunately , is not quite physical but brings about some considerable simplification of the dynamical equations is to send @xmath852 while @xmath456 is fixed at some finite value , in such a way that @xmath853 converges to some finite vector - valued function of @xmath270 , which of course implies that @xmath854 .
the semi - relativistic maxwell lorentz equations are again given by ( [ eq : mldivb ] ) , ( [ eq : mldive ] ) , ( [ eq : mlrote ] ) , ( [ eq : mlrotb ] ) , and the charge density is still ( [ eq : alcharge ] ) , but the current density vector now simplifies to @xmath855 moreover , the newtonian equation of motion ( [ eq : aleqt ] ) reduces to @xmath856 { f_{\text{e}}}(|{{\boldsymbol { { x } } } } -{{{\boldsymbol { { q } } } } } ( t)| ) \
, { \mathrm{d}}^3x \ , , \label{eq : aleqtinfi}\ ] ] with @xmath857 still given by ( [ eq : pdefnewton ] ) , and the eulerian equation ( [ eq : aleqr ] ) reduces to @xmath858 { f_{\text{e}}}(|{{\boldsymbol { { x } } } } -{{{\boldsymbol { { q } } } } } ( t)| ) \ , { \mathrm{d}}^3x \ , .\label{eq : aleqrinfi}\ ] ] in this limit , the maxwell lorentz equations ( [ eq : mldivb ] ) , ( [ eq : mldive ] ) , ( [ eq : mlrote ] ) , ( [ eq : mlrotb ] ) , equipped with the charge density ( [ eq : alcharge ] ) and current density vector ( [ eq : alcurrentnogyro ] ) , and the newtonian equation of motion ( [ eq : aleqtinfi ] ) , form a non - linearly coupled , _ closed _ system of equations , with initial data at @xmath841 given by : @xmath835 and @xmath836 ( equivalently : @xmath837 ) for the particle ; for the fields , the data are @xmath859 satisfying ( [ eq : mldivb ] ) at @xmath841 , and @xmath860 satisfying ( [ eq : mldive ] ) at @xmath841 .
spin in turn adjusts passively to this combined field - particle dynamics according to a linear equation ( [ eq : aleqrinfi ] ) , which has to be supplemented by the initial data @xmath839 ( now of course no longer equivalent to prescribing @xmath838 ) .
consequently , in this limit spin drops from the conservation laws with the exception of the law for the angular momentum .
thus , the conserved energy @xmath861 , linear momentum @xmath862 , and angular momentum @xmath863 are now given by @xmath864 @xmath865 and @xmath866 notice that the laws for energy and linear momentum do not feature @xmath526 , but the conservation law for the angular momentum does .
furthermore , the continuity equation simplifies to @xmath867 but this does not affect the law of charge conservation , which is still given by ( [ eq : totalcharge ] ) . since , in the limit @xmath868 , spin is of no dynamical relevance , apart from the law of angular conservation , ( [ eq : aleqrinfi ] ) is frequently omitted from the system of equations .
the resulting subset of equations defines the semi - relativistic massive led ` without spin . '
this model without spin is still invariant under rotations , and by e. noether s theorems there exists a conserved quantity associated with this invariance ; only this quantity does not qualify as the physical angular momentum in the conventional sense .
the observation that discarding the gyroscopic equation ( [ eq : aleqrinfi ] ) violates the conventional law of angular momentum conservation ( [ eq : totaldrehimpulsinfi ] ) was made in @xcite .
the semi - relativistic massive led without spin is the most thoroughly understood version of led .
in particular , we mention the recent rigorous treatments in @xcite , where for technical reasons the einsteinian linear momentum @xmath869 with @xmath832 is used instead of the newtonian ( [ eq : pdefnewton ] ) ( see also @xcite for work on a simpler scalar field - particle theory ) . in this case , of course also the newtonian kinetic energy in ( [ eq : totalenergy ] ) is to be replaced by its einsteinian counterpart @xmath870 in @xcite , the global existence and uniqueness for the cauchy problem of a particle without spin and @xmath832 was proven
. moreover , in @xcite it was shown that the motion is stable if @xmath408 and unstable if @xmath871 .
the papers @xcite address the long time asymptotics of the slow motion of a spinless particle with @xmath408 in slowly varying external fields . using center manifold theory , these authors derive effective equations of motion of second order in the time derivative .
originally these effective second - order equations were obtained in landau
lifshitz @xcite by applying a heuristic closure argument to the ( in)famous third - order term in the abraham lorentz dirac ( ald ) equation ( [ eq : aldeq ] ) , which works equally well for the relativistic and for the semi - relativistic version .
as far as we know , the rigorous work of spohn and collaborators gives the first clean derivation of the landau
lifshitz closure for the ald equation , so far in its semi - relativistic setting .
abraham s @xcite purely electromagnetic model obtains from the massive semi - relativistic led with spin in the formal double limit @xmath457 and @xmath872 together with the assumption that the velocity @xmath873 and angular velocity @xmath874 of the particle remain bounded . the maxwell
lorentz equations are not affected by this limit and still given by ( [ eq : mldivb ] ) , ( [ eq : mlrote ] ) , ( [ eq : mldive ] ) , ( [ eq : mlrotb ] ) together with ( [ eq : alcharge ] ) and ( [ eq : alvfield ] ) . however , the bare newtonian momentum @xmath875 and bare eulerian spin @xmath876 identically in this limit . as a consequence , the evolution equations ( [ eq : aleqtmb ] ) and ( [ eq : aleqrmb ] ) degenerate into @xmath877 { f_{\text{e}}}(|{{\boldsymbol { { x } } } } -{{{\boldsymbol { { q } } } } } ( t)| ) \ , { \mathrm{d}}^3x \label{eq : aleqt}\ ] ] and @xmath878 { f_{\text{e}}}(|{{\boldsymbol { { x } } } } -{{{\boldsymbol { { q } } } } } ( t)| ) \ , { \mathrm{d}}^3x \ , .
\label{eq : aleqr}\ ] ] the conservation laws for energy , linear and angular momentum
simplify accordingly , because @xmath879 and @xmath526 vanish quadratically fast with @xmath457 and @xmath872 while @xmath880 and @xmath874 remain bounded ( by hypothesis ! ) .
thus , the conserved energy @xmath861 , linear momentum @xmath862 , and angular momentum @xmath863 are now given by @xmath881 @xmath882 and @xmath883 the continuity equation ( [ eq : continuitylaw ] ) and its offspring , the law for the charge ( [ eq : totalcharge ] ) , still hold as stated .
this purely electromagnetic abraham model with spin is not well - studied at all .
in fact , it is hardly mentioned in textbooks and monographs on classical electromagnetism , with the notable exception of the book by miller @xcite .
subjecting equations ( [ eq : aleqt ] ) and ( [ eq : aleqr ] ) to closer scrutiny reveals that abraham s purely electromagnetic semi - relativistic led with spin is qualitatively a very different dynamical theory from the massive semi - relativistic led with spin . in fact ,
since the coefficients in front of the time derivative of both @xmath880 and @xmath874 vanish , these equations constitute a _ singular limit _ of the regular evolution equations ( [ eq : aleqtmb ] ) and ( [ eq : aleqrmb ] ) , a limit in which cauchy data for the dynamical particle variables @xmath880 and @xmath874 can no longer be prescribed freely ; also the field data are severely restricted .
only the initial particle position can be prescribed arbitrarily in abraham s model .
the situation is similar as in nodvik s relativistic model , only worse .
more specifically , choose the initial time to be @xmath449 and the initial particle position to be @xmath884 , which can always be achieved by at most a time and a space translation .
as before , denote the initial fields by @xmath885 and @xmath886 .
then ( [ eq : aleqt ] ) demands that at @xmath449 , we have @xmath887 while ( [ eq : aleqr ] ) demands ( as for nodvik s model ) that at @xmath449 we have @xmath888 here , the averaging is defined by @xmath889 , where we assume with abraham that @xmath890 ; furthermore , @xmath891 now means the three - tensor product .
one readily deduces from ( [ eq : alinitt ] ) and ( [ eq : alinitr ] ) that only the following subset of the originally admissible field data is now admitted .
namely , if @xmath892 , then also @xmath893 must hold ; if @xmath894 , then @xmath895 must hold ; if @xmath896 , then @xmath897 must be in the orthogonal complement of the left kernel space of @xmath898 , unless this tensor is the zero tensor , in which case @xmath897 must vanish also ; if @xmath899 is in the left kernel space of @xmath900 , then also @xmath901 must hold . _
remark : _ the above set of conditions on the initial fields imply in particular that the abraham equations admit _ no solution _ for the following setup : @xmath902 , @xmath903 , and @xmath904 , i.e. the linear superposition of the charge s radial symmetric coulomb field with a homogeneous external field . in this case
no choice of @xmath905 and @xmath906 will rescue ( [ eq : alinitt ] ) ; in short , these perfectly physical initial conditions are _ inconsistent _ for the abraham model .
we continue with the discussion of consistent initial data .
given now consistent data for the fields in accordance with the restrictions just explained , the initial data for @xmath905 and @xmath906 are arbitrary only if @xmath896 and @xmath907 , respectively ; otherwise they must be of the following form . if @xmath908 , then @xmath909 which leaves only one degree of freedom to choose from instead of three , viz .
the coefficient @xmath328 is at our disposal . if @xmath899 , then @xmath910 which again leaves only one degree of freedom to choose from instead of three , here the coefficient @xmath911 is at our disposal . the extent to which such consistently restricted initial data now launch a unique regular solution of the system of abraham equations ( [ eq : aleqt ] ) , ( [ eq : aleqr ] ) , ( [ eq : mldivb ] ) , ( [ eq : mlrote ] ) , ( [ eq : mldive ] ) , and ( [ eq : mlrotb ] ) , with ( [ eq : alcharge ] ) and ( [ eq : alvfield ] ) ,
is currently not known . the purely electromagnetic abraham model without spin @xcite obtains from the abraham model with spin by simply discarding ( [ eq : aleqr ] ) and setting @xmath912 identically in ( [ eq : alvfield ] ) , so that ( [ eq : aleqt ] ) goes over into @xmath913 { f_{\text{e}}}(|{{\boldsymbol { { x } } } } -{{{\boldsymbol { { q } } } } } ( t)| ) \ , { \mathrm{d}}^3x\ , , \label{eq : aleqtnogyro}\ ] ] while the maxwell
lorentz equations for this model are the same as for the massive model without spin , i.e. again given by ( [ eq : mldivb ] ) , ( [ eq : mldive ] ) , ( [ eq : mlrote ] ) , ( [ eq : mlrotb ] ) , with the charge density ( [ eq : alcharge ] ) and with the current density vector ( [ eq : alcurrentnogyro ] ) .
the continuity equation for the charge is the simpler ( [ eq : continuitylawinfi ] ) , while the law of charge conservation is of course just ( [ eq : alcharge ] ) .
the conserved energy , linear and angular momentum are given again by the purely electromagnetic expressions ( [ eq : totalenergyal ] ) , ( [ eq : totalimpulsal ] ) , and ( [ eq : totaldrehimpulsal ] ) that apply to the abraham model with spin .
the abraham model without spin can also be obtained as the singular limit @xmath457 of the massive semi - relativistic model without spin under the hypothesis that @xmath880 remains bounded . for the abraham model without spin a similar assessment holds as for the abraham model with spin . again choosing the initial time to be @xmath449 and the initial particle position to be @xmath914 ( after at most a time and a space translation ) , the restrictions for the field data
are now determined by ( [ eq : alinitt ] ) with @xmath915 ( and ignoring the discarded ( [ eq : alinitr ] ) ) .
we conclude that , if @xmath916 then also @xmath917 must hold , and if @xmath899 , then @xmath918 must hold .
although these are milder restrictions than those for the abraham model with spin , these restrictions on the field data in the abraham model without spin still rule out various perfectly physical initial data , for instance those remarked on in the previous subsection , i.e. @xmath902 , @xmath903 , and @xmath904 . as for the initial velocity of the particle @xmath905 in given consistently restricted initial fields , the following rules hold .
the initial velocity @xmath905 is at our disposal only if @xmath916 , and if @xmath919 it is of the form @xmath920 which , once again , leaves us only the choice of @xmath911 .
whether such a restricted , consistent set of initial data can now be continued into a unique regular solution of ( [ eq : aleqtnogyro ] ) , ( [ eq : mldivb ] ) , ( [ eq : mldive ] ) , ( [ eq : mlrote ] ) , and ( [ eq : mlrotb ] ) , with ( [ eq : alcharge ] ) and ( [ eq : alcurrentnogyro ] ) , is currently not rigorously known . as a matter of fact , there are good reasons to doubt that the evolution problem is well - defined even given consistent initial data . for notice that ( [ eq : aleqtnogyro ] ) states that at @xmath921 for _ all _ @xmath270 , where the averages here are now defined by @xmath922 hence , we can repeat our discussion of the conditions on the initial data essentially verbatim to conclude that
: * if at any time @xmath270 we have @xmath923 , then ( [ eq : aleqtnogyrot ] ) has no solution @xmath924 at that instant @xmath270 unless also @xmath925 ; * if @xmath926 , then ( [ eq : aleqtnogyrot ] ) has no solution @xmath924 at @xmath270 unless also @xmath927 ; * if @xmath926 and @xmath927 , then ( [ eq : aleqtnogyrot ] ) states that @xmath924 is of the form @xmath928 with an _ a priori _ undetermined coefficient @xmath929 . while there is hardly any literature on the abraham model with spin , the abraham model without spin
is discussed in most major texts since abraham @xcite , e.g. @xcite , and also in the more leisurely text of peierls @xcite . however
, none of these texts makes the elementary observation that the initial value problem for the purely electromagnetic abraham model(s ) is singular in the sense explained above , nor that abraham s ` evolution ' equation ( [ eq : aleqtnogyrot ] ) preserves this singular structure in time ; some of the main conclusions drawn in @xcite are even in open conflict with the singular nature of abraham s evolution equations .
a re - assessment of these traditional treatments of the abraham model seems to be called for .
abraham , m. , `` prinzipien der dynamik des elektrons , '' _ phys .
z. _ * 4 * , ( 1902 ) 57 ; _ ann .
* 10 * , ( 1903 ) 105 .
abraham , m. , `` theorie der elektrizitt , '' vol .
ii , teubner , leipzig , 1905 .
lorentz , h.a .
, `` electromagnetic phenomena in a system moving with any velocity less than that of light , '' _ proc .
amsterdam , _ * 6 * ( 1904 ) , 809 .
poincar , h. , `` sur la dynamique de llectron , '' _ comptes - rendus _ * 140 * , ( 1905 ) 1504 ; _ rendiconti del circolo matematico di palermo _ * 21 * , ( 1906 ) 129 .
miller , a. i. , `` albert einstein s special theory of relativity , '' springer , new york , 1998 .
einstein , a. , `` zur elektrodynamik bewegter krper , '' _ ann .
* 17 * , ( 1905 ) 891 .
einstein , a. , `` ist die trgheit eines krpers von seinem energieinhalt abhngig ?
, '' _ ann .
phys . _ * 18 * , ( 1905 ) 639 .
minkowski , h. , `` die grundgleichungen fr elektromagnetische vorgnge in bewegten krpern , '' _ nachr .
knigl . ges .
gttingen , _ math .
( 1908 ) 53 .
uhlenbeck , g.e . , and goudsmit , s.a .
, `` spinning electrons and the structure of spectra , '' _ nature _ * 117 * , ( 1926 ) 264 .
tomonaga , s. , `` the story of spin , '' univ .
chicago press , chicago , 1997 .
glimm , j. , and jaffe , a. , `` quantum physics , '' 2nd . ed . ,
springer , new york , 1980 .
fernandez , r. , frhlich , j. , and sokal , a. , `` random walks , critical phenomena , and triviality in quantum field theory , '' springer , new york , 1992 .
lorentz , h.a .
, `` the theory of electrons and its applications to the phenomena of light and radiant heat , '' reprint of the 2nd ed . , 1915 , by dover , new york , 1952 .
landau , l. , and lifshitz , e.m .
, `` the theory of classical fields , '' pergamon press , oxford , 1962 .
nodvik , j.s .
, `` a covariant formulation of classical electrodynamics for charges of finite extension , '' _ ann .
_ * 28 * , ( 1964 ) 225 .
misner , c.w . , thorne , k.s . , and wheeler , j.a . , `` gravitation , '' w.h .
freeman co. , new york , 1973 .
streater , r.f . , and wightman , a. , `` pct , spin and statistics , and all that , '' princeton univ .
press , princeton , 1964 .
barut , a.o .
, `` electrodynamics and classical theory of fields and particles , '' dover , new york , 1964 .
gittel , h .-
p . , kijowski , j. , and zeidler , e. , `` the relativistic dynamics of the combined particle - field system in renormalized classical electrodynamics , '' _ commun .
_ * 198 * , ( 1998 ) 711 .
thomas , l.h .
, `` the motion of the spinning electron , '' _ nature _ * 117 * , ( 1926 ) 514 ; `` on the kinematics of an electron with an axis , '' _ phil . mag . _ * 3 * , ( 1927 ) 1 .
bargmann , v. , michel , l. , and telegdi , v.l . , `` precession of the polarization of particles moving in a homogeneous electromagnetic field , '' _ phys .
* 2 * , ( 1955 ) 435 .
fermi , e. , `` ber einen widerspruch zwischen der elektrodynamischen und der relativistischen theorie der elektromagnetischen masse , '' _ physik .
zeitschr . _
* 23 * , ( 1922 ) 340 .
rohrlich , f. , `` self - energy and stability of the classical electron , '' _ am .
_ * 28 * , ( 1960 ) 639 .
panofsky , w.k.h . , and phillips , m. , `` classical electricity and magnetism , '' 2nd edition , addison - wesley , reading , mass . , 1962 .
feynman , r.p . ,
leighton , r.b . , and sands , m. , `` the feynman lectures in physics , '' addison - wesley , ma , 1963 .
jackson , j.d .
, `` classical electrodynamics , '' 3rd ed . , wiley , new york , 1999 .
rohrlich , f. , `` classical charged particles , '' 2nd .
ed . , addison wesley , redwood city , ca , 1990 .
schwinger , j. , `` electromagnetic mass revisited , '' _ found .
* 13 * , ( 1983 ) 373 .
yaghjian , a.d .
, `` relativistic dynamics of a charged sphere , '' lect . notes phys . *
m11 * , springer , berlin , 1992 .
barut , a.o . , and pavi , `` radiation reaction and the electromagnetic energy - momentum of moving relativistic charged membranes , '' _ phys .
b _ * 331 * , ( 1994 ) 45 . weinberg , s. , `` gravitation and cosmology : principles and applications of the general theory of relativity , '' wiley , new york , 1972 .
christodoulou , d. , `` the action principle and partial differential equations , '' ann . math . stud . * 146 * , princeton univ . press , princeton , 2000 .
appel , w. , and kiessling , m.k .- h .
, in preparation for _ lett .
_ ( 2000 ) . hawking , s.w . , and ellis , g.f.r . ,
`` the large scale structure of space - time , '' cambridge univ . press , cambridge , 1973 .
christodoulou , d. and klainerman , s. , `` the global nonlinear stability of the minkowski space , '' princeton math . ser . * 41 * , princeton univ . press , princeton , 1993 .
thorne , k. , price , r.h . , and mcdonald , d.a .
, `` black holes : the membrane paradigm , '' yale univ . press , yale , 1986 .
reed , m. , and simon , b. , `` scattering theory , '' ( methods of modern mathematical physics iii ) , acad . press , orlando , 1979 .
sigal , i.m . and soffer , a. , `` the @xmath930-particle scattering problem : asymptotic completeness for short - range systems , '' _ ann .
_ * 126 * , ( 1987 ) 35 .
sigal , i.m . and soffer , a. , `` asymptotic completeness for @xmath930-particle long - range scattering , ''
soc . _ * 7 * , ( 1994 ) 307 .
drr , d. , goldstein , s. , teufel , s. and zangh , n. `` scattering theory from microscopic first principles , '' _ physica _ * a 279 * , ( 2000 ) 416 .
kunze , m. , `` instability of the periodic motion of a particle interacting with a scalar wave field , '' _ comm .
phys . _ * 195 * , ( 1998 ) 509 .
soffer , a. , and weinstein , m. , `` resonances , radiation damping , and instability in hamiltonian nonlinear wave equations , '' _ inv . math . _ * 136 * , ( 1999 ) 9 .
spohn , h. , `` dynamics of charged particles and their radiation field , '' monograph preprint ( 131 pages ) , tu mnchen , 1999 .
cottingham , w.n . , `` the isolated electron , '' _ in _ `` electron a centenary volume , '' ( m. springford , ed . ) , p. 24
, cambridge univ . press ,
cambridge , 1997 .
dyson , f. , `` the radiation theories of tomonaga , schwinger , and feynman , '' _ phys .
rev . _ * 75 * , ( 1949 ) 486 ; `` the s matrix in quantum electrodynamics , '' _ ibid . _
dyson , f. , `` divergence of perturbation theory in quantum electrodynamics , '' _ phys .
rev . _ * 85 * , ( 1952 ) 631 .
balaban , t. , _ private communications _ ( 1999 ) .
wightman , a. _ private communication _
wilczek , f. , `` the future of particle physics as a natural science , '' _ in _ `` critical problems in physics , '' ( v. fitch , d. marlow , m. dementi , eds . ) , p. 281
, princeton univ . press , princeton 1997 .
witten , e. , `` vistas in theoretical physics , '' _ in _ `` critical problems in physics , '' ( v. fitch , d. marlow , m. dementi , eds . ) , p. 271
, princeton univ . press , princeton 1997 .
connes , a. , `` noncommutative geometry , '' academic press , new york , 1994 .
schweber , s. , `` qed and the man who made it : dyson , feynman , schwinger , and tomonaga , '' princeton univ . press , princeton 1994 .
griesemer , m. lieb , e.h . , and loss , m. , `` ground states in non - relativistic qed , '' lanl e - print arxive math - ph/0007014 ( jul .
kontsevich , m. , `` deformation quantization of poisson manifolds , '' lanl e - print arxive //http://xxx.lanl.gov / q - alg/9709040 ( 1997 ) .
nyborg , p. , `` on classical theories of spinning particles , '' _
nuovo cim . _
* 23 * , ( 1962 ) 47 .
benguria , r. , `` electrodynamics and the equations of motion for charged particles with spin , '' _ phys .
d. _ * 13 * , ( 1976 ) 274 .
glassey , r. , and schaeffer , j. , `` global existence for the relativistic vlasov
maxwell system with nearly neutral initial data , '' _ comm .
phys . _ * 119 * , ( 1988 ) 353 .
glassey , r. , and strauss , w. , `` absence of shocks in an initially dilute collisionless plasma , '' _ comm .
phys . _ * 113 * , ( 1987 ) 191 .
glassey , r. , and strauss , w. `` singularity formation in a collisionless plasma could occur only at high velocities , '' _ arch .
* 92 * , ( 1986 ) 59 .
vlasov , a.a .
, `` many - particle theory and its application to plasma , '' gordon and breach , new york , 1961 ; published originally by state publishing house for technical - theoretical literature , moscow and leningrad , 1950 .
dirac , p.a.m .
, `` classical theory of radiating electrons , '' _ proc .
a _ * 167 * , ( 1938 ) 148 .
carati , a. , delzanno , p. , galgani , l. , and sassarini , j. , `` nonuniqueness properties of the physical solutions of the lorentz dirac equation , '' _ nonlinearity _ * 8 * , ( 1995 ) 65 .
gaftoi , v. , lpez - bonilla , j. l. , morales , j. , and rosales , m. a. , `` the lorentz
dirac equation in minkowski space , '' _ j. math
. phys . _ * 35 * , ( 1994 ) 3482 .
luca , j. de , `` the lorentz
dirac equation for linear and nonlinear potentials , '' _ braz .
j. phys . _
* 27 * , ( 1997 ) 285 .
peierls , r. , `` more surprises in theoretical physics , '' princeton univ . ,
princeton , 1991 .
bambusi , d. , and noja , d. , `` some rigorous results on the pauli fierz model of classical electrodynamics , '' _ lett .
math phys . _
* 37 * , ( 1996 ) 449 .
bambusi , d. , and galgani , l. , `` some rigorous results on the pauli fierz model of classical electrodynamics , '' _ ann .
h. poincar , phys .
theor . _ * 58 * , ( 1993 ) 155 .
noja , d. , and posilicano , a. , `` on the point limit of the pauli fierz model , '' _ ann .
h. poincar , phys .
theor . _ * 71 * , ( 1999 ) 425 .
noja , d. , and posilicano , a. , `` delta interactions and electrodynamics of point particles , '' los alamos e - print //http://xxx.lanl.gov
/ math - ph/9907009 ( 1999 ) .
lorentz , h.a .
, `` la thorie lectromagnetique de maxwell et son application aux corps mouvants , '' _ arch .
exactes nat . _ * 25 * , ( 1892 ) 363 .
kiessling , m. k .- h .
, `` classical electron theory and conservation laws , '' _ phys .
a _ * 258 * , ( 1999 ) 197 .
bauer , g. , and drr , d. , `` the maxwell lorentz system of a rigid charge distribution , '' preprint , lmu mnchen , 1999 .
komech , a. , and spohn , h. , `` long - time asymptotics for the coupled maxwell lorentz equations , '' _ commun
. pde _ * 25 * , ( 2000 ) 559 .
kunze , m. , and spohn , h. , `` adiabatic limit of the maxwell
lorentz equations , '' ann . inst
. h. poincar , phys .
thor . , ( 2000 ) in press .
kunze , m. , and spohn , h. , `` radiation reaction and center manifolds , '' _ siam j. math . anal .
_ , ( 2000 ) in press .
kunze , m. , and spohn , h. , slow motion of charges interacting through the maxwell field , _ commun .
* 212 * , ( 2000 ) 437 .
komech , a. , spohn , h. , and kunze , m. , `` long - time asymptotics for a classical particle interacting with a scalar wave field , '' _ commun
. pde _ * 22 * , ( 1997 ) 307 .
komech , a. , kunze , m. , and spohn , h. , `` effective dynamics for a mechanical particle coupled to a wave field , '' _ commun .
phys . _ * 203 * , ( 1999 ) 1 .
lorentz , h.a .
, `` versuch einer theorie der elektrischen und optischen erscheinungen in bewegten krpern , '' teubner , leipzig , 1909 ; orig . :
brill , leyden , 1895 .
lorentz , h.a .
, `` weiterbildung der maxwellschen theorie : elektronentheorie . , '' _ in _ encyklopdie d. mathematischen wissenschaften @xmath931 , art .
14 , p. 145
sommerfeld , a. , `` electrodynamics , '' academic press , new york , 1952 . | lorentz objection to the uhlenbeck - goudsmit proposal of a spin magnetic moment of the electron , namely that the electron s equatorial rotation speed would exceed the speed of light by a factor @xmath0 , mutated into an objection against lorentz electrodynamics ( led ) itself when the spin magnetic moment became established .
however , lorentz renormalization calculation , based on the early @xmath1 century s notion of a purely electromagnetic electron , does not qualify as proper from a modern perspective .
this paper shows that renormalization treated properly does lead to a mathematically consistent and physically viable led . a new , relativistically covariant massive led
is presented in which the bare particle has a finite positive bare rest mass and moment of inertia .
the particle s electromagnetic self - interaction renormalizes its mass and spin .
most crucially , the renormalized particle is a _ soliton _ : after any scattering process its rest mass and spin magnitude are dynamically restored to their pre - scattering values .
this guarantees that `` an electron remains an electron , '' poetically speaking . a renormalization flow study of the limit of vanishing bare rest mass is conducted for this model .
this limit yields a purely electromagnetic classical field theory with ultra - violet cutoff at about the electron s compton wavelength ! the renormalized limit model matches the empirical electron data as orderly as one can hope for at the level of lorentz theory .
in particular , no superluminal equatorial speeds occur . ' '' ''
\(2000 ) the authors . reproduction for non - commercial purposes of this article , in its entirety and by any means , is permitted . |
the discovery of kilohertz quasi - periodic brightness oscillations ( qpos ) from many neutron - star low - mass x - ray binaries ( lmxbs ) has given us a sensitive probe of the conditions near accreting neutron stars ( see van der klis 1998 for a review of the properties of kilohertz qpos ) .
a variety of models have been suggested for this phenomenon , including beat - frequency models ( strohmayer et al .
1996 ; miller , lamb , & psaltis 1998 ) , relativistic precession models ( stella & vietri 1998 ) , and disk oscillation models ( e.g. , osherovich & titarchuk 1999 ) . in any such model , the brightness oscillations can be produced in two general ways . in a pure _ luminosity
_ oscillation , the total luminosity of the source changes quasi - periodically but the angular distribution of the radiation remains fixed . in
a pure _ beaming _ oscillation , the total luminosity remains constant but the angular distribution of the radiation changes and hence the flux in the direction of the observer is modulated .
an example of a luminosity oscillation is the beat - frequency oscillation in the magnetospheric beat - frequency model proposed for the so - called horizontal branch oscillation seen from z sources ( alpar & shaham 1985 ; lamb et al .
1985 ; shibazaki & lamb 1987 ) .
an example of a beaming oscillation is the modulation created by an accretion - powered pulsar , in which , as the neutron star rotates , the hot spots on the magnetic polar cap pass into and out of our view , creating an observed modulation . here
we focus on beaming oscillations . in some models of the kilohertz qpos ( e.g. , miller et al .
1998 ) , the qpo peak with the highest frequency in a given source is a beaming oscillation caused by the movement of bright arcs of impact around the star , at the orbital frequencies in a special range of radii in the accretion disk . only the fundamental for such an oscillation has been observed , and the amplitudes at any overtones are down by at least factors of several .
moreover , in no kilohertz qpo source has there been a strong brightness oscillation observed at the putative stellar spin frequency during the persistent emission between type 1 x - ray bursts ( kilohertz qpos have not been observed from the millisecond x - ray pulsar sax j1808 - 3658 ; see wijnands & van der klis 1998 ) . for some of the higher - luminosity sources , such as sco x-1 , the upper limit on the pulsed fraction is as low as 0.3% ( vaughan et al .
this constrains beat frequency models , in which the lower - frequency qpo peak is created in part by a modulation at the spin frequency .
it is therefore necessary to explain how the spin modulation can be important close to the star but undetectable by current instruments at infinity . to understand this
, we need to consider various ways in which beaming oscillations can be attenuated on their way to us .
two general effects are ( 1 ) integration over the visible surface of the star , which tends to smooth out variations in intensity , and ( 2 ) scattering in a surrounding hot central corona , whose existence has been inferred from previously existing spectral and temporal modeling of neutron - star lmxbs ( lamb 1989 ; miller & lamb 1992 ; psaltis , lamb , & miller 1995 ; psaltis & lamb 1997 ) .
previous treatments of the attenuation of beaming oscillations by scattering ( brainerd & lamb 1987 ; kylafis & phinney 1989 ) assumed for simplicity that the source of radiation was a pencil beam at the center of a scattering cloud with uniform electron density ( although see wang & schlickeiser 1987 for a discussion of offset emission and varying electron density in the diffusion approximation ) , and have implicitly assumed that photons follow straight - line trajectories ( although for a treatment of general relativistic light - bending without scattering , see wood , ftaclas , & kearney 1988 ; mszros , riffert , & berthiaume 1988 ) . with the large amounts of data available from the _ rossi _ x - ray timing explorer ( rxte ) , it is now important to examine some of the deviations expected from this idealization .
here we calculate the attenuation of beaming oscillations near neutron stars , taking into account the finite size of the star , the varying electron number density near the star , and general relativistic light deflection . in
2 we explain our numerical method . in 3
we discuss the decrease in the modulation amplitude of beaming oscillations caused by integration over the visible surface of the star . in 4 we treat several aspects of the attenuation of beaming oscillations by scattering , including the presence of a finite object , varying electron number density , modulation from multiple harmonics of emission , the angular dependence of the specific intensity , and general relativistic light deflection . in 5 we calculate the modulation amplitude close to the star , within the scattering cloud . in 6 we summarize and give our conclusions .
the numerical results in this paper are generated using a monte carlo code .
we assume that a star of radius @xmath0 is in the center of a scattering cloud .
we also define optical depths @xmath1 and @xmath2 such that if the locally measured electron number density were uniform then , if the star were not present , the optical depth from the center of the sphere to the radius of the star would be @xmath1 , and the optical depth from the center to the outer boundary of the scattering sphere would be @xmath2 .
photons are introduced into the sphere at some radius and angle , traveling in some given direction , and they are followed individually as they scatter in the sphere . when the photons escape , their direction is stored .
after all the photons ( @xmath3 to @xmath4 , depending on the run ) have been tracked , rms amplitudes or beaming ratios can be computed .
we now discuss how photon paths between scatterings and after escape are followed for straight - line propagation and light deflection .
the number @xmath5 of mean free paths traversed between two scatterings is selected randomly according to @xmath6 .
if the scattering cloud has a uniform density then the physical distance traveled is proportional to @xmath5 , otherwise the distance must be calculated from the density distribution .
once the distance between two given scatterings is ascertained , the code checks whether ( 1 ) the photon hits the star , or ( 2 ) the photon escapes .
if the photon escapes , the current direction of propagation of the photon is stored .
if the photon hits the star , then in the next iteration the photon is assumed to be emitted in an outward direction from the point of impact , selected from an isotropic distribution .
otherwise , the new location is calculated by adding the vector of the photon path to the previous position . when light deflection is included , the propagation of photons is slightly more complicated than it is if photons propagate in straight lines .
although light deflection occurs in newtonian gravity as well as in general relativity , light deflection is only important when gravity is strong enough to make general relativistic corrections to newtonian gravity significant .
hence , we consider an external spacetime that is the schwarzschild spacetime , which is appropriate outside spherically symmetric , nonrotating neutron stars .
this spacetime includes most of the important features of , e.g. , light deflection around neutron stars , and the results are easier to evaluate than the results of scattering around rotating neutron stars . in the schwarzschild spacetime
, any geodesic can be treated as an equatorial geodesic by an appropriate rotation of the coordinate system ( a statement that is not true for general spacetimes with rotation ) .
this feature means that a simple way to treat photon propagation between scatters is to follow the path in that temporary equatorial plane " and then rotate back into the global coordinate system .
hence , there are three new tasks brought up by the inclusion of light deflection : ( 1 ) calculation of temporary equatorial planes , ( 2 ) computation of the curved trajectory of the photon in that plane , and ( 3 ) determination of the additional curvature of the photon trajectory after it escapes from the scattering cloud .
we now treat these in order .
let the current angular location of the photon have a colatitude @xmath7 and an azimuthal angle @xmath8 in the global coordinate system .
we represent the unit vector in this direction by @xmath9 , which in cartesian coordinates is as usual the three - vector @xmath10 , where the first , second , and third components are along the global @xmath5 , @xmath11 , and @xmath12 directions , respectively .
let the initial direction of propagation of the photon , also in the global coordinate system , be @xmath13 .
we set up the new equatorial plane as follows .
the @xmath14-axis is in the direction @xmath9 , and since this is properly normalized we assign @xmath15 .
the @xmath16 axis is perpendicular to the plane containing @xmath9 and @xmath13 , and is thus in the direction @xmath17 .
the remaining axis is in the direction @xmath18
. if the angle between @xmath9 and @xmath13 is @xmath19 , then the unit vectors are @xmath20\\ { \hat z}={1\over{\sin\psi}}(\theta,\phi)\times(\alpha,\beta)\ ; , \\
\end{array}\ ] ] and the equatorial plane is defined by @xmath14 and @xmath21 .
thus , if we use @xmath22 to denote the azimuthal angle in this new equatorial plane [ where @xmath23 at the original angular location @xmath9 ] , then the location in the original , global cartesian coordinates at an arbitrary radius @xmath25 and angle @xmath22 is just @xmath26 .
the next task is to follow the propagation of a photon in this new equatorial plane between scatterings . to do this
, we note that the locally measured spacelike components of the photon four - velocity are just @xmath27 here we use geometrized units in which @xmath28 . in this and
subsequent equations , hatted quantities such as @xmath29 are measured in a local tetrad , in contrast to unhatted quantities such as @xmath30 , which are measured in the global boyer - lindquist coordinate system .
the components of the four - velocity in global boyer - lindquist coordinates are given by the transformation of these quantities from the local to global frames ( see , e.g. , abramowicz , ellis , & lanza 1993 ; miller & lamb 1996 ) : @xmath31 the photon path is followed by moving a small distance @xmath32 along the ray ( we found @xmath33 in units of the mean free path gives sufficient accuracy ) and recalculating the propagation angle @xmath19 : @xmath34 ( see , e.g. , abramowicz et al . 1993 or
miller & lamb 1996 ) . using this formula
, we can determine the total deflection @xmath22 and rotate back from the temporary equatorial plane to the global coordinate system .
the last task is to follow the deflection of the photon after it has escaped .
this is done straightforwardly using the approach of , e.g. , pechenick , ftaclas , & cohen ( 1983 ) . in this approach
, we define the impact parameter @xmath35 , and let @xmath36 . the total deflection angle from radius @xmath25
to infinity is ( pechenick et al .
1983 , eq .
[ 2.12 ] ) @xmath37^{-1/2}du\ ; .\ ] ] note that , as defined , this is actually the difference between the global azimuthal angle at infinity @xmath38 and the global azimuthal angle @xmath39 at @xmath25 .
that is , even without general relativistic light deflection @xmath40 can be nonzero .
for example , if @xmath41 ( so that the photon is emitted tangentially to the radial vector ) , then if @xmath42 then @xmath43 . after computing @xmath40 , the angular location in the global coordinate system can be calculated as before , by rotating from the temporary equatorial plane to the global system .
the initial location and direction of propagation of the photons depend on the run .
the default condition is that the photons start on the surface of the star and are beamed directly outward ( we will refer to this as a pencil beam " specific intensity ) .
we will , however , consider other specific intensities , such as one that is isotropic outwards or one that has the slightly beamed pattern appropriate for radiation that was generated deep in the star and that propagated outward via isotropic scattering ( see chandrasekhar 1960 , chapter 3 ) .
the individual scatters are assumed to be locally isotropic .
note that we make a distinction between the angular width of the specific intensity and the angular pattern of emission on the star .
the former is what an observer standing on the star would measure from a particular emitting point , whereas the latter is the variation in total intensity ( integrated over local angles ) as a function of position on the star .
before treating the effects of scattering , we first note that a beaming pattern of high amplitude at the surface of the star may appear to an observer at infinity to have a low or zero amplitude .
this could happen because the angular width of the specific intensity from radiating points on the stellar surface is nonzero , and thus an observer at infinity sees light from everywhere on the visible portion of the star .
if the specific intensity has a wide beaming pattern ( e.g. , if the pattern is isotropic ) , then the intensity seen by the observer integrates over much of the star , and radiation patterns with large numbers of lobes are smeared out at infinity , leading to low amplitudes of intensity modulation as the star rotates . in this section
we derive expressions for the modulation amplitude seen at infinity , with no scattering , under various assumptions about the specific intensity and the angular pattern of emission on the star .
we first treat the case of straight - line photon propagation , and show that for an isotropic specific intensity and an odd number @xmath44 of lobes in the stellar emission pattern the intensity seen at infinity is constant .
thus , the relative modulation amplitude measured at infinity can be much less than the relative amplitude measured at the source even without light deflection .
we then consider general relativistic light deflection .
we confirm that , as demonstrated before ( e.g. , pechenick et al .
1983 ) , light deflection has a tendency to decrease the modulation amplitude .
however , we also show that for some harmonics of the stellar spin frequency , the modulation seen at infinity can actually have a _ higher _ amplitude when light deflection is included , compared to the modulation that would be observed if the photon trajectories were straight . in our treatments of both the straight and curved photon trajectories
, we assume that the emission pattern on the star is a thin equatorial belt , with a half - thickness @xmath45 much less than @xmath0 .
this is intended to model the intensity distribution expected in beat - frequency models of kilohertz qpos .
assume that the specific intensity is isotropic outwards , and that the emission intensity at an azimuthal angle @xmath8 is @xmath46 here @xmath47 gives the @xmath47th harmonic ; @xmath48 is the fundamental , @xmath49 is the first overtone , and so on .
the flux observed at infinity from a short segment of the equatorial belt is proportional to the product of the projected area of the segment ( which is proportional to the cosine of the angle @xmath50 between the line of sight and the surface normal ) and the emission intensity of the segment .
let the observer be at an angle @xmath51 , and assume that the segment of interest is at the angle @xmath52 .
then @xmath53 , and thus emission is observed between @xmath54 and @xmath55 .
if the star rotates with angular frequency @xmath56 , then at a time @xmath57 the star is at a rotational phase @xmath58 and the observed intensity is @xmath59 \cos\phi\,d\phi\\ = \sin\theta\left[2i_0+i_1{\pi\over 2}\cos\omega t+ \sum_{m=1}^\infty i_{2m}{2\over{4m^2 - 1}}\cos 2m\omega t\right]\ ; . \end{array}\ ] ] hence , the modulation amplitude vanishes for odd - numbered harmonics other than the fundamental , and decreases as @xmath60 for the even harmonics .
if the specific intensity is proportional to @xmath61 , then the variable part of the observed flux vanishes if @xmath62 is odd and @xmath63 , and is finite otherwise .
thus , contingent on the angle dependence of the specific intensity , either odd or even harmonics can integrate to zero .
when light deflection is included , the observer can see more of the star
. a larger fraction of the emission is thus observed , and hence one might expect that the general tendency will be for light deflection to reduce the observed amplitude modulation . indeed , this is the case at , e.g. , the fundamental of the rotation frequency when the specific intensity is isotropic .
however , since the constancy of the flux when @xmath63 and @xmath62 is odd comes from the vanishing of the integral when the limits are exactly @xmath64 and @xmath65 , the change of these limits by light deflection means that the amplitudes at these harmonics are actually _ larger _ when light deflection occurs than they would be in the limit of straight - line photon propagation .
we show this effect in figure 1 , in which we plot the rms amplitude vs. @xmath66 for different numbers @xmath47 of lobes in the emission pattern ( i.e. , different harmonic numbers @xmath47 ) .
shortly after low - frequency qpos were first discovered in neutron - star lmxbs ( van der klis et al .
1985 ; for a review see van der klis 1989 ) , a beat - frequency model was suggested for them in which the frequency of the qpo is equal to the difference between the stellar spin frequency and the frequency of a keplerian orbit near the main magnetospheric gas pickup radius ( alpar & shaham 1985 ; lamb et al .
1985 ; shibazaki & lamb 1987 ) .
in such a model , one would also expect to see a peak in the power density spectrum at the stellar spin frequency itself , but no such peak was observed . an early qualitative idea ( lamb et al .
1985 ; lamb 1986 ) , backed up by more quantitative calculations ( brainerd & lamb 1987 ; wang & schlickeiser 1987 ; kylafis & phinney 1989 ) , was that whereas the beat frequency oscillation is a luminosity oscillation and thus not easily attenuated , the brightness oscillation expected at the stellar spin frequency is a beaming oscillation and is thus relatively easily isotropized and attenuated by scattering through a hot central corona .
the existence of such a scattering corona near the neutron stars in lmxbs has been inferred by detailed comparisons of spectral models with observations ( see , e.g. , psaltis , lamb , & miller 1995 ) .
these model fits give radial optical depths to thomson scattering of @xmath67 and scattering corona radii of @xmath68 cm ( see miller et al .
1998 ) . with such radii and optical depths
, the beaming oscillation at the spin frequency is expected to be attenuated by a factor of several , which would lower the observed amplitudes to values consistent with current upper limits .
now that rxte has observed a number of these sources and discovered high - frequency qpos , we are once again faced with the quantitative question of why the spin frequency or other harmonics of the kilohertz keplerian frequency are not observed in the power spectra from these lmxbs . due to the high quality of rxte data and the more stringent upper limits it provides , we need to consider a number of the effects that previous treatments neglected for simplicity .
we do this in this section , where we consider ( 1 ) a beaming pattern that originates from the surface of a finite - sized neutron star instead of from the center of the scattering cloud , ( 2 ) a corona with a number density that changes with radius , ( 3 ) beaming patterns for different harmonics , ( 4 ) different specific intensity distributions , and ( 5 ) the effects of general relativistic light deflection .
our most important results are that the finite size of the neutron star decreases the expected attenuation of a beaming oscillation as seen at infinity even if the corona is much larger than the star , and that for many emission geometries _ all _ overtones ( not just the even harmonics ) are attenuated far more than is the fundamental . to follow the treatment of brainerd & lamb ( 1987 ) , consider a plane - parallel slab that goes from optical depth @xmath69 to @xmath2 , and assume that there is a source of intensity @xmath70 pointing forward ( towards @xmath71 ) at @xmath72 .
if we solve this problem in the two - stream approximation , where @xmath73 is the forward intensity at @xmath74 and @xmath75 is the backward intensity at @xmath74 , the equations are @xmath76 with the boundary conditions @xmath77 . solving these equations , we find that @xmath78 a measure of the asymmetry of the emission is the beaming ratio @xmath79 this confirms the result of wang & schlickeiser ( 1987 ) , who showed that in the diffusion approximation ( @xmath80 , @xmath81 ) the anisotropy of radiation depends approximately on the ratio @xmath82 .
brainerd & lamb ( 1987 ) effectively considered the special case @xmath83 .
note that the offset of the emission from the center of the scattering cloud multiplies the modulation amplitude seen at infinity by the constant factor @xmath84 , and does _ not _ asymptote to the @xmath83 amplitude even when @xmath85 .
the results of the numerical calculation with a finite object are shown in table 1 , where both the star and the scattering cloud are assumed to be spherical . in this table ,
@xmath3 photons were used for each calculation , and each one started out on the surface of the star at an optical depth @xmath1 from the center and was directed radially outward ( thus , this is a result for an initial pencil beam ) .
the numbers are the ratios of forward to backward intensity , defined as the ratio of the emergent intensity through the northern hemisphere ( @xmath86 ) to the emergent intensity through the southern hemisphere .
the numbers in parentheses are for no central star but a photon that starts at radius @xmath1 ; this comparison indicates the importance of a hard surface vs. the importance of an initial offset from the center .
this table shows that the presence of a central star increases the beaming ratio significantly , when compared to the beaming ratio produced without a finite central star , both by the offset of the initial emission from the center of the scattering cloud and by the presence of a hard surface . the analytical estimate for the beaming ratio matches the numerical results well when there is no star but the photon starts at an optical depth @xmath1 .
now consider what happens when the electron number density is not constant with radius .
we explore three different density profiles .
first , a density that goes as @xmath87 , where @xmath88 is the radius of the scattering region .
second , a density that is @xmath89 for @xmath90 and @xmath91 for @xmath92 .
third , a density that is @xmath91 for @xmath90 and @xmath89 for @xmath93 . in table 2
we compare the beaming ratios for these three different density profiles ( quad " , 2@xmath941 " , and 1@xmath942 " , respectively ) , for @xmath83 and the same total optical depth @xmath95 from the center to @xmath88 . in each case
we assume a pencil beam specific intensity originating from the center of the scattering cloud .
this table shows that , compared to a uniform - density scattering cloud , a cloud with an edge concentration yields a higher anisotropy as seen at infinity .
this is because the lower the number density is close in , the farther photons can travel before scattering . for a pencil - beam specific intensity
this is effectively similar to having the source of the radiation be offset from the center by a relatively large distance , implying that the anisotropy of the emergent radiation is larger than it would have been for uniform density .
conversely , a cloud with a central concentration yields a lower anisotropy at infinity than if the cloud had uniform density . using the same treatment as brainerd & lamb ( 1987 )
, we find that as @xmath96 the scattering green s function approaches @xmath97\ ; , \ ] ] where @xmath7 is the angular distance between the observer and the point of emission and there is no dependence on the longitude @xmath8 of the observer .
we wish to determine the intensity seen at infinity if the emission pattern on the surface has some number of lobes ( equivalently , for some harmonic ) .
consider first the lobe structure used by brainerd & lamb , which is that the intensity on the surface is proportional to @xmath98 .
this intensity pattern is appropriate for an emission that has a favored axis , such as a centered dipole .
if the center of the emission is at an angle @xmath99 and the observer is at an angle @xmath9 , then the cosine of the angle between them is just @xmath100 .
the intensity seen at @xmath9 is @xmath101 where @xmath102 is the intensity at the stellar surface at the colatitude @xmath103 .
the @xmath8 term will therefore always integrate to zero , regardless of the number of lobes .
the intensity is then @xmath104 ( 1+\cos n\theta^\prime)\sin\theta^\prime d\theta^\prime\\ & = \left\ { \begin{array}{ll } { 1\over{4\pi}}\left[1+\left(2\over{4-n^2}\right)\left(1+\tau_r\over { 1+\tau_c}\right)\cos\theta\right]i_0\ ; , & n{\rm \ odd } \\ { 1\over{4\pi}}i_0\ ; , & n{\rm \ even } \end{array } \right .
\end{array}\ ] ] in the limit @xmath80 .
here we have assumed @xmath105 .
note that a fraction @xmath106 of the photons will escape directly , and hence for finite @xmath2 the amplitude at even harmonics is nonzero but small .
this extends the result of brainerd & lamb ( 1987 ) to photons offset from the center of the scattering cloud : for emission symmetric around an axis , even - lobed patterns are attenuated much more rapidly than are odd - lobed patterns , and odd - lobed patterns with @xmath44 are more rapidly attenuated than is the @xmath48 pattern .
consider now a pattern that is symmetric about the equatorial plane but is not axisymmetric .
this is the pattern of interest for any rotationally or orbitally modulated emission , such as the emission believed to produce the higher - frequency qpo peaks observed during persistent emission from neutron - star lmxbs , which is symmetric about the rotational equator but is otherwise arbitrary .
specifically , consider an emission pattern @xmath107 $ ] , where @xmath108 . in this case , the intensity seen at infinity in the direction @xmath9 is @xmath109 where @xmath19 is the angle between @xmath9 and @xmath110 : @xmath111 the @xmath112dependent terms in this integral are proportional to either @xmath113 or @xmath114 .
however , note that for @xmath115 @xmath116_0^{2\pi}=0\ ; , \ { \rm and } \\
\int_0^{2\pi}\cos\phi^\prime\cos n\phi^\prime d\phi^\prime= \left[{\sin(n-1)\phi^\prime\over{2(n-1)}}+{\sin(n+1)\phi^\prime \over{2(n+1)}}\right]_0^{2\pi}=0\ ; . \end{array}\ ] ] note also that the @xmath117 term integrates to zero , due to the symmetry of @xmath118 .
therefore , in the diffusion limit ( @xmath80 ) , @xmath119 for @xmath44 .
this result means that for _ any _ emission pattern that is symmetric about the rotational equator , _ all _ overtones are attenuated extremely rapidly , not just the even harmonics .
in particular , note that it is not necessary to have the same @xmath22 dependence at all latitudes ; an arbitrary emission pattern symmetric about the equator can be built up using pairs of rings of the form @xmath120)h(\phi^\prime)$ ] , where @xmath121 is some function of @xmath22 ( @xmath122 can , therefore , be fourier decomposed into terms proportional to @xmath123 ) .
this is a strong reason why , even if the fundamental is strong at the sonic - point keplerian frequency , we do not expect to see significant peaks in the power spectrum at overtones of @xmath124 .
figure 2 shows the fractional rms amplitude of different harmonics as a function of the optical depth of the scattering cloud .
here we assume a pencil - beam specific intensity and straight - line photon propagation , and we assume that the central object has a radius of @xmath125 optical depths .
therefore , @xmath126 would mean that there is no scattering cloud ; @xmath127 would mean that the radial optical depth from the stellar surface to the edge of the cloud is @xmath128 , and so on .
as in figure 1 , @xmath47 is the number of lobes in the radiation emission pattern from the surface ; thus @xmath48 is the fundamental of the spin frequency , @xmath49 is the first overtone , and so on . for harmonic number
@xmath47 we assume an emission intensity proportional to @xmath129 ; for an intensity actually proportional to @xmath130 , with @xmath131 , the magnitude of the amplitude must be multiplied by @xmath132 .
this figure shows that all overtones of the fundamental oscillation frequency are attenuated rapidly by scattering .
+ heretofore we have considered only a pencil beam type of specific intensity . in reality , the angular dependence of the specific intensity is likely to differ from a pencil beam . in figure
3 we show the rms amplitude vs. optical depth for a pencil beam , for the beamed pattern appropriate for radiation generated at great depth ( see chandrasekhar 1960 , chapter 3 ) , and for an isotropic beam .
these different specific intensity distributions have different uses .
a pencil beam may be considered to set an upper limit on the anisotropy , but it is unlikely to be of direct significance in the physical situations considered here . the beamed pattern appropriate for radiation generated at great depth that propagates to the surface via isotropic scattering is likely to represent well the emergent radiation pattern from an x - ray burst . an isotropic specific intensity may be a good model for the radiation produced by accretion , since the energy is likely to be released in a shallow layer . from this figure
, we see that , as expected , the more beamed the pattern the greater the rms amplitude of variation for a given surface intensity distribution .
we also see that the chandrasekhar - type specific intensity gives a modulation amplitude closer to an isotropic beam than to a pencil beam .
+ as we described in 2 , the code has the capability to follow curved photon trajectories .
we show the effects of light deflection in a scattering cloud in figure 4 and figure 5 .
figure 4 plots the rms amplitude vs. @xmath66 for different numbers of nodes for a pencil beam with @xmath125 , @xmath133 .
figure 5 compares the amplitude vs. @xmath66 for @xmath125 , @xmath133 for a pencil beam ( solid line ) , an isotropic beam ( dotted line ) , and a pencil beam which undergoes no light deflection after it escapes from the scattering cloud ( dashed line ) .
the curves in figure 5 , especially those for the pencil beam with and without deflection after escape , demonstrate that there are actually two effects on the amplitude that compete with each other .
the deflection of light after escape spreads out the beam , and hence decreases the amplitude .
the deflection of light between scatters , however , can have the opposite effect .
for a given distance traveled after one scatter , the subsequent scatter is angularly closer to the first because the photon travels in a curved trajectory .
this tends to increase the number of scatters , because the coordinate distance traveled is less between scatterings .
however , for a fixed number of scatters the angular distance traveled is smaller , and hence the effective isotropization is diminished .
the overall effect is always that the modulation amplitude decreases with increasing compactness @xmath66 , but figure 5 shows that for @xmath134 the amplitude would increase with increasing @xmath66 if there were no deflection after escape .
another important question is how great a modulation amplitude is to be expected _ inside _ a scattering cloud . for example , in the sonic - point model ( miller et al .
1998 ) the lower - frequency qpo in a pair is generated by the interaction of radiation at the surface with clumps of matter in the disk .
the radiation is modulated at the stellar spin frequency , and thus the mass accretion rate from the clump is modulated at the difference between the sonic - point keplerian frequency and the stellar spin frequency . from no source with kilohertz qpos has there been a strong peak at the stellar spin frequency detected in the power density spectrum , and hence it is important to determine if it is possible that the modulation at the spin frequency is strong near the star , where the beat frequency is generated , yet weak enough at infinity that no peak in the power density spectrum is evident .
it is not easy to compute the modulation in mass accretion rate that results from a given modulation amplitude in the radiation force or any of the components of the radiation stress - energy tensor , because the effects of radiation drag can be nonlinear .
for example , since extra radiation drag increases the radial velocity of gas and thus decreases the optical depth from the stellar surface ( radiation - induced transparency " ; see miller & lamb 1996 ; miller et al .
1998 ) , a small increase in radiation drag can in principle lead to a large increase in the accretion rate .
nonetheless , as a proof of principle we can calculate the radiation energy density , to determine if it is modulated significantly near the star .
figure 6 shows the result of this calculation .
here we have chosen @xmath125 and @xmath133 , for illustrative purposes , and we assume a central star of radius @xmath135 . the solid line shows the fractional rms amplitude of the modulation in the radiation energy density , and the horizontal dotted line shows the fractional rms amplitude in the energy density as observed at infinity .
it is clear from this figure that the modulation amplitude can be much higher near the star than at infinity .
for example , the rms amplitude is more than ten times as high at the surface of the star as it is at infinity .
this confirms that a brightness oscillation can have strong effects near the star , but be relatively weak at infinity .
the amplitude of beaming oscillations as seen at infinity depends on a number of variables , including the radiation pattern on the surface of the star , the angular width of the specific intensity , the size of the star , and the size , optical depth , and density profile of the scattering cloud .
some of the trends evident from this paper are ( 1 ) when there is a scattering cloud , the observed amplitudes of all overtones are much less than the observed amplitude of the fundamental , ( 2 ) a brightness oscillation that is weak or undetectable with current instruments at infinity can nonetheless have a significant amplitude near the star , and ( 3 ) the presence of a finite - sized star in the center of a scattering cloud decreases the attenuation of beaming oscillations compared to the attenuation expected when the photons are emitted at the center of the cloud .
these results are particularly useful in the study of the kilohertz qpos from neutron - star lmxbs .
spectral models ( see lamb 1989 ; miller & lamb 1992 ; psaltis et al .
1995 ) suggest that many of the neutron stars in lmxbs are surrounded by hot central coronae with radii of @xmath136 cm and optical depths @xmath137 .
the observed amplitudes of kilohertz qpos , combined with a detailed model of them , can be used to place further restrictions on the radius and optical depth of the central corona , and possibly even on the compactness of the neutron star . in conclusion ,
the results presented here on attenuation of beaming oscillations explain several features of the kilohertz qpos , including why no overtones of a beaming oscillation have been detected and why the beaming oscillation at the stellar spin frequency can be strong enough near the star to generate a beat - frequency qpo ( see miller et al .
1998 ) , yet too weak to detect at infinity .
the attenuation factors at frequencies such as the keplerian qpo frequency , the spin frequency , and the overtones and sidebands of these fundamental frequencies depend on the radius and optical depth of the scattering cloud and on the redshift at the surface of the neutron star , and hence detections of qpos at these frequencies , or strong upper limits on their amplitudes , provide a valuable source of information about the conditions near neutron stars .
it is a pleasure to thank fred lamb and dimitrios psaltis for discussions about the optical depths and radii of coronae in lmxbs , and don lamb and carlo graziani for their comments on early versions of the ideas contained in this paper .
this work was supported in part by nasa grant nag 5 - 2868 , by nasa atp grant number nra-98 - 03-atp-028 , and through the gro fellowship program by nasa grant nas 5 - 2687 .
ccccccc + @xmath2&@xmath83&@xmath125&@xmath139&@xmath140&@xmath141&@xmath142 + 1&0.465&&&&& + 2&0.301&0.697(0.605)&&&& + 3&0.246&0.528(0.459)&0.810(0.712)&&& + 4&0.199&0.448(0.364)&0.678(0.566)&0.877(0.774)&& + 5&0.174&0.380(0.311)&0.601(0.486)&0.769(0.643)&0.906(0.814)& + 10&0.094&0.218(0.183)&0.337(0.273)&0.453(0.366)&0.577(0.461)&0.699(0.556 ) + 20&0.056&0.121(0.085)&0.184(0.145)&0.254(0.205)&0.332(0.234)&0.391(0.299 ) + ccccc + + @xmath95&constant&quad&2@xmath941&1@xmath942 + 1&0.465&0.453&0.462&0.487 + 2&0.301&0.273&0.268&0.334 + 3&0.246&0.206&0.206&0.286 + 4&0.199&0.173&0.152&0.224 + 5&0.174&0.127&0.134&0.195 + 10&0.094&0.082&0.070&0.109 + 20&0.056&0.041&0.033&0.055 + | observations with the _ rossi _ x - ray timing explorer have revealed kilohertz quasi - periodic brightness oscillations ( qpos ) from nearly twenty different neutron - star low - mass x - ray binaries ( lmxbs ) .
these frequencies often appear as a pair of kilohertz qpos in a given power density spectrum .
it is extremely likely that the frequency of the higher - frequency of these qpos is the orbital frequency of gas at some radius near the neutron star .
it is also likely that the qpos are caused by the movement of bright arcs or luminous clumps around the star , which produce a modulation in the observed x - ray intensity as they are periodically occulted by the star or as they present a different viewing aspect to the observer at infinity .
if this picture is correct , it means that this type of qpo is a beaming oscillation . in such models
it is expected that there will also be beaming oscillations at the stellar spin frequency and at overtones of the orbital frequency , but no strong qpos have been detected at these frequencies .
we therefore examine the processes that can attenuate beaming oscillations near neutron stars , and in doing so extend the work on this subject that was initiated by the discovery of lower - frequency qpos from lmxbs .
we consider attenuation by scattering , attenuation by light deflection , and the decrease in modulation caused by integration over the visible surface of the neutron star .
our main results are ( 1 ) in a spherical scattering cloud , _ all _ overtones of rotationally modulated beaming oscillations are attenuated strongly , not just the even harmonics , ( 2 ) the amount of attenuation is diminished , and hence the observed modulation amplitude is increased , by the presence of a central , finite - sized star , even if the scattering cloud is much larger than the star , and ( 3 ) if the specific intensity of radiating points on the star has a large angular width , then even with zero optical depth from the stellar surface to the observer , and even in the approximation of straight - line photon propagation , the modulation amplitude seen at infinity is decreased significantly by integration over the visible portion of the surface .
we also compare the modulation of flux as seen at infinity with the modulation near the star , and show that ( 4 ) it is possible to have a relatively high - amplitude modulation near the star at , e.g. , the stellar spin frequency , even if no peak at that frequency is detectable in a power density spectrum taken at infinity . |
if the standard model ( sm ) of particle interactions is extended to accommodate @xmath8 , then the conventional assignment of @xmath9 implies the well - known result that @xmath10 and @xmath11 .
there must then be higgs bidoublets @xmath12 both transforming as @xmath13 , yielding lepton dirac mass terms @xmath14 and similarly in the quark sector .
this results in the appearance of phenomenologically undesirable tree - level flavor - changing neutral currents from higgs exchange , as well as inevitable @xmath15 mixing .
if supersymmetry is imposed , then @xmath16 can be eliminated , but then @xmath17 as well as @xmath18 , contrary to what is observed .
hence the prevalent thinking is that @xmath19 is actually broken down to @xmath20 at a very high scale from an @xmath1 higgs triplet @xmath21 which provides @xmath22 at the same time with a large majorana mass from @xmath23 .
the canonical seesaw mechanism for neutrino mass is thus implemented and everyone should be happy .
but wait , no remnant of the @xmath1 gauge symmetry is detectable at the tev scale and we will not know if @xmath22 really exists .
is there a natural way to lower the @xmath24 breaking scale ?
the answer was already provided 23 years ago @xcite in the context of the superstring - inspired supersymmetric @xmath25 model .
the fundamental fermion representation here is decomposed under @xmath26 $ ] as @xmath27 under its maximum subgroup @xmath28 , the is organized instead as @xmath29 , i.e. @xmath30 it was realized @xcite in 1987 that there are actually two left - right options : ( a ) let @xmath25 break down to the fermion content of the conventional @xmath31 , given by @xmath32 , which is the usual left - right model which everybody knows .
( b ) let @xmath25 break down to the fermion content given by @xmath33 instead , thereby switching the first and third rows of @xmath34 and the first and third columns of @xmath35 .
thus @xmath36 becomes @xmath0 and @xmath37 is not the mass partner of @xmath38 .
this is referred to by the particle data group as the alternative left - right model ( alrm ) . here
the usual left - handed lepton doublet is part of a bidoublet : @xmath39 in this supersymmetric model , @xmath38 is still the dirac mass partner of @xmath22 and gets a seesaw mass , whereas @xcite @xmath37 ( which couples to @xmath40 through @xmath41 ) mixes with the usual neutralinos , the lightest of which is a dark - matter candidate .
two simpler nonsupersymmetric versions of the alrm with @xmath37 as dark matter have recently been proposed @xcite .
we call them dark left - right models ( dlrm and dlrm2 ) .
we impose a global u(1 ) symmetry @xmath42 , so that under @xmath43 , where @xmath44 , a generalized lepton number is conserved , such that @xmath45 in dlrm and @xmath46 in dlrm2 .
the resulting dark - matter fermion @xmath37 has @xmath47 ( majorana ) in dlrm and @xmath48 ( dirac ) in dlrm2 .
this talk is on dlrm2 , with particle content @xcite under @xmath43 given below : @xmath49 as a result , the yukawa terms @xmath50 , @xmath51 , @xmath52 , @xmath53 , @xmath54 , @xmath55 are allowed , whereas @xmath56 , @xmath57 are forbidden together with the bilinear terms @xmath58 , @xmath59 .
the breaking of @xmath60 leaves @xmath46 unbroken , so that @xmath61 [ @xmath62 has @xmath47 ] , but @xmath64 [ @xmath65 has @xmath66 .
the former contributes to @xmath67 and @xmath68 , whereas the latter means that @xmath38 and @xmath37 are not dirac mass partners and can be completely different particles .
in fact , @xmath69 , @xmath70 come from @xmath71 , and @xmath72 , @xmath73 from @xmath74 .
this structure guarantees the absence of tree - level flavor - changing neutral currents . as
for the gauge bosons and their interactions , let @xmath75 , @xmath76 , @xmath77 , then @xmath78 to avoid @xmath79 mixing at tree level , the condition @xmath80 must be imposed . in that case ,
note that @xmath41 does not mix with @xmath82 because they have different @xmath3 parity . in fig . 1
, the present tevatron bound on @xmath83 is shown for various values of @xmath84 , showing a typical bound of about 1 tev .
vs @xmath83 compared against tevatron data.,scaledwidth=80.0% ]
the usual leptons have @xmath85 as expected , but there are now new particles also with lepton number : @xmath86 have @xmath85 and @xmath87 has @xmath88 as well as @xmath89 .
thus they all have odd @xmath3 parity , i.e. @xmath90 , even though the model is nonsupersymmetric .
the scotino @xmath2 has @xmath48 and thus also odd @xmath3 .
the lightest @xmath2 is a dark - matter candidate , and will be considered below in the context of recent data from the cdms - ii collaboration @xcite .
two possible dark - matter signal events were observed with an expected background of @xmath91 .
the most stringent bound on the elastic spin - independent scattering cross section of @xmath92 occurs at @xmath93 gev , and it is @xmath94 pb . in the dlrm2 , @xmath95 where @xmath96 , @xmath97 , @xmath98 .
let @xmath99 , @xmath100 , then @xmath101 ^ 2 \over a^2}.\ ] ] using @xmath102ge , i.e. @xmath103 and @xmath104 , as an estimate , the cdms bound ( of @xmath94 pb at @xmath93 gev ) implies a bound on @xmath83 . in fig . 2 ,
the resulting lower bounds on @xmath83 from the tevatron search and from cdms are plotted as functions of @xmath84 .
vs @xmath84 from the tevatron search ( red solid line ) and from the cdms search at @xmath93 gev ( blue dashed line ) .
the dotted segments assume a simple extrapolation of the tevatron data.,scaledwidth=80.0% ] to obtain the correct dark - matter relic abundance , the annihilation of @xmath105 sm fermions is considered .
the thermally averaged cross section multiplied by the relative velocity of the annihilating particles is given by @xmath106 where @xmath107 .
fixing the above at 1 pb , the values of @xmath72 and @xmath83 are constrained as a function of @xmath84 . for @xmath93
gev , there is no solution , but if @xmath72 is greater than about 300 gev , solutions exist which are consistent with the tevatron bound as well as the cdms bound . in the range @xmath108 tev , the latter is well approximated by @xmath109 pb ( @xmath72/1 tev)@xmath110 . in fig . 3 , the cases @xmath111 and 600 gev are shown .
( blue dashed line ) and its value ( red solid line ) from @xmath112 pb vs @xmath84 for @xmath113 and 600 gev.,title="fig:",scaledwidth=45.0% ] ( blue dashed line ) and its value ( red solid line ) from @xmath112 pb vs @xmath84 for @xmath113 and 600 gev.,title="fig:",scaledwidth=45.0% ] the @xmath114 annihilation to leptons through @xmath5 exchange also contributes , i.e. @xmath115 but it is subdominant and has been neglected .
at the lhc ( @xmath116 gev ) , @xmath4 may be discovered with 10 dilepton events . using the cuts * @xmath117 gev for each lepton , *
@xmath118 for each lepton , * @xmath119 , the sm background is negligible . with an integrated luminosity of 1 fb@xmath120 ,
the discovery reach of the @xmath4 of the dlrm2 is about 2 tev , as shown in fig .
discovery by 10 dielectron events at lhc .
small circles are tevatron limits.,scaledwidth=80.0% ] to distinguish this @xmath4 from others , the following ratios @xcite may be considered : @xmath121 where @xmath122 . in the conventional left - right model , the numerator for @xmath123
is changed to @xmath124 , i.e. 13.6 larger ( @xmath125 ) . in the alrm ,
the denominator for both is changed to @xmath126 , i.e. 2.6 larger ( @xmath125 ) .
there are also important loop effects @xcite on rare processes from the new interactions @xmath127 , @xmath128 , @xmath129 , @xmath130 , etc .
the anomalous magnetic moment of the muon receives a contribution of order @xmath131 , below the present experimental sensitivity of @xmath132 . the flavor - changing radiative decay @xmath133 has the branching fraction @xmath134 where @xmath135 and @xmath136 it is suppressed by @xmath137 and for @xmath138 tev ( corresponding to @xmath139 tev )
, the current experimental bound of @xmath140 implies @xmath141 for @xmath125 .
a more sensitive probe of the existence of these new interactions is @xmath142 or @xmath7 conversion in a nucleus or muonic atom @xcite .
the reason is that there is an effective @xmath143 vertex from @xmath128 and @xmath129 , given by @xmath144 where @xmath145 , which is not suppressed if @xmath146 is not small , which holds even if the @xmath1 scale is much greater than the electroweak scale .
this unusual ( nondecoupling ) property depends crucially on the @xmath128 vertex , which is not available in other extensions of the sm , including all @xmath147 models .
the current experimental bound of @xmath148 on @xmath149 implies @xmath150 for @xmath125 .
the presence of @xmath22 is unavoidable in a left - right gauge extension of the standard model .
however , it does not have to be the dirac mass partner of @xmath38 . in that case
, it should be renamed @xmath37 and could function as a scotino , i.e. a dark - matter fermion . in the context of the recently proposed new dark left - right model ( dlrm2 ) , latest cdms observations are shown to be consistent with the lightest @xmath2 at about a few hundred gev in mass with the new @xmath4 gauge boson at less than 2 tev .
the latter should then be accessible directly at the lhc , while the @xmath5 gauge bosn may contribute indirectly to enhancing rare lepton - flavor - changing processes such as @xmath142 and @xmath7 conversion in a nucleus or muonic atom .
this work was supported in part by the u. s. department of energy under grant no .
de - fg03 - 94er40837 .
i thank d. delepine and the other organizers of the vi international workshop on the dark side of the universe for their great hospitality and a stimulating meeting in leon .
9 e. ma , phys . rev . *
d36 * , 274 ( 1987 ) .
e. ma , phys . rev . *
d62 * , 093022 ( 2000 ) .
s. khalil , h .- s .
lee , and e. ma , phys . rev . * d79 * , 041701(r ) ( 2009 ) .
s. khalil , h .- s .
lee , and e. ma , phys . rev . *
d81 * , 051702(r ) ( 2010 ) .
z. ahmed _ et al .
_ , science * 327 * , 1619 ( 2010 ) . s. godfrey and t. a. w. martin , phys .
* 101 * , 151803 ( 2008 ) .
m. yamanaka , these proceedings . | the standard model of particle interactions is extended to include fermion doublets @xmath0 transforming under the gauge group @xmath1 such that @xmath2 is a dirac scotino ( dark - matter fermion ) , with odd @xmath3 parity .
based on recent cdms data , it is shown how this new dark left - right model ( dlrm2 ) favors a @xmath4 gauge boson at around 1 or 2 tev and be observable at the lhc .
the new @xmath5 gauge bosons may also contribute significantly to lepton - flavor - changing processes such as @xmath6 and @xmath7 conversion in a nucleus or muonic atom .
address = physics and astronomy department , university of california , riverside , california 92521 , usa |
during the past decade , there has been a growing interest in various electronic systems driven out of equilibrium by the injection of spin - polarized carriers .
such systems can be realized by current - biasing structures consisting of ferromagnetic and nonferromagnetic ( e.g. , superconducting ) layers , due to the difference in population of majority- and minority- spin subbands.@xcite the concept of spin - polarized current nowadays has attracted considerable interest in ferromagnetic heterostructures , in particular for applications in spintronics.@xcite charge transport through a normal metal / superconductor ( ns ) junction , with an insulating barrier of arbitrary strength at the interface , has been studied by blonder , tinkham , and klapwijk ( btk),@xcite and the andreev reflection is recognized as the mechanism of normal - to - supercurrent conversion.@xcite the btk theory has been extended by tanaka and kashiwaya to include the anisotropy of the pair potential in @xmath0-wave superconductors.@xcite the modification of the andreev reflection by the spin injection from a ferromagnetic metal into a superconductor in ferromagnet / superconductor ( fs ) junctions was first analyzed by de jong and beenakker.@xcite more recently , the effects of unconventional @xmath0-wave and @xmath1-wave pairing and of the exchange interaction in fs systems , such as the zero - bias conductance peak and the virtual andreev reflection , have been clarified by kashiwaya _
et al_.@xcite and yoshida _ et al_.@xcite the fermi velocity mismatch between two metals can also significantly affect the andreev reflection by altering the subgap conductance,@xcite which is similar to the presence of an insulating barrier.@xcite in experiments , a superconductor is used to determine the spin polarization of the current injected from ( or into ) a ferromagnet by measuring the differential conductance .
these measurements have been performed on tunnel junctions in an external magnetic field,@xcite metallic point contacts,@xcite nano - contacts formed by microlithography,@xcite and fs junctions with @xmath0-wave superconductors , grown by molecular beam epitaxy.@xcite in diffusive fs junctions , the excess resistance may be induced by spin accumulation near the insulating interface@xcite and by the proximity effect.@xcite when interfaces act incoherently , the btk model can be successfully applied to normal metal / superconductor / normal metal ( nsn ) or ferromagnet / superconductor / ferromagnet ( fsf ) double junctions.@xcite however , the properties of coherent quantum transport in clean superconducting heterostructures are strongly influenced by size effects , which are not included in the btk model .
well - known examples are the current - carrying andreev bound states@xcite and multiple andreev reflections@xcite in superconductor / normal metal / superconductor ( sns ) junctions . since early experiments by tomasch,@xcite the geometric resonance nature of the differential conductance oscillations in sns and
nsn tunnel junctions has been ascribed to the electron interference in the central film.@xcite recently , mcmillan - rowell oscillations were observed in sns edge junctions of @xmath0-wave superconductors and used for measurements of the superconducting gap and the fermi velocity.@xcite in this paper we study coherent electronic transport in fsf double - barrier junctions ( and nsn as a special case ) within the framework of bcs theory .
we limit ourselves to clean conventional ( isotropic and @xmath2-wave ) superconductors , and neglect , for simplicity , the self - consistency of the pair potential@xcite and nonequlibrium effects of charge and spin accumulation at the interfaces.@xcite when two interfaces are recognized by electrons simultaneously , characteristic features of finite size and coherency are the subgap transport of electrons and oscillations of both charge and spin differential conductances above the gap .
one consequence of the quasiparticle interference is the periodic vanishing of the andreev reflection at the energies of geometrical resonances .
the other is the existence of a nontrivial spin polarization of the current not only for the parallel ( p ) , but also for the antiparallel ( ap ) alignment of the electrode magnetizations .
previous analysis of incoherent transport in fsf double junctions in ap alignment predict the absence of spin current and suppression of superconductivity with increasing voltage , as a result of spin imbalance in the superconducting film.@xcite
we consider an fsf double junction consisting of a clean superconducting layer of thickness @xmath3 , connected to ferromagnetic electrodes by thin , insulating interfaces . for the ferromagnetic metal
we adopt the stoner model , describing the spin - polarization effect by the usual one - electron hamiltonian with an exchange potential .
the quasiparticle propagation is described by the bogoliubov de gennes equation @xmath4 with @xmath5 , where @xmath6 and @xmath7 are the hartree and the chemical potential , respectively .
the interface potential is modeled by @xmath8 , where the @xmath9 axis is perpendicular to the layers and @xmath10 is the dirac @xmath11 function . neglecting the self - consistency of the superconducting pair potential , @xmath12 is taken in the form @xmath13 , where @xmath14 is the heaviside step function and @xmath15 is the bulk superconducting gap . in eq .
( [ bdg ] ) , @xmath16 is the quasiparticle spin ( @xmath17 and @xmath18 ) , @xmath19 is the energy with respect to @xmath7 , @xmath20 is the exchange potential given by @xmath21\theta(z - l)\}$ ] for the p [ ap ] alignment , and @xmath22 is @xmath23 ( @xmath24 ) for spins up ( down ) .
the electron effective mass @xmath25 is assumed to be the same for the whole junction . here , @xmath26 is the fermi energy of the superconductor , @xmath27 , or the mean fermi energy of a ferromagnet , @xmath28 .
moduli of the fermi wave vectors , @xmath29 and @xmath30 , can be different in general , and in the following , the fermi wave vector mismatch ( fwvm ) will be taken into account through the parameter @xmath31 .
the parallel component of the wave vector @xmath32 is conserved , and the wave function @xmath33 satisfies appropriate boundary conditions .
four independent solutions of eq .
( [ bdg ] ) correspond to the four types of injection : an electron or a hole from either the left or from the right electrode.@xcite for the injection of an electron from the left , with energy @xmath34 , spin @xmath16 , and angle of incidence @xmath35 ( measured from the @xmath9 axis ) , solution for @xmath36 in various regions has the following form : in the left ferromagnet ( @xmath37 ) @xmath38 in the superconductor ( @xmath39 ) , @xmath40 and in the right ferromagnet ( @xmath41 ) , for the p [ ap ] alignment of the magnetizations , @xmath42}z)\left(\begin{array}{c } 1 \\ 0 \\ \end{array}\right)+ d_{\sigma}(e,\theta)\exp(-ik^-_{{\bar\sigma}[\sigma]}z)\left(\begin{array}{c } 0 \\ 1 \\
\end{array}\right ) .
\label{psir}\ ] ] here , @xmath43 and @xmath44 are the bcs coherence factors , and @xmath45 .
the @xmath9 components of the wave vectors are @xmath46 and @xmath47 where @xmath48 .
the coefficients @xmath49 , @xmath50 , @xmath51 , and @xmath52 are , respectively , the probability amplitudes of ( 1 ) andreev reflection as a hole of the opposite spin ( ar ) , ( 2 ) normal reflection as an electron ( nr ) , ( 3 ) transmission to the right electrode as an electron ( te ) , and ( 4 ) transmission to the right electrode as a hole of the opposite spin ( th )
. processes ( 1 ) and ( 4 ) are equivalent to the formation of a cooper pair in the superconductor by taking one more electron from either the left or the right electrode , respectively .
amplitudes of the bogoliubov electronlike and holelike quasiparticles , propagating in the superconducting layer , are given by the coefficients @xmath53 through @xmath54 .
neglecting small terms @xmath55 and @xmath56 in the wave vectors , except in the exponents @xmath57 solutions for the probability amplitudes can be written in the compact form for the general case ( see the appendix ) . in the following
we use the approximated wave - vector components in units of @xmath58 : @xmath59 , @xmath60 and @xmath61 , where @xmath62 , @xmath63 , and @xmath64 is measuring fwvm .
dimensionless parameter measuring the strength of each interface barrier is @xmath65 .
solutions for the other three types of injection can be obtained by the same procedure . in particular , if a hole with energy @xmath66 , spin @xmath16 , and angle of incidence @xmath35 is injected from the left , the substitution @xmath67 holds , and the scattering probabilities are the same as for the injection of an electron with @xmath19 , @xmath16 , and @xmath35 . therefore , in order to include the description of both electron and hole injections , the calculated probabilities should be regarded as even functions of @xmath19 .
also , for an electron or a hole , injected from the right , the probabilities are the same as for the injection from the left , except @xmath68 for the ap alignment . from the probability current conservation ,
the probabilities of outgoing particles satisfy the normalization condition @xmath69 where , @xmath70}}{\tilde{k}_{\sigma}}\right ) \re\left(\frac{\tilde{k}_{{\bar{\sigma}}[\sigma]}}{\tilde{k}_{\sigma}}\right ) it follows from the general solution that @xmath71 when @xmath72 for @xmath73 , independently of @xmath74 , @xmath75 , and @xmath76 . therefore ,
both direct and crossed andreev reflection vanishes at the energies of geometrical resonances in quasiparticle spectrum .
the absence of ar and th processes means that all quasiparticles with energies satisfying eq .
( [ resonance ] ) will pass unaffected from one electrode to another , without creation or annihilation of cooper pairs .
the effect is similar to the over - the - barrier resonances in the simple problem of one - particle scattering against a step - function potential,@xcite the superconducting gap playing the role of a finite - width barrier.@xcite characteristic features of coherent electronic transport through the superconducting layer are the subgap transmission ( without conversion to supercurrent ) and oscillations of the scattering probabilities . for @xmath77 ,
the subgap transmission of electrons or holes suppresses the andreev reflection . for @xmath78 ,
all probabilities oscillate with @xmath19 and @xmath3 due to the interference of incoming and outgoing particles .
these effects are illustrated in figs .
[ thin ] and [ thick ] for an fsf double junction , with @xmath79 and @xmath80 , in p alignment .
taking @xmath81 , in a thin superconducting film , @xmath82 , the andreev reflection is strongly suppressed , since the subgap transmission of electrons is considerable , fig .
[ thin ] . in this case , the oscillations are less pronounced , with the period much larger than @xmath15 . for a thick film , @xmath83 ,
the subgap tunneling is irrelevant [ except for small `` tails '' in @xmath84 and @xmath85 at @xmath86 , and the oscillations above the gap are more pronounced with the period on the order of @xmath15 , fig .
[ thick ] .
the scattering probabilities for ap and p alignment differ very slightly in the case of normal incidence , @xmath87 .
although spin independent , @xmath84 is even more suppressed due to the exchange interaction .
in contrast with nsn junction with transparent interfaces , @xmath88 and @xmath89 become nontrivial . the spin - dependent
@xmath88 has zeros at the same energies as @xmath84 and @xmath89 , so that maxima in @xmath85 at geometrical resonances are still equal to unity due to the interface transparency . the insulating barriers ( @xmath90 ) and fwvm ( @xmath64 ) reduce ar and te and enhance nr and
th probabilities , as well as the exchange interaction ( @xmath91 ) .
in contrast to the positions of zeros of @xmath84 , given by eq .
( [ resonance ] ) , the positions of maxima of @xmath84 , as well as that of zeros and maxima of @xmath88 , @xmath85 , and @xmath89 , are dependent on @xmath74 , @xmath75 , and @xmath76 . approaching the tunnel limit ( @xmath92 ) , peaks in the scattering probabilities gradually split into two spikes belonging to consecutive pairs with positions defined by the quantization conditions @xmath93 giving the bound - state energies of an isolated superconducting film .
( [ resonance ] ) and ( [ n ] ) are simply connected by @xmath94 .
the exception is the spike at the gap edge , originating from the singularity in the bcs density of states .
this gives the correspondence between the electronic transport through resonances in metallic junctions and through bound states in the tunnel junctions .
when voltage @xmath95 is applied to the junction symmetrically , the charge current density can be written in the form@xcite @xmath96\delta f({\bf k},v)\\ & = & \frac{e{k^{(s)}_{f}}^2}{\pi h}\int\limits_{-\infty}^{\infty}{\rm d}e \sum_{\sigma=\uparrow,\downarrow}\lambda^2_\sigma \int\limits_{0}^{\pi/2}{\rm d}\theta \sin\theta \cos\theta \left[a_\sigma(e,\theta)+c_\sigma(e,\theta)\right]\delta f({\bf k},v),\end{aligned}\ ] ] where @xmath97 is the asymmetric part of the nonequilibrium distribution function of current carriers . in the last equality the normalization condition , eq .
( [ abcd ] ) , was taken into account . without solving the suitable transport equation ,
we take @xmath98 , where @xmath99 is the fermi - dirac equilibrium distribution function.@xcite in this approach , the charge current per orbital transverse channel is given by @xmath100g_q(e),\ ] ] where the differential charge conductance at zero temperature is @xmath101.\ ] ] on the other hand , the probability current per orbital transverse channel is given by @xmath102\delta f({\bf k},v)\\ & = & \frac{e{k^{(s)}_{f}}^2}{\pi h}\int\limits_{-\infty}^{\infty}{\rm d}e \sum_{\sigma=\uparrow,\downarrow}\rho_\sigma\lambda^2_\sigma \int\limits_{0}^{\pi/2}{\rm d}\theta \sin\theta \cos\theta \left[1-a_\sigma(e,\theta)-b_\sigma(e,\theta)\right]\delta f({\bf k},v).\end{aligned}\ ] ] the corresponding spin current is then @xmath103g_s ( e),\ ] ] where the differential spin conductance at zero temperature is @xmath104\nonumber\\ & = & \frac{e^2}{h}\sum_{\sigma=\uparrow,\downarrow}\rho_\sigma \lambda^2_\sigma \int\limits_{0}^{\pi/2}{\rm d}\theta~\sin\theta\cos\theta \left[c_\sigma(e,\theta)+d_\sigma(e,\theta)\right].\end{aligned}\ ] ] the upper limit of integration over @xmath35 is determined by total reflection .
following the conservation of @xmath105 , transmission of an electron ( hole ) with @xmath106 , injected from the left electrode into the superconductor , is possible only for angles of incidence @xmath35 satisfying @xmath107 , where @xmath108 is the angle of total reflection . then , @xmath109 and @xmath110 for @xmath111 . on the other hand , @xmath112 , which corresponds to the hole ( electron ) created by the andreev reflection , is real only for @xmath113 .
the virtual andreev reflection occurs for @xmath114 , since @xmath112 becomes imaginary in that case.@xcite for injection of an electron ( hole ) with @xmath115 , transmission into the superconductor is possible for any @xmath116 , and @xmath117 is always real .
the influence of the exchange interaction on the conductance spectra is illustrated for @xmath118 and @xmath79 , for thin ( fig .
[ l3 ] ) and thick ( fig .
[ gz0 ] ) superconducting films . the spin - polarized subgap transmission of quasiparticles , and strong suppression of the andreev reflection as a consequence is significant in thin superconducting films,@xcite whereas the conductance oscillations above the gap are pronounced in the thick films .
the magnetoresistance is apparent , as charge and spin conductances are larger for the p than for the ap alignment .
the effect of interface resistance for weak nontransparency ( @xmath119 ) is illustrated in fig .
it can be shown that the fwvm has an influence similar to the nontransparency , due to the enhancement of normal reflection .
suppression of the conductance is more significant for @xmath120 . in figs .
[ l3][gz1 ] we have indicated the values of normal conductances @xmath121 and @xmath122 of the corresponding fnf double planar junction , obtained by setting @xmath123 and @xmath124 in eqs .
( [ 3d q ] ) and ( [ 3d s ] ) , where @xmath125 is given by eq .
( [ bn ] ) .
note that amplitudes of the conductance oscillations in metallic fsf junctions ( @xmath79 ) are few orders of magnitude greater than in the corresponding fnf junctions .
an important consequence of the coherency is the nontrivial spin conductance for the ap alignment , which approaches the normal value @xmath126 either for @xmath127 or in the tunnel limit for all energies .
incoherent transport through an fsf double junction is described as a transport through the corresponding fs and sf junctions in series . in that case , the conductance spectra are calculated using the generalized btk probabilities , obtained from eqs .
( [ a fs ] ) and ( [ b fs ] ) .
numerical results for the incoherent transport are also presented in figs .
[ gz0 ] and [ gz1 ] for comparison .
it is evident that in thick films the only difference comes from the interference - effect oscillations for the energies above the gap .
in contrast with the coherent transport , @xmath128 for the ap alignment , and nonequilibrium spin density accumulation changes the chemical potential of two spin subbands in the superconductor .
this reduces the superconducting gap with increasing voltage and destroys the superconductivity at a critical voltage on the order of @xmath129.@xcite
we have analyzed coherent electronic transport properties of an fsf double - barrier junction taking into account the influence of the exchange interaction , the resistance of the interfaces , and the fermi velocity mismatch on the scattering probabilities and conductance spectra .
the exchange potential , the insulating barriers at the interfaces , and the fermi velocity mismatch reduce the andreev reflection due to the enhancement of normal reflection .
it is shown that subgap electronic transmission and oscillations of differential conductances are the main features of the coherent quantum transport through a superconducting layer in fsf ( and nsn ) double junctions . in metallic junctions ,
the subgap transmission suppresses the excess current through thin superconducting films .
the scattering probabilities and conductances oscillate as a function of the layer thickness and of the quasiparticle energy above the gap .
periodic vanishing of the andreev reflection ( and the excess current ) at the energies of geometrical resonances is found as an important consequence of the quasiparticle interference . in principle , oscillations of differential conductances with the period of geometrical resonances could be used for reliable spectroscopy of quasiparticle excitations in superconductors.@xcite finite - size effects , along with the difference between coherent and incoherent transport , are essential for spin currents in fsf junctions .
besides the spin - polarized subgap tunneling in thin superconducting films , pronounced oscillations of spin conductance are found in thick films . as a consequence of the interference
, a nontrivial spin current without the excess spin accumulation and without the destruction of superconductivity by voltage is found even for the antiparallel alignment of the electrode magnetizations .
we are grateful to ivan boovi for pointing out the significance of the problem treated in this paper and for help at the initial stage of this work .
furthermore , we thank irena kneevi for useful discussions . this work has been supported by the serbian ministry of science , project @xmath130 .
the wave functions @xmath131 , eqs .
( [ psil])([psir ] ) , satisfy the boundary conditions @xmath132 neglecting @xmath133 and @xmath134 in the wave vectors , except in the exponents , eq .
( [ zeta ] ) , solutions of eqs .
( [ bc1])([bc4 ] ) for ar , nr , te , and th amplitudes can be written in the form @xmath135 , \\
\label{b general } b_\sigma(e,\theta)&=&\frac{1}{\gamma}[{\cal a}^r_+{\cal c}_+\delta^2 - \left({\cal a}^r_+{\cal c}_+e^2
+ { \cal b}^r_+{\cal d}_+\omega^2 \right)\cos(\zeta_- ) + \left({\cal a}^r_-{\cal c}_- + { \cal b}^r_-{\cal d}_-\right)\omega^2\cos(\zeta_+ ) \nonumber\\ & ~ & + i\left({\cal b}^r_+{\cal c}_+ + { \cal a}^r_+{\cal d}_+\right)e\omega\sin(\zeta_- ) - i\left({\cal b}^r_-{\cal c}_- + { \cal a}^r_-{\cal d}_-\right)\omega^2\sin(\zeta_+ ) ] , \\ \label{c general } c_\sigma(e,\theta)&=&\frac{4 ( \tilde{k}_{\sigma}/\tilde{q}_\sigma ) \omega e^{-ik^+_{\bar{\sigma}}l } } { \gamma}\times \nonumber\\ & ~&\times\ { i\left[{\cal f}_+ \cos(\zeta_+/2)+i { \cal e}_+ \sin(\zeta_+/2)\right]e\sin(\zeta_-/2 ) - \left[{\cal e}_+\cos(\zeta_+/2)+i { \cal f}_+\sin(\zeta_+/2 ) \right]\omega\cos(\zeta_-/2)\ } , \\ \label{d general } d_\sigma(e,\theta)&=&\frac{4 ( \tilde{k}_{\sigma}/\tilde{q}_\sigma ) \delta\omega e^{ik^-_{\sigma}l}}{\gamma}\times \nonumber\\ & ~&\times i\left[{\cal f}_- \cos(\zeta_+/2)+i { \cal e}_-\sin(\zeta_+/2)\right]\sin(\zeta_-/2),\end{aligned}\ ] ] where @xmath136 in eqs .
( [ a general])([gamma ] ) , @xmath137 with @xmath138}+iz}{\tilde{q}_\sigma } , \\
k^r_2 & = & \frac{\tilde{k}_{{\bar{\sigma}}[\sigma]}-iz}{\tilde{q}_\sigma},\end{aligned}\ ] ] for the p [ ap ] alignment . here , @xmath139 is the complex conjugate of @xmath140 .
general solutions for nsn double - barrier junctions correspond to @xmath141 . in the corresponding fnf double junction ,
ar and th processes are absent , @xmath142 , and the expression for nr amplitude , eq .
( [ b general ] ) , reduces to @xmath143 where @xmath144 . setting @xmath80 and @xmath87 in eq .
( [ bn ] ) , @xmath145 reduces to the result of zheng _
et al._@xcite for the transmission coefficient . to complete our considerations ,
we also present the probability amplitudes for an fs single junction in the same notation , @xmath146 note that @xmath147 and @xmath148 now describe the transmission of the bogoliubov electronlike and holelike quasiparticle , respectively . for @xmath80 and @xmath87 , eqs .
( [ a fs ] ) and ( [ b fs ] ) reduce to the results of zheng _
et al_.@xcite the well - known btk results can be reproduced by taking @xmath141 , @xmath80 , and @xmath87 in eqs .
( [ a fs])([d fs]).@xcite in thin superconducting films the spatial variation of the self - consistent pair potential is negligible and solutions retain the same form .
the only difference coming from the proximity effect is @xmath15 smaller then the bulk value , especially for transparent interfaces , see m. boovi and z. radovi , cond - mat/0409544 . , @xmath88 , @xmath85 , and @xmath89 for an fsf double junction with thin superconducting film , @xmath149 , for @xmath118 , @xmath79 , @xmath80 , @xmath81 , and p alignment .
solid curves : injection of an electron with @xmath106 . dashed curves : injection of an electron with @xmath115 . here
, @xmath84 and @xmath89 are spin independent due to the singlet - state pairing and transparent interfaces ( ref .
48).,width=377 ] , @xmath88 , @xmath85 , and @xmath89 for an fsf double junction with thick superconducting film , @xmath150 , for @xmath118 , @xmath79 , @xmath80 , @xmath81 , and p alignment .
solid curves : injection of an electron with @xmath106 . dashed curves : injection of an electron with @xmath115 . here
, @xmath84 and @xmath89 are spin independent due to the singlet - state pairing and transparent interfaces ( ref .
48).,width=377 ] and @xmath151 of an fsf double planar junction with thin superconducting film , @xmath149 , for @xmath118 , @xmath79 , @xmath80 , @xmath81 , in p and ap alignment .
conductances of the corresponding nsn junction are shown for comparison ( dotted lines ) .
arrows indicate @xmath121 and @xmath122 values . here , @xmath152 is the conductance quantum.,width=377 ] and @xmath151 of an fsf double planar junction with thick superconducting film , @xmath150 , for @xmath118 , @xmath79 , @xmath80 , @xmath81 , in p and ap alignment .
dashed curves represent the generalized btk results for the same parameters .
arrows indicate @xmath121 and @xmath122 values . here , @xmath152 is the conductance quantum.,width=377 ] and @xmath151 of an fsf double planar junction with thick superconducting film , @xmath150 , for @xmath118 , @xmath119 , @xmath80 , @xmath81 , in p and ap alignment .
dashed curves represent the generalized btk results for the same parameters .
arrows indicate @xmath121 and @xmath122 values . here , @xmath152 is the conductance quantum.,width=377 ] | coherent quantum transport in ferromagnet / superconductor / ferromagnet ( fsf ) double - barrier junctions is studied .
analytic expressions for charge and spin conductance spectra are derived for the general case of insulating interfaces ( from metallic to tunnel limit ) , the fermi velocity mismatch , and for parallel ( p ) and antiparallel ( ap ) alignment of the electrode magnetizations .
we focus on two characteristic features of finite size and coherency : subgap electronic transport , and oscillations of the differential conductance .
periodic vanishing of the andreev reflection at the energies of geometrical resonances above the superconducting gap is a striking consequence of the quasiparticle interference .
in contrast with the case of incoherent transport , a non - trivial spin - polarization without the excess spin accumulation is found for the ap alignment . |
quantum entanglement has become a fundamental resource for quantum information science , as it takes on extensive potential in the application of quantum computation and quantum communication @xcite . in particular , because of the relative simplicity and high efficiency in the generation , manipulation and detection of optical continuous - variable ( cv ) states @xcite , cv entanglement can offer an advantage in quantum information processing @xcite , and it has been an important part of quantum information theory @xcite . therefore , more and more theoretical and experimental efforts have been devoted to the generation of cv entanglement @xcite .
meanwhile , on the aspect of theory , simon @xcite and duan et al @xcite have proposed inseparability criterions for cv states , separately .
it has been proven to be efficient ways for generating the cv entangled beams that using of nondegenerate parametric down conversion ( npdc ) in a crystal @xcite .
besides , the preparation of the cv entangled light based on the interaction of two - mode cavity fields with atoms coherently driven by laser fields has also been investigated extensively . for example , li et al . @xcite considered the generation of two - mode entangled states of the cavity field via the four - wave mixing process , by means of the interaction of
properly driven v - type three - level atoms with two cavity modes .
subsequently , tan et al @xcite extended the analysis of @xcite and studied the generation and evolution of entangled light by taking into account the effects of spontaneously generated interference between two atomic decay channels . in an earlier study ,
qamar et al .
@xcite proposed a scheme for generating of two - mode entangled states in a quantum beat laser @xcite .
the system consists of a v - type three - level atom interacting with two modes of the cavity field in a doubly resonant cavity , and the atom is driven into a coherent superposition of the upper two levels by a strong classical field .
they numerically studied the property of entanglement for different values of rabi frequencies in the presence of cavity losses .
and fang et al @xcite extended the analysis of @xcite , investigating the influence of phase and rabi frequency of the classical driving field , cavity loss , and the purity and nonclassicality of the initial state of the cavity field on the property of the resulting two - mode entangled state .
moreover , in recent years , more and more theoretical and experimental efforts @xcite have been devoted to the generation of entanglement in macroscopic light based on the single - atom laser .
following by the work @xcite and @xcite , we study the generating and evolution of two - mode cv entangled states from system of v - type atom in a quantum beat laser @xcite by taking into account the effects of spontaneously generated interference . in our scheme , the two transitions in the v - type atom independently interact with the two cavity modes and the two upper levels of the atom are driven by a strong classical field .
we show that the cv entanglement with large mean number of photons can be generated in our scheme and by properly modulating the frequency detuning of the fields , the property of the filed entanglement can be adjusted .
more interesting , it is found that the entanglement can be significantly enhanced by the spontaneously generated interference , in the given situation .
let us consider the atomic system for the quantum beat laser which is proposed by scully and zubairy @xcite .
it is consist of a three - level atom with the v configuration interacting with two ( nondegenerate ) cavity modes and the two upper levels of the atom are driven by a strong classical field . in fig . 1
, we show the the atomic level scheme .
the atom is pumped at a rate @xmath0 into the level @xmath1 . use a strong magnetic field with rabi frequency @xmath2 to drive the transition between @xmath1 and @xmath3 , which is electric - dipole forbidden . while the two nondegenerate cavity modes of frequencies @xmath4 and @xmath5 independently interact with the transitions @xmath6 ( with @xmath7 resonant frequency ) and @xmath8 ( with @xmath9 resonant frequency ) , respectively .
@xmath10 is the detuning of the field @xmath11 from the corresponding atomic transition @xmath6 .
@xmath12 is the detuning of the field @xmath13 from the corresponding atomic transition @xmath8.for simplicity , we ll note @xmath14 .
schematic diagram of the three - level atom system in a v configuration .
two ( nondegenerate ) cavity modes with coupling constant @xmath11 and @xmath13 interact with the transition @xmath6 and @xmath8 , respectively , while the atom transition @xmath15 is driven by a strong magnetic field with rabi frequency @xmath2 . @xmath16 and @xmath17 correspond the frequency detunings.,width=340 ] then , under the dipole and rotating wave approximation , the total interaction hamiltonian of our system can be given in the interaction picture by ( @xmath18 ) @xmath19 where the symbol @xmath20 means the hermitian conjugate and
we have taken the ground state @xmath21 as the energy origin for the sake of simplicity . @xmath22
describe the rabi frequencies of the strong classical field and denote the phase of classical fields with @xmath23 , and @xmath11 and @xmath13 are the atom@xmath24field coupling constants .
@xmath25 ( @xmath26 ) is the annihilation ( creation ) operator of the corresponding cavity modes . considering the vacuum damping of the atom and the cavity modes ,
the reduced density equations of the cavity fields ( taking a trace over the atom degrees of freedom @xcite ) can be obtained from the hamiltonian ( 1 ) : @xmath27+tr_{atom}(l_f\rho+l_a\rho)\nonumber\\ & = & -ig_1[a_1^{\dag},\rho_{ac}]-ig_2[a_2^{\dag},\rho_{bc}]+\sum_{j=1}^{2}\kappa_j[a_j,\rho_f a_j^{\dag}]+h.c.,\end{aligned}\ ] ] with @xmath28+h.c.,\\ \label{eq4 } & & l_a\rho=\gamma_1[s_{ca},\rho s_{ac}]+\gamma_2[s_{cb},\rho s_{bc}]+\gamma_{12}([s_{ca},\rho s_{bc}]+[s_{cb},\rho s_{ac}])+h.c.,\end{aligned}\ ] ] where @xmath29 , @xmath30 are the vacuum damping of the cavity modes and atom , respectively .
@xmath31 , and @xmath32 are the decay rates from the states @xmath1 to @xmath21,and @xmath3 to @xmath21 , respectively .
meanwhile , @xmath33 represents the spontaneously generated interference which is resulted from the cross coupling between the transitions @xmath6 and @xmath8 , and @xmath34 . here
@xmath35 and @xmath36 represent the atomic dipole polarizations and @xmath37 is the angle between the two dipole moments . from the expressions of the parameter p and @xmath38 we can find that the spontaneously generated interference dependents on the angle between the two dipole moments .
when the two dipole moments are perpendicular to each other the interference effect disappears ( p=0 ) , and it will be maximal ( p=1 ) if the two dipole moments are parallel to each other . note that @xmath39 ( @xmath40 ) are the damping constants of two cavity modes based on the standard methods of laser theory in @xcite , considering the spontaneously decay of atom , @xmath41 and @xmath42 can be evaluated to the first order in the coupling constants @xmath11 and @xmath13 as @xmath43 where the density matrix elements @xmath44 can be obtained by the corresponding zeroth - order equations @xmath45 plugging the steady - state solution of @xmath44 into eqs .
( [ eq5a]-[eq5b ] ) , we find the steady - state solution for @xmath41 and @xmath42 can be described as : @xmath46 with the explicit expressions of the coefficients @xmath47 being given in appendix . by substituting eqs.([eq7a]-[eq7b ] ) to the eq .
( [ eq2 ] ) , the reduced master equation govern the evolution of the cavity field can be obtained as @xmath48 here we remain all orders in the rabi frequency @xmath2 whereas only consider second order in the coupling constants @xmath11 , @xmath13 due to that the coupling constants of two cavity modes are smaller than other system parameters in our scheme .
thus we can ignore the saturation effects and operate in the regime of linear amplification .
in this section , we use the sufficient inseparability criterion proposed by duan et al @xcite . to verify that cv entanglement with large mean number of photons can be obtained in our model , and study the property of entanglement under the given conditions . according to duan s criterion @xcite ,
the two cavity modes are entangled if and only if the sum of the variances of the two einstein - podolsky - rosen ( epr ) type operators @xmath49 and @xmath50 satisfies the following inequality @xmath51 with the pair quadrature operators @xmath52 and @xmath53 ( @xmath40 ) the local operators which correspond to the mode at the frequency @xmath54 . by substituting @xmath55 and @xmath56 into equation ( [ eq9 ] )
, we can express the total variance of the operators @xmath57 and @xmath58 in terms of the operators @xmath25 and @xmath59 and achieve @xmath60.\end{aligned}\ ] ] with the help of equation ( [ eq8]),we can obtain the equations of motion for the expectation values of the field operators in equation ( [ eq10 ] ) as @xmath61 by numerically solving these equations , we can give a few numerical results for the time evolution of the total photon numbers @xmath62 and @xmath63 with different values of parameter while the cavity modes are assumed to be in the coherent state @xmath64 , as illustrated in figs .
it is easy to find that , the cv entanglement with large mean number of photons from system of v - type atom in a quantum beat laser can be generated , and the entanglement can be enhanced by the spontaneously generated interference . for simplicity ,
all the parameters used here are scaled with @xmath65 , and we have chosen @xmath66 during our numerical calculations . by above knowable , @xmath33 represents the spontaneously generated interference .
therefore , in order to check the effect of the spontaneously generated interference on the time evolution of entanglement , we can study the property of entanglement for different values of atomic decay rates and the parameter p. in fig .
2 , we show the plot of the time development of @xmath67 and @xmath68 for different values of atomic decay rates , when the cavity field is initially in the coherent state @xmath64 . from fig .
2(a ) , one can find that the intensity and period of entanglement can be enlarged by increasing spontaneous emission decay rates of the atom level , while fig .
2(b ) illustrates that the effect of the atomic decay rates @xmath69 ( @xmath40 ) on the maximum mean photon numbers give a small extent .
[ fig2 ] the time evolution of @xmath67 ( shown in fig .
2(a ) ) and the total mean photon numbers @xmath68 ( shown in fig .
2(b ) ) for different atomic decay rates @xmath70 ( @xmath71 ) , when the cavity field is initially in the coherent state @xmath64 .
the other parameters are @xmath72 , @xmath73 , @xmath74 , @xmath75 , @xmath76 and @xmath66.,title="fig:",width=264 ] the time evolution of @xmath67 ( shown in fig .
2(a ) ) and the total mean photon numbers @xmath68 ( shown in fig .
2(b ) ) for different atomic decay rates @xmath70 ( @xmath71 ) , when the cavity field is initially in the coherent state @xmath64 .
the other parameters are @xmath72 , @xmath73 , @xmath74 , @xmath75 , @xmath76 and @xmath66.,title="fig:",width=264 ] [ fig3 ] the time evolution of @xmath67 ( shown in fig .
3(a ) ) and the total mean photon numbers @xmath68 ( shown in fig .
3(b ) ) different values of p , when the cavity field is initially in the coherent state @xmath64 .
the other parameters are @xmath72 , @xmath77 , @xmath74 , @xmath75 , @xmath76 and @xmath66.,title="fig:",width=264 ] the time evolution of @xmath67 ( shown in fig .
3(a ) ) and the total mean photon numbers @xmath68 ( shown in fig .
3(b ) ) different values of p , when the cavity field is initially in the coherent state @xmath64 .
the other parameters are @xmath72 , @xmath77 , @xmath74 , @xmath75 , @xmath76 and @xmath66.,title="fig:",width=264 ] in order to check the influence of p on the entanglement property , we numerically simulate the time evolution of @xmath67 and @xmath68 for different values of p. as shown in fig .
3(a ) , when the cavity field is initially in coherent state @xmath64 , the intensity of entanglement between the two cavity modes slightly enhances with the increase of the value of p. with the same set of the parameters , fig .
3(b ) shows that the maximum mean photon numbers @xmath68 become more pronounced with the increase of the value of p. up to now , we have investigated the the influence of spontaneous emission decay rates of the atom level and the parameter p on the time evolution of entanglement .
it is easy to find that , the entanglement can be enhanced by no matter increasing the spontaneous emission decay rates of the atom level or increasing the value of p. then we conclude that , the spontaneously generated interference which depends on @xmath69 ( @xmath40 ) and p , can strengthen the entanglement .
we plot the influence of frequency detuning of the cavity field on the time evolution of entanglement under condition that the cavity field is initially in the coherent state @xmath64 ( shown in fig .
4 ) . as shown in fig .
4(a ) that the intensity and period of entanglement between the two cavity modes can be enlarged at one time by decreasing the frequency detuning of the pump field @xmath78(@xmath14 ) .
in addition , fig .
4(b ) illustrate that with the increased of the detuning @xmath78 , the maximum mean photon number is enlarged .
this result implicates that in order to obtain the entanglement of cavity modes with high intensity and longer period we can do that by properly adjusting the frequency detuning .
[ fig4 ] the time evolution of @xmath67 ( shown in fig.4(a ) ) and the total mean photon numbers @xmath68 ( shown in fig .
4(b ) ) for different frequency detuning @xmath78(@xmath14 ) , when the cavity field is initially in the coherent state @xmath64 .
the other parameters are @xmath79 , @xmath72 , @xmath73 , @xmath74 , @xmath76 and @xmath66.,title="fig:",width=264 ] the time evolution of @xmath67 ( shown in fig.4(a ) ) and the total mean photon numbers @xmath68 ( shown in fig .
4(b ) ) for different frequency detuning @xmath78(@xmath14 ) , when the cavity field is initially in the coherent state @xmath64 .
the other parameters are @xmath79 , @xmath72 , @xmath73 , @xmath74 , @xmath76 and @xmath66.,title="fig:",width=264 ] before conclusion , we should note that our scheme is drastically different from the conventional scheme of cv entanglement generation @xcite . in our scheme , with decreasing the frequency detuning , our numerical results showed that a long entanglement time and strong entanglement intensity can be synchronously achieved .
it illustrates that the entanglement of cavity modes with higher intensity and longer period can be realized in our scheme with a low - q cavity . and
it can strengthen the entanglement by increasing the spontaneous emission decay rates of the atom level and the value of the parameter p. the physical reason can be explained that no matter larger the atom decay rates @xmath69 ( @xmath40 ) or larger the value of p results in increasing the spontaneously generated interference which can enhance the entanglement .
all these distinguish advances illustrate that our scheme is drastically different from the conventional scheme .
in summary , we have proposed a new scheme to generate the cv entanglement and investigated the evolution of it from a system of v - type atom in a quantum beat laser @xcite . in this scheme ,
the two transitions in the v - type atom independently interact with the two cavity modes while the two upper levels of the atom are driven by a strong classical field . by taking into account the effects of
spontaneously generated interference between two atomic decay channels , and using the standard methods of laser theory @xcite , we show that , in the given conditions , the cv entanglement with large mean number of photons can be realized in our scheme . and by properly modulating the frequency detuning of the field can adjust the entanglement period , intensity and the total mean photon numbers of two cavity modes .
different from the conventional scheme @xcite , the cv entanglement of cavity modes with higher intensity and longer period can be realized in our scheme with a low - q cavity .
furthermore , our results showed that the entanglement can be significantly enhanced by the spontaneously generated interference .
we would like to thank prof .
peng xue for her enlightening discussions .
the research is supported in part by national natural science foundation of china under grant nos .
11374050 and 61372102 , by qing lan project of jiangsu , and by the fundamental research funds for the central universities under grant no .
here we give the expressions of the coefficients @xmath80 and @xmath81 ( @xmath82 ) in eqs .
( [ eq7a]-[eq8 ] ) , @xmath83/d_1,\\ \label{eqa2 } & & \alpha_{12}=-g_1g_2[(r+i\delta)\beta_{ab}+(i\omega-\gamma_{12})\beta_{bb}]/d_1,\\ \label{eqa3 } & & \alpha_{22}=-g_2 ^ 2[(r+i\delta)\beta_{bb}+(i\omega^*-\gamma_{12})\beta_{ab}]/d_1,\\ \label{eqa4 } & & \alpha_{21}=-g_1g_2[(r+i\delta)\beta_{ba}+(i\omega^*-\gamma_{12})\beta_{aa}]/d_1,\\ \label{eqa5 } & & \beta_{aa}=(2r^2-\gamma_{12}^2+|\omega|^2)r\gamma_a / d_2,\\ \label{eqa6 } & & \beta_{bb}=(\gamma_{12}+i\omega)(\gamma_{12}-i\omega^*)r\gamma_a / d_2,\\ \label{eqa7 } & & \beta_{ab}=-(\gamma_{12}+i\omega)(2r^2-i\gamma_{12}\omega - i\gamma_{12}\omega^*)r\gamma_a/(2d_2),\\ \label{eqa8 } & & \beta_{ba}=-(\gamma_{12}-i\omega^*)(2r^2+i\gamma_{12}\omega+i\gamma_{12}\omega^*)r\gamma_a/(2d_2),\end{aligned}\ ] ] | we investigate the generation and the evolution of two - mode continuous - variable ( cv ) entanglement from system of a microwave - driven v - type atom in a quantum beat laser . by taking into account the effects of spontaneously generated quantum interference between two atomic decay channels ,
we show that the cv entanglement with large mean number of photons can be generated in our scheme , and the property of the filed entanglement can be adjusted by properly modulating the frequency detuning of the fields .
more interesting , it is found that the entanglement can be significantly enhanced by the spontaneously generated interference . |
a _ knot _ is an equivalence class of embeddings of the circle @xmath0 into the 3-sphere @xmath1 , where two embeddings are said to be _ equivalent _ if an ambient isotopy of @xmath1 deforms one to the other . in knot theory
, it is a fundamental and important problem to determine whether given two representatives of knots are equivalent , and furthermore to describe how one can be deformed to the other .
in particular , a simplification to a minimal position " is of great interest .
let @xmath2 be the standard height function , that is , the restriction of @xmath3 to @xmath1 .
morse position _ of a knot @xmath4 is a representative @xmath5 such that @xmath5 is disjoint from the poles of @xmath1 and the critical points of @xmath6 are all non - degenerate and have pairwise distinct values . since @xmath5 is a circle ,
there are the same number of local maxima and local minima . in @xcite ,
schubert introduced the notion of bridge position and bridge number for knots .
a _ bridge position _ of @xmath4 is a morse position @xmath5 where all the local maxima are above all the local minima with respect to @xmath7 .
a level @xmath8-sphere @xmath9 separating the local maxima from the local minima is called a _ bridge sphere _ of @xmath5 .
if @xmath5 intersects @xmath9 in @xmath10 points , then @xmath5 is called an _ @xmath11-bridge position _ and @xmath11
is called the _ bridge number _ of @xmath5 .
the minimum of the bridge number over all bridge positions of @xmath4 is called the _ bridge number _ of @xmath4 . a knot with the bridge number @xmath11
is called an _
@xmath11-bridge knot_. the bridge number is a fundamental geometric invariant of knots as well as the crossing number . in @xcite
, gabai introduced the notion of width for knots .
suppose that @xmath5 is a morse position of a knot @xmath4 , let @xmath12 be the critical levels of @xmath6 such that @xmath13 for @xmath14 , and choose regular levels @xmath15 of @xmath6 so that @xmath16 .
the _ width _ of @xmath5 is defined as @xmath17 , and the _ width _ of @xmath4 is the minimum of the width over all morse positions of @xmath4 .
two morse positions of a knot are _ isotopic _ if an ambient isotopy of @xmath1 deforms one to the other keeping it a morse position except for exchanging two levels of local maxima or two levels of local minima .
such an isotopy preserves the width of a morse position and the bridge number of a bridge position .
the following two types of moves change the isotopy class of a morse position .
suppose that @xmath5 is a morse position of a knot @xmath4 and let @xmath18 as above .
we say that the level 2-sphere @xmath19 is _ thick _ if @xmath20 is a locally minimal level of @xmath6 and @xmath21 is a locally maximal level of @xmath6 , and that @xmath19 is _ thin _ if @xmath20 is a locally maximal level and @xmath21 is a locally minimal level .
a _ strict upper _ ( resp . _ lower _ ) _ disk _ for a thick sphere @xmath9 is a disk @xmath22 such that the interior of @xmath23 does not intersect with @xmath5 and any thin sphere , the interior of @xmath23 contains no critical points with respect to @xmath7 , and @xmath24 consists of a subarc @xmath25 of @xmath5 and an arc in @xmath26 .
note that the arc @xmath25 has exactly one local maximum ( resp .
minimum ) .
first suppose that there exist a strict upper disk @xmath27 and a strict lower disk @xmath28 for a thick sphere @xmath9 such that @xmath29 consists of a single point of @xmath30 .
then @xmath5 can be isotoped along @xmath27 and @xmath28 to cancel the local maximum in @xmath31 and the local minimum in @xmath32 . in @xcite , schultens called this move a _ type i move_. the inverse operation of a type i move is called a _ stabilization _
( @xcite , @xcite ) or a _ perturbation _ ( @xcite , @xcite ) and
the resultant position is said to be _ stabilized _ or _
perturbed_. next suppose that there exist a strict upper disk @xmath27 and a strict lower disk @xmath28 for a thick sphere @xmath9 such that @xmath33 .
then @xmath5 can be isotoped along @xmath27 and @xmath28 to exchange the two levels of the local maximum in @xmath31 and the local minimum in @xmath32 .
schultens ( @xcite ) called this move a _ type ii move_. the following are fundamental theorems for bridge positions and morse positions which correspond to reidemeister s theorem ( @xcite , @xcite ) for knot diagrams .
[ fundamental1 ] two knots are equivalent if and only if their two bridge positions can be related by a sequence of type i moves and the inverse operations up to isotopy .
[ fundamental2 ] two knots are equivalent if and only if their two morse positions can be related by a sequence of type i and type ii moves and the inverse operations up to isotopy .
we say that a bridge position of a knot @xmath4 is _ globally minimal _ if it realizes the bridge number of @xmath4 , and a bridge position is _ locally minimal _ if it does not admit a type i move .
similarly , we say that a morse position of a knot @xmath4 is _ globally minimal _ if it realizes the width of @xmath4 , and a morse position is _ locally minimal _ if it does not admit a type i move nor a type ii move .
note that if a bridge ( resp .
morse ) position of a knot is globally minimal , then it is locally minimal .
otal ( later hayashi shimokawa , the first author ) proved the converse for bridge positions of the trivial knot .
[ otal ] a locally minimal bridge position of the trivial knot is globally minimal .
this implies that even complicated bridge positions of the trivial knot can be simplified into the @xmath34-bridge position only by type i moves .
furthermore , otal ( later scharlemann tomova ) showed that the same statement for 2-bridge knots is true ( @xcite , @xcite ) , and the first author showed that the same statement for torus knots is also true ( @xcite ) .
then , the following problem is naturally proposed .
[ question ] is any locally minimal bridge position of any knot globally minimal ? in this paper , we give a negative answer to this problem .
it implies that a bridge position can not always be simplified into a minimal bridge position only by type i moves .
[ main ] a @xmath35-bridge position @xmath36 of a knot @xmath37 in figure [ example ] is locally minimal , but not globally minimal .
@xmath36 + -bridge position of a knot.,title="fig : " ] @xmath38 + to prove this theorem , we show that the hempel distance of the 4-bridge position is greater than @xmath34 by the method developed by the second author ( @xcite ) .
this guarantees that the 4-bridge position is locally minimal ( see lemma [ minimal ] ) .
on the other hand , zupan showed that locally minimal , but not globally minimal morse positions exist even if the knot is trivial .
there exists a locally minimal morse position of the trivial knot which is not globally minimal .
we remark that this answers scharlemann s question ( * ? ? ?
* question 3.5 ) . by using this example
, zupan showed that there exist infinitely many locally minimal , but not globally minimal morse positions for any knot .
note that figure [ example ] displays a @xmath39-bridge position of @xmath37 after a @xmath40-rotation of @xmath36 , and so the @xmath35-bridge position is not globally minimal . to prove that the @xmath35-bridge position is locally minimal , we apply the following : [ minimal ] an @xmath11-bridge position is locally minimal if it has hempel distance greater than @xmath34 . [ criterion ] an @xmath11-bridge position has hempel distance greater than @xmath34 if a bridge diagram of it satisfies the well - mixed condition . here
@xmath11 is an integer greater than @xmath8 .
the notions of _ hempel distance _ , _ bridge diagram _ and _ well - mixed condition _ are described in the following subsections .
suppose that @xmath5 is an @xmath11-bridge position of a knot @xmath4 and @xmath9 is a bridge sphere of @xmath5 .
let @xmath41 be the @xmath39-balls divided by @xmath9 , and @xmath42 be the @xmath11 arcs @xmath43 for each @xmath44 .
consider a properly embedded disk @xmath45 in @xmath46 .
we call @xmath45 an _ essential disk _ of @xmath47 if @xmath45 is disjoint from @xmath42 and @xmath48 is essential in the @xmath10-punctured sphere @xmath49 . here , a simple closed curve on a surface is said to be _ essential _ if it neither bounds a disk nor is peripheral in the surface .
the essential simple closed curves on @xmath49 form a @xmath34-complex @xmath50 , called the _ curve graph _ of @xmath49 .
the vertices of @xmath50 are the isotopy classes of essential simple closed curves on @xmath49 , and a pair of vertices spans an edge of @xmath50 if the corresponding isotopy classes can be realized as disjoint curves .
the _ hempel distance _ of @xmath5 is defined as @xmath51,[\partial e_-])\mid e_\varepsilon \text { is an essential disk of } ( b_\varepsilon , \tau _
\varepsilon ) \text { for each } \varepsilon = \pm .\ } , \ ] ] where @xmath52,[\partial e_-])$ ] is the minimal distance between @xmath53 $ ] and @xmath54 $ ] measured in @xmath50 with the path metric .
assume that @xmath5 has hempel distance @xmath55 . by the definition ,
there exist essential disks @xmath56 of @xmath57 , @xmath58 , respectively , such that @xmath59 , which requires that @xmath5 is split . since the circle @xmath5 is connected , the hempel distance is at least @xmath34 .
the hempel distance is @xmath34 if there exist essential disks @xmath56 of @xmath60 , respectively , such that @xmath61 .
we can find such disks for a not locally minimal bridge position as follows : assume that the @xmath11-bridge position @xmath5 is not locally minimal . by definition , there exist a strict upper disk @xmath62 and a strict lower disk @xmath63 for @xmath9 such that @xmath64 consists of a single point of @xmath30 .
note that @xmath65 is @xmath11 arcs each of which has a single local minimum .
we can choose strict lower disks @xmath66 for @xmath9 such that @xmath67 are pairwise disjoint .
let @xmath68 denote a closed regular neighborhood of @xmath69 in @xmath1 .
by replacing subdisks of @xmath70 with subdisks of @xmath71 , we can arrange that @xmath67 are disjoint from @xmath27 except for the two points of @xmath30 .
since we assumed @xmath72 , one of the strict lower disks , denoted by @xmath28 , is disjoint from @xmath27 .
the boundary of a regular neighborhood in @xmath1 of each @xmath73 intersects @xmath46 in an essential disk of @xmath47 .
they guarantee that the hempel distance is @xmath34 .
we continue with the above notation .
there are @xmath11 pairwise disjoint strict upper ( resp .
lower ) disks @xmath74 ( resp .
@xmath75 , @xmath76 ) for @xmath9 .
the knot diagram of @xmath4 obtained by projecting @xmath5 into @xmath9 along these disks is called a _
bridge diagram _ of @xmath5 . in the terminology of @xcite , @xmath77
are the overpasses and the underpasses of @xmath5 .
now let us describe how we can obtain a bridge diagram of the @xmath35-bridge position @xmath36 .
isotope @xmath36 as in figure [ braid ] , and start with a bridge sphere @xmath78
. there are canonical strict upper disks @xmath79 and @xmath80 .
figure [ diagram-1 ] illustrates a view of the arcs @xmath81 and @xmath82 on @xmath9 from @xmath83 side . shifting the bridge sphere @xmath9 to @xmath84 ,
the arcs can be seen as in figure [ diagram-2 ] . shifting @xmath9
further to @xmath85 and to @xmath86 , the arcs are as in figure [ diagram-3 ] and [ diagram-4 ] , respectively . by continuing this process ,
the arcs are as in figure [ diagram-5 ] when @xmath9 is at @xmath87 .
the picture grows more and more complicated as @xmath9 goes down .
we include huge pictures in the back of this paper .
figure [ diagram-6 ] illustrates the arcs when @xmath9 is at @xmath88 , and finally figure [ diagram-7 ] when @xmath9 has arrived at @xmath89 .
then we can find canonical strict lower disks @xmath90 and obtain a bridge diagram of @xmath36 .
.,title="fig : " ] + @xmath91@xmath92@xmath93@xmath80 + @xmath94 + @xmath84 + @xmath85 + @xmath86 + @xmath87 + @xmath88 + @xmath89 + @xmath95@xmath96@xmath97@xmath98 @xmath99@xmath100@xmath101@xmath82 + .,title="fig : " ] . ] . ] . ] . ]
suppose again that @xmath5 is an @xmath11-bridge position of a knot @xmath4 with @xmath72 and @xmath9 is a bridge sphere of @xmath5 .
let @xmath41 be the @xmath39-balls divided by @xmath9 , and @xmath42 be the @xmath11 arcs @xmath102 for each @xmath44 .
let @xmath74 and @xmath67 be strict upper and lower disks for @xmath9 determining a bridge diagram of @xmath5 .
let @xmath103 be a loop on @xmath9 containing the arcs @xmath104 such that the arcs are located in @xmath103 in that order .
we can assume that @xmath74 have been isotoped so that the arcs @xmath105 have minimal intersection with @xmath103 .
for the bridge diagram of figure [ diagram-7 ] , it is natural to choose @xmath103 to be the one which is seen as a horizontal line .
let @xmath106 be the hemi - spheres divided by @xmath103 and let @xmath107 ( @xmath108 ) be the component of @xmath109 which lies between @xmath110 and @xmath111 .
( here the indices are considered modulo @xmath11 . )
let @xmath112 be the collection of components of @xmath113 separating @xmath107 from @xmath114 in @xmath115 for a distinct pair @xmath116 and @xmath117 .
for example , figure [ ( 1,2,+ ) ] roughly displays @xmath118 for the bridge diagram of figure [ diagram-7 ] .
note that @xmath112 consists of parallel arcs in @xmath115 . of arcs , which looks like gray bands.,title="fig : " ] + @xmath119 + @xmath120@xmath121@xmath122@xmath123@xmath124@xmath125@xmath126@xmath127@xmath120 + @xmath128 +
1 . a bridge diagram satisfies the _ @xmath129-well - mixed condition _ if in @xmath130 , a subarc of @xmath131 is adjacent to a subarc of @xmath132 for every distinct pair @xmath133 .
2 . a bridge diagram satisfies the _ well - mixed condition _ if it satisfies the @xmath129-well - mixed condition for every combination of a distinct pair @xmath116 and @xmath117 . by making figure [ ( 1,2,+ ) ]
detailed , one can check the @xmath134-well - mixed condition for the bridge diagram of figure [ diagram-7 ] .
one can also check the @xmath129-well - mixed condition for all the other @xmath129 to complete the well - mixed condition . by theorem [ criterion ] ,
the hempel distance of @xmath36 is greater than @xmath34 . by lemma [ minimal ] , we conclude the proof of theorem [ main ] .
we would like to remark that the hempel distance of @xmath36 is exactly @xmath8 .
notice that the boundary of a regular neighborhood in @xmath9 of the closure of @xmath126 is a simple closed curve disjoint from both @xmath91 and @xmath95 .
note that the boundary of a regular neighborhood in @xmath1 of each @xmath135 intersects @xmath46 in an essential disk of @xmath47 .
they guarantee that the hempel distance is at most @xmath8 .
figure [ example ] shows that the bridge number of @xmath37 is at most @xmath39 . since any locally minimal bridge position of any 2-bridge knot is globally minimal ( @xcite , @xcite ) , the bridge number of @xmath37 is equal to @xmath39 .
since @xmath37 has a @xmath35-bridge position with the hempel distance @xmath8 , it is a prime knot by the following : let @xmath5 be an @xmath11-bridge position of a composite knot and @xmath9 be a bridge sphere of @xmath5 .
let @xmath41 be the @xmath39-balls divided by @xmath9 , and @xmath42 be the @xmath11 arcs @xmath43 for each @xmath44 . by the arguments in @xcite , @xcite
, it follows that any decomposing sphere for @xmath5 can be isotoped so that it intersects @xmath9 in a single loop .
then , in the opposite sides of the decomposing sphere , there are two essential disks @xmath56 of @xmath60 respectively such that @xmath61 .
this shows that the hempel distance is @xmath34 .
furthermore , @xmath37 is hyperbolic since any locally minimal bridge position of any torus knot is globally minimal ( @xcite ) .
thus , @xmath37 is a hyperbolic 3-bridge knot which admits a @xmath35-bridge position with the hempel distance 2 .
we expect that not only @xmath36 may be an example for theorem [ main ] but also many @xmath35-bridge positions of knots with the same projection image as that of figure [ example ] .
however only finitely many knots have the same projection image , and we would like to ask the following problem .
we further expect that for some integers @xmath137 , we can find a locally minimal @xmath11-bridge position of an @xmath138-bridge knot which has a similar projection image as that of figure [ example ] .
however it seems difficult to find more than two locally minimal bridge positions of such a knot , and we would like to ask the following problem .
it should be remarked that there exist only finitely many bridge positions of given bridge numbers for a hyperbolic knot ( @xcite ) . in particular , there are finitely many globally minimal bridge positions of a hyperbolic knot .
it should be also remarked that multiple bridge surfaces restrict hempel distances ( @xcite ) .
composite knots are a simple example of knots with essential surfaces properly embedded in the exteriors of their representatives . theorem [ composite ] suggests that essential surfaces restrict hempel distances .
schleimer showed it in general .
[ bound ] let @xmath139 be an orientable essential surface properly embedded in the exterior of a bridge position @xmath5 of a knot . then the hempel distance of @xmath5 is bounded above by twice the genus of @xmath139 plus @xmath140 . by theorem
[ bound ] , if a knot exterior contains an essential annulus or an essential torus , then the hempel distance of a bridge position is at most 2 .
therefore , if there exists a bridge position of a knot with the hempel distance at least 3 , then the knot is hyperbolic .
the properties of our knot @xmath37 can be compared with it .
a knot without an essential surface with meridional boundary in the exterior of its representative is called a _ meridionally small knot_. for example , the trivial knot , 2-bridge knots and torus knots are known to be meridionally small . as we mentioned in section [ intro ] , these knots also have the nice property that any nonminimal bridge position is stabilized .
we say that a knot is _ destabilizable _ if it has this property .
zupan showed that any cabled knot @xmath141 of a meridionally small knot @xmath4 is also meridionally small , and that if @xmath4 is destabilizable , then @xmath141 is also destabilizable ( @xcite ) .
then , the following problem is naturally proposed . on the other hand , it is not always true that if the knot exterior contains an essential surface with meridional boundary , then a bridge position has hempel distance @xmath34 .
for example , ( * ? ? ?
* example 5.1 ) shows that a @xmath39-bridge position of @xmath142 has hempel distance greater than @xmath34 , but the knot exterior contains an essential surface with meridional boundary .
theorem [ fundamental1 ] allows us to define a distance between two bridge positions of a knot , which we call the _
birman distance_. that is to say , the birman distance between two bridge positions is the minimum number of type i moves and the inverse operations relating the bridge positions up to isotopy .
for example , the birman distance between an @xmath11-bridge position and an @xmath138-bridge position of the trivial knot is always @xmath143 by theorem [ otal ] .
the birman distance between @xmath36 and the @xmath39-bridge position of @xmath37 is at least @xmath39 since @xmath36 is locally minimal .
in fact , we can see that it is at most @xmath144 by observing the @xmath40-rotation of @xmath36 .
tomova gave an upper bound for the birman distance between two bridge positions with high hempel distance which are obtained from each other by flipping , namely the rotation of @xmath1 exchanging the poles .
for an integer @xmath145 , if an @xmath11-bridge position @xmath5 of a prime knot has hempel distance at least @xmath146 , then the birman distance between @xmath5 and the flipped bridge position of @xmath5 is @xmath147 .
we remark that the @xmath10-bridge position is not locally minimal , and hence it does not answer problem [ question ] .
it turns out that there are two @xmath149-bridge positions such that the birman distance is at least @xmath151 .
the following are major problems .
10 j. w. alexander and g. b. briggs , _ on types of knotted curves _ ,
ann . of math .
( 2 ) * 28 * ( 1926/27 ) , no . 1 - 4 , 562586 .
j. s. birman , _ on the stable equivalence of plat representations of knots and links _
, canad . j. math .
* 28 * ( 1976 ) , no .
2 , 264290 . d. bachman and s. schleimer , _ distance and bridge position _ , pacific j. math . * 219 * ( 2005 ) , no .
2 , 221235 .
a. coward , _ algorithmically detecting the bridge number of hyperbolic knots _ , arxiv:0710.1262 .
r. h. crowell and r. h. fox , introduction to knot theory , reprint of the 1963 original , graduate texts in mathematics , * 57 * , springer - verlag , new york - heidelberg , 1977 .
d. gabai , _ foliations and the topology of 3-manifolds . iii _ , j. differential geom .
* 26 * ( 1987 ) , no .
3 , 479536 . c. hayashi , _ stable equivalence of heegaard splittings of @xmath34-submanifolds in @xmath39-manifolds _ , kobe j. math . *
15 * ( 1998 ) , no .
2 , 147156 . c. hayashi and k. shimokawa , _ heegaard splittings of the trivial knot _ , j. knot theory ramifications * 7 * ( 1998 ) , no . 8 , 10731085 . c. hayashi and k. shimokawa , _ thin position of a pair ( 3-manifold , 1-submanifold ) _ , pacific j. math .
* 197 * ( 2001 ) , no .
2 , 301324 .
d. j. heath and t. kobayashi , _ essential tangle decomposition from thin position of a link _ , pacific j. math .
* 179 * ( 1997 ) , no .
1 , 101117 .
j. johnson and m. tomova , _ flipping bridge surfaces and bounds on the stable bridge number _ , algebr .
geom . topol .
* 11 * ( 2011 ) , no .
4 , 19872005 . j .-
otal , _ prsentations en ponts du nud trivial _ , c. r. acad .
paris sr .
* 294 * ( 1982 ) , no .
16 , 553556 .
otal , _ presentations en ponts des nuds rationnels _ , low - dimensional topology ( chelwood gate , 1982 ) , 143160 , london math .
lecture note ser .
, 95 , cambridge univ . press ,
cambridge , 1985 .
m. ozawa , _ bridge position and the representativity of spatial graphs _ ,
topology and its appl .
* 159 * ( 2012 ) , no .
4 , 936947 .
m. ozawa , _ non - minimal bridge positions of torus knots are stabilized _ , math .
cambridge philos .
* 151 * ( 2011 ) 307317 .
k. reidemeister , _ elementare begrndung der knotentheorie _ , abh .
hamburg * 5 * ( 1927 ) 2432 .
m. scharlemann , _ thin position in the theory of classical knots _ , handbook of knot theory , 429459 , elsevier b. v. , amsterdam , 2005 .
m. scharlemann and m. tomova , _ uniqueness of bridge surfaces for 2-bridge knots _ , math .
cambridge philos .
* 144 * ( 2008 ) , no .
3 , 639650
. h. schubert , _ ber eine numerische knoteninvariante _ , math . z.
* 61 * ( 1954 ) , 245288 .
j. schultens , _ additivity of bridge numbers of knots _ , math .
cambridge philos .
* 135 * ( 2003 ) , no .
3 , 539544 .
j. schultens , _ width complexes for knots and 3-manifolds _ , pacific j. math .
* 239 * ( 2009 ) , no .
1 , 135156 .
k. takao , _ bridge decompositions with distances at least two _
, hiroshima math . j.
* 42 * ( 2012 ) , no .
2 , 161168 .
m. tomova , _ thin position for knots in a 3-manifold _ , j. lond .
( 2 ) * 80 * ( 2009 ) , no .
1 , 8598 .
m. tomova , _ multiple bridge surfaces restrict knot distance _
* 7 * ( 2007 ) , 9571006 .
a. zupan , _ properties of knots preserved by cabling _ , comm .
* 19 * ( 2011 ) , no .
3 , 541562 .
a. zupan , _ unexpected local minima in the width complexes for knots _ , algebr .
geom . topol .
* 11 * ( 2011 ) , no .
2 , 10971105 .
-bridge position , which is decomposed into the following 14 pages.,title="fig : " ] + @xmath153@xmath154@xmath155@xmath156@xmath157@xmath158@xmath159 + @xmath160@xmath161@xmath162@xmath163@xmath164@xmath165@xmath166 + | we give a locally minimal , but not globally minimal bridge position of a knot , that is , an unstabilized , nonminimal bridge position of a knot .
it implies that a bridge position can not always be simplified so that the bridge number monotonically decreases to the minimal . |
the principle of sufficient reason ( psr ) , although famously espoused and advocated by rationalist philosophers such as spinoza and leibniz in the xvii century , has an illustrious history that pervades the whole western thought ( see schopenhauer 2012 ) .
anaximander is usually credited as the first thinker that invoked the psr in philosophical reasoning on the occasion of his argument for the earth to be at rest at the center of the universe .
anaximander claimed that since space is isotropic there is no reason for the earth to move in any direction .
then , it does not move .
in other words , a symmetry should not be broken if there is not a sufficient reason for breaking it ( see mckirahan 1994 on anaximander s view ) .
the principle was later adopted by parmenides , leucippus and other ancient philosophers . in the middle ages , aquinas argued from the psr to reject an infinite regress : he held that there must be some reason for the whole chain of causes .
spinoza stated the principle in his famous major work , the _ ethics _ ( spinoza 1985 ) .
in e1p11d2 , we read : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ for each thing there must be assigned a cause , or reason , both for its existence and for its nonexistence . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ for spinoza not only there must be a reason for what there is , but also for what there is not .
this seems to be a particularly strong version of the principle .
leibniz introduced the expression ` principle of sufficient reason ' and he is its best known exponent and defender . in the _ monadology _ , sec .
32 , he wrote : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ there can be no fact real or existing , no statement true , unless there be a sufficient reason , why it should be so and not otherwise , although these reasons usually can not be known by us .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ and in his second letter to samuel clarke , he simplifies ( leibniz and clarke 2000 ) : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the principle of sufficient reason , namely , that nothing happens without a reason .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ this is not far from the only extant fragment of leucippus ( taylor 1999 ) : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ nothing happens in vain , but everything from reason and necessity . _
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the psr was under attack in the xviii century by the empiricists , especially david hume .
hume critique of causality can be easily extended to sufficient reason .
logical positivists and modern analytic philosophers have also distrusted of the psr , in part because of its alleged theological implications and in part because its dubious nomological status ( see pruss 2010 for a broad review of contemporary criticisms ) . in this brief paper
i want to clarify the role of the psr in science .
i maintain that this principle , properly understood , plays an important role in scientific research .
far from being an obscure tinge from an outdated rationalism in search of theological justifications , i submit that the psr is a fundamental working hypothesis in the toolkit of any research scientist .
but before discussing epistemological issues related to the principle , we should recognise that under the single acronym ` psr ' there are a variety of statements of different import and strength , as well as some vagueness that must be dispelled . let s see
in his book devoted to the psr , alexander pruss ( 2010 ) collects the several ways in which the principle has been enunciated ( see also oppy 2009 for a rational reconstruction ) .
i distinguish four different main forms of the principle ( pruss 2010 ) : * psr 1 . everything that is the case must have a reason why it is the case . + * psr 2 .
necessarily every true proposition has an explanation . + * prs 3 .
every event has a cause . + * psr 4 . _
ex nihilo nihil fit _
( nothing comes from nothing ) .
+ these statements are certainly not equivalent .
they have presumably different strength and meaning .
but in order to compare them and their import , vagueness should be removed from some terms that appear in the statements .
words such as ` reason ' , ` explanation ' , ` cause ' , and ` nothing ' should be carefully defined in the present context before we can discuss the principle and its place in science . in the next section
i proffer elucidations of these terms .
the word ` reason ' is polysemous .
i differentiate two main meanings ( e.g. bunge 2003 ) : ( 1 ) a mental faculty consisting in thinking in a cogent way , and ( 2 ) an ontological justification of the occurrence of an event or a state of affairs .
psr 1 does not refer to properties of the brain , so we better try to refine ( 2 ) so as to make of psr 1 a meaningful statement . an ontological justification for events and states of affairs might be the specification of a sufficient system of causes .
in such a case psr 1 @xmath0 psr 3 .
i discuss causes and psr 3 below .
but there is another possible meaning of ` ontological justification ' : the specification of a system of laws and facts such that given a number of conditions a , then the event or state of affairs b follows .
for instance , the specification of the law of gravitation and the masses of all objects in the solar system , plus some adequate initial conditions , justify the state of motion of the earth with respect to the sun . in this sense
, we can say that there is a ` reason ' for the earth motion around the sun .
i call this type of justification ` nomological justification ' . under this interpretation ,
psr 1 is not a law , but a metanomological statement ( bunge 1961 , 1967 ) .
since laws can be understood as constant relations among properties of things , psr 1 would be tantamount to * psr 1a .
all events are lawful .
let us now turn to psr 2 .
this version refers to propositions , i.e. classes of statements , and not to the world . as it has been enunciated , it is a statement about our uses of language .
since explanation is an epistemic operation and not a semantic one , in its current form psr 2 is meaningless .
it can be minimally modified , nevertheless , to become a meaningful statement , namely : * psr 2a .
necessarily every true proposition satisfies truth conditions .
this is trivially true , but says nothing about the world . since the proponents of the psr think that they are saying something about the way the world actually is , we can attempt a different approach replacing propositions by what they represent : facts . we obtain : * psr 2b .
necessarily every fact has an explanation .
i should say a bit more here about explanation if we are going to make any sense of this statement . to explain a fact is to show how it happens .
this ` to show ' means to make explicit a mechanism that produces the fact ( see bunge 2006 ) .
a mechanism is a system of processes that occur lawfully .
so , psr 2b can be rendered into : * psr 2c .
every fact results from lawful processes .
since there is no need for the word ` necessary ' given the unrestricted universal quantification , i dropped it in this reformulation .
psr 2c is very similar to psr 1a .
if we define facts as either events ( changes in the state of things ) or states , and admit that events are related to either previous states or other events , then both statements have the same import .
i turn now to psr 4 . in the formulation given above , psr 4
is defective since ` nothing ' is not a thing but a concept : an empty domain of quantification .
i propose the following reformulations of psr 4 : * psr 4a .
there are not bare facts .
psr 4a means that all facts are part of a system of facts , the world , where no event occurs isolated .
although this implies a nomic determinism , it is certainly _ not _ a causal determinism , as the one required by psr 3 . using unrestricted quantification . ]
we can rewrite psr 4a as : * psr 4b .
every fact results lawfully from previous facts .
this form is quite similar to psr 2c . in order to discuss psr 3
it is convenient to say a few words about causality before .
causation is a mode of event generation ( bunge 1979 , romero and prez 2013 ) .
it is not the only way of generating events .
particle decays , as those of elementary particles , such as muons and taus , generate events without causal origination : the existence of no previous event is necessary for the occurrence of these processes . the same is valid for the decay of composed particles such as the various mesons , the neutron , or any spontaneous decay , from radioactive nuclei to the implosion of a star .
the events of decay are legal ( they occur in conformity to the probabilistic laws of quantum interactions , or other laws or even complexes of laws ) , but these events and processes are not causal : there is no prior event that we can call the ` cause ' of these occurrences .
i suggest the following definition of causal interaction between events : two events @xmath1 and @xmath2 are causally related iff there is at least a process @xmath3 such that @xmath2 is a component of @xmath3 and @xmath1 is a component as well , and it is never the case that @xmath1 is not a component of @xmath3 .
then i say that @xmath1 is a cause of @xmath2 .
the event @xmath2 is an effect of @xmath1 . in symbols :
@xmath4 the process @xmath3 involving @xmath2 can never occur without the existence of @xmath1 .
the world is legal and determinate , but not strictly causal .
there are events that are not causally related and processes that are not causally originated . if these is correct , then psr 3 is a false statement . from our analysis of different proposals for the formulation of the psr we conclude that , once all the terms have been conveniently defined , the different statements collapse into the following one : + * psr@xmath5 .
every fact results lawfully from previous facts .
+ in this formulation ` fact ' means ` a state or a change of state of a thing ' .
an event is a change of state , so a fact is either a state or an event in a thing .
i propose psr@xmath5 as the only version of the psr that is compatible with modern science . before discussing the ontological and epistemological status of psr@xmath5
, i will briefly comment on the system of all things , to which the principle is applied .
the psr is equated sometimes to the statement , likely inspire by hegel , that `` reality is rational '' .
this can adopt occasionally the form ( 1 ) `` the world is rational '' or ( 2 ) `` the universe is reasonable '' .
i submit that all these sentences are nonsense .
reality is a concept : the set of all real entities . as all sets
, it lacks of independent existence , it is a fiction , albeit a convenient one . the word ` rational ' , to the contrary , qualifies a type of behaviour : the one that is guided by reason , i.e. by cogent thinking .
sets do not think , so reality can not be rational , hence ( 1 ) makes no sense .
the world , on the other hand , can be understood as the system of all events ; the universe , as the system of all things ( see romero 2013 for a detailed characterisation ) .
both world and universe are concrete entities , but the faculty of thinking , and of thinking reasonably and rationally , is not among their known properties . to the best of our current knowledge ,
only beings endowed with brains of notable plasticity are able to think .
it is difficult to understand what would mean for the universe to think , and even more difficult for the world ( since the world is changeless
romero 2013 ) .
at best , sentences ( 1 ) and ( 2 ) are false statements . in the worst case , they are not even statements .
perhaps what is meant by this type of talk about the world is that it is comprehensible for us , humans .
this , in turn , means that we can produce conceptual representations of all aspects of the world .
although we can assume we can do that as a guiding methodological principle for our research ( `` there are not forbidden topics '' ) , there is no certainty , i think , that we will ever be able to develop the conceptual tools for a full representation of the world and the means to test such representations ( see rescher 1999 ) .
this should not be a hindrance to our attempts at deepening our understanding of reality .
it is in this enterprise where the psr becomes prominent .
i propose that the correct enunciation of the psr is psr@xmath5 : every fact results lawfully from previous facts .
this is a general statement about facts and laws .
it is a statement neither necessary nor obviously true . since it claims that laws cover the whole range of facts , it is a metanomological statement .
it is a condition upon law statements : they ought to cover all the realm of reality .
the epistemological status of such a statement is methodological : it is guiding principle for generating knowledge . in every situation where apparent brute facts seem to appear ,
the psr@xmath5 recommends the search for deeper laws .
any working scientist adopts this principle when an apparent inconsistency appears in the data at hand . instead of simply assuming brute facts , the responsible scientist proposes a revision of the data or , as a last resource , a modification of the accepted ontology .
for instance , the non - conservation of energy , momentum , and spin in some particle decays led the physicist wolfgang pauli to postulate the existence of the neutrino in 1930 .
recently , the apparent violation of special relativity in neutrino experiments led some scientists to speculate about some exotic explanations and , ultimately , to find a mistake in the interpretation of the experimental data due to some systematics not originally taken into account . in these an many other instances of scientific inquiry
the researchers are guided by the non - explicit assumption of psr@xmath5 : there must be a lawful explanation of each experimental or observational situation . not being the psr a necessary truth , the theological scruples of some philosophers are groundless .
the principle reflects our disposition to solve problems in science , but can not be used to make direct predictions .
it is too a general statement for that .
predictions can be made from law statements plus a set of conditions obtained from information about particular states of affairs .
we can not infer the existence of something , e.g. the neutrino , from the psr alone .
the actual process is that we _
propose _ the existence of the neutrino to satisfy a well tested law ( e.g. momentum conservation ) .
we are _ motivated _ by the psr to demand a fully lawful situation .
ultimately , it will be the experiment that will confirm whether the neutrino exists or not .
it is important to notice that quantum transitions and other intrinsically probabilistic phenomena are not violations of the psr@xmath5 and _ do not require _ any special interpretation of quantum mechanics .
transitions and decays occur in perfect agreement with the law statements of quantum mechanics . actually ,
the probabilistic predictions of quantum mechanics are extraordinarily well corroborated , to the point that most of our modern technology is based on them .
i can not think of a worse attempt to rebut the psr@xmath5 than invoking quantum mechanics .
amazingly , some philosophers have tried to do it ... in papers written with computers that operate in accordance with the laws of quantum mechanics .
the principle of sufficient reason is a metanomological statement that provides a useful guide in the pursuit of scientific knowledge .
it is not a law of nature nor a necessary statement .
it is , nonetheless , used by scientist in their everyday work , and has been assumed in many of the most important discoveries of science .
this is reason enough for a principle of sufficient reason to be well coveted into the toolbox of any serious researcher .
i thank mario bunge and daniela prez for stimulating discussions . 9 m. bunge ( 1961 ) .
laws of physical laws , _ american journal of physics _ * 29 * , 518 - 529 . m. bunge ( 1967 ) .
_ foundations of physics_. new york : springer - verlag .
m. bunge ( 1977 ) .
_ treatise of basic philosophy .
ontology i : the furniture of the world_. dordrecht : reidel .
m. bunge ( 1979 ) . _
causality in modern science_. new york : dover , 2nd ed .. m. bunge ( 2003 ) . _
philosophical dictionary_. prometheus books , 2nd ed .. m. bunge ( 2006 ) .
_ chasing reality : strife over realism_. toronto : university of toronto press
leibniz ( 2005 ) . _ discourse on metaphysics and the monadology_. mineola , ny : dover publications .
leibniz and s. clarke ( 2000 ) .
_ correspondence _
, edited by r. ariew .
indianapolis : hackett publishing co .. r.d .
mckirahan ( 1994 ) . _ philosophy before socrates_. indianapolis : hackett publishing co .. g. oppy ( 2009 ) .
_ philosophical perspectives on infinity_. cambridge : cambridge university press .
a. r. pruss ( 2010 ) . _
the principle of sufficient reason : a reassessment_. cambridge : cambridge university press .
n. rescher ( 1999 ) . _ the limits of science_. pittsburgh : university of pittsburgh press ; revised edition .
romero ( 2013 ) .
from change to spacetime : an eleatic journey , _ foundations of science _ , * 18 * , 139 - 148 .
g.e . romero and d. prez ( 2012 ) .
new remarks on the cosmological argument , _ int .
j. philos .
_ , * 72 * , 103 - 113 .
a. schopenhauer ( 2012 ) .
_ on the fourfold root of the principle of sufficient reason and other writings : volume 4_. cambridge : cambridge university press .
b. spinoza ( 1985 ) . _ the collected works of spinoza ( volume 1 ) _ , translated and edited by edwin curley , princeton : princeton university press .
taylor ( 1999 ) . _ the atomists : leucippus and democritus . fragments .
_ toronto : university of toronto press .
full professor of relativistic astrophysics at the university of la plata and chief researcher of the national research council of argentina . a former president of the argentine astronomical society
, he has published more than 300 papers on astrophysics , gravitation and the foundations of physics .
romero has authored or edited 9 books ( including _ introduction to black hole astrophysics _ , with g.s .
vila , springer , 2013 ) .
his main current interest is on black hole physics and ontological problems of space - time theories . | i offer an analysis of the principle of sufficient reason and its relevancy for the scientific endeavour .
i submit that the world is not , and can not be , rational only some brained beings are .
the principle of sufficient reason is not a necessary truth nor a physical law .
it is just a guiding metanomological hypothesis justified _ a posteriori _ by its success in helping us to unveil the mechanisms that operate in nature .
example.eps gsave newpath 20 20 moveto 20 220 lineto 220 220 lineto 220 20 lineto closepath 2 setlinewidth gsave .4 setgray fill grestore stroke grestore _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ what necessity would have stirred it up to grow later rather than earlier , beginning from nothing? ... nor will the force of conviction ever permit anything to come to be from what is not .
+ _ parmenides . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ |
in 1985 mikhail gromov proved his famous non - squeezing theorem saying that a ball @xmath10 of radius @xmath4 , in a symplectic vector space @xmath11 can not be symplectically embedded into @xmath12 unless @xmath13 ( both sets are equipped with the usual symplectic structure induced from @xmath14 on @xmath11 ) .
this motivated the definition of the invariant called the gromov width .
consider the ball of capacity @xmath15 @xmath16 with the standard symplectic form @xmath14 .
the * gromov width * of a @xmath17-dimensional symplectic manifold @xmath18 is the supremum of the set of @xmath15 s such that @xmath19 can be symplectically embedded in @xmath18 .
it follows from darboux theorem that the gromov width is positive unless @xmath20 is a point .
let @xmath2 be an integer greater then three and let @xmath21 be positive real numbers .
the * polygon space * @xmath1 is the space of closed piecewise linear paths in @xmath3 such that the @xmath22-th step has norm @xmath23 , modulo rigid motions .
these moduli spaces are also the symplectic reduction of the product of @xmath2 spheres , of radii @xmath21 , by the diagonal action of @xmath24 , @xmath25 the length vector @xmath4 is * generic * if and only if the scalar quantity @xmath26 is not zero for any @xmath27 in this case , the polygon space @xmath1 is a smooth symplectic ( in fact , khler ) manifold of dimension @xmath28 . observe that for any permutation @xmath29 manifolds @xmath1 and @xmath30 are symplectomorphic .
note that the existence of an index set @xmath31 such that @xmath32 is equivalent to the existence of an element @xmath33 in @xmath34 that lies completely on a line .
the stabilizer of such an element is non - trivial : it is the @xmath35 of rotations along that line . therefore the associated symplectic reduction has a singularity . an index set @xmath31 is called * short * if @xmath36 , and long of its complement is short . moreover @xmath31 is * maximal short * if it is short and is not contained in any other short set . from an algebro - geometric point of view
, polygon spaces are identified with the git quotient of @xmath37 by @xmath38 .
this git quotient is a compactification of the configuration space of @xmath2 points in @xmath39 and , via the gelfand
macpherson correspondence , relates polygon spaces to the symplectic reductions of the grassmannian of @xmath40-planes in @xmath41 by the maximal torus @xmath42 of the unitary group @xmath43 . in this work
we analyze the gromov width of polygon spaces @xmath1 for @xmath44 generic , @xmath45 , and prove the following results .
[ theo:5gon ] the gromov width of @xmath46 is equal to @xmath47 [ theo:6gon ] the gromov width of @xmath48 is at least @xmath49 moreover let @xmath50 be such that @xmath51 , if one of the following holds : * @xmath52 and @xmath53 are short for @xmath54 , or * @xmath53 and @xmath55 are long for @xmath54 , or * @xmath56 and @xmath57 are short for @xmath54 then the gromov width of @xmath48 is equal to @xmath49 [ theo : projective ] assume that there exists a maximal @xmath4-short index set @xmath58 . in this case , @xmath1 is symplectomorphic to @xmath59 where @xmath60 denotes the usual fubini - study symplectic structure and its gromov width is @xmath61 in this case , as @xmath58 is maximal short , the above value is @xmath62 .
we conjecture that the gromov width of polygon spaces , for any @xmath45 and any @xmath4 generic is @xmath63 note that as @xmath1 and @xmath30 are symplectomorphic for all @xmath29 , we can always assume that @xmath64 . with this assumption @xmath65 @xmath66
an important tool in the proof of the above results is a toric action , called the bending action , defined on a dense open subset of @xmath1 ( possibly on the whole @xmath1 ) .
let @xmath67 be a choice of a diagonal of the polygons in @xmath1 .
the circle action associated to @xmath68 rotates the piecewise linear path @xmath69 along the axis of the diagonal @xmath68 .
this action is defined the dense open subset of @xmath1 consisting of polygons @xmath70 for which the diagonal @xmath68 does not vanish . in this way
any system of @xmath71 non intersecting diagonals gives a toric action of @xmath72 on a dense open subset of @xmath1 ( where the respective diagonals do not vanish ) . for many @xmath4 s and for appropriate choices of diagonals
the action can be defined on the whole @xmath1 . using the flow of this action one can construct symplectic embeddings of balls and thus obtain lower bounds for the gromov width .
the bending action has a central role also in determining upper bounds for the gromov width , as certain tools ( for example , @xcite ) are available for toric manifolds which are fano , or blow ups of fano toric manifolds at toric fixed points . using a moser - type continuity argument
we obtain upper bounds for the gromov width of some non - toric spaces @xmath1 .
the upper bounds coincide with the lower bounds we have determined by embedding techniques , and so we determine an explicit formula for the gromov width of ( some ) polygon spaces .
one should mention here that there are efficient methods for finding the gromov width ( and for solving the more general problem of ball packings ) of @xmath73-dimensional manifolds that do not require the use of toric geometry .
in particular , the gromov width of the spaces of @xmath6-gons could also be found using propositions 1.9 , 1.10 and remark 1.11 of @xcite by mcduff . these methods , however , are specific for dimension @xmath73 . therefore , instead of using these tools , we use some tools from toric geometry as those can be applied in any dimension . * organization .
* we start with describing the tools for finding the gromov width in section [ section gromov width ] . in section
[ section polygon spaces ] we carefully define the polygon spaces and various toric actions on them ( or on their subsets ) . in section
[ section projective ] we prove the gromov with of these @xmath1 which are symplectomorphic to a projective space .
sections [ section 5gons ] and [ section 6gons ] are devoted to the gromov width of spaces of @xmath6-gons and @xmath7-gons , respectively .
* acknowledgements : * the authors would like to thank tara holm , yael karshon , dominic joyce , dusa mcduff , felix schlenk and kazushi ueda for helpful discussions .
the research leading to these results has received funding from the european research council under the european union s seventh framework programme ( fp7/2007 - 2013 ) / erc grant agreement no .
the second author was supported by the fundao para a cincia e a tecnologia ( fct , portugal ) : fellowship sfrh / bpd/87791/2012 and projects ptdc / mat/117762/2010 , excl / mat - geo/0222/2012 .
we start with describing techniques for finding a lower bound for the gromov width .
if the manifold @xmath18 is equipped with a hamiltonian ( so effective ) action of a torus @xmath74 , one can use this action to construct explicit embeddings of balls and therefore to calculate the lower bound for the gromov width .
such construction was provided by karshon and tolman in @xcite .
if additionally the action is * toric * , that is , @xmath75 , then more constructions are available ( see for example : @xcite , @xcite,@xcite ) . in what follows we use results of latschev , mcduff and schlenk , @xcite , presented here as proposition [ prop diamond ] and proposition [ prop diamondlike ] .
as we are to calculate a numerical invariant , we need to fix a way of identifying the lie algebra of @xmath76 with the real line @xmath77 .
we think of the circle as @xmath78 . with this convention
the moment map for the standard @xmath76-action on @xmath79 by rotation with speed @xmath80 is given ( up to an addition of a constant ) by @xmath81 .
define @xmath82 when the dimension is understood from the context we simply write @xmath83 if @xmath84 is toric , @xmath85 is the associated moment map and @xmath86 is a subset of the interior of the moment map image , then a subset of @xmath87 is symplectomorphic to @xmath88 with the symplectic structure induced from the standard one on @xmath89 .
below we present a result of latschev , mcduff and schlenk , ( see ( * ? ? ?
* lemma 4.1 ) ) which , though stated in dimension @xmath73 , holds also in higher dimensions .
note that the authors are using the convention where @xmath90 and therefore the proposition below looks differently than ( * ? ? ?
* lemma 4.1 ) . to translate the conventions note that @xmath91 is symplectomorphic to @xmath92 .
* lemma 4.1 ) [ prop diamond ] for each @xmath93 the ball @xmath94 of capacity @xmath95 symplectically embeds into @xmath96 . therefore ,
if @xmath97 for a toric manifold @xmath98 with moment map @xmath85 , then the gromov width of @xmath99 is at least @xmath100 .
a more general result is true .
let @xmath101 be real numbers such that @xmath102 , @xmath103 .
we build a , not necessarily symmetric , cross whose arms are open intervals of length @xmath15 and take the convex hull of it .
this way we obtain a `` diamond - like '' open subset @xmath104 .
@xmath105 , @xmath106.,width=151 ] ( * ? ? ?
* section 4.2 ) [ prop diamondlike ] for each @xmath93 the ball @xmath94 of capacity @xmath95 symplectically embeds into @xmath107 .
therefore , if @xmath108 for a toric manifold @xmath98 with moment map @xmath85 , then the gromov width of @xmath109 is at least @xmath100 . note that a simplex is a particular case of a diamond - like shape .
therefore the above proposition also implies the following result ( * ? ? ?
* proposition 1.3 ) ( * ? ? ?
* proposition 2.5 ) ( * ? ? ?
* lemma 5.3.1 ) let @xmath110 be the @xmath2-dimensional simplex .
for any connected proper ( not necessarily compact ) hamiltonian @xmath111 space @xmath20 let @xmath112 where @xmath113 is some choice of moment map .
the lower bound for gromov width of @xmath20 is @xmath114 .
note that @xcite , @xcite and @xcite use different identification of @xmath115 with @xmath116 than @xcite and we here .
it was already observed by gromov that one can use j - holomorphic curves to find upper bounds for the gromov width ( see proposition [ thm ub jholom ] ) .
many tools for finding upper bounds are based on a similar idea : non - vanishing of a certain gromov - witten type of invariant implies some upper bound for the gromov width .
we start this section with explaining the above observation in more details .
later we recall some tools for finding upper bounds for the gromov width constructed by lu ( @xcite ) , making use of a toric action . here
we set the essential definitions and notations to be able to use pseudoholomorphic curves and gromov - witten invariants to compute the upper bound of the gromov width .
we refer to mcduff
salamon @xcite for a comprehensive exposition of the subject , and to caviedes @xcite , zoghi @xcite , karshon - tolman @xcite where these techniques are used to determine the gromov width of certain coadjoint orbits . an almost complex structure on a symplectic manifold @xmath99 is a smooth operator @xmath117 such that @xmath118 .
a almost complex structure @xmath119 is @xmath120-compatible if @xmath121 defines a riemannian metric .
denote by @xmath122 the space of all @xmath120-compatible almost complex structures .
let @xmath123 be the riemann sphere equipped with its standard complex structure @xmath22 and let @xmath119 be a @xmath120-compatible almost complex structures on @xmath20 .
a * @xmath119-holomorphic curve * is a map @xmath124
satisfying @xmath125 . an important feature of @xmath119-holomorphic curves is that they come in families , which combine together in a moduli space as follows . given a homology class @xmath126 , let @xmath127 denote the moduli space of simple @xmath119-holomorphic curves : @xmath128=a , u \text { is simple } \}.\ ] ] consider the evaluation map @xmath129 the group @xmath130 acts on naturally on @xmath39 and by reparametrization on @xmath131 .
the evaluation map descends to the quotient and we obtain the map @xmath132 the following result ( quoted here from the work of caviedes @xcite ) explains how j - holomorphic curves can be used to obtain the upper bounds for the gromov width .
the idea goes back to gromov and was used by him to prove his famous non - squeezing theorem .
the proof can be found , for example , in @xcite .
* theorem 2.3)[thm ub jholom ] let @xmath99 be a compact symplectic manifold .
given @xmath133 , if for a dense open subset of @xmath120-compatible almost complex structures @xmath119 the evaluation map @xmath134 is onto , then for any symplectic embedding @xmath135 one has @xmath136 where @xmath137 is the symplectic area of @xmath138 . in particular
, it follows that the gromov width of @xmath99 is at most @xmath137 . one way to prove that the evaluation map is onto is via the gromov - witten invariants .
fix @xmath133 , and let @xmath139 be an element of @xmath140 where @xmath141 + 2k-6)=2 \dim m^k.\ ] ] the gromov - witten invariant @xmath142 counts " the number of j - holomorphic curves @xmath143 in the homology class @xmath138 which meet each of the cycles @xmath144 . the precise definition of the gromov - witten invariant involves some delicate and technical tools that go beyond what is needed for the purpose of this work ,
so we refer the reader to @xcite .
we want to stress that gromov
witten invariants are symplectic invariants and independent of the choice of a ( generic ) almost complex structure @xmath145 .
let @xmath146 $ ] denote the poincar dual of the fundamental class of a point .
if @xmath147,\alpha_2,\ldots , \alpha_k ) \neq 0 $ ] for some classes @xmath148 then the evaluation map is onto and we can apply the above theorem . in section [ section 6gons ] we will apply this method to @xmath7-dimensional ( so semipositive ) symplectic manifolds @xmath18 with chosen homology classes @xmath138 for which @xmath149=2 $ ]
. then one can take @xmath150 and consider @xmath147 ) \in { \mathbb{z}}$ ] .
when @xmath147)\neq 0 $ ] the above theorem implies that the gromov width of @xmath18 is not greater than @xmath137 .
let @xmath151 be a delzant polytope with primitive inward facets normals @xmath152 and let @xmath153 be the smooth toric symplectic manifold corresponding to it .
let @xmath154 be the fan associated to @xmath155 , and let @xmath156 denote the generators of the @xmath80-dimensional cones of @xmath157 .
a well - known construction in algebraic geometry assigns to @xmath157 a toric variety @xmath158 ( no symplectic structure yet ) . here
@xmath158 is compact and smooth because @xmath157 is smooth and its support is the whole @xmath159 .
moreover , our @xmath158 is projective and therefore there is a one - to - one correspondence between khler forms on @xmath158 and strictly convex support functions @xmath160 on @xmath157 . recall that a piecewise linear function @xmath160 on @xmath157 is called a strictly convex support function for @xmath157 if * it is upper convex , i.e. , @xmath161 for all @xmath162 , and * the restrictions of it to any two different @xmath2-dimensional cones @xmath163 ,
@xmath164 are two different linear functions , ( see ( * ? ? ?
* section 2 ) ) . given a support function @xmath160 on @xmath157 the symplectic toric manifold @xmath165 has moment map image @xmath166 defined by inequalities @xmath167 for all @xmath168 .
therefore , the symplectic toric manifold , @xmath153 , obtained from @xmath155 via delzant construction is @xmath165 where @xmath169 .
to the pair @xmath170 lu associates @xmath171 and use it to describe an upper bound for gromov width of toric fano manifolds .
a toric manifold is * fano * if the anticanonical divisor is ample .
we refer the reader to , for example , @xcite or @xcite , for more information about fano varieties . here
we only mention the properties that will be relevant to our results . for compact symplectic toric manifolds
one can determine whether it is fano by looking at the moment map image .
as the property of being fano is a property of the underlying toric variety , not of the symplectic structure , it is enough to analyze the fan associated to the moment map image .
a compact symplectic toric manifold @xmath84 , with associated fan @xmath157 , is fano if and only if there exists a monotone polytope @xmath172 ( vectors @xmath173 are primitive inward normals to the facets of @xmath174 ) , whose fan is also @xmath157 .
this follows from theorem 8.3.4 of @xcite .
another way to see that is by observing that the dual @xmath175 is exactly equal to the convex hull of points @xmath176 and applying proposition 3.6.7 of @xcite .
in particular all monotone compact symplectic toric manifolds are fano .
we now quote a result of lu which we will use for finding the upper bounds of gromov width .
* theorem 1.2)[theorem lufanoupperbounds ] if @xmath153 is fano then the gromov width of @xmath153 is at most @xmath177 as we will see later , in the case of polygon spaces the expressions that may appear in the above set are @xmath178 and @xmath179
. of course not all polygon spaces are toric and fano .
some of the not fano ones can be obtained from some toric fano manifold by a sequence of toric blow ups . in these situations
we can apply another theorem of lu .
* theorem 6.2)[theorem lublowupupperbounds ] let @xmath180 be a toric manifold obtained from a toric fano manifold @xmath158 by a sequence of blow ups at toric fixed points . then the generators of @xmath80-dim cones of associated fans satisfy @xmath181 .
moreover any strictly convex support function @xmath160 for @xmath157 is also strictly convex for @xmath182 and it holds that the gromov width of @xmath183 is not greater than @xmath184 . note the typo in @xcite : there is an extra @xmath185 appearing in his formulation of the above theorem .
the moduli space @xmath1 , @xmath44 , of closed spatial polygons is the space of closed piecewise linear paths in @xmath3 with the @xmath22-th step of length @xmath23 , modulo rigid motions in @xmath3 ( i.e. rotations and translations ) .
the space @xmath1 inherits a symplectic structure by means of symplectic reduction , as we describe below . for any choice of @xmath2 positive real numbers
@xmath186 let @xmath187 be the sphere in @xmath3 of radius @xmath188 and center the origin , equipped with the symplectic volume form .
the product @xmath189 where @xmath190 is the projection on the @xmath191-th factor , is a compact smooth symplectic manifold .
the group @xmath24 acts diagonally on @xmath192 via the coadjoint action ( thinking of each sphere @xmath193 as of a @xmath24-coadjoint orbit ) .
this action is hamiltonian with moment map @xmath194 the symplectic quotient @xmath195 is the space of @xmath2-gons of fixed sides length @xmath196 modulo rigid motions , and is usually called polygon space .
when it generates no confusion we will use the name polygon for both : an element in @xmath197 and its class in @xmath1 . note that
if @xmath198 then the closing condition can not be satisfied , if @xmath199 then @xmath1 is either empty or a point , depending on whether @xmath200 , and if @xmath201 then @xmath1 is either empty or a point , depending on whether @xmath202 satisfy a triangle inequality . in our study of the gromov width of polygon spaces we omit these trivial cases and assume that @xmath45 .
a polygon is _ degenerate _ if it lies completely on a line .
the moduli space @xmath1 is a smooth manifold if and only if the lengths vector @xmath4 is _ generic _ , i.e. for each @xmath203 the quantity @xmath26 is non - zero .
equivalently , @xmath4 is generic if and only if in @xmath1 there are no degenerate polygons .
in fact , if there exists a polygon @xmath70 on a line ( or an index set @xmath31 such that @xmath204 ) then its stabilizer is @xmath35 since the polygon @xmath70 is fixed by rotations around the axis it defines .
therefore the @xmath24-action on @xmath205 is not free and the quotient , @xmath206 , has singularities .
note that , for @xmath4 generic , the polygon space @xmath1 inherits a symplectic form by symplectic reduction .
observe moreover that for any permutation @xmath29 , the manifolds @xmath1 and @xmath30 are symplectomorphic .
for any @xmath4 generic , an index set @xmath31 is said to be * short * if @xmath36 , and * long * if @xmath207 , i.e. if its complement is short .
an index set @xmath31 is * maximal short * if it is short and maximal with respect to the inclusion on the collection of short sets for @xmath4 , i.e. any index set containing @xmath31 as a non - trivial subset is long . in @xcite ,
hausmann and knutson prove that polygon spaces are also realized as symplectic quotients of the grassmannians @xmath208 of @xmath40-planes in @xmath41 , obtaining the gelfand
macpherson s correspondence .
the construction goes as follows .
let @xmath42 be the maximal torus of diagonal matrices in the unitary group @xmath43 and consider the action by conjugation of @xmath209 on @xmath210 .
as the diagonal circle @xmath211 acts trivially , let us consider the effective action of @xmath212 on @xmath210 .
let @xmath213 , with @xmath214 , denote the coordinates in @xmath210 .
the hamiltonian action of @xmath215 on @xmath210 @xmath216= ( a^{-1 } q_1 e^{i\theta_1 } , \ldots , a^{-1 } q_n e^{i\theta_n}),\ ] ] with @xmath217 , has moment map @xmath218 where @xmath219 denotes the traceless part : @xmath220 the polygon space @xmath1 is then symplectomorphic to the symplectic reduction of @xmath210 by @xmath215 : @xmath221 ( cf @xcite ) . in fact , performing the reduction in stages and taking first the quotient by @xmath42 at the @xmath4-level set , one obtains the product of spheres @xmath192 .
the residual @xmath222 action is the coadjoint action described above , and one recovers the polygon space @xmath223 as the symplectic quotient @xmath224 performing the reduction in stages in the opposite order , one obtains the gelfand
macpherson correspondence .
in fact , one first obtains the grassmannian @xmath208 of complex planes in @xmath41 as the reduction @xmath225 .
then the quotient of @xmath208 by the residual @xmath226 action is isomorphic to the moduli space of @xmath2 points in @xmath39 , cf .
@xcite , and is also isomorphic to the polygon space @xmath223 , see @xcite .
this is summarized in the following diagram : @xmath227_{u(2 ) } \ar[dr]^{u(1)^n } & { } \\
gr(2,n ) \ar[dr]_{u(1)^n / u(1 ) } & { } & \prod_{j=1}^n s_{r_j}^2 \ar[dl]^{u(2)/u(1 ) \simeq so(3 ) } \\ { } & m(r ) & { } } \ ] ] the chambers of regular values in the moment polytope @xmath228 are separated by walls @xmath229 of critical values . note that an index set @xmath31 and its complement @xmath230 determine the same wall , and if @xmath31 has cardinality @xmath80 or @xmath231 , then the associated wall @xmath232 is an external wall .
a chamber @xmath233 is called external if it contains in its closure an external wall .
in particular , if @xmath4 is in an external chamber , then there is a maximal short index set @xmath31 of cardinality one . in this case
the polygon space @xmath223 is diffeomorphic to the projective space @xmath9 , @xcite .
let @xmath234 be generic .
for any polygon @xmath70 in @xmath1 of edges @xmath235 and verteces @xmath236 , consider the system of @xmath237 non - intersecting diagonals @xmath238 from the first vertex to the remaining non - adjacent vertices , i.e. @xmath239 . following nohara
ueda @xcite , we call this system of diagonals * caterpillar system*. the lengths of the @xmath237 diagonals @xmath240 are continuous functions on @xmath1 and are smooth on the subset where they are not zero . their image is a convex polytope in @xmath241 , which we denote by @xmath155 , consisting of points @xmath242 that satisfy the following triangle inequalities @xmath243 where @xmath244 and we use the notation @xmath245 , @xmath246 .
the functions @xmath247 give rise to hamiltonian flows , called bending flows , ( cf @xcite ) . the circle action associated with a given diagonal @xmath248
is defined on the dense open subset @xmath249 in the following way .
the first @xmath250 edges bend along the diagonal @xmath248 at a constant speed while the remaining edges do not move .
putting together the actions coming from @xmath71 non - intersecting diagonals we obtain a toric action of @xmath72 on the dense open subset @xmath251 .
the action - angle coordinates are given by the lengths @xmath252 and the angles of rotation respectively .
if each @xmath247 does not vanish on @xmath1 then @xmath1 is a symplectic toric manifold . in this case , the moment map is given by and the moment polytope is @xmath155 described by inequalities .
the choice of another system of @xmath237 non - intersecting diagonals gives rise to a different hamiltonian system on an open dense subset of @xmath223 ( possibly the whole @xmath1 ) .
these actions were investigated in @xcite , where it is shown that any bending action on polygon spaces is induced by an integrable system of gelfand - cetlin type on the grassmannian @xmath208 .
in this section we analyze the gromov width of @xmath1 in the cases when @xmath1 is diffeomorphic to @xmath9 , @xmath253 , and prove theorem [ theo : projective ] .
we assume that @xmath44 is generic ( so @xmath1 is a smooth manifold ) and that @xmath254 . as shown in (
* proposition 4.2 ) , @xmath255 is diffeomorphic to @xmath9 if and only if there is a maximal short set of cardinality 1 . using the assumption that the @xmath188 s are ordered non - decreasingly , this is equivalent to @xmath256 being long , i.e. , @xmath257 .
let @xmath4 be generic , ordered non - decreasingly and such that @xmath258 is long .
then the symplectic volume of @xmath1 is @xmath259 this proposition , together with ( * ? ? ?
* proposition 4.2 ) recalled above , proves theorem [ theo : projective ] : it shows that the gromov width of the above @xmath1 is @xmath260 , which in this case is exactly @xmath61 from ( * ? ? ?
* section 2.5.1 , page 210 ) , we know that the symplectic volume of the polygon space @xmath1 is given by @xmath261 where @xmath262 , @xmath263 and @xmath264 if @xmath265 and @xmath266 if @xmath267 . for the long set @xmath268
one gets a contribution to the volume of @xmath269 note that any other long set contains @xmath2 .
let us analyze the contributions to the coefficient of a generic element @xmath270 , for some @xmath271 , given by longs sets @xmath31 that are different from @xmath272 .
the index sets @xmath31 that contains @xmath2 and @xmath273 contribute by @xmath274 note that the right hand side of can be rewritten as @xmath275 similarly , long sets @xmath31 that contains @xmath2 and @xmath276 elements in @xmath277 contribute , up to a sign , by @xmath278 continuing this way we prove @xmath279 hence , the volume of @xmath1 is @xmath280 here is an alternative way of finding the gromov width in this case
. one can show that if @xmath4 is generic , ordered non - decreasingly and @xmath258 is long , then the bending action coming from the caterpillar system of diagonals is toric on @xmath1 .
the moment map image is the set @xmath155 determined by the following inequalities from : @xmath281 after appropriate translation , this set is @xmath282 -equivalent to a simplex @xmath283 , namely @xmath284 where @xmath285 @xmath286x + \left ( \begin{array}{c } r_1+r_2 \\r_1+r_2+r_3 \\ \vdots \\ r_1+\ldots + r_{n-3}\\ r_n - r_{n-1 } \end{array } \right).\ ] ] this proves that in this case the manifold @xmath1 is symplectomorphic to @xmath287 and its gromov width is @xmath288 .
let @xmath289 be generic and without loss of generality assume that the lengths are non - decreasingly ordered .
on @xmath1 consider the bending action along the diagonal @xmath290 .
the diagonal @xmath68 does not vanish if @xmath291 or @xmath292 , which is always the case by the genericity assumption .
thus the bending action is defined on the whole @xmath293 making it a toric symplectic @xmath40-dimensional manifold .
in particular hausmann and knutson in @xcite show that they are diffeomorphic to @xmath39 .
the moment map image is then the interval @xmath294 = [ \max(r_2-r_1,r_4-r_3),r_1+r_2]\ ] ] of length @xmath295 .
therefore the gromov width of @xmath1 is @xmath296 , as claimed in theorem [ theo : projective ] .
in this section we analyze the gromov width of @xmath1 for generic @xmath297 . for this purpose
we use the bending action along the caterpillar system of diagonals starting from the first vertex , as in figure [ fig : diag5 ] .
note that the caterpillar bending action on @xmath1 is toric if and only if @xmath291 and @xmath298 . since @xmath1 is symplectomorphic to @xmath299 for any permutation @xmath300 of the lengths vector , one can use this symplectomorphism to define a toric action on any @xmath1 with @xmath301 and @xmath302 for some @xmath300 .
the image of bending flow functions ( which are the moment map in the toric case ) is the polytope @xmath303 in @xmath304 given by the intersection of the rectangle of vertices @xmath305 with the non - compact region @xmath306 as in figure [ fig : catmomentpolytope ] , cf @xcite .
note that the diffeotype of @xmath1 is uniquely determined by the chamber @xmath308 or , in other words , by the collection of @xmath4-short sets , cf .
moreover , for @xmath4 in any fixed chamber @xmath309 , all toric @xmath1 have the same shape " of the moment map image with respect to the bending action along a fixed system of diagonals , i.e. the moment polytopes have the same collection of facet normals , though the lattice lengths of the edges of the polytopes may vary .
note that for a non - trivial reshuffling @xmath54 , @xmath300 , of the length vector @xmath4 , @xmath54 is in a different chamber than @xmath4 .
nevertheless @xmath1 and @xmath30 are symplectomorphic .
the bending action along the caterpillar system of diagonals on @xmath30 induces a hamiltonian system on @xmath1 which may , or may not , correspond to bending along a different system of diagonals . in section [ section 5gons upper ] , when needed , we use reshuffling of the length vector @xmath4 to obtain an action defined on the whole @xmath1 , making it a toric manifold . to determine the moment image , @xmath155
, the following chart will be useful .
the chart describes when the vertices of the rectangle @xmath310 satisfy the inequalities defining the region @xmath311 in the language of short sets . for simplicity
we assume partial ordering on the length vector @xmath4 : @xmath312 , @xmath313 .
all our reshuffled length vectors from section [ section 5gons upper ] will satisfy this partial ordering assumption .
+ [ 6gon ub toric2 ] let @xmath8 be generic , ordered non - decreasingly and such that @xmath314 is short .
assume additionally that @xmath315 and @xmath53 are long .
then the gromov width of the symplectic toric manifold @xmath1 is at most @xmath316 . reshuffle the length vector @xmath4 to @xmath317 note that @xmath54 is partially ordered and thus we can use the above chart ( applied to @xmath54 ) to analyze the set @xmath318 , which is the image of @xmath30 by @xmath319 . as
@xmath320 and @xmath52 are short , the hyperplanes @xmath321 , @xmath322 do not cut any vertex of the cuboid , and thus @xmath323 .
note that the assumption @xmath53 long implies that @xmath324 are not in @xmath325 .
the vertex @xmath326 is always in @xmath325 as we are assuming that @xmath314 is short .
depending on whether @xmath327 , @xmath80 , or @xmath40 of sets @xmath328 and @xmath329 are short , ( corresponding to @xmath327 , @xmath80 , or @xmath40 of the vertices @xmath330 being in @xmath325 ) , the set @xmath155 is a simplex with @xmath80 , @xmath40 or @xmath331 corners chopped off , respectively ( one corner is always chopped off as @xmath326 is in @xmath325 ) . more precisely : the simplex is bounded by hyperplanes @xmath332 , @xmath333 , @xmath334 and @xmath335 .
the vertex @xmath336 of this simplex is chopped off in @xmath155 by the hyperplane @xmath337 ( as @xmath338 ) .
let @xmath339 denote the above simplex with one corner chopped .
note that the vectors @xmath340 and @xmath341 are among the inward normals to the facets of @xmath339 .
the vertices @xmath342 and @xmath343 of simplex may also be chopped in @xmath155 depending on whether @xmath328 and @xmath329 are short . if @xmath344 , @xmath345 , and @xmath346 then the bending action on @xmath30 is toric and @xmath155 is the moment map image .
this implies that @xmath30 with the bending action is @xmath347 blown up at @xmath80 , @xmath40 or @xmath331 points . in other words , it is a toric fano manifold corresponding to the polytope @xmath339 , or a blow up of this manifold at @xmath80 or @xmath40 toric fixed points .
applying theorem [ theorem lufanoupperbounds ] or [ theorem lublowupupperbounds ] we get that the gromov width of @xmath30 is at most @xmath348 since @xmath30 and @xmath1 are symplectomorphic , then the gromov width of @xmath1 is also at most @xmath316 .
if at least one of @xmath344 , @xmath345 , and @xmath346 is not satisfied , then the bending action is defined only on an open dense subset of @xmath30 and the above argument does not apply .
in that situation , one can use moser s trick as in the case of @xmath6-gons described in detail in section [ section 5gons ] .
let @xmath349 for @xmath350 small , the polygon space @xmath351 with the bending action induced using the symplectomorphism @xmath352 is toric .
moreover , if @xmath315 and @xmath53 were long for @xmath1 , then @xmath315 and @xmath53 are also long for @xmath351 and the gromov width of @xmath351 is @xmath353 by proposition [ 6gon ub toric2 ] .
assume that there exists a symplectic embedding of a ball of capacity @xmath354 into @xmath1 . use moser s trick argument to show that for @xmath93 there exists an embedding of a symplectic ball of capacity @xmath355 into @xmath351 for @xmath350 small enough ( see the end of section [ section 5gons ] for details ) .
taking @xmath356 small enough so that @xmath357 we obtain a contradiction .
therefore there can not exist an embedding of a ball of capacity @xmath354 into @xmath1 .
reshuffle the length vector @xmath4 to @xmath317 note that @xmath54 is partially ordered and thus we can use the above chart ( applied to @xmath54 ) to analyze the set @xmath318 , which is the image of @xmath30 by @xmath319 . as
@xmath52 is long ( by assumption ) , then all @xmath73element sets are long , and all @xmath40element sets are short . thus each hyperplane is cutting at least one vertex .
our assumptions guarantee that each of them cuts exactly one vertex ( out of the vertices @xmath360 , @xmath326 , @xmath361 ) .
therefore @xmath155 is the cuboid @xmath362 with three non - adjacent corners chopped off . if @xmath344 , @xmath345 , and @xmath346 then the bending action on @xmath30 is toric , and @xmath30 is symplectomorphic to the blow up of the toric fano manifold @xmath363 ( corresponding to the cuboid @xmath362 ) , at three toric fixed points . applying theorem
[ theorem lublowupupperbounds ] we get that the gromov width of @xmath30 is at most @xmath348 in the cases not covered by propositions [ 6gon ub toric2 ] , [ 6gon ub 3bu ] the polygon space @xmath1 equipped with the bending action along the system of diagonals as in figure [ fig:6gonsystem ] is not obtained from a toric fano manifold by blowing up at toric fixed points , and so theorem [ theorem lublowupupperbounds ] can not be applied .
note however that all @xmath1 are obtained by a sequence of symplectic cuts from the manifold associated to the cuboid @xmath362 , which is @xmath364 , with some scaling of fubini - study symplectic forms on each @xmath39 factor .
it seems very natural to expect that the gromov width of a compact symplectic manifold would not increase under the symplectic cut operation .
this would imply that the gromov width of @xmath1 would be bounded above by the gromov width of the manifold corresponding to @xmath362 , which is @xmath316 . together with proposition [ proposition 6lowerbounds ]
that would prove that the gromov width of @xmath1 is exactly @xmath316 .
we now use a different argument to obtain the upper bound for the gromov width of @xmath7-gons , under different restrictions on the lengths @xmath188 s .
when @xmath155 contains a whole facet @xmath365 of the cuboid @xmath362 , where one of the side lengths of @xmath365 is @xmath366 , then we obtain that the gromov width of the associate polygon space @xmath1 is at most @xmath367 .
we do this by showing the non - vanishing of some gromov witten invariant , as explained below .
we are grateful to dusa mcduff for for suggesting us this approach .
suppose that the moment map image @xmath155 for the toric manifold @xmath1 contains a whole facet of the cuboid @xmath362 , where one of the side length is @xmath366 .
call this facet @xmath365 , and let @xmath368 .
note that as @xmath4 is generic , some neighborhood of @xmath365 in @xmath362 is also in @xmath155 .
therefore some neighborhood of @xmath369 in @xmath1 is symplectomorphic to a neighborhood of @xmath370 in the symplectic manifold @xmath371 corresponding to the cuboid @xmath362 .
this means we can choose a compatible almost complex structure @xmath119 on @xmath1 such that near @xmath369 @xmath119 is a product @xmath372 , where each @xmath373 is a complex structure on the respective copy of @xmath39 .
let @xmath374 be the homology class corresponding to the preimage ( under the moment map @xmath319 ) of the edge of length @xmath366 .
note that @xmath375=2,\ ] ] where @xmath376 denotes the chern class of the tangent bundle of @xmath1
. therefore the gromov - witten invariant , @xmath377)$ ] , associated to the homology class @xmath138 and evaluated on the poincar dual to the fundamental class of a point , is an element of @xmath378 .
moreover , @xmath379 ) = 1 \neq 0.\ ] ] indeed , since @xmath380 is a product , each @xmath119-holomorphic curve in @xmath369 must project to a @xmath119-holomorphic curve in each factor , and hence it must be of the form @xmath381 . therefore there is one such curve through every point @xmath382 . in general
there might be other @xmath119-holomorphic curves in the manifold @xmath1 that go through the designated point but do not lie in @xmath369 , which could count positively or negatively .
note however that @xmath369 is @xmath119-holomorphic , and @xmath383 = 0 $ ] .
therefore positivity of intersections of @xmath119-holomorphic submanifolds ( see ( * ? ? ?
* example 2.6.1 ) which , though stated in dimension @xmath73 , also holds for higher dimensions ) tells us that every @xmath119-holomorphic curve must lie entirely in @xmath369 .
hence there are no other @xmath119-holomorphic curves and @xmath377 ) = 1 $ ] .
the non - vanishing of the above gromov - witten invariant ( for chosen @xmath119 ) implies that for a generic choice of an almost complex structure @xmath384 , the evaluation map is onto , and thus the gromov width of @xmath1 is at most @xmath385 ( theorem [ thm ub jholom ] ) . using this argument
, we show that when the bending action on @xmath1 is toric and @xmath155 contains one facet @xmath365 of the cuboid @xmath362 as above , with one edge of length @xmath366 then the gromov width of @xmath1 is at most @xmath386 .
we show that if @xmath53 and @xmath52 are short then the top facet of the cuboid @xmath362 is in @xmath155 .
assume first that @xmath292 .
consider the following reshuffling of @xmath4 @xmath388 the bending action on @xmath30 associated to the choice of diagonals in figure [ fig:6gonsystem ] is toric as @xmath292 ( so also @xmath389 and @xmath390 ) .
the moment map image @xmath155 contains the `` top '' facet of the cuboid @xmath362 as it contains the vertices @xmath391 ( see the chart on page applied to @xmath54 ) .
then the image @xmath155 of @xmath30 contains the vertices @xmath391 if and only if @xmath392 , @xmath393 and @xmath394 are short for @xmath54 , or , equivalently , if and only if @xmath53 , @xmath395 and @xmath52 are short for @xmath4 .
( note that @xmath53 short implies that @xmath395 is also short . ) since the top facet of @xmath155 has shortest edge of length @xmath396 , by the argument above , if @xmath53 and @xmath52 are short then the gromov width of @xmath1 is less or equal then @xmath316 . if @xmath397 then the bending action on @xmath30 is not toric and the above argument does not apply . in that case
we proceed as before : consider the family @xmath398 and use the continuity argument ,
moser s trick " .
note that for @xmath8 ordered non - decreasingly , and such that @xmath399 none of our results for the upper bound applies .
hence in these cases we only have the lower bound as in proposition [ proposition 6lowerbounds ] .
a. klyachko , _ spatial polygons and stable configurations of points in the projective line _
, algebraic geometry and its applications ( yaroslavl , 1992 ) , 6784 , aspects math . , e25 , vieweg , braunschweig , 1994 .
d. mcduff , d. salamon , _ j - holomorphic curves and symplectic topology _ , second edition .
american mathematical society colloquium publications , 52 .
american mathematical society , providence , ri , 2012 .
xiv.+726 pp .
isbn 978 - 0 - 8218 - 8746 - 2 . | for generic @xmath0 the space @xmath1 of @xmath2gons in @xmath3 with edges of lengths @xmath4 is a smooth , symplectic manifold .
we investigate its gromov width and prove that the expression @xmath5 is the gromov width of all ( smooth ) @xmath6gons spaces and of @xmath7gons spaces , under some condition on @xmath8 .
the same formula constitutes a lower bound for all ( smooth ) spaces of @xmath7gons .
moreover , we prove that the gromov width of @xmath1 is given by the above expression when @xmath1 is symplectomorphic to @xmath9 , for any @xmath2 . |
relativistic dissipative hydrodynamics has been quite successful in explaining the spectra and azimuthal anisotropy of particles produced in heavy - ion collisions at the rhic and lhc .
apart from its applications , theoretical formulation of relativistic dissipative hydrodynamics is quite a challenging task by itself .
the first - order dissipative fluid dynamics commonly known as relativistic navier - stokes ( ns ) theory involves parabolic differential equations and suffers from acausality and instability .
the second - order or israel - stewart ( is ) theory @xcite with its hyperbolic equations restores causality but may not guarantee stability .
the correct formulation of relativistic dissipative fluid dynamics is still unresolved and is currently under intense investigation @xcite .
it is essential to note that all formulations of the second - order dissipative hydrodynamics employing the boltzmann equation ( be ) make a strict assumption of locality in the configuration space for the collision term @xcite .
this means , the collisions that increase or decrease the number of particles with a given momentum @xmath0 , in an infinitesimal space - time volume element , are assumed to occur at the same point @xmath1 .
this makes the collision integral a purely local functional of the single - particle phase - space distribution function @xmath2 independent of the derivatives @xmath3 .
although @xmath2 may not vary significantly over the length scale of a single collision event , its variation over the length scales extending over many inter - particle spacings may not be negligible . including the gradients of @xmath2 in the collision term
gives rise to different evolution equations for the dissipative quantities . in this article
, we provide a new formulation of the dissipative hydrodynamic equations within kinetic theory by using a nonlocal collision term in the boltzmann equation : @xmath4 $ ] .
for two - body elastic collisions , the collision term is @xmath5= \frac{1}{2 } \int dp ' dk \ dk ' \
w_{pp ' \to kk ' } ( f_k f_{k ' } \tilde f_p \tilde f_{p ' } - f_p f_{p ' } \tilde f_k \tilde f_{k'}),\ ] ] where @xmath6 is the collisional transition rate , @xmath7 , @xmath8 with @xmath9 for fermi , bose , and boltzmann gas and @xmath10 $ ] .
the first and second terms in eq .
( [ coll ] ) refer to the processes @xmath11 and @xmath12 , respectively .
these processes are traditionally assumed to occur at the same space - time point @xmath1 with an underlying assumption that @xmath2 is constant not only over a region characterizing a single collision , but also over an infinitesimal fluid element of size @xmath13 , large compared to the interparticle separation .
it is important to note that the space - time points at which the above two processes occur may be separated by a small interval @xmath14 within @xmath15 ( see fig .
[ flelmnt ] ) . with this realistic viewpoint , the second term in eq .
( [ coll ] ) involves @xmath16 , which on taylor expansion up to second order in @xmath14 , results in the modified be @xcite @xmath17 + \partial_\mu(a^\mu f ) + \partial_\mu\partial_\nu(b^{\mu\nu}f).\ ] ] and @xmath12 separated by a spacetime interval @xmath14 within an infinitesimal fluid element of size @xmath13 , containing a large number of particles represented by dots . ] the momentum dependence of coefficients , @xmath18 can be made explicit by expressing them in terms of the available tensors @xmath19 and the metric @xmath20 as @xmath21 and @xmath22 .
the scalar coefficients @xmath23 , @xmath24 and @xmath25 are functions of @xmath1 .
the conserved particle current , @xmath26 , and the energy - momentum tensor , @xmath27 have the standard form @xcite : @xmath28 conservation of current , @xmath29 and energy - momentum tensor , @xmath30 , yield the fundamental evolution equations for @xmath31 , @xmath32 and @xmath33 @xmath34 conservation of current and energy - momentum implies vanishing zeroth and first moments of the collision term @xmath35 $ ] leading to three constraint equations for the coefficients ( @xmath36 ) @xcite , @xmath37 in order to obtain the second - order evolution equations for dissipative quantities , we consider second moment of the modified be , eq .
( [ mbe ] ) @xmath38 + p^\gamma \partial_\gamma(af ) + \partial^2(b_1f_0 ) + ( p \cdot \partial)^2 ( b_2f_0 ) \big].\ ] ] using grad s 14-moment approximation for the single particle distribution function in orthogonal basis , @xmath39 we finally obtain the following evolution equations for the dissipative fluxes @xmath40 in the usual notation @xcite .
the 8 terms " ( `` 9 terms '' ) involve second - order , linear scalar ( vector ) combinations of derivatives of @xmath41 .
we observe that in the above equations , the nonlocal coefficients ( @xmath42 ) modify the standard ns as well as the is terms .
this method is also able to generate all possible second - order terms allowed by symmetry . )
normalized to their initial values ( left panel ) , and anisotropy parameter @xmath43 ( right panel ) .
initial values are @xmath44 fm / c , @xmath45 mev , @xmath46 , @xmath47.,width=340,height=226 ]
to illustrate the numerical significance of the non - locality in the collision term through the new dissipative hydrodynamic equations derived here , we consider bjorken evolution of a massless boltzmann gas ( @xmath48 ) at vanishing net baryon number density @xcite . in terms of the coordinates ( @xmath49 ) where @xmath50 and @xmath51 , the initial four - velocity becomes @xmath52 .
in this scenario @xmath53 and the equation for @xmath54 reduces to @xmath55 where the coefficients are @xmath56 the coupled differential equations ( [ evol ] ) , ( [ param ] ) and ( [ bjpi ] ) are solved simultaneously for a variety of initial conditions relevant for rhic and lhc .
figure [ evolu ] ( left panel ) , shows the evolution of several quantities for a particular choice of initial conditions .
@xmath57 decreases monotonically to the crossover temperature @xmath58 mev at time @xmath59 fm / c .
parameters @xmath24 , @xmath25 vary smoothly to zero at large times indicating reduced but still significant presence of nonlocal effects at late times .
this is also evident in fig .
[ evolu ] ( right panel ) , where the pressure anisotropy @xmath60 shows marked deviation from is .
although the shear pressure @xmath61 vanishes rapidly indicating approach to ideal fluid dynamics , the @xmath43 is far from unity .
faster isotropization for initial @xmath62 may be attributed to a smaller effective shear viscosity in the modified ns equation .
figure [ evol2 ] shows the evolution of @xmath43 for isotropic initial pressure configuration , at various @xmath63 for the lhc energy regime . with small initial corrections ( @xmath64% to ns and @xmath65% to the is terms ) due to @xmath66 , nonlocal hydrodynamics ( solid lines ) exhibits appreciable deviation from the ( local ) is theory ( dashed lines ) .
the above results clearly demonstrate the importance of the nonlocal effects , which should be incorporated in transport calculations as well .
in is ( dashed lines ) and the present work ( solid lines ) for isotropic initial pressure configuration.,width=283,height=219 ]
to summarize , we have derived viscous hydrodynamic equations by introducing a nonlocal generalization of the collision term in the boltzmann equation .
the navier - stokes as well as israel - stewart equations are modified and new terms are obtained in the evolution equations of the dissipative quantities .
the method presented is able to generate all possible terms that are allowed by symmetry . within one - dimensional scaling expansion ,
we find that nonlocality of the collision term has a rather strong influence on the evolution of the viscous medium via hydrodynamic equations . 50 h. song et .
al . , phys .
lett . * 106 * , 192301 ( 2011 ) .
w. israel and j. m. stewart , annals phys . * 118 * , 341 ( 1979 ) .
g. s. denicol , t. koide and d. h. rischke , phys .
lett . * 105 * , 162501 ( 2010 ) .
a. jaiswal , r. s. bhalerao and s. pal , phys .
b * 720 * , 347 ( 2013 ) j. d. bjorken , phys .
d * 27 * , 140 ( 1983 ) . | generalizing the collision term in the relativistic boltzmann equation to include nonlocal effects , and using grad s 14-moment approximation for the single - particle distribution function , we derive evolution equations for the relativistic dissipative fluid dynamics and compare them with the corresponding equations obtained in the standard israel - stewart and related approaches .
significance of this generalization on hydrodynamic evolution is demonstrated in the framework of one - dimensional scaling expansion . * relativistic hydrodynamics from boltzmann equation with modified collision term * + * amaresh jaiswal , rajeev s. bhalerao and subrata pal * + tata institute of fundamental research , mumbai , india |
in algebra and geometry , a group action is a way of describing symmetries of objects using groups .
the essential elements of the object are described by a set and the symmetries of the object are described by the symmetric group of this set , which consists of bijective transformations on the set .
the symmetries play an important role in the classical and quantum mechanics .
it is well known that the modular group @xmath8 has finite presentation @xmath9 where @xmath10 are elliptic transformations and their fixed points in the upper half plane are @xmath11 and @xmath12 .
+ coset diagrams have been used to study quotients , orbits , number of subgroups and structure of the finitely generated groups .
the concept of coset diagrams seems to originate from the work by schreier and reidemeister in the 1920s .
the interest has been risen in the last decades as the possibilities to use the coset diagram techniques in combination with mathematical software has been improved .
novotny and hrivnak @xcite consider the action of the finitely generated group @xmath13 on the ring @xmath14 and determined the orbits for @xmath15 arbitrary natural number .
shabbir and khan @xcite discussed conformally flat- but non flat bianchi type i and cylindrically symmetric static space - times according to proper projective symmetry by using some algebraic and direct integration techniques .
it is shown that the special class of the above space - times admit proper projective vector fields .
+ torstensson used coset diagrams to study the quotients of the modular group in @xcite .
the number of subgroups of @xmath8 when acting on @xmath16 has been discussed in @xcite and the subgroups of the classical modular group has been discussed in @xcite .
higman and mushtaq ( 1983 ) introduced the concept of the coset diagrams for the modular group @xmath8 and laid its foundation .
mushtaq ( 1988 ) showed that for a fixed non - square positive integer @xmath15 , there are only a finite number of ambiguous numbers in @xmath17 and that the ambiguous numbers in the coset diagram for the orbit @xmath18 form a closed path and it is the only closed path contained in it . by using the coset diagrams for the orbit of the modular group @xmath9 acting on the real quadratic fields m. a. malik et al determined the exact number of ambiguous numbers in @xmath17 in @xcite , as a function of @xmath15 .
the ambiguous length of an orbit @xmath18 is the number of ambiguous numbers in the same orbit .
m. a. malik et al . in @xcite proved that @xmath19 , @xmath20 , splits into at least two orbits namely @xmath21 and @xmath22 , and it was also proved by the same authors that @xmath19 , @xmath23 , splits into at least two orbits namely @xmath21 and @xmath24 @xcite .
in @xcite it was proved that there exist two proper @xmath0-subsets of @xmath17 when @xmath25 and four @xmath0-subsets of @xmath17 when @xmath26 . in @xcite
we generalized these result for @xmath27 , where @xmath28 are distinct odd primes .
we also proved for @xmath29 there are exactly two @xmath0-orbits of @xmath30 namely @xmath31 and @xmath32 .
+ by a circuit @xmath33 we always mean a closed path in which @xmath34 triangles have one vertex outside the circuit and @xmath35 triangles have one vertex inside the circuit and so on .
this circuit induces an element @xmath36 of @xmath0 which fixes a particular vertex @xmath37 , where @xmath37 must be an ambiguous number , the detail is given in @xcite .
+ example : by a circuit @xmath38 we mean the transformation + @xmath39 which fixes a particular vertex @xmath37 , that is @xmath40 as shown in figure 1 .
+ + fig : 1 * * theorem 1.1**@xcite + let @xmath41 and @xmath42 with @xmath43 . then @xmath44 is a @xmath0-subset of @xmath17 for each @xmath45 and @xmath46
. + + * * theorem 1.2**@xcite + let @xmath7 be an odd prime and @xmath47 ; @xmath48 , take @xmath49 with @xmath43 .
then + @xmath50 and + @xmath51 are exactly two @xmath0-subsets of @xmath52 , depending upon the classes @xmath53(mod~p^k)$ ] +
in @xcite m. a. malik and m. riaz obtained @xmath0-subsets of @xmath17 by the help of @xmath54 acting on @xmath17 , when @xmath25 .
we also proved for @xmath29 there are exactly two @xmath0-orbits of @xmath30 namely @xmath31 and @xmath32 . in @xcite m. a. malik
et al . obtained @xmath0-subsets of @xmath17 when @xmath55 and @xmath56 .
thus it becomes interesting to know about more @xmath0-subsets and to know about the exact number of @xmath0-orbits under the action of the modular group @xmath0 . in the present studies , we use coset diagrams and modular arithmetic to determine the exact number of @xmath0-orbits of @xmath1 , @xmath2 , @xmath3 , and in general @xmath4 , for all @xmath5 and @xmath6 , for each odd prime @xmath7 .
+ + * theorem 2.1 * + if @xmath20 and @xmath6 then there are exactly two @xmath0-orbits of @xmath1 namely @xmath57 and @xmath58 . +
* + if @xmath20 then by @xcite , @xmath59 and @xmath22 are two distinct @xmath0-orbits of @xmath19 .
then by @xcite and @xcite the classes @xmath53~(mod p^k$ ] ) with @xmath60 or @xmath61 quadratic residues of @xmath62 lie in the orbit @xmath57 and similarly the classes @xmath53~ ( mod p^k$ ] ) , with @xmath60 or @xmath61 quadratic non - residues of @xmath62 lie in the orbit @xmath58 . if @xmath20 then by @xcite , @xcite @xmath63 and none of @xmath64 and @xmath65 is contained in @xmath58 .
thus it is clear that @xmath57 lies in the @xmath66 and @xmath58 lies in @xmath67 .
moreover the @xmath0-subsets @xmath66 and @xmath67 are transitive . + in the closed path lying in the orbit @xmath57 ,
the transformation @xmath68 given by @xmath69 fixes @xmath70 , that is @xmath71 , and so gives the quadratic equation @xmath72 , the zeros , @xmath73 , of this equation are fixed points of the transformations @xmath68 . similarly in the closed path lying in the orbit @xmath58 , the transformation @xmath74 given by @xmath75 fixes @xmath76 , this proves the result .
@xmath77 + + * example 2.2 * + there are exactly two @xmath0-orbits of @xmath78 namely @xmath79 , in the closed path lying in the orbit @xmath80 , the transformation @xmath81 fixes @xmath82 , that is @xmath71 , and so gives the quadratic equation @xmath83 , the zeros , @xmath84 , of this equation are fixed points of the transformations @xmath68 .
similarly in the closed path lying in the orbit @xmath85 , the transformation @xmath86 fixes @xmath87 .
@xmath77 + * theorem 2.3 * + if @xmath23 and @xmath6 then there are exactly two @xmath0-orbits of @xmath1 namely @xmath57 and @xmath88 , for each odd prime @xmath7 . + * proof . *
+ if @xmath23 then by @xcite @xmath59 and @xmath24 are two distinct @xmath0-orbits of @xmath19 .
by @xcite it is clear that @xmath57 lies in the @xmath66 and @xmath88 lies in @xmath67 .
moreover the @xmath0-subsets @xmath66 and @xmath67 are transitive . + in the closed path lying in the orbit @xmath89 ,
the transformation @xmath90 given by @xmath91 fixes @xmath64 , that is @xmath92 , and so gives the quadratic equation @xmath72 , the zeros , @xmath93 , of this equation are fixed points of the transformations @xmath90 . similarly in the closed path lying in the orbit @xmath88 , the transformation @xmath94 given by @xmath95 fixes @xmath65 , this proves the result .
@xmath77 + + * example 2.4 * + there are exactly two @xmath0-orbits of @xmath96 namely @xmath97 , in the closed path lying in the orbit @xmath98 , the transformation @xmath99 fixes @xmath100 , that is + @xmath101 , + and so gives the quadratic equation @xmath102 , the zeros , @xmath103 , of this equation are fixed points of the transformations @xmath68 .
similarly in the closed path lying in the orbit @xmath104 , the transformation + @xmath105 + fixes @xmath104 . @xmath77 + * theorem 2.5 * + if @xmath20 and @xmath6 then there are exactly two @xmath0-orbits of @xmath2 namely @xmath106 and @xmath107 . + * proof .
* + since @xmath108 is the only quadratic residue of @xmath109 and there is no quadratic non - residue of @xmath109 .
thus the quadratic residues and quadratic non residues of @xmath62 and @xmath110 are the same .
then by @xcite and @xcite the classes @xmath53 $ ] ( modulo @xmath62 ) with @xmath60 or @xmath61 quadratic residues of @xmath62 lie in the orbit @xmath89 and similarly the classes @xmath53 $ ] ( modulo @xmath62 ) , with @xmath60 or @xmath61 quadratic non - residues of @xmath62 lie in the orbit @xmath107 .
if @xmath20 then by @xcite @xmath111 and none of @xmath112 and @xmath113 is contained in @xmath58 . thus it clear that @xmath106 and @xmath107 are distinct orbits .
+ by @xcite it is clear that @xmath57 lies in the @xmath66 and @xmath58 lies in @xmath67 .
+ in the closed path lying in the orbit @xmath106 , the transformation @xmath68 given by @xmath69 fixes @xmath70 , that is @xmath71 , and so gives the quadratic equation @xmath114 , the zeros , @xmath115 , of this equation are fixed points of the transformations @xmath68 . similarly in the closed path lying in the orbit @xmath107 , the transformation @xmath74 given by @xmath75 fixes fixes @xmath113 , this proves the result .
@xmath77 + * theorem 2.6 * + if @xmath23 and @xmath6 then there are exactly two @xmath0-orbits of @xmath2 namely @xmath106 and @xmath116 , for each odd prime @xmath7 . + * proof .
* + since @xmath108 is the only quadratic residue of @xmath109 and there is no quadratic non - residue of @xmath109
. thus the quadratic residues and quadratic non residues of @xmath62 and @xmath110 are the same .
the classes @xmath53 $ ] ( modulo @xmath110 ) with @xmath60 or @xmath61 quadratic residues of @xmath110 lie in the orbit @xmath117 and similarly the classes @xmath53 $ ] ( modulo @xmath110 ) , with @xmath60 or @xmath61 quadratic non - residues of @xmath110 lie in the orbit @xmath116 . if @xmath23 then by @xcite @xmath64 and @xmath88 are distinct orbits .
+ by @xcite it is clear that @xmath57 lies in the @xmath66 and @xmath88 lies in @xmath67 .
+ in the closed path lying in the orbit @xmath117 , the transformation @xmath90 given by @xmath91 fixes @xmath64 , that is @xmath92 , and so gives the quadratic equation @xmath114 , the zeros , @xmath118 , of this equation are fixed points of the transformations @xmath68 .
similarly in the closed path lying in the orbit @xmath116 , the transformation @xmath94 given by @xmath95 fixes @xmath113 , this proves the result .
@xmath77 + * theorem 2.7 * + if @xmath20 and @xmath6 then there are exactly two @xmath0-orbits of @xmath3 namely @xmath119 and @xmath120 . + * proof .
* + the proof is analogous to the proof of theorem 2.5 . +
* theorem 2.8 * + if @xmath23 and @xmath6 then there are exactly two @xmath0-orbits of @xmath3 namely @xmath106 and @xmath116 , for each odd prime @xmath7 . + * proof . *
+ the proof is analogous to the proof of theorem 2.6 .
+ + the results becomes interesting in modulo @xmath121 and we observe that there are exactly four @xmath0-orbits when @xmath122 , @xmath5 . + + *
theorem 2.9 * + let @xmath6 and @xmath5 , then there are exactly four @xmath0-orbits of @xmath4 namely @xmath123 , @xmath124 , @xmath125 and @xmath126 for each odd prime @xmath7 . + * proof .
* + we know by theorem 1.1 @xcite that if @xmath41 then @xmath127 is a @xmath0-subset of @xmath52 for each @xmath45 and @xmath46 , we represent these @xmath0-subsets by @xmath66 , @xmath67 , @xmath128 and @xmath129 .
+ thus it is easy to see that the orbits @xmath123 , @xmath124 , @xmath125 and @xmath130 are contained in @xmath66 , @xmath67 , @xmath128 and @xmath129 respectively .
moreover it is clear that @xmath0-subsets given by @xmath66 , @xmath67 , @xmath128 and @xmath129 are transitive .
+ in the closed path lying in the orbits @xmath131 , the transformation @xmath132 fixes @xmath133 . in the closed path lying in
the orbits @xmath125 and @xmath126 , the transformation @xmath134 fixes @xmath135 , this proves the result .
@xmath77 + + * example 2.10 * + there are exactly four @xmath0-orbits of @xmath136 namely @xmath137 , @xmath137 , @xmath138 and @xmath139 .
+ similarly there are exactly four @xmath0-orbits of @xmath140 namely @xmath141 , @xmath141 , @xmath142 and @xmath143 .
+ + * conclusion * the action of the modular group @xmath8 on @xmath1 , @xmath6 is intransitive .
@xmath4 splits into exactly two @xmath0-orbits for each @xmath5 and @xmath144 , moreover @xmath4 splits into exactly four @xmath0-orbits for each @xmath5 and @xmath144 or @xmath145 . | we look at some ways in which coset diagrams have been used to find the orbits , number of subgroups and structure of the finitely generated groups . in this paper
we use coset diagrams and modular arithmetic to determine the exact number of @xmath0-orbits of @xmath1 , @xmath2 , @xmath3 , and in general @xmath4 , for each @xmath5 and @xmath6 , for each odd prime @xmath7 .
+ ams mathematics subject classification ( 2000 ) : 05c25 , 11e04 , 20g15 + * keywords : * finitely generated groups ; coset diagrams ; orbits . + |
because the neutral ( @xmath4 ) members of the @xmath5 octet of vector currents , @xmath6 ( with @xmath7 the usual gell - mann matrices ) , couple to fermions in the standard model , it is possible to use experimental data on the spectral functions associated with correlators involving these currents to determine certain quantities of phenomenological interest .
for example , defining the scalar correlators , @xmath8 , by means of @xmath9 and the corresponding spectral functions , @xmath10 , as usual , by @xmath11 , one finds that ( 1 ) integrating the difference @xmath12 with the weight function occuring naturally ( due to kinematics ) in the finite energy sum rule ( fesr ) treatment of hadronic @xmath13 decays@xcite produces a sum rule from which one can , in principle , determine the running strange quark mass , @xmath14@xcite , and ( 2 ) integrating the same difference @xmath12 with weight function @xmath15 produces a sum rule from which one can extract one of the @xmath2 order low - energy constants ( lec s ) , @xmath3 , appearing in the @xmath2 order version of the effective chiral lagrangian@xcite .
( see ref .
@xcite for a discussion of chiral perturbation theory ( chpt ) and the method of effective chiral lagrangians in general , ref .
@xcite for the form of the @xmath16 terms in the effective lagrangian in the most general case , and ref .
@xcite for both a discussion of the subset of these terms surviving when one restricts one s attention to vacuum correlators , and a definition of @xmath3 . ) of course , @xmath17 and @xmath18 do not couple separately in the standard model , but only in the combination @xmath19 which gives the light quark ( @xmath20 ) part of the electromagnetic ( em ) current .
thus , what is measured in @xmath21 , is not the desired quantities , @xmath22 and @xmath23 , separately , but the combination @xmath24 in the isospin symmetry limit , @xmath25 would vanish and , since one could then classify the final hadronic states according to their g - parity , it would be straightforward to separate the isovector ( @xmath26 ) and isoscalar ( @xmath27 ) components of the em spectral function . in the presence of isospin breaking , however , this process is no longer so straightforward .
the most obvious experimental signature of the presence of isospin - breaking in @xmath21 is the interference shoulder in the @xmath28 cross - section in the @xmath29-@xmath30 region@xcite .
the @xmath31 contribution to @xmath32 is clearly , to leading order in isospin - breaking , to be associated with @xmath25 and hence is usually removed explicitly in analyzing the data .
this removal is accomplished by ( 1 ) fitting the parameters of a model for the total @xmath28 amplitude , consisting of @xmath29 , @xmath30 and possible background contributions , to the experimental data , ( 2 ) removing the @xmath30 contribution once the fit has been performed and ( 3 ) squaring the modulus of the remaining @xmath29 contribution and identifying this result with the @xmath29 contribution to @xmath22@xcite .
while this procedure _ does _ remove one source of isospin breaking contamination from the nominal @xmath22 so extracted , it is easy to see that other such contaminations still remain . indeed ,
once one allows isospin - breaking , the physical @xmath29 and @xmath30 are admixtures of pure isovector and isoscalar states , the size of the admixture of the `` wrong '' isospin component being governed by the scale of isospin breaking . as a consequence , the intermediate @xmath29 contribution to @xmath25 ,
for example , does not vanish .
in fact , if one denotes the pure isovector @xmath29 state by @xmath33 and the pure isoscalar @xmath30 state by @xmath34 , one expects @xmath29 contributions to @xmath25 from two sources : ( 1 ) that due to @xmath33-@xmath34 mixing ( a one particle reducible contribution , with coupling of the isovector current to the @xmath33 component and the isoscalar current to the @xmath34 component of the @xmath29 ) , and ( 2 ) that due to the `` direct '' ( one particle irreducible , @xmath35 ) coupling of the @xmath33 component to the isoscalar current ( such a coupling being unavoidable in any hadronic effective lagrangian based on qcd ) .
thus , removing the contribution due to the intermediate state @xmath30 from the @xmath28 cross - section , while removing part of the @xmath25 contribution , does not remove it all .
one is then left with , _ not _ the desired quantity , @xmath22 , but rather with a combination of @xmath22 and the residual part of @xmath25 associated with the intermediate @xmath29 state ( plus possible additional such contaminations from elsewhere in the spectrum ) .
similar isospin - breaking flavor @xmath36 contributions exist for @xmath37 , complicating the extraction of the isoscalar spectral function .
corrections for such isospin breaking effects , _ which are unavoidable so long as no process exists in which only one of the two neutral flavor currents couples _ , are thus necessary if one wishes to perform phenomenological analyses of the type mentioned above .
such corrections would also be important in performing precision tests of cvc , which involve a comparison of @xmath22 and the charged isovector spectral function @xmath38 , measured in hadronic @xmath13 decays ( see , for example , ref .
@xcite ) .
it is easy to see that , to be able to make these corrections ( at least in the region below @xmath39 gev@xmath40 , where the em spectral function is , experimentally , resonance dominated ) , it is sufficient to determine the isospin - breaking vector meson decay constants
. let us first clarify notation .
we define the flavor @xmath41 and @xmath42 vector meson decay constants via @xmath43 where @xmath44 , @xmath45 is the vector meson polarization vector , and @xmath4 .
@xmath46 , @xmath47 and @xmath48 are non - zero in the limit of isospin symmetry ; @xmath49 , @xmath50 and @xmath51 zero in the absence of isospin breaking .
the experimentally determined em decay constants , @xmath52 , are then given by @xmath53 thus , for example , the broad @xmath29 contribution to @xmath32 , usually taken to be associated purely with @xmath22 , consists not only of a flavor @xmath26 contribution proportional to @xmath54 ^ 2 $ ] , but also of a flavor @xmath36 contribution proportional to @xmath55 .
the @xmath30 contribution to @xmath32 , similarly , contains both a flavor @xmath27 part proportional to @xmath56 ^ 2 $ ] and a flavor @xmath36 part proportional to @xmath57 .
the flavor @xmath36 parts , in both cases , are present only due to isospin breaking , and have to be removed from the experimental @xmath29 and @xmath30 contributions to @xmath32 in order to obtain the corresponding @xmath29 contribution to @xmath22 and @xmath30 contribution to @xmath23 .
it is important to stress at this point that the conventional `` few - percent '' rule - of - thumb for estimating the size of isospin - breaking effects , which might lead one to expect such effects to be numerically negligible , is inapplicable in the cases involving @xmath12 discussed above .
this is true for a number of reasons .
first , because the difference of spectral functions is itself flavor - breaking , the _ relative _ importance of isospin breaking is enhanced by a factor @xmath58 , characteristic of the inverse of the scale of flavor - breaking .
second , the effect of @xmath29-@xmath30 mixing naturally produces corrections for the @xmath29 contribution to @xmath22 and @xmath30 contribution to @xmath23 which are opposite in sign ; the effects therefore add when the difference is taken .
finally , there is a natural numerical enhancement which makes the size of the correction needed to remove the @xmath25 part of the @xmath30 contribution to @xmath32 , and hence isolate @xmath23 , larger than naively expected@xcite .
the latter two points are discussed in somewhat more detail in section ii below . in what follows ,
we evaluate the isospin - breaking vector meson decay constants by performing a qcd sum rule analysis of the isospin - breaking vector current correlator @xmath59 .
the vector meson spectral contributions are , in this case , proportional to @xmath60 , so that a determination of this product , in combination with the experimental determination of @xmath52 , given in terms of @xmath61 and @xmath62 above , allows a separate determination of @xmath61 and @xmath62 .
the rest of the paper is organized as follows . in section
ii we discuss qualitative expectations for the pattern of isospin - breaking corrections based on the structure of the leading ( chiral ) order terms in the vector meson effective chiral lagrangian , as well as semi - quantitative expectations for their probable scale which can , using this perspective , be obtained from experimental data . in section iii , we discuss briefly the form of qcd sum rules employed ( a version of fesr ) , and the advantages of this approach . in section iv , we discuss the input used for the hadronic and ope sides of the sum rules employed , and present our results . some advantages of the approach , in particular _
viz a vis _ the handling of the @xmath1 terms in the ope of the @xmath36 correlator , will also be discussed here . finally , in section
v we summarize , and make some brief comments on the phenomenological significance of our results .
chpt provides both an underlying conceptual framework and systematic procedure@xcite for writing down the most general effective lagrangian relevant to a given set of hadronic states which fully incorporates the symmetries of qcd and implements the broken symmetries ( such as chiral symmetry ) with the same pattern of symmetry breaking as occurs in qcd . although the resulting effective lagrangian , @xmath63 , is non - renormalizable , and hence contains a infinite number of low - energy constants ( lec s ) ( coefficients of those terms allowed by the symmetry arguments that go into constructing @xmath63 ) , it is possible to formulate the theory in such a way that only a finite number of such terms appear to a given order in the chiral , or low - energy expansion .
( for so - called `` heavy '' fields , those whose masses are non - vanishing in the chiral limit , this requires a re - formulation in terms of velocity - dependent fields@xcite . ) the leading order terms in this expansion ( in which light quark masses , @xmath64 , @xmath65 , count as @xmath66 , with @xmath67 representing some soft external momentum ) , incorporate the leading constraints associated either with chiral symmetry , or the symmetry pattern of its breaking . the heavy field implementation of chpt given in ref .
@xcite for the vector mesons and their interactions with the members of the pseudo - goldstone boson pseudoscalar octet provides some useful information about the pattern of isospin - breaking mixing in the vector meson sector .
one must bear in mind that , in general , mixing in field theory is more complicated than in hamiltonian quantum mechanics .
this is because , in an effective field theory , one in general has ( and can not , in fact , avoid ) momentum - dependent mixing terms .
this means that there are off - diagonal elements of the wave - function renormalization matrix and , as a consequence , the relation between the original , unmixed fields and the final , renormalized and diagonalized fields , involves the product of a symmetric matrix and a rotation ( and hence is not itself a rotation ) .
( see , for example , the discussion of the effects of @xmath68-@xmath69 mixing in ref .
@xcite ; an expanded version is also given as part of the discussion of the treatment of the mixed - isospin axial current correlator in ref . @xcite . ) from the heavy vector meson field formulation , however , one sees that the leading ( in chiral order ) term in @xmath63 generating isospin - breaking mixing involves one power of the quark mass matrix and no derivatives@xcite , and hence produces no off - diagonal contributions to the wave - function renormalization matrix .
the _ leading order _ mixing effect thus results in a physical @xmath29 and @xmath30 basis which is related by a rotation to the original pure isospin @xmath33 , @xmath34 basis , just as in the quantum mechanical case .
the `` wrong '' isospin @xmath34 admixture in the physical @xmath29 state is thus equal in magnitude , but opposite in sign , to that of the @xmath33 admixture in the physical @xmath30 state , at this order in the chiral expansion .
while higher ( chiral ) order corrections exist , this pattern should remain approximately valid , even at higher order .
let us now consider the vector meson decay constants . since the chiral limit is also @xmath5 symmetric , the leading order term generating the vector meson decay constants is necessarily @xmath5-symmetric .
when one now considers the effects of flavor- and isospin - symmetry breaking ( recalling that both are generated by the quark mass matrix , and hence both are produced by the same set of terms in the effective lagrangian ) , there are two potential sources of flavor and isospin breaking in the vector meson decay constants .
the first is that associated with higher order terms , involving at least one power of the quark mass matrix , coupling the external photon field to the vector meson nonet , the second that induced by the leading quark - mass - dependent term , responsible for mixing , discussed above .
the leading order mixing effect simply reproduces the standard leading order @xmath5 mixing analysis@xcite , leading to near ideal mixing in the vector meson sector .
as is well - known , the combination of ideal mixing and neglect of flavor - breaking in the em couplings of the unmixed states leads to the prediction that the vector meson em decay constants , measured experimentally in @xmath70,@xcite should be in the proportions @xmath71 , where the superscript @xmath72 indicates that the couplings refer to the ideally mixed , but isospin pure vector meson states . that this prediction is borne out by experiment
represents empirical evidence that , despite the potential @xmath5-breaking photon coupling contributions being of the same formal order as effects induced by mixing , the former are numerically suppressed relative to the latter .
since flavor breaking and isospin breaking are generated by the same terms in the effective lagrangian , this implies that isospin breaking in the vector meson decay constants should also be dominated by mixing effects .
if we take this point of view then , up to sub - leading corrections , we find , for the physical @xmath29 and @xmath30 decay constants , now including isospin breaking and taking into account the relation @xmath73 , @xmath74 where @xmath75 is the leading order mixing angle , defined via @xmath76 we note two relevant features of these results : ( 1 ) because of the dominance by mixing , the corrections required to convert the pure isovector @xmath77 coupling to the experimental @xmath78 coupling is opposite in sign to that required to convert the pure isoscalar @xmath79 coupling to the @xmath80 and , ( 2 ) because of the pattern of ideal mixing and the numerical suppression of the isoscalar current relative to the isovector current in @xmath81 , the magnitude of the correction is a factor of @xmath82 larger in the @xmath30 than in the @xmath29 case .
this does not mean that , somehow , isospin breaking has become huge in the @xmath30 case , but rather reflects the expected natural suppression of the @xmath30 coupling relative to that of the @xmath29 , which is confirmed by experiment .
in general , but particularly in light of the large numerical enhancement just discussed , it is crucial to find a means of estimating the isospin - breaking contributions to the vector meson em decay constants . in view of the discussion above , although there exist additional effects beyond those associated with mixing at leading order , one can get a rough idea of the size expected for the isospin - breaking vector meson decay constants by considering experimental information on @xmath29-@xmath30 mixing , and ignoring all non - mixing effects .
although crude , this estimate provides an additional qualitative constraint for our later sum rule analysis . in order to obtain the parameter @xmath75 describing @xmath29-@xmath30 mixing at leading order , it is sufficient to determine the off - diagonal element , @xmath83 , of the vector meson self - energy matrix . in the past ,
values for @xmath83 in the range @xmath84 mev@xmath40 have been quoted , based on simplified analyses of @xmath28 data in the interference region which effectively assume that the one - particle irreducible @xmath85 vertex is zero , even in the presence of isospin breaking . since effective operators which generate such a coupling exist in the vector meson effective lagrangian , however , this assumption is unphysical ( in the sense of being incompatible with qcd ) .
once one includes contributions to the @xmath86 amplitude generated both by @xmath29-@xmath30 mixing _ and _ the 1pi vertex ( whose strength we will denote by @xmath87 ) , the analysis of the experimental data is somewhat more complicated but , in principle , allows a separate determination of both @xmath83 and the isospin - breaking ratio of couplings of the isospin pure states @xmath88@xcite . an important feature of the analysis framework developed in ref .
@xcite is that , from it , one understands that the smallness of previously quoted errors for @xmath83 is an artifact of the unphysical assumption @xmath89 , and does not survive the more general treatment .
it is worth outlining the reason why this is the case since , in doing so , it will become clear that it is a difficult task to improve the experimental situation sufficiently to really pin down the mixing contribution . the contribution of the _ physical _
( i.e. , mixed - isospin ) @xmath30 to the amplitude for @xmath28 is obtained experimentally by determining the timelike pion form factor , @xmath90 , in the interference region and fitting it to a form @xmath91 \ , + \ , { \rm background } \label{fpi}\ ] ] where @xmath92 are the complex pole positions , @xmath93 , and the fit parameter , @xmath94 , is known as the `` orsay phase '' . the @xmath30 contribution in eq .
( [ fpi ] ) is generated by the coupling of the physical @xmath30 to @xmath95 which , as discussed above , has two sources : 1pr ( @xmath33-@xmath34 mixing ) , and 1pi ( associated with the @xmath96 vertex ) .
the physical coupling is given , in terms of these contributions , by @xmath97 where , as usual , the superscript @xmath72 indicates couplings of the unmixed isospin pure states .
in the ( physically plausible ) approximation that one assumes saturation of the imaginary part of @xmath83 by @xmath98 intermediate states , one finds @xmath99 and hence , in the narrow @xmath29-@xmath30 interference region , @xmath100 where @xmath101 is now real .
the mixing angle @xmath75 is then given by@xcite @xmath102 where @xmath103 one then finds , upon substitution of eq .
( [ epsilon ] ) into eq .
( [ gomega ] ) , that @xmath104 \ , g^{(0)}_{\rho\pi\pi}\ . \label{renard}\ ] ] in many places in the literature , one finds the approximation @xmath105 employed .
the deviation of the quantity @xmath106 from @xmath107 represents the error made in employing this approximation .
since @xmath108 and @xmath109 is small ( @xmath110 to @xmath111 ) , one might think it rather safe to set @xmath112 in the above analysis ( this approximation was , in fact , made uniformly in analyses previous to the discussion of ref .
@xcite ) . if it were true that this approximation were reliable , then the effect of @xmath113 in eq .
( [ renard ] ) would cancel exactly@xcite , and the experimental data would determine the real part of @xmath83 in the interference region with the usually quoted errors ( @xmath114 mev@xmath40 ; see ref .
@xcite and earlier references cited therein ) .
unfortunately , it turns out that the approximation is both misleading and unreliable .
the reason is that , although @xmath106 is approximately real and near @xmath107 , @xmath115 is dominantly imaginary . since the denominator of the second term in eq .
( [ renard ] ) is also dominantly imaginary , the two terms add nearly constructively . were the phases of these terms to be actually identical , of course , it would be impossible to use experimental information to separate them , no matter how precise that data .
fortunately , there is a small phase difference which , at least in principle , means that a determination , with sufficient accuracy , of both the magnitude , @xmath116 , and phase , @xmath94 , of the @xmath30 contribution to @xmath117 , would allow separate determination of @xmath113 and @xmath101 . from this
, one would be able to reconstruct @xmath83 and hence determine @xmath75 .
there is , however , a practical difficulty , which limits the accuracy obtainable using present experimental data .
this is a result of the fact that the phase difference of the two terms is not large , so one is rather close to being in the `` identical phase '' situation , for which a separate extraction of the two terms would be impossible . if one takes the updated numerical analyses of ref .
@xcite , for example , one sees that values of @xmath101 between @xmath118 and @xmath119 mev@xmath40 are allowed by present data ( with a central value @xmath120 mev@xmath40 ) , and that , while the central extraction for @xmath113 is moderately large @xmath121 , @xmath89 is only @xmath122 distant .
a significant improvement in this situation would require a significant reduction of the errors in the determination of the orsay phase .
unfortunately , the prospects for seeing such an improvement in the near future are remote , at present .
although , with present experimental accuracy , the errors in the determination of @xmath75 are disappointingly large , we can , nonetheless , as explained above , use the range of values obtained in refs .
@xcite to help set a rough scale for the expected size of those corrections required to go from @xmath78 to @xmath123 , and @xmath80 to @xmath124 . using the central values for the four fits given in table 1 of ref .
@xcite , one finds that @xmath123 is _ less than
_ @xmath78 by between @xmath125 and @xmath126 ( the former corresponding to fixing @xmath89 by hand , the latter to the mow and a solutions contained in the table of ref .
@xcite ) and @xmath124 _ greater than
_ @xmath80 by between @xmath127 and @xmath128 .
we will see that the solutions obtained below via the sum rule analysis satisfy these rather loose constraints , and , in fact , provide an alternate determination which is to be favored since it both has considerably smaller errors and is free of the uncertainties associated with effectively truncating at leading chiral order ( which goes into the estimate just discussed ) .
as is well - known , the properties of unitarity and analyticity lead to the existence of various ( appropriately subtracted ) dispersion relations for typical hadronic correlators , @xmath129 . by the term `` dispersion relations '' we mean here both those relations based on the cauchy representation theorem , and those based on cauchy s theorem ( the vanishing of the integral of an analytic function over a closed contour entirely in the region of analyticity ) .
the generic term `` qcd sum rules '' describes those versions of these relations obtained by making kinematic restrictions which allow one to take advantage of the asymptotic freedom of qcd . in the first case ,
one simply writes the representation theorem for @xmath129 with @xmath130 large and spacelike ; in the second , a choice of the integration contour is made so that at least part of the contour lies in the region of large spacelike momenta . in either case , the point of the kinematic restriction is to take advantage of the fact that , in the region of large spacelike momenta , computational techniques based on the operator product expansion ( ope)/perturbative qcd ( pqcd ) are applicable . the generic form of the sum rules generated using the representation theorem is then , up to possible subtractions , @xmath131 where the @xmath132 is the lowest physical threshold in the channel in question and the spectral function , @xmath29 , is defined as usual by @xmath133 .
the lhs is to be computed using the ope / pqcd , while the rhs is given in terms of measured spectral data and/or some spectral anstz ( involving a limited number of unknown parameters ) for the unmeasured part of the spectral function .
as first pointed out by shifman , vainshtein and zakharov ( svz)@xcite , the utility of such relations is greatly improved by borel transformation .
the effect of the borel transform is ( 1 ) to replace the weight @xmath134 in the hadronic spectral integral on the rhs of eq .
( [ normsr ] ) with @xmath135 , where @xmath136 , the borel mass , is a parameter of the transformation , ( 2 ) to kill all subtraction terms and ( 3 ) to create a factorial suppression of the contributions of higher dimensional operators on the ope side of the equation ( @xmath137 ) . ideally , on the hadronic side
, one would like to choose @xmath136 as small as possible , in order to suppress the contributions from the large-@xmath138 part of the integral , for which the spectral function will typically be complicated , and difficult to model .
in contrast , to improve the convergence of the ope when truncated at operators of relatively low dimension , one wants @xmath136 as large as possible . to be usable
, svz sum rules thus require a compromise : one must hope to find a `` stability window '' in @xmath136 , i.e. , a region of @xmath136 values for which both contributions from the complicated part of the spectral distribution ( on the hadronic side ) and from the highest dimension operators ( on the ope side ) are not too large .
typically , because of the compromise nature of the choice of stability window , this means that neither the contributions from the large-@xmath138 part of the spectrum , nor that from the highest dimension operator retained on the ope side , are negligible@xcite .
for the case of sum rules based on cauchy s theorem , a common choice of integration contour ( and the one we will employ in the analysis below ) is that shown in fig . 1 .
the radius , @xmath139 , of the circular part of the contour is to be taken large enough that the ope , to the order available , is reliable in the spacelike region of the circle .
the resulting sum rule is then generically of the form @xmath140 where @xmath141 is any function of @xmath130 analytic in the region of integration , and @xmath142 denotes the circular part of the contour , traversed counterclockwise ( from above to below the cut ) .
the ope is then to be used on @xmath142 ( see also the discussion to follow ) , while spectral data and/or a spectral anstz is to be employed in the `` hadronic '' integral on the rhs .
such sum rules are called , generically , finite energy sum rules ( fesr s ) , and have usually been employed with integer - power weights ( @xmath143 , with @xmath144)@xcite , though the standard theoretical treatment of hadronic @xmath13 decays , which is of the fesr type , involves a more complicated weight , as determined by kinematics@xcite .
it is useful , for the discussion which follows , to recall at this point the distinction between `` local '' and `` semi - local '' duality . for a typical hadronic correlator , @xmath129 , as above
, the ope is expected to be reliable , not only for @xmath130 large and spacelike , but also for those @xmath130 on any circle of sufficiently large radius in the complex @xmath130-plane , apart possibly from some region whose size is hadronic in scale about the timelike real axis ( where the effects of confinement are expected to be important)@xcite . the term `` local duality '' , in this context , refers to the postulate ( underlying , for example , all statements giving the number of subtractions required to make a given dispersion relation convergent in qcd ) that , once one is in a region @xmath145 for which the separation between adjacent resonances is small compared to the typical resonance width , the region of validity of the ope on the circle of radius @xmath139 extends all the way down to the real timelike axis .
this is equivalent to saying that the hadronic spectral function can then be calculated using the ope .
this is what allows , for example , the hadroproduction ratio , @xmath146 , in @xmath147 scattering to be computed theoretically in the asymptotic region using perturbative methods .
similarly , `` semi - local duality '' refers to the idea that , at somewhat lower ( `` intermediate '' ) scales , where local duality is no longer valid , nonetheless , averaged over some range of ( timelike ) momenta , the mean values given by using either the actual hadronic spectral function or the ope version thereof should be the same .
it is important to understand that , empirically , the condition that resonance spacing be _ much smaller _ than typical resonance widths is crucial to the validity of local duality .
indeed , one can test local duality through the application of various fesr s in the case of the isovector vector channel , for which the hadronic spectral function is very accurately measured in hadronic @xmath13 decays@xcite .
one finds that , even at scales as large as @xmath148 gev@xmath40 , and even though the experimental spectral function appears rather featureless in this region , nonetheless , local duality is rather poorly satisfied@xcite .
( note that in this channel , at these scales , the separation of resonances is comparable to their widths . ) in light of the above discussion , we distinguish three ( timelike ) kinematic regimes : `` low '' ( the region of narrow , well - isolated resonances ) , `` intermediate '' ( the region of the validity of semi - local duality ) and `` high '' ( the region of the validity of local duality , for which the hadronic spectral function can be reliably obtained using pqcd / ope methods ) .
the distinction is important because of the ways , described above , in which sum rules are practically implemented .
typically , one attempts to use known information , contained in the ope ( i.e. , @xmath149 , and the values of various vacuum condensates ) , to place constraints on the parameters of the ( hopefully , physically - motivated ) anstz for the corresponding unknown ( or not fully - known ) spectral function .
this attempt will , of course , be successful only if ( 1 ) one is able to work at scales for which the ope , truncated at operators of relatively low dimension , and at a given perturbative order for the wilson coefficients of these operators , is well - converged , and ( 2 ) one has a _ qualitative _ form for the hadronic spectral anstz which is both physically plausible and does not involve a large number of unknown parameters .
the latter condition can , realistically , be satisfied only if the scale up to which one needs the spectral anstz is not too far into the intermediate region .
once the higher resonances start to overlap , and various multiparticle background processes start to become important , the general form of the spectral function will become increasingly difficult to anticipate in advance ( until , that is , one is at sufficiently large scales that local duality finally becomes valid ) . in the svz version of qcd sum rules , as explained above , it is only rarely possible to choose the stability window for the analysis in such a way as to avoid contributions from the region where either one does not know the qualitative form of the spectral function , or where , even if one does , to implement it fully would involve the use of more free parameters than could be determined reliably , given the limited amount of information available in the truncated ope@xcite .
conventionally , this problem is dealt with by employing a spectral anstz in which ( 1 ) the low-@xmath138 region is assumed to be dominated by one or two low - lying resonance contributions and ( 2 ) the intermediate- and high-@xmath138 region is approximated using the local duality version of the spectral function , which one assumes to start at some `` continuum threshold '' , @xmath139 .
it is well - known that this form of `` continuum anstz '' represents a rather crude approximation , and hence can create significant uncertainties in the analysis if the contributions from the continuum part of the spectral integral are large for @xmath136 values in the stability window .
typically , one attempts to minimize this problem by avoiding @xmath136 values for which this is the case , but the cost of doing so is poorer convergence on the ope side of the sum rule and hence increased uncertainties associated with neglect of higher dimension contributions . a similar problem exists for the integer - power weighted version of fesr s .
one advantage of the fesr approach is that , in contrast to the svz method , the choice of scale appearing on the ope side of the sum rule ( @xmath139 ) is not constrained by the requirement of working in a stability window .
one is , of course , still forced to choose @xmath139 not too far into the intermediate region , since otherwise the spectral anstz would be too complicated to allow a reliable analysis .
the possibility of working at such intermediate scales in the fesr approach can , nonetheless , represent a practical advantage over the svz approach in certain cases .
this is true of those channels for which the stability window of the svz analysis lies at relatively low @xmath136 ( _ e.g. _ , @xmath150 gev@xmath40 , as found for many applications in the literature ) . in such cases ,
the larger scale of the fesr analysis leads to an improvement of both the convergence by operator dimension and convergence by perturbative order of the wilson coefficients on the ope side of the sum rule .
unfortunately , this advantage is usually more than offset by an increase in the difficulties associated with the use of the local duality approximation in the intermediate region .
the reason is obvious : with integer - power weights , the region near the timelike real axis where the circular part of the contour joins the cut , does not have the exponential suppression present for the `` continuum '' contributions in the svz approach .
the problems that result can be quantified in the case of the isovector vector channel , where the hadronic spectral function is known experimentally . as shown in ref .
@xcite , the errors in integer - power weighted fesr s , even at scales as high as @xmath151 , can be very large , despite the fact that the ope at this scale is both dominated by the leading ( @xmath152 ) perturbative term , and rather rapidly converging .
( a more detailed discussion of why it is that one can not judge the degree of validity of the local duality approximation simply by looking at the degree of convergence of the ope in a given region is given in ref .
@xcite . )
these problems , encountered in employing the conventional , integer - power - weighted version of fesr s , are , however , not intrinsic to the fesr approach .
indeed , there is at least one example of a non - integer - power - weighted fesr that is very well - satisfied : that giving the hadronic @xmath13 decay widths in terms of an integral over the circle of radius @xmath153 of the product of the ope for the isovector vector current correlator and the weight function @xmath154 ( where the dominant input parameter in the ope representation is @xmath155 , which can be taken as obtained by running the value measured at the @xmath156 mass down to the @xmath13 scale ) .
the reason for the success of this sum rule is well - known : the juncture of the cut and circular portions of the fesr contour corresponds to the edge of hadronic phase space and hence , because of kinematics , the weight function @xmath157 has a ( double ) zero at @xmath158
. this zero suppresses contributions from the portion of the circular part of the contour near the real timelike axis for which the ope representation of the correlator is not reliable ( at intermediate scales like @xmath151)@xcite .
this immediately suggests that the appropriate way to implement fesr s in other channels is , not with integer - power weights , but rather with weights having a zero at @xmath159 . in ref .
@xcite it was shown that , in the isovector vector channel , where one can check the procedure explicitly , weight functions of either the form @xmath160 having a single zero at @xmath159 , or @xmath161 having a double zero at @xmath159 , produce extremely well - satisfied fesr s for a wide range of values of @xmath139 and the continuous parameter , @xmath116 .
in addition , using _ only _ the ope representation , for a range of @xmath116 and @xmath139 , and fitting the parameters of a sum - of - resonances anstz to this representation , results in a very good reconstruction of the hadronic spectral function , including a determination of the @xmath29 decay constant accurate to within a few @xmath162@xcite . in order to have a compact terminology for use in describing the weight families in eqs .
( [ singlepinch ] ) and ( [ doublepinch ] ) , we will refer to @xmath163 and @xmath164 as single - pinch and double - pinch weights , respectively .
the freedom to vary @xmath116 plays a role analogous to that of the variation of @xmath136 within the stability window in an svz - style analysis , in that it strengthens the sum rule constraints on the parameters of the spectral anstz by working with a range of different weight profiles .
an additional advantage , at least if one wishes to determine not just the parameters of the lowest resonance in the channel but also those associated with higher resonances , is that the weight function can actually be arranged to be larger in the second resonance region than in the first . in what follows , in light of its success in the isovector vector channel , we will employ the fesr framework , using both the single- and double - pinch weight families , to investigate the isospin - breaking vector current correlator , @xmath59 , defined above . as usual in the fesr framework , we will work at scales as high as possible , compatible with the constraint of having a tractable and physically sensible spectral anstz for @xmath165 .
since little is known about the vector meson resonance spectrum beyond the second excited resonance region , and since including even the second excited resonance region would lead to a spectral anstz with more parameters than are generally tractable for the present analysis , we are forced to work at scales no higher than @xmath166 gev@xmath40 . since the separation of the first and second excited vector meson resonance regions is comparable to the resonance widths ( the @xmath167 and @xmath168 lie at @xmath169 and @xmath170 mev , the @xmath171 and @xmath172 at @xmath173 mev and @xmath174 mev , respectively@xcite ) , it is clear that , at these scales , we are not yet in the region of the validity of local duality .
this makes the use of the single- and double - pinch families crucial to the reliability of the fesr analysis . in order to maintain as good convergence as possible on the ope side of the two sum rule families , while at the same time allowing enough variation in @xmath139 to get a good determination of the parameters of the spectral anstz , we also restrict our attention to scales , @xmath139 , greater than @xmath175 gev@xmath40 .
since the general framework to be employed in the analysis has been outlined in the previous section , it remains only to discuss the input required on the hadronic and ope sides of the various sum rules .
we begin with the hadronic side .
we take , as our anstz for the hadronic spectral function , a sum of resonance contributions . for the scales used in the analysis ,
the resonances present in the region of the hadronic spectral integral are the @xmath29 , @xmath30 , @xmath94 , @xmath167 and @xmath168 .
( although the tails of the @xmath171 and @xmath172 intrude slightly into the hadronic integration region for @xmath139 near @xmath176 gev@xmath40 , their contributions are strongly suppressed by the zeros in the weight functions .
we have checked that including an effective , combined @xmath171-@xmath172 contribution in the spectral anstz has negligible effect on the extracted @xmath29 , @xmath30 and @xmath94 spectral strength parameters . )
we thus include contributions , written in terms of breit - wigner resonance forms , for all these resonances . because the separation of the @xmath167 and @xmath168 is much smaller than either of their widths , and also to reduce the number of free parameters in the spectral anstz , we have combined the latter two contributions .
the strong overlap of the two resonances would , in any case , prevent one from being able to sensibly extract separate @xmath167 and @xmath168 strength parameters by means of any sum rule analysis of @xmath59 .
the spectral anstz then has the form @xmath177\ , \label{specansatz}\ ] ] where @xmath178 ( with @xmath179 the width of the resonance in question ) .
this expression reduces to @xmath180 in the narrow width approximation ( nwa ) . the minus sign in front of @xmath181 and the factor of @xmath182
are conventional ; inclusion of the former ensures that @xmath181 and @xmath183 become equal in the limit that the spectral contributions in the @xmath29-@xmath30 region are generated entirely by leading order @xmath29-@xmath30 mixing . for the combined @xmath167-@xmath168 contribution we have taken average values for the effective mass and width .
@xmath183 , @xmath181 , @xmath184 and @xmath185 are free parameters , to be determined from the matching of hadronic and ope sides of the single- and double - pinch sum rules , for a range of @xmath139 , @xmath116 values .
a few comments are in order concerning the form of the anstz above , and the physical meaning of the parameters to be extracted from the analysis which follows .
the first concerns the need for the inclusion of a @xmath94 contribution .
note that the correlator @xmath59 is very closely related to that , @xmath186 , obtained by dropping the strange part of the hypercharge current from @xmath59 ( the ope s for the two correlators are , in fact , identical to three - loop order ) .
the latter correlator has been studied in a number of earlier svz - style analyses@xcite . in the earliest of these ,
the nwa was employed for all resonances , and no @xmath94 contribution was included in the spectral anstz@xcite . as pointed out in ref .
@xcite , however , the existence of significant cancellations between the nwa @xmath29 and @xmath30 contributions ( which would be exact in the limit of mixing dominance and equality of @xmath29 and @xmath30 masses ) means that a @xmath94 contribution , even if significantly smaller than the _ individual _ @xmath29 and @xmath30 contributions , could nonetheless be important . performing the sum rule analysis with a @xmath94 contribution
included shows that this is , indeed the case@xcite .
including the @xmath94 contribution in the spectral anstz also cured an unphysical feature of the solutions obtained earlier , which did not include it@xcite . the analysis of ref .
@xcite , however , still employed the nwa for all resonances .
the second point concerns the need to incorporate the @xmath29 width into the analysis . because , again , of the high degree of cancellation between the nwa @xmath29 and @xmath30 contributions , it was pointed out that the precise degree of this cancellation might well be sensitive to whether or not the difference between the @xmath29 and @xmath30 widths was retained in the spectral anstz@xcite .
the subsequent analysis of ref .
@xcite showed that this is , indeed , the case : the spectral parameters , @xmath187 , decrease by factors @xmath188 when one employs the physical widths in place of the nwa .
the third point concerns the interpretation of the higher resonance strength parameters , @xmath184 and @xmath185 .
it is , of course , very natural to take the spectral function to be resonance dominated .
moreover , the near - threshold region of the spectral function has been computed to two loops in chpt@xcite , and one can see from this result that the corresponding low-@xmath138 background contribution to the relevant spectral integrals is tiny compared to that from the @xmath29-@xmath30 region .
the case of the @xmath94 , however , is less clear , since background contributions above the @xmath29-@xmath30 region are not amenable to as reliable estimates , and hence might not be similarly negligible . in the anstz as written , such physical contributions , if present , could only be mocked up ( approximately ) by additional effective contributions to the @xmath94 and @xmath167-@xmath168 strengths .
thus , one must use some caution in interpreting , for example , the extracted @xmath184 in terms of the physical resonance parameters @xmath189 and @xmath190 some portion of @xmath184 could actually correspond to an averaged version of background contributions in the region between @xmath29-@xmath30 and @xmath167-@xmath168 .
the quality of the agreement between the hadronic and ope sides of our sum rules is , however , _
post facto _ evidence in favor of resonance dominance , and hence also in favor of the possibility of interpreting @xmath184 in terms of @xmath189 and @xmath190 .
let us turn , then , to the input on the ope side of the sum rules .
we will discuss the contributions , in turn , by operator dimension . since the correlator in question is isospin - breaking , the only dimension @xmath152 contribution to @xmath59 is electromagnetic ( we adhere , here , to common usage , according to which the leading mass - dependent perturbative terms are labelled @xmath191 ) .
we retain only the leading order ( 2-loop ) graph in this case .
the @xmath191 contributions are dominated by the strong interaction terms proportional to @xmath192 . to 3-loop order ,
the results for these terms follow from the 3-loop expressions for the correlator involving a flavor - non - diagonal current and its conjugate@xcite , since the perturbative contributions involving two quark loops and a purely gluonic intermediate state ( present for flavor diagonal currents but not for flavor - non - diagonal currents ) do not enter until 4-loop order .
the resulting expressions are given in the appendix .
to evaluate them , we require the running masses , @xmath193 , and running strong coupling , @xmath194 .
these can be obtained once the values are determined at any fixed scale , @xmath195 .
since the 4-loop @xmath196@xcite and @xmath197@xcite functions for qcd are now known , we have employed these when running the masses and coupling ( explicitly , we solve the rg equations exactly , using the truncated 4-loop @xmath196 and @xmath197 functions as input ) . as input for the running coupling ,
we take @xmath198 and use the latest ( 1998 ) value for @xmath199 obtained by the aleph collaboration in their analysis of non - strange hadronic @xmath13 decays@xcite .
( the analysis of the strange decays employed previous theoretical results for the @xmath191 terms , proportional to @xmath200 , which turn out to be in error@xcite ; the value obtained in the global aleph analysis must , therefore , be excluded . ) the situation with the light quark mass difference @xmath201 is somewhat more complicated .
we first write @xmath202 the isospin - breaking mass ratio , @xmath203 , is known , from a number of chpt analyses , to be @xmath204@xcite , which would allow one to determine @xmath205 if @xmath206 were known .
the most recent determination of @xmath206 is that based on an integer - power - weighted fesr analysis of the isovector pseudoscalar channel@xcite . in this analysis ,
the pion pole contribution to the spectral function is known experimentally but the continuum contribution is not .
the authors of refs .
@xcite , therefore , constructed an anstz for the unmeasured continuum contribution .
it turns out that the continuum portion of the resulting model spectral function provides roughly @xmath207 of the contribution to the extracted value of @xmath208 .
unfortunately , it has recently been shown , using the fesr framework discussed above , that this continuum anstz is unphysical@xcite , so one can not employ the values of refs .
@xcite . if @xmath209 were known ( at some scale )
, then one could straightforwardly determine @xmath206 ( at that same scale ) using the known ( scale - independent ) ratio of masses , @xmath210 , obtained by leutwyler@xcite using chpt .
unfortunately , the situation is also somewhat complicated for @xmath209 .
a number of recent analyses produce values of @xmath211 ( in the @xmath212 scheme ) ranging from @xmath213 mev to @xmath214 mev , often with rather large errors@xcite .
because the analyses based either on flavor breaking in hadronic @xmath13 decays@xcite or narison s @xmath13-decay - like sum rule for @xmath215@xcite have rather large errors resulting from experimental uncertainties which are unlikely to be significantly improved in the near future , the most favorable approach would appear to be that based on various sum rule treatments of the strange scalar channel , where the dominant @xmath216 part of the spectral function is in principle determined , via the omnes representation of the timelike scalar @xmath216 form factor , in terms of experimental @xmath216 phase shifts and @xmath217 data@xcite .
the most recent analyses of this channel@xcite employ the svz framework , and produce values @xmath218 mev@xcite , and @xmath219 mev@xcite .
( the same low-@xmath138 part of the spectral function is used in both analyses ; the only difference between the two lies in the treatment of the `` continuum '' .
the results of ref.@xcite , in addition , show no stability window for @xmath209 ) .
preliminary work using the fesr framework discussed above indicates that residual errors associated with the use of the local duality approximation in the continuum region remain , for this channel , when one uses the svz approach .
( see , e.g. , the results of ref . @xcite . from these one can see ( 1 ) that using the central values for the parameters describing the fit to the @xmath216 phases from refs .
@xcite , together with the central value from the @xmath209 range from ref .
@xcite , one obtains rather poorly satisfied families of fesr s , and ( 2 ) that using the spectral function of refs .
@xcite , again with central values for the fit parameters , the fesr analysis , in fact , produces @xmath209 values larger by @xmath220 mev than those obtained in the analysis of ref .
although work on the extraction of @xmath209 is still in progress@xcite , we conclude already from the preliminary results noted above , that @xmath221 mev , probably with errors @xmath222 mev or less .
the chpt ratio then produces @xmath223 mev , with errors @xmath224 mev . for any value in this range
it turns out that the @xmath191 contributions are at the @xmath225 ( or less ) level of the @xmath226 contributions , and the resulting errors lead , therefore , to very small ( @xmath162 level ) uncertainties in the final results .
since these uncertainties are much smaller than those generated by the uncertainty in the isospin - breaking mass ratio @xmath203 , we have employed the central value @xmath227 mev , and retained only the uncertainty in @xmath203 , in the analysis which follows .
the @xmath226 contributions are much more straightforward .
although in principle both those @xmath226 terms proportional to the isospin - breaking mass difference , @xmath205 , and those proportional to the isospin - breaking condensate difference , @xmath228 , appear in the ope of @xmath59 , the latter are numerically tiny compared to the former .
the dominant @xmath226 contribution can then be written in terms of @xmath203 and the combination @xmath229 , which we can take from the gmo relation @xmath230 the dominant uncertainties for the @xmath226 terms thus result from those on @xmath203 .
the phenomenological situation is not so favorable in the case of the @xmath1 condensates .
usually , in the absence of pre - existing determinations of the relevant condensates , one makes estimates based on the vacuum saturation approximation ( vsa ) .
it is well known that , in situations where it has been possible to perform phenomenological checks by extracting the total @xmath1 contribution from data , the vsa has proven to significantly underestimate these contributions@xcite . usually one simply replaces the factor @xmath231 , which is produced by the vsa , by an effective scale - independent factor , written @xmath232 .
the parameter , @xmath167 , then represents the deviation from the vsa .
ideally , it should either be possible to determine @xmath167 from data , or the @xmath1 contributions should be small , for the sum rule in question . in our case ,
neither of these conditions holds .
in particular , because we are forced to work at scales as low as @xmath175 gev@xmath40 in order to constrain the spectral parameters , the @xmath1 contributions can , for @xmath233 gev@xmath40 , and certain values of @xmath116 employed in our analysis , approach @xmath234 of the leading @xmath226 term . fortunately , it turns out , as we will see explicitly below , that by working with both the single- and double - pinch weight families , we can actually obtain a rather good determination of the @xmath1 contribution to the correlator ( albeit it as a function of @xmath203 ) by insisting on the consistency of the results obtained from the two different sum rule families . in the appendix
, it is shown that the vsa leads to an expression for the @xmath1 contribution to @xmath59 proportional to @xmath235 where @xmath236 is the average of the @xmath237 and @xmath238 condensates , and @xmath239 describes isospin - breaking in the light quark condensates . in order to compare the deviation from the vsa in the isospin - breaking channel with that in the analogous isospin - conserving isovector vector ( @xmath240 ) channel
, we write the re - scaled version of the rhs of eq.([vacsat ] ) in the form @xmath241 where @xmath29 is the parameter describing the deviation from the vsa in the @xmath26 channel , and one has , phenomenologically,@xcite @xmath242 with this definition , @xmath243 reduces to @xmath107 in the limit that the deviation from the vsa is the same in the @xmath26 and @xmath36 channels . the consistency procedure for fixing the @xmath1 contribution to @xmath59 , together with the phenomenological input of eq .
( [ value33rho ] ) , of course , determines only the product @xmath244 . in presenting our results for @xmath243 below ,
we have taken @xmath245 , which represents an average of the previous determinations listed in ref.@xcite , bar one .
( we omit the value based on an analysis of baryon splittings because it implies ( via the 1-loop chpt relations between flavor - breaking and isospin - breaking in the light quark condensate@xcite ) @xmath246 , which appears unphysical ) .
we will discuss the determination of @xmath243 in more detail below when we present the results of the analysis .
the last point in need of discussion concerns the way in which we handle the integrals on the ope side of the various fesr s .
two options exist in the literature . the first , sometimes called the `` fixed order expansion '' , involves first expanding @xmath194 and the mass factors , generically @xmath193 , in terms of @xmath247 and @xmath248 .
the coefficients of the perturbative expansions in powers of @xmath247 are then polynomials in @xmath249@xcite , and the desired contour integrals can thus be written in terms of elementary integrals involving logarithms and powers of @xmath138 .
the integrated ope expressions which result involve @xmath250 , multiplied by a power series in @xmath247 .
there is , of course , in this expression , the usual residual dependence on the choice of scale @xmath139 for the expansions discussed above , which results from truncating the full perturbative series at fixed order .
the second alternative , often referred to as `` contour improvement '' , involves numerically integrating the factors @xmath251^k \left[\alpha_s(q^2)\right]^j s^\ell$ ] around the circular contour in the @xmath252 plane@xcite .
it is known that this has the effect of simultaneously improving the convergence of the perturbative series and reducing the residual scale dependence@xcite . as a result
, we have evaluated all the integrals on the ope sides of our sum rules using this approach .
let us now turn to the results , which are presented in the table . as explained above
, the dominant uncertainty is due to that in the chpt determination of @xmath203 .
we have , therefore , tabulated the results for the range of values corresponding to the errors on @xmath203 quoted by leutwyler@xcite .
all results are based on matching the hadronic and ope sides of the two sum rule families for @xmath139 in the range @xmath253 to @xmath176 gev@xmath40 , and with @xmath116 in the range @xmath175 to @xmath254 for the single - pinch case and @xmath41 to @xmath255 in the double - pinch case .
the choice of range of @xmath116 in each case has been made so as to keep the convergence of the perturbative series for the @xmath191 term under control .
it is worth mentioning that the quality of the match between the ope and hadronic sides which results after the fitting of the spectral parameters is significantly better for leutwyler s central value of @xmath203 .
the value of the @xmath1 vsa - violating parameter , @xmath243 , given in the table , is determined by requiring that the valueds of @xmath183 obtained using the single- and double - pinch weight families are the same .
the sensitivity of @xmath183 to variations in @xmath243 ( true also of the other @xmath187 ) , as well as the difference in the @xmath243 dependence of @xmath183 for the single- and double - pinch analyses , is shown in figure 2 .
the fact that , once @xmath243 has been determined by the requirement of the consistency of the two output @xmath183 values , all the rest of the spectral parameters , determined using either the single- or double - pinch weights , also become consistent is strong evidence in favor of the reliability of the analysis .
note that ( 1 ) the possibility of determining the correction to the vsa for the @xmath1 operators , and ( 2 ) the inclusion of both the @xmath191 terms and the @xmath256 contributions to the wilson coefficient of the @xmath226 term , are features not present in previous analyses of the analogous isospin - breaking @xmath186 correlator .
although the value of @xmath243 , determined as just described , depends somewhat on @xmath203 , this dependence is not strong , and we obtain @xmath257 .
the first error corresponds to that in eq .
( [ value33rho ] ) , the second to that on @xmath203 .
we see that the violation of the vsa is very similar in both the @xmath26 and @xmath36 channels .
the importance , in reducing the errors on the determinations of the spectral parameters , @xmath187 , of being able to determine @xmath243 is also evident from fig . 2 .
having determined the @xmath1 contributions by self - consistency , the errors on the extracted values of @xmath187 are determined solely by those on @xmath203 , and are @xmath258 , completely correlated with @xmath203 .
having extracted the parameters @xmath187 , it is straightforward to determine the isospin violating decay constants .
one finds @xmath259 where the errors reflect those on the input isospin - breaking mass ratio , @xmath203 .
a number of useful general observations follow from the analysis above . first
, we have found that the violation of the vsa for the @xmath1 condensates is very similar in the isospin - breaking ( @xmath36 ) and isospin - conserving ( @xmath26 ) vector current channels .
second , we have seen that , although the analysis can be performed successfully for any @xmath203 in the range given by leutwyler , the central value of that range is preferred , in the sense of giving the best match between ope and hadronic sides of both the single- and double - pinch families of sum rules .
finally , we have demonstrated that the fesr method , particularly when implemented using both the single- and double - pinch weight families , is very effective , allowing a determination of the @xmath25 spectral parameters , @xmath187 , with rather small errors .
these errors ( between @xmath260 and @xmath261 for @xmath183 , @xmath181 and @xmath184 ) are a factor of @xmath41 smaller than those obtained in the earlier analysis of ref .
@xcite based on results of an svz analysis of @xmath186@xcite . in order to obtain this level of reduction , the ability to self - consistently determine @xmath243 was crucial .
let us now turn to the phenomenological consequences of our results .
first note that the corrections required to convert the measured contribution of the vector meson , @xmath262 , to the em spectral function , @xmath32 , into the corresponding contribution to either @xmath22 ( for @xmath263 ) or @xmath23 ( for @xmath264 ) are given by the ratios @xmath265 ^ 2&= & 0.982\pm 0.0021 \nonumber \\ \left [ \frac{f^{8}_\omega}{\sqrt{3}\ , f^{em}_\omega}\right]^2&= & 1.154\pm 0.017 \nonumber \\ \left [ \frac{f^{8}_\phi}{\sqrt{3}\ , f^{em}_\phi}\right]^2&= & 1.009\pm 0.001\ , \label{corrections}\end{aligned}\ ] ] where the numerical values follow from those in eq .
( [ decayconst ] ) .
the size of the deviations of the @xmath29 and @xmath30 corrections from @xmath107 are reduced by @xmath266 from those obtained in the earlier analysis @xcite ; that for the @xmath94 is increased , but remains small . in all cases
the errors have been reduced by a factor of @xmath41 or more .
note that the first of these corrections is the one relevant to precision tests of cvc .
note also that , as claimed above , the corrections given in eqs .
( [ corrections ] ) , for both the @xmath29 and @xmath30 , lie in the corresponding ranges produced by the estimate of section ii . with the results given in eq .
( [ corrections ] ) , it is now possible to correct the em data used as input to the inverse moment chiral sum rule for the @xmath2 order lec , @xmath3 .
the sum rule is given by@xcite @xmath267\ , \label{invchmom}\end{aligned}\ ] ] where @xmath268 is the renormalization scale of the effective chiral theory and @xmath269 are the usual renormalized @xmath270 order lec s of gasser and leutwyler@xcite .
this sum rule was evaluated in ref .
@xcite using as input em data for the isoscalar spectral function and both em and @xmath13 decay data for the isovector spectral function .
the corrections above , required for the em data , were not considered in this analysis .
it is not clear , from our reading of the discussion of ref .
@xcite , exactly what the relative weightings of @xmath13 decay and em data in the determination of the @xmath271 contribution to the lhs above actually were .
since , however , @xmath13 decay data is considerably more precise than electroproduction data , we have assumed in what follows that the determination is dominated by @xmath13 decay data .
to the extent that this is true , we need only make corrections to the ( nominally ) isoscalar @xmath30 and @xmath94 contributions .
the result of this exercise is a shift of @xmath272 from @xmath273 to @xmath274 ( for reference , making , instead , the somewhat perverse assumption that the determination of the @xmath271 contribution was dominated by em data , one would find @xmath275 .
the full correction is dominated by that to the @xmath30 contribution .
the reason this correction is so much larger than the others has been discussed above . )
one should bear in mind , in interpreting these results , that ( 1 ) there are , in principle , additional corrections to be made to the nominal isovector and isoscalar contributions at higher @xmath138 , and ( 2 ) the @xmath276 contributions were taken to be purely isoscalar in the analysis of ref . @xcite . because the separations within the higher isovector and isoscalar resonance pairs , @xmath167-@xmath168 and @xmath171-@xmath172 , are much smaller than the resonance widths , it is not possible to use sum rule methods to extract the individual isospin - breaking decay constants of these resonances .
as such , we are unable to estimate the size of the former corrections . were the @xmath276 states to have a significant isovector component , the effect would be to raise @xmath272 . an alternate method for determining @xmath3
is based on the observation that @xmath3 occurs not only in the inverse chiral moment sum rule above , but also in the 2-loop chpt expression for @xmath277@xcite .
it is , thus , possible to make an independent estimate by using the fitted spectral anstz for @xmath59 to compute @xmath277 , assuming negligible contribution from the portion of the spectrum above @xmath176 gev@xmath40 .
one obtains , from this exercise , @xmath278 since there is no positivity constraint on @xmath279 one does not know in which direction this result would be changed by corrections due to the small higher-@xmath138 part of the spectral integral .
the results of a study of the effect of including two combined spectral contributions , one for the @xmath280-@xmath281 and one for the @xmath171-@xmath172 , however , shows negligible change in @xmath272 , suggesting that such corrections are unlikely to be numerically significant . since the two independent determinations of @xmath3 are completely consistent , within errors , the conclusions that @xmath282 is considerably strengthened .
the last phenomenological application of our results concerns the effect on narison s @xmath13-decay - like sum rule for @xmath209 . since a detailed discussion of the way in which one implements the isospin - breaking corrections is given in ref .
@xcite , we report here only the results of employing the improved determinations of the correction factors determined above . in doing
so we will also take the opportunity to update the input parameters to the analysis of ref .
@xcite , employing the newer ( 1998 ) aleph value of @xmath199@xcite .
one finds , for example , using @xmath13 decay data for the isovector input , that the average over the values of @xmath283 extracted using the set of scales @xmath284 , @xmath285 and @xmath286 gev@xmath40 in the sum rule analysis is shifted from @xmath287 mev to @xmath288 mev ( @xmath289 mev if one retains the 1997 aleph value of @xmath290 as input ) .
unfortunately , the errors on this value associated with uncertainties in the experimental input are still very large , @xmath291 mev at least , and this uncertainty can not be appreciably reduced without a significant improvement in the accuracy of the determination of the experimental @xmath292 and @xmath293 widths . as such , although the central value is brought into better agreement with that discussed above , little more can learned from the narison sum rule , at the present time .
km would like to acknowledge the ongoing support of the natural sciences and engineering research council of canada , and to thank terry goldman , andreas hcker , heath oconnell , derek leinweber , tony thomas , and tony williams for useful discussions .
the explicit form of the ope for @xmath59 , keeping terms only up to dimension six , and to @xmath294 , can be obtained , as explained in the text , from the relevant expressions for the flavor - non - diagonal case given in the literature ( see ref .
@xcite and the paper by braaten , narison and pich ( bnp ) in ref .
we list the results by operator dimension .
.15 in dimension 0:.075 in the only isospin - breaking contribution at dimension 0 is that due to em , and is given by@xcite @xmath295_{d=0 } = -{\frac{\alpha } { 16\pi^3}}{\frac{1}{4\sqrt{3}}}ln(q^2)\ ] ] where @xmath296 is the usual em coupling .
.15 in dimension 2:.075 in the @xmath191 term consists of the leading mass - dependent part of the perturbative contribution to the ope , and follows from the expression given in ref .
one finds @xmath297_{d=2}&= & { \frac{3}{2\pi^2}}{\frac{1}{4\sqrt{3 } } } \left[(m_d^2-m_u^2)(q^2)\right ] \left [ 1+{\frac{8}{3 } } a(q^2)\right .
\nonumber \\ & & \qquad \left .
+ \left ( { 17981\over 432}+{62\over 27 } \zeta(3 ) -{1045\over 54 } \zeta(5)\right ) a^2(q^2)\right ] \label{d2}\end{aligned}\ ] ] where @xmath298 , and @xmath299 is the riemann zeta function .
further details on how the running of the coupling and the masses is handled can be found in section iv .
.15 in dimension 4:.075 in our expression for the @xmath226 contribution also follows from that given in ref .
only the @xmath300 and quark condensate terms survive once one takes the relevant isospin - breaking difference .
the former are numerically tiny compared to the latter , and hence have not been written down explicitly .
we then find @xmath301_{d=4 } = { \frac{2 \left(\langle m_u\bar{u}u\rangle - \langle m_d\bar{d}d\rangle\right ) } { 4\sqrt{3 } } } \left [ 1+{1\over 3}a(q^2)+{11\over 2}a^2(q^2)\right ] \ .\ ] ] the scale - invariant @xmath302 difference can be written in terms of the @xmath303 , @xmath304 , and the averages of the @xmath237 and @xmath238 quark masses and condensates .
since isospin - breaking in the condensates is much smaller than in the masses , the term proportional to @xmath303 dominates numerically .
it can be recast in terms of the isospin - breaking quark mass ratio , @xmath203 , and @xmath305 , @xmath306 , as explained in the text .
.15 in dimension 6:.075 in the 4-quark operators are the dominant operators at dimension 6 .
their contribution to @xmath59 can be obtained from the expressions given in the appendix of bnp@xcite . since lack of phenomenological information on the various condensates forces one to work with the re - scaled version of the vsa
, one must , for consistency , drop the terms of @xmath307 contained there .
( see the discussion of this point contained in bnp . )
one then finds @xmath308_{d=6 } & = & -8\pi^2 a(q^2)\left(\langle\bar{u}\gamma_{\mu } \gamma_5 t^a u\bar{u}\gamma^{\mu}\gamma_5 t^a u\rangle -\langle \bar{d}\gamma_{\mu}\gamma_5
t^a d\bar{d}\gamma^{\mu}\gamma_5 t^a d\rangle \right)\\ & & -{\frac{16\pi^2}{9 } } a(q^2)\sum_k \left(\langle\bar{u}\gamma_{\mu } t^a u\bar{q}_k\gamma^{\mu}t^a q_k\rangle -\langle \bar{d}\gamma_{\mu}t^a d\bar{q}_k\gamma^{\mu}t^a q_k\rangle \right)\ , \end{aligned}\ ] ] where @xmath309 is an @xmath310 generator . implementing the re - scaled vsa
, this expression reduces to @xmath308_{d=6 } & = & { \frac{448\pi}{81}}\rho_{red}\gamma(\rho\alpha_s\langle\bar{q}q\rangle^2)\ , \end{aligned}\ ] ] where @xmath196 , @xmath29 and @xmath243 are as defined in the text .
.results for the @xmath1 vsa - violation parameter , @xmath243 , and the spectral strength parameters , @xmath311 , @xmath181 , @xmath184 and @xmath312 as a function of the isospin - breaking mass ratio , @xmath203 .
the first line , for each value of @xmath203 , corresponds to the results obtained using the single - pinch weight family , the second line to those obtained using the double - pinch family . [ cols="^,^,^,^,^,^",options="header " , ] e. braaten , phys .
lett . * 60 * , 1606 ( 1988 ) and phys .
rev . * d39 * , 1458 ( 1989 ) ; s. narison and a. pich , phys . lett . *
b211 * , 183 ( 1988 ) ; e. braaten , s. narison and a. pich , nucl . phys . *
b373 * , 581 ( 1992 ) ; f. le diberder and a. pich , phys . lett . *
b286 * , 147 ( 1992 ) and * b289 * , 165 ( 1992 ) ; s. narison and a. pich , phys . lett . *
b304 * , 359 ( 1993 ) ; a. pich , hep - ph/9704453 , to appear in heavy flavors ii , eds . a.j .
buras and m. lindner , world scientific .
s. narison , phys . lett . *
b358 * , 113 ( 1995 ) .
e. golowich and j. kambor , phys . rev .
* d53 * , 2651 ( 1996 ) . j. gasser and h. leutwyler , nucl
b250 * , 465 ( 1985 ) .
fearing and s. scherer , phys .
* d53 * , 315 ( 1996 ) .
e. golowich and j. kambor , nucl .
b447 * , 373 ( 1995 ) .
et al . _ , nucl .
b256 * , 365 ( 1985 ) and references cited therein .
a. hcker , private communication .
r. barate _
( the aleph collaboration ) , z. phys . *
c76 * , 15 ( 1997 ) .
k. maltman , phys . rev . *
d58 * , 014008 ( 1998 ) .
callen , s. coleman , j. wess and b. zumino , phys .
rev . * 177 * , 2247 ( 1969 ) .
e. jenkins and a.v .
manohar , phys .
* b255 * , 558 ( 1991 ) and * b259 * , 353 ( 1991 ) .
wise , phys .
rev . * d45 * , 2188 ( 1992 ) ; g. burdman and j.f .
donoghue , phys .
b280 * , 287 ( 1992 ) ; t .- m . yan _
et al . _ , phys .
rev . * d46 * , 1148 ( 1992 ) ; e. jenkins , a.v .
manohar and m.b .
wise , phys .
* 75 * , 2272 ( 1995 ) .
k. maltman , phys .
b351 * , 56 ( 1995 ) .
particle data group , eur .
j. * 3 * , 1 ( 1998 ) .
k. maltman , h.b .
oconnell and a.g .
williams , phys .
b376 * , 19 ( 1996 ) .
oconnell , a.w . thomas and a.g .
williams , nucl .
* a623 * , 559 ( 1997 ) .
renard , in springer tracts in modern physics * 63 * , 98 ( 1972 ) ( springer - verlag ) ; m. gourdin , l. stodolsky and f.m .
renard , phys . lett . *
30b * , 347 ( 1969 ) . h.b .
oconnell , b.c .
pearce , a.w .
thomas and a.g .
williams , phys .
b354 * , 14 ( 1995 ) .
s.a . coon and r.g .
barrett , phys .
rev . * c36 * , 2189 ( 1987 ) .
shifman , a.i .
vainshtein and v.i .
zakharov , nucl .
* b147 * , 385 , ( 1979 ) ; _ ibid .
_ , 448 ( 1979 ) .
reinders , h. rubinstein and s. yazaki , phys .
rep . * 127 * , 1 ( 1985 ) .
s. narison , _ qcd spectral sum rules _ ( world scientific , singapore , 1989 ) .
d. leinweber , ann .
( n.y . ) * 254 * , 328 ( 1997 ) .
r. shankar , phys . rev . * d15 * , 755 ( 1977 ) ; r.g .
moorhose , m.r . pennington and g.g .
ross , nucl .
b124 * , 285 ( 1977 ) ; k.g .
chetyrkin and n.v .
krasnikov , nucl . phys . *
b119 * , 174 ( 1977 ) ; k.g .
chetyrkin , n.v .
krasnikov and a.n .
tavkhelidze , phys . lett . *
76b * , 83 ( 1978 ) ; n.v .
krasnikov , a.a .
pivovarov and n.n .
tavkhelidze , z. phys . *
c19 * , 301 ( 1983 ) ; e.g. floratos , s. narison and e. de rafael , nucl . phys . *
b155 * , 115 ( 1979 ) ; r.a .
bertlmann , g. launer and e. de rafael , nucl . phys .
* b250 * , 61 ( 1985 ) . j. bijnens , j. prades and e. de rafael , phys . lett . *
b348 * , 226 ( 1995 ) .
e. poggio , h. quinn and s. weinberg , phys . rev . *
d13 * , 1958 ( 1976 ) .
k. maltman , `` constraints on hadronic spectral functions from continuous families of finite energy sum rules '' , york univ .
preprint , july 1998 , to appear phys . lett . * b*. m.a .
shifman , a.i .
vainshtein and v.i .
zahkarov , nucl . phys .
* b147 * , 519 ( 1979 ) . t. hatsuda , e.m .
henley , t. meissner and g. krein , phys .
* c49 * , 452 ( 1994 ) . k. maltman , phys . rev . * d53 * , 2563 ( 1996 ) .
iqbal , x .-
jin and d.b .
leinweber , phys . lett .
* b386 * , 55 ( 1996 ) . k. maltman , phys . rev . * d53 * , 2573 ( 1996 ) .
k.g . chetyrkin and a. kwiatkowski , z. phys . *
c59 * , 525 ( 1993 ) .
chetyrkin , phys .
b404 * , 161 ( 1997 ) .
t. van ritbergen , j.a.m .
vermaseren and s.a .
larin , phys .
lett . * b400 * , 379 ( 1997 ) .
r. barate _
et al . _ ( the aleph collaboration ) , eur .
j. * c4 * , 409 ( 1998 ) .
k. maltman , phys . rev . *
d58 * , 093015 ( 1998 ) .
chetyrkin , j.h .
kuhn and a.a .
pivovarov , hep - ph/9805335 .
a. pich and j. prades , j. high energy phys . * 06 * , 013 ( 1998 ) .
h. leutwyler , phys . lett . *
b374 * , 163 ( 1996 ) ; phys . lett . *
b378*,313 ( 1996 ) and hep - ph/9609467 j. prades , nucl .
. suppl . * 64 * , 253 ( 1998 ) .
m. jamin and m. mnz , z. phys .
* c66 * , 633 ( 1995 ) .
chetyrkin , d. pirjol and k. schilcher , phys . lett .
* b404 * , 337 ( 1997 ) .
p. colangelo , f. de fazio , g. nardulli and n. paver , phys .
b408 * , 340 ( 1997 ) .
s. narison , phys . lett . *
b358 * , 113 ( 1995 ) .
k. maltman , phys . lett . *
b428 * , 179 ( 1998 ) .
m. jamin , nucl .
. suppl . * 64 * , 250 ( 1998 ) .
k. maltman , r. gupta and t. bhattacharya , work in progress .
g. launer , s. narison and r. tarrach , z. phys . * c26 * , 433 ( 1984 ) ; y. chung _ et al .
_ z. phys . *
c25 * , 151 ( 1984 ) ; m. kremer , n.a . papadopoulos and k. schilcher , phys . lett . *
b143 * , 476 ( 1984 ) ; c.a .
dominguez and j. sola , z. phys .
* c40 * , 63 ( 1988 ) ; v. gimenez , j.a .
penarroch and j. bordes , phys . lett . *
b223 * , 245 ( 1989 ) .
s. narison , phys . lett . *
b361 * , 121 ( 1995 ) .
kniehl , z. phys . *
c72 * , 437 ( 1996 ) . f. le diberder and a. pich , phys
* b286 * , 147 ( 1992 ) and * b289 * , 165 ( 1992 ) .
e. golowich and j. kambor , phys .
rev . * d53 * , 2651 ( 1996 ) . | the isospin - breaking vector meson decay constants are determined from a qcd sum rule analysis of the vector current correlator @xmath0 , using a recently proposed implementation of the finite energy sum rule approach .
the analysis employs the three - loop version of the ope , and two different families of weight functions .
it is shown that the requirement of consistency between results obtained using these two different weight families leads to a rather good determination of the parameter describing the deviation of the @xmath1 condensate term in the ope from its vacuum saturation value , and that the ability to determine this value has non - trivial numerical consequences to the analysis .
the phenomenological relevance of the results to the extraction of the strange quark mass and the determination of the @xmath2 order chiral low energy constant , @xmath3 , is also briefly discussed . |
properties of exotic nuclei are not only interesting from the viewpoint of nuclear structure , but also have important implications in nuclear astrophysics , such as in nova and supernova explosions , x - ray bursts associated with explosive hydrogen burning , rapid proton capture processes , etc . in view of the recent rapid progress in the discovery of exotic nuclei driven by the advent of the new generation of radioactive ion beam facilities ,
an important problem is to understand how the shell structure evolves from stable to exotic nuclei , in which the role of the tensor force has received wide attention .
the effect of the tensor force on the shell evolution has been explored by otsuka and his collaborators within the shell model .
they found that the tensor force strongly affects the evolution of the single - particle energy level spacing and shell structure @xcite . in the case of skyrme interactions , a zero - range tensor component was actually present in the original papers of skyrme @xcite .
however , this component was omitted when realistic skyrme parameter sets were determined , and it has been systematically neglected for several decades in practical applications . until very recently , the tensor component was added to the existing skyrme parametrizations or included by fitting new parametrizations to investigate the contributions of the tensor force to the spin - orbit splitting @xcite , evolution of single - particle energies @xcite , shell evolution @xcite , stability of superheavy nuclei @xcite , binging energy @xcite , nuclear deformability @xcite , properties of excited states @xcite , and the spin and spin - isospin instabilities of nuclear matter @xcite .
the concept of pseudospin was supported by experimental observations about 40 years ago that the single - nucleon doublet levels with quantum numbers ( @xmath2 , @xmath3 , @xmath4 ) and ( @xmath5 , @xmath6 , @xmath7 ) lie very close in energy , where @xmath2 , @xmath3 , @xmath8 are the single nucleon radial , orbital , and total angular momentum quantum numbers , respectively .
this single - nucleon doublet pair can be relabelled as a pseudospin doublet : ( @xmath9 , @xmath10 , @xmath11 ) .
then the two states in the doublet are almost degenerate with respect to pseudospin .
this symmetry has been used to explain a number of phenomena in nuclear structure including the deformation @xcite , superdeformation @xcite , magnetic moment @xcite and identical bands @xcite .
it is also pointed out that the conservation of the pseudospin symmetry plays an essential role in stabilizing the neutron halo structure and generating the neutron shell effects when the neutron number approaches the neutron drip line @xcite .
the details about the pseudospin symmetry can be found in refs .
it is interesting to discuss the influence of the tensor force on the pseudospin symmetry . in this work
, we shall discuss the effect of the tensor component of the two - body effective interaction on the pseudospin energy splitting by employing a skyrme - hartree - fock approach , taking a long chain of sn isotopes as an example .
in addition , the effect of tensor interaction on shell correction energy will be investigated .
the liquid drop model explains well the gross features of nuclear fission and mass , however , there are systematic deviations from the smooth mass . the shell correction , which is a fluctuation in the binding energy , can be supplemented to the liquid - drop model to improve the description of the nuclear masses and fission barriers @xcite . for superheavy nuclei , as an important property , the shell correction energy is widely investigated because it is relevant to the stability of a superheavy nucleus .
the strutinsky method @xcite , which is widely used today , will be employed to extract the shell correction energy here . with
the single - particle spectra obtained from the skyrme - hartree - fock approach , the shell correction energies for heavy and superheavy nuclei can be calculated and one can analyze the effect of the tensor force .
this paper is organized as follows : in section ii , a brief theoretical method is presented .
the effects of the tensor interaction on the pseudospin energy splitting and shell correction are discussed in sections iii and iv , respectively .
finally , a brief summary is provided in section v.
the tensor part of the skyrme effective interaction is written as @xcite @xmath12 \delta ( % \mathbf{r}_{1}-\mathbf{r}_{2 } ) \notag \\ & & + \frac{t}{2}\left [ ( \mathbf{\sigma } _ { 1}\cdot \mathbf{k})(\mathbf{\sigma } % _ { 2}\cdot \mathbf{k})-\frac{1}{3}(\mathbf{\sigma } _ { 1}\cdot \mathbf{\sigma } % _ { 2})\mathbf{k}^{2}\right ] \delta ( \mathbf{r}_{1}-\mathbf{r}_{2 } ) \notag \\ & & + u\left [ ( \mathbf{\sigma } _ { 1}\cdot \mathbf{k}^{\prime } ) \delta ( \mathbf{r}% _ { 1}-\mathbf{r}_{2})(\mathbf{\sigma } _ { 2}\cdot \mathbf{k})\right ] \notag \\ & & -\frac{u}{3}(\mathbf{\sigma } _ { 1}\cdot \mathbf{\sigma } _ { 2})\left [ \mathbf{% k}^{\prime } \cdot \delta ( \mathbf{r}_{1}-\mathbf{r}_{2})\mathbf{k}\right ] , \end{aligned}\ ] ] where @xmath13 acts on the right and @xmath14 acts on the left . @xmath15 and
@xmath16 provide the intensity of the tensor force in even and odd states of relative motion , respectively .
the spin - orbit potential including the tensor contributions is given by @xcite @xmath17 where @xmath18 denotes the like ( unlike ) particles and the @xmath19 is the proton or neutron spin - orbit density given as @xmath20 r_{i}^{2}(r).\ ] ] where @xmath21 is the bcs occupation probability of each orbit .
the orbital with @xmath22 gives a positive contribution to @xmath19 while the orbital with @xmath23 gives a negative contribution to @xmath19 .
the strengths of these contributions @xmath24 and @xmath25 are expressed as @xmath26 and @xmath27 . @xmath28 and
@xmath29 are related to the central exchange part of the interaction and can be written in terms of the usual skyrme parameters @xmath30 @xmath31 and @xmath32 are related to the tensor part in the following way @xmath33 the central exchange and tensor contributions modify the energy density @xmath34 by @xmath35 one set of the parameters employed in the present work is the same as that in ref .
@xcite : @xmath36 mev @xmath37 and @xmath38 mev @xmath37 added perturbatively to the skyrme force sly5 @xcite ( marked by sly5+tf ) , which can fairly well explain the isospin - dependence of the experimental energy differences between the single - proton states outside the @xmath39 core for sn isotopes , and the single - neutron states outside the @xmath40 core for @xmath40 isotones both quantitatively and qualitatively . recently , zou _
et al_. analyzed the evolution of the spin - orbit splittings in the ca isotopes and @xmath41 isotones with this sly5+tf and they found that adding the tensor contribution can qualitatively explain in most cases the empirical trends @xcite . and @xmath42 single - proton states along the sn isotopes .
the experimental data are taken from ref .
for the proton - rich isotopes , one of or both of these two levels lie on the positive energy states ( unbound ) , so we do not present them here.,scaledwidth=50.0% ]
measured by @xmath43 for @xmath44 and by @xmath45 for @xmath46 in the sn isotopes.,scaledwidth=50.0% ] apart from affecting the spin symmetry , the tensor force influences the pseudospin symmetry .
the shift of single - particle levels might have different origins and mechanisms . in the present study , we focus on the effect of the tensor force .
the calculations are performed in a coordinate space using a box size of 20 fm and a mesh size 0.1 fm .
the pairing correlation is accounted in the bcs formalism with an energy gap determined by the observed odd - even mass differences for an open shell and being zero for a closed shell .
in addition to the sly5+tf , the parameter set t31 including the tensor force ( t31+tf ) is employed in the present study , which also explains the experimental energy differences between the @xmath47 and @xmath42 single - proton states along the sn isotopes , while the interaction t31 without the tensor component fails to reproduce the trend of the experimental data , as shown in fig .
1 . note that the calculations with the sly5+tf were firstly performed in ref .
@xcite using other pairing formalisms , yet their results coincide with our calculated ones , which is therefore consistent with the conclusion that the pairing effect is negligible for the spin - orbit splitting suggested in ref .
we display the proton pseudospin energy splitting of @xmath44 @xmath48 and @xmath46 @xmath49 of the sn isotopes in fig .
the results with the parameter sets sly5 ( sly5+tf ) and t31 ( t31+tf ) show the same trend .
the proton spin current does not contribute to isospin dependence of the spin - orbit potential for an isotope chain , however , the neutron spin current is uniquely responsible for the isospin dependence and affected greatly by the sign of the parameter @xmath32 . from @xmath50 to @xmath51 , @xmath52 is reduced as the @xmath42 neutron orbit is gradually filled . because of a positive value of @xmath32 , the absolute values of the proton spin - orbit splittings are enlarged . therefore , the energy level of the proton orbit @xmath53 is pushed up while that of the @xmath54 is pulled down , leading to the reduction of the pseudospin orbit splitting of @xmath44 .
when the neutron number @xmath55 changes from 66 to 70 , the filling of the @xmath56 neutron orbit also reduces the pseudospin orbit splitting of @xmath44 .
the pseudospin energy splitting of @xmath46 should increase in this region , nevertheless , the lower panel of fig .
2 shows an opposite trend .
the reason for this result is that the centroid of the @xmath57 levels shifts down more rapidly than that of the @xmath58 levels as the neutron number increases , which submerges the weak isospin - dependence of the tensor effect in this region .
moreover , from @xmath59 to @xmath60 , the @xmath47 orbit is gradually filled , which reduces the spin - orbit splitting , so that the @xmath53 and @xmath42 orbits shift downwards , but the @xmath54 and @xmath61 shift upwards .
accordingly , the pseudospin energy splitting of @xmath44 is enhanced , however , that of the @xmath46 is reduced as the neutron number increases .
although the proton spin current does not affect isospin dependence of the spin - orbit potential for a given isotope chain , being spin - unsaturated for protons , the filling of the proton orbits affects nucleonic spin - orbit splitting as compared to that without the tensor force .
the negative value of @xmath31 and positive value of @xmath62 enhance the absolute value of the spin - orbit potential and hence increase the spin - orbit splitting . from @xmath50 to @xmath63 ,
the tensor effects resulting from the proton spin current ( @xmath31 term ) and neutron spin current ( @xmath32 term ) affect the spin - orbit splitting in the same way .
consequently , the pseudospin energy splitting of @xmath44 is reduced and @xmath46 is enlarged evidently compared to those without the tensor force . from @xmath64 to @xmath60 , the @xmath31 and @xmath65 terms vary in an opposite way , and the contribution of the @xmath31 term is cancelled out gradually .
accordingly , the difference of the pseudospin energy splitting between that with and without the tensor force is reduced gradually . on the whole ,
compared to that without the tensor force , the pseudospin symmetry is recovered to a large extent for the pseudospin doublet of @xmath44 but is broken for @xmath46 .
the tensor interaction thus brings a distinct change in pseudospin symmetry . for neutrons ,
the role of the tensor force is more complicated due to the variation of neutron number , but the obvious alternations caused by the tensor component are distinguished , as shown in fig .
the tensor force leads to the energy inversion of some pseudospin doublets , such as @xmath48 when @xmath55 varies from @xmath50 to 60 as can be seen in fig .
3 . for the pseudospin doublets except ( n@xmath66 , n@xmath67 ) , such as ( 1@xmath68 , 1@xmath69 ) and ( 1@xmath70 , 1@xmath71 ) in the spin - unsaturated heavy nucleus ,
they are all composed of a @xmath22 orbit and a @xmath23 orbit .
accordingly , their pseudospin symmetry could be affected considerably by the tensor effect because the two orbits are shifted by the tensor component in an opposite direction . for the pseudospin doublets ( n@xmath66 , n@xmath67 )
, only the n@xmath67 orbit is shifted due to the tensor force ( the n@xmath66 orbit is not affected due to its zero orbit angular momentum ) , and hence the tensor force also affects these pseudospin splittings . our aim is to illustrate the significant influence of the tensor force on the pseudospin symmetry . because the pseudospin symmetry is affected by the tensor force , a number of phenomena @xcite being related to this symmetry can be accordingly affected .
in addition , the rotation bands of the deformed nuclei based on the single - particle levels should be influenced by the tensor force , which needs investigation but it is not going to be easy because of the breaking of the time reversal symmetry .
once the single - particle spectra are obtained , the shell correction energies can be calculated .
calculating such a correction requires the subtraction of the smoothing varyingpart of the sum of single - particle energies @xmath72 where @xmath73 is the neutron ( proton ) number , and @xmath74 are the single - particle energies .
the neutrons and protons are treated quite separately . the smoothed fermi level @xmath75
is obtained from the equation @xmath76 .
taking into account the @xmath77-fold degeneracies of the single particle orbits , the smoothed level density @xmath78 takes the form of @xmath79 where @xmath80 is the smoothing range in units of @xmath81 mev .
the folding function is usually taken as @xmath82 , where @xmath83 is an associated laguerre polynomial . here the energy cutoff of the single - particle spectra is selected to be 40 mev . for @xmath0sn.,scaledwidth=50.0% ]
it remains to determine the order of the associated laguerre polynomial @xmath84 and the smoothing range @xmath80 .
ideally , the calculated @xmath85 should be not dependent on the specific values in a broad range of reasonable values , namely , the so - called plateau condition @xmath86 in our calculations , it is found that @xmath87 is an optimal value just like that in ref .
@xcite , in which the associated laguerre polynomial takes the form of @xmath88 . fig .
4 presents the neutron and proton shell correction energies as a function of the smoothing range @xmath80 taking the nucleus @xmath0sn as an example , where the skyrme - hartree - fock equations are solved with box sizes of @xmath89 fm with the parameter sets sly5 and t31 .
one can find that the plateau condition is well fulfilled within the range @xmath90 for neutrons and @xmath91 for protons in the case of @xmath92 fm .
the plateau vanishes gradually as @xmath93 fm . and @xmath92 fm is large enough in our calculations to achieve the convergence for the total binding energy with an accuracy of @xmath94 . when @xmath92 fm , the calculated shell correction energies for @xmath0sn with the parameter sets sly5 and t31 in the cases with and without the tensor force , are presented in the insets of fig .
this shell correction shows a parameter dependence . without the tensor force ,
the shell correction energies for neutron ( proton ) are about -12.0 mev ( -7.0 mev ) and -10.3 mev ( -5.5 mev ) by adopting the parameter sets sly5 and t31 , respectively .
the tensor force shifts the single - particle levels , thereby the single - particle level density undergoes a corresponding change , so that the shell correction energy is modified .
it can be found that the shell correction energies for both neutrons and protons are reduced by the tensor force , the reason for which is partly because the tensor interaction makes the energy level density near the fermi surface become lower . in order to show a clearer change in the shell correction energy caused by the tensor force , we plot in fig .
5 the net effect of the tensor force on the shell correction energy @xmath95 , namely , the difference between the shell correction energies with and without the tensor force .
as can be seen , the @xmath95 presents a more broad range of @xmath96 from 1.5 to 3.5 to fulfill the plateau conditionwith @xmath92 fm . with @xmath97 fm , though the neutron and proton shell correction energies do not show the plateau as presented in fig .
4 , the net contribution of the tensor force to the shell correction energies shows the plateau , especially for protons and nearly provides the same result as that with @xmath92 fm .
this indicates the reliability of our calculations to a large extent .
the error caused by the box size can be cancelled out to a large extent when one computes the contributions from the tensor force .
one can find that due to the tensor force , the neutron shell correction energy is lowered by about 2.3 mev and 3.0 mev , and the proton shell correction energy is lowered by about 1.1 mev and 1.8 mev with the parameter sets sly5+tf and t31+tf , respectively .
in other words , the effect of the tensor force on the shell correction is considerable . by the way
, the shell correction energy at the saddle point turns out to be too important to be neglected @xcite , so the tensor force may play a non - negligible role in fission processes . for @xmath98114.,scaledwidth=50.0% ] 114 near the fermi surface .
the calculations are performed with ( sly5+tf and t31+tf ) and without ( sly5 and t31 ) the tensor interaction in the spin - orbit potential.,scaledwidth=50.0% ] now we turn to the shell correction for superheavy nuclei .
the existence of the superheavy island was predicted in 1960s . however , there is no consensus among theorists with regard to what should be the next doubly magic nucleus beyond @xmath99pb .
nearly all of modern calculations predict the existence of a closed neutron shell at @xmath100 . however , they differ in predicting the position of the closed proton shell .
modern calculations predict that they appear at @xmath101 , 120 , 124 , or 126 @xcite depending on the models employed .
we investigate the tensor effect on the shell correction of superheavy nuclei taking @xmath98114 as an example .
the influence of the tensor force on the neutron and proton shell correction energies ( @xmath102 fm ) is illustrated in fig .
6 . with the parameter sets used above
, we find that the shell correction energy for protons is altered by @xmath103 mev due to the inclusion of the tensor component , and is much larger than the @xmath104 mev for neutrons .
this implies that @xmath98114 becomes more stable due to the presence of the tensor force with the present parameter sets since the shell correction energy can reflect the stability of nuclei .
we plot in fig .
7 the single - particle levels near the fermi surface to show the effect of the tensor force ( @xmath105 fm ) .
as can be seen , the shell gap at @xmath106 is obviously enlarged due to the tensor force , and the shell gap at @xmath100 is changed slightly , which is consistent with the above discussion about the shell correction energy and is agreement with the results in ref .
the difference between the @xmath107 and @xmath108 neutron orbits determines the size of the gap at @xmath100 , yet the tensor force only directly affects the @xmath108 orbit . in consideration of the fact that the @xmath109 and @xmath110 terms in eq .
( [ aa ] ) take effect in an opposite way , their contributions to the spin - orbit splitting are cancelled out to a certain extent , hence the predicted neutron shell structures is modified only slightly by inclusion of the tensor component .
however , for @xmath106 , the two proton orbits @xmath111 and @xmath112 shift in an opposite direction due to the tensor component , thus the gap is changed obviously .
the discussions here aim at displaying the effect of the tensor force on the structure and stability of the superheavy nucleus .
the theoretical results at least imply that the tensor force should play an important role in the shell structure of the superheavy nucleus .
we would like to say that many - body approaches with various parameter sets tend to predict different locations of the shell gaps , which perhaps results from the uncompleted knowledge about the tensor force to a large extent .
a detailed investigation for the effect of the tensor force on the pseudospin energy splitting have been carried out taking sn isotope chain as an example in the framework of the skyrme - hartree - fock approach with the parameter sets sly5+tf and t31+tf combined with bcs method , where the sign of @xmath31 and @xmath32 plays a crucial role .
the contributions of the tensor force to the pseudospin symmetry are significant since all pseudospin doublets except ( n@xmath66 , n@xmath67 ) are composed of a @xmath113 orbit and a @xmath114 orbit which are affected in an opposite way by the tensor force . for ( n@xmath66 , n@xmath67 )
, only n@xmath67 orbit is shifted due to the tensor force , and hence the tensor force also affects the pseudospin splitting .
thus a lot of phenomena being related to the pseudospin symmetry should be accordingly affected .
in addition , the effect of the tensor force on the shell correction energy is calculated taking a magic nucleus @xmath0sn and a superheavy nucleus @xmath1 as examples .
it was shown that the influence of the tensor force on the shell correction energy is considerable .
for @xmath1 , the absolute values of calculated proton shell correction energy is enlarged due to the inclusion of the tensor component , which is consistent with the result that the shell gap at @xmath106 is obviously enhanced with the present parameter sets .
the modification of the shell correction energies by the tensor component are important for superheavy nuclei , which are related to their stability .
all the conclusions here actually originate from the shifts of the single - particle levels ( i.e. modified spin - orbit splittings ) on account of the presence of the tensor force .
j. m. dong would like to thank profs . w. h. long and s. g. zhou for their helpful discussions .
this work was supported by the national natural science foundation of china ( with grant nos .
10875151,10575119,10975190,10947109 ) , the major state basic research developing program of china under grant nos . 2007cb815003 and 2007cb815004 ,
the knowledge innovation project ( kjcx2-ew - n01 ) of chinese academy of sciences , cas / safea international partnership program for creative research teams ( cxtd - j2005 - 1 ) , the fundamental research funds for the central university under grant no .
lzujbky-2010 - 160 and the funds for creative research groups of china under grant no .
11021504 . c. l. bai , h. sagawa , h.q .
zhang , x. z. zhang , g. col , f. r. xu , phys .
b * 675 * , 28 ( 2009 ) ; c. l. bai , h. q. zhang , x. z. zhang , f. r. xu , h. sagawa , and g. col , phys .
c * 79 * , 041301(r ) ( 2009 ) ; c. l. bai , h. q. zhang , h. sagawa , x. z. zhang , g. col , and f. r. xu , phys .
lett . * 105 * , 072501 ( 2010 ) . | in the framework of a skyrme - hartree - fock approach combined with bcs method , the role of the tensor force on the pseudospin energy splitting for tin isotope chain is investigated .
the tensor force turns out to obviously affect the pseudospin energy splitting of the spin - unsaturated nuclei .
since the tensor force shifts the single - particle levels , it modifies the single - particle level density and the shell correction energy thereof .
the influence of the tensor interaction on shell correction energy is considerable according to our analysis taking a magic nucleus @xmath0sn as well as a superheavy nucleus @xmath1 as examples .
this modification of the shell correction energy due to the tensor component affects the stability of the superheavy nuclei . |
in ground - state density functional theory , the noninteracting kinetic energy ( ke ) @xmath7 of a system of n electrons may be treated as an exact functional of the occupied kohn - sham ( ks ) @xcite orbitals @xmath8 , and only the exchange - correlation energy has to be approximated . however finding an accurate orbital - free ke functional @xcite will simplify and speed up by orders of magnitude any ks self - consistent calculation . in general , a kinetic energy density ( ked ) is any function which integrates to the non - interacting kinetic energy @xmath7 : @xmath9=\int d{\mathbf{r}}\ ; \tau , \label{e355}\ ] ] where @xmath10 and @xmath11 are the electron spin densities and @xmath12,{\mathbf{r}})+\tau_\downarrow([{n_{\downarrow}}],{\mathbf{r}})$ ] . because of the local version of the spin - scaling relation @xcite @xmath13,{\mathbf{r}})=(1/2)\tau([n=2n_{\sigma}],{\mathbf{r } } ) , \label{e3551}\ ] ] we will only need to show our expressions for spin - unpolarized systems with @xmath14
. there are two important forms of the ked : one which depends on the laplacian of the kohn - sham orbitals and follows directly from the kohn - sham self - consistent equations : @xmath15 and another which is positive definite : @xmath16 where @xmath17 is the electronic density .
( we use atomic units , with @xmath18 ) .
@xmath19 has been proposed @xcite as the closest analog of a classical ked . while a generalized gradient approximation ( gga ) uses only @xmath1 and @xmath6 , a meta - gga ( mgga ) , such as the accurate nonempirical tpss @xcite density functional for the exchange - correlation energy ,
is constructed from local ingredients @xmath20 , @xmath21 , and @xmath22 . in this work
, we will present evidence that @xmath22 and @xmath23 can carry essentially the same information beyond that carried by @xmath20 and @xmath21 . to do so , and to make an improved semi - local density functional for @xmath24
, we shall construct a laplacian - level meta - gga for @xmath22 .
our semi - local functional recovers the fourth - order gradient expansion ( ge4 ) ked for a slowly - varying density and the von weizscker ked for a rapidly - varying density . the gradient expansion ( ge ) , which becomes exact for densities that vary slowly over space @xcite , is @xcite @xmath25 with @xmath26 where @xmath27 and @xmath28 are dimensionless derivatives of the density , @xmath29 is the thomas - fermi ked @xcite , and @xmath30 is the enhancement factor of the 2n - th term of the gradient expansion .
the zero - th order enhancement factor is @xcite @xmath31 , and the second - order one is @xcite @xmath32 .
the term linear in @xmath33 is a key ingredient in our mgga .
although this term integrates to zero in eq .
( [ e355 ] ) , it is @xcite important for the ked .
this term also indicates rapidly - varying density regimes ( e.g. , near a nucleus where @xmath34 and in the tail of the density where @xmath35 ) .
the fourth - order enhancement factor is @xcite @xmath36 this enhancement factor is a simplified expression obtained with green s theorem ( integration by parts in eq .
( [ e355 ] ) ) under the assumption that @xmath37 and its gradients vanish as @xmath38 .
a full expression for @xmath39 , involving derivatives of the density of higher than second order , is given in ref.@xcite .
although the integration by parts that leads to eq .
( [ ef4 ] ) is inappropriate for some non - analytic densities @xcite , we shall incorporate eq .
( [ ef4 ] ) into our mgga .
the sixth - order term @xcite , even if it provides a useful correction to the fourth - order gradient expansion for the formation energy of a monovacancy in jellium @xcite , diverges for atoms after the integration of eq .
( [ e355 ] ) , and requires higher derivatives of the density than we would like to use . for later use , we define @xmath40 the von weizscker expression @xcite ( @xmath41 ) is exact for any one- or two - electron ground state , is accurate in nearly iso - orbital regions , satisfies the exact nuclear cusp condition ( @xmath42 where @xmath43 is the nuclear charge ) @xcite , is exact in the @xmath38 asymptotic region ( where the density matrix behaves like eq .
( 8) of ref.@xcite ) , and gives a rigorous lower bound @xcite @xmath44 or @xmath45 .
the semi - local bound of eq .
( [ rlb ] ) is one of the most important constraints in the construction of our functional .
the von weizscker enhancement factor @xcite @xmath46 is simple .
but the von weizscker ke functional gives , in general , very poor approximate atomization kinetic energies ( see table iii of ref.@xcite ) , and this has been attributed @xcite to its strong violation of eq .
( 8) of ref .
@xcite .
recently tao , perdew , staroverov and scuseria ( tpss ) @xcite have constructed a nonempirical meta - generalized gradient approximation for the exchange - correlation energy .
this functional , which satisfies as many exact constraints as a meta - gga can ( see table 1 of ref.@xcite , or ref.@xcite for a detailed explanation ) , provides a good overall description of atoms , molecules , solids and surfaces @xcite-@xcite .
we construct a laplacian - level tpss ( ll - tpss ) by replacing @xmath0 by @xmath47 in the tpss exchange - correlation energy per particle .
laplacian - level functionals for exchange and correlation have been advocated in ref .
@xcite , and proposed in ref .
recent interest in them is driven in part by the observation of ref .
@xcite that @xmath5 can be used along with @xmath1 to mimic some exact exchange - correlation energy densities .
the paper is organized as follows . in section [ sec2 ]
, we present the ked meta - gga functional . in section [ sec3 ] ,
we test our functional for several physical systems and models , and further explain its behavior . in section [ sec4 ] , we construct the ll - tpss exchange - correlation functional , and present numerical and analytic evidence that @xmath5 and @xmath0 carry essentially the same information beyond that carried by @xmath1 and @xmath6 . in section [ sec5 ] ,
we summarize our conclusions .
our meta - gga for the ked is an interpolation between a modified gradient expansion and the von weizscker expression .
there are many ways to satisfy the inequality of eq .
( [ rlb ] ) , so we will have to rely on empiricism to select one of them . first , we construct a modified fourth - order gradient expansion enhancement factor @xmath48 which has the following properties : \(1 ) for small @xmath49 and @xmath50 , @xmath51 so that it recovers the fourth - order gradient expansion for a slowly - varying density .
\(2 ) when @xmath52 , @xmath53 which is the correct limit for a uniform density perturbed by a small - amplitude , short - wavelength density wave @xcite .
when @xmath54 , @xmath55 , which is reasonable for other rapid density variations .
\(3 ) the modified enhancement factor of eq .
( [ fsge4 ] ) satisfies a uniform damping property : @xmath56 this is desirable because for large @xmath49 and @xmath50 spuriously large values of @xmath57 can arise from truncation of the gradient expansion .
we shall assume that the condition @xmath58 indicates the need for @xmath59 .
our meta - gga interpolates between @xmath60 and @xmath61 .
the smooth interpolating function is @xmath62 where @xmath63 and @xmath64 are parameters .
this function is plotted in fig .
[ functiong ] .
the meta - gga ked is defined as follows : @xmath65 where @xmath66 when @xmath67 , i.e. , when @xmath68 , eq .
( [ pr6 ] ) makes @xmath69 . when @xmath70 , i.e. , when @xmath71 , eq .
( [ pr6 ] ) makes @xmath72 . in between
, @xmath73 is an interpolation between @xmath74 and @xmath75 .
our positive meta - gga ked keeps all the correct features of @xmath76 , tends to be exact in iso - orbital regions , and satisfies the important constraint of eq .
( [ rlb ] ) : @xmath77 the meta - gga depends on the empirical parameters @xmath78 and @xmath79 .
these parameters were found numerically , by minimizing the following expression @xmath80 where m.a.r.e .
atoms " is the mean absolute relative error ( m.a.r.e . ) of the integrated kinetic energy of 50 atoms and ions , m.a.r.e .
clusters " is the m.a.r.e . of the ke of @xmath81 , @xmath82 , @xmath83 , @xmath84 , @xmath85 , @xmath86 , @xmath87 , @xmath88 , and @xmath89
neutral spherical jellium clusters ( with bulk parameter @xmath90 which corresponds to na ) and m.a.r.e .
ldm(n=8 ) `` is the m.a.r.e .
of the ke of n=8 jellium spheres for @xmath91 = 2 , 4 , and 6 , calculated in the liquid drop model ( ldm ) @xmath92 where @xmath93 is the fermi wavevector , @xmath94 is the radius of a sphere which contains on average one electron , @xmath1 is the bulk density , and @xmath95 is the surface ke .
the ' ' exact " ldm value is one computed with the exact @xmath95 . because the relative errors of surface kinetic energies are much larger than those of the atoms and spherical jellium clusters , we use the ldm approach for calculating the jellium surface ke errors ; ldm gives m.a.r.e .
comparable to that of atoms and clusters ( see table [ errors ] ) .
the densities and orbitals we use are analytic hartree - fock @xcite for atoms and ions , and numerical kohn - sham for clusters and surfaces ( with the local density approximation for the exchange - correlation potential ) . in the process of optimization
, we observed that the m.a.r.e .
decreases very slowly when the parameter @xmath96 increases , but large values of @xmath79 deteriorate the ked .
so , we chose the following set of parameters : @xmath97 and @xmath98 . as we can see in table [ errors ] , this set of parameters gives an accuracy close to ( but better than ) that of the fourth - order gradient expansion . .
mean absolute relative error ( m.a.r.e . ) of integrated kinetic energies of 50 atoms and ions , of neutral spherical jellium na clusters ( @xmath81 , @xmath82 , @xmath83 , @xmath84 , @xmath85 , @xmath86 , @xmath87 , @xmath88 , and @xmath89 ) and of jellium surfaces ( with @xmath99 , @xmath100 , and @xmath101 ) incorporated into the liquid drop model ( ldm ) for a jellium sphere with n=8 electrons ( see eq . ( [ ldm ] ) ) .
the atoms and ions are : h , he , @xmath102 , be , @xmath103 , li , @xmath104 , ne , @xmath105 , @xmath106 , @xmath107 , ar , @xmath108 , @xmath109 , @xmath110 , @xmath111 , @xmath112 , zn , @xmath113 , @xmath114 , @xmath115 , @xmath116 , kr , @xmath117 , @xmath118 , @xmath119 , @xmath120 , xe , @xmath121 , @xmath122 , @xmath123 , @xmath124 , @xmath125 , @xmath126 , @xmath127 , @xmath128 , @xmath129 , @xmath130 , @xmath131 , @xmath132 , @xmath133 , cu , @xmath134 , @xmath135 , @xmath136 , @xmath137 , @xmath138 , @xmath139 , @xmath140 and n. [ cols= " < , < , < , < , < , < " , ] [ molecules ]
the exchange - correlation meta - gga uses as ingredients the spin densities and their gradients , and the positive kohn - sham kinetic energy densities : @xmath141 where @xmath142 is the exchange energy per electron of a uniform , spin - unpolarized density n. the local density approximation ( lda ) is recovered for @xmath143 and @xmath144 , @xmath145 . our laplacian - level non - interacting ked functionals @xmath146 and @xmath147 can replace the exact kohn - sham kinetic energy densities @xmath148 and @xmath149 in eq .
( [ mgga1ll ] ) .
the resulting ll - mgga is an explicit density functional , while eq .
( [ mgga1ll ] ) is an implicit one .
we have constructed in this way the laplacian - level ( ll - tpss ) of the tao , perdew , staroverov and scuseria exchange - correlation meta - gga @xcite . like tpss ,
ll - tpss recovers the fourth - order gradient expansion @xcite for the exchange energy in the slowly - varying limit .
moreover , like tpss , ll - tpss has a finite exchange potential at a nucleus .
the need for an ingredient beyond @xmath1 and @xmath6 ( e.g. , @xmath5 ) to satisfy this exact constraint was emphasized in refs . @xcite and @xcite .
the nuclear cusp of an atom can be defined by @xmath34 and @xmath150 , so the constraint used in the construction of the tpss exchange enhancement factor @xcite @xmath151 where @xmath152 , becomes @xmath153 such constraints can be satisfied by a laplacian - level meta - gga , but not by a gga ( using only @xmath1 and @xmath6 ) . in fig .
[ fig1 ] we show the exchange enhancement factor @xmath154 as a function of the inhomogeneity parameter @xmath155 for several values of the reduced laplacian @xmath33 .
the enhancement factor interpolates in an orderly way between the exact slowly - varying limit ( for small @xmath156 and @xmath50 ) and a rapidly - varying limit ( for large @xmath156 ) , while satisfying eq.([q401 ] ) . versus reduced gradient @xmath155 , for several values of the reduced laplacian : @xmath157 . ] in figs .
[ fig2 ] and [ fig4 ] we compare the tpss and the ll - tpss enhancement factors ( for exchange and exchange - correlation ) for the zn atom , which was also studied in ref .
the ll - tpss exchange - correlation energies are close to the tpss values for this and other atoms .
( for the h atom , @xmath158 and @xmath159 ; for the other 49 atoms and ions , the m.a.r.e . of ll - tpss with respect to tpss is 0.00132 for exchange , 0.0098 for correlation , and 0.00133 for the combined exchange - correlation energy . ) and @xmath160 versus radial distance r for the zn atom .
@xmath161 a. u. , @xmath162 a.u .. note that @xmath163 is 0.65 for the @xmath164 and 2.26 for the @xmath165 electrons . ] , @xmath166 , @xmath167 and @xmath168 versus radial distance r for the zn atom .
@xmath169 a.u .
, @xmath170 a.u.,@xmath171 a.u .. note that @xmath163 is 0.65 for the @xmath164 and 2.26 for the @xmath165 electrons .
pbe is the non - empirical gga of ref.@xcite ] in fig .
[ fig5 ] we show that the ll - tpss exchange functional , like the tpss , shows a strong enhancement in the @xmath172 region of an atom but is elsewhere not so different from the second - order gradient expansion for exchange , as discussed in ref .
@xcite . versus radial distance r for the xe atom .
gea is the second - order gradient expansion for exchange .
@xmath173 a.u . and @xmath174 a.u .. ]
we have constructed a laplacian - level kinetic energy density ( ked ) meta - gga functional which depends on two empirical parameters .
these parameters control an interpolation between a modified fourth - order gradient expansion and the von weizscker expression .
this interpolation is designed to respect the exact constraint of eq .
( [ rlb ] ) . from our tests and results of section [ sec3 ]
, we conclude that the ked meta - gga functional is a successful approximation to the positive kohn - sham ked .
our functional uses a simplified expression for the fourth - order gradient expansion @xcite of @xmath0 .
we also tested the meta - gga with the full fourth - order terms @xcite , but did not find a considerable improvement in ked or integrated ke .
we have built a laplacian - level meta - gga for exchange and correlation ( ll - tpss ) which seems to imitate faithfully the tpss @xcite exchange - correlation meta - gga .
several constraints of the tpss meta - gga are exactly satisfied by the ll - tpss , but others , like @xmath175 for any one - electron system , are approximately satisfied .
it appears to us that @xmath5 and @xmath0 carry essentially the same information beyond that carried by @xmath1 and @xmath6 .
either @xmath5 or @xmath0 can be and are used to recover the fourth - order gradient expansion of the exchange energy in the slowly - varying limit and to make the exchange potential finite at a nucleus .
integration of our laplacian - level meta - gga for the kinetic energy density yields an orbital - free density functional for the kinetic energy that seems to improve upon the fourth - order gradient expansion , especially for rapidly- varying densities .
we have made all our tests for electron densities constructed from orbitals , and do not know what might be found from a selfconsistent solution of the orbital - free euler equation for the electron density .
it was argued in ref .
@xcite that ke functionals employing only @xmath1 and @xmath6 can not yield both accurate integrated energies and accurate functional derivatives .
we believe that accurate results could be expected in some rapidly - varying regions ( near nuclei and in density tails ) from ke functionals that employ @xmath1 , @xmath6 , and @xmath5 ; our specific expressions however may encounter a problem due to the sharper features in fig .
[ efactor ] ( also visible in fig .
[ fig1 ] ) .
it is generally believed that correct quantum density oscillations and shell - structure oscillations can only be found from a fully - nonlocal orbital - free density functional for the kinetic energy @xcite .
the authors thank l.m .
almeida for help with the surface code , and lisa pollack for early tests like those of fig .
[ lisa ] . this work was supported in part by the national science foundation under grants no .
dmr 01 - 35678 and no .
dmr 05 - 01588
. [ vo ] , phys .
@xmath176 , a1133 ( 1965 ) . for a review , see e.v . lude@xmath177a and v.v .
karasiev , in a celebration of the contributions of robert g. parr ( reviews of modern quantum chemistry , vol .
i ) , ed . k.d .
sen ( world scientific , new jersey , 2002 ) .
a @xmath178 , 397 ( 1979 ) . , j. phys .
g : nucl . phys .
@xmath179 , 455 ( 1991 ) .
@xmath180 , 146401 ( 2003 ) . ,
@xmath181 , 223002 ( 2006 ) .
a @xmath182 , 4601 ( 1997 ) .
cambridge phil .
@xmath183 , 542 ( 1926 ) , e. fermi , rend .
lizei @xmath184 , 602 ( 1927 ) .
jetp @xmath185 , 64 ( 1957 ) , d.a .
kirzhnitz , field theoretical methods in many - body systems " , pergamon , oxford , 1967 .
@xmath186 , 1 ( 1976 ) .
a @xmath187 , 4586 ( 1986 ) . , can .
@xmath188 , 1428 ( 1973 ) . , phys
b @xmath187 , 686 ( 1986 ) .
rev . a @xmath189 , 1862 ( 1981 ) .
, z. phys .
@xmath190 , 431 ( 1935 ) .
pure appl .
@xmath191 , 151 ( 1957 ) .
, j. chem .
@xmath192 , 2798 ( 1996 ) .
, j. phys .
b @xmath193 , 17 ( 1978 ) . , in preparation .
a @xmath194 , 052508 ( 2001 ) . , j. phys . a @xmath195 , 10607 ( 2003 ) .
b. @xmath196 , 075102 ( 2003 ) . ,
@xmath197 , 6898 ( 2004 ) . , j. chem
@xmath198 , 12129 ( 2003 ) .
j. quantum chem .
@xmath199 , 506 ( 2005 ) . , j. chem . phys .
@xmath200 , 114102 ( 2005 ) .
@xmath201 , 196403 ( 2005 ) .
, j. phys .
a @xmath202 , 11006 ( 2005 ) ; ibid .
@xmath202 , 11015 ( 2005 ) . ,
j. chem . phys .
@xmath203 , 044103 ( 2006 ) .
a @xmath204 , 012502 ( 2004 ) .
b @xmath205 , 205104 ( 2006 ) .
, phys . rev .
a @xmath206 , 3827 ( 1994 ) .
a @xmath188 , 3571 ( 1995 ) .
, phys . rev .
b @xmath207 , 081202 ( 2006 ) . , j. phys .
c @xmath208 , 1322 ( 1971 ) .
, atomic data nucl .
data tables @xmath209 , 177 ( 1974 ) . ,
phys . rev .
a @xmath210 , 2010 ( 1985 ) .
a @xmath194 , 062501 ( 2001 ) .
a @xmath211 , 5002 ( 1988 ) .
a @xmath212 , 3561 ( 1993 ) .
, j. chem .
@xmath213 , 1290 ( 1994 ) . , in recent developments and applications of modern density functional theory " , theoretical and computational chemistry , vol .
4 , pp . 295 ( 1996 ) .
@xmath214 , 3487 ( 1998 ) .
a @xmath215 , 052511 ( 2000 ) .
, z. naturforsch .
@xmath216 , 1176 ( 1972 ) . ,
density functional theory " , springer - verlag ( 1990 ) . , j. phys
. condens .
matter @xmath217 , 1239 ( 2000 ) .
a @xmath218 , 155 ( 1990 ) .
a @xmath211 , 3098 ( 1988 ) .
, in _ electronic structure of solids 91 _ , edited by p. ziesche and h. eschrig ( akademie verlag , berlin , 1991 ) . , phys .
b @xmath219 , 17402 ( 1996 ) . ,
@xmath220 , 3865 ( 1996 ) .
a @xmath194 , 052501 ( 2001 ) . for a review , see y.a .
wang and e.a .
carter , in theoretical methods in condensed phase chemistry ( theoretical methods in chemistry and physics ) , ed .
schwartz ( kluwer , dordrecht , 2000 ) .
, j. phys .
@xmath200 , 214106 ( 2005 ) . | we construct a laplacian - level meta - generalized gradient approximation ( meta - gga ) for the non - interacting ( kohn - sham orbital ) positive kinetic energy density @xmath0 of an electronic ground state of density @xmath1 .
this meta - gga is designed to recover the fourth - order gradient expansion @xmath2 in the appropiate slowly - varying limit and the von weizscker expression @xmath3 in the rapidly - varying limit .
it is constrained to satisfy the rigorous lower bound @xmath4 .
our meta - gga is typically a strong improvement over the gradient expansion of @xmath0 for atoms , spherical jellium clusters , jellium surfaces , the airy gas , hooke s atom , one - electron gaussian density , quasi - two dimensional electron gas , and nonuniformly - scaled hydrogen atom .
we also construct a laplacian - level meta - gga for exchange and correlation by employing our approximate @xmath0 in the tao , perdew , staroverov and scuseria ( tpss ) meta - gga density functional .
the laplacian - level tpss gives almost the same exchange - correlation enhancement factors and energies as the full tpss , suggesting that @xmath0 and @xmath5 carry about the same information beyond that carried by @xmath1 and @xmath6 .
our kinetic energy density integrates to an orbital - free kinetic energy functional that is about as accurate as the fourth - order gradient expansion for many real densities ( with noticeable improvement in molecular atomization energies ) , but considerably more accurate for rapidly - varying ones . |
we consider an _ array of size @xmath4 _ to be an infinite matrix @xmath5 of real - valued elements satisfying @xmath6 the array is called _ binary _ if @xmath7 takes values only in @xmath8 for all @xmath9 satisfying @xmath10 and @xmath11 , and is called _ ternary _ if @xmath7 takes values only in @xmath12 .
given integers @xmath13 and @xmath14 , the _ aperiodic autocorrelation _ of @xmath5 at displacement @xmath15 is defined to be @xmath16 we refer to an array of size @xmath17 as a _ sequence of length @xmath18 _ , abbreviating the array @xmath19 to @xmath20 and the aperiodic autocorrelation @xmath21 to @xmath22 .
binary arrays with small out - of - phase aperiodic autocorrelation have a wide range of applications in digital communications and storage systems , including radar @xcite and steganography @xcite .
ideally , we would like to find a binary array @xmath23 of size @xmath4 satisfying @xmath24 in which case , @xmath23 is called a _ barker array
_ @xcite .
however , it was recently shown by davis , jedwab , and smith @xcite that a ( truly two - dimensional ) barker array must have size @xmath25 .
( barker sequences of length @xmath18 , namely @xmath17 barker arrays , are known for @xmath26 , and any other barker sequence must have even length @xcite greater than @xmath27 @xcite . ) since the barker array criterion is too restrictive for array dimensions exceeding @xmath25 , it is natural to define a measure for the collective smallness of the aperiodic autocorrelation values of a binary array .
one such measure is the _ merit factor _ , which is defined for a binary array @xmath5 of size @xmath4 with @xmath28 to be @xmath29 ^ 2}.\ ] ] let @xmath30 denote the maximum value of @xmath31 taken over all @xmath32 binary arrays @xmath23 of size @xmath4 , and abbreviate @xmath33 to @xmath34 .
we note that the mean of @xmath35 , taken over all binary sequences @xmath23 of length @xmath18 , equals @xmath36 @xcite .
the argument of @xcite easily generalises to two dimensions : the mean of @xmath35 , taken over all @xmath32 binary arrays @xmath23 of size @xmath4 , equals @xmath37 .
it follows that @xmath38 , which asymptotically equals @xmath39 and provides a first benchmark result . a number of theoretical and computational results on @xmath34 are known ( see @xcite for a survey ) .
one line of research is to calculate @xmath34 for small values of @xmath18 . at present , @xmath34 has been calculated for all @xmath40 ( see ( * ? ? ?
* fig . 1 ) , for example ) .
the largest values of @xmath34 currently known are @xmath41 and @xmath42 , which are attained by barker sequences .
the computational analysis of @xmath34 quickly becomes infeasible as @xmath18 grows .
another line of research is therefore to construct particular infinite families of binary sequences of increasing length and to calculate their asymptotic merit factor . the only non - trivial infinite families of binary sequences for which the asymptotic value of the merit factor is known are rudin - shapiro sequences @xcite , legendre sequences @xcite , and @xmath43-sequences @xcite , together with some generalisations of these families @xcite , @xcite , @xcite , @xcite , @xcite .
the largest proven asymptotic merit factor of a binary sequence family is @xmath44 , which is attained by cyclically rotated legendre sequences ( see theorem [ thm : hj ] ) .
there is also considerable numerical evidence , though currently no proof , that an asymptotic merit factor greater than @xmath45 can be achieved for a family of binary sequences related to legendre sequences @xcite .
much less is known about the value of @xmath30 for @xmath46 .
eggers ( * ? ? ?
4.2 ) computed @xmath30 for @xmath47 and found lower bounds on @xmath30 for @xmath48 .
although the data supplied in @xcite are very limited , it is apparent that @xmath30 tends to be smaller than @xmath49 .
the largest value of @xmath30 for @xmath46 reported in @xcite equals @xmath50 .
however , an elementary construction technique gives binary arrays with larger merit factor .
given two sequences @xmath51 and @xmath52 of length @xmath18 and @xmath43 , respectively , we follow @xcite in defining the _ product array _ @xmath53 .
a straightforward calculation shows that @xmath54 from which we deduce @xmath55 let @xmath23 and @xmath56 be barker sequences of length @xmath57 and @xmath58 , respectively .
it follows from that @xmath59 , @xmath60 , and @xmath61 .
another consequence of is @xmath62 currently , no theoretical results on the asymptotic merit factor of families of binary truly two - dimensional arrays are known .
bmer and antweiler @xcite analysed the merit factor of several binary array families numerically . among the investigated families ,
two types of array families related to the quadratic character appeared to have largest merit factor .
the arrays in the first family were proposed by calabro and wolf @xcite and have size @xmath0 , and the arrays in the second family were proposed by bmer , antweiler , and schotten @xcite and have size @xmath63 , where @xmath1 and @xmath2 are ( not necessarily distinct ) odd primes .
both families can be considered as two - dimensional generalisations of legendre sequences .
the authors of @xcite successively applied three operations , namely rotations of rows and columns , stairlike rotations of rows and columns , and proper decimations , and computed the maximum value of the merit factor for arrays of small sizes taken from these families .
they then remarked @xcite that _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ `` , for large arrays , the acf [ aperiodic autocorrelation function ] merit factors of both classes [ the above mentioned array families of square size ] appear to tend to @xmath64 . '' , _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ and asked for a theoretical explanation of this observation .
in this paper we study the merit factor of two families of binary arrays .
the arrays in the first family , called _ legendre arrays _ , have size @xmath0 , where @xmath1 and @xmath2 are ( not necessarily distinct ) odd primes , and contain as a special case the arrays proposed by calabro and wolf @xcite .
the arrays in the second family , called _ quadratic - residue arrays _ , have size @xmath63 , where @xmath1 is an odd prime , and contain as a special case the arrays proposed by bmer , antweiler , and schotten @xcite .
we calculate , under certain conditions on the growth rate of @xmath1 relative to @xmath2 , the asymptotic merit factor at all rotations of rows and columns for both array families .
in particular , we show that for both families the asymptotic merit factor equals @xmath65 for an optimal rotation of rows and columns .
although we only maximise the merit factor with respect to the first operation considered in @xcite , namely rotations of rows and columns , this result does not support the conclusion of @xcite quoted above .
for all other ( non - optimal ) rotations of rows and columns , the asymptotic merit factor of quadratic - residue arrays is larger than that of square legendre arrays .
given an odd prime @xmath1 and a positive integer @xmath43 , let @xmath66 be the finite field containing @xmath67 elements .
whenever convenient , we treat integers after reduction modulo @xmath1 as elements in @xmath68 .
the _ quadratic character of @xmath66 _ is the function @xmath69 defined by @xmath70 this function is multiplicative : @xmath71 if @xmath72 , then @xmath73 is the _
legendre symbol_. a _ legendre sequence _
@xmath74 of prime length @xmath75 is defined by @xmath76 if the initial element in a legendre sequence is changed to zero , so that @xmath77 then we call @xmath78 a _ ternary legendre sequence_. in what follows , we present two families of binary arrays , which can be considered as two - dimensional generalisations of legendre sequences .
let @xmath1 and @xmath2 be two ( not necessarily distinct ) odd primes , and let @xmath79 be the set of ternary arrays @xmath80 of size @xmath0 satisfying @xmath81 then @xmath79 contains @xmath82 arrays , each having @xmath83 nonzero elements
. given ternary legendre sequences @xmath78 and @xmath84 of length @xmath1 and @xmath2 , respectively , we define a _ legendre array _
@xmath85 of size @xmath0 to be a binary array of size @xmath0 that can be written as @xmath86 for example , the array @xmath87 of size @xmath0 , given by @xmath88 is a legendre array of size @xmath0 .
this particular array was originally defined by calabro and wolf @xcite , and its merit factor properties were investigated numerically in @xcite . in the original paper @xcite
such an array was called a `` quadratic - residue array '' .
we use the term _ legendre array _ to distinguish it from our second family of binary arrays .
let @xmath1 be an odd prime , let @xmath89 be the quadratic character of @xmath90 , and let @xmath91 be a basis for @xmath90 over @xmath68 . following bmer , antweiler , and schotten @xcite , we define a _ quadratic - residue array _ @xmath92 of size @xmath63 to be a binary array of size @xmath63 satisfying @xmath93 the class of quadratic - residue arrays @xmath94 satisfying @xmath95 was defined by bmer , antweiler , and schotten @xcite , and its merit factor properties
were investigated numerically in @xcite . in our analysis
it will be convenient to change the leading element in a quadratic - residue array to zero .
accordingly , we define the _ ternary quadratic - residue array _ of size @xmath63 to be the ternary array @xmath96 of size @xmath63 given by @xmath97 next we define an operation acting on an array to produce a new array of the same size . given an array @xmath5 of size @xmath4 and real numbers @xmath98 and @xmath99 , the _ rotation _ @xmath100 is the array @xmath101 of size @xmath4 given by @xmath102 if @xmath23 is a sequence of length @xmath18 , we abbreviate @xmath103 to @xmath104 .
the asymptotic merit factor of a legendre sequence was calculated for all rotations by hholdt and jensen @xcite .
[ thm : hj ] let @xmath78 be the legendre sequence of prime length @xmath75 , and let @xmath98 be a real number satisfying @xmath105 .
then @xmath106 the constraint @xmath107 in theorem [ thm : hj ] is for notational convenience only since by definition @xmath104 is the same as @xmath108 for every sequence @xmath23 and all real @xmath98 .
the maximum asymptotic merit factor of a rotated legendre sequence @xmath109 is @xmath44 , which occurs for @xmath110 .
given a positive integer @xmath18 , let @xmath111 be a primitive @xmath112th root of unity .
let @xmath5 be an array of size @xmath4 .
the _ generating function _ of @xmath23 is defined to be the power series @xmath113 if @xmath23 is a sequence of length @xmath18 , we write @xmath114 for @xmath115 . the next lemma shows how the merit factor of @xmath23 can be computed from the values @xmath116 .
this approach generalises to two dimensions the method of hholdt and jensen @xcite to compute the asymptotic merit factor of a sequence .
[ lem : cq ] let @xmath23 be an array of size @xmath4 .
then @xmath117 ^ 2=\frac{1}{4nm}\sum_{i=0}^{2n-1}\sum_{j=0}^{2m-1}{\big\lverta({\zeta}_n^i,{\zeta}_m^j)\big\rvert}^4.\ ] ] straightforward manipulation shows that @xmath118 using this identity , an elementary calculation gives @xmath119 ^ 2,\end{aligned}\ ] ] as required .
in this section we compute the asymptotic merit factor of legendre arrays for all rotations , subject to certain conditions on the growth rate of the dimensions .
we first record a result on the aperiodic autocorrelation of rotated ternary legendre sequences , which arises as an immediate corollary of ( * ? ? ? * thm .
[ prop : flegendre ] let @xmath78 be the ternary legendre sequence of prime length @xmath75 , and let @xmath98 be a real number satisfying @xmath107 . then ,
as @xmath120 , @xmath121 ^ 2=\tfrac{7}{6}+8\left({\lverts\rvert}-\tfrac{1}{4}\right)^2+o(p^{-1}(\log p)^2).\ ] ] we note that , as explained after ( * ? ? ?
3 ) , we can recover theorem [ thm : hj ] from proposition [ prop : flegendre ] .
we also need the following bound for the magnitude of a polynomial over @xmath122 at a @xmath123th root of unity , in terms of its magnitudes at @xmath124th roots of unity .
[ lem : poly - bound ] let @xmath125 be odd , and let @xmath126 $ ] have degree at most @xmath127 . then @xmath128 the bound is trivial in the case that @xmath129 is even .
we may therefore take @xmath129 to be odd , writing @xmath130 for some integer @xmath131 so that @xmath132 .
it is then sufficient to bound @xmath133 .
now by lagrange interpolation we have @xmath134 and so , since @xmath124 is odd , @xmath135 the result follows from the inequality @xmath136 ( which holds since @xmath124 is odd , see @xcite , for example ) .
the next theorem gives , under certain conditions on the growth rate of @xmath1 relative to @xmath2 , the asymptotic merit factor of all @xmath82 legendre arrays of size @xmath0 for all rotations .
[ thm : farray1 ] let @xmath137 be an infinite set of products of two not necessarily distinct odd primes , and let @xmath138 take values only in @xmath137 . write @xmath139 for odd primes @xmath1 and @xmath2 , and suppose that @xmath140 let @xmath85 be a legendre array of size @xmath0 , and let @xmath98 and @xmath99 be real numbers satisfying @xmath141
. then @xmath142\left[\tfrac{7}{6}+8\left({\lvertt\rvert}-\tfrac{1}{4}\right)^2\right]-1 .
\label{eqn : f_legendre}\ ] ] let @xmath78 and @xmath84 be the ternary legendre sequences of length @xmath1 and @xmath2 , respectively , and write @xmath143 .
notice that @xmath144 and @xmath145 .
then from @xmath146 ^ 2=\sum_u[c_{l_s}(u)]^2\cdot \sum_v[c_{k_t}(v)]^2.\ ] ] the condition implies that @xmath1 and @xmath2 grow without bound as @xmath147
. we can therefore apply proposition [ prop : flegendre ] to give @xmath148 ^ 2=\left[\tfrac{7}{6}+8\left({\lverts\rvert}-\tfrac{1}{4}\right)^2\right]\left[\tfrac{7}{6}+8\left({\lvertt\rvert}-\tfrac{1}{4}\right)^2\right]+o(1)\quad\mbox{as $ n\to\infty$}. \label{eqn : ss_ternary_l_array}\ ] ] define @xmath149 ^ 2-\sum_{u , v}[c_{t_{s , t}}(u , v)]^2\bigg\rvert}. \label{eqn : def_delta}\ ] ] we claim that @xmath150 the theorem then follows from and and the fact @xmath151 .
it remains to prove the claim . by the definition of a legendre array
, there exists an array @xmath152 such that @xmath153 . from lemma [ lem : cq ]
we then have @xmath154 now for @xmath155 the identity @xmath156 ^ 2 + 2\,{\lverta\rvert}^2\cdot{\lvertb\rvert}^2 + 4\,{\lverta\rvert}^2\cdot\re(a\overline{b})+4\,{\lvertb\rvert}^2\cdot\re(a\overline{b})\ ] ] gives the inequality @xmath157 apply this bound to to obtain @xmath158 given a ternary legendre sequence @xmath23 of length @xmath124 , it is well known ( see @xcite , for example ) that @xmath159 for each integer @xmath160 .
it is easily verified that this implies @xmath161 for each integer @xmath160 and all real @xmath162 .
therefore , since @xmath145 , lemma [ lem : poly - bound ] gives @xmath163 substitute into to give @xmath164 where @xmath165 from and a straightforward calculation we obtain @xmath166 the cauchy
schwarz inequality gives @xmath167 , and since @xmath168 for @xmath169 , we also have @xmath170 for @xmath171 . substitution into gives @xmath172 now , using the condition , we readily verify our claim . there is no loss of generality in theorem [ thm : farray1 ] from the restriction @xmath141 since @xmath100 is the same as @xmath173 and @xmath174 for every array @xmath23 and all real @xmath98 and @xmath99
we note that the condition can be relaxed for particular legendre arrays .
for example , let @xmath78 and @xmath84 be the legendre sequences of length @xmath1 and @xmath2 , respectively .
then @xmath175 is a legendre array . from and theorem [ thm : hj ] we conclude that holds under the relaxed condition @xmath120 and @xmath176 as @xmath147 .
in this section our goal is to calculate the asymptotic merit factor of a quadratic - residue array of size @xmath63 at all rotations .
we shall assume throughout this section that @xmath1 is an odd prime .
write the @xmath1th roots of unity as @xmath177 then , since @xmath1 is odd , we have @xmath178 therefore , given an array @xmath23 of size @xmath63 , lemma [ lem : cq ] asserts that @xmath179 ^ 2\\ = \frac{1}{4p^2}\sum_{0\le i , j < p}\left({\lverta({\epsilon}_i,{\epsilon}_j)\rvert}^4+{\lverta(-{\epsilon}_i,{\epsilon}_j)\rvert}^4+{\lverta({\epsilon}_i,-{\epsilon}_j)\rvert}^4+{\lverta(-{\epsilon}_i,-{\epsilon}_j)\rvert}^4\right ) .
\label{eqn : cq}\end{gathered}\ ] ] our objective is to find an asymptotic expression for the sum on the right - hand side of the identity , where @xmath23 is a rotated ternary quadratic - residue array .
since a ternary quadratic - residue array and a quadratic - residue array differ in only one element , this will be sufficient to compute the asymptotic merit factor of a rotated quadratic - residue array .
before we analyse the sum in , we shall need several technical results , which we state in the next subsection .
the following lemma evaluates the generating function of a ternary quadratic - residue array at @xmath1th roots of unity .
[ lem : gauss_sum ] let @xmath89 be the quadratic character of @xmath90 , and let @xmath180 be a ternary quadratic - residue array of size @xmath63 , as defined in . then there exists a basis @xmath181 for @xmath90 over @xmath68 such that @xmath182 let @xmath183 be the trace function given by @xmath184 and for @xmath185 , let @xmath186 be the additive character of @xmath90 given by @xmath187 it is readily verified that @xmath188 we choose @xmath181 such that @xmath91 ( appearing in the definition of @xmath180 ) and @xmath181 are dual bases , that is , @xmath189 such a basis is guaranteed to exist @xcite . given integers @xmath190 , we then have by and @xmath191 and therefore @xmath192 by putting @xmath193 .
the above sum is called a _ gaussian sum _ , and it is well known that @xmath194 ( see @xcite , for example ) .
this proves the lemma .
our next lemma bounds a certain character sum and evaluates it in special cases .
let @xmath89 be the quadratic character of @xmath90 , and define @xmath195 and @xmath196 then @xmath197 and @xmath198 since @xmath89 is multiplicative by , @xmath199 ^ 2\\ & = \sum_{x\in\operatorname{gf}(p^2)\setminus\{0,-\kappa\}}1\end{aligned}\ ] ] using @xmath200 and @xmath201 ^ 2=1 $ ] for all nonzero @xmath202 .
this proves . to prove
, we use a special case of a result on multiplicative character sums with polynomial arguments ( * ? ? ?
* thm . 5.41 ) , which can be stated as follows .
if @xmath203 $ ] is a monic polynomial of positive degree @xmath124 that is not a square ( that is , @xmath204 can not be written as @xmath205 ^ 2 $ ] for some polynomial @xmath206 $ ] ) , then @xmath207 for all @xmath208 , the polynomial @xmath209 is not a square .
using the multiplicativity of @xmath89 , application of gives .
the following technical lemma can be obtained from the results of @xcite and @xcite .
[ lem : part_frac_exp ] define @xmath210 then @xmath211 \dfrac{p^2}{2 } \cdot \dfrac{1}{{\epsilon}_k\,{\lvert1-{\epsilon}_k\rvert}^2 } & \mbox{for $ k\not\equiv 0\pmod p$ } \end{cases } \label{eqn : gammakk0}\ ] ] and @xmath212 the identity was established in ( * ? ? ?
* , cases 4 and 5 ) .
the bound follows from the inequality @xmath213 ( see @xcite , for example ) since @xmath214 we are now in a position to analyse asymptotic behaviour of the sum on the right - hand side of the identity , where @xmath23 is a rotated ternary quadratic - residue array .
we split the analysis into the following three lemmas .
[ lem : qsum0 ] let @xmath180 be a ternary quadratic - residue array of size @xmath63 , and let @xmath98 and @xmath99 be real numbers satisfying @xmath141 . then
, as @xmath120 , @xmath215 lemma [ lem : gauss_sum ] implies @xmath216 then , using the easily verified identity @xmath217 we find that @xmath218 as required .
[ lem : qsum1 ] let @xmath180 be a ternary quadratic - residue array of size @xmath63 , and let @xmath98 and @xmath99 be real numbers satisfying @xmath141 .
then , as @xmath120 , @xmath219 given an array @xmath23 , let @xmath220 denote the transpose of @xmath23 . since @xmath221 and @xmath222 is again a ternary quadratic - residue array , it is sufficient to prove the first statement in the lemma .
let @xmath223 $ ] be a polynomial of degree @xmath224 , and let @xmath225 and @xmath129 be integer .
we shall make use of the lagrange interpolation formula @xmath226{hoholdt1988 } , for example ) .
it follows that } z_{s , t}(-{\epsilon}_i,{\epsilon}_j ) & = \frac{2}{p}\sum_{k=0}^{p-1 } z_{s , t}({\epsilon}_k,{\epsilon}_j ) \frac{{\epsilon}_k}{{\epsilon}_k+{\epsilon}_i}. \nonumber \intertext{set $ s:=\lfloor ps\rfloor$ and $ t:=\lfloor pt\rfloor$ , and use~\eqref{eqn : xst_x } and lemma~\ref{lem : gauss_sum } to obtain } z_{s , t}(-{\epsilon}_i,{\epsilon}_j ) & = 2(-1)^{\frac{p+1}{2 } } { \epsilon}_j^{-t } \sum_{k=0}^{p-1 } { \epsilon}_k^{-s } \chi(k\beta+j\beta ' ) \frac{{\epsilon}_k}{{\epsilon}_k+{\epsilon}_i } , \nonumber \intertext{where $ \chi$ is the quadratic character of $ \operatorname{gf}(p^2)$ and $ \{\beta,\beta'\}$ is some basis for $ \operatorname{gf}(p^2)$ over $ \operatorname{gf}(p)$. since we also have } \overline{z_{s , t}(-{\epsilon}_i,{\epsilon}_j ) } & = 2(-1)^{\frac{p+1}{2 } } { \epsilon}_j^t \sum_{k=0}^{p-1 } { \epsilon}_k^s \chi(k\beta+j\beta ' ) \frac{{\epsilon}_i}{{\epsilon}_k+{\epsilon}_i } , \nonumber\end{aligned}\ ] ] we find that @xmath227 use the definition of @xmath228 to write @xmath229
so that after variable relabeling @xmath230 put @xmath231 and note that @xmath232 ranges over @xmath90 as @xmath233 and @xmath129 range from @xmath234 to @xmath224 . by the definition of @xmath235 we therefore obtain @xmath236 let the set @xmath237 be as defined in , and write @xmath238 where @xmath239 using , the sum @xmath23 can be bounded as @xmath240 using symmetry of @xmath235 and @xmath228 with respect to interchanging their arguments , the sum @xmath56 can be written as @xmath241 application of and to evaluate @xmath242 and @xmath243 then gives @xmath244 we wish to apply the identity @xmath245 ( see @xcite , for example ) . since @xmath246
, we have @xmath247 for all sufficiently large @xmath1 , which allows us to apply to to obtain , for all sufficiently large @xmath1 , @xmath248 by the definition of @xmath249 , we have @xmath250 as @xmath120 , and therefore , @xmath251 the proof is completed by substituting and in .
[ lem : qsum2 ] let @xmath180 be a ternary quadratic - residue array of size @xmath63 , and let @xmath98 and @xmath99 be real numbers satisfying @xmath141 . then
, as @xmath120 , @xmath252 the idea of the proof is similar to that of the proof of lemma [ lem : qsum1 ] .
the main difference is that we now have to apply interpolation of @xmath253 in both indeterminates .
let @xmath225 and @xmath129 be integer . by the interpolation formula we have @xmath254 we obtain @xmath255 use ,
relabel the summation indices , and use the definition of @xmath235 to give @xmath256 let @xmath237 be the set defined in , and write @xmath257 where @xmath258 & \hspace{5em}\omega(k\beta+k'\beta',\ell\beta+\ell'\beta',m\beta+m'\beta')\,\gamma(k,\ell , m)\,\gamma(k',\ell',m')\\[1ex ] b&=\frac{64}{p^{10}}\sum_{{\genfrac{}{}{0pt}{}{0\le k,\ell , m , k',\ell',m'<p}{(k\beta+k'\beta',\ell\beta+\ell'\beta',m\beta+m'\beta')\in i}}}{\epsilon}_{k-\ell+m}^s\,{\epsilon}_{k'-\ell'+m'}^t\,{\epsilon}_\ell\,{\epsilon}_{\ell'}\\[1ex ] & \hspace{5em}\omega(k\beta+k'\beta',\ell\beta+\ell'\beta',m\beta+m'\beta')\,\gamma(k,\ell , m)\,\gamma(k',\ell',m').\end{aligned}\ ] ] from we have the bound @xmath259 we use symmetry of @xmath235 and @xmath228 with respect to interchanging their arguments to partition the sum @xmath56 further as @xmath260 where @xmath261 ^ 2\\ b_2&=\frac{64}{p^{10}}\sum_{k=1}^{p-1}\omega(k\beta , k\beta,0)\,\gamma(k , k,0)\,\gamma(0,0,0)(2{\epsilon}_k+{\epsilon}_k^{2s})\\
b_3&=\frac{64}{p^{10}}\sum_{k'=1}^{p-1}\omega(k'\beta',k'\beta',0)\,\gamma(0,0,0)\,\gamma(k',k',0)(2{\epsilon}_{k'}+{\epsilon}_{k'}^{2t})\\ b_4&=\frac{64}{p^{10}}\sum_{1\le k , k'<p}\omega(k\beta+k'\beta',k\beta+k'\beta',0)\,\gamma(k , k,0)\,\gamma(k',k',0)(2{\epsilon}_k{\epsilon}_{k'}+{\epsilon}_k^{2s}{\epsilon}_{k'}^{2t}).\end{aligned}\ ] ] application of and to evaluate @xmath262 and @xmath243 then gives @xmath263 we complete the proof by evaluating the asymptotic behavior of the terms @xmath264 , @xmath265 , @xmath266 , and @xmath267 .
_ the term @xmath264 _ : : the term @xmath264 becomes @xmath268 _ the terms @xmath265 and @xmath266 _ : : since @xmath269 implies that @xmath247 for all sufficiently large @xmath1 , we can use the identity to write @xmath270 .
\nonumber \intertext{then , since $ s = ps+o(1)$ as $ p\to\infty$ , we obtain } b_2&=\tfrac{1}{12}+\tfrac{4}{3}({\lverts\rvert}-\tfrac{1}{4})^2+o(p^{-1})\quad\mbox{as $ p\to\infty$}. \label{eqn : term_b } \intertext{we proceed similarly for the term $ b_3 $ and find that } b_3&=\tfrac{1}{12}+\tfrac{4}{3}({\lvertt\rvert}-\tfrac{1}{4})^2+o(p^{-1})\quad\mbox{as $ p\to\infty$}. \label{eqn : term_c}\end{aligned}\ ] ] _ the term @xmath267 _ : : the term @xmath267 can be rewritten as @xmath271.\ ] ] noting that @xmath247 and @xmath272 for all sufficiently large @xmath1 , application of gives @xmath273\left[\frac{p^2}{2}\left(\frac{{\lvert2t-1\rvert}}{p}-\frac{1}{2}\right)^2-\frac{p^2 + 2}{24}\right].\end{gathered}\ ] ] since @xmath250 and @xmath274 as @xmath120 , we finally obtain @xmath275\left[8({\lvertt\rvert}-\tfrac{1}{4})^2-\tfrac{1}{6}\right]+o(p^{-1})\quad\mbox{as $ p\to\infty$}. \label{eqn : term_d}\ ] ] the result now follows by substituting the asymptotic forms , , , and into and then and into .
we are now able to prove the main result of this section .
let @xmath276 be a quadratic - residue array of size @xmath63 , and let @xmath98 and @xmath99 be real numbers satisfying @xmath141
. then @xmath277\left[\tfrac{1}{2}+8\left({\lvertt\rvert}-\tfrac{1}{4}\right)^2\right].\ ] ] let @xmath180 be a ternary quadratic - residue array of size @xmath63 .
from we have @xmath278 ^ 2&=\frac{1}{4p^6}\sum_{0\le i , j < p}{\big\lvertz_{s , t}({\epsilon}_i,{\epsilon}_j)\big\rvert}^4+\frac{1}{4p^6}\sum_{0\le i , j < p}{\big\lvertz_{s , t}(-{\epsilon}_i,{\epsilon}_j)\big\rvert}^4\\ & + \frac{1}{4p^6}\sum_{0\le i , j < p}{\big\lvertz_{s , t}({\epsilon}_i,-{\epsilon}_j)\big\rvert}^4+\frac{1}{4p^6}\sum_{0\le i , j <
p}{\big\lvertz_{s , t}(-{\epsilon}_i,-{\epsilon}_j)\big\rvert}^4.\end{aligned}\ ] ] using lemmas [ lem : qsum0 ] , [ lem : qsum1 ] , and [ lem : qsum2 ] to determine the asymptotic behavior of the sums on the right - hand side , we obtain @xmath279 ^ 2\\ = \tfrac{10}{9}+\left[\tfrac{1}{2}+8\left({\lverts\rvert}-\tfrac{1}{4}\right)^2\right]\left[\tfrac{1}{2}+8\left({\lvertt\rvert}-\tfrac{1}{4}\right)^2\right]+o\left(p^{-1}(\log p)^8\right)\quad\mbox{as $ p\to\infty$}. \label{eqn : ss_z}\end{gathered}\ ] ] now , since the binary array @xmath276 differs from the ternary array @xmath180 only in a single position , we have @xmath280 where @xmath281 then , by the cauchy
schwarz inequality , @xmath282 ^ 2-\sum_{u , v}[c_{z_{s , t}}(u ,
v)]^2\bigg\rvert}\\ \le\sum_{u , v}[\delta(u , v)]^2 + 2\bigg(\sum_{u , v}[c_{z_{s , t}}(u , v)]^2\bigg)^{1/2}\bigg(\sum_{u , v}[\delta(u , v)]^2\bigg)^{1/2},\end{gathered}\ ] ] so that by and @xmath283 ^ 2-\sum_{u , v}[c_{z_{s , t}}(u , v)]^2\bigg\rvert}=o(p^{-1})\quad\mbox{as $ p\to\infty$}.\ ] ] the theorem follows from by noting that @xmath284 .
we have computed the asymptotic value of the merit factor of legendre arrays ( under certain conditions on the growth rate of their dimensions ) and of quadratic - residue arrays , for all rotations of rows and columns .
the asymptotic merit factor of rotated legendre arrays and rotated quadratic - residue arrays is shown in figures [ fig : fx ] and [ fig : fz ] , respectively .
the maximum asymptotic merit factor , taken over all rotations , equals @xmath65 for both array families , which occurs at the rotations @xmath285 , where @xmath286 . however , at all other rotations , the asymptotic merit factor of quadratic - residue arrays is larger than that of legendre arrays . on the other hand ,
an advantage of legendre arrays is that they are not restricted to be square .
( under conditions specified in theorem [ thm : farray1 ] ) , where @xmath85 is a legendre array.,width=453 ] , where @xmath276 is a quadratic - residue array.,width=453 ] in @xcite , the authors exhibited a family of sequences @xmath23 obtained by appending an initial fraction of a rotated legendre sequence to itself . based on partial explanations and extensive numerical computations ,
it was conjectured in @xcite that this sequence family has asymptotic merit factor greater than @xmath45 . under the assumption that this conjecture is correct
, the corresponding square product array @xmath287 has asymptotic merit factor greater than @xmath288 , by .
this suggests that the maximum asymptotic value of the merit factor of the two array families considered in this paper , namely @xmath65 , can be surpassed by another array family . in closing
, we remark that instead of studying the merit factor of two - dimensional arrays of size @xmath4 , gulliver and parker @xcite studied the ( suitably generalised ) merit factor of @xmath124-dimensional arrays of size @xmath289 . in @xcite ,
the merit factor of several families of such arrays was calculated .
in particular , the largest asymptotic value , as @xmath290 , of the merit factor of a family of @xmath124-dimensional arrays considered in @xcite equals @xmath64 .
i would like to thank jonathan jedwab for valuable discussions on the subject and for careful comments on a draft of this paper .
j. jedwab , _ a survey of the merit factor problem for binary sequences _ , in `` sequences and their applications , '' ( eds .
t. helleseth et al . ) , lecture notes in computer science , vol . 3486 , springer verlag , ( 2005 ) , 3055 .
h. e. jensen and t. hholdt , _ binary sequences with good correlation properties _ , in `` applied algebra , algebraic algorithms and error - correcting codes '' , ( eds .
l. huguet and a. poli ) , lecture notes in computer science , vol .
356 , springer - verlag , ( 1989 ) , 306320 .
r. g. van schyndel , a. z. tirkel , i. d. svalbe , t. e. hall , and c. f. osborne , _ algebraic construction of a new class of quasi - orthogonal arrays for steganography _ , proceedings of spie * 3657 * ( 1999 ) , 354364 . | we calculate the asymptotic merit factor , under all cyclic rotations of rows and columns , of two families of binary two - dimensional arrays derived from the quadratic character .
the arrays in these families have size @xmath0 , where @xmath1 and @xmath2 are not necessarily distinct odd primes , and can be considered as two - dimensional generalisations of a legendre sequence .
the asymptotic values of the merit factor of the two families are generally different , although the maximum asymptotic merit factor , taken over all cyclic rotations of rows and columns , equals @xmath3 for both families .
these are the first non - trivial theoretical results for the asymptotic merit factor of families of truly two - dimensional binary arrays . |
in treatment planning ( tp ) of radiotherapy , a patient body is commonly modeled to be h@xmath0o of variable effective density , which corresponds to electron density for high - energy photons or stopping - power ratio for charged particles .
the effective density is normally converted from computed - tomography ( ct ) number for attenuation of kilovoltage ( kv ) x ray .
the standard approach to construct such conversion functions involves experimental modeling of the x ray and stoichiometric analysis of body tissues.@xcite the effective density is , though practically successful,@xcite an approximate concept for radiations undergoing complex interactions . with the advancement of computer technology , monte carlo ( mc ) simulation of radiotherapy is becoming feasible , where a radiation is handled as a collection of particles individually interacting with matter of known composition according to the basic laws of physics . for the modeling of body tissues , schneider , bortfeld , and schlegel ( sbs ) applied the stoichiometric calibration to construct functions to convert ct number to mass density and elemental weights.@xcite their conversion functions , though commonly used for research,@xcite are not applicable to the other ct systems .
calibration and maintenance of the complex one - to - many relations may prevent mc simulation from applying to tp practice . for patient dose calculation , tp systems commonly use a selectable function to convert ct number to effective density of interest .
recently , a practical two - step approach was proposed for tp with proton and ion beams,@xcite where ct number is only converted to electron density that is automatically converted to interaction - specific effective densities using invariant relations . in the present study
, we extend the two - step approach to promote mc - based tp practice .
we used the standard body tissue data in icrp publication 110,@xcite which is the latest compilation of the kind . in the publication
are mass density @xmath1 and elemental weights @xmath2 for 53 standard tissues that fully comprise 141 organs of reference male and female , for which we derived electron density @xmath3 and mass fraction @xmath4 per person ( 50% male/50% female ) .
ignoring tiny mass of air , below 0.90 g/@xmath5 was only a lung at 0.384 g/@xmath5 occupying 1.4% of body mass , which we chose for elemental composition of air - containing tissues . in the 0.901.00 g/@xmath5 region
were an adipose tissue at 0.95 g/@xmath5 ( 33.9% ) and medullary cavities including bone marrow all at 0.98 g/@xmath5 ( 0.4% ) . in the 1.001.07 g/@xmath5 region
were a muscle at 1.05 g/@xmath5 ( 34.5% ) , spongiosa tissues ( 0.3% ) , and many other general organs ( 11.7% ) . in the 1.071.101 g/@xmath5 region were a skin ( 4.9% ) , a cartilage ( 0.6% ) , and spongiosa tissues ( 0.7% ) , which are miscellaneously epithelium , connective , and fatty bone tissues . in the 1.1011.25 g/@xmath5 region were only spongiosa tissues ( 5.7% ) .
above 1.25 g/@xmath5 were a mineral bone at 1.92 g/@xmath5 ( 5.7% ) and a tooth at 2.75 g/@xmath5 ( 0.1% ) . among the material properties , we adopted mass density @xmath1 as the independent variable and examined its correlation to the dependent variables : electron density @xmath3 and elemental densities @xmath6 of six major elements
@xmath7 = \{h , c , n , o , p , ca}. to define the regional representative tissues , we took mass - weighted mean for the set of densities @xmath8 over tissue @xmath9 in region @xmath10 by @xmath11 for the standard tissues and the regional representative tissues , we calculated the residual weight and the mean residual atomic number by adding weights and averaging atomic number @xmath12 over residual element @xmath13 , @xmath14 with which the residual mass could be approximately included in mc simulation .
compared to muscle / organ tissues , adipose / marrow tissues have high concentration of fat .
teeth have high concentration of minerals , connective tissues have high concentration of collagen , and bones are in between them .
the concentration varies among and within individual tissues and are generally correlated with density .
anatomically , adipose tissues are neighboring to muscles and organs , muscles are connected to bones via connective tissues , and teeth are connected to jaw bones . at these interfaces ,
tissue mixing may occur in an image of finite spatial resolution .
therefore , we modeled an arbitrary tissue of mass density @xmath1 to be mixture of representative tissues 1 and 2 that comprise a density segment of @xmath15 .
the other densities are interpolated from those of the representative tissues by @xmath16 in a mass - weighting manner.@xcite in other words , we assigned the representative tissues to the polyline points for conversion from mass density to the other densities . for fatty tissues lighter than the representative adipose / marrow tissue , we extended the soft - tissue region down to 0.90 g/@xmath5 , which is the mass density of human fat at 37@xmath17c.@xcite similarly , we extended the region for air - containing tissues up to 0.80 g/@xmath5 , leaving a transition segment of 0.800.90 g/@xmath5 to avoid discontinuity of the material properties and to cope with body fat that escape over the segment boundary as viewed in ct image .
we also extended the hard - tissue region up to 3.00 g/@xmath5 for teeth heavier than their representative . to define the extended boundaries
, we applied eq .
[ eq:3 ] for extrapolation although , for the fat , the n , p , and ca weights were forced to zero and the c weight was adjusted to sum up to 100% to comply with general composition of fatty acid . for the polyline tissues , we compiled mass fraction , mass and electron densities , elemental and residual weights , and mean residual atomic number .
hnemohr _ et al .
_ formulated a bi - line relation between mass and electron densities for the body tissues compiled in refs .
.@xcite from the same dataset , sbs selected their representative tissues for mixing of air / lung , adipose tissue / adrenal gland , small intestine / connective tissue , and marrow / cortical bone,@xcite from which we derived their relations between mass density and elemental densities as broken - line functions .
we compared our results with those preceding formulations .
) and female ( @xmath18 ) tissues plotted with the polyline function ( solid lines ) , the bi - line function ( dotted lines ) and `` adipose 3 '' ( @xmath19 ) at 0.93 g/@xmath5 by hnemohr
_ et al . _
( 2014)@xcite , and embedded subplots for the box - shaped areas.,width=8 ] ) and female ( @xmath18 ) tissues plotted with the polyline functions ( solid lines ) and the broken - line functions ( dotted lines ) according to schneider , bortfeld , and schlegel ( 2000)@xcite.,width=8 ] ) and female ( @xmath18 ) tissues with thickness varied with residual mass and ( b ) mass - weighted histogram of mean residual atomic number.,width=8 ] rlddddddddddd & tissue type & & & & & & & & & & & + 1 & air & & 0.001 & 0.001 & 0.00 & 0.01 & 75.52 & 23.17 & 0.00 & 0.00 & 1.30 & 18.0 + 2 & lung & 1.4 & 0.384 & 0.380 & 10.3 & 10.7 & 3.2 & 74.6 & 0.2 & 0.0 & 1.0 & 15.9 + 3 & air - containing & & 0.80 & 0.793 & 10.3 & 10.7 & 3.2 & 74.6 & 0.2 & 0.0 & 1.0 & 15.9 + 4 & fat & & 0.90 & 0.907 & 12.09 & 84.39 & 0.00 & 3.52 & 0.00 & 0.00 & 0.00 & + 5 & adipose / marrow & 34.3 & 0.950 & 0.952 & 11.40 & 58.92 & 0.74 & 28.64 & 0.00 & 0.00 & 0.30 & 14.7 + 6 & muscle / organ & 46.5 & 1.049 & 1.040 & 10.25 & 14.58 & 3.20 & 70.87 & 0.21 & 0.02 & 0.87 & 16.8 + 7 & miscellaneous & 6.3 & 1.090 & 1.077 & 9.94 & 20.90 & 3.84 & 63.73 & 0.45 & 0.27 & 0.87 & 15.5 + 8 & spongiosa & 5.7 & 1.137 & 1.116 & 9.30 & 39.15 & 2.22 & 41.71 & 2.36 & 4.60 & 0.66 & 14.9 + 9 & mineral bone & 5.7 & 1.92 & 1.784 & 3.6 & 15.9 & 4.2 & 44.8 & 9.4 & 21.3 & 0.8 & 13.1 + 10 & tooth & 0.1 & 2.75 & 2.518 & 2.2 & 9.5 & 2.9 & 42.1 & 13.7 & 28.9 & 0.7 & 12.0 + 11 & extra tooth & & 3.00 & 2.739 & 1.93 & 8.27 & 2.65 & 41.58 & 14.53 & 30.37 & 0.67 & 11.8 + [ tab:1 ] figure [ fig:1 ] shows the correlation between mass and electron densities , where @xmath20 is the electron density of water .
the two fitting functions , the polyline and the bi - line by hnemohr _
_ , were generally consistent with the standard tissues except around the fat at 0.90 g/@xmath5 and around the tooth at 2.75 g/@xmath5 , for which the bi - line function gave 0.892 ( @xmath21 ) and 2.486 ( @xmath22 ) , respectively .
the discontinuity in the bi - line at their `` adipose 3 '' was settled by the polyline with better fitting .
figure [ fig:2 ] shows the correlation between mass density and elemental densities .
the polyline functions and the broken - line functions according to sbs were both generally consistent with the standard tissues except for the c , n , and o densities in the 1.01.1 g/@xmath5 region .
as shown in fig .
[ fig:3 ] , the highest and lowest mean residual atomic numbers were 20.6 at 1.04 g/@xmath5 for the thyroid and 12.0 at 2.75 g/@xmath5 for the tooth . the global mean of 15.95 approximately corresponded to element s. table [ tab:1 ] shows the resultant material properties of the polyline tissues .
a concise yet complete dataset of the polyline tissues may be suited to stoichiometric analysis for ct - number calibration.@xcite in the cases where electron density is available , the most likely mass and elemental densities can be determined by eq .
[ eq:3 ] with variables @xmath23 replaced to @xmath24 .
the resultant mass density and composition including residual weight for element s will constitute a volumetric patient model for mc simulation .
the general agreement between the formulations indicates the consistency between the compilations of body - tissue data .
the mass - weighting approach of this study took advantage of the publication that focused on the computational phantoms of reference adult male and female rather than on variations in age , physical status , or individual.@xcite the poor fitting in c , n , and o densities was due to undifferentiated inclusion of spongiosa tissues occupying 1% of body mass in the 10.1.1 g/@xmath5 region .
they could potentially be resolved by anatomical identification or independent quantitative imaging , in which case either an extended spongiosa / mineral - bone segment or a separate marrow / spongiosa segment should be applied to them .
recently , megavoltage ( mv ) ct and dual - energy ( de ) ct have been investigated for direct electron - density imaging.@xcite hnemohr _
_ formulated the composition of body tissues as a two - variable function of electron density and effective atomic number from dect with improved accuracy.@xcite the dect with kv and mv x rays would be ideal.@xcite nevertheless , use of mvct or dect will potentially mitigate metal artifact and beam hardening , which limits the accuracy of kvct.@xcite in conclusion , we have formulated the invariant polyline relations between mass and electron densities and from mass to elemental densities for body tissues .
the formulation enables mc simulation in tp practice without additional burden with ct - number calibration .
http://dx.doi.org/10.1088/0031-9155/41/1/009[u .
schneider ] , e. pedroni , and a. lomax , `` the calibration of ct hounsfield units for radiotherapy treatment planning , '' phys . med . biol . * 41 * , 111124 ( 1996 ) . http://dx.doi.org/10.1088/0031-9155/48/8/307[n .
kanematsu ] , n. matsufuji , r. kohno , s. minohara , and t. kanai , `` a ct calibration method based on the polybinary tissue model for radiotherapy treatment planning , '' phys .
biol . * 48 * , 10531064 ( 2003 ) . http://dx.doi.org/10.1118/1.4870980[n .
kanematsu ] , y. koba , r. ogata , and t. himukai , `` influence of nuclear interactions in polyethylene range compensators for carbon - ion radiotherapy , '' med . phys . * 41 * , 071704 - 18 ( 2014 ) .
http://dx.doi.org/10.1088/0031-9155/60/1/421[t .
inaniwa ] , n. kanematsu , y. hara , and t. furukawa , `` nuclear - interaction correction of integrated depth dose in carbon - ion radiotherapy treatment planning , '' phys .
* 60 * , 421435 ( 2015 ) .
http://dx.doi.org/10.1088/0031-9155/45/2/314[w .
schneider ] , t. bortfeld , and w. schlegel , `` correlation between ct numbers and tissue parameters needed for monte carlo simulation of clinical dose distributions , '' phys .
* 45 * , 459478 ( 2000 ) .
http://dx.doi.org/10.1088/0031-9155/53/17/023[h .
paganetti ] , h. jiang , k. parodi , r. slopsema , and m. engelsman , `` clinical implementation of full monte carlo dose calculation in proton beam therapy , '' phys .
biol . * 53 * , 48254853 ( 2008).x http://dx.doi.org/10.1088/0031-9155/58/8/2471[a .
mairani ] , t. t. bhlen , a. schiavi , t. tessonnier , s. molinelli , s. brons , g. battistoni , k. parodi , and v. patera , `` a monte carlo - based treatment planning tool for proton therapy , '' phys .
biol . * 58 * , 24712490 ( 2013 ) . http://dx.doi.org/10.1118/1.3679339[n .
kanematsu ] , t. inaniwa , and y. koba , `` relationship between electron density and effective densities of body tissues for stopping , scattering , and nuclear interactions of proton and ion beams , '' med . phys . * 39 * , 10161020 ( 2012 ) . http://dx.doi.org/10.1088/0031-9155/59/22/7081[p .
farace ] , `` experimental verification of ion stopping power prediction from dual energy ct data in tissue surrogates , '' phys
biol . * 59 * , 70817084 ( 2014 ) .
http://www.icrp.org/publication.asp?id=icrp%20publication%20110[international commission on radiological protection ] , _ icrp publication 110 : adult reference computational phantoms , _ ann .
icrp 39(2 ) , ( icrp , ottawa , 2009 ) .
http://dx.doi.org/10.1088/0031-9155/60/11/4243[d .
r. warren ] , m. partridge , m. a. hill , and k. peach , `` improved calibration of mass stopping power in low density issue for a proton pencil beam algorithm , '' phys .
biol . * 60 * , 42434261 ( 2015 ) . http://dx.doi.org/10.1259/0007-1285-60-717-907[d .
r. white ] , h. q. woodard , and s. m. hammond , `` average soft - tissue and bone models for use in radiation dosimetry , '' br . j. radiol .
* 60 * , 907913 ( 1987 ) .
http://dx.doi.org/10.1118/1.4875976[n .
hnemohr ] , h. paganetti , s. greilich , o. jkel , and j. seco , `` tissue decomposition from dual energy ct data for mc based dose calculation in particle therapy , '' med . phys . *
41 * , 061714 - 114 ( 2014 ) .
http://dx.doi.org/10.1088/0031-9155/45/4/404[k .
j. ruchala ] , g. h. olivera , e. a. schloesser , r. hinderer , and t. r. mackie , `` calibration of a tomotherapeutic mvct system , '' phys .
* 45 * , n27n36 ( 2000 ) .
http://dx.doi.org/10.1016/j.radonc.2011.08.029[g .
landry ] , b. reniers , p. v. granton , b. van rooijen , l. beaulieu , j. e. wildberger , and f. verhaegen , `` extracting atomic numbers and electron densities from a dual source dual energy ct scanner : experiments and a simulation model , '' radiother .
oncol . * 100 * , 375379 ( 2011 ) . http://dx.doi.org/10.1088/0031-9155/53/9/015[m .
bazalova ] , j .- f .
carrier , l. beaulieu , and f. verhaegen , `` dual - energy ct - based material extraction for tissue segmentation in monte carlo dose calculations , '' phys .
biol . * 53 * , 24392456 ( 2008 ) . http://dx.doi.org/10.1118/1.4901551[m
. wu ] , a. keil , d. constantin , j. star - lack , l. zhu , and r. fahrig , `` metal artifact correction for x - ray computed tomography using kv and selective mv imaging , '' med .
41 * , 121910 - 115 ( 2014 ) . | * purpose : * for monte carlo simulation of radiotherapy , x - ray ct number of every system needs to be calibrated and converted to mass density and elemental composition .
this study aims to formulate material properties of body tissues for practical two - step conversion from ct number .
* methods : * we used the latest compilation on body tissues that constitute reference adult male and female .
we formulated the relations among mass , electron , and elemental densities into polylines to connect representative tissues , for which we took mass - weighted mean for the tissues in limited density regions .
we compared the polyline functions of mass density with a bi - line for electron density and broken lines for elemental densities , which were derived from preceding studies .
* results : * there was generally high correlation between mass density and the other densities except of c , n , and o for light spongiosa tissues occupying 1% of body mass .
the polylines fitted to the dominant tissues and were generally consistent with the bi - line and the broken lines .
* conclusions : * we have formulated the invariant relations between mass and electron densities and from mass to elemental densities for body tissues .
the formulation enables monte carlo simulation in treatment planning practice without additional burden with ct - number calibration . |
the cdf and d0 collaborations at the tevatron have successfully collected numerous data for the @xmath26 and @xmath27 meson @xcite . the lhc shall also provide a good platform to study the properties of the @xmath26 and @xmath27 meson , e.g. , recent results and review papers on the @xmath26 and @xmath27 meson measurements and searches by the lhcb , cms , and atlas collaborations at the lhc can be found in refs .
@xcite .
the @xmath26 and @xmath27 meson form an interesting experiment and theory for the study of the quantum chromodynamics ( qcd ) and new physics ( np ) , which can be used to study those interesting topics as the qcd model building , the electroweak symmetry breaking mechanism , charge - parity ( cp ) violation , new physics beyond the standard model ( sm ) and etc .
@xcite . taking into account the prospects of the @xmath26 or @xmath27 meson production and decay at the tevatron and at the running lhc ,
the future numerous data require more accurate theoretical predictions on its hadronic production and indirect production through weak - decay processes . for the @xmath26 or @xmath27 meson direct production ,
various investigations has been carried out studied in detail in refs .
the @xmath26 or @xmath27 meson indirect production through the @xmath28 boson and high excited @xmath29 meson decays has been measured and discussed in refs .
although in literature the direct hadronic production of the @xmath26 or @xmath27 meson has been thoroughly studied in refs .
@xcite and cdf discovered the meson which just comes from the hadroproduction .
as a compensation to understanding the production mechanisms , and also testing the perturbative quantum chromodynamics ( pqcd ) and the qcd factorization formula .
it is quite interesting that to study the indirectly production of the @xmath26 or @xmath27 through top quark or antitop quark decays , especially on considering the fact that numerous @xmath21-quark or @xmath22-quark may be produced at the lhc . at the lhc running at
the center - of - mass energy @xmath20 tev and luminosity raising up to @xmath19 @xcite , about @xmath30 @xmath31 quark per year will be produced .
more @xmath21-quark or @xmath22-quark rare decays can be adopted for precise studies at the lhc .
this paper is devoted to study the indirect production of the @xmath26 or @xmath27 meson via @xmath21-quark or @xmath22-quark decays .
the theoretical studies of the bound states is based on the factorization formula , which includes the color - singlet qcd factorization formula ( csqcdff ) @xcite , the nonrelativistic quantum chromodynamics ( nrqcd ) @xcite , the @xmath32-factorization @xcite and etc .
the csqcdff is a useful method for studying on the production and decays of the quarkinum .
according to the csqcdff , the processes of the production and decays of the quarkinum can be factorized as two parts of the perturbatively calculable short - distance coefficients and the nonperturbative long - distance factors . in present paper
it is investigated a systematic study of the indirect production of the @xmath0\rangle$ ] or @xmath1\rangle$ ] quarkonium via @xmath33-quark or @xmath34-quark decays within the csqcdff .
to derive the analytical expression of the pqcd calculable conventional squared amplitudes becomes too complex and lengthy for more ( massive ) particles in the final states and for the excited states to be generated for the emergence of massive - fermion lines in the feynman diagrams , especially to derive the squared amplitudes of the @xmath10-wave states . in refs .
@xcite , the `` improved trace technology '' is suggested and developed to solve the problem .
here we adopt the improved trace technology " to derive the analytical expression for the above - mentioned processes . without confusing and for simplifying the statements , later on we will not distinguish the @xmath26 and @xmath27 meson unless necessary . because of the production of the @xmath0\rangle$ ] or @xmath1\rangle$ ] meson through @xmath21-quark or @xmath22-quark decays channels , i.e. , @xmath35\rangle+w^{+}s$ ] and @xmath36\rangle+w^{-}\bar{s}$ ] , and @xmath37\rangle+w^{+}b$ ] and @xmath38\rangle+w^{-}\bar{b}$ ] , are symmetric in the interchange from particle to anti - particle .
the rest is organized as follows . in sec .
ii , we introduce the calculation techniques for the above - mentioned @xmath21-quark and @xmath22-quark decays to the @xmath0\rangle$ ] quarkonium . in order to calculate the production of the @xmath0\rangle$ ] quarkonium and its excited states
, we evaluate the masses of the @xmath0\rangle$ ] quarkonium with the energy eigenvalue of the nonrelativistic schr@xmath4dinger equation . in sec.iii , we calculate and tabulate the masses of the @xmath0\rangle$ ] quarkonium and its values of the schr@xmath4dinger radial wave functions at the origin .
then , we investigate a systematic study on the production of @xmath0\rangle$ ] through @xmath21-quark and @xmath22-quark decays channels , i.e. , @xmath35\rangle+w^{+}s$ ] and @xmath39\rangle+w^{-}\bar{b}$ ] , where @xmath0\rangle$ ] stands for @xmath40\rangle$ ] , @xmath8\rangle$ ] , @xmath11\rangle$ ] , and @xmath13\rangle$ ] ( with @xmath41 $ ] ) .
we make some discussions on the various differential distributions and the uncertainties of the decays widths by the masses of the @xmath0\rangle$ ] quarkonium and the correspondingly values of the schr@xmath4dinger radial wave functions at the origin .
the fourth section is reserved for a summary .
the idea of improved trace technology " is put in the appendix finally .
for the production of the @xmath0\rangle$ ] meson through @xmath21-quark and @xmath22-quark decays channels , @xmath43\rangle(q_3 ) + w^{+}(q_2 ) + s(q_1)$ ] and @xmath44\rangle(q_3 ) + w^{-}(q_2 ) + \bar{b}(q_1)$ ] , where @xmath45 and @xmath46 ( @xmath47 ) are the momenta of the corresponding particles , according to the csqcdff , its total decay widths @xmath48 can be factorized as @xcite @xmath49\rangle+ w^{+}s ) \langle{\cal o}^h(n ) \rangle,\nonumber\\ d\gamma&=&\sum_{n } d\hat\gamma(\bar{t}\to |(b\bar{s})[n]\rangle+w^{-}\bar{b } ) \langle{\cal o}^h(n ) \rangle.\end{aligned}\ ] ] where the nonperturbative matrix element @xmath50 describes the hadronization of a @xmath51 pair into the observable quark state @xmath52 and is proportional to the transition probability of the perturbative state @xmath51 into the bound state @xmath0\rangle$ ] quarkonium .
the matrix elements @xmath50 for the color - singlet components can be directly related to the schr@xmath4dinger wave functions at the origin for the @xmath5-wave states and the first derivative of the wave functions at the origin for the @xmath10-wave states , which can be computed via the potential models @xcite .
the nonperturbative matrix element @xmath50 can also be obtained via potential nrqcd ( pnrqcd ) @xcite and/or lattice qcd @xcite .
the hard - scattering amplitudes for specified processes can be dealt with @xmath53\rangle+w^{+}s,\nonumber\\ & & \bar{t}\rightarrow|(b\bar{s})[n]\rangle+w^{-}\bar{b}.\end{aligned}\ ] ] \rangle(q_3 ) + w^{+}(q_2)+ s(q_1)$ ] , where @xmath0\rangle$ ] stands for @xmath54\rangle$ ] , @xmath55\rangle$ ] , @xmath56\rangle$ ] , and @xmath57\rangle$ ] ( with @xmath58 $ ] ) , respectively.,scaledwidth=50.0% ] \rangle(q_3 ) + w^{-}(q_2)+ \bar{b}(q_1)$ ] , where @xmath0\rangle$ ] stands for @xmath54\rangle$ ] , @xmath55\rangle$ ] , @xmath56\rangle$ ] , and @xmath57\rangle$ ] ( with @xmath58 $ ] ) , respectively.,scaledwidth=50.0% ] the feynman diagrams of the two processes , @xmath59\rangle(q_3 ) + w^{+}(q_2)+s(q_1)$ ] and @xmath60\rangle(q_3 ) + w^{-}(q_2)+\bar{b}(q_1)$ ] , are presented in figs .
[ feyn1 ] and [ feyn2 ] , where the intermediate gluon should be hard enough to produce a @xmath61 pair or @xmath62 pair , so the amplitudes are pqcd calculable .
these amplitudes of the above - mentioned channels can be generally expressed as @xmath63 where @xmath64 stands for the number of feynman diagrams , @xmath65 and @xmath66 are spin states , and @xmath67 and @xmath68 are color indices for the outing quark and the initial top quark or antitop quark , respectively .
the overall factor @xmath69 stands for the specified quarkonium in the color - singlet , here @xmath70 is the cabibbo - kobayashi - maskawa ( ckm ) matrix element , where @xmath71 stands for the processes @xmath72\rangle + w^{+ } s $ ] , and @xmath73 for @xmath74\rangle + w^{-}\bar{b}$ ] . while for production the @xmath75 ( @xmath76 ) states of the @xmath77\rangle$ ] quarkonium via the decays channel @xmath59\rangle(q_3 ) + w^{+}(q_2)+s(q_1)$ ] , the amplitudes @xmath78 can be written as @xmath79_{q=0 } , \label{a1}\\ { \cal a}_2 & = & \left[{\gamma_\alpha } \frac{\pi^{0(\nu)}_{q_3}(q)}{(q_1 + { q_{32}})^2}\slashed\epsilon(q_2)({1-\gamma^5 } ) \frac{(\slashed{q}_2+\slashed{q}_{31 } ) + { m_t}}{(q_2+q_{31 } ) ^2 - m_t^2 } \gamma_\alpha \right]_{q=0}. \label{a2}\end{aligned}\ ] ] for the @xmath80-wave states , @xmath78 can be written as @xmath81_{q=0 } , \label{a3}\\ { \cal a}^{s=0,l=1}_2 & = & \varepsilon_l^{\mu}(q_3 ) \frac{d}{dq_\mu } \left[{\gamma_\alpha } \frac{\pi^{0}_{q_3}(q)}{(q_1 + { q_{32}})^2}\slashed\epsilon(q_2)({1-\gamma^5 } ) \frac{(\slashed{q}_2+\slashed{q}_{31 } ) + { m_t}}{(q_2+q_{31 } ) ^2 - m_t^2 } \gamma_\alpha\right]_{q=0 } , \label{a4}\end{aligned}\ ] ] and the @xmath82-wave states ( @xmath83 ) @xmath84_{q=0 } , \label{a5}\\ { \cal a}^{s=1,l=1}_2 & = & \varepsilon^{j}_{\mu\nu}(q_3 ) \frac{d}{dq_\mu } \left[{\gamma_\alpha } \frac{\pi^{\nu}_{q_3}(q)}{(q_1 + { q_{32}})^2}\slashed\epsilon(q_2)({1-\gamma^5 } ) \frac{(\slashed{q}_2+\slashed{q}_{31 } ) + { m_t}}{(q_2+q_{31 } ) ^2 - m_t^2 } \gamma_\alpha\right]_{q=0}. \label{a6}\end{aligned}\ ] ] here @xmath85 stands for the relative momentum between the two constituent quarks in the @xmath0\rangle$ ] quarkonium . @xmath86 and @xmath87 are the momenta of the two constituent quarks , i.e. , @xmath88 where @xmath89 is implicitly adopted to ensure the gauge invariance of the hard scattering amplitude .
@xmath90 is the polarization vector of @xmath91 boson . @xmath92 and @xmath93 are the polarization vectors relating to the spin and the orbit angular momentum of the @xmath0\rangle$ ] quarkonium , @xmath94 is the polarization tensor for the spin triplet @xmath10-wave states ( with @xmath58 $ ] ) .
the covariant form of the projectors can be conveniently written as @xmath95 and @xmath96 here @xmath97 stands for the unit color matrix with @xmath98 for the qcd . after substituting these projectors into the above amplitudes and doing the possible simplifications , the amplitudes
then can be squared , summed over the freedoms in the final state and averaged over the ones in the initial state .
the selection of the appropriate total angular momentum quantum number is done by performing the proper polarization sum . if defining @xmath99 the sum over polarization for a spin triplet @xmath5-state or a spin singlet @xmath10-state is given by @xmath100 where @xmath101 or @xmath102 respectively .
in the case of @xmath82 states , the sum over the polarization is given by @xcite @xmath103 the amplitudes @xmath104 of the @xmath74\rangle+ w^{-}\bar{b}$ ] in the formulas are similar in the above - mentioned decays channels , so do not listed here to shorten the paper .
finally , the short - distance decay widths are given by @xmath105\rangle+w^{+}s)&= & \frac{1}{2 k^0 } \overline{\sum } |m|^{2 } d\phi_3,\nonumber\\ d\hat\gamma(\bar{t}\to |(b\bar{s})[n]\rangle+w^{-}\bar{b})&= & \frac{1}{2 k^0 } \overline{\sum } |m|^{2 } d\phi_3.\end{aligned}\ ] ] where @xmath106 means that one needs to average over the spin and color states of the initial particle and to sum over the color and spin of all the final particles . in the @xmath33-quark or @xmath34-quark rest frame ,
the three - particle phase space can be expressed as @xmath107 the @xmath108 phase space with massive quark / antiqark in the final state have been calculated to simplify procedures in refs .
@xcite . to shorten the paper ,
we shall not present it here and the interested reader may turn to these two references for the detailed technology .
we can not only derive the whole decay widths but also obtain the corresponding differential decay widths that are helpful for experimental studies with the help of the formulas listed in refs .
@xcite , such as @xmath109 , @xmath110 , @xmath111 , and @xmath112 , where @xmath113 , @xmath114 , @xmath115 stands for the angle between @xmath116 and @xmath117 , and @xmath118 for @xmath116 and @xmath119 . for the present
considered the squared amplitude of top quark semi - exclusive decay process , @xmath120(q_3 ) + w^{+}(q_2 ) + s(q_1)$ ] , there are two massive spinors of top quark and @xmath65-quark .
so to achieve the analytical expression for the squared amplitude becomes too complex and lengthy in the conventional way , especially for @xmath10-states of the @xmath0\rangle$ ] meson . for solved the problem , we adopt ` new trace amplitude approach ' to deal with the hard scattering amplitude ( [ amplitude ] ) for such complicated processes at the amplitude level . in a difference from the helicity amplitude approach @xcite , only the coefficients of the basic lorentz structures are numerical at the amplitude level .
however , by using the improved trace technology " in refs .
@xcite , one can sequentially obtain the squared amplitudes , and the numerical efficiency can also be greatly improved .
detailed processes of the approach can be found in refs .
@xcite , and here for self - consistency , we shall present its main idea in appendix .
finally , the partial decay widths over @xmath121 and @xmath122 can be expressed as @xmath123 where @xmath124 is the mass of the top quark . according to the csqcdff @xcite ,
the color - singlet nonperturbative matrix element @xmath125 can be related to the schr@xmath4dinger wave function @xmath126 at the origin and the first derivative @xmath127 for @xmath5 and @xmath10-wave states of the @xmath0\rangle$ ] quarkonium .
@xmath128\rangle}(0)|^2,\nonumber\\ \langle{\cal o}^h(np ) \rangle & \simeq & |\psi^\prime_{\mid(b\bar{s})[np]\rangle}(0)|^2.\end{aligned}\ ] ] as the spin - splitting effects are small , we will not distinguish the difference between the wave function parameters for the spin - singlet @xmath40\rangle$ ] and spin - triplet @xmath8\rangle$ ] states , also not discriminate the difference among the wave function parameters for @xmath11\rangle$ ] and @xmath13\rangle$ ] ( with @xmath58 $ ] ) in following calculation .
the mass spectrum of the @xmath0\rangle$ ] bound states has the form @xmath129\rangle)=m_b+m_s+e_n(m_b , m_s , v(\vec{r})).\end{aligned}\ ] ] where @xmath130 and @xmath131 are the current quark masses in the @xmath132 scheme , and @xmath133 is the energy eigenvalue of the nonrelativistic schr@xmath4dinger equation with a flavor - dependent potential models @xmath134 @xcite . according to the csqcdff @xcite
, the color - singlet nonperturbative matrix element @xmath125 can be related to the wave function at the origin . in the rest frame of the @xmath0\rangle$ ] quarkonium , it is can be separated the schr@xmath4dinger wave function into radial and angular pieces @xmath135 , where @xmath136 is the principal quantum number , and @xmath137 and @xmath64 are the orbital angular momentum quantum number and its projection . @xmath138 and @xmath139 are the radial wave function and the spherical harmonic function accordingly .
the wave function at the origin @xmath140\rangle}(0)$ ] and the first derivative @xmath141\rangle}(0)$ ] are related to the radial wave function at the origin @xmath142\rangle}(0)$ ] and the first derivative @xmath143\rangle}(0)$ ] , accordingly .
@xmath144\rangle}(0)&=&\sqrt{{1}/{4\pi}}r_{|(b\bar{s})[ns]\rangle}(0),\nonumber\\ \psi'_{|(b\bar{s})[np]\rangle}(0)&=&\sqrt{{3}/{4\pi}}r'_{|(b\bar{s})[np]\rangle}(0).\end{aligned}\ ] ]
the values of the wave function at the origin for the @xmath0\rangle$ ] quarkonium are related to the number of flavor quark @xmath145 , the regge slope @xmath146 , the @xmath147-loop beta - function coefficient @xmath148 and the @xmath149-loop beta - function coefficient @xmath150 , the renormalization scale parameters @xmath151 , and so forth , where @xmath132 is the modified minimal subtraction scheme . in calculating the wave function at the origin of the @xmath152\rangle$ ]
quarkonium within the five potentials @xcite , we adopt the regge slope @xmath153 , and the scale parameters @xmath151 as @xmath154=0.386 gev , @xmath155=0.332 gev , @xmath156=0.231 gev @xcite .
the current quark masses , i.e. , @xmath157 gev and @xmath158 gev , are chosen as derived in ref .
@xcite for calculating masses of the @xmath0\rangle$ ] quarkonium .
the masses of the @xmath0\rangle$ ] quarkonium ( @xmath2 meson ) under the buchm@xmath159ller and tye potential ( b.t .
potential ) @xcite are presented in table [ tabrpa ] , that can be derived through solving the energy eigenvalue of the schr@xmath4dinger equation @xcite .
the values of the constituent quark masses @xmath130 and @xmath131 for the @xmath0\rangle$ ] quarkonium and the accordingly quantities @xmath160\rangle}(0)|^2 $ ] , @xmath161\rangle}(0)|^2 $ ] , and @xmath162\rangle}(0)|^2 $ ] are presented in table [ tabrpb ] for the five potential models @xcite .
the five potentials are the b.t .
potential , the john l. richardson potential ( j. potential ) @xcite , the k. igi and s. ono potential(i.o .
potential ) @xcite , the yu - qi chen and yu - ping kuang potential(c.k .
potential ) , and coulomb - plus - linear potential ( the so - called cor .
potential ) @xcite , respectively . during the following calculation
, we adopt the values of wave functions at the origin and the constituent quark masses @xmath130 and @xmath131 for the @xmath0\rangle$ ] quarkonium under the b.t .
potential in table [ tabrpb ] ( with @xmath163 ) as the central values .
the input other parameters are adopted as the following values @xcite : @xmath164=80.399 gev , @xmath165 gev , @xmath166=0.88 , @xmath167=0.0406 , @xmath168 .
we set the renormalization scale to be @xmath169 of the @xmath170 quarkonium for leading - order @xmath171 running , which leads to @xmath172=0.26 .
furthermore , we adopt the constituent quark mass @xmath173 gev and @xmath174 gev for the @xmath175\rangle$ ] states , and @xmath176 gev and @xmath177 gev for the @xmath178\rangle$ ] states . as the masses of @xmath65-quark is greater than @xmath151 , so
the above - mentioned processes is pqcd calculable . to ensure the gauge invariance of the hard amplitude , the @xmath0\rangle$ ] quarkonium mass @xmath179
is set to @xmath180 , where @xmath181 and @xmath182 stand for constituent quark mass . as a reference for branching fractions
, we calculate the decay widths for the basic processes @xmath183 and @xmath184 .
their decay widths can be written as @xmath185.\end{aligned}\ ] ] where @xmath71 and @xmath186 are for processes @xmath183 , and @xmath73 and @xmath187 for @xmath184 .
@xmath188 stands for the relative momentum between the final two particles in the rest frame of the top quark or antitop quark .
gev and @xmath190 mev .
the decay widths and branching fractions for the @xmath0\rangle$ ] quarkonium states through the two decay channels , @xmath72\rangle+w^{+}s$ ] and @xmath74\rangle+w^{-}\bar{b}$ ] , are listed in tables [ tabrpc ] and [ tabrpd ] within the b.t .
potential .
.mass of the @xmath0\rangle$ ] quarkonium under the b.t . potential @xcite .
[ cols="^,^,^,^,^,^,^,^,^",options="header " , ] [ tabrpf ] and @xmath191 for @xmath35\rangle + sw^{+}$ ] , where contributions from the color - singlet @xmath10-wave states have been summed up . the up shaded band , the middle shaded band , the down shaded band are for the uncertainties of @xmath8\rangle$ ] , @xmath6\rangle$ ] , and @xmath178\rangle$ ] , respectively.,title="fig:",scaledwidth=40.0% ] and @xmath191 for @xmath35\rangle + sw^{+}$ ] , where contributions from the color - singlet @xmath10-wave states have been summed up . the up shaded band , the middle shaded band , the down shaded band are for the uncertainties of @xmath8\rangle$ ] , @xmath6\rangle$ ] , and @xmath178\rangle$ ] , respectively.,title="fig:",scaledwidth=40.0% ] and @xmath192 for @xmath35\rangle + sw^{+}$ ] , where contributions from the color - singlet @xmath10-wave states have been summed up . the up shaded band , the middle shaded band , the down shaded band are for the uncertainties of @xmath8\rangle$ ] , @xmath6\rangle$ ] and @xmath178\rangle$ ] , respectively.,title="fig:",scaledwidth=40.0% ] and @xmath192 for @xmath35\rangle + sw^{+}$ ] , where contributions from the color - singlet @xmath10-wave states have been summed up . the up shaded band , the middle shaded band , the down shaded band are for the uncertainties of @xmath8\rangle$ ] , @xmath6\rangle$ ] and @xmath178\rangle$ ] , respectively.,title="fig:",scaledwidth=40.0% ] the uncertainties are drawn by shaded bands in the figs . [ ds1ds2sum ] and [ dcos12cos13sum ] .
in the figs . [ ds1ds2sum ] and [ dcos12cos13sum ] , the up shaded band is the uncertainties of the @xmath8\rangle$ ] meson for the varying constituent quark masses , the middle shaded band is that of the @xmath6\rangle$ ] meson , the down shaded band is that of the @xmath178\rangle$ ] . in the above two shaded bands in the figs .
[ ds1ds2sum ] and [ dcos12cos13sum ] , where the solid - line in the center of the shaded bands is for @xmath173 gev and @xmath174 gev , the upper edge of the band is for @xmath193 gev and @xmath194 gev , which the correspondingly radial wave function at the origin is @xmath195\rangle}|^2 = 0.4133~gev^3 $ ] . and the lower edge of the bands is for @xmath196 gev and @xmath197 gev , and the correspondingly @xmath195\rangle}|^2 = 0.9404~gev^3 $ ] . in the lowest shaded bands ,
the solid - line in the center of the shaded bands is for @xmath176 gev and @xmath177 gev of the @xmath178\rangle$ ] constituent quark masses , the upper edge of the band is for @xmath198 gev and @xmath199 gev , and the correspondingly @xmath200\rangle}(0)|^2=1.999\times 10^{-2}~gev^5 $ ] , and the lower edge of the band is for @xmath201 gev and @xmath202 gev , and @xmath200\rangle}(0)|^2=4.472\times 10^{-2}~gev^5 $ ] , where the contributions of @xmath178\rangle$ ] form the color - singlet of @xmath11\rangle$ ] and @xmath13\rangle$ ] ( with @xmath203 $ ] ) states have been summed up .
adding all the uncertainties caused by the constituent quark masses @xmath204 gev and @xmath205 gev in quadrature and the corresponding wave functions at the origin for the process @xmath35\rangle
+ sw^{+}$ ] , we can obtain @xmath206\rangle + w^{+}s)}&=&4528^{+1149}_{-798.0}\;{\rm kev},\nonumber\\ \gamma{(t\to |(b\bar{s})[1 ^ 3s_1]\rangle + w^{+}s)}&=&8614^{+2799}_{-2155}\;{\rm kev},\nonumber\\ \gamma{(t\to |(b\bar{s})[1p]\rangle + w^{+}s)}~~&=&793.6^{+343.5}_{-192.9}\;{\rm kev}.\end{aligned}\ ] ] if the excited @xmath0\rangle$ ] quarkonium states decay to the ground spin - singlet @xmath5-wave state @xmath6\rangle$ ] with @xmath207 efficiency via electromagnetic or hadronic interactions , we can obtain the total decay width of the top quark decay channels under the b.t .
@xmath206\rangle + w^{+}s)}&=&14.19^{+4.36}_{-3.20}\;{\rm mev } \label{twbs1}.\end{aligned}\ ] ]
in the present paper , for studying the @xmath0\rangle$ ] quarkonium production through @xmath33-quark or @xmath34-quark decays , we calculate the masses of the @xmath0\rangle$ ] quarkonium under the b.t .
potential and the values of the schr@xmath4dinger radial wave function at the origin of the @xmath0\rangle$ ] quarkonium within the five potential models , and made a detailed study on the @xmath0\rangle$ ] quarkonium production via top quark and antitop quark semiexclusive decays channels , @xmath35\rangle + w^{+}s$ ] and @xmath74\rangle+w^{-}\bar{b}$ ] , within the csqcdff .
results for the @xmath0\rangle$ ] quarkonium states , i.e. , @xmath6\rangle$ ] , @xmath8\rangle$ ] , @xmath11\rangle$ ] , and @xmath13\rangle$ ] ( with @xmath208 $ ] ) have been presented . and to provide the analytical expressions as simply as possible , we have adopted the `` improved trace technology '' to derive lorentz - invariant expressions for top quark and antitop quark semiexclusive decay processes at the amplitude level .
such a calculation technology shall be very helpful for dealing with processes with massive spinors .
numerical results show that excited @xmath178\rangle$ ] or @xmath209\rangle$ ] states in addition to the ground @xmath210-wave states can also provide sizable contributions to the @xmath0\rangle$ ] or @xmath0\rangle$ ] quarkonium production through top quark decays , so one needs to take the excited states into consideration for a sound estimation .
if all the excited states decay to the ground state @xmath6\rangle$ ] , we can obtain the total decay width for the @xmath170 quarkonium production through top quark decays as shown by eq .
( [ twbs1 ] ) . at the lhc , due to
its high collision energy and high luminosity , sizable @xmath0\rangle$ ] or @xmath1\rangle$ ] quarkonium events can be produced in @xmath33-quark or @xmath34-quark decays ; i.e. , about @xmath211 @xmath170 or @xmath212\rangle$]@xmath213 quarkonium events per year can be obtained .
* acknowledgements * : we are grateful to professor xing - gang wu for many enlightening discussions .
this work was supported in part by the natural science foundation of china under grant no.11347024 , the scientific and technological research program of chongqing municipal education commission under grant no .
kj1401313 and the research foundation of chongqing university of science and technology under grant no .
ck2016z03 , the natural science foundation project of cqcstc under grant no . 2014jcyja00030 .
first , we introduce a masses spinor with negative helicity @xmath214 , which satisfies the following projection : @xmath215 where @xmath216 is an arbitrary light - like momentum , @xmath217 , @xmath218 . then , we construct the positive helicity states @xmath219 where @xmath220 is an arbitrary space - like momentum , @xmath220 is defined as @xmath221 and @xmath222 .
@xmath223 is satisfied with the projection : @xmath224 where @xmath225 . using the above identities
, one can write down the amplitude @xmath228 with four possible spin projections in the trace form for @xmath229-matrices : @xmath230\nonumber\\ m_{-s - s'}&=&{n}tr[(\slashed{k}+m_t)\omega_{+}\slashed{k}_0((\slashed{q}_1+m_s)\cdot{a})]\nonumber\\ m_{-ss'}&=&{n}tr[(\slashed{k}+m_t)\omega_{-}\slashed{k}_0\slashed{k}_1((\slashed{q}_1+m_s)\cdot{a})]\nonumber\\ m_{s - s'}&=&{n}tr[(\slashed{k}+m_t)\omega_{+}\slashed{k}_1\slashed{k}_0((\slashed{q}_1+m_s)\cdot{a})].\end{aligned}\ ] ] where @xmath231 is the amplitudes of the above - mentioned decays channels .
@xmath232 is the normalization constant .
it is easy check that @xmath228 is orthogonal for each other .
then , the squared amplitude @xmath179 can be written as @xmath233 where the four amplitude @xmath234 with ( @xmath235 ) are defined as @xmath236 then , the resultant @xmath234 are , @xmath243\\ m_2 & = & -{l_2 } \times tr [ ( \slashed{k } + m_{t}){\gamma _ 5}(\slashed{q}_{1}+{m_s})a]\\ m_3 & = & m_{3 ' } + { n_0 } [ ( { k } \cdot { q_3}){m_s } -({q_1 } \cdot { q_3 } ) m_{t}]m_2 \\ m_4 & = & m_{4 ' } - { n_0 } [ ( { k } \cdot { q_3}){m_s } + ( { q_1 } \cdot { q_3 } ) m_{t}]m_1\end{aligned}\ ] ] where @xmath244 and @xmath245 , \\ m_{4 ' } & = & \frac{n_0}{4l_1 } tr\left [ { ( \slashed{k } + m_{t}){\slashed{q_3}}({\slashed{q}_1 } + { m_s})a } \right ] .\end{aligned}\ ] ] r. aaij _ et al . _ ( lhcb collaboration ) , phys . rev . lett .
* 108 * , 101803 ( 2012 ) .
r. aaij _ et al . _
( lhcb collaboration ) , phys .
108 * , 201601 ( 2012 ) .
r. aaij _ et al . _
( lhcb collaboration ) , phys .
b * 707 * , 497 ( 2012 ) .
r. aaij _ et al . _
( lhcb collaboration ) , phys .
b * 713 * , 378 ( 2012 ) .
p. nason , s. dawson and r.k .
ellis , nucl .
b * 327 * , 49 ( 1989 ) ; nucl .
b * 335 * , 260 ( 1990 ) .
w. beenakker , h. kuijf , w. l. van neerven and j. smith , phys .
d * 40 * , 54 ( 1989 ) .
zhang , z.y .
fang , c.h .
chang , x.g .
wu , t. zhong and y. yu , phys .
d * 79 * , 114012 ( 2009 ) .
zhang , z.y .
fang , x.g .
wu , t. zhong , y. yu and j. jiang , eur .
j. c * 73 * , no . 6 , 2464 ( 2013 ) .
m. artuso _ et al . _
( cleo collaboration ) , phys .
* 95 * , 261801 ( 2005 ) .
huang _ et al .
_ ( cleo collaboration ) , phys . rev .
d * 75 * , 012002 ( 2007 ) .
r. sia and s. stone , phys .
d * 74 * , 031501 ( 2006 ) ; phys .
d * 80 * , 039901 ( 2009 ) .
drutskoy , f.k .
guo , f.j .
llanes - estrada , a.v .
nefediev , and j.m .
torres - rincon , eur .
j. a * 49 * , 7 ( 2013 ) . c. oswald and t.k .
pedlar , mod .
a * 28 * , 1330036 ( 2013 ) .
g. abbiendi _ et al .
_ ( opal collaboration ) , eur .
j. c * 11 * , 587 ( 1999 ) .
c.h . chang and y.q
. chen , phys .
d * 48 * , 4086 ( 1993 ) ; c.h .
chang , y.q .
chen , g.p .
han , and h.t .
jiang , phys .
b * 364 * , 78 ( 1995 ) ; c.h .
chang and x.g .
j. c * 38 * , 267 ( 2004 ) ; r.m .
thurman - keup , a.v .
kotwal , m. tecchio , and a.b .
wagner , rev .
phys * 73 * , 267 ( 2001 ) . q.l . liao and g.y .
xie , phys .
d * 90 * , 054007 ( 2014 ) .
e. eichten , k. gottfried , t. kinoshita , k.d . lane , and t.m .
yan , phys .
d * 17 * , 3090 ( 1978 ) ; * 21 * , 313 ( e ) ( 1980 ) ; * 21 * , 203 ( 1980 ) ; e. eichten and f. finberg , phys .
d * 23 * , 2724 ( 1981 ) .
chang , c. driouich , p. eerola , and x.g .
wu , comput .
. commun . * 159 * , 192 ( 2004 ) ; c.h .
chang , j.x .
wang , and x.g .
wu , comput .
phys . commun . * 174 * , 241 ( 2006 ) ; * 175 * , 624 ( 2006 ) ; x.y .
wang and x.g .
wu , comput .
* 183 * , 442 ( 2006 ) . | in this work we evaluate the masses of the @xmath0\rangle$ ] or @xmath1\rangle$ ] quarkonium ( @xmath2 or @xmath3 meson ) under the b.t .
potential , and the values of the schr@xmath4dinger radial wave function at the origin of the @xmath0\rangle$ ] or @xmath1\rangle$ ] quarkonium within the five potential models .
then we investigate a systematic study on the production of the @xmath0\rangle$ ] or @xmath1\rangle$ ] quarkonium via top quark or antitop quark decays in the color - singlet qcd factorization formula ( csqcdff ) , i.e. , the two @xmath5-wave states , @xmath6 \rangle$ ] ( or @xmath7 \rangle$ ] ) and @xmath8 \rangle$ ] ( or @xmath9 \rangle$ ] ) , and its four @xmath10-wave excited states , @xmath11 \rangle$ ] ( or @xmath12 \rangle$ ] ) and @xmath13 \rangle$ ] ( or @xmath14 \rangle$ ] ) ( with @xmath15 $ ] ) . for deriving compact analytical results for complex processes ,
the improved trace technology " is adopted to deal with the decay channels at the amplitudes . moreover
, various differential distributions and uncertainties of the concerned processes are analyzed carefully . by adding the uncertainties caused by the @xmath16 and @xmath17-quark masses in quadrature
, we obtain @xmath18\rangle + w^{+}s)}=14.19^{+4.36}_{-3.20}$ ] mev . at the lhc with the luminosity @xmath19 and
the center - of - mass energy @xmath20 tev , sizable @xmath0\rangle$ ] or @xmath1\rangle$ ] meson events can be produced through @xmath21-quark or @xmath22-quark decays ; i.e. , about @xmath23 @xmath24 or @xmath25 events per year can be obtained .
+ * pacs numbers : * 12.38.bx , 14.65.ha , 14.80.bn , 14.40.pq |
binaries are believed to constitute a significant fraction of the local stellar population ( e.g. , duquennoy a. & mayor 1991 , suchkov & mcmaster 1999 ) .
because of that they contribute substantially to the statistics of age , metallicity , and kinematics of the main sequence stars in the solar neighborhood
. since the evolution of the local galactic disk is deduced mostly from these statistics ( e.g. , twarog 1980 , carlberg et al .
1985 , marsakov et al .
1990 , meusinger et al .
1991 , edvardsson et al . 1993 ,
caloi et al . 1999 ) , the presence of binaries may affect inferences regarding the history of star formation and metal enrichment in the disk if the population of these stars is statistically different from that of the single stars . in this paper
we compare the kinematics and metallicity of binary and single f stars to see if they belong to the same statistical ensemble .
we use a sample of the _ hipparcos _ f stars ( @xmath0 ) that have photometry in hauck & mermilliod ( 1998 ) , and treat separately the stars marked in _
hipparcos _ as single stars , unresolved binaries , and resolved binaries .
the latter two groups of stars will be referred to as _ known _ binaries to differentiate them from the binary candidates ( c binaries ) discussed below .
all the stars have tangential velocity obtained from the _ hipparcos _ parallax and proper motion . about a third of them have also radial velocities ( taken mostly from ochsenbein f. 1980 , barbier - brossat et al .
1994 , and duflot et al .
1995 ) , which have been used to compute total spatial velocity .
metallicity has been derived from the data using the metallicity calibration from carlberg et al .
effective temperature , @xmath1 , has been computed from the @xmath2 color , with the algorithm based on moon ( 1985 ) , moon & dworetsky ( 1985 ) .
the sample is constrained to the stars within 200 pc , metallicity range @xmath3}$}\ } < 0.5 $ ] , magnitude accuracy better then 0.003 ( in @xmath4 ) , parallax accuracy better than 3 mas , and proper motion accuracy better than 3 mas in both coordinates .
many of the nearby single stars are in fact yet unidentified binaries ( e.g. , duquennoy & mayor 1991 ) .
numerous binary candidates , called c binaries , were identified in suchkov & mcmaster ( 1999 ) .
the criterion used to isolate c binaries involves the difference between the star s absolute magnitude derived from the _ hipparcos _ parallax and the johnson @xmath4 magnitude , @xmath5 , and the absolute magnitude computed from the dereddened color index @xmath6 , @xmath7 ( the computation of @xmath7 utilizes the algorithm based on moon 1985 , moon & dworetsky 1985 ) .
the discrepancy between the two magnitudes , @xmath8 ( @xmath5 too bright for the star s @xmath9 ) , appears to be a strong signature of a binary system .
so , to reduce the contamination in the sample of the single stars with binaries , we have excluded the stars with @xmath10 ( c binaries at the 3-@xmath11 level ) from that sample .
the same rejection criterion has been applied to the sample of the binary stars to avoid any differential bias .
the anomalously bright `` single '' stars , @xmath10 , have been isolated into a sample of c binaries .
a summary of kinematics and metallicity of all groups of stars considered here is given in table 1 .
lcccr + sample & @xmath12 & @xmath11 & [ fe / h ] & n + & ( mag ) & ( ) & ( @xmath13 ) & + & + single & @xmath14 & 17.1 @xmath15 0.2 & @xmath16 & 6482 + binary & @xmath14 & 15.2 @xmath15 0.4 & @xmath17 & 591 + `` single '' & @xmath18 & 19.7 @xmath15 0.3 & @xmath19 & 1581 + binary & @xmath18 & 17.0 @xmath15 0.5 & @xmath20 & 641 + & + single & @xmath14 & 22.0 @xmath15 0.4 & @xmath21 & 1414 + binary & @xmath14 & 15.9 @xmath15 0.9 & @xmath22 & 147 + `` single '' & @xmath18 & 23.4 @xmath15 0.9 & @xmath23 & 332 + binary & @xmath18 & 17.0 @xmath15 0.8 & @xmath24 & 203 +
figure [ fig - knm ] compares tangential velocity dispersion of the single f stars , known unresolved binaries , and c binaries , revealing a dramatic difference between the three populations .
first , the known binary f stars have significantly smaller velocity dispersions at all distances from the sun .
the latter fact rules out a possible selection bias against distant binaries as a cause of the kinematic differences between the three populations . as seen in table 1 ,
the difference in spatial velocity dispersion is even more dramatic .
we have checked if the sample color selection criterion , @xmath25 , introduces a bias favoring hotter binaries and/or cooler single stars , which could make the sample binary stars younger just because of different age sampling at different effective temperatures . such a bias can arise if , for whatever reasons , binaries are hotter than single stars at the same @xmath26 color
. if this is the case , the relationship between the effective temperature , @xmath27 ( derived from spectroscopic analysis ) , and @xmath26 must be different for the binaries and the single stars .
this is tested in figure [ fig - tef ] , where @xmath27 is from cayrel de strobel et al .
comparison of the upper and lower panels shows that within the uncertainties the two relationships are identical , which rules out any age bias associated with the color selection criterion . among other
known selection effects none was found to introduce a bias favoring low - velocity binary stars .
so , the kinematic discrepancy between the known binaries and the single stars in figure [ fig - knm ] must be intrinsic . for the stellar populations in the solar neighborhood
, smaller velocity dispersion implies younger age .
therefore , the above result means that the known binary f stars are , on average , younger than single f stars .
this conclusion is further supported by the respective discrepancy in the mean metallicity of the two populations .
figure [ fig - feh ] shows that the binary stars are , on average , more metal rich , implying that they are statistically younger than single stars . as with the kinematics ,
the difference in metallicity is present at all distances , and none of the known selection effects was found to introduce a bias favoring metal rich binary stars .
this argues that the metallicity difference and the corresponding age discrepancy between the known binaries and the single stars are real .
if the population of known binaries is indeed younger than that of single f stars , the questions is : why ?
an obvious place to look for an answer is the interaction of the binary components .
depending on separation , the components of close binaries can strongly interact with each other , including mass loss and/or mass transfer , which drastically changes their evolution .
the interaction effects are believed to make the cm diagrams of star clusters look much different from what is expected on the basis of the standard stellar evolution theory as applied to a coeval population of single stars ( see , e.g. , pols & marinus 1994 , and references therein ) .
so , it would be natural to associate the younger , on average , age of the binary f stars with the interaction of binary components in tight pairs , which somehow reduces the main sequence lifetime of the primary component .
this hypothesis is supported by figure [ fig - i1i2mh ] , which provides evidence that component separation is relevant to how long the primary star may live on the ms .
figure [ fig - i1i2mh ] compares the velocity dispersions of the spatially resolved binary f stars ( in fact , these are the binary components having individual entries in the _ hipparcos _ catalog ) with those of the unresolved double stars .
statistically , the spatially resolved binary pairs have obviously wider separations than the unresolved binaries . therefore ,
if there is any impact of component interaction on stellar evolution , it must be less significant among the resolved binaries , so the populations of resolved and unresolved binaries must exhibit statistical differences arising from different stellar evolution in wide and tight binary pairs .
figure [ fig - i1i2mh ] shows that this is indeed the case .
the kinematics of the unresolved binaries turns out to be younger ( smaller velocity dispersion ) than that of the resolved binaries at any distance from the sun , indicating that binary pairs with smaller component have shorter lifetimes .
a crude estimate of the age difference between the binary and single f stars can be obtained using age
velocity relation ( avr ) . to illustrate this , we have derived the avr for the subsample of f stars constrained to the range of @xmath12 shown by the best known single stars within 25 pc , @xmath28 , and metallicity range @xmath29}$}\ } \leq 0.01 $ ] .
the isochrones for isochrone fitting are from demarque et al .
1996 , for the composition matching the metallicity range of the selected stars : @xmath30 , @xmath31 .
this avr is shown in figure [ fig - avr ] .
the slope of the linear regression in figure [ fig - avr ] indicates that a 1 difference in tangential velocity dispersion , @xmath32 , corresponds to a @xmath33 gyr difference in average age .
the unresolved binaries have @xmath32 about @xmath34 smaller than @xmath32 of the single stars ( table 1 ) , which means that they are statistically _ younger _ by at least @xmath35 gyr .
in contrast to the known binaries , the c binaries turn out to be kinematically _ older _ than the single f stars . one can infer this from figure [ fig - knm ] and table 1 , where higher velocity dispersions of the c binaries indicate older age . from the slope of the avr in figure [ fig - avr ] and the average difference in tangential velocity dispersions of @xmath35 ( table 1 )
, we conclude , that the age difference between the two population is @xmath35 gyr .
there is evidence that the older age of c binaries is associated with retarded stellar evolution of a primary f star in a binary system , with @xmath12 being a measure of retardation ( suchkov 1999 ) ; the retardation probably results from component interaction in tight binary pairs , which is suggested by a strong correlation between @xmath12 and binary components separation found for known binaries .
however , if c binaries are double star systems similar to known binaries , the question is , why they evolve differently .
we believe that the solution to this puzzle will eventually be found in the details of interaction in tight binary pairs
. the outcome of the binary components interaction in terms of its impact on stellar evolution should depend on the combination of the binary system parameters , such as the component mass ratio , separation , orbit eccentricity , etc .
it seems plausible that depending on that combination , the interaction of binary components can alter the standard evolution of the main sequence primary either by affecting the hydrogen - burning core and prolonging the star s main sequence lifetime , as implied in some scenarios of blue straggler formation ( cf .
wheeler 1979 ; see also livio 1991 ) , or by various mass loss / transfer events that interrupt the normal main sequence evolution ( e.g. , pols & marinus 1994 ) or both .
the old age of c binaries suggests that the former mechanism statistically dominates the evolution of these stars .
one can speculate that this is because the secondary in many of these systems is probably too small ( substellar ) to evoke appreciable mass loss / transfer but is very close to the primary to tidally induce instabilities leading to enhanced internal mixing , which provides a fresh supply of hydrogen to the core .
in the case of known binaries , both components are stellar , so both of the above mechanisms can operate in these systems .
this means that although stellar evolution in a tight pair can be slower , the primary can prematurely cease to exist as a main sequence f star because of a strong mass loss occurring at some point in the evolution of the binary .
in fact , the latter scenario predicts that the known binaries with far evolved primaries should be , on average , older than expected from the standard stellar evolution models .
this prediction can be verified by comparing velocity dispersions of binaries with different @xmath12 . as mentioned above
, @xmath12 appears to provide a measure of stellar evolution retardation .
the larger @xmath12 the larger amount of extra time spent by the star on the main sequence , so the stars with large @xmath12 must be statistically older than those with small @xmath12 .
this seems to be indeed the case .
based on the age
velocity relation in figure [ fig - avr ] , the difference in tangential velocity dispersion of @xmath36 between the two groups of known binaries ( see table 1 ) implies that those with a strong signature of retardation , @xmath10 , are older ( statistically ) by @xmath35 gyr , this discrepancy is an indicator that such a primary evolves at a much slower pace than expected , so the standard isochrones underestimate its age . ] . in summary
, we conclude that the populations of binary and single f stars do not belong to the same statistical ensemble .
the main difference is in the mean age of the populations , which has important implications for the modeling of both stellar evolution and galactic evolution . | we have compared the kinematics and metallicity of the main sequence binary and single f stars from the _ hipparcos _ catalog to see if the populations of these stars originate from the same statistical ensemble .
the velocity dispersions of the known unresolved binary f stars have been found to be dramatically smaller than those of the single f stars .
this suggests that the population of these binaries is , in fact , younger than that of the single stars , which is further supported by the difference in metal abundance : the binaries turn out to be , on average , more metal rich than the single stars .
so , we conclude that the population of these binaries is indeed _ younger _ than that of the single f stars .
comparison of the single f stars with the c binaries ( binary candidates identified in suchkov & mcmaster 1999 ) has shown , on the other hand , that the latter stars are , on average , _ older _ than the single f stars .
we suggest that the age difference between the single f stars , known unresolved binaries , and c binaries is associated with the fact that stellar evolution in a binary systems depends on the binary components mass ratio and separation , with these parameters being statistically very different for the known binaries and c binaries ( e.g. , mostly substellar secondaries in c binaries versus stellar secondaries in known binaries ) .
in general we conclude that the populations of known binaries , c binaries , and single f stars do not belong to the same statistical ensemble .
the implications of the discovered age difference between these populations along with the corresponding differences in kinematics and metallicity should be important not only for understanding the evolution of stars but also for the history of star formation and the evolution of the local galactic disk . |
a percolating cluster in an infinite system is defined as an unbounded cluster ; that is , a cluster which is not the subset of a compact observation window .
the percolation threshold in the intensity @xmath6 is the intensity at which a percolating cluster emerges , more precisely , it is the infimum of all intensities @xmath5 at which a percolating cluster appears ( almost surely ) .
alternatively , the percolation threshold @xmath7 in the occupied area fraction can be considered ( if the mean area of a typical grain is strictly positive ) .
a phase transition occurs if the percolation threshold takes a non - trivial value @xmath8 ( e.g. , for boolean models with a bounded grain size ) . above the percolation threshold , the probability that the cluster spans an observation window converges in the limit of an infinitely large window to unity .
- or in @xmath1-direction , respectively .
the figure shows a boolean model of fully aligned rectangles .
the two percolating clusters in @xmath0-direction , which span the finite sample horizontally , are colored blue and red .
there is no connected path spanning the system vertically.,scaledwidth=55.0% ] in a finite sample , no infinite cluster can appear .
the percolating cluster is therefore defined as a cluster that connects opposite sides of the sample . in other words ,
a cluster that percolates in @xmath0- or @xmath1-direction spans the system horizontally or vertically , respectively .
figure [ fig_perco_sample ] shows a finite sample in which two clusters span the system in @xmath0- but none in @xmath1-direction .
phase transitions occur only in the thermodynamic limit .
however , there is an apparent percolation threshold close to which the probability for the appearance of a percolating cluster steeply increases . if the elongated particles have an orientation bias in @xmath0-direction , a finite anisotropic sample percolates more likely in @xmath0- than in @xmath1-direction .
in other words , the occupied volume fraction at which the probability that the system percolates in @xmath0-direction with a given probability is lower than the corresponding volume fraction in @xmath1-direction .
however , we show in this section that the difference vanishes in the limit of an infinite system size , no matter how anisotropic the system may be .
we relate this observation to the uniqueness of the percolating cluster .
the ( isotropic ) boolean model is not only a benchmark model of continuum percolation but also a popular model of heterogeneous materials @xcite .
examples include fractured materials or hydrating cement - based materials @xcite , sedimentary rock @xcite and wood composites @xcite .
figure [ fig_boolean_model_def ] visualizes the definition of the boolean model .
first , points are randomly placed in space distributed uniformly .
they form a completely independent point process , the so - called poisson point process .
the number of points in a finite observation window follows a poisson distribution . in the second step ,
each point is decorated with a grain .
the shape and orientation follows the so - called grain distribution .
we here only consider stationary processes , that is , statistically homogeneous models for which both intensity and orientation distribution of the grains are spatially constant .
we use grains with a constant shape , namely , rectangles with a fixed size and aspect ratio . to explore a range of different degrees of anisotropy in our simulations
, we use a parametric orientation distribution from @xcite .
the anisotropy can be tuned by adjusting the orientation distribution of the grains .
the orientation of a grain is defined by the angle @xmath9 between its main axis and the @xmath0-axis .
its probability density function is @xmath10 see figure [ fig_mil_orientation_distribution ] .
the parameter @xmath11 controls the degree anisotropy of the system : @xmath12 results in a uniform distribution and hence to an isotropic system ; the larger @xmath11 the stronger the anisotropy ; we define that @xmath13 corresponds to a full alignment with the @xmath0-axis .
the orientation of a single grain is parametrized by the angle @xmath9 between the main axis of the grain and the @xmath0-axis of the coordinate system .
the probability density function @xmath14 of this angle determines the anisotropy of the boolean model .
two samples are depicted on the left hand side of the figure.,scaledwidth=84.0% ] the expected area fraction of the grain phase is called the occupied area fraction @xmath15 .
it is independent of the orientation distribution and given by @xmath16 where @xmath17 is the area of a single grain @xcite .
a dependency on the intensity always corresponds to an dependency on the occupied area fraction .
percolation in an infinitely large system is a geometric phase transition . above the critical occupied area fraction
@xmath7 there is a percolating cluster with probability one , and below this threshold there is almost surely no percolation of the grains .
the probability that there is a percolating cluster as a function of the occupied area fraction @xmath15 is called the connectivity @xmath18 .
it is a step function that changes its functional value at @xmath7 from zero to unity , see figure [ fig_perco_geom_phase_trans ] .
we are interested in the behavior of finite systems . in this case , there is no phase transition .
depending on statistical fluctuations , a system spanning cluster may appear or it may not appear . for all occupied area fractions @xmath19
there is a finite probability that there is a percolating cluster .
the connectivity is expected to be a smooth and to take on values @xmath20 .
the connectivity can be estimated from simulated boolean models simply by the fraction of samples that contain a percolating cluster .
we use polygons for an exact representation of the boolean model to avoid strong pixelation errors that would occur in pixelated boolean models even for fine pixelations ( because unconnected clusters can become connected or vice versa ) .
we simulate anisotropic boolean models at different expected area fractions and for various sizes of the simulation box , and determine in each case the connectivity .
figure [ fig_perco_finite_system_connectivity ] shows ( for a boolean model of aligned rectangles with aspect ratio @xmath21 ) , as expected , a smooth increase of the connectivity for an increasing occupied area fraction @xmath15 . in figure [ fig_perco_finite_system_connectivity ]
, we distinguish the connectivity in @xmath0- and @xmath1-direction in a boolean model with fully horizontally aligned rectangles ( @xmath13 ) .
the connectivity @xmath22 is the probability that there is a cluster that spans the system in @xmath0-direction , and accordingly the connectivity @xmath23 considers clusters that span the system in @xmath1-direction .
more particles are needed for vertical than for horizontal percolation , and a fluctuation that spans the system horizontally is more likely than vertically .
the system is more likely to percolate horizontally than vertically ( at least close to the percolation transition ) . even for relatively large system sizes ,
there are occupied volume fractions at which the system most likely percolates in @xmath0- but not in @xmath1-direction .
since there is no phase transition in a finite system , the percolation threshold @xmath7 is not well - defined . instead an _ effective percolation threshold _
@xmath24 must be defined .
a simple and intuitive definition is via @xmath25 a trivial estimate inverting the estimated connectivity as a function of the occupied area fraction would not be robust against statistical fluctuations .
a more robust estimate has been defined in @xcite .
the derivative of the connectivity @xmath26 as a function of the occupied area fraction @xmath15 is the probability that the system starts to percolate at @xmath15 .
another possibility to define an effective percolation threshold is @xmath27 @xcite .
because of the phase transition , the behavior of the system is close to the percolation threshold and for large systems dominated by the universal scaling . if the derivative of the connectivity @xmath26 is approximated by a gaussian distribution , the connectivity can be approximated by an error function with a constant offset and prefactor : @xmath28 fitting this rescaled error - function to the connectivity as a function of @xmath15 , where @xmath29 and @xmath24 are the fit parameters , see figure [ fig_perco_finite_system_connectivity ] , provides a numerically robust estimate of the effective percolation threshold @xmath24 . in the limit of infinite system size , the error function converges to a step function and @xmath30 .
we adopt this procedure , but we use the intensity @xmath5 as an input instead of the occupied area fraction @xmath15 . in a finite system , there are statistical fluctuations in @xmath15 . to avoid systematic errors ,
we replace the volume fractions in ( [ eq : erffit ] ) by the corresponding intensities .
the definition of the effective percolation threshold via ( [ eq : erffit ] ) has the advantage that it can be easily interpreted and implemented . however , there are two major disadvantages .
first , unnecessary samples are simulated especially at high thresholds ; they are computationally expensive to simulate because of the large number of grains .
second , the precision is limited by unknown systematic errors .
although the error function is a good approximation of the connectivity as a function of the occupied area fraction , high precision estimates of the connectivity reveal in finite systems statistically significant deviations from an error function .
the statistical errors in our simulation are of the order @xmath31 , whereas systematic errors are of @xmath32 .
there are more efficient algorithms using inhomogeneous boolean models for more precise estimates of the percolation threshold of homogeneous models ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in future research
, these algorithms could be generalized to anisotropic systems . for our analysis of anisotropic continuum percolation ,
the procedure based on ( [ eq : erffit ] ) is sufficient .
an error function is fitted to each curve in figure [ fig_perco_finite_system_connectivity ] .
the resulting estimates of the effective percolation thresholds are depicted by vertical dashed lines .
the effective percolation thresholds in @xmath0-direction are distinctly smaller than in @xmath1-direction for all simulated systems sizes @xmath33 .
however , the differences in the effective percolation thresholds depend on the system sizes .
the differences for horizontal and vertical directions become smaller with increasing system size . because percolation is a phase transition ,
the scaling of the effective percolation thresholds with the system size is universal for large system sizes .
the scaling exponent does not depend on the grain distribution , it is even the same for continuum percolation and lattice percolation ; see ( e.g. , * ? ? ?
* ; * ? ? ?
* ) : @xmath34 with the universal critical exponent @xmath35 @xcite . , see ( [ eq : scaling ] ) , for an isotropic boolean model ( @xmath12 ) with aspect ratio @xmath36 .
the unit of length for the linear system size @xmath33 is the long semiaxis of the rectangle .
the color bands represent the bands of one and two standard deviations.,scaledwidth=47.9% ] the percolation threshold @xmath7 can be estimated from the effective thresholds @xmath24 by an extrapolation from the finite system sizes @xmath33 to @xmath37 .
figure [ fig_perco_scaling_isotropic ] shows for an isotropic boolean model with rectangles with aspect ratio @xmath36 the numerical estimates of @xmath24 , which are derived from fits of error functions to the connectivity ( in @xmath0-direction ) .
it also plots the extrapolation to @xmath7 and the corresponding bands of one or two standard deviations .
+ + for anisotropic models , we have to consider that the proportionality constant is nonuniversal .
it varies for the effective percolation thresholds @xmath38 in @xmath0- and @xmath39 in @xmath1-direction .
therefore , we a priori distinguish also in the limit of infinite system size the percolation thresholds in @xmath0- or in @xmath1-direction .
in other words , we estimate the thresholds in @xmath0- or in @xmath1-direction separately : @xmath40 however , we will show that even for the most anisotropic systems @xmath41 .
we simulate boolean models with rectangles for different values of @xmath11 , that is , for different degrees of anisotropy .
their aspect ratios @xmath42 are fixed within one model , but we simulate several models with different ratios @xmath43 and anisotropy parameters @xmath44 .
the models thus vary from isotropic orientation distributions to full alignment . for each system
, we determine the connectivity @xmath18 as a function of the intensity for differently large simulation boxes with linear system sizes @xmath45 .
as discussed above , the more anisotropic a boolean model the larger the simulation boxes must be so that the connectivity is accurately described by an error function ( so that the finite size scaling can be applied ) .
for each system , we fit error functions separately to the both the connectivity in @xmath0- and in @xmath1-direction .
the fit is applied to the five largest system sizes which are simulated for the according boolean model .
we thus estimate the effective percolation thresholds @xmath38 and @xmath39 .
next we extrapolate the effective percolation thresholds to the limit of infinite system size using ( [ eq : scaling_xy ] ) with fit parameters @xmath46 and @xmath47 or @xmath48 and @xmath49 . figure [ fig_perco_scaling ] shows for the six exemplary boolean models the effective percolation thresholds @xmath38 and @xmath39 as a function of the rescaled system size .
it also plots the extrapolation and the corresponding bands of one or two standard deviations . within statistical significance
, the percolation threshold is isotropic . even for the most anisotropic systems ,
we find @xmath41 .
similarly , balberg et al .
@xcite found isotropic percolation thresholds in two- and three - dimensional anisotropic systems of sticks .
this isotropy of the percolation threshold , which means that the boolean model either percolates simultaneously in all directions or not at all , can be expected for boolean models for at least three reasons : first , consider the extreme case of aligned rectangles .
a dilation can transform a boolean model with aligned squares into a boolean model with aligned rectangles of arbitrary aspect ratios .
however , the affine transformation neither changes the area fraction nor the topology of the model .
therefore , the percolation threshold for aligned rectangles must be independent of the aspect ratio . moreover , because the model can either be dilated in @xmath0- or @xmath1-direction , the percolation threshold must be isotropic .
balberg and binenbaum @xcite provide a similar argument for the anisotropic systems of sticks .
they also discuss a renormalization - group argument why in an anisotropic square lattice an isotropic percolation threshold is expected .
another plausibility argument for the isotropic percolation threshold is that , as indicated in section [ sec : percolation_connectivity ] , percolation in @xmath0-direction is almost surely not possible along a straight line but only along a winded path .
this requires that also grains with different @xmath1-coordinates are connected .
the most fundamental reason for the isotropy of the percolation threshold is that it is closely related to the uniqueness of the percolating cluster in continuum percolation .
meester and roy @xcite have proven for boolean models of spheres with varying size that there can be at most one unbounded component , but the results can be generalized to grains with general shape , see also @xcite . either with probability one or zero
there is either exactly one or no percolating cluster , in other words , the percolating cluster is ( almost surely ) unique .
similar proofs could be derived for other percolation models ( e.g. , * ? ? ?
* ; * ? ? ?
but not in @xmath1-direction , there would appear with finite probability two percolating clusters ( dark blue and cyan ) not connected to each other ( indicated by the red line that lies wholly in the void phase).,scaledwidth=84.0% ] the isotropy of the percolation threshold follows by a heuristic argument from the uniqueness of the percolating cluster .
if we assume that the unbounded cluster spans the system only in @xmath0- but not in @xmath1-direction , two percolating clusters that are not connected to each other can appear with finite probability , see figure [ fig : two_percolating_clusters ] .
assume that the percolating ( dark blue ) cluster is bounded in @xmath1-direction , that is , the cluster lies completely below some bound ( maximum of red line ) . because of the statistical homogeneity of the system , the probability that there is another percolating cluster in the domain above the bound ( colored cyan in figure [ fig : two_percolating_clusters ] ) should be nonzero
. however , this would lead to a contradiction of the uniqueness of the unbounded cluster .
therefore , we conjecture that in a statistically homogeneous system with almost surely at most one unique unbounded component this percolating cluster must be unbounded both in @xmath0- and in @xmath1-direction , and the percolation thresholds must coincide : @xmath50 the percolation threshold is by plausibility argument and empirical findings confirmed to be isotropic .
so the finite size scaling from ( [ eq : scaling_xy ] ) can be replaced by @xmath51 where the percolation threshold in the infinite system @xmath7 is isotropic .
it is because of the nonuniversal prefactors @xmath52 and @xmath53 that the effective percolation thresholds differ in @xmath0- and @xmath1-direction . of boolean models with orientation distributions with different degrees of anisotropy ( @xmath12 randomly oriented and @xmath13 aligned rectangles ) as a function of their aspect ratio @xmath42 .
samples of rectangles ( left ) or squares ( right ) are depicted that are aligned ( top ) or randomly oriented ( bottom ) . ]
we simultaneously extrapolate the effective percolation thresholds @xmath38 and @xmath39 to the same @xmath7 using a maximum likelihood fit described in [ sec : appendix ] .
the final results for the isotropic percolation thresholds @xmath7 are plotted in figure [ fig : percolation_thresholds_isotropic ] .
if the grains are aligned , the percolation threshold does not depend on the aspect ratio of the grains , because in this special case changes in the aspect ratio of the single grains correspond to dilations of the boolean model , as mentioned above .
the latter do not change the topology and therefore not the percolation threshold .
for all other orientation distributions of the grains , the percolation threshold decreases with a decreasing ratio of smaller to larger side length .
put differently , the more elongated a rectangle with rotational degrees of freedom the smaller the percolation threshold . in the limit of stick percolation , that is , for vanishing aspect ratio @xmath54 , the percolation threshold vanishes as well @xmath55 .
comparing different orientation distribution , we find in agreement with previous findings @xcite , the more anisotropic the orientation distribution , the larger is the percolation threshold .
numerical estimates of the percolation threshold can get computationally very expensive .
however , applications might need the threshold for some specific model parameters for which no simulation results are yet available . a possibility to predict the percolation threshold without the need for expensive simulations
would be of great importance ; that is , an explicit formula approximating the percolation threshold
. such an explicit formula should be easy computable and as accurate as possible @xcite .
an example is the scaling relation for varying dimensions from @xcite , which is based on a lower bound @xcite and a conjecture that no convex shape provides higher thresholds than the hyperspheres .
a common approximation is the excluded area approximation ( or excluded volume in three dimensions ) @xcite .
it tries to predict percolation thresholds based on simulation results of similar models by approximating the dependence of the critical intensity on the model parameters by the change in the excluded area .
the excluded area for convex objects can be expressed explicitly by a so - called mixed intrinsic volume from integral geometry . in section [ sec : excluded - area - approx ] , we compare the approximation of the excluded area to the percolation thresholds for the models with different degrees of anisotropy discussed above .
in contrast to the excluded area approximation , which uses an empirical parameter , mecke and wagner @xcite suggested a purely geometrical approximation .
it uses an additive topological constant from integral geometry , the euler characteristic . for a compact set
, it is given by the number of connected components minus the number of holes .
the critical intensity is approximated by the zero of its mean value ( as a function of the intensity ) .
the euler characteristic is positive for intensities below this zero and negative above . intuitively speaking ,
this indicates a change from a predominant role of single clusters to a network - like structure .
explicit formulas for the mean euler characteristic are known for very general boolean models @xcite , that potentially allow for estimates of the percolation threshold .
at least for the parametric boolean models studied here , the approximation is a lower bound , see section [ sec : euler ] .
we expect it to be a lower bound for even more general boolean models , because the euler characteristic does not distinguish holes in connected clusters from the percolating void phase . moreover , we find that the trend , i.e. , the qualitative behavior , is very well described by the zero of the euler characteristic . for not too small aspect ratios of the rectangles ,
this finding allows us to very accurately estimate the percolation thresholds of boolean models if the percolation threshold is known for squares that follow the same orientation distribution as the rectangles .
the euler characteristic belongs to a class of so - called minkowski functionals ( or intrinsic volumes ) from integral geometry @xcite . in two - dimensions ,
the two other minkowski functionals are area and perimeter .
we investigate the potential of further approximations of the percolation threshold that do not use an empirical parameter , namely , the extremal points of the variances and covariances of the minkowski functionals . in section [ sec_percolation_cov ] , we given an outlook on candidates for close upper bounds on the percolation thresholds .
of the two rectangles @xmath56 and @xmath57 is the area of the green octagon : if the center of @xmath57 is within this region , @xmath56 and @xmath57 intersect .
@xmath58 , where @xmath59 is the angle between the two rectangles , each with side lengths @xmath60 and @xmath61.,scaledwidth=40.0% ] the excluded area is defined for two particles @xmath56 and @xmath57 .
it is the area of the region defined by all those center positions of the second particle @xmath57 which leads to intersection of the two particles @xcite . in the following , the number of grains that intersect an object
is called its number of bonds . in other words
, the second particle @xmath57 can not enter this area without intersecting @xmath56 .
figure [ fig_perco_a_ex ] visualizes the excluded area of two rectangles .
the excluded area @xmath62 of two convex grains @xmath56 and @xmath57 can be explicitly expressed using the so - called mixed functional @xmath63 from integral geometry @xcite , @xmath64 note that our notation differs from @xcite , in which the same quantity is denoted by @xmath65 , but @xmath66 indicates the total excluded area .
averaging @xmath67 with the orientation distribution @xmath68 of the grains @xmath56 and @xmath57 yields the averaged excluded area for single grains @xmath69 where @xmath70 denotes a rotation by the angle @xmath9 .
for the isotropic orientation distribution @xmath71 , the average mixed intrinsic volume separates . in this special case , it is proportional to the product of the perimeters of the grains . for arbitrary dimensions ,
a general formula for the exclusion volume of identical randomly oriented convex particles is given is given in @xcite .
therefore , the excluded area in two dimensions can , as it is well - known @xcite , be expressed by the areas @xmath17 and the perimeters @xmath72 of the single grains @xmath73 also for rectangles with arbitrary orientations , there is an explicit formula for the mixed intrinsic volume in ( [ eq : a_ex ] ) @xcite .
the excluded area of two rectangles @xmath74 with orientations @xmath75 and @xmath76 and side lengths @xmath60 and @xmath61 is given by @xmath77 this equation allows for an explicit expression approximating the percolation thresholds of boolean models with rectangles . the product of the critical intensity @xmath6 and the average excluded area @xmath66 is called the total excluded area .
it is equivalent to the average number of bonds per object at the critical intensity @xmath78 , because the average number of bonds of a grain is equal to the average number of intersecting grains and thus equal to the average number of particles within the excluded area @xcite @xmath79 which was confirmed numerically ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
balberg et al .
@xcite proposed that @xmath78 is , at least approximately , an invariant quantity for similar systems , following a similar argument from lattice percolation @xcite .
actually , the average number of bonds per object @xmath78 is not a universal constant but depends on the grain shape and orientation distribution .
however , it varies to reasonable approximation only slightly for similar systems ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . if the average number of bonds per object at the critical intensity @xmath78 is empirically known ( e.g. , for a boolean model of squares ) and assumed to be the same for similar systems ( e.g. , boolean models with rectangles following the same orientation distribution ) the critical intensity of these similar models can be approximated by the explicit expresssion @xmath80 in numerous simulation studies , the assumption of a system invariant total excluded area is shown to be a useful approximation @xcite for which the critical intensity scales correctly with the particle size @xcite and which approximates the functional dependence of the percolation threshold on the anisotropy of the model @xcite .
the excluded area for overlapping rectangles with anisotropic orientation distribution is also calculated in @xcite , where the orientations are random within the interval @xmath81 $ ] ( following @xcite ) .
the approximation is compared to monte carlo simulations of isotropic models . for increasing anisotropy , an increase of the percolation threshold
is predicted , which we have confirmed by our simulations above .
here we compare the excluded area approximation to the monte carlo estimates of the anisotropic parametric boolean model studying the variety of models with different degrees of anisotropy from section [ sec : empirics ] . at the critical intensity ,
the average number of bonds @xmath78 for boolean models can be derived from the percolation threshold @xmath7 using ( [ eq : phi - gamma ] ) and ( [ eq_perco_bc_def ] ) : @xmath82 where @xmath17 is the area of a single grain . using the numerical results for the percolation thresholds @xmath7 from section [ sec : empirics ] , we estimate @xmath78 for overlapping squares for the different orientation distributions . using ( [ eq_perco_aex_approx ] ) , we finally approximate the critical intensities of the overlapping rectangles with varying aspect - ratio . figure [ fig : excluded_area_approximation ] compares this excluded area approximation to the simulation results from figure [ fig : percolation_thresholds_isotropic ] . for aligned rectangles ,
the independence of the percolation thresholds from the aspect ratio is correctly reproduced , because the excluded area of aligned rectangles @xmath83 is proportional to the area of a single grain ( which is here chosen to be unity ) . for the other anisotropic or isotropic boolean models
, the excluded area approximation seems to be a close upper bound for the rectangles if the empirical parameters of a square following the same orientation distribution are used . however , there are significant deviations from the prediction , for example , of the percolation threshold of randomly orientated rectangles with aspect ratio @xmath84 . as a function of the aspect ratio @xmath42 of the rectangles for boolean models with different degrees of anisotropy ,
see ( [ eq_perco_aex_approx ] ) .
a constant average number of bonds per object @xmath78 is assumed for the same orientation distribution , and the average excluded areas are calculated by ( [ eq : avaex ] ) and ( [ eq_perco_aex_rectangles ] ) .
the estimates of @xmath78 are for each curve derived from the percolation thresholds of squares with the corresponding anisotropy parameter @xmath11 , see figure [ fig : percolation_thresholds_isotropic ] and ( [ eq_perco_bc_of_phic ] ) .
in other words , the critical intensities @xmath6 of the squares are used as empirical parameters , so that the functional dependence of @xmath6 on the aspect ratio can be approximated . at the left- and right - hand side samples of the boolean models
are depicted . ]
in contrast to the excluded area approximation that requires the input of an empirical parameter ( @xmath78 ) , mecke and wagner @xcite suggested a purely geometrical approximation , namely , the zero of the mean euler characteristic as a function of the intensity .
they were motivated by the idea that a vanishing euler characteristic implies a balance between clusters and holes .
moreover , for self - dual matching lattices , like the square - lattice , the zero of the euler characteristic and the critical point coincide exactly @xcite .
they compared the critical intensity @xmath6 to the zero of the euler characteristic @xmath85 , for example , for overlapping discs and spheres , and suggested that @xmath85 could be a close bound for @xmath6 ; see also @xcite .
mecke and seyfried @xcite found in monte carlo simulations of a large class of discretized boolean models a good qualitative agreement , in that the dependence on the shape of the constituents is captured well , but they found also quantitative deviations .
neher et al .
@xcite analyzed two - dimensional lattice graphs , for which the euler characteristic provides a good approximation of the critical point .
the criterion has , for example , been applied to study the percolation in level sets of gaussian random fields @xcite , biopolymeric networks @xcite , flow in heterogeneous soil @xcite , or surface roughness of medical implants @xcite .
moreover , the connection between percolation and the euler characteristic triggered predictions of flow in porous media that are based on the minkowski functionals @xcite .
more precisely , we here refer to the density of the euler characteristic @xmath86 , that is , before taking the limit of an infinitely large observation window the functional value is rescaled by the size of the window @xcite .
explicit formulas are known for the density of the euler characteristic of boolean models as a function of the intensity @xcite .
the criterion is thus easily applicable to a great variety of continuum percolation models .
( marks ) are approximated by the zero of the euler characteristic ( lines ) for boolean models with rectangles .
the single plots show the percolation thresholds as a function of the aspect ratio for orientation distributions with different degrees of anisotropy , from randomly oriented ( far left ) to aligned ( far right ) rectangles .
the zero of the euler characteristic is a lower bound on the percolation thresholds .
although there is a significant offset , the qualitative behavior , i.e. , the dependence on anisotropy and grain shape , is described quite accurately . ] here we calculate this approximation for the boolean models with varying anisotropy that are studied above , and we compare it to the simulation results for the percolation thresholds from section [ sec : empirics ] .
we investigate how well the dependence of the threshold on the grain shape and orientation distribution is captured by the zero of the euler characteristic .
the density @xmath86 of the euler characteristic of a boolean model with rectangles as a function of the intensity @xmath5 is given by @xmath87 see @xcite , where @xmath88 is the average of the mixed intrinsic volume @xmath89 of two rectangles , @xmath90 see ( [ eq : a_ex ] ) and ( [ eq : avaex ] ) .
the zero of the euler characteristic then follows in a straightforward way : @xmath91 which appears similar to the excluded area approximation in ( [ eq_perco_aex_approx ] ) .
the occupied area fraction @xmath92 at which the mean euler characteristic changes its sign is according to ( [ eq : phi - gamma ] ) then given by @xmath93 where @xmath17 is the area of a single grain .
( marks ) for different aspect ratios @xmath42 is well described by the change in the zero of the euler characteristic @xmath85 ( lines ) ; not only in the trivial case of aligned rectangles ( red , see also samples depicted at the top of the right- and left - hand side ) , but a remarkably good approximation is also provided for randomly oriented rectangles with aspect ratios @xmath94 ( blue , see also samples depicted at the bottom of the right- and left - hand side ) . for details , see figure [ fig : percolation_thresholds_isotropic ] . ]
figure [ fig_perco_euler_zero ] compares @xmath92 from ( [ eq_perco_phi0 ] ) to the estimates of the isotropic percolation thresholds @xmath7 from figure [ fig : percolation_thresholds_isotropic ] . at least for these boolean models ,
the zero of the euler characteristic is a lower bound on @xmath7 .
there is a significant quantitative deviation , but similar to the findings for discretized boolean models @xcite , we find that the qualitative behavior of @xmath7 , i.e. , the dependence on anisotropy and grain shape , is well described by @xmath92 . to further investigate this accurate qualitative description , we plot in figure [ fig : offset_eulerzero ] the difference of the critical intensity of rectangles and the critical intensity of squares with the same orientation distribution .
we compare these differences to the differences of the corresponding zeros of the euler characteristic . for the boolean models studied here
, we find a remarkable agreement . only for an intermediate regime of strongly preferred orientation , but no full alignment ( @xmath95 ) , we find distinct ( but still small ) deviations . however , there is an excellent agreement of the difference in the critical intensities to the difference of the zeros of the euler characteristic not only for the special case of full alignment but also for randomly oriented rectangles .
if the critical intensity of squares is used as an empirical parameter , similar to the excluded area approximation , the zero of the euler characteristic allows at least for these boolean models a quite precise prediction of the percolation threshold , compare figure [ fig : excluded_area_approximation ] . like the critical average number of bonds @xmath78 , the offset between the critical intensity @xmath6 and the zero @xmath85 of the euler characteristic depends on the shape of the grains ; it is different for randomly oriented squares , 0.19688(1 ) @xcite , and discs , 0.128085(2 ) @xcite .
note , however , that assuming a constant offset to the number of grains per area of a single grain would lead to a nonphysical behavior in the limit of vanishing aspect ratios . for grains with unit area , both the critical intensity and the zero of the euler characteristic trivially vanish in the limit of extreme anisotropy , i.e. , for sticks , because of the choice of units .
for example , for randomly oriented sticks of length unity , the zero of the euler characteristic ( @xmath96 ) is as expected a lower bound on the critical intensity ( @xmath97 @xcite ) .
the euler characteristic is one of the so - called minkowski functionals ( or intrinsic volumes ) from integral geometry @xcite . in two - dimensions ,
the minkowski functionals are the area @xmath98 , the perimeter @xmath99 , and a functional proportional to the euler characteristic @xmath100 .
these three minkowski functionals characterize , according to hadwiger s theorem , the complete `` additive shape information '' of a compact and convex set @xmath56 in the following sense : they form a basis of all functionals @xmath101 that are motion invariant , continuous , and additive @xcite .
the latter means that @xmath102 for two disjoint sets @xmath56 and @xmath33 . the second moments of minkowski functionals and generalizations thereof are known analytically @xcite .
remarkably , a behavior similar to thermodynamical quantities in statistical physics has been found for the second moments of minkowski functionals @xcite . as an outlook
, we here examine the prospects of further approximations of the percolation threshold based on the second moments of the minkowski functionals without using empirical parameters .
we compare the extremal points of the variances and covariances of the minkowski functionals to the empirical estimates of the critical intensity .
an interesting example , which can serve as a motivation of this approach , is cell percolation in planar poisson voronoi tessellations .
voronoi cells of a poisson point process are colored black with some probability @xmath103 , and a percolating cluster is an unbounded and connected collection of black cells
. the critical probability of this model of continuum percolation is @xmath104 @xcite .
last and ochsenreither @xcite have proven that this coincides with the global maximum of the asymptotic variance of the euler characteristic
. moreover , hug et al .
@xcite found for overlapping discs that both the local minimum of the variance of the perimeter and the local minimum of the covariance of area and euler characteristic are good approximations of the percolation threshold .
for example , the local minimum of the asymptotic covariance @xmath105 is found at @xmath106 compared to the critical intensity @xmath107 @xcite , where the unit of area is again defined by the grain . of the minkowski functionals as functions of the intensity @xmath5 of the boolean models with squares ( left ) or rectangles with aspect ration @xmath108 ( right ) , the rectangles are either fully aligned ( top ) or follow an isotropic orientation distribution ( bottom ) .
the marks represent numerical estimates from simulations , the solid lines represent the analytic curves , and the dashed lines are guides to the eye .
the critical intensity @xmath6 is indicated by the black vertical line , the estimate for the aligned squares is taken from @xcite and for the isotropic squares and rectangles from @xcite or @xcite , respectively.,title="fig:",scaledwidth=48.0% ] of the minkowski functionals as functions of the intensity @xmath5 of the boolean models with squares ( left ) or rectangles with aspect ration @xmath108 ( right ) , the rectangles are either fully aligned ( top ) or follow an isotropic orientation distribution ( bottom ) . the marks represent numerical estimates from simulations , the solid lines represent the analytic curves , and the dashed lines are guides to the eye .
the critical intensity @xmath6 is indicated by the black vertical line , the estimate for the aligned squares is taken from @xcite and for the isotropic squares and rectangles from @xcite or @xcite , respectively.,title="fig:",scaledwidth=48.0% ] + of the minkowski functionals as functions of the intensity @xmath5 of the boolean models with squares ( left ) or rectangles with aspect ration @xmath108 ( right ) , the rectangles are either fully aligned ( top ) or follow an isotropic orientation distribution ( bottom ) .
the marks represent numerical estimates from simulations , the solid lines represent the analytic curves , and the dashed lines are guides to the eye .
the critical intensity @xmath6 is indicated by the black vertical line , the estimate for the aligned squares is taken from @xcite and for the isotropic squares and rectangles from @xcite or @xcite , respectively.,title="fig:",scaledwidth=48.0% ] of the minkowski functionals as functions of the intensity @xmath5 of the boolean models with squares ( left ) or rectangles with aspect ration @xmath108 ( right ) , the rectangles are either fully aligned ( top ) or follow an isotropic orientation distribution ( bottom ) .
the marks represent numerical estimates from simulations , the solid lines represent the analytic curves , and the dashed lines are guides to the eye .
the critical intensity @xmath6 is indicated by the black vertical line , the estimate for the aligned squares is taken from @xcite and for the isotropic squares and rectangles from @xcite or @xcite , respectively.,title="fig:",scaledwidth=48.0% ] of the minkowski functionals as functions of the intensity @xmath5 of the boolean models with squares ( left ) or rectangles with aspect ration @xmath108 ( right ) , the rectangles are either fully aligned ( top ) or follow an isotropic orientation distribution ( bottom ) . for details , see figure [ fig_perco_var].,title="fig:",scaledwidth=48.0% ] of the minkowski functionals as functions of the intensity @xmath5 of the boolean models with squares ( left ) or rectangles with aspect ration @xmath108 ( right ) , the rectangles are either fully aligned ( top ) or follow an isotropic orientation distribution ( bottom ) . for details , see figure [ fig_perco_var].,title="fig:",scaledwidth=48.0% ] + of the minkowski functionals as functions of the intensity @xmath5 of the boolean models with squares ( left ) or rectangles with aspect ration @xmath108 ( right ) , the rectangles are either fully aligned ( top ) or follow an isotropic orientation distribution ( bottom ) . for details , see figure [ fig_perco_var].,title="fig:",scaledwidth=48.0% ] of the minkowski functionals as functions of the intensity @xmath5 of the boolean models with squares ( left ) or rectangles with aspect ration @xmath108 ( right ) , the rectangles are either fully aligned ( top ) or follow an isotropic orientation distribution ( bottom ) . for details ,
see figure [ fig_perco_var].,title="fig:",scaledwidth=48.0% ] in our outlook here , we compare for four different boolean models with rectangles the extremal points of the variance - covariance structure of the minkowski functionals to the percolation threshold .
we investigate whether the second moments could allow for accurate predictions of the thresholds without the need for empirical parameters , especially testing whether the local minima of @xmath105 or @xmath109 are close to the percolation threshold for anisotropic models .
figures [ fig_perco_var ] and [ fig_perco_cov ] show the different asymptotic variances or covariances for a boolean models with either squares or rectangles ( with aspect ratio @xmath108 ) .
these are either fully aligned or follow an isotropic orientation distribution .
for these four boolean models very accurate estimates of the percolation thresholds can be found in literature @xcite .
we use these precise estimates with at least four significant number of digits because we indeed find some remarkable agreement between the extremal points and the percolation threshold . like for the overlapping discs , a local minimum point of the variance of the perimeter
is an excellent approximation of the percolation threshold for aligned squares or rectangles with aspect ratio @xmath108 .
the critical intensity of aligned squares is @xmath110 @xcite , the local minimum point of the asymptotic variance of the perimeter is @xmath111 \approx 1.0994 $ ] , which is derived by numerically minimizing @xmath112 using @xcite .
while this is a remarkably good approximation , there is a statistically significant difference .
moreover , while the critical intensity must not depend on the aspect ratio of the aligned rectangles , see section [ sec_iso_perco ] , the local minimum point of the asymptotic variance of the perimeter slightly depends on the aspect ratio . for an aspect ratio of @xmath108 ,
the local minimum point is at @xmath113 \approx 1.0952 $ ] , and for an aspect ratio of @xmath114 , it is @xmath115 \approx 1.0727 $ ] .
the local minimum point of the variance is also neither an upper nor a lower bound for the percolation threshold . for aligned rectangles
, the minimum points can be below the critical intensity .
however , the minimum points are above the critical intensities for aligned squares or for the randomly oriented squares and rectangles , see figure [ fig_perco_var ] .
another candidate for an approximation of the percolation threshold is the local minimum point of the covariance of area and euler characteristic .
figure [ fig_perco_cov ] shows that it is a very good approximation for aligned squares and rectangles with aspect ratio @xmath108 and a close upper bound if the grains are randomly oriented . only for one of the boolean models , a vanishing correlation can serve as a good approximation of the percolation threshold : for randomly oriented squares the first zero of the covariance of perimeter and euler characteristic ( @xmath116 ) seems to be in good agreement with the critical intensity , see figure [ fig_perco_cov ] .
however , for randomly oriented rectangles there is a distinct difference .
moreover , for aligned squares or rectangles , as well as for overlapping discs , this zero crossing from negative to positive correlation does not exist .
we have investigated percolation in anisotropic systems , both for finite simulation boxes and in the limit of infinite system size .
we have especially elucidated how the anisotropy of the boolean model , which is a geometric property , influences percolation , which is a topological property .
there is both universal and non - universal behavior .
the isotropy of the percolation threshold is of the first kind . in infinite systems , there is no difference in effective percolation thresholds for different directions .
independent of the details of the model , the difference between the effective percolation thresholds observed in finite systems vanishes in the thermodynamic limit .
this has here been related to the uniqueness of the percolating cluster , see figure [ fig : two_percolating_clusters ] . in a finite sample of an anisotropic boolean model ,
a cluster is more likely to span the system in the preferred than in the perpendicular direction , see figure [ fig_perco_sample ] .
this anisotropic percolation in finite samples is demonstrated by determining the effective percolation thresholds , see ( [ eq_perco_def_phieff ] ) and ( [ eq : erffit ] ) , by a fit of an error function to the connectivity , see figure [ fig : connectivity ] .
although a distinct difference remains even for very large system sizes , the extrapolation to infinite system sizes shows that even the most anisotropic model percolates simultaneously in all directions , see figure [ fig_perco_scaling ] .
this finding we here refer to as the isotropy of the percolation threshold .
our results are in agreement with previous findings @xcite . for boolean models with different degrees of anisotropy
, we have estimated the isotropic percolation thresholds by a simultaneous fit of the finite size scaling in @xmath0- and @xmath1-direction , see ( [ eq : scaling_xy_equiv])([eq : percolation_fit ] ) . while the isotropy of the percolation threshold is universal , its value is non - universal , that is , it depends on the grain distribution ( or the degree of orientation bias ) . in agreement with previous findings
@xcite , the more anisotropic the orientation distribution , the larger is the percolation threshold , see figure [ fig : percolation_thresholds_isotropic ] .
there are efficient algorithms that use inhomogeneous boolean models for high - precision estimates of percolation thresholds in homogeneous and isotropic models , e.g. , see @xcite .
the isotropy of the percolation threshold implies that these algorithms can be directly generalized to anisotropic systems .
different directions of the gradient in the intensity should yield ( within statistical errors ) the same estimate of the percolation threshold .
such methods could strongly increase the precision of estimates of the percolation thresholds in anisotropic models . in two dimensions ,
the percolation of grain or void phase are complementary phenomena , that is , there is ( almost surely ) a percolating cluster of the void phase at occupied area fractions below the threshold , but it vanishes ( almost surely ) above the threshold .
we can therefore directly apply the numerical results and the approximations also to void percolation and the critical porosity .
the minkowski functionals help to study the relation between geometry , e.g. , quantified by area or perimeter , and topology , e.g. , quantified by the euler characteristic .
they thus allow for insights into percolation as a topological phase transition . their generalization , the mixed volumes ,
provide an explicit formula for the well - known excluded area approximation of percolation thresholds , see ( [ eq : a_ex])([eq_perco_aex_approx ] ) . if the percolation thresholds of squares with different orientation distributions are used as empirical parameters , we find that the excluded area approximation is in relatively good agreement with the percolation thresholds of the rectangles , see fig . [
fig : excluded_area_approximation ] . for the systems studied here
, it provides a close upper bound , which implies that the critical average number of bonds per grain slightly decreases for the rectangles compared to squares with the same orientation distribution .
in contrast to the excluded area approximation , the zero of the euler characteristic serves as a purely geometric approximation of the percolation threshold . we show that it provides a lower bound that captures well the qualitative dependence on the system parameters , see figure [ fig_perco_euler_zero ] . if an empirical parameter is introduced ( similar to the excluded area approximation )
, it allows for quite precise predictions of @xmath6 within a narrow range of parameters studied here , see figure [ fig : offset_eulerzero ] . using the second moments of the minkowski functionals , we suggest and compare new candidates for accurate approximations of the percolation thresholds , see figs .
[ fig_perco_var ] and [ fig_perco_cov ] .
some are found to be in surprisingly good agreement for the four systems that are tested here , for example , the minimum point of the covariance of area and euler characteristic .
more examples and further research are needed to test this suggestion .
an important prospect for applications is that no empirical parameters would be needed because analytic formulas are known for the second moments of the minkowski functionals @xcite .
similar techniques could also easily be applied to percolation on lattices for which the covariances of the minkowski functionals can be calculated straightforwardly @xcite . moreover
, our results can also lead to more fundamental questions .
why is there such a good agreement between the percolation threshold and some of the extremum points of the second moments of the minkowski functionals as a function of the intensity ? is it a coincidence or is there a connection between local fluctuations and the global topology ?
we thank the german research foundation ( dfg ) for the grants no .
hu1874/3 - 2 , no .
la965/6 - 2 , and no .
me1361/11 awarded as part of the dfg - forschergruppe for 1548 `` geometry and physics of spatial random systems . ''
based on the finite size scaling in ( [ eq : scaling_xy_equiv ] ) , we determine the isotropic percolation threshold @xmath7 via a simultaneous extrapolation of the effective percolation thresholds in @xmath0- and in @xmath1-direction .
more precisely , we estimate the percolation threshold via a maximum likelihood fit of the parameters @xmath7 , @xmath52 , and @xmath53 . the effective percolation thresholds in @xmath0- and in @xmath1-directions could be correlated , but the correlation can not be considered in this analysis .
we neglect the correlation of @xmath117 and @xmath118 in the simulation and use the likelihood function @xmath119\cdot\mathrm{prob}\left[\phi_{c , i}^{\mathrm{eff},y}|\phi_c , m^y\right]\end{aligned}\ ] ] with @xmath120=\frac{1}{\sqrt{2\pi}\sigma_{z , i}}\exp\left[{-\frac{\left(\phi_{c , i}^{\mathrm{eff},z}-\phi_c - m^z\cdot l_i^{-1/\nu}\right)^2}{2\sigma_{z , i}^2}}\right]\;,\end{aligned}\ ] ] where for each system size @xmath121 the error of the effective percolation thresholds @xmath117 and @xmath118 are given by @xmath122 and @xmath123 , respectively .
the log - likelihood function can be written as @xmath124 with @xmath125 and @xmath126 where the covariance matrix @xmath127 is assumed to be diagonal @xmath128 therefore , the maximum likelihood estimation equals a minimum least square fit with block matrices .
schrder - turk g e , mickel w , kapfer s c , klatt m a , schaller f m , hoffmann m j f , kleppmann n , armstrong p , inayat a , hug d , reichelsdorfer m , peukert w , schwieger w and mecke k 2011 _ adv . mater . _ * 23 * 2535 klatt m a 2016 _ morphometry of random spatial structures in physics _ ph.d .
thesis friedrich - alexander - universitt erlangen - nrnberg ( fau ) . parts of this article are from this phd thesis of one of the authors .
stoyan d and mecke k 2005 the boolean model : from matheron till today _ space , structure and randomness _
( new york : springer - verlag ) michel bilodeau , fernand meyer and michel schmitt ( eds . ) , pp .
151182 mecke k r 2000 additivity , convexity , and beyond : applications of minkowski functionals in statistical physics _ statistical physics and spatial statistics _
( _ lecture notes in physics _
vol 554 ) ed mecke k r and stoyan d ( berlin , heidelberg : springer ) pp 111184 isbn 978 - 3 - 540 - 67750 - 5 rodrguez - valverde m a , ramn - torregrosa p j and cabrerizo - vlchez m a 2010 estimation of percolation threshold of acid - etched titanium surfaces using minkowski functionals ( _ microscopy book series _
vol 3 ) ed a mndez - vilas j d ( formatex ( badajoz , spain ) ) p 1978 microscopy : science , technology , applications and education | we examine the interplay between anisotropy and percolation , i.e. , the spontaneous formation of a system spanning cluster in an anisotropic model .
we simulate an extension of a benchmark model of continuum percolation , the boolean model , which is formed by overlapping grains . here
we introduce an orientation bias of the grains that controls the degree of anisotropy of the generated patterns .
we analyze in the euclidean plane the percolation thresholds above which percolating clusters in @xmath0- and in @xmath1-direction emerge . only in finite systems ,
distinct differences between effective percolation thresholds for different directions appear .
if extrapolated to infinite system sizes , these differences vanish independent of the details of the model . in the infinite system ,
the uniqueness of the percolating cluster guarantees a unique percolation threshold . while percolation is isotropic even for anisotropic processes , the value of the percolation threshold depends on the model parameters , which we explore by simulating a score of models with varying degree of anisotropy .
to which precision can we predict the percolation threshold without simulations ?
we discuss analytic formulas for approximations ( based on the excluded area or the euler characteristic ) and compare them to our simulation results .
empirical parameters from similar systems allow for accurate predictions of the percolation thresholds ( with deviations of @xmath2 in our examples ) , but even without any empirical parameters , the explicit approximations from integral geometry provide , at least for the systems studied here , lower bounds that capture well the qualitative dependence of the percolation threshold on the system parameters ( with deviations of @xmath3@xmath4 ) .
as an outlook , we suggest further candidates for explicit and geometric approximations based on second moments of the so - called minkowski functionals .
= 4 _ keywords _ : anisotropy , continuum percolation , percolation threshold , uniqueness of percolating cluster , excluded area , euler characteristic percolation is a geometric phase transition in which formerly local clusters spontaneous form a system spanning cluster . dating back to the work by flory @xcite and stockmayer @xcite and later by broadbent and hammersley @xcite
, percolation theory describes in general the critical behavior of connected clusters , either on lattices or in the continuum @xcite .
it is a general concept that can be applied to a multitude of very different physical systems and complex networks @xcite . in this paper , we are interested in the interplay of topology and geometry , more precisely , of percolation and anisotropy .
we investigate continuum percolation , that is , percolation on a continuous space , in statistically homogeneous ( also called stationary ) two - phase random media .
more specifically , we study a benchmark model of continuum percolation @xcite , the so - called boolean model .
it consists of overlapping grains that are distributed randomly in space .
the intensity @xmath5 , which is also called number density , gives the expectation for the number of grain centers per unit area .
it is one of the parameters of the system .
the overlapping grains form clusters of connected components . starting from a dilute arrangement of grains ,
the mean size of these clusters grows with increasing area fraction of the grain phase , called the occupied area fraction . below the critical area fraction , the so - called percolation threshold , these clusters are almost surely ( i.e. , with probability one ) bounded . at occupied volume fractions above the percolation threshold ,
almost surely exactly one unbounded cluster emerges that spans the whole system .
it is called the percolating cluster .
percolation is a geometric phase transition with universal scaling of geometrical properties , for example , the size of the percolating cluster .
anisotropy appears in many man - made or natural materials , from geological formations @xcite over fiber composites and laminates @xcite to porous media @xcite .
moreover , anisotropy manifests itself in different forms
. a heterogeneous medium might be isotropic w.r.t .
the distribution of the area but anisotropic w.r.t . of the orientation of the interface , or vice
versa @xcite .
anisotropic two - phase random media can be modeled by the boolean model with an orientation bias of the grains @xcite . in these systems
, percolation exhibits an interesting qualitatively different behavior in finite samples or in the idealized infinite limit .
although there is no phase transition in a finite system , an effective percolation threshold can be defined for the ensemble . as expected , it is anisotropic . in other words ,
the system is more likely to percolate in the preferred direction than in the perpendicular direction .
however , this difference vanishes in the thermodynamic limit .
even the most anisotropic system simultaneously percolates in all directions .
the percolation threshold is isotropic , see also @xcite , which we here link to the uniqueness of the percolating cluster .
there have already been numerous studies of anisotropic continuum percolation , some using analytic calculations ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) others are based on monte carlo simulations ( e.g. , * ? ? ?
* ; * ? ? ?
* ) or experiments ( e.g. , * ? ? ?
* ) .
there are also detailed studies of anisotropic lattice percolation @xcite .
while the isotropy of the percolation threshold is independent of the details of the model , the value of the percolation threshold is a nonuniversal constant , that is , it depends on the details of the system and specifically on its anisotropy . for boolean models
an increasing degree of anisotropy typically also increases the percolation threshold of the grain phase @xcite . because of the complexity of the geometric problem , analytic results for the percolation thresholds of boolean models are hard to derive .
precise numerical estimates require complex and time - consuming calculations .
explicit formulas for estimates of percolation thresholds help to adjust models to experimental systems and predict their physical properties .
we determine , from simulations , the percolation thresholds for a score of models with varying degree of anisotropy .
how well can we predict the dependence of the critical intensity on the model parameters ?
first , we compare the simulation results to the well - known excluded - area approximation , which uses an empirical parameter @xcite .
then , we examine a purely geometric lower bound without any empirical parameter , which is based on a topology index from integral geometry , the euler characteristic @xcite .
the simulation results suggest that the lower bound qualitatively captures well how the critical intensity depends on the model parameters .
we also suggest new candidates of purely geometric approximations based on further integral geometric shape descriptors . in section [ sec : empirics ]
, we simulate anisotropic boolean models ( see section [ sec_anisotropic_boolean_models ] ) and examine the behavior of continuum percolation both in finite samples and extrapolated to infinite systems . for finite systems , we define the effective percolation threshold in section [ sec : percolation_connectivity ] and extrapolate it to infinite system sizes in section [ sec_iso_perco ] .
there we also provide several heuristic explanations of the empirical finding of an isotropic percolation threshold . in section [ sec_perco_thresh_estimate ] ,
we examine how the value of the percolation threshold changes with the anisotropy of the system . in section [ sec : approx ] , we study analytic approximations of the percolation thresholds based on the excluded area ( section [ sec : excluded - area - approx ] ) , the euler characteristic ( section [ sec : euler ] ) , or the second moments of minkowski functionals ( section [ sec_percolation_cov ] ) . in section [ sec_conclusion_and_outlook ] , besides concluding remarks , we given an outlook to open problems and generalizations as well as to how our results suggest techniques for future research @xcite . |
there has been tremendous interest over the last several years in the brane - world scenario where we are assumed to be living on a four - dimensional hypersurface ( brane ) in a higher - dimensional spacetime ( bulk ) @xcite .
the simplest example of this scenario is proposed by randall and sundrum where there is a brane embedded in a five - dimensional anti - de sitter ( ads ) spacetime @xcite .
they showed that the four - dimensional gravity is recovered at low energies , because gravity is confined to a single positive - tension brane even if the extra dimension is not compact .
therefore , in order to see the deviation from the conventional four - dimensional gravity , it is necessary to investigate the phenomena occurred at high energies .
one of the most promising tools to extract the information of the extra - dimension is the primordial density fluctuations generated in the period of inflation in the early universe . in the inflation model where inflation is driven by an inflaton field confined to the brane
, the amplitude of the curvature perturbation is calculated in the extremely slow - roll limit where the coupling between inflaton field fluctuation and bulk perturbations can be neglected @xcite .
this work is proceeded to go beyond the zero - th order slow - roll approximation by solving the bulk metric perturbations classically @xcite .
it is shown that we could have significant effects from the back - reaction due to the coupling to five - dimensional gravitational perturbations .
on the other hand , it was pointed out that the localized matter on a brane can induce gravity on the brane via quantum loop corrections at high energies @xcite .
this induced gravity can act as the ultra - violet cut - off for the inflaton perturbations .
based on the induced gravity scenario , dvali , gabadadze , and porrati ( dgp ) proposed a brane - world model with induced gravity @xcite in which the four - dimensional brane is embedded in a five - dimensional minkowski spacetime .
( for a review of the phenomenology of dgp model , see @xcite . ) in the presence of the induced gravity , according to the embedding of the brane in the bulk , there appear two branches of background solutions . in @xmath4 branch ,
the size of the four - dimensional hypersurface becomes minimum at the brane while , in @xmath7 branch , it becomes maximum at the brane .
inflation models in this scenario were studied in @xcite . in this paper
, we study the behavior of five - dimensional metric perturbations excited by the inflaton perturbations confined to the brane in dgp model .
the solutions in @xmath7 branch are regarded as the high energy limit of the randall - sundrum model including the curvature corrections on a brane , depending on the parameters .
then our analysis can be used to discuss the generation of primordial fluctuations in these models . the solutions in @xmath4 branch
are largely used to explain the present cosmic acceleration without introducing dark energy in the dgp model @xcite , but it turns out that they are plagued by the ghost instability @xcite . however , the study of the gravitational property of this model in the presence of matter and cosmic expansion is still very limited @xcite so it is also worth considering the solutions in this branch .
the structure of the rest of the paper is as follows . in section
ii , we derive the basic equations for the five - dimensional metric perturbation and the boundary condition imposed at the brane and then summarize the solution for the vacuum brane obtained in @xcite . in section iii , we provide the solutions for curvature perturbations on the brane in the presence of the scalar field on the brane .
next we discuss the appearance and disappearance of van dam - veltman - zakharov discontinuity in section iv . in section
v we show the comparison of the behavior of perturbations on small scales with the brans - dicke theory .
the solutions for perturbations on large scales are discussed in section vi .
section vii is devoted to conclusions .
technical details of the calculation of the @xmath6 mode and the brief summary of the cosmological perturbation in four - dimensional brans - dicke gravity are presented in the appendix a and b , respectively .
we consider a four - dimensional brane - world model with a five - dimensional infinite - volume bulk .
the action is given by @xmath10 , \label{bulkbrane_action}\end{aligned}\ ] ] where @xmath11 is the fundamental five - dimensional gravitational constant , @xmath12 is the trace of extrinsic curvature @xmath13 on the brane and @xmath14 is a lagrangian for brane - localized matter .
in addition , we consider the last term , an intrinsic curvature term on the brane which plays a crucial role in this model . here , @xmath15 is the four - dimensional gravitational constant .
we define a crossover scale @xmath16 as @xmath17 we are interested in the gravitational property of this model in a cosmological setting . by assuming the flat friedmann - robertson - walker metric on the brane and neglecting the effect from the five - dimensional weyl tensor in the background ,
the following friedmann like equation is derived @xcite ; @xmath18 where @xmath19 is a hubble parameter on the brane and @xmath20 , which is the parameter related to the embedding of the brane in the bulk .
we call the case with @xmath21 @xmath4 branch , while @xmath22 @xmath7 branch . in this paper
, we restrict ourselves to de sitter background @xmath23 constant . in this case , the energy density is constant and given by @xmath24 the 5d solution for the metric with the four - dimensional de sitter brane can be obtained as @xmath25 where @xmath26 is the metric for the de sitter spacetime and the brane is located at @xmath27 . during a slow - roll inflation ,
the hubble scale is varying slowly , thus we can approximate the background geometry by ( [ 5dmetric ] ) .
now let us consider the scalar perturbation .
in order to solve the perturbations in this background , it is convenient to use five - dimensional longitudinal gauge , given by @xmath28,\end{aligned}\ ] ] where @xmath29 . in the absence of bulk matter perturbations , five - dimensional perturbed einstein equations
@xmath30 are solved in a minkowski background if the metric perturbations are derived from a ` master variable ' , @xmath31 @xcite ( see also @xcite ) ; @xmath32 where dot and prime denote derivatives with respect to @xmath33 and @xmath34 , respectively .
the perturbed five - dimensional einstein equations yield a single wave equation governing the evolution of the master variable @xmath31 in the bulk : @xmath35 solutions for the master equation can be separated into eigenmodes of the time - dependent equation on the brane and bulk mode equation : @xmath36 where @xmath37g_m = 0,\\ f_m '' & - & 2\frac{n'}{n } f'_m + \frac{m^2}{n^2 } f_m = 0.\end{aligned}\ ] ] in order to solve @xmath31 , we must specify the boundary condition for @xmath31 from the junction conditions at the brane . the junction conditions in this model have been found already in literatures ( see @xcite ) , but , we re - derive these so that the property of linearized gravity is transparent . for simplicity , we consider the case where the anisotropic stress of the matter on the brane does not exist .
there are two important contributions to the junction conditions .
one is that , from eq .
( [ def_tilde_t_munu ] ) , we must take into account not only the matter energy momentum , but also the contribution from the induced gravity on the brane .
the other is that we should take into account the brane bending @xmath38 @xcite .
the metric perturbations in five - dimensional longitudinal gauge evaluated at @xmath27 is _ not _ the induced metric perturbations on the brane .
this is due to the anisotropic stress coming from the induced gravity .
the junction conditions are given by @xmath39 - 2 \ddot{\xi } - 4h \dot{\xi } + 6h^2 \xi - \frac{4}{3}\frac{k^2}{a^2}\xi = \kappa^2(\delta p_{\rm m}+ \delta p_{\rm g } ) , \label{junc_1_ii}\\ & & -\frac{1}{a}\dot{{\cal{f } } } + 2 \dot{\xi } -2h \xi = \kappa^2 ( \delta q_{\rm m}+ \delta q_{\rm g } ) , \label{junc_1_0i}\\ & & -\frac{1}{a}\left[3h \dot{{\cal{f } } } + \frac{k^2}{a^2}{\cal{f } } \right ] + 6h \dot{\xi } -6h^2 \xi + 2 \frac{k^2}{a^2}\xi = \kappa^2 ( \delta \rho_{\rm m}+ \delta \rho_{\rm g } ) \label{junc_1_00},\end{aligned}\ ] ] where @xmath40 here the brane bending mode is given by @xmath41 where @xmath42 and @xmath43 are the induced metric perturbations on the brane that are related with the metric perturbations in five - dimensional longitudinal gauge as @xmath44 , \label{a_b_ito_ar}\\ { \cal{r}}_{(b ) } & = & \frac{1}{1 -2 \epsilon h r_c } \big [ ( 1- \epsilon h r_c ) { \cal{r } } + \epsilon h r_c a \big ] , \label{r_b - ito_ar}\end{aligned}\ ] ] where @xmath45 and @xmath46 are metric perturbations in five - dimensional longitudinal gauge evaluated at @xmath27 , which can be expressed by @xmath31 as @xmath47 , \label{r - ito_fomega}\\ a & = & \frac{1}{6 a } \big [ -3 \epsilon h { \cal f } -3 \ddot{\omega } + 6h \dot{\omega } -3 h^2 \omega -2 \frac{k^2}{a^2 } \omega \big ] .
\label{a - ito_fomega}\end{aligned}\ ] ] @xmath48 and @xmath49 are the components of the perturbed energy momentum tensor of the matter on the brane and @xmath50 , \label{def_rho_g}\\ & & \delta q_{\rm g } = \frac{2}{\kappa_4 ^ 2 } ( h a_{(b ) } - \dot{{\cal{r}}}_{(b ) } ) , \label{def_q_g}\\ & & \delta \rho_{\rm g } = \frac{2}{\kappa_4 ^ 2 } \left[3h(- \dot{{\cal{r}}}_{(b ) } + h a_{(b ) } ) - \frac{k^2}{a^2 } { \cal{r}}_{(b)}\right],\end{aligned}\ ] ] are contributions from the induced gravity . if we do not take into account the induce gravity term which corresponds to the limit @xmath51 , the quantities @xmath52 and @xmath53 can be neglected and the junction conditions in randall - sundrum type brane - world @xcite are reproduced .
defining @xmath54 as @xmath55 the junction conditions ( [ junc_1_ii]),([junc_1_0i]),([junc_1_00 ] ) are amazingly simplified as @xmath56 before discussing the case with the matter perturbations on the brane , we briefly summarize the case for a vacuum brane , @xmath57 considered in @xcite .
we will only consider a positive tension brane which means @xmath58 for @xmath4 branch and @xmath59 for @xmath7 branch .
the solutions obtained here will serve as homogeneous solutions when we include matter perturbations . for the vacuum brane , the boundary conditions for @xmath31 given by eqs .
( [ junc_2_gen_00 ] ) , ( [ junc_2_gen_0i ] ) and ( [ junc_2_gen_ii ] ) reduce to a single boundary condition on the master variable , @xmath60 for both branches , we find a tower of continuous massive modes starting from @xmath61 , which are the kaluza - klein ( kk ) modes of the spin-2 perturbations . in @xmath4 branch ( @xmath21 ) ,
in addition to continuous modes , there are two discrete modes which are normalizable .
one is the mode with @xmath5 which is the spin-0 perturbation , and the other is the mode with @xmath62 which is the helicity-0 excitation of the spin-2 perturbations @xcite . for @xmath63 ,
the mass @xmath64 is in the range @xmath65 .
if @xmath66 , the helicity-0 excitation of the spin-2 perturbation has mass @xmath67 , which coincides with the mass of the spin-0 perturbation . then , there is a resonance between these two modes .
it was also shown that the spin-2 helicity-0 mode becomes a ghost for @xmath58 . in figure 1 , we summarize the mass spectrum for @xmath68 in @xmath4 branch .
branch @xcite .
spin-2 perturbation has continuous modes with @xmath69 and a discrete mode with @xmath70 while spin-0 perturbation has @xmath6 . in the limit @xmath71 ,
both the helicity-0 excitation of spin-2 perturbation and the spin-0 perturbation have mass @xmath6 and there is a resonance @xcite .
, width=226 ]
until now , we have not specified the matter on the brane . in the following , we model the matter on the brane as a canonical scalar field @xmath72 with potential @xmath73 whose homogeneous part gives approximately de sitter universe on the brane @xcite .
in this case , the junction conditions ( [ junc_1_ii ] ) , ( [ junc_1_0i ] ) , ( [ junc_1_00 ] ) reduce to @xmath74 , \label{junc_2_00}\\ & & \dot{{\cal{g } } } = \kappa^2 a \dot{\phi } \delta \phi , \label{junc_2_0i}\\ & & \ddot{{\cal{g } } } + 2h \dot{{\cal{g } } } = \kappa^2 a [ \dot{\phi } \dot{\delta \phi } - { v,}_\phi \delta \phi ] .
\label{junc_2_ii}\end{aligned}\ ] ] these equations can be thought as the boundary conditions for @xmath31 .
combining the junction conditions ( [ junc_2_00 ] ) , ( [ junc_2_0i ] ) , ( [ junc_2_ii ] ) , we get an evolution equation for @xmath54 : @xmath75 this gives the boundary condition for the time dependence of the master variable @xmath31 . assuming that @xmath72 is slow - rolling , so that @xmath76 in eq . ( [ evol_eq ] ) , which is valid for the de sitter universe ,
the solution for @xmath54 is @xmath77 we use the formulae for summation of bessel functions , @xmath78 these show that an infinite sum of mode functions @xmath79 can satisfy the boundary condition imposed on @xmath54 , where the spectrum of kk modes is given by @xmath80 figure 2 shows the mass spectrum of the scalar perturbation supported by the scalar field on the brane .
since there is a mode with @xmath6 regardless of the value of @xmath81 in the @xmath4 branch , the resonance inevitably appears .
this gives qualitative difference between @xmath7 branch and @xmath4 branch .
in addition , for @xmath82 , that is , on the self - accelerating universe , the spin-0 and spin-2 modes are already degenerates .
thus the behavior of gravity is again qualitatively different from @xmath58 .
, there is a mode with @xmath6 regardless of the value of @xmath81 .
since there is another mode with @xmath6 by the spin-0 perturbation in the @xmath4 branch , there is a resonance .
in addition , for @xmath83 , the helicity-0 mode of spin-2 perturbations also has @xmath84 .
, width=453 ] in order to examine the behavior of the master variable @xmath31 , we will consider @xmath7 branch ( @xmath22 ) and @xmath4 branch ( @xmath21 ) separately in the following .
since there is a horizon at @xmath85 in this branch , we should choose the solution in the @xmath34-direction so that the metric perturbations remain small as @xmath86 . by imposing the normalizability condition on @xmath87 , @xmath88 we obtain the solution for @xmath31 in the bulk ( for @xmath89 ) as @xmath90 in @xmath4 branch
, the bulk is infinite .
we should choose the solution in the @xmath34-direction so that the metric perturbations remain small as @xmath91 and the mode is normalizable for a single brane .
we impose the normalizability condition on @xmath87 as @xmath92 because of the existence of the resonance , we must treat the mode with @xmath6 separately .
we also treat @xmath83 case separately as the spin-2 and spin-0 degenerate ( see appendix a for derivations ) .
we obtain the solution for @xmath31 in the bulk ( for @xmath93 0 ) as @xmath94 where @xmath95 for @xmath96 and @xmath97,\\ \label{til_beta_def_2 } \tilde{\beta}(-k\eta ) & = & \beta(-k\eta ) + \frac{1}{2 } \int^{-k\eta } d(-k\bar{\eta } ) ( -k\bar{\eta})^{-2 } [ \alpha(-k\bar{\eta } ) \{\cos(-2k\bar{\eta})-1\ } -\beta ( -k\bar{\eta } ) \sin(-2k\bar{\eta } ) ] , \end{aligned}\ ] ] for @xmath82 . since we have obtained the general solutions for @xmath31 , from eqs .
( [ a_b_ito_ar ] ) , ( [ r_b - ito_ar ] ) , ( [ r - ito_fomega ] ) and ( [ a - ito_fomega ] ) the induced metric perturbations on the brane can be obtained .
we concentrate on the behavior of the curvature perturbation on the brane @xmath43 . as in the previous section
, we will consider @xmath7 branch ( @xmath22 ) and @xmath4 branch ( @xmath21 ) separately . in @xmath7 branch ( @xmath22 )
, the curvature perturbation on the brane is given as @xmath98\nonumber\\ & & -\frac{\sqrt{2 \pi}}{6 } \frac{c_2}{k } \frac{h^2}{(2 h r_c + 1 ) } \sum_{\ell = 0 } ^{\infty } \frac{(-1)^{\ell } ( 2 \ell + \frac{3}{2})}{2(\ell + 1)\{(2\ell + 3)h r_c + 1\ } } \nonumber\\ & & \times [ 6(\ell + 1)(2\ell + 3)h r_c ( -k \eta)^{-\frac{1}{2 } } j_{2\ell + \frac{3}{2}}(-k \eta ) + 2 ( 2\ell + 2 ) ( -k \eta)^{\frac{1}{2 } } j_{2\ell + \frac{1}{2 } } ( -k \eta ) - ( -k \eta)^{\frac{3}{2 } } j_{2\ell - \frac{1}{2 } } ( -k \eta ) ] .
\nonumber\\ \label{r_brane_minus}\end{aligned}\ ] ] in @xmath4 branch ( @xmath21 ) , the curvature perturbation on the brane is given as @xmath99\nonumber\\ & & -\frac{\sqrt{2 \pi}}{6 } \frac{c_2}{k } \frac{h^2}{(2 h r_c -1 ) } \sum_{\ell = 0 } ^{\infty } \frac{(-1)^{\ell } ( 2 \ell + \frac{3}{2})}{(2\ell + 1)\{2\ell h r_c + 1\ } } \nonumber\\ & & \times [ \{6\ell ( 2\ell + 1)h r_c + 12\ell + 9\ } ( -k \eta)^{-\frac{1}{2 } } j_{2\ell + \frac{3}{2}}(-k \eta ) -4 ( \ell + 1 ) ( -k \eta)^{\frac{1}{2 } } j_{2\ell + \frac{1}{2 } } ( -k \eta ) + ( -k \eta)^{\frac{3}{2 } } j_{2\ell - \frac{1}{2 } } ( -k \eta)],\nonumber\\ \label{r_brane_plus}\end{aligned}\ ] ] where @xmath100 , \label{rb_m2_2h2}\end{aligned}\ ] ] for @xmath101 and @xmath102 , \label{rb_m2_2h2_hrc1}\end{aligned}\ ] ] for @xmath103 .
we should note that there are homogeneous solutions that satisfy @xmath104 .
these homogeneous solutions also induce curvature perturbations on the brane .
since the background cosmology given by eq .
( [ effective_friedman ] ) recovers the conventional four - dimensional cosmology based on the general relativity at high energies , we expect that the gravity behaves like four - dimensional general relativity in this limit . here , we show this fact analytically using our solutions . if we take the limit @xmath105 for fixed @xmath106 , by using the formula eq .
( [ tri_formula ] ) , the sum of the infinite ladder of the discrete modes can be performed and we obtain @xmath107 in both branches , which agree with the solutions in general relativity eqs .
( [ curv_pert_4dgr_ds ] ) , ( [ lapse_func_4dgr_ds ] ) .
therefore , we see that four - dimensional general relativity is recovered in this high energy limit . this confirms the result first obtained in ref .
@xcite . in this limit
, the brane bending vanishes @xmath108 .
we can understand this result in a very simple way as follows . in terms of @xmath46 and @xmath109 , @xmath42 and @xmath110
are given by @xmath111 then at high energies @xmath112 , the brane bending mode is shielded by the curvature of the brane @xcite . however , this argument holds only for @xmath113 . for a finite @xmath81 ,
if we do _ not _ fix @xmath114 and take large enough @xmath114 , we can no longer perform the summation of the infinite ladder of the mode and obtain the solutions ( [ rb_hihg_energy ] ) .
thus the theory deviates from general relativity .
this is related with the well - known van dam - veltman - zakharov ( vdvz ) discontinuity @xcite first discussed in the context of the pauli - fierz theory for a massive spin two field @xcite and also in dgp model @xcite .
let us consider small scales limit @xmath115 . in this region ,
@xmath116 component of the junction conditions eqs .
( [ junc_1_00 ] ) and the solution for the brane bending give @xmath117 on small scales , we can assume @xmath118 , then , eq . ( [ junc_1_00_dvz ] ) and ( [ junc_1-ij_dvz ] ) becomes completely the same as the one obtained from brans - dicke ( bd ) theory ( [ field_eq_bd_00_s ] ) and ( [ field_eq_bd_ij_s ] ) @xcite , provided that @xmath119 thus in general we do not recover four - dimensional general relativity but bd theory .
this is due to the fact that four - dimensional gravity is recovered by a continuum of massive states of five - dimensional graviton . in the high - energy limit
, we have essentially shown that the bd parameter becomes infinity for @xmath113 .
this results remind us that in the case of massive gravity , if we introduce a cosmological constant @xmath120 , the vdvz discontinuity disappears if the limits @xmath121 is taken for fixed @xmath120 @xcite .
for a finite @xmath81 , the theory is not described exactly by general relativity and this difference shows up at @xmath122 .
then we need to know what kind of bd theory is realized for a fixed @xmath81 .
the value of the bd parameter @xmath123 is obtained only by solving the equation of the motion for the bd scalar field ( [ field_eq_bd_varphi ] ) because it is the only equation which includes the bd parameter . in the bd theory ,
the bd scalar obeys a simple four - dimensional equation of motion .
however , in the brane - world model , the behavior of the brane bending mode can not be obtained unless we solve the five - dimensional perturbations .
now , since we have a solution for @xmath31 , we can find the solutions for the brane bending mode . in the next section ,
we identify the bd parameter using the solution for the brane bending .
in this section , we derive the solutions for the brane bending mode and identify the bd parameter for the effective theory on small scales . here
, we show the solutions for the brane bending mode @xmath38 in @xmath7 branch and @xmath4 branch separately . in @xmath7 branch ( @xmath22 )
, the brane bending @xmath38 is given as @xmath124\nonumber\\ & & -\frac{\sqrt{2 \pi}}{3 } \frac{c_2 h}{k } \frac{h r_c}{(2 h r_c + 1 ) } \sum_{\ell = 0 } ^{\infty } \frac{(-1)^{\ell } ( 2 \ell + \frac{3}{2})}{2(\ell + 1)\{(2\ell + 3)h r_c + 1\ } } \nonumber\\ & & \times [ 3(\ell+1)(2\ell + 3)(-k\eta)^{-\frac{1}{2}}j_{2\ell + \frac{3}{2}}(-k\eta ) -2 ( 2\ell + 2 ) ( -k \eta)^{\frac{1}{2 } } j_{2\ell + \frac{1}{2 } } ( -k \eta ) + ( -k \eta)^{\frac{3}{2 } } j_{2\ell - \frac{1}{2 } } ( -k \eta ) ] .
\label{xi_minus}\end{aligned}\ ] ] in @xmath4 branch ( @xmath21 ) , the brane bending @xmath38 is given as @xmath125 \nonumber\\ & & - \frac{\sqrt{2\pi}}{3}\frac{c_2 h}{k } \frac{h r_c}{2 h r_c -1 } \sum_{\ell = 0}^{\infty } \frac{(-1)^{\ell } ( 2 \ell + \frac{3}{2})}{(2 \ell+1)\{2 \ell hr_c + 1\}}\nonumber\\ & & \times [ -3 ( \ell + 1)(2 \ell + 3 ) ( -k\eta)^{-\frac{1}{2 } } j_{2\ell + \frac{3}{2 } } + 2 ( 2\ell + 2)(-k \eta)^{\frac{1}{2 } } j_{2 \ell + \frac{1}{2}}- ( -k \eta)^{\frac{3}{2 } } j_{2 \ell -\frac{1}{2 } } ] , \label{xi_plus}\end{aligned}\ ] ] where for @xmath96 @xmath126,\end{aligned}\ ] ] and for @xmath82 @xmath127,\end{aligned}\ ] ] there are homogeneous solutions that satisfy @xmath128 .
these are given by @xmath129 where @xmath130 is a linear combination of bessel functions and @xmath131 is determined by the spectrum of massive modes . on small scales , regardless of the value of @xmath131 , the solution for the brane bending from the homogeneous solutions are given by @xmath132 where @xmath133 and @xmath134 are arbitrary constants .
now we examine the behavior of @xmath135 in the bd theory in the presence of the scalar field @xmath72 , whose dynamics is completely the same as the brane - world model .
since we are interested in the behavior of @xmath135 on sub - horizon scale , we take the small scale limit .
then , from eqs .
( [ junc_2_0i ] ) and ( [ g_junc_cond ] ) , @xmath136 can be evaluated as @xmath137 , \label{dphi_deltadphi}\end{aligned}\ ] ] where in order to obtain eq .
( [ dphi_deltadphi ] ) , we used the fact that @xmath138 in the de sitter background and subhorizon scale .
this term serves as a source term in the equation of motion of the perturbed brans - dicke scalar field ( [ field_eq_bd_varphi ] ) which can be simplified in the de sitter background as @xmath139 where @xmath140 is a constant related with the bd parameter @xmath123 as @xmath141 .
then by solving eq .
( [ field_eq_bd_ds_varphi ] ) and keeping leading terms , we obtain @xmath142 si ( -2 k \eta)\nonumber\\ & & + \frac{1}{2 } \left[-{\bar{c}}_1 ( -k \eta ) \sin ( -k \eta ) -{\bar{c}}_2 ( -k \eta ) \cos ( -k\eta)\right ] ci ( -2 k \eta ) \nonumber\\ & & + \frac{1}{2}\left [ { \bar{c}}_1 ( -k \eta ) \ln ( -k \eta ) \sin ( -k \eta ) -{\bar{c}}_2 ( -k \eta ) \ln ( -k \eta)\cos ( -k \eta)\right ] , \label{deltaphi_bd}\end{aligned}\ ] ] where @xmath143 and @xmath144 are integration constants related with the homogeneous solutions and @xmath145 and @xmath146 are given by @xmath147 note that the homogeneous solutions behave in the same ways as the brane bending mode coming from homogeneous solutions for @xmath31 ( [ homo ] ) .
these homogeneous solutions are fixed by appropriate boundary conditions . here
, we identify the value of the brans - dicke parameter @xmath123 . for this purpose
, we evaluate eqs .
( [ xi_minus ] ) , ( [ xi_plus ] ) and ( [ deltaphi_bd ] ) numerically . in practice , we must approximate the infinite sum to proceed the calculations . in our previous paper @xcite
, we checked that we can approximate the infinite summation by introducing a cut - off @xmath148 into the summation because bessel function satisfies @xmath149 for @xmath150 . therefore ,
as long as we start from a finite time @xmath151 , we can approximate the infinite ladder of the modes by introducing sufficiently large cut - off @xmath148 and the result is insensitive to @xmath148 . here
, we compare the part of the numerical solutions including @xmath152 of eqs .
( [ xi_plus ] ) with ( [ deltaphi_bd ] ) .
the same result applies to the solution including @xmath153 . for the numerical calculation , we construct the following dimensionless quantities ; @xmath154 where the subscript @xmath155 denotes that they correspond to the part including @xmath152 .
( solid line ) and @xmath156 ( dotted line ) as functions of @xmath157 . the brans - dicke parameter is @xmath158 .
here we take @xmath159 .
the dotted line completely coincides with the solid line on small scales ( large @xmath160 ) .
the deviation become obvious around @xmath161.,width=377 ] figure 3 shows @xmath162 ( solid line ) and @xmath156 ( dotted line ) as functions of @xmath106 in the scale smaller than the hubble scale ( @xmath163 ) .
the brans - dicke parameter is chosen as @xmath158 . here
we take @xmath159 and choose the amplitudes of the homogeneous solutions for @xmath156 , @xmath143 and @xmath144 so that it agrees to @xmath162 with @xmath164 at @xmath165 .
we see that the behavior of perturbations on scales much smaller than the hubble scale is well described by the brans - dicke theory with @xmath166 .
it can also be seen that , at least for @xmath167 , the deviation of @xmath162 from @xmath156 become large and the solution is no longer well described by the brans - dicke theory with @xmath168 .
figure 4 also shows @xmath162 ( solid line ) and @xmath156 ( dotted line ) with @xmath8 as functions of @xmath106 for @xmath169 .
we also choose the homogeneous solutions so that both coincide at @xmath165 in the same way .
again we see the solution is well described by the bd theory with @xmath166 , but the deviation of @xmath170 from @xmath171 become large for @xmath172 .
this is natural because , for @xmath173 , @xmath16 becomes smaller than the horizon scale ( corresponding to @xmath174 ) .
( solid line ) and @xmath156 ( dotted line ) as functions of @xmath157 . the brans - dicke parameter is @xmath158 .
here we take @xmath173 .
the dotted line completely coincides with the solid line on small scales ( large @xmath160 ) .
the deviation become obvious around @xmath175.,width=377 ] these results are consistent with ref . @xcite . in ref .
@xcite , it is shown that the linear growth of density contrast in non - relativistic spherical collapse of dust is described the brans - dicke theory with @xmath168 for @xmath176 when @xmath81 is greater than @xmath155 . even though we do not show the results explicitly here
, we can confirm that , in the @xmath4 branch , similar results can be obtained that the perturbations can be described by the brans - dicke theory with @xmath9 on small scale .
this is also consistent with @xcite .
in the brans - dicke theory , the brans - dicke scalar @xmath177 becomes a ghost if the brans - dicke parameter satisfies @xmath178 .
this result implies that the brane bending @xmath38 is a ghost for @xmath58 in @xmath4 branch .
in this section , we study the behavior of perturbations on large scales @xmath179 . in this limit , the solution for @xmath31 is dominated by the @xmath5 mode . from eq .
( [ junc_2_0i ] ) in the limit @xmath179 , @xmath152 is related with the scalar field perturbation as @xmath180 in both branches . in the following ,
we express the curvature perturbation on the brane as the perturbation of the scalar field in this limit in both branches . in this limit ,
the @xmath5 mode of @xmath181 is , @xmath182 where we have used the fact that @xmath183 as @xmath184 .
then we can relate the perturbation of the scalar field and the curvature perturbation as @xmath185 which is the same as the standard four - dimensional result ( see eq .
( [ curv_pert_4dgr_ds_ls_fluc ] ) ) , except for the overall normalization of the effective gravitational constant .
if we take the limit @xmath186 , @xmath187 coincides with @xmath188 and we recover four - dimensional results . we can also show that @xmath189 . in this limit , for @xmath96 , the @xmath5 mode of @xmath181 becomes @xmath190 where we have used the fact that @xmath191 strictly speaking , the asymptotic value of @xmath192 and @xmath193 contain an arbitrary integration constants ( see eqs .
( [ alpha_def ] ) and ( [ beta_def ] ) .
it was shown in ref .
@xcite that this degree of freedom corresponds to a radion and dark radiation . in the following , we neglect this contribution .
then , for @xmath96 , we can relate the perturbation of the scalar field and the curvature perturbation as @xmath194 on the other hand we get @xmath195 again , if we take the limit @xmath105 , we recover the four - dimensional results . similarly , for @xmath83 , @xmath181 becomes @xmath196 where we have used the fact that @xmath197 as well as eqs .
( [ asym_alpha ] ) and ( [ asym_beta ] ) . in terms of the perturbation of the scalar field , this is given by @xmath198 we also get @xmath199 the solution for @xmath110 is diverging on large scales limit @xmath200 . in the next section , we study whether this means the physical instability on the brane or not . in order to see the origin of the difference between @xmath4 branch and @xmath7 branch , it is useful to consider effective covariant gravitational field equations on the brane derived by projecting the five - dimensional einstein equations and applying the israel junction conditions with reflection symmetry at the brane @xcite , @xmath201 where @xmath202 g_{\mu\nu } , \label{def_pi_munu}\end{aligned}\ ] ] and @xmath203 is the trace - free projection of the five - dimensional weyl tensor . in the background ,
we take @xmath204 by assuming a minkowski bulk .
however , perturbations of @xmath205 are not necessarily vanishing . for perturbations ,
the effective equations are @xmath206 @xmath207 @xmath208 substituting the solution for @xmath110 and @xmath42 , we find the perturbations of weyl tensor vanish in @xmath7 branch . in @xmath4 branch , @xmath209 , but there is a non - trivial anisotropic stress ; @xmath210 since @xmath211 , even though @xmath212 diverges as @xmath213 , the anisotropic stress from perturbations of @xmath203 does not diverge in the limit @xmath214
. since gauge invariant perturbations of weyl tensor do not diverge , the divergence of the curvature perturbation in longitudinal gauge eq .
( [ rb_ito_dphi_delphi_large_hrc1 ] ) does not imply the instability of the spacetime itself and the divergence is caused simply by a bad choice of the gauge . in order to show this ,
let us consider a four - dimensional gauge transformation , @xmath215 by choosing @xmath216 , the metric perturbations are transformed as @xmath217 dx^i dx^j\},\end{aligned}\ ] ] where @xmath218 for @xmath96 , we can eliminate the singular part in @xmath110 in the limit @xmath71 by choosing @xmath219 then the resultant metric perturbations are given by @xmath220 for @xmath83 , we can also eliminate the growing part of @xmath110 by choosing @xmath221 then , the resultant metric perturbations become @xmath222 at late times @xmath214 , @xmath223 behaves as @xmath224 thus we can find a gauge where all perturbations remain small and so there is no gravitational instability .
in this paper , we studied inflaton perturbations confined to a de sitter brane with induced gravity in a five - dimensional minkowski spacetime . for a vacuum brane
, the spin-0 mode appears as a discrete bulk mode with @xmath6 in the @xmath4 branch , while in the @xmath7 branch there are no normalizable solutions for the spin-0 modes .
since there is another discrete bulk mode ( helicity-0 mode of spin-2 perturbation ) with mass @xmath225 in the @xmath4 branch , there is a resonance between the spin-0 mode and the helicity-0 mode of spin-2 perturbation for @xmath82 . in this paper we introduced inflaton perturbations on a brane .
then an infinite ladder of discrete modes with @xmath226 and @xmath227 are excited .
since there is a mode with @xmath6 regardless of the value of @xmath81 , in the @xmath4 branch the resonance inevitably appears .
we obtained the solutions for the curvature perturbation on the brane @xmath43 and studied their behavior . at high energies , @xmath105 ,
we have confirmed that , in both branches , the four - dimensional general relativity solutions are recovered from the infinite sum of the modes .
this results remind us that in the case of massive gravity , if we introduce a cosmological constant @xmath120 , the vdvz discontinuity disappears if the limits @xmath121 is taken for fixed @xmath120 @xcite .
however , this argument holds only for @xmath113 . for a finite @xmath81 ,
if we do @xmath228 fix @xmath106 and take large enough @xmath106 , we can no longer perform the summation of the infinite ladder of the modes and the theory deviate from four - dimensional general relativity . in fact , on small scales , for a fixed @xmath81 we can see that the junction condition becomes the same as the gravitational field equations in brans - dicke theory where the brane bending mode acts as the bd scalar .
then , in order to identify the bd parameter for the effective bd theory , we compared the solutions for brans - dicke scalar field and the brane bending mode .
we have confirmed that , on sufficiently small scales , gravity is well described by the brans - dicke theory with @xmath229 in @xmath4 branch and @xmath168 in @xmath7 branch .
this is consistent with the previous work @xcite and the existence of the ghost in @xmath4 branch is confirmed .
we also studied the large scales perturbations . in @xmath7 branch ,
the solutions agree with four - dimensional gr with a modified gravitational constant . in @xmath4 branch
, there is an additional contribution to @xmath110 due to the resonance , which diverges on large scales for @xmath83 .
we identified this contribution as the effect of the weyl anisotropic stress and showed that the anisotropic stress itself does not diverge .
in fact , we can find a suitable gauge where all metric perturbations remain small , so the resonance does not lead to a gravitational instability on the brane .
we make comments on future applications of our results . in @xmath7 branch ,
if @xmath230 , small scale perturbations @xmath231 can be described by the four - dimensional brans - dicke theory . if the perturbations approaches to @xmath16 , the gravity becomes five - dimensional and we expect significant effects from the coupling to five - dimensional metric perturbations @xcite . unlike the randall - sundrum model where the small scales perturbations are always coupled to five - dimensional perturbations and a quantum vacuum state is hard to be specified @xcite , we can specify a vacuum state without ambiguity based on the four - dimensional bd theory on sufficiently small scales .
then we can estimate the effect of the coupling to five - dimensional gravity without ambiguity .
finally we should emphasize that our analysis is limited to linear perturbations .
it has been shown that we need to take into account the non - linearity of the brane bending modes before the gravity becomes non - linear .
actually , we can recover four - dimensional general relativity due to this non - linearity @xcite . ref .
@xcite suggested that this non - linear scale becomes very small at high energies by the factor @xmath232 , which is the same factor seen in eq .
( [ bendcurve ] ) .
the non - linear interactions of the brane bending mode in de sitter spacetime deserve further study .
sm is grateful to the icg , portsmouth for their hospitality when this work was initiated .
sm is supported in part by the japan society for promotion of science ( jsps ) research fellowship .
kk is supported by pparc .
in order to obtain the mode corresponding to @xmath6 , we must prepare the solution other than @xmath234 , since the junction conditions can not be satisfied by this solution .
we find that depending on whether @xmath82 or not , there are two types of the solutions .
first , we consider the case @xmath96 . in this case , we assume the following form of the solution : @xmath235 by substituting eq .
( [ omega_m^2_2h^2 ] ) into the junction condition ( [ g_junc_cond ] ) , we obtain the following relation ; @xmath236 it is worth noting that in eq .
( [ a_b_junc_cond ] ) , since @xmath237 is no longer @xmath238 , @xmath239 is not a constant but a differential operator . on the other hand , we find that the equation of the motion in the bulk is satisfied if the following conditions hold : @xmath240 by eliminating @xmath237 from eqs .
( [ a_b_junc_cond ] ) and ( [ a_bulk_cond ] ) , we get @xmath241 and by substituting eq .
( [ b_conc ] ) into eq .
( [ a_bulk_cond ] ) we obtain @xmath242 where @xmath243 next , we consider the case @xmath66 . in this case , we assume the following form of the solution : @xmath244 by substituting eq .
( [ omega_m^2_2h^2_hrc1 ] ) into the junction condition ( [ g_junc_cond ] ) , we can obtain the following relation : @xmath245 it is worth noting that in eq .
( [ a_b_junc_cond_hrc1 ] ) , since @xmath237 is no longer @xmath238 , @xmath239 is not a constant but a differential operator . on the other hand
, we find that the equation of the motion in the bulk is satisfied if the following conditions hold : @xmath246 by eliminating @xmath237 and @xmath247 from eqs .
( [ a_bulk_cond_hrc1 ] ) , ( [ b_bulk_cond_hrc1 ] ) and ( [ c_bulk_cond_hrc1 ] ) we obtain @xmath248 by substituting ( [ c_conc_hrc1 ] ) into ( [ b_bulk_cond_hrc1 ] ) , we obtain @xmath249 where @xmath192 and @xmath193 are defined as eqs .
( [ alpha_def ] ) and ( [ beta_def ] ) .
furthermore , by substituting ( [ c_conc_hrc1 ] ) and ( [ b_conc_hrc1 ] ) into ( [ a_bulk_cond_hrc1 ] ) , we get @xmath250 where @xmath251,\\ \label{til_beta_def } \tilde{\beta}(-k\eta ) & = & \beta(-k\eta ) + \frac{1}{2 } \int^{-k\eta } d(-k\bar{\eta } ) ( -k\bar{\eta})^{-2 } [ \alpha(-k\bar{\eta } ) \{\cos(-2k\bar{\eta})-1\ } -\beta ( -k\bar{\eta } ) \sin(-2k\bar{\eta } ) ] .\end{aligned}\ ] ]
in this appendix , we briefly summarize the results for cosmological perturbations in four - dimensional brans - dicke theory ( see @xcite for a review of bd theory ) . in this theory , we start with the following action : @xmath252 where @xmath177 is a brans - dicke scalar field and @xmath123 is a brans - dicke coupling constant . from eq .
( [ action_bd ] ) we can derive the field equations , @xmath253 where @xmath140 is a constant related with @xmath123 as @xmath254 for perturbations , the metric is taken as @xmath255 when we consider a canonical scalar field with a potential @xmath73 as the matter , the field equations reduce to @xmath256 = -{\phi'}^2 a + \phi ' \delta \phi ' + { v,}_\phi a^2 \delta \phi -6{\cal{h } } \delta \varphi ' + 2\nabla^2 \delta \varphi , \label{field_eq_bd_00}\\ & & 4 \varphi_0 ( { \cal{h } } a - { \cal{r } } ' ) = \phi ' \delta \phi + 2 \delta \varphi ' - 2 { \cal{h } } \delta \varphi , \label{field_eq_bd_0i}\\ & & 2 \delta \varphi ( 2{\cal{h } } ' + { \cal{h}}^2 ) -4 \varphi_0 [ ( 2{\cal{h } } ' + { \cal{h}}^2)a + { \cal{h } } a ' - { \cal{r}}''- 2{\cal{h } } { \cal{r } } ' + \frac{1}{2 } \nabla^2 ( a + { \cal{r } } ) ] \nonumber\\ & & \;\;\;\;\;\;\;\;\;\;= { \phi'}^2 a - \phi ' \delta \phi ' + { v,}_\phi a^2 \delta \phi -2 \delta \varphi '' -2{\cal{h } } \delta \varphi ' + 2 \nabla^2 \delta \varphi , \label{field_eq_bd_ii}\\ & & \varphi_0 ( a+{\cal{r } } ) = - \delta \varphi , \label{field_eq_bd_ij}\\ & & \delta \varphi '' + 2{\cal{h } } \delta \varphi ' -\nabla^2 \delta \varphi = 2 \zeta^2 [ { \phi'}a - \phi ' \delta \phi + 2 { v,}_\phi a^2 \delta \phi ] \label{field_eq_bd_varphi},\end{aligned}\ ] ] where the prime here denotes the derivative with respect to conformal time @xmath257 and @xmath258 if we require eqs .
( [ field_eq_bd_00 ] ) , ( [ field_eq_bd_0i ] ) , ( [ field_eq_bd_ii ] ) and ( [ field_eq_bd_ij ] ) reproduces the results of gr in the limit of @xmath259 and @xmath260 , the value of @xmath261 is set to be @xmath262 . for the scale much smaller than the hubble radius and
the scalar field evolves slowly in the background , eqs .
( [ field_eq_bd_00 ] ) , ( [ field_eq_bd_0i ] ) , ( [ field_eq_bd_ii ] ) and ( [ field_eq_bd_ij ] ) can be simplified as @xmath263 then the equation for curvature perturbation is obtained as @xmath264 if we take general relativity limit , we get @xmath265 and its fourier modes of @xmath45 are expressed as @xmath266 where @xmath267 and @xmath268 are integration constants @xcite .
the lapse function @xmath46 is obtained as @xmath269 if we take a large scale limit , that is , @xmath214 , from eq .
( [ lapse_func_4dgr_ds ] ) , we can obtain @xmath270 from einstein equations , the curvature perturbation is related with the scalar field fluctuation as @xmath271 k. koyama , s. mizuno and d. wands , jcap * 0508 * , 009 ( 2005 ) [ arxiv : hep - th/0506102 ] .
a. d. sakharov , dokl .
nauk sssr * 177 * , 70 , ( 1967 ) [ sov .
dokl.*12 * , 1040 ( 1968 ) ] ; k. akama , y. chikashige , t. matsuki and h. terazawa , prog .
theor . phys . * 60 * , 868(1978 ) ; k. akama , prog . theor . phys . * 60 * , 1900(1978 ) ; a. zee , phys .
. lett . * 42 * , 417 ( 1979 ) ; s. l. adler , phys .
lett . * 44 * , 1567 ( 1980 ) ; s. l. adler , phys .
* 54 * , 729 ( 1982 ) [ erratum - ibid . * 55 * , 837 ( 1983 ) ] .
a. lue , phys .
rept . * 423 * , 1 ( 2006 ) [ arxiv : astro - ph/0510068 ] .
m. bouhmadi - lopez , r. maartens and d. wands , phys .
d * 70 * , 123519 ( 2004 ) [ arxiv : hep - th/0407162 ] .
e. papantonopoulos and v. zamarias , jcap * 0410 * , 001 ( 2004 ) [ arxiv : gr - qc/0403090 ] .
h. s. zhang and r. g. cai , jcap * 0408 * , 017 ( 2004 ) [ arxiv : hep - th/0403234 ] . c. deffayet , phys .
b * 502 * , 199 ( 2001 ) [ arxiv : hep - th/0010186 ] .
m. a. luty , m. porrati and r. rattazzi , jhep * 0309 * , 029 ( 2003 ) [ arxiv : hep - th/0303116 ] .
a. nicolis and r. rattazzi , jhep * 0406 * , 059 ( 2004 ) [ arxiv : hep - th/0404159 ] .
k. koyama , phys .
d * 72 * , 123511 ( 2005 ) [ arxiv : hep - th/0503191 ] .
d. gorbunov , k. koyama and s. sibiryakov , phys .
d * 73 * , 044016 ( 2006 ) [ arxiv : hep - th/0512097 ] . c. charmousis , r. gregory , n. kaloper and a. padilla , arxiv : hep - th/0604086 .
a. lue and g. starkman , phys .
d * 67 * , 064002 ( 2003 ) [ arxiv : astro - ph/0212083 ] ; a. lue , r. scoccimarro and g. d. starkman , phys .
d * 69 * , 124015 ( 2004 ) [ arxiv : astro - ph/0401515 ] .
k. koyama and r. maartens , jcap * 0601 * , 016 ( 2006 ) [ arxiv : astro - ph/0511634 ] .
n. kaloper , phys .
d * 71 * , 086003 ( 2005 ) [ erratum - ibid .
d * 71 * , 129905 ( 2005 ) ] [ arxiv : hep - th/0502035 ] ; n. kaloper , phys .
* 94 * , 181601 ( 2005 ) [ erratum - ibid . *
95 * , 059901 ( 2005 ) ] [ arxiv : hep - th/0501028 ] .
s. mukohyama , phys .
d * 62 * , 084015 ( 2000 ) .
h. kodama , a. ishibashi and o. seto , phys .
d * 62 * , 064022 ( 2000 ) . c. deffayet , phys .
d * 66 * , 103504 ( 2002 ) [ arxiv : hep - th/0205084 ] . c. deffayet , phys .
d * 71 * , 103501 ( 2005 ) [ arxiv : gr - qc/0412114 ] .
j. garriga and t. tanaka , phys .
lett . * 84 * , ( 2000 ) 2778 .
t. tanaka , phys .
d * 69 * , 024001 ( 2004 ) [ arxiv : gr - qc/0305031 ] .
k. koyama , jcap * 0409 * , 010 ( 2004 ) [ arxiv : astro - ph/0407263 ] .
k. koyama and k. koyama , phys . rev .
d * 72 * , 043511 ( 2005 ) [ arxiv : hep - th/0501232 ] .
h. van dam and m. j. g. veltman , nucl .
b * 22 * , 397 ( 1970 ) ; v. i. zakaharov , jetp . lett .
* 12 * , 312 ( 1970 ) ; y. iwasaki , phys .
d 2 , 2255 ( 1970 ) .
m. porrati , phys .
b * 498 * , 92 ( 2001 ) [ arxiv : hep - th/0011152 ] .
a. higuchi , nucl .
b * 282 * , 397 ( 1987 ) ; a. higuchi , nucl .
b * 325 * , 745 ( 1989 ) .
i. i. kogan , s. mouslopoulos and a. papazoglou , phys .
b * 503 * , 173 ( 2001 ) [ arxiv : hep - th/0011138 ] .
u. gen and m. sasaki , prog .
* 105 * , 591 ( 2001 ) [ arxiv : gr - qc/0011078 ] .
h. yoshiguchi and k. koyama , phys . rev .
d * 70 * , 043513 ( 2004 ) [ arxiv : hep - th/0403097 ] .
t. shiromizu , k. i. maeda and m. sasaki , phys .
d * 62 * , 024012 ( 2000 ) .
k. i. maeda , s. mizuno and t. torii , phys .
d * 68 * , 024033 ( 2003 ) [ arxiv : gr - qc/0303039 ] .
k. koyama , a. mennim and d. wands , phys .
d * 72 * , 064001 ( 2005 ) [ arxiv : hep - th/0504201 ] .
h. yoshiguchi and k. koyama , phys .
d * 71 * ( 2005 ) 043519 [ arxiv : hep - th/0411056 ] . c. deffayet , g. r. dvali , g. gabadadze and a. i. vainshtein , phys .
d * 65 * , 044026 ( 2002 ) [ arxiv : hep - th/0106001 ] .
a. gruzinov , new astron .
* 10 * , 311 ( 2005 ) [ arxiv : astro - ph/0112246 ] .
i. giannakis and h. c. ren , phys .
b * 528 * , 133 ( 2002 ) [ arxiv : hep - th/0111127 ] .
m. porrati , phys . lett .
b * 534 * , 209 ( 2002 ) [ arxiv : hep - th/0203014 ] .
d. k. park , phys .
b * 562 * , 316 ( 2003 ) [ arxiv : hep - th/0304056 ] .
y. fujii and k. maeda , _ the scalar - tensor theory of gravitation _ ( cambridge univesity press , cambridge , uk , 2003 ) m. sasaki , prog .
. phys . * 76 * , 1036 ( 1986 ) ; v. f. mukhanov , sov .
jetp * 67 * , 1297 ( 1988 ) [ zh .
* 94n7 * , 1 ( 1988 ) ] .
v. f. mukhanov , h. a. feldman and r. h. brandenberger , phys .
* 215 * , 203 ( 1992 ) . | we study cosmological perturbations in the brane models with an induced einstein - hilbert term on a brane .
we consider an inflaton confined to a de sitter brane in a five - dimensional minkowski spacetime .
inflaton fluctuations excite kaluza - klein modes of bulk metric perturbations with mass @xmath0 and @xmath1 where @xmath2 is an integer .
there are two branches ( @xmath3 branches ) of solutions for the background spacetime . in the @xmath4 branch , which includes the self - accelerating universe
, a resonance appears for a mode with @xmath5 due to a spin-0 perturbation with @xmath6 .
the self - accelerating universe has a distinct feature because there is also a helicity-0 mode of spin-2 perturbations with @xmath6 . in the @xmath7 branch , which can be thought as the randall - sundrum type brane - world with the high energy quantum corrections ,
there is no resonance . at high energies
, we analytically confirm that four - dimensional einstein gravity is recovered , which is related to the disappearance of van dam - veltman - zakharov discontinuity in de sitter spacetime . on sufficiently small scales
, we confirm that the lineariaed gravity on the brane is well described by the brans - dicke theory with @xmath8 in @xmath7 branch and @xmath9 in @xmath4 branch , respectively , which confirms the existence of the ghost in @xmath4 branch .
we also study large scale perturbations . in @xmath4 branch ,
the resonance induces a non - trivial anisotropic stress on the brane via the projection of weyl tensor in the bulk , but no instability is shown to exist on the brane . |
interesting results obtained at the cern sps @xcite have heightened theoretical interest in @xmath4 production in relativistic heavy ion collisions . a new generation of models has appeared in which both dissociation and recombination of @xmath9 pairs play a significant role .
there are now model predictions for rhic central collisions ranging from the almost complete disappearance of @xmath4 @xcite , to @xmath4 enhancement @xcite .
in this paper we present the first results from phenix on @xmath4 production at rhic .
phenix @xcite detects electrons in the pseudo - rapidity range @xmath10 , and muons in the range @xmath11 . in the central arms ,
electrons are identified by matching charged particle tracks to energy deposits in the electromagnetic calorimeter and rings in the ring imaging cherenkov detector . in the muon arms ,
muon candidates are found by identifying roads in the muon identifier , and matching them to tracks found in the muon tracker . during the 2001/2002 rhic run , phenix recorded data from both @xmath2 and @xmath3 collisions . only the south muon arm was operational during that run .
[ [ pp ] ] pp ~~ data from @xmath3 collisions were recorded for an integrated luminosity of 150 @xmath12 .
the minimum bias interaction trigger for the present analysis was the beam - beam counter ( bbc ) , with at least one hit per arm .
interaction rates for @xmath3 varied from about 5 @xmath13 to 30 @xmath13 .
@xmath4 events were selected using level-1 triggers @xcite in coincidence with the minimum bias interaction trigger . the @xmath14 trigger used in this analysis ( about half of our @xmath3 data sample ) required that at least 700 mev be deposited in the electromagnetic calorimeter .
the @xmath15 trigger required at least one deep muon and one shallow muon in the muon identifier .
the level-1 triggers allowed all @xmath4 events to be recorded at the highest interaction rates .
after vertex cuts ( @xmath16 35 cm for @xmath17 and @xmath16 38 cm for @xmath18 ) and selection of good runs , about @xmath19 events were sampled for @xmath17 and @xmath20 events were sampled for @xmath18 .
the muon arm data were subdivided into two rapidity bins .
detector acceptance , level-1 trigger efficiency , and detector efficiency were estimated for a typical run using simulations of single @xmath4 decays with a realistic detector response .
an average correction for run - to - run variations in detector active area was then determined from the azimuthal distribution of electrons / event for all runs used in the analysis . for the @xmath17 case , two models of the transverse momentum distribution and a range of values for the mean p@xmath21 were used to establish the systematic uncertainty in the acceptance .
this was necessary because the p@xmath21 distribution is not well defined by the data due to low statistics .
this systematic uncertainty was estimated to be 10% . for both the @xmath17 and @xmath18 cases ,
pythia simulations were used to estimate the bbc trigger efficiency for inelastic @xmath3 events ( needed to calculate the total number of events ) , and to estimate the fraction of events containing a @xmath4 that fired the bbc trigger ( ie . the bbc trigger efficiency for @xmath4 detection ) . for cross section calculations ,
the rhic luminosity was assumed to have a systematic uncertainty of 20% in this preliminary analysis .
this uncertainty is expected to be reduced considerably with more study .
systematic errors were added in quadrature to produce an overall systematic uncertainty of 35% for @xmath22 and 29% for @xmath23 cross section data .
data from @xmath2 collisions were recorded for an integrated luminosity of 24 @xmath24 .
the minimum bias trigger used for @xmath2 collisions was a coincidence between the bbc ( 2 hits per arm ) and the zero degree calorimeter ( zdc ) . at the highest luminosities , where only a fraction of the interactions could be recorded ,
level-2 triggers @xcite were used to select @xmath4 events and minimum bias triggers were allocated only @xmath25 20% of the data bandwidth .
interaction rates varied from about 100 @xmath26 to 1600 @xmath26 . at the end of the @xmath2 run ,
the data samples from minimum bias triggers and level-2 triggers were approximately equal . in the present preliminary analysis of the @xmath2 data ,
only the @xmath0 data have been analyzed , and only minimum bias triggers were used . after applying a vertex cut of @xmath27 cm and selecting good runs , @xmath25 26 million events were analyzed .
yields were extracted using 7 different procedures for yield and background estimation , to establish the systematic errors .
the acceptance and detector efficiency were estimated using simulations of single @xmath4 decays with a realistic detector response .
the minimum bias trigger is estimated @xcite to have unit efficiency for the most central 90% of interactions used in this analysis .
the event centrality was estimated @xcite using the combined data from the bbc and zdc detectors .
the data were separated into three centrality bins .
the centrality dependence of the efficiency for offline @xmath4 reconstruction was estimated by embedding simulated single electrons into measured events and then taking the square of the single electron reconstruction efficiency as the @xmath4 reconstruction efficiency due to occupancy . combining systematic errors quadratically produced an overall systematic uncertainty per centrality bin of 26% ( 0 - 20% ) 27%(20 - 40% ) and 26%(40 - 90% ) .
the dielectron invariant mass spectrum from @xmath2 ( minimum bias ) is shown in fig . 1 , and that from @xmath3 is shown in fig .
3 shows the @xmath3 dimuon invariant mass spectrum .
invariant mass spectrum ( points ) , with unlike - sign pair background ( histogram).,scaledwidth=100.0% ] invariant mass spectrum ( points ) , with unlike - sign pair background ( histogram).,scaledwidth=100.0% ] invariant mass spectrum ( points ) , with unlike - sign pair background ( histogram).,scaledwidth=100.0% ] + the @xmath4 rapidity distribution was obtained by combining the data from the @xmath17 and @xmath18 channels , and is shown in fig .
the total cross section was obtained by fitting both a gaussian and the pythia prediction for the shape of the rapidity distribution , using the parton distribution function grv94lo .
both gave essentially the same integral , and the average of the two results for the total cross section is 3.8 @xmath16 0.6 ( stat ) @xmath16 1.3 ( sys ) @xmath8b .
this value agrees well with the prediction of the color evaporation model @xcite .
branching ratio times dn / dy scaled by n@xmath28 .
the flat line is the best fit binary scaling value .
the curve is a normal nuclear absorption model calculation @xcite.,scaledwidth=100.0% ] branching ratio times dn / dy scaled by n@xmath28 .
the flat line is the best fit binary scaling value .
the curve is a normal nuclear absorption model calculation @xcite.,scaledwidth=100.0% ] the @xmath0 yield versus the number of participants ( n@xmath29 ) is shown in fig . 5 for the @xmath3 data and for the @xmath2 data divided into three centrality bins .
rather limited conclusions can be drawn from the data in fig . 5 because of the large statistical and systematic uncertainties associated with this preliminary analysis .
we have fitted a flat line to the data to see if the data are consistent with binary scaling . the best fit , shown in fig . 5 , has a confidence level of 16% based on the statistical uncertainties alone .
thus the binary scaled @xmath4 yields probably trend down with increasing n@xmath29 , but that is not a strong statement .
we also show in fig .
5 a comparison of the n@xmath29 dependence with a calculation from a simple normal nuclear absorption model assuming a 7.1 mb absorption cross section @xcite . here the confidence level is found to be 80% .
however we wish to stress that the uncertainties in the data are too large to allow discrimination between any but the most extreme scenarios .
we expect to approximately double the number of measured @xmath30 when we add level-2 triggered events to this analysis .
we also hope to extract a comparable sample of @xmath15 data at forward rapidity from the existing data set .
m. c. abreu et al .
( na50 collaboration ) , phys .
lett * b 521 * 195 ( 2001 ) .
see for example b. zhang et al .
c * 62 * , 054905 ( 2000 ) .
see for example r. l. thews et al .
c * 63 * 054905 ( 2001 ) . the phenix collaboration , d. p. morrison et al .
, nucl . phys . *
a638 * , 565c ( 1998 ) .
the phenix collaboration , nucl .
instr . and meth .
, to be published .
k. adcox et al .
( the phenix collaboration ) , phys . rev . lett . * 86 * , 3500 ( 2001 ) .
j. f. amundson et al . , phys .
b * 390 * , 323 ( 1997 ) .
j. l. nagle and m. j. bennett , phys .
b * 465 * 21 ( 1999 ) . | first results are presented from @xmath0 and @xmath1 measurements by the phenix experiment at rhic .
these results are from a preliminary analysis of dielectrons from @xmath2 and @xmath3 data and dimuons from @xmath3 data measured in the rhic run 2 .
a total cross section for @xmath4 production in @xmath5 @xmath3 collisions of @xmath6 ( stat ) @xmath7 ( sys ) @xmath8b is found .
the @xmath0 yield is presented versus number of binary collisions . |
the experimental setup consists of two erbium doped fiber ring lasers ( edfrls ) coupled with two passive coupling lines .
each edfrl has approximately of erbium - doped fiber , the active medium , and approximately of passive single ( transverse ) mode fiber , making the total length of each cavity approximately .
the lengths of the cavities are matched within of each other .
though the doping density and the lengths of the active media are the same in both lasers , defects and imperfections in the fiber make the lasers nonidentical .
the erbium ions in the active medium are pumped with identical semiconductor lasers at a pump power of . the lasing threshold for both lasers
is approximately .
this results in approximately of light circulating within each ring .
an optical isolator is inserted within each ring cavity to ensure unidirectional propagation within the rings .
each laser contains 4 fiber - optic evanescent field couplers placed at similar locations in both rings .
these consist of two 70/30 couplers which input and output light between the lasers , one 90/10 coupler for monitoring , and a 95/5 coupler as an extra port for applications that are beyond the scope of this paper .
the locations of the couplers are shown in figure [ fig : expsetup ] .
the ports of the couplers not in use are angle cleaved to ensure that there are no back reflections and were monitored to ensure that light was propagating in the correct direction within the cavities .
the lasers are connected via two injection lines , which consist of passive single mode optical fiber , one splitter , and a variable attenuator . in this configuration
, we have the ability to monitor and control the injection amplitude between the lasers , through the splitter and variable attenuator respectively , and to observe it on an oscilloscope .
the coupling strength is defined to be the ratio of the power of light in the injection line to the power of light in the source ring .
( note that both lasers have been adjusted to have the same power . )
the lowest coupling strength that can be resolved in our system is 0.0001 .
the injection lines are approximately long , corresponding to a travel time between the two lasers of approximately , and again they are matched within . in the experiments presented here , the two coupling strengths are always the same ( symmetric coupling ) , though the electric field from each laser undergoes different phase and polarization changes due to fiber imperfections along their separate paths .
the electric field intensity of each laser is monitored using a bandwidth photodetector and a digital oscilloscope .
the optical spectrum of each laser is also monitored , and the spectra of the uncoupled lasers are matched to within .
we use the model of @xcite with the modifications introduced in @xcite for coupled fiber ring lasers . only a single polarization mode in each laser is considered here .
the model is characterized by the total population inversion @xmath8 ( averaged over the length of the fiber amplifier ) and the electric field @xmath9 in each laser .
two delay times occur in the model : @xmath10 , the cavity round - trip time , and @xmath11 , the delay in the coupling between lasers .
the equations for the model dynamics are as follows : @xmath12 e^{\text{fdb}}_j(t ) \nonumber \\ & & + \xi_j(t ) \label{eequ } \\
\frac{dw_j}{dt } & = & q -1-w_j(t ) \nonumber \\ & & -\left|e^{\text{fdb}}_j(t)\right|^2 \left\ { \exp\left[2\gamma w_j(t)\right]-1 \right\ } , \label{wequ}\end{aligned}\ ] ] where @xmath13 @xmath14 is the complex envelope of the electric field in laser @xmath15 , measured at a given reference point inside the cavity .
@xmath16 is a feedback term that includes optical feedback within laser @xmath15 and optical coupling with the other laser .
time is dimensionless , measured in units of the decay time of the atomic transition , @xmath17 .
the active medium is characterized by the dimensionless detuning @xmath18 between the transition and lasing frequencies and by the dimensionless gain @xmath19 , where @xmath20 is the material gain , @xmath21 the active fiber length , and @xmath22 the population inversion at transparency .
the ring cavity is characterized by its return coefficient @xmath23 , which represents the fraction of light remaining in the cavity after one round - trip , and the average phase change @xmath24 due to propagation of light with wavelength @xmath25 along the passive fiber of length @xmath26 and index of refraction @xmath27 .
energy input is given by the pump parameter @xmath28 .
the electric field is perturbed by a complex gaussian noise source @xmath29 with standard deviation @xmath30 .
coupling between the lasers is characterized by the coupling strength @xmath31 , which was varied from 0 to 0.014 .
the same @xmath31 is used in both lasers ( symmetric mutual coupling ) .
values of the parameters in the model , similar to those in the experimental system , are given in table [ tab : parameters ] .
.[tab : parameters]parameters used in the coupled fiber laser model . [ cols="^,^,^,<",options="header " , ] eqns .
[ eequ]-[wequ ] consist of a delay differential equation for @xmath8 coupled to a map for @xmath9 .
we integrated eqn .
[ wequ ] numerically using heun s method .
the time step for integration was @xmath32 , where @xmath33 .
this step size corresponds to dividing the ring cavity into @xmath0 spatial elements . because of the feedback term @xmath34 in eqn .
[ eequ ] , one can think of eqn .
[ eequ ] as mapping the electric field on the time interval @xmath35 $ ] to the time interval @xmath36 $ ] in the absence of coupling ( @xmath37 ) .
equivalently , because the light is traveling around the cavity , eqn .
[ eequ ] maps the electric field at all points in the ring at time @xmath38 to the electric field at all points in the ring at time @xmath39
. we can thus construct spatiotemporal plots for @xmath9 or the intensity @xmath40 by unwrapping @xmath9 into segments of length @xmath10 .
the experimental measurements were of the intensity of light from each laser after passing through a 125 mhz bandwidth photodetector .
to correspond with the experiment , we computed intensities from model and applied a low pass filter with @xmath41 mhz , multiplying the fourier transform by the transfer function @xmath42 \right\ } ^{-1}.\ ] ] it should be noted that the time step for integration is an order of magnitude smaller than the smallest timescale allowed by the filtering function .
we expect that the results are insensitive to the integration time step as long as it is sufficiently small .
test runs with a finer mesh , @xmath43 time steps per round - trip , yielded similar behavior to that described here .
all subsequent analysis of the model properties is based on the filtered intensity .
in this section we discuss the types of dynamics observed in simulations and the experimental system .
after transients have died out , each laser settles into a pattern that is approximately repeated every round - trip . the pattern may shift or change over time intervals of tens or hundreds of round - trips . due to
the presence of noise in the model , even the uncoupled lasers ( @xmath37 ) are not fully periodic for these parameter values .
typical spatiotemporal plots and time traces for the model are shown in figure [ fig : st_sim ] .
spatiotemporal plots are constructed by displaying each round - trip as a row in the diagram , colored according to the laser intensity , with subsequent round - trips forming subsequent rows .
the system displays approximately steady behavior for small coupling ( @xmath44 ) ( fig .
[ fig : st_sim]a ) .
as the coupling increases ( @xmath45 ) , it then enters a regime where pulsing occurs , in which the intensity alternates between pulsing and dropping to a low value while the inversion builds up ( fig .
[ fig : st_sim]b ) .
although the spatiotemporal plot may appear more periodic during the pulses , the intensity is actually highly irregular .
when the coupling increases further ( @xmath46 ) , the system leaves the pulsing regime and displays complicated behavior . traveling wave solutions
are commonly observed ( e.g. , fig .
[ fig : st_sim]c ) . as the coupling increases , the traveling waves become less prominent ( e.g. , fig . [ fig : st_sim]d ) .
time traces of the intensity are also shown ( figure [ fig : st_sim]e - h ) for two adjacent round - trips .
approximate repetition of the intensity pattern from one round - trip to the next can be seen .
spatiotemporal plots for the experiment are shown in figure [ fig : st_exp ] .
because the round - trip time is not an integer multiple of the 1 ns sampling time , spacetime plots were constructed by the following procedure .
the experimental intensity time trace was expanded to 10 times as many time points by linear interpolation .
subsequent round - trips were aligned by shifting the relative position of the rows to maximize correlation between the rows .
spatiotemporal plots for the experiment are normalized so that the intensities range from 0 to 1 .
the experimental system displays similar behavior to the model . when @xmath37 , approximately steady behavior is observed ( figure [ fig : st_exp]a ) .
when very small coupling is turned on , higher frequency structures emerge in the spatiotemporal pattern ( figure [ fig : st_exp]b ) .
complicated spatiotemporal patterns are seen at stronger coupling ( figure [ fig : st_exp]c ) .
pulsing was not observed in this experimental study .
however , pulsing in other fiber ring laser systems has been commonly reported in the literature ( e.g. , @xcite ) . as in the model ,
the experimental time traces show the approximate repetition of the intensity pattern in subsequent round - trips ( figure [ fig : st_exp]d - f ) .
we next consider synchronization between the two lasers .
because the lasers are mutually delay - coupled , with a delay time @xmath11 , we expect the coupling to cause lag synchronization in the lasers so that @xmath47 or @xmath48 . in figure
[ fig : synctimetraces ] , we display time traces of the two lasers with an offset of @xmath11 . for both the experiment and the model , synchronization can be seen with a time shift in _ either _ direction .
we studied the synchronization quantitatively using two metrics , the _ cross - correlation _
@xmath49 for amplitude synchronization and the _ mean phase coherence _
@xmath23 for phase synchronization .
let @xmath50 denote the correlation computed for laser 1 leading laser 2 and @xmath51 the correlation for laser 2 leading laser 1 .
thus for time series of length @xmath52 time points , @xmath53 \left [ i_b(t_j ) -\left
< i_b \right > \right ] } , \ ] ] where @xmath54 denotes averaging and @xmath55 are standard deviations in intensity for lasers @xmath20 and @xmath56 .
we define @xmath57 and @xmath58 similarly .
the phase of a time series is defined as the hilbert phase @xmath59 , computed from the hilbert transform of the intensity .
the mean phase coherence for two time series of length @xmath52 is then defined as follows @xcite : @xmath60 where @xmath61 is the phase difference between the @xmath15th points of the time series when laser @xmath20 leads laser @xmath56 .
@xmath23 ranges between 0 ( no synchronization ) and 1 ( complete phase synchronization ) .
the synchronization metrics were calculated for each round - trip , with each laser leading .
we divided the time series into round - trips and shifted each short time series of length @xmath10 by the delay @xmath11 in either direction .
( hanging ends were omitted , so that only @xmath62 was used in the calculation . )
@xmath49 and @xmath23 were then computed
. when analyzing the experimental data , we compensated for the discrete sampling by using linear interpolation to expand the data to 10 times as many time points and then calculating @xmath49 and @xmath23 for a 2.5 ns range of coupling delays ( centered around the estimated @xmath11 ) .
the maximal @xmath50 , @xmath51 , @xmath57 , and @xmath58 values for each round - trip ( maximized over possible coupling delays ) were recorded .
sample results for phase synchrony are shown in figure [ fig : sync_vs_t ] .
the synchronization fluctuates over time , and there are changes in which laser leading gives the maximum synchronization .
the amplitude synchronization behaves similarly . to determine the mean amplitude synchronization @xmath63 for a given coupling value , we take the maximum of @xmath50 and @xmath51 for each round - trip and then average over all round - trips .
we similarly define the mean phase synchronization @xmath64 over all round - trips .
mean synchronizations were computed for each value of the coupling constant @xmath31 .
experimental data were averaged over 5 - 10 separate runs of 4500 round - trips each .
simulations were averaged over 6 runs of @xmath65 round - trips each ( with runs separated by at least 0.1 s of simulated time to insure independence ) .
results are shown in figure [ fig : sync_vs_kappa ] .
error bars are the standard deviations in @xmath49 and @xmath23 over all round - trips , so they are a measure of fluctuations in synchrony . in both the model and experiments , amplitude and phase synchronization increase as the coupling increases . however , the shapes of the synchronization vs. coupling curves are different .
the experimental system is fairly well synchronized even at the smallest nonzero coupling , while the model synchronizes more gradually .
the model also exhibits more fluctuations in synchrony , as exemplified by the larger standard deviations . for both model and experiment , the amplitude synchronization ( cross - correlation )
is typically larger than the phase synchronization .
we next consider which laser leads the other during synchronization , and switches in the leader / follower . in figure
[ fig : sync_vs_t ] , the two curves are the synchrony computed with laser 1 leading ( @xmath57 ) and with laser 2 leading ( @xmath58 ) . examining the figure qualitatively , we see that at certain times the synchrony is clearly greater with one laser leading than with the other ( e.g. , round - trips 1300 - 1900 ) , and at other times it is more difficult to determine which laser is leading ( e.g. , round - trips 2500 - 2900 ) because the difference in synchrony with time shifts in either direction is small compared to the fluctuations in the synchrony .
the latter portions of figure [ fig : sync_vs_t ] are enlarged in figure [ fig : sync_vs_t_zoom ] .
we develop an algorithm to compute the leader and follower . for each round - trip
, we compute the synchronization with either laser leading .
let @xmath66 be the difference between the correlations for each round - trip @xmath67 within a run , and let @xmath68 be the standard deviation of the @xmath69 .
when @xmath70 for some cutoff factor @xmath71 , we say that the synchronization is substantially greater with one laser leading than with the other , and a clear leader and follower can be identified .
leader and follower for phase synchronization are defined similarly .
figure [ fig : cutoff ] shows the fraction of round - trips for which a leader and follower can be identified for a range of cutoff factors @xmath71 .
similar curves are obtained for the model and experiment for most values of @xmath31 .
we select @xmath72 to use as the cutoff factor in this study . for most values of @xmath31 , for both the model and experimental data
, the choice @xmath72 leads to approximately 30% of round - trips having an identified leader and follower for amplitude and phase synchronization .
after leader and follower have been computed , we next search for switches . the following algorithm is used to identify changes in leader and follower .
let @xmath73 be the set of round - trip numbers @xmath74 for which @xmath75 . whenever the sign of @xmath76 is opposite the sign of @xmath77 , a switch is identified at round - trip @xmath74 . to obtain statistics over several runs ,
let @xmath78 be the total number of round - trips and @xmath79 total number of switches in amplitude synchronization observed .
then the average number of round - trips between switches is @xmath80 .
we assume square root error in the number of observed switches , leading to an error of @xmath81 in the number of round - trips between switches .
switches in phase synchronization are located by the same procedure .
an example of switching is given in figure [ fig : sync_vs_t ] and figure [ fig : sync_vs_t_zoom ] : locations of switches identified by our algorithm are marked by arrows .
note that switches are defined only when the leader and follower change ; for example , no switch is identified around round - trip 2000 when @xmath57 and @xmath58 come together with no defined leader and then separate with the same leader and follower as before .
switches were studied in the model and experiment for a range of coupling strengths , using the same data sets as in section [ sec : synchronization ] .
results are shown in figure [ fig : switches ] .
we find that the time between switches in phase synchronization decreases as the coupling increases for both the experiment and model .
results for the amplitude synchronization are less clear . for the model ,
switching in amplitude synchronization follows the same trend as for phase synchronization .
however , the experimental system exhibits fewer switches in amplitude synchronization at larger coupling strengths .
we compared experimental measurements with model predictions for a pair of mutually coupled fiber ring lasers as the coupling strength was varied .
the coupling has similar effect on the dynamics in both model and experiment .
quantitative measures of amplitude and phase synchronization were made , the latter by using the hilbert phase .
approximately equal synchronization was found for forward and reverse time shifts equal to the coupling delay .
we notice that the lasers spend considerable periods of time in synchronized states with no discernable leader or follower , since the synchronization values are very similar for time shifts in either direction .
this is in contrast to the mutually coupled system with delay but no optical feedback described in @xcite , where leader and follower could be identified when the cross - correlation was computed for a short time series .
we defined an algorithm to locate subtle switches of leader / follower between the two lasers .
our statistical measures allow a leader and follower to be identified when the synchronization is sufficiently larger in one direction of the time shift . although the trend of increasing synchronization with increasing coupling was consistent between model and experiment , the shape of the synchronization vs. coupling curves differed , with the experimental system synchronizing more quickly .
in addition , the fluctuations in synchronization were generally larger in the model .
the model we have used includes only one polarization mode in each laser .
future extensions to the model should include both polarization modes .
this may explain the discrepancy between the experimental observations and computational prediction in the shapes of figure [ fig : sync_vs_kappa ] . in @xcite ,
the authors note that for long time series , forward and reverse time shifts yielded similar cross - correlations , although leader and follower could be identified from cross - correlations of short time series .
our cross - correlations were computed over the natural time scale of a round - trip , and similar results are obtained for shorter time series with length equal to the delay time ( data not shown ) .
our observation that the system often does not have a clear leader and follower seems robust for reasonable choices of window size for averaging , although further study of appropriate window size may be conducted .
the method for choosing the cutoff to identify clear regions of leader and follower is based on an internal definition for a given time series .
the choice of the cutoff factor @xmath71 to be one standard deviation is somewhat arbitrary , but preliminary results indicate that the trends of the switching frequency vs. coupling are preserved for different cutoff factors .
the algorithm for identifying switches may in the future be applied to other systems in which the difference in synchronization for forward and reverse time shifts is more substantial .
we typically observe a higher frequency of switching at higher coupling values .
there are very similar trends in the frequency of switches of leader / follower for the phase synchronization in experiment and theory .
this effect might be expected since the lasers mutually influence each other more effectively with stronger coupling and thus can respond to each other more quickly .
however , we can not explain the differences in corresponding plots for the amplitude synchronization .
further study of this phenomenon is needed . in this study , we considered variations in the coupling strengths only .
the noise strength for the numerical computations was chosen somewhat arbitrarily .
the main guiding principle was to have a reasonable correspondence between the uncoupled lasers in theory and experiment .
future studies will consider changes in the noise strength .
another parameter to be varied is the detuning between lasers , because its value is not known exactly for the experimental system .
the inclusion of a second polarization mode will introduce another set of parameters whose variation may also be studied .
we are indebted to jordi garcia ojalvo for developing the original version of the model and the simulation code , and we acknowledge david deshazer s assistance in generating spatiotemporal plots of experimental data .
this research was supported by the office of naval research .
lbs is currently a national research council post doctoral fellow .
ear was supported in part by a national science foundation fellowship . | a pair of coupled erbium doped fiber ring lasers is used to explore the dynamics of coupled spatiotemporal systems .
the lasers are mutually coupled with a coupling delay less than the cavity round - trip time .
we study synchronization between the two lasers in the experiment and in a delay differential equation model of the system .
because the lasers are internally perturbed by spontaneous emission , we include a noise source in the model to obtain stochastic realizations of the deterministic equations .
both amplitude synchronization and phase synchronization are considered .
we use the hilbert transform to define the phase variable and compute phase synchronization .
we find that synchronization increases with coupling strength in the experiment and the model . when the time series from two lasers are time - shifted in either direction by the delay time , approximately equal synchronization is frequently observed , so that a clear leader and follower can not be identified .
we define an algorithm to determine which laser leads the other when the synchronization is sufficiently different with one direction of time shift , and statistics of switches in leader and follower are studied . the frequency of switching between leader and follower increases with coupling strength , as might be expected since the lasers mutually influence each other more effectively with stronger coupling . *
the main goal of this paper is to explore the synchronization of mutually delay - coupled spatiotemporal systems and develop techniques to identify the evolving phase relationships between them .
we illustrate our ideas in the specific case of mutually coupled fiber ring lasers , which exhibit a remarkable ability to adapt their dynamics so that much of the time there is no clear leader or follower , even though signals propagate from one system to the other with an appreciable time delay .
the lasers are described by a system of stochastic delay differential equations , so as to include the effect of spontaneous emission in the erbium doped fiber amplifiers that constitute the active media .
we use statistical measures to study phase and amplitude synchrony in model and experimental delay - coupled systems from their time series . *
one fascinating area of dynamics in nature is the study of how systems respond to each other due to their interactions .
when systems behave dynamically , interactions between them may cause the systems to operate in a similar , or coherent , manner .
if the measured signals are similar enough over time , the correlated motion is termed synchronized . there now exist several excellent reviews on synchronization dynamics in the literature , such as refs .
@xcite , and they cover a wide range of applications from many fields of science . in general , synchronization between interacting systems may be quantified by examining and comparing the output time series from each dynamical system .
although the dynamics may be considered for the general coupling between @xmath0 systems , in this paper we restrict the number of dynamical systems to @xmath1 .
then we may consider only two coupling schemes ; either mutual ( bidirectional ) , or unidirectional .
suppose @xmath2 denote the vector output time series measured from each dynamical system .
then several types of synchronization may be classified depending on the type of coherence measure .
the systems are in complete synchronization if @xmath3 .
complete synchronization occurs in coupled phase oscillators @xcite as well as in coupled chaotic oscillators @xcite . in this case ,
amplitudes and phases are identical .
if the amplitudes are uncorrelated but the phases are locked , or entrained , between the two signals , then the systems are said to be in phase synchrony @xcite .
one other type of synchronization deals solely with the unidirectional coupling between two oscillators of drive and response type , and is termed generalized synchronization @xcite . in generalized synchronization , there exists a functional relationship between the drive and response , where there might exist a function @xmath4 such that @xmath5 in a more general setting , this may also be thought of as a generalized entrainment in dynamics , whereby one system is entrained functionally to another .
many examples of entrained systems occur in singularly perturbed problems , and specifically in systems with highly different time scales @xcite .
if there is a parameter mismatch or noise in the dynamical systems , complete synchronization may not be possible , and other measures of synchronization are needed .
one possible example of phase synchronization occurs when the amplitudes are correlated but locked in phase at a value other than zero .
such a system exhibits lag synchronization when for some @xmath6 we have @xmath7 that is , the two outputs of the coupled dynamical systems appear shifted in time with respect to one another .
for mutually coupled chaotic systems such as rossler attractors that are mismatched in frequency , lag synchronization is one of the routes to complete synchrony as coupling is increased @xcite .
it should be noted that lag synchronization may occur without the presence of delay in the coupling terms .
on the other hand , if delay is introduced into the coupling terms to model finite time signal propagation , then synchronized behavior may still occur .
when there is a clear time lag between the delay - coupled dynamics , the systems are said to exhibit achronal synchronization .
achronal synchronization exhibits a clear leading time series which is followed by a lagging time series .
heil _
et al . _
@xcite showed the existence of achronal synchronization in a delay - coupled semiconductor laser experiment , as well as in a single mode model of the delay - coupled lasers in which stochastic effects modeling spontaneous emission are included .
the time series shows a clear leader with delay equal to the coupling delay time .
other groups have also considered leader - follower synchronization in single - mode semiconductor equations @xcite .
one of the non - intuitive facets of interacting systems which synchronize with delay is that of anticipation in systems with short time delay .
first observed in unidirectionally coupled systems , in contrast to lag synchronization , anticipatory synchronization occurs when a response in a system s state is replicated not simultaneously but anticipated by the response system @xcite .
an example of anticipation in synchronization is found in coupled semiconductor lasers @xcite . here , the author followed ref .
@xcite in the design of a unidirectional coupling arrangement , in which two single mode semiconductor lasers with delayed optical feedback and delayed injection coupling were modeled .
cross - correlation statistics between the two intensities showed clear maxima at delay times consisting of the difference between the feedback and the coupling delay .
anticipatory responses in the presence of stochastic drives , equally applied to transmitter and receiver , have been observed in models of excitable media as well @xcite .
given that both lag and anticipatory dynamics may be observed in delay - coupled systems which are deterministic , it is natural to ask whether the systems may exhibit coexisting features . in single mode
mutually coupled lasers , this is indeed the case @xcite . in fact , in the absence of noise ( the authors include a noise term in their model , but turn it off for the simulations ) , switching between leading and following state is observed . in a similar model with spontaneous emission included as a noise source , theory and experiment have exhibited achronal synchronization @xcite , with switching between leader and follower .
the authors conjecture that for semiconductor lasers , changes in leader - follower roles may occur during sharp dropouts of the laser intensity . in the above discussion
, it is clear that the role of signal propagation time is important in the dynamics of leader and follower in coupled systems . when noise is sufficiently large , it is not easy to detect changes in leader and follower statistically .
this is especially the case when the systems are stochastic and close to a synchronized state . in ref .
@xcite , the authors compute statistics on the phase differences and show that they might correspond to a random walk . in this paper
, we develop methods to extract statistical behavior of the switches in leader and follower in mutually coupled fiber lasers . the chaotic dynamics of fiber ring lasers have been studied in the past .
an experiment on the coupling between polarization modes was set up and modeled using delay differential equations in @xcite .
other experiments on synchronization with fiber lasers have been reported in @xcite , and noise - induced generalized synchronization in fiber ring lasers has been reported in @xcite .
modeling the ring laser yields a system of equations which consists of coupled difference and differential delay equations . to obtain better agreement with experiment , it was found that inclusion of spontaneous emission effects was necessary in the modeling , which resulted in a stochastic difference - differential system of equations @xcite , and it is this approach we follow here . for mutual coupling of two ring lasers with delay ,
we analyze both experimental results and a delay differential equation model of the system .
synchronization with delay occurs and increases with coupling strength .
generally , approximately equal synchronization is observed between the two lasers when time - shifted in either direction by the delay time , so that a clear leader and follower can not be identified .
we define an algorithm to determine which laser leads the other when the synchronization is sufficiently different , and statistics of switches in leader and follower are studied .
we introduce the experimental setup and corresponding model in sec .
[ sec : experiment]-[sec : model ] .
the dynamics of the system are outlined in sec .
[ sec : dynamics ] .
synchronization of the coupled lasers is discussed in sec .
[ sec : synchronization ] and switching of leader and follower in sec .
[ sec : switching ] . |
newly formed stars are observed to have infrared ( ir ) excess due to their circumstellar disk composed of dust and gas ( @xcite ) .
most older main - sequence ( ms ) stars , on the other hand , have photospheric emission with no excess in the ir .
it is intuitive to conclude that the circumstellar disk evolves with time , gradually getting rid of the ir excess .
one of the main questions in stellar astrophysics is how this happens .
observational studies , as well as theoretical simulations , have demonstrated the interaction between star and disk .
the stellar radiation facilitates disk evolution in terms of photoevaporation ( e.g. , @xcite ) or dust growth and settling ( e.g. , @xcite ) . in the other direction
, mass is accreted from the disk to the star following magnetic field lines ( e.g. , @xcite ) .
a diversity of stellar temperatures , luminosities and masses among young stars has been known and studied for decades . facilitated by new ir and ( sub-)millimeter observations , a great variety of disk shapes , structures and masses
is now being actively studied .
the next step is to try to connect stellar and disk characteristics in order to understand the evolution of these systems .
the study of a single object , however , is unlikely to provide unambiguous information regarding the evolutionary stage of the associated disk .
most studies to date refer to samples of young stars scattered across the sky , or to sources distributed across large star - forming clouds like taurus .
in addition to evolutionary stage , the specific environment in which the stars are formed may influence the evolution of disks by dynamical and radiative interaction with other stars or through the initial conditions of the starting cloud , making it difficult to separate the evolutionary effects ( e.g. , @xcite ) .
for this reason , clusters of stars are very often used as laboratories for calibrating the evolutionary sequence ( e.g. , @xcite ) .
the power of this method , to gain statistical information on disk composition in coeval samples , was found to be very successful for loose associations of older , pre - main sequence stars such as the 8 myr old @xmath2 cha @xcite and the 10 myr old tw hydrae association @xcite .
identifying clusters of even younger disk populations is a natural step towards the completion of the empirical calibration of the evolution of disks surrounding young low - mass stars .
this paper analyzes the inner disk properties of a flux - limited , complete unbiased sample of young stars with ir excess in the serpens molecular cloud ( d = 259 @xmath3 37 pc , @xcite ) which has a mean age @xmath4 myr , @xcite with an yso population in clusters and also in isolation .
it has been recently argued ( l. loinard , private communication ) that the distance to serpens could be considerably higher than previously calculated .
this would imply a rather younger median age for this cloud .
the spitzer legacy program `` from molecular cores to planet - forming disks '' ( c2d ) has uncovered hundreds of objects with ir excess in five star - forming clouds ( cha ii , lupus , ophiuchus , perseus and serpens ) , and allowed statistical studies within a given cloud @xcite .
the c2d study of serpens with irac ( 3.6 , 4.5 , 5.8 and 8.0 @xmath0 m , @xcite ) and mips ( 24 and 70 @xmath0 m , @xcite ) data has revealed a rich population of mostly previously unknown young stellar objects ( ysos ) associated with ir excess , yielding a diversity of disk seds ( @xcite ) . because of the compact area in serpens mapped by spitzer ( 0.89 deg@xmath5 ) , this impressive diversity of disks presents itself as an excellent laboratory for studies of early stellar evolution and planet formation .
indeed , the serpens core ( cluster a , located in the northeastern part of the area studied by c2d ) has been well studied in this sense ( e.g. , @xcite ) , whereas only some of the objects in group c ( formerly known as cluster c ) were studied with isocam data @xcite .
because of its wavelength coverage , sensitivity and mapping capabilities , the spitzer space telescope has offered an opportunity to study many of these systems ( star@xmath6disk ) in unprecedented detail .
spitzer s photometry in the mid - ir , where the radiation reprocessed by the dust is dominant , gives information on the shape of the disks and , indirectly , its evolutionary stage ( assuming an evolution from flared to flat disks ) .
follow - up mid - ir spectroscopic observations with the infrared spectrograph ( irs , 5 38 , @xmath0 m @xcite ) on - board spitzer probe the physical and chemical processes affecting the hot dust in the surface layers of the inner regions of the disk . the shapes and strengths of the silicate features provide information on dust grain size distribution and structure ( e.g. , @xcite ) .
these , in turn , reflect dynamical processes such as radial and vertical mixing , and physical processes such as annealing . a smooth strong si o stretching mode feature centered at 9.8
@xmath0 m is indicative of small amorphous silicates ( like those found in the ism ) while a structured weaker and broader feature reveals bigger grains or the presence of crystalline silicates .
polycyclic aromatic hydrocarbon ( pah ) features are a probe of the uv radiation incident on the disk , whereas their abundance plays a crucial role in models of disk heating and chemistry ( e.g. , @xcite ) .
the shape and slope of the mid - infrared excess provides information on the flaring geometry of the disks @xcite , while ice bands may form for highly inclined sources ( edge - on ) where the light from the central object passes through the dusty material in the outer parts of the disk @xcite .
thus , the wavelength range probed by the irs spectra enables analysis of the geometry of individual disks .
it also probes the temperature and dust size distributions as well as crystallinity of dust in the disk surface at radii of 0.1 few au . statistical results from a number of sources help the understanding of the progression of disk clearing and possibly planet formation .
our group has been conducting multi - wavelength observing campaigns of serpens .
optical and near - ir wavelength data , where the stellar radiation dominates , are being used to characterize the central sources of these systems @xcite .
effective temperatures , luminosities , extinctions , mass accretion rates , as well as relative ages and masses are being determined . in this paper , we present a complete flux - limited set of spitzer irs spectra for this previously unexplored young stellar population in serpens .
we analyze these spectra in terms of common and individual characteristics and compare the results to those of taurus , one of the best studied molecular clouds to date and dominated by isolated star formation .
a subsequent paper will deal with the full sed fitting for the disk sources in serpens and their detailed analyzes .
our ultimate goal is to statistically trace the evolution of young low - mass stars by means of the spectroscopic signatures of disk evolution , discussed above .
section [ sdata ] describes our spitzer irs data . in
[ sres ] we divide our sample into categories based on their irs spectra : the background contaminants are presented in
[ sbg ] separated according to the nature of the objects ; embedded sources are presented in [ semb ] , and disk sources in [ sdisks ] ( with emphasis on pah and silicate emission ) . in [ sdis ]
we discuss disk properties in relation to environment and to another cloud , taurus , for comparison . in
[ scon ] we present our conclusions .
@xcite describe the selection criteria used by the c2d team to identify yso candidates based on the spitzer irac and mips data , band - merged with the 2mass catalog .
they used a number of color - color and color - magnitude diagrams to separate ysos ( both embedded ysos and young stars with disks ) from other types of sources , such as background galaxies and stars . in this manner , a set of 235 ysos was identified in serpens , in both clusters and in isolation , as described in [ scf ] .
these criteria , however , are not fail - proof .
@xcite treated this problem very carefully in a statistical sense , but the locus region for ysos and , for instance , background galaxies overlap somewhat in color - color diagrams with spitzer photometry . therefore , some contamination may well still be present in the sample which was impossible to disentangle from photometry alone .
an additional criterion was applied to this sample in order to guarantee irs spectra with sufficient quality to allow comparative studies of solid - state features , namely s / n @xmath1 30 on the continuum
. a lower limit flux cutoff of 3 mjy at 8@xmath0 m was imposed .
this flux limit ensures coverage down to the brown dwarf limit ( @xmath7 0.01 l@xmath8 ) and leads to a final sample of 147 objects , the same as in @xcite .
this is a complete flux - limited sample of ir excess sources in the c2d mapped area , except in the serpens core .
it is also important to note that this sample , by definition , does not include young stars without ir excess ( class iii ) .
figure [ objirac ] shows the distribution of our observed objects in the serpens molecular cloud .
table [ tab1 ] lists the sample together with their 224 @xmath0 m spectral slope
. observed objects over - plotted on the c2d irac4 ( 8.0 @xmath0 m ) map of serpens .
the contours ( 5 , 10 , 15 , 20 and 25 mag ) of visual extinction are derived from the c2d extinction maps @xcite .
blue squares are background stars , cyan ellipses are redshifted galaxies , red circles are embedded sources , and stars represent disks ( yellow disks with pah emission and green other disk sources ) .
the pink triangle is the pn candidate , see [ sbgpn ] for details.,scaledwidth=45.0% ] out of the 147 objects in this sample , 22 were observed between september 2004 and june 2006 , as part of the c2d irs 2nd look campaign .
the reduced high resolution spectra ( r = @xmath9 = 600 ; short high [ sh ] , 9.9
19.6 @xmath0 m and long high [ lh ] , 18.7 37.2 @xmath0 m ) , complemented with short low spectra ( sl , r = 50 100 ; 5.2 14.5 @xmath0 m ) are public to the community through the c2d delivery website @xcite .
the other 125 targets were observed as part of a spitzer space telescope s cycle 3 program ( go3 # 30223 , pi : pontoppidan ) in the low resolution module ( sl and long low [ ll ] , 14.0
38.0 @xmath0 m ) in 35 aors , between october 2006 and april 2007 .
all objects were observed in irs staring mode and extracted from the ssc pipeline ( version s12.0.2 for the c2d irs 2nd look program , and version s15.3.0 for the go3 program ) basic calibration data ( bcd ) using the c2d reduction pipeline @xcite .
this process includes bad - pixel correction , optimal psf aperture extraction , defringing and order matching .
technical information about each object can be found in table [ tab1 ] .
the large number of objects studied here allows statistical studies of each stage of disk evolution in a single star - forming region . by simply looking at the spectra
, a diversity of spectral shapes can be noted immediately .
this diversity of objects can be divided into several categories , each of which is thought to be related to a different evolutionary stage of the star@xmath6disk system , described below . due to the low galactic latitude of serpens ( @xmath10=030.4733 , @xmath11=+05.1018 ) the contamination by galactic background sources
is expected to be higher for this region than for other studied star - forming regions ( e.g. , taurus , lupus , chamaeleon ) . with the help of the irs spectra , background contaminants can readily be identified , even when seen through a large column density of molecular cloud material .
these are presented and discussed in appendix [ abg ] . to further study the ysos , the 32 contaminants are removed from the sample , which then consists of 115 objects for further analysis .
a newly formed protostar , still embedded in an envelope of gas and dust , will show an excess in the mid - ir that rises with wavelength . in this cold environment ,
a significant fraction of the molecules are found in ices .
many of these ices are observed as absorption features in the wavelength range of the irs instrument ( e.g. , @xcite ) . of our sample
, 21 sources have spectra consistent with embedded sources .
these spectra are shown in figure [ irsemb ] .
the most prominent ice feature , co@xmath12 at 15.2 @xmath0 m , can readily be seen in all spectra .
weaker features , such as h@xmath12o at 6.0 and 6.85 @xmath0 m and ch@xmath13oh at 9.7 @xmath0 m , need high s / n .
those features are detected in objects # 26 , 27 , 28 , 44 , 45 , 47 , 49 , 67 , 72 , 73 , 135 and 140 . as can be seen in figure [ objirac ] ,
the embedded objects seem to be concentrated in high extinction regions ( @xmath14 mag ) .
the embedded objects observed as part of the c2d irs 2nd look program have been studied in detail in a series of papers on ices around ysos ( h@xmath12o in @xcite ; co@xmath12 in @xcite ; ch@xmath15 at 7.7 @xmath0 m in @xcite ; and nh@xmath13 at 9.0 @xmath0 m and ch@xmath13oh in @xcite ) and are not discussed further in this paper .
irs spectra of the embedded sources in serpens with rising mid - ir spectra .
the prominent ice absorption features ( at 6.0 , 6.85 , 7.7 , 9.0 and 15.2 @xmath0 m ) are marked , together with the silicate absorption feature at 9.7 @xmath0 m .
the object # is indicated at the top right of each panel . ]
young stellar objects with ir excess that have a flat or negative slope in the near- to mid - ir are identified as disk sources . following @xcite and @xcite , flat sources have @xmath16 , class ii sources have @xmath17 , and class iii sources have @xmath18 .
many of these sources also have silicate emission bands and a few show pah features . in the following ,
these different features are analyzed and discussed in detail for the total of 94 disk sources .
disk geometry is inferred from the flux ratio between 30 and 13 @xmath0 m ( @xmath19 ) .
high ratios ( @xmath20 ) yield edge - on disks , intermediate values ( @xmath21 and @xmath22 ) identify cold disks and flared disks with considerable excess in the ir , respectively , and low ratios ( @xmath23 ) select flat , settled disks with little ir excess @xcite .
figure [ irsshape ] shows the distribution of @xmath19 for the disk population in serpens .
distribution of the flux ratio between 30 and 13 @xmath0 m , used as a proxy for disk geometry.,scaledwidth=50.0% ] cold disks , sometimes referred to as transitional disks , as studied in @xcite and @xcite , present a peculiar spectral energy distribution ( sed ) , with inner gaps or holes producing a region with no ir excess in the near- to mid - ir but a substantial excess at mid- to far - ir wavelengths .
binaries @xcite , planet formation @xcite and photoevaporation @xcite are some of the suggested mechanisms for this removal of warm dust .
following @xcite , @xmath19 is used to differentiate cold disks from the general disk population .
those wavelengths were chosen to avoid the silicate features and to probe the spectra steeply rising at mid- to far - ir following a deficit at shorter wavelengths , characteristic of cold disks .
eight objects ( # 9 , 21 , 69 , 82 , 90 , 113 , 114 and 122 ) in this sample are classified as cold disks based on @xmath24 , corresponding to 8.5% of the disk population . these spectra are shown in figure [ irscd ] .
for clarity , the original spectra are shown in grey , while in black a binned version is overplotted .
the binning was done using a savitzky - golay @xcite filter of order 0 and degree 3 , for illustrative purposes only .
this is the most unbiased survey of cold disks to date , based entirely on spectroscopy in the critical mid - ir range , and should thus give a representative fraction of such disks .
these objects are also part of a larger cold disk sample , studied by @xcite , where the authors have selected cold disks from photometric criteria in all five c2d clouds .
irs spectra of the cold disks in serpens .
object # is indicated in each panel . in grey
are the original spectra , while in black a binned version is overplotted .
the dotted line represents the stellar photosphere , for comparison . ] polycyclic aromatic hydrocarbons ( pah ) are large molecules whose emission is excited by uv radiation out to large distances from the star .
thus , these features are good diagnostics of the amount of stellar uv intercepted by the disk surface .
almost half of the sample of herbig ae / be stars ( young stars of intermediate masses ) studied by @xcite and @xcite show pah emission .
@xcite , on the other hand , show that pah emission is not as common among low - mass ysos as it is for more massive stars .
the authors argue that the lack of features is not only due to their weaker uv radiation fields , but also to a lower pah abundance compared with the general ism .
irs spectra of the disk sources which show clear pah features .
the dashed lines indicate the positions of the most prominent bands ( at 6.2 , 7.7 , 8.6 , 11.2 , 12.8 and 16.4 @xmath0 m ) .
the object # and its spectral type ( from @xcite ) are indicated at the top right of each panel .
note that the ratios of different line strengths differ between sources .
, scaledwidth=50.0% ] in the current sample , 5 sources have clear pah emission , presented in figure [ irspah ] .
although a great variation in shapes and strengths is seen , all sources show the features at 6.2 , 7.7 , 8.6 , 11.2 and 12.8 @xmath0 m .
objects # 52 , 97 and 131 also show a feature that could be the pah band at 16.4 @xmath0 m , although the feature does not match perfectly in position .
the central stars in these 5 sources were studied with optical spectroscopy in @xcite .
spectral types and luminosities were derived , yielding ages and masses for each object .
objects # 52 , 98 , 120 and 131 have masses between 1.9 and 2.7 @xmath25 ( spectral type g3 to a2 ) , while object # 97 has lower mass , of around 0.45 @xmath25 ( m2 ) .
objects # 97 and 98 are spatially very close to each other and have very similar pah features .
irac band 4 ( 8 @xmath0 m ) images show that there is extended nebulosity surrounding both positions so that the pahs are likely associated with this nebula rather than the disks . discarding these 2 sources
, this sample indicates that pahs are present in 3.2% of the disk population studied , which is below the 1114% of pah detection rate found by @xcite
. this discrepancy could be due to the two samples being differently selected , with the current sample having a larger fraction of late - type stars .
the percentage of pahs present in disks in serpens is , however , very comparable with that in taurus ( 4% of the disk population ,
[ stau ] ) .
interestingly , none of the cold disks show pahs , different from expected if grain settling enhances pah features @xcite and observed for other cold disk samples studied which do show pahs @xcite . in addition , the objects with pah emission have standard @xmath19 flux ratios , falling in the ` flat ' and ` flared ' regions of figure [ irsshape ] .
only object # 120 shows a silicate feature ( at 20 @xmath0 m , as discussed in
[ ssi ] ) , in addition to pahs .
this alludes to an anti - correlation between pahs and silicates that could be simply explained in terms of contrast , i.e. , it is hard to detect pah features when a strong 10 @xmath0 m silicate feature is present @xcite . in this sense , pahs should be more easily detected in disks without silicates , since the falling spectrum provides an increase in the feature to continuum ratio improving contrast .
silicate emission has been observed around numerous circumstellar disks ( e.g. , @xcite ) .
some of this material has different characteristics in disks compared to primitive interstellar material .
the strength and the shape of silicate features ( at 10 and 20 @xmath0 m ) contain important information on the dust sizes in the surface layer of the disk .
the 89 disk sources in this sample were inspected for their silicate characteristics .
their spectra can be seen in figures [ irsdisk1 ] to [ irsdisk3 ] ( completed with figures [ irscd ] and [ irspah ] ) , arranged in order of decreasing strength of the 10 @xmath0 m feature and decreasing @xmath26 spectral slope .
no extinction correction has been applied to these spectra ( see [ ssi_ext ] ) .
irs spectra of the disk sources in serpens .
object # is indicated in each panel . in grey
are the original spectra , while in black a binned version ( using a savitzky - golay @xcite filter of order 0 and degree 3 ) is overplotted .
] irs spectra of the disk sources in serpens , contd . ]
irs spectra of the disk sources in serpens , contd . ]
five of these sources ( 5% of the disk population ) have spectra that are featureless , i.e. , they have neither the 10 nor the 20 @xmath0 m silicate features in emission , although they present ir excess ( objects # 38 , 42 , 59 , 83 , and 87 ) .
it is found that when the 10 @xmath0 m feature appears , the 20 @xmath0 m band is also present ( by visual inspection of the continuum - subtracted spectra , see [ ssi_ssr ] ) .
the contrary is not true , as objects # 13 , 20 , 24 , 32 , 62 , 64 , 69 , 70 , 109 , 116 , 127 , 139 , 143 and 145 have an observed 20 @xmath0 m feature but no apparent 10 @xmath0 m feature ( also reported by @xcite for t cha , sr21 and hd 135344b ) .
also , object # 120 shows the 20 @xmath0 m band and no 10 @xmath0 m feature , as this region is dominated by pah features in this source ( see [ spah ] ) .
when present , the 10 @xmath0 m feature ( @xmath27 ) is usually stronger than the 20 @xmath0 m ( @xmath28 ) one as is common in disk sources .
crystalline silicate features ( as discussed by e.g. @xcite ) are seen in several sources .
a detailed analysis in terms of statistics and crystalline fraction of these features is non - trivial and beyond the scope of this paper , but will be subject of a future publication .
as first shown for herbig ae stars @xcite , there is a dependence between grain size and silicate shape and strength for both 10 and 20 @xmath0 m features : as the sizes of the grains grow , the silicate features appear more and more flattened and shift to longer wavelengths .
a statistical analysis of these two features can be done using the following quantities : @xmath27 is a measure of the strength of the feature centered at 10 @xmath0 m and @xmath29 is a proxy for the shape of the feature ( the smaller this value , the more peaked the feature is ) ; @xmath30 measures the strength of the 20 @xmath0 m feature , while @xmath31 is a proxy for this feature shape . following @xcite ,
@xmath32 is defined as : @xmath33 in this paper , we use @xmath34 m .
@xcite found a global trend ( best fit to observations ) for both features ( figure 9 in their paper and [ si ] in this one ) , indicating that the feature shape and strength are physically related , as expected if the silicate features are indeed tracers of grain size .
the relationship is much tighter for the 10 @xmath0 m feature , as also seen by the great spread in the correlation for the 20 @xmath0 m feature .
following their procedure , we first subtracted the continuum of each spectrum by fitting a second order polynomial to the following regions : blue - ward of the 10 @xmath0 m feature , between the 10 and the 20 @xmath0 m features , and red - ward of the 20 @xmath0 m feature ( 6.87.5 , 12.513.5 and 3036 @xmath0 m ) .
the peak fluxes at 9.8 , 10 , 11.3 , 19 , 20 and 23.8 @xmath0 m were then determined ( see table [ tsi ] ) .
figure [ si ] shows the results for our sample of disks , excluding those with pah emission .
the best fit found by @xcite for the c2d irs first look sample of ysos has been indicated for comparison ( solid line ) .
the agreement of the serpens data with the best fit of @xcite is excellent , and more evident for the 10 @xmath0 m than for the 20 @xmath0 m feature due to the larger scatter in the latter group . in the left panel ,
black dots show the ratio of peaks at 11.3 to 9.8 @xmath0 m ( @xmath29 ) plotted against the peak at 10 @xmath0 m ( @xmath27 ) . in the right panel ,
the ratio of peaks at 23.8 to 19 @xmath0 m ( @xmath31 ) is plotted against the peak at 20 @xmath0 m ( @xmath35 ) .
the best fit found by @xcite for the c2d irs first look sample of ysos ( grey dots ) has been indicated for comparison ( solid black line ) .
green squares are the cold disks in this sample ( see [ scd ] ) . colored curves are derived from theoretical opacities for different mixtures by @xcite .
the open circles correspond to different grain sizes , from left to right 6.25 , 5.2 , 4.3 , 3.25 , 2.7 , 2.0 , 1.5 , 1.25 , 1.0 and 0.1 @xmath0 m .
typical uncertainties for peak strength are @xmath360.1 , while typical uncertainties for the peak ratios are @xmath360.12 .
] @xcite generated synthetic 10 @xmath0 m features calculated for different grain sizes and compositions . their models are generated for amorphous silicates of olivine and pyroxene stoichiometry and a 50:50 mixture , with grain size varying between 0.1 and @xmath366 @xmath0 m .
those models are overplotted in the left panel of figure [ si ] , with open symbols corresponding to different grain sizes .
the comparison of the data presented here with these models shows that the majority of our sample lies in a region consistent with an opacity dominated by grains with sizes larger than 2.0 @xmath0 m , having no features consistent with grain sizes smaller than 1.5 @xmath0 m .
a considerable fraction of the sample is consistent with grains as large as 6 @xmath0 m ( the largest grain size modeled ) .
the precise sizes depend on composition and treatment of the opacities , but sizes of a few @xmath0 m are clearly implicated .
it is important to note that crystallinity also affects the shape of silicate features .
however , as shown by @xcite and @xcite , the effect of crystallinity is orthogonal to that of grain sizes in the strength versus shape plot ( figure [ si ] ) .
grain sizes are found to be the dominant parameter changing the shape of the 10 @xmath0 m silicate feature , whereas crystallinity introduces scatter ( figure 13 of @xcite ) .
all cold disks present silicate features in emission . according to the models of grain sizes ( left panel of figure [ si ] ) ,
the cold disks in this sample have grains bigger than 2.7 @xmath0 m , with the bulk of the sample presenting grains as big as 6.0 @xmath0 m , according the these models . the amount of foreground extinction is , for some sources , substantial enough to affect the appearance of silicate emission features in the mid - infrared .
because the extinction law has strong resonances from silicates , the strength and shape of any silicate emission feature may be significantly affected if a large column of cloud material is present in front of a given source . here , the potential effects of extinction on statistical results , such as those presented in figure [ si ] , are discussed .
the model presented by @xcite for an absolute to selective extinction of @xmath37 has been found to be a reasonable match to the observed dark cloud extinction law @xcite , and it is assumed that this law holds for serpens .
one caveat is the lack of resonances due to ices in this model , but given other uncertainties in the determination of extinction laws , this is probably a minor contribution for the purposes of this work . the opacity caused by a silicate resonance is defined as ( e.g. @xcite ) : @xmath38 where @xmath39 is the optical depth of the `` continuum '' opacity without the silicate resonance .
the relation between optical extinction and @xmath40 is , using the fact that the optical depth is proportional to the extinction coefficient @xmath41 : @xmath42 using equations [ es ] and [ x ] and by approximating @xmath43 , the silicate strength corrected for extinction is : @xmath44 for the @xmath37 extinction law of @xcite , the relevant values for @xmath45 are 19.6 , 20.9 and 41.6mag for 9.8 , 10.0 and 11.3@xmath0 m , respectively . using this relationship , it is possible to correct the position of sources with known values for @xmath46 in figure [ si ] .
@xcite measured @xmath46 values for 49 of the 89 disks studied here albeit with some uncertainty .
the resulting extinction - corrected strength - shape distribution is shown in figure [ si_cor ] ( in green ) and compared with the uncorrected distribution ( in black ) .
it is seen that extinction moves points along curves that are almost parallel to the relation between shape and strength .
most points , having @xmath47mag , are not changed enough to alter the slope of the relation , and the median strength is unaffected .
however , a few highly extincted disks are corrected by large amounts , as can be seen by the extinction arrows in figure [ si_cor ] .
comparison between the observed values ( in black ) and the extinction corrected values ( in green ) for the 10 @xmath0 m silicate feature .
the arrows give an indication of the effects of extinction on the feature , however their length depends on the starting point ( see text for details ) .
, scaledwidth=50.0% ] there are 40 ysos in the sample for which @xcite do not estimate @xmath46
. some of these may be highly extincted ( @xmath48 mag ) sources ( not being bright enough for optical spectroscopy using a 4-m class telescope ) .
the correction for extinction , as derived above , would be greater for such objects , possibly introducing a displacement between the distributions of extinction corrected and uncorrected features .
however , it is important to note that this assumes that the @xmath37 extinction law is valid for the densest regions of molecular clouds .
@xcite recently found that at @xmath49mag , the relationship between optical depth and extinction derived for diffuse ism is no longer valid , possibly due to grain growth , and that @xmath50 is significantly larger in this regime .
if true , then the @xcite extinction law represents a `` worst case '' scenario .
since most of the measured extinctions are small , and given the uncertainty in extinction laws and the lack of @xmath46 data for a significant fraction of the sample , the uncorrected silicate feature strengths are used in the remainder of the paper .
figure [ sipeak ] shows the strengths of the observed 10 and 20 @xmath0 m features with no apparent correlation . to quantify a correlation ,
a kendall @xmath51 rank correlation coefficient is used to measure the degree of correspondence between two populations and to assess the significance of said correspondence .
in other words , if there is a correlation ( anti - correlation ) between two datasets , the kendall @xmath51 rank coefficient is equal to 1 ( -1 ) .
if the datasets are completely independent , the coefficient has value 0 .
for the strengths of the 10 and 20 @xmath0 m silicate features , @xmath52 , meaning at best a weak correlation .
the lack of apparent correlation between the 10 and the 20 @xmath0 m features , in both strength and appearance , suggests that they are emitted by different regions in the disk .
indeed , several authors have shown that the 10 @xmath0 m feature is emitted by warm dust in the inner region , while the 20 @xmath0 m feature originates from a colder component , further out and deeper into the disk ( e.g. , @xcite ) .
therefore , the absence of the 10 @xmath0 m feature but presence of the 20 @xmath0 m band for a given source implies that such disk lacks warmer small dust grains close to the star .
this reminisce the disk around the herbig ae star hd 100453 studied by @xcite and @xcite , who find that the absence of the 10 @xmath0 m silicate feature can be fitted with a model deprived in warm small grains .
16% of the disk sources shown in this sample fit this scenario .
a possible explanation for such a scenario is that those disks have holes or gaps on such a small scale ( @xmath53 1 au ) that the holes do not produce a strong signature in the spectra probed by our data , as the cold disks do .
relative strength of the 10 vs. 20 @xmath0 m silicate features .
typical error bars are shown in the bottom right corner .
no obvious correlation is found .
, scaledwidth=50.0% ]
in this section we discuss the properties of the yso sample presented in sections [ semb ] and [ sdisks ] with respect to environment , and compare the results with those for another nearby star - forming region , taurus @xcite , as well as the full c2d irs sample ( @xcite ) .
comparison of the disk properties between the cluster and field populations determines the importance of environment in the evolution of these systems .
if the evolution of the disks in clusters is found to follow a different pace than that for more isolated stars , this can set important constraints on disk evolution theories .
clusters and groups in serpens as defined according to the criteria in @xcite .
the volume density contours are overlaid on the serpens extinction map @xcite .
the red dots are the ysos in serpens from @xcite .
black contours indicate volume densities of 0.125 , 0.25 , 0.5 , 2.0 and 4.0 m@xmath8pc@xmath54 , blue contour corresponds to volume density of 1 m@xmath8pc@xmath54 and yellow contours to a volume density of 25 m@xmath8pc@xmath54.,scaledwidth=50.0% ] as discussed in @xcite , the brightest ysos in serpens appear to be concentrated in clusters , but a more extended young stellar population exists outside these clusters . for the determination of clusters and their boundaries ,
we follow the method developed by @xcite for ophiuchus and perseus . using volume density and number of ysos as criteria ,
the method consists of a nearest neighbor algorithm , and has as input the complete sample of ysos in serpens @xcite .
the associations are divided into loose ( volume density of 1 m@xmath8pc@xmath54 , blue contour in figure [ clusters ] ) and tight ( volume density of 25 m@xmath8pc@xmath54 , yellow contours in figure [ clusters ] ) .
furthermore , the associations are divided into clusters ( more than 35 members ) and groups ( less than 35 members ) .
the results of this method in serpens yields 2 clusters ( a and b ) and 3 groups ( c , d and e ) , that can be seen in figure [ clusters ] .
individual memberships are marked in table [ tsi ] . with the exception of one ,
virtually all ysos in serpens are within the blue contour , i.e. , all ysos belong to , at least , a loose association .
this differs , for instance , from ophiuchus and perseus @xcite , where some ysos are completely isolated . for number statistics of clustering in all c2d clouds ,
see @xcite . table [ tcluster ] lists the number of objects in each tight association .
it appears that the ratio of disk to embedded sources in a given cluster increases with distance from the densest part of the cluster , where group c is located ( see figure [ clusters ] ) . to compare the populations , all objects belonging to any of the tight associations
were grouped into the `` cluster population '' , for better number statistics .
objects not belonging to any cluster were grouped into the `` field population '' .
figure [ irscluster ] shows the comparison between these two populations for 3 quantities derived from the irs spectra . in the left panel ,
the flux ratio between 30 and 13 @xmath0 m ( @xmath19 ) is an indication of disk geometry .
the middle and right panels show the peak intensity of the silicate features at 10 and 20 @xmath0 m , respectively . for all 3 quantities , the two populations ( cluster and field ) are statistically indistinguishable .
a two sample kolmogorov - smirnov test ( ks - test ) was performed for each quantity and the results show that the null hypothesis that the two distributions come from the same parent population can not be rejected to any significance ( 18% , 13% and 79% for @xmath19 , @xmath27 and @xmath28 , respectively ) .
comparison between the clustered ( dot - dashed black line ) and field populations ( solid blue line ) of serpens , with data from the c2d irs program ( dashed red line , @xcite ) . in the left panel ,
the flux ratio between 30 and 13 @xmath0 m ( @xmath19 ) is an indication of the disk geometry .
the middle and right panels show the strength of the silicate features at 10 and 20 @xmath0 m , respectively .
, scaledwidth=90.0% ] differences between the cluster vs. field populations will be further investigated with ancillary data ( e.g. , relative stellar ages and masses ) and modeling ( e.g. , disk sizes ) . however , the irs spectra allow the conclusion that no significant differences are found for disk geometry or the grain size distribution in the upper layers of circumstellar disks in clustered compared with field stars .
the latter result only applies to the inner disk , as traced by silicate features .
the outer disk may still be different between cluster and field populations .
the young stars with disks observed by the c2d team with irs spectroscopy ( c2d irs sample ) are scattered across the sky in the 5 molecular clouds studied @xcite .
the other 4 clouds studied by the c2d are chamaeleon ii , lupus , perseus and ophiuchus .
all clouds are nearby ( within 300 pc ) , span a range of star - formation activity , have typical median ages of a few myr and have a spread between more ( perseus and ophiuchus ) and less ( cha ii and lupus ) clustered yso populations @xcite .
the results from this sample are also compared to the results in serpens in figures [ si ] and [ irscluster ] .
a conspicuous similarity is seen between the samples .
note that @xcite do not analyze the 20 @xmath0 m silicate feature in this same manner and , therefore , @xmath55 for the c2d irs sample is missing from the right panel in figure [ irscluster ] . the flux ratio between 30 and 13 @xmath0 m ( @xmath19 ) plotted against the peak at 10 @xmath0 m ( @xmath27 , black dots ) , compared against the t tauri feps sample ( red triangles , @xcite ) and the c2d irs sample ( blue stars , @xcite ) .
typical uncertainties for the 10 @xmath0 m peak strength are @xmath360.1 , while typical uncertainties for the flux ratio are @xmath360.12 .
, scaledwidth=60.0% ] the models of @xcite conclude that , if sedimentation is the unique reason for the variety of observed strength and shape of the 10 @xmath0 m silicate feature , then this feature is strong for weak excess in mid- to far - ir ( as probed by @xmath19 ) , and vice - versa
. however , when studying a small sample of t tauri stars from the spitzer legacy program `` the formation and evolution of planetary systems : placing our solar system in context '' ( feps ) sample , @xcite found the opposite : a trend in which weak @xmath19 correlates with a weak feature .
a confirmation of this trend for a larger sample implies that sedimentation alone can not be the sole cause for the diversity of observed silicate features .
@xcite argue that dust coagulation must then play a vital role in producing different silicate profiles . in figure [ feps ]
, we populate this diagram with the serpens sample ( black dots ) and the c2d irs sample ( blue stars ) , as well as the bouwman feps sample ( red triangles ) .
this large combined sample shows no correlation between the strength of the 10 @xmath0 m silicate feature and @xmath19 ( @xmath56 ) and therefore does not support either the correlation ( as seen by @xcite ) or the anti - correlation ( as modeled by @xcite for sedimentation alone ) between the ir flux excess and the strength of the 10 @xmath0 m silicate feature .
the taurus molecular cloud is the best characterized star - forming region to date , due to its proximity and relatively low extinction . with young stars and their surrounding disks studied for more than two decades ( e.g. , @xcite ) ,
its members have been well characterized at a wide range of wavelengths .
taurus has thus become the reference for comparison of star - forming regions . compared to serpens
, taurus seems to be a somewhat younger ( 2.0 myr median age , @xcite vs. 4.7 myr @xcite ) and has lower star forming rate .
due to uncertainties in pre - main sequence age determinations ( e.g. , @xcite ) , this difference may not be significant .
however , cluster ages are statistically more likely to be different than the same . in a campaign similar to that presented here , irs spectra were obtained for a sample of 139 ysos in taurus , as part of a larger irs guaranteed - time observing program .
these data were presented in two separate papers : @xcite treated the disk sources , while @xcite presented the embedded population in taurus . because the young stellar population in serpens has been discovered and characterized using ir observations
, this sample does not include young stars without disks ( and therefore no ir excess , also called class iii sources , as defined in @xcite ) .
for this reason , a comparison between the class iii sources of the two clouds does not make sense .
26 class iii objects were studied in taurus .
these objects present featureless spectra with very little ir excess at longer wavelengths .
it is , however , possible to compare the ir excess populations of both regions . out of the entire yso sample
studied in taurus , embedded sources ( or class i ) amount to 20% while in serpens they amount to 18% , a very comparable number .
this also matches the typical percentage of class i sources in the global yso statistics of the five clouds from the c2d photometric sample @xcite . in serpens only 5 objects
have featureless spectra . excluding the embedded sources to analyze the disk population alone , objects having both 10 and 20 @xmath0 m silicate features amount to 72% in taurus , comparable to 73% of the serpens disk population . also comparable is the percentage of disks with pah emission : 3% in serpens and 4% in taurus . in both regions , all disks with a 10 @xmath0 m silicate feature
also show the 20 @xmath0 m one .
however , a difference is found for disks with only the 20 @xmath0 m feature in emission : they amount to 17% of the disk population in serpens and only 4.5% in taurus .
these statistics are summarized in table [ tstats ] .
top : distribution of the 10 @xmath0 m peak strength for serpens ( black ) and taurus ( red ) .
bottom : strength and shape of the 10 @xmath0 m silicate feature for both serpens ( black ) and taurus ( red ) .
, scaledwidth=50.0% ] irs data on the 85 disks sources presented in @xcite were obtained from the spitzer archive and reduced in the same manner as our serpens data . the same method described in section [ ssi_ssr ] was applied to these data in order to compare the processes affecting the dust in both regions .
figure [ si_taurus ] ( bottom ) shows the strength versus shape of the 10 @xmath0 m silicate feature for both serpens ( black dots ) and taurus ( red dots ) , as well as the distribution of peak strength for both regions ( top ) .
serpens and taurus present remarkably similar distributions , with populations clustered around flatter features and bigger grains , and almost no strongly peaked silicate emission sources .
a ks - test shows that the hypothesis that the distributions are drawn from the same population can not be rejected ( 82% ) .
comparisons between disks around stars that were formed in clusters or in isolation in the serpens cloud , explored in [ scf ] , indicate very similar populations .
thus , statistically no differences are found in terms of both disk geometry ( probed by the ratio @xmath19 ) and processes affecting the dust ( probed by the silicate features ) .
this suggests that local environment does not affect the evolution path and timescale of disks .
even more remarkable is the agreement between the yso populations of serpens and taurus , as well as the c2d irs sample spread over 5 clouds , in terms of silicate features , as shown in
[ sc2d ] and [ stau ] .
even though there are clear differences in silicate features from source to source within a cloud , the overall _ distribution _ of feature shapes is statistically indistinguishable : in all three samples , each containing of the order of 100 disks , the bulk of the sources has a rather flat silicate profile , and a tail of peaked shapes .
if the difference in median ages is significant , the similarity seen in figure [ si_taurus ] indicates that a 2 - 3 myr difference in age is not reflected in a concurrent evolution of the average disk surface dust properties .
this indicates that the dust population in the disk surface is an equilibrium between dust growth and destruction processes , which is maintained at least as long as the disk is optically thick .
the process of grain growth through coagulation and settling to the midplane has been shown to be much too short to be consistent with disk observations @xcite .
that means that the small grain population must be replenished somehow , which could happen by fragmentation of bigger aggregates and turbulent mixing .
it is widely accepted in the literature that the 10 @xmath0 m silicate feature is representative of the dust in the surface layers of the disk at a few au from the star @xcite .
significant evolution could still take place in the disk mid - plane , which is not traced with these observations .
indeed , millimeter observations have indicated that disk midplanes are abundant in grains with mm / cm sizes @xcite .
that means that , if this population of small grains in the disk surface must be replenished by fragmentation of bigger grains followed by turbulent mixing , the surface dust is an indirect tracer of midplane grains .
if the age difference between serpens and taurus is significant , then there exists a statistical equilibrium of processes of dust coagulation and fragmentation in the disk lasting at least a few million years that is independent of which yso population is being studied .
the observable ir dust characteristics of a cluster population of protoplanetary disks does not appear to depend on cluster properties ( age , density ) within the first 5 myr , for relatively low - mass clusters .
no specific property or event that may influence the infrared dust signatures of a single disk ( stellar luminosity , spectral type , presence of a companion , disk - planet interactions , disk instabilities etc . ) produces detectable temporal or spatial evolutionary effects visible in the distribution of disk properties of an entire cluster within this time frame .
this is consistent with a picture in which a specific disk may change its appearance on short time scales ( much less than 1 myr ) , but in a _
reversible _ way .
that is , the properties of disk surfaces may oscillate between different states @xcite .
if some evolution of the disk surface is _ irreversible _ , and happens at a given rate , a stable distribution of surface properties would not be seen over time .
this requires that the effect on the disk surface of any reprocessing events has to be erased .
another option is that the disk surface properties are determined by a single parameter , such as the initial conditions of the formation of such disks . in this scenario those properties
should be kept `` frozen '' over the observed time scale of @xmath36few myr . however
, this can not be the case for objects like ex lup , where real time changes of the disk surface properties after the initial formation of the disk have been observed on time scales of just a few years ( namely crystallization through thermal annealing of the dust in the disk surface , @xcite ) .
for this theory to work , the putative oscillation of states proposed here must be stable over disk lifetimes of @xmath365 myr .
as long as the disks are gas - rich and optically thick , as in the three samples studied here , we do not see a distinguishable evolution
. it will be important to search for surface evolution indicators in even older clusters .
such comparisons will yield constraints on the global time scale of disk evolution , which in turn will help constrain the importance of the processes that play a vital role in the dissipation of the disks .
the scenario in which a considerable population of small particles is still present after bigger aggregates are formed and altered through multiple events is consistent with evidence from primitive chondrites in our own solar system .
chondritic meteorites are observed to contain fine - grained dust - like matrices formed after coarse - grained materials such as chondrules and calcium - aluminum - rich inclusions ( cais ) .
this indicates that the solar nebula underwent violent reprocessing events in the feeding zones of parent bodies at 3 - 4 au ( see @xcite for a recent review on solar system dust in an astrophysical context ) .
these events included systemic melting of large dust aggregates and possibly even evaporation and re - condensation of silicate grains .
in fact , presolar material is present in primitive chondrites only at the trace level : @xmath36100 ppm @xcite , testifying to the complete evaporative destruction of ism dust at a few au in protoplanetary dust .
we present spitzer / irs spectra from a complete and unbiased flux limited sample of ir excess sources found in the serpens molecular cloud , following the c2d mapping of this region . * among our total of 147 irs spectra
, 22% are found to be background contamination ( including background stars , redshifted galaxies and a planetary nebula candidate ) .
this high number is not surprising given the position of serpens , close to the galactic plane .
* excluding the background objects from the sample , the bona fide set of ysos amounts to 115 objects .
the embedded to disk source ratio is 18% , in agreement with the ratios derived from photometry @xcite for the five c2d clouds .
* disks with pah in emission amount to 2% of the yso population , but 3% of the disk population . only g and a star show pah emission . in the disk population ,
73% show both silicate emission features , at 10 and 20 @xmath0 m , while 17% show only the 20 @xmath0 m feature .
4% of the disks show featureless mid - ir spectra .
only one source , # 120 , shows both pah and silicate in emission . *
our yso population in serpens is very similar to that in taurus , also studied with irs spectra . in both regions about 70% of the disk sources present both 10 and 20
@xmath0 m silicate features in emission and the 10 @xmath0 m feature is never seen without the 20 @xmath0 m feature .
the only significant difference between the two populations is in the number of sources lacking the 10 @xmath0 m feature but showing the 20 @xmath0 m feature .
this may be indicative of small holes ( @xmath53 1 au ) in these sources . *
the silicate features in the irs spectra measure the grain sizes that dominate the mid - ir opacity .
the relationships between shape and strength of these features present distributions very similar to those obtained from other large samples of young stars .
comparison with the models of @xcite yield grains consistent with sizes larger than a few @xmath0 m . *
no significant differences are found in the disk geometry or the grain size distribution in the upper layers of circumstellar disks ( probed by the silicate features ) in clustered or field stars in the cloud , indicating that the local environment where a star is born does not influence the evolution of its harboring disk .
* quantitatively , the shape and strength of the 10 @xmath0 m silicate feature were used to compare both serpens and taurus as well as a large sample of disks across five clouds , indicating remarkably similar populations .
this implies that the dust population in the disk surface results from an equilibrium between dust growth and destruction processes that are maintained over a few million years .
support for this work , part of the spitzer space telescope legacy science program , was provided by nasa through contract numbers 1256316 , 1224608 and 1230780 issued by the jet propulsion laboratory , california institute of technology under nasa contract 1407 and by the spanish grant aya 2005 - 0954 .
astrochemistry at leiden is supported by a spinoza grant from the netherlands organization for scientific research ( nwo ) and by the netherlands research school for astronomy ( nova ) grants .
support to kmp was provided by nasa through hubble fellowship grant # 01201.01 , awarded by the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, for nasa , under contract nas 5 - 26555 . the authors are very grateful to l. allamandola , r. overzier , p. beiro , r. demarco , j. green , and a. kspal for fruitful discussions , and would like to acknowledge the anonymous referee for suggestions that have improved the manuscript considerably .
brahm , p. , et al .
2009 , , 459 , 224 acke , b. , & van den ancker , m. e. 2004 , , 426 , 151 alexander , r. d. , clarke , c. j. , & pringle , j. e. 2006 , , 369 , 216 alexander , r. d. 2008 , , 391 , l64 apai , d. , pascucci , i. , bouwman , j. , natta , a. , henning , t. , & dullemond , c. p. 2005
, science , 310 , 834 baraffe , i. , chabrier , g. , allard , f. , & hauschildt , p. h. 1998 , , 337 , 403 baraffe , i. , chabrier , g. , & gallardo , j. 2009 , , 702 , l27 bary , j. s. , leisenring , j. m. , & skrutskie , m. f. 2007 , bulletin of the american astronomical society , 38 , 1007 bernard - salas , j. , peeters , e. , sloan , g. c. , gutenkunst , s. , matsuura , m. , tielens , a. g. g. m. , zijlstra , a. a. , & houck , j. r. 2009 , , 699 , 1541 boogert , a. c. a. , et al . 2004 , , 154 , 359 boogert , a. c. a. , et al .
2008 , , 678 , 985 bottinelli , s. , et al . in preparation bouwman , j. , meeus , g. , de koter , a. , hony , s. , dominik , c. , & waters , l. b. f. m. 2001 , , 375 , 950 bouwman , j. , lawson , w. a. , dominik , c. , feigelson , e. d. , henning , t. , tielens , a. g. g. m. , & waters , l. b. f. m. 2006 , , 653 , l57 bouwman , j. , et al .
2008 , , 683 , 479 brauer , f. , dullemond , c. p. , johansen , a. , henning , t. , klahr , h. , & natta , a. 2007 , , 469 , 1169 brauer , f. , dullemond , c. p. , & henning , t. 2008 , , 480 , 859 brauer , f. , henning , t. , & dullemond , c. p. 2008
, , 487 , l1 brown , j. m. , et al .
2007 , , 664 , l107 chabrier , g. 2003 , , 115 , 763 chapman , n. l. , mundy , l. g. , lai , s .-
, & evans , n. j. 2009 , , 690 , 496 chapman , n. l. , & mundy , l. g. 2009 , , 699 , 1866 chiar , j. e. , et al .
2007 , , 666 , l73 clarke , c. j. , gendrin , a. , & sotomayor , m. 2001 , , 328 , 485 condon , j. j. 1992 , , 30 , 575 condon , j. j. , cotton , w. d. , greisen , e. w. , yin , q. f. , perley , r. a. , taylor , g. b. , & broderick , j. j. 1998 , , 115 , 1693 djupvik , a. a. , andr , p. , bontemps , s. , motte , f. , olofsson , g. , glfalk , m. , & florn , h .-
2006 , , 458 , 789 dominik , c. , & tielens , a. g. g. m. 1997 , , 480 , 647 draine , b. t. 2003 , , 41 , 241 dullemond , c. p. , & dominik , c. 2004 , , 421 , 1075 dullemond , c. p. , & dominik , c. 2005 , , 434 , 971 dullemond , c. p. , henning , t. , visser , r. , geers , v. c. , van dishoeck , e. f. , & pontoppidan , k. m. 2007 , , 473 , 457 dullemond , c. p. , & dominik , c. 2008 , , 487 , 205 eiroa , c. , & casali , m. m. 1992 , , 262 , 468 eiroa , c. , torrelles , j. m. , curiel , s. , & djupvik , a. a. 2005 , , 130 , 643 evans , n. j. , ii , et al . 2003 , , 115 , 965 evans , n. j. , ii , et al .
2007 , delivery documentation evans , n. j. , et al .
2009 , , 181 , 321 fazio , g. g. , et al .
2004 , , 154 , 10 furlan , e. , et al .
2006 , , 165 , 568 furlan , e. , et al .
2008 , , 176 , 184 forrest , w. j. , et al .
2004 , , 154 , 443 geers , v. c. , et al .
2006 , , 459 , 545 geers , v. c. 2007 , ph.d .
thesis ghavamian , p. , raymond , j. c. , blair , w. p. , long , k. s. , tappe , a. , park , s. , & winkler , p. f. 2009 , , 696 , 1307 gorti , u. , & hollenbach , d. 2009 , , 690 , 1539 greene , t. p. , wilking , b. a. , andre , p. , young , e. t. , & lada , c. j. 1994 , , 434 , 614 haisch , k. e. , jr
. , lada , e. a. , & lada , c. j. 2001 , , 553 , l153 hartmann , l. , ballesteros - paredes , j. , & bergin , e. a. 2001 , , 562 , 852 harvey , p. m. , et al .
2006 , , 644 , 307 harvey , p. m. , et al .
2007 , , 663 , 1139 harvey , p. , mern , b. , huard , t. l. , rebull , l. m. , chapman , n. , evans , n. j. , ii , & myers , p. c. 2007 , , 663 , 1149 hillenbrand , l. a. 2008 , physica scripta volume t , 130 , 014024 hillenbrand , l. a. 2009 , iau symposium , 258 , 81 houck , j. r. , et al .
2004 , , 154 , 18 ireland , m. j. , & kraus , a. l. 2008 , , 678 , l59 johansen , a. , brauer , f. , dullemond , c. , klahr , h. , & henning , t. 2008 , , 486 , 597 jrgensen , j. k. , johnstone , d. , kirk , h. , myers , p. c. , allen , l. e. , & shirley , y. l. 2008 , , 683 , 822 kaas , a. a. , et al .
2004 , , 421 , 623 kenyon , s. j. , & hartmann , l. 1987 , , 323 , 714 kenyon , s. j. , & hartmann , l. 1995 , , 101 , 117 kessler - silacci , j. , et al . 2006 , , 639 , 275 kessler - silacci , j. e. , et al .
2007 , , 659 , 680 lada , c. j. 1987 , star forming regions , 115 , 1 lada , e. a. , & lada , c. j. 1995 , , 109 , 1682 lahuis , f. , kessler - silacci , j. e. , evans , n. j. , i. , et al .
2006 , c2d spectroscopy explanatory supplement , tech .
, pasadena : spitzer science center lodders , k. , & amari , s. 2005 , arxiv : astro - ph/0501430 lommen , d. , et al .
2007 , , 462 , 211 magnelli , b. , elbaz , d. , chary , r. r. , dickinson , m. , le borgne , d. , frayer , d. t. , & willmer , c. n. a. 2009 , , 496 , 57 mcclure , m. 2009 , , 693 , l81 meeus , g. , waters , l. b. f. m. , bouwman , j. , van den ancker , m. e. , waelkens , c. , & malfait , k. 2001 , , 365 , 476 meeus , g. , bouwman , j. , dominik , c. , waters , l. b. f. m. , & de koter , a. 2002 , , 392 , 1039 mern , b. , brown , j.m . , oliveira , i. , et al .
, submitted to mern , b. , et al .
2007 , , 661 , 361 muzerolle , j. , hillenbrand , l. , calvet , n. , briceo , c. , & hartmann , l. 2003 , , 592 , 266 muzerolle , j. , et al .
2009 , , 704 , l15 natta , a. , testi , l. , muzerolle , j. , randich , s. , comern , f. , & persi , p. 2004
, , 424 , 603 natta , a. , testi , l. , neri , r. , shepherd , d. s. , & wilner , d. j. 2004 , , 416 , 179 naylor , t. 2009 , , 1142 berg , k. i. , boogert , a. c. a. , pontoppidan , k. m. , blake , g. a. , evans , n. j. , lahuis , f. , & van dishoeck , e. f. 2008 , , 678 , 1032 oliveira , i. , et al .
2009 , , 691 , 672 olofsson , j. , et al .
2009 , , 507 , 327 papovich , c. , et al .
2004 , , 154 , 70 pascucci , i. , apai , d. , luhman , k. , henning , t. , bouwman , j. , meyer , m. r. , lahuis , f. , & natta , a. 2009 , , 696 , 143 prantzos , n. , & aubert , o. 1995 , , 302 , 69 pontoppidan , k. m. , dullemond , c. p. , van dishoeck , e. f. , blake , g. a. , boogert , a. c. a. , evans , n. j. , ii , kessler - silacci , j. e. , & lahuis , f. 2005 , , 622 , 463 pontoppidan , k. m. , et al .
2008 , , 678 , 1005 pontoppidan , k. m. , & brearly , a. j. 2010 , book by d. apai przygodda , f. , van boekel , r. , brahm , p. , melnikov , s. y. , waters , l. b. f. m. , & leinert , c. 2003 , , 412 , l43 quillen , a. c. , blackman , e. g. , frank , a. , & varnire , p. 2004
, , 612 , l137 richling , s. , & yorke , h. w. 1998 , , 340 , 508 richling , s. , & yorke , h. w. 2000 , , 539 , 258 rieke , g. h. , et al .
2004 , , 154 , 25 rieke , g. h. , alonso - herrero , a. , weiner , b. j. , prez - gonzlez , p. g. , blaylock , m. , donley , j. l. , & marcillac , d. 2009 , , 692 , 556 rodmann , j. , henning , t. , chandler , c. j. , mundy , l. g. , & wilner , d. j. 2006 , , 446 , 211 sanders , d. b. , & mirabel , i. f. 1996 , , 34 , 749 sandstrom , k. , bolatto , a. , stanimirovi , s. , smith , j. d. , simon , j. d. , & leroy , a. 2009 , the evolving ism in the milky way and nearby galaxies , sargent , b. a. , et al .
2009 , , 182 , 477 savitzky , a. , & golay , m. j. e. 1964 , analytical chemistry , 36 , 1627 siess , l. , dufour , e. , & forestini , m. 2000 , , 358 , 593 smith , j. d. t. , et al .
2007 , , 656 , 770 sterzik , m. f. , morfill , g. e. , & dubrulle , b. 1995 , , 224 , 567 straizys , v. , cernis , k. & bartasiute , s. 1996 , baltic astron . , 5 , 125 strom , s.w .
1992 , , 359 , 360 testi , l. , & sargent , a. i. 1998 , , 508 , l91 tielens , a. g. g. m. 2008 , , 46 , 289 uchida , k. i. , et al .
2004 , , 154 , 439 van boekel , r. , waters , l. b. f. m. , dominik , c. , bouwman , j. , de koter , a. , dullemond , c. p. , & paresce , f. 2003 , , 400 , l21 van boekel , r. , waters , l. b. f. m. , dominik , c. , dullemond , c. p. , tielens , a. g. g. m. , & de koter , a. 2004 , , 418 , 177 van boekel , r. , min , m. , waters , l. b. f. m. , de koter , a. , dominik , c. , van den ancker , m. e. , & bouwman , j. 2005 , , 437 , 189 vandenbussche , b. , dominik , c. , min , m. , van boekel , r. , waters , l. b. f. m. , meeus , g. , & de koter , a. 2004 , , 427 , 519 varnire , p. , blackman , e. g. , frank , a. , & quillen , a. c. 2006 , , 640 , 1110 visser , r. , geers , v. c. , dullemond , c. p. , augereau , j .- c . , pontoppidan , k. m. , & van dishoeck , e. f. 2007 , , 466 , 229 watson , d. m. , et al .
2009 , , 180 , 84 weidenschilling , s. j. 1980 , icarus , 44 , 172 weingartner , j. c. , & draine , b. t. 2001 , , 548 , 296 white , r. j. , & basri , g. 2003 , , 582 , 1109 winston , e. , et al .
2007 , , 669 , 493 winston , e. , et al .
2009 , , 137 , 4777 zhang , c. y. , laureijs , r. j. , & clark , f. o. 1988 , , 196 , 236 zinner , e. , et al .
2007 , , 71 , 4786 r r c c c r l c c 1 & 1 & 18275383 - 0002335 & @xmath571.42 & go3 & 17891072 & 2007 - 04 - 19 & 1.31 & disk + 2 & 2 & 18280503 + 0006591 & @xmath572.29 & go3 & 17883648 & 2007 - 04 - 24 & 0.35 & bg star + 3 & 3 & 18280845 - 0001064 & @xmath570.70 & go3 & 17883136 & 2007 - 04 - 24 & 1.52 & disk + 4 & 4 & 18281100 - 0001395 & @xmath572.13 & go3 & 17888256 & 2007 - 04 - 29 & 0.27 & bg star + 5 & & 18281315 + 0003128 & @xmath571.84 & c2d & 13210368 & 2005 - 05 - 20 & 0.27 & bg star + 6 & 5 & 18281350 - 0002491 & @xmath571.08 & go3 & 17883136 & 2007 - 04 - 24 & 1.71 & disk + 7 & 6 & 18281501 - 0002588 & @xmath570.05 & go3 & 17883136 & 2007 - 04 - 24 & 2.54 & disk + 8 & 7 & 18281519 - 0001407 & @xmath571.25 & go3 & 17889536 & 2007 - 04 - 30 & 2.43 & disk + 9 & 8 & 18281525 - 0002434 & @xmath570.85 & go3 & 17882368 & 2007 - 04 - 25 & 5.37 & cold disk + 10 & 9 & 18281629 - 0003164 & @xmath571.13 & go3 & 17884160 & 2007 - 04 - 19 & 1.83 & disk + 11 & & 18281757 - 0006474 & 0.65 & go3 & 17890560 & 2007 - 04 - 28 & 3.36 & galaxy + 12 & & 18281757 + 0016065 & 0.53 & go3 & 17888768 & 2007 - 05 - 05 & 4.72 & galaxy + 13 & 11 & 18281981 - 0001475 & @xmath570.89 & go3 & 17889792 & 2007 - 04 - 30 & 4.09 & disk + 14 & 13 & 18282143 + 0010411 & @xmath571.44 & go3 & 17888000 & 2007 - 04 - 30 & 1.38 & disk + 15 & 14 & 18282159 + 0000162 & @xmath570.99 & go3 & 17883136 & 2007 - 04 - 24 & 2.56 & disk + 16 & 15 & 18282432 + 0034545 & @xmath572.01 & go3 & 17887744 & 2007 - 04 - 24 & 0.32 & bg star + 17 & & 18282720 + 0044450 & 1.82 & go3 & 17884928 & 2007 - 04 - 29 & 11.04 & pn candidate + 18 & 16 & 18282738 - 0011499 & @xmath571.98 & go3 & 17883136 & 2007 - 04 - 24 & 0.32 & bg star + 19 & 17 & 18282741 + 0000239 & @xmath571.97 & go3 & 17887744 & 2007 - 04 - 24 & 0.43 & bg star + 20 & 18 & 18282849 + 0026500 & @xmath571.12 & go3 & 17890048 & 2007 - 04 - 29 & 1.13 & disk + 21 & 19 & 18282905 + 0027560 & @xmath571.06 & go3 & 17884928 & 2007 - 04 - 29 & 10.38 & cold disk + 22 & & 18283000 + 0020147 & @xmath571.60 & c2d & 13210624 & 2005 - 05 - 20 & 0.29 & bg star + 23 & & 18283736 + 0019276 & @xmath572.19 & go3 & 17887744 & 2007 - 04 - 24 & 0.27 & bg star + 24 & 20 & 18284025 + 0016173 & @xmath570.96 & go3 & 17890048 & 2007 - 04 - 29 & 1.20 & disk + 25 & 21 & 18284053 + 0022144 & @xmath571.91 & go3 & 17887744 & 2007 - 04 - 24 & 0.42 & bg star + 26 & 22 & 18284187 - 0003215 & 0.53 & go3 & 17883136 & 2007 - 04 - 24 & 3.39 & emb + 27 & 23 & 18284403 + 0053379 & 0.70 & go3 & 17882880 & 2007 - 04 - 28 & 3.37 & emb + 28 & 24 & 18284479 + 0051257 & 1.04 & go3 & 17885184 & 2007 - 04 - 24 & 10.18 & emb + 29 & 25 & 18284481 + 0048085 & @xmath571.01 & go3 & 17888512 & 2007 - 05 - 05 & 2.76 & disk + 30 & 27 & 18284497 + 0045239 & @xmath571.35 & go3 & 17886720 & 2007 - 04 - 24 & 0.77 & disk + 31 & 38 & 18284559 - 0007132 & @xmath570.77 & go3 & 17882368 & 2007 - 04 - 25 & 0.51 & disk + 32 & 29 & 18284614 + 0003016 & @xmath570.64 & go3 & 17889024 & 2007 - 04 - 25 & 0.66 & disk + 33 & & 18284632 - 0011103 & 0.53 & go3 & 17890560 & 2007 - 04 - 28 & 11.17 & galaxy + 34 & & 18284828 - 0005300 & @xmath572.43 & go3 & 17888256 & 2007 - 04 - 29 & 0.47 & bg star + 35 & & 18284938 - 0006046 & @xmath572.01 & c2d & 13210624 & 2005 - 05 - 20 & 0.28 & bg star + 36 & 32 & 18285020 + 0009497 & @xmath570.24 & c2d & 13461505 & 2006 - 04 - 31 & 1.98 & disk + 37 & 33 & 18285039 - 0012552 & @xmath571.40 & go3 & 17884160 & 2007 - 04 - 19 & 0.32 & bg star + 38 & & 18285060 + 0007540 & @xmath572.31 & c2d & 13460736 & 2005 - 09 - 09 & & disk + 39 & 34 & 18285122 + 0019271 & 0.45 & go3 & 17883904 & 2007 - 04 - 29 & 8.98 & emb + 40 & 36 & 18285249 + 0020260 & @xmath570.15 & go3 & 17883136 & 2007 - 04 - 24 & 1.46 & disk + 41 & 37 & 18285276 + 0028466 & 0.07 & go3 & 17887232 & 2007 - 04 - 27 & 9.50 & disk + 42 & 38 & 18285362 + 0019302 & @xmath570.64 & go3 & 17887488 & 2007 - 04 - 25 & 2.09 & disk + 43 & 39 & 18285395 + 0045530 & @xmath571.16 & go3 & 17886208 & 2007 - 09 - 30 & 1.96 & disk + 44 & 40 & 18285404 + 0029299 & 1.35 & c2d & 13461249 & 2006 - 04 - 20 & 18.21 & emb + 45 & 42 & 18285450 + 0029520 & 0.15 & c2d & 13460993 & 2006 - 04 - 19 & 35.49 & emb + 47 & 43 & 18285489 + 0018326 & 0.88 & go3 & 17883392 & 2007 - 04 - 29 & 7.58 & emb + 48 & 44 & 18285529 + 0020522 & @xmath570.29 & go3 & 17884160 & 2007 - 04 - 19 & 1.43 & disk + 49 & 45 & 18285580 + 0029444 & 1.81 & go3 & 17883392 & 2007 - 04 - 29 & 99.75 & emb + 50 & 46 & 18285660 + 0030080 & 1.84 & go3 & 17889792 & 2007 - 04 - 30 & 21.37 & emb + 51 & 47 & 18285719 + 0048359 & @xmath570.32 & go3 & 17888768 & 2007 - 05 - 05 & 11.27 & emb + 52 & 48 & 18285808 + 0017244 & @xmath572.11 & go3 & 17888000 & 2007 - 04 - 30 & 1.23 & disk + 53 & 49 & 18285860 + 0048594 & @xmath571.06 & go3 & 17887488 & 2007 - 04 - 25 & 1.55 & disk + 54 & 50 & 18285946 + 0030029 & @xmath570.59 & c2d & 13461249 & 2006 - 04 - 20 & 0.62 & disk + 55 & 51 & 18290025 + 0016580 & @xmath571.01 & go3 & 17884416 & 2007 - 04 - 30 & 1.50 & disk + 56 & 52 & 18290057 + 0045079 & @xmath570.94 & go3 & 17886208 & 2007 - 09 - 30 & 3.22 & disk + 57 & 53 & 18290082 + 0027467 & @xmath570.36 & go3 & 17885440 & 2007 - 04 - 24 & 3.41 & disk + 58 & 54 & 18290088 + 0029315 & @xmath570.37 & c2d & 13210112 & 2005 - 04 - 17 & 1.82 & disk + 59 & 55 & 18290107 + 0031451 & @xmath570.49 & c2d & 13461249 & 2006 - 04 - 20 & 1.14 & disk + 60 & 56 & 18290122 + 0029330 & @xmath570.39 & c2d & 13461505 & 2006 - 04 - 21 & 1.58 & disk + 61 & 58 & 18290175 + 0029465 & @xmath570.84 & c2d & 13461249 & 2006 - 04 - 20 & 2.30 & disk + 62 & 59 & 18290184 + 0029546 & @xmath570.77 & c2d & 13210112 & 2005 - 04 - 17 & 1.32 & disk + 63 & 60 & 18290210 + 0031210 & 0.29 & go3 & 17887232 & 2007 - 04 - 27 & 32.23 & emb + 64 & & 18290215 + 0029400 & @xmath570.86 & go3 & 17885184 & 2007 - 04 - 24 & 2.89 & disk + 65 & 61 & 18290286 + 0030089 & @xmath570.23 & c2d & 13460993 & 2006 - 04 - 21 & 2.43 & disk + 66 & 62 & 18290393 + 0020217 & @xmath570.82 & go3 & 17883136 & 2007 - 04 - 24 & 1.19 & disk + 67 & 63 & 18290436 + 0033237 & 0.95 & c2d & 13461505 & 2006 - 06 - 21 & 2.54 & emb + 68 & & 18290442 - 0020018 & 0.52 & go3 & 17890560 & 2007 - 04 - 28 & 5.61 & galaxy + 69 & 64 & 18290518 + 0038438 & @xmath571.24 & go3 & 17889792 & 2007 - 04 - 30 & 5.64 & cold disk + 70 & 65 & 18290575 + 0022325 & @xmath572.12 & go3 & 17886976 & 2007 - 04 - 29 & 1.76 & disk + 71 & 66 & 18290615 + 0019444 & @xmath571.52 & go3 & 17888512 & 2007 - 05 - 05 & 2.49 & disk + 72 & 67 & 18290620 + 0030430 & 1.69 & c2d & 13461249 & 2006 - 04 - 21 & 17.07 & emb + 73 & 68 & 18290680 + 0030340 & 1.67 & c2d & 13460993 & 2006 - 04 - 21 & 75.97 & emb + 74 & 69 & 18290699 + 0038377 & @xmath570.63 & go3 & 17884416 & 2007 - 04 - 30 & 2.90 & disk + 75 & 71 & 18290765 + 0052223 & @xmath571.01 & go3 & 17888512 & 2007 - 05 - 05 & 2.01 & disk + 76 & 72 & 18290775 + 0054037 & @xmath571.06 & go3 & 17888000 & 2007 - 04 - 30 & 1.34 & disk + 77 & 73 & 18290808 - 0007371 & @xmath572.22 & go3 & 17888256 & 2007 - 04 - 29 & 0.27 & bg star + 78 & & 18290843 - 0026187 & @xmath572.28 & go3 & 17885696 & 2007 - 04 - 28 & 0.31 & bg star + 79 & 75 & 18290910 + 0031300 & 0.24 & go3 & 17884928 & 2007 - 04 - 29 & 51.02 & emb + 80 & 76 & 18290956 + 0037016 & @xmath571.13 & go3 & 17882368 & 2007 - 04 - 25 & 0.51 & disk + 81 & 77 & 18290980 + 0034459 & @xmath570.97 & c2d & 13210624 & 2005 - 05 - 20 & 0.85 & disk + 82 & 78 & 18291148 + 0020387 & @xmath571.34 & go3 & 17887232 & 2007 - 04 - 27 & 11.90 & cold disk + 83 & 79 & 18291249 + 0018152 & @xmath571.06 & go3 & 17889536 & 2007 - 04 - 30 & 2.06 & disk + 84 & 81 & 18291407 + 0002589 & @xmath572.19 & go3 & 17885696 & 2007 - 04 - 28 & 0.30 & bg star + 85 & & 18291477 - 0004237 & @xmath572.29 & c2d & 13210112 & 2005 - 04 - 17 & 0.24 & bg star + 86 & 83 & 18291508 + 0052124 & @xmath571.89 & go3 & 17882880 & 2007 - 04 - 28 & 0.45 & disk + 87 & 84 & 18291513 + 0039378 & @xmath571.50 & go3 & 17889024 & 2007 - 04 - 25 & 1.01 & disk + 88 & 85 & 18291539 - 0012519 & @xmath571.72 & go3 & 17885952 & 2006 - 10 - 23 & 0.50 & disk + 89 & 86 & 18291557 + 0039119 & @xmath570.56 & go3 & 17884416 & 2007 - 04 - 30 & 2.51 & disk + 90 & 87 & 18291563 + 0039228 & @xmath570.89 & go3 & 17885184 & 2007 - 04 - 24 & 6.52 & cold disk + 91 & 88 & 18291617 + 0018227 & 0.45 & c2d & 13210112 & 2005 - 04 - 17 & 2.11 & emb + 92 & 89 & 18291969 + 0018031 & @xmath571.23 & go3 & 17885440 & 2007 - 04 - 24 & 2.10 & disk + 93 & 90 & 18292001 + 0024497 & @xmath572.04 & go3 & 17886720 & 2007 - 04 - 24 & 0.29 & bg star + 94 & & 18292050 + 0047080 & 1.29 & go3 & 17888768 & 2007 - 05 - 05 & 25.76 & disk + 95 & 93 & 18292094 + 0030345 & @xmath571.73 & go3 & 17882880 & 2007 - 04 - 28 & 0.40 & bg star + 96 & 94 & 18292184 + 0019386 & @xmath570.99 & go3 & 17886976 & 2007 - 04 - 29 & 1.54 & disk + 97 & & 18292250 + 0034118 & @xmath571.65 & go3 & 17889280 & 2007 - 04 - 24 & 0.29 & disk + 98 & & 18292253 + 0034176 & @xmath572.48 & go3 & 17889024 & 2007 - 04 - 25 & 1.34 & disk + 99 & 95 & 18292616 + 0020518 & @xmath572.26 & go3 & 17883648 & 2007 - 04 - 24 & 0.37 & bg star + 100 & 96 & 18292640 + 0030042 & @xmath570.98 & go3 & 17888768 & 2007 - 05 - 05 & 1.89 & disk + 101 & 97 & 18292679 + 0039497 & @xmath570.60 & go3 & 17885440 & 2007 - 04 - 24 & 4.13 & disk + 102 & 99 & 18292736 + 0038496 & 0.29 & go3 & 17886464 & 2007 - 04 - 29 & 3.95 & emb + 103 & 100 & 18292824 - 0022574 & @xmath570.41 & c2d & 13210368 & 2005 - 05 - 20 & 1.17 & disk + 104 & & 18292833 + 0049569 & @xmath570.48 & go3 & 17887488 & 2007 - 04 - 25 & 1.85 & disk + 105 & & 18292864 + 0042369 & @xmath570.97 & go3 & 17890048 & 2007 - 04 - 29 & 1.94 & disk + 106 & 102 & 18292927 + 0018000 & @xmath571.24 & go3 & 17889792 & 2007 - 04 - 30 & 1.15 & disk + 107 & & 18293056 + 0033377 & @xmath570.60 & go3 & 17889536 & 2007 - 04 - 30 & 1.46 & disk + 108 & 105 & 18293254 - 0013233 & @xmath572.58 & go3 & 17884672 & 2007 - 04 - 28 & 1.36 & bg star + 109 & 106 & 18293300 + 0040087 & @xmath571.35 & go3 & 17888512 & 2007 - 05 - 05 & 2.16 & disk + 110 & 107 & 18293319 + 0012122 & @xmath571.81 & go3 & 17883648 & 2007 - 04 - 24 & 0.34 & bg star + 111 & & 18293337 + 0050136 & @xmath570.26 & go3 & 17888768 & 2007 - 05 - 05 & 3.80 & disk + 112 & 109 & 18293381 + 0053118 & @xmath572.06 & go3 & 17886720 & 2007 - 04 - 24 & 0.29 & bg star + 113 & 110 & 18293561 + 0035038 & @xmath571.53 & go3 & 17883904 & 2007 - 04 - 29 & 5.76 & cold disk + 114 & 111 & 18293619 + 0042167 & @xmath570.92 & go3 & 17883904 & 2007 - 04 - 29 & 7.52 & cold disk + 115 & 114 & 18293672 + 0047579 & @xmath571.00 & go3 & 17886976 & 2007 - 04 - 29 & 3.19 & disk + 116 & 116 & 18293882 + 0044380 & @xmath571.23 & go3 & 17890048 & 2007 - 04 - 29 & 1.55 & disk + 117 & 118 & 18294020 + 0015131 & 0.87 & go3 & 17883392 & 2007 - 04 - 29 & 2.72 & disk + 118 & & 18294067 - 0007033 & @xmath572.53 & go3 & 17884672 & 2007 - 04 - 28 & 0.38 & bg star + 119 & 119 & 18294121 + 0049020 & @xmath571.39 & go3 & 17889280 & 2007 - 04 - 24 & 1.55 & disk + 120 & 123 & 18294168 + 0044270 & @xmath571.42 & go3 & 17884416 & 2007 - 04 - 30 & 3.13 & disk + 121 & & 18294301 - 0016083 & @xmath572.18 & go3 & 17890304 & 2006 - 10 - 25 & 0.33 & bg star + 122 & 126 & 18294410 + 0033561 & @xmath571.68 & go3 & 17886464 & 2007 - 04 - 29 & 10.92 & cold disk + 123 & 129 & 18294503 + 0035266 & @xmath571.26 & go3 & 17886976 & 2007 - 04 - 29 & 3.71 & disk + 124 & 132 & 18294725 + 0039556 & @xmath571.32 & go3 & 17887232 & 2007 - 04 - 27 & 2.22 & disk + 125 & 133 & 18294726 + 0032230 & @xmath571.31 & go3 & 17886976 & 2007 - 04 - 29 & 1.50 & disk + 126 & 140 & 18294962 + 0050528 & @xmath572.00 & go3 & 17890816 & 2007 - 04 - 24 & 0.35 & bg star + 127 & 143 & 18295001 + 0051015 & @xmath571.68 & go3 & 17889280 & 2007 - 04 - 24 & 1.23 & disk + 128 & & 18295012 + 0027229 & 0.14 & go3 & 17888768 & 2007 - 05 - 05 & 2.60 & galaxy + 129 & 144 & 18295016 + 0056081 & @xmath570.76 & go3 & 17886208 & 2007 - 09 - 30 & 1.86 & disk + 130 & 145 & 18295041 + 0043437 & @xmath571.32 & go3 & 17886720 & 2007 - 04 - 24 & 2.14 & disk + 131 & 148 & 18295130 + 0027479 & @xmath572.42 & go3 & 17889024 & 2007 - 04 - 25 & 2.09 & disk + 132 & 149 & 18295206 + 0036436 & @xmath570.06 & go3 & 17886464 & 2007 - 04 - 29 & 4.04 & emb + 133 & & 18295240 + 0035527 & @xmath570.33 & go3 & 17884416 & 2007 - 04 - 30 & 3.07 & emb + 134 & 153 & 18295244 + 0031496 & @xmath570.59 & go3 & 17889792 & 2007 - 04 - 30 & 1.48 & disk + 135 & 154 & 18295252 + 0036116 & 0.78 & go3 & 17883392 & 2007 - 04 - 29 & 24.39 & emb + 136 & 156 & 18295304 + 0040105 & @xmath571.16 & go3 & 17889536 & 2007 - 04 - 30 & 1.50 & disk + 137 & 157 & 18295305 + 0036065 & 0.07 & go3 & 17882624 & 2006 - 10 - 23 & 10.57 & disk + 138 & 161 & 18295322 + 0033129 & @xmath571.99 & go3 & 17886720 & 2007 - 04 - 24 & 2.35 & bg star + 139 & 165 & 18295422 + 0045076 & @xmath572.35 & go3 & 17889024 & 2007 - 04 - 25 & 1.03 & disk + 140 & 166 & 18295435 + 0036015 & @xmath570.19 & go3 & 17887232 & 2007 - 04 - 27 & 9.65 & emb + 142 & 172 & 18295592 + 0040150 & @xmath570.85 & go3 & 17888000 & 2007 - 04 - 30 & 1.21 & disk + 143 & 173 & 18295620 + 0033391 & @xmath570.54 & go3 & 17889024 & 2007 - 04 - 25 & 1.21 & disk + 144 & 177 & 18295701 + 0033001 & @xmath571.15 & go3 & 17882880 & 2007 - 04 - 28 & 0.96 & disk + 145 & 178 & 18295714 + 0033185 & @xmath570.82 & go3 & 17883648 & 2007 - 04 - 24 & 1.26 & disk + 146 & 182 & 18295772 + 0114057 & @xmath570.44 & c2d & 9828352 & 2004 - 09 - 02 & 1.41 & disk + 147 & 189 & 18295872 + 0036205 & @xmath571.11 & go3 & 17888000 & 2007 - 04 - 30 & 0.72 & disk + 148 & 206 & 18300178 + 0032162 & @xmath571.25 & go3 & 17886976 & 2007 - 04 - 29 & 1.90 & disk + 149 & 210 & 18300350 + 0023450 & @xmath571.22 & go3 & 17888000 & 2007 - 04 - 30 & 1.81 & disk + r c c c c c c c c c c 1 & & & 2.53 & 2.13 & 2.59 & 2.00 & 1.52 & 1.96 & & k2 + 3 & & & 1.49 & 1.35 & 1.50 & 1.33 & 1.23 & 1.26 & d & m0 + 6 & & & 2.12 & 1.73 & 2.13 & 1.61 & 1.33 & 1.55 & d & k5 + 7 & & & 1.38 & 1.25 & 1.38 & 1.39 & 1.34 & 1.36 & d & m0 + 8 & & & 1.15 & 1.43 & 1.14 & 1.21 & 1.45 & 0.82 & d & + 9 & & & 1.64 & 1.48 & 1.68 & 2.24 & 1.97 & 2.27 & d & + 10 & & & 1.25 & 1.29 & 1.28 & 1.47 & 1.29 & 1.51 & d & + 13 & & & & & & 1.33 & 1.87 & 1.29 & d & + 14 & & & 1.45 & 1.38 & 1.45 & 1.65 & 1.43 & 1.53 & & m2 + 15 & & & 2.67 & 2.19 & 2.68 & 2.59 & 2.07 & 2.46 & d & + 20 & & & & & & 0.90 & 0.85 & 1.13 & & + 21 & & & 1.09 & 1.17 & 1.10 & 1.45 & 1.57 & 1.34 & & + 24 & & & & & & 1.69 & 1.29 & 1.30 & & + 26 & y & & & & & & & & & + 27 & y & & & & & & & & & + 28 & y & & & & & & & & & + 29 & & & 1.84 & 1.59 & 1.89 & 1.56 & 2.03 & 1.73 & & m2 + 30 & & & 1.32 & 1.38 & 1.34 & 1.22 & 1.18 & 1.33 & & m1 + 31 & & & 1.76 & 1.72 & 1.76 & 0.99 & 0.92 & 0.99 & & + 32 & & & & & & 0.89 & 1.13 & 0.72 & & + 36 & & & 1.47 & 1.35 & 1.48 & 1.41 & 1.31 & 1.41 & & k5 + 38 & & & & & & & & & & + 39 & y & & & & & & & & e & + 40 & & & 1.06 & 1.20 & 1.09 & 1.29 & 1.26 & 1.25 & e & m7 + 41 & & & 1.73 & 1.77 & 1.93 & 3.45 & 4.18 & 3.74 & c & k2 + 42 & & & & & & & & & e & + 43 & & & 1.32 & 1.40 & 1.35 & 1.18 & 1.18 & 1.28 & & m0.5 + 44 & y & & & & & & & & c & m6 + 45 & y & & & & & & & & c & + 47 & y & & & & & & & & e & m5 + 48 & & & 1.22 & 1.21 & 1.26 & 1.12 & 1.05 & 1.18 & e & m5.5 + 49 & y & & & & & & & & c & + 50 & y & & & & & & & & c & + 51 & y & & & & & & & & & + 52 & & y & & & & & & & e & g3 + 53 & & & 1.19 & 1.32 & 1.25 & 1.26 & 1.04 & 1.18 & & m2.5 + 54 & & & 1.13 & 1.26 & 1.21 & 1.19 & 1.47 & 1.07 & c & + 55 & & & 1.88 & 1.68 & 1.89 & 1.53 & 1.18 & 1.35 & e & k2 + 56 & & & 1.56 & 1.49 & 1.55 & 1.32 & 1.23 & 1.41 & & + 57 & & & 1.39 & 1.47 & 1.43 & 2.01 & 1.81 & 1.90 & c & + 58 & & & 1.09 & 1.15 & 1.16 & 1.13 & 1.20 & 1.33 & c & k7 + 59 & & & & & & & & & c & + 60 & & & 1.00 & 1.08 & 1.01 & 1.27 & 1.25 & 1.26 & c & m0.5 + 61 & & & 1.24 & 1.25 & 1.25 & 1.28 & 1.32 & 1.26 & c & m0 + 62 & & & & & & 1.21 & 1.31 & 1.26 & c & k0 + 63 & y & & & & & & & & c & + 64 & & & & & & 1.41 & 1.27 & 1.36 & c & + 65 & & & 1.00 & 1.10 & 1.00 & 1.21 & 1.12 & 1.06 & c & + 66 & & & 1.19 & 1.18 & 1.20 & 1.15 & 1.07 & 1.11 & e & k5 + 67 & y & & & & & & & & c & + 69 & & & & & & 2.14 & 2.17 & 2.27 & b & + 70 & & & & & & 1.29 & 1.36 & 1.31 & e & a3 + 71 & & & 1.50 & 1.49 & 1.48 & 2.48 & 2.86 & 2.51 & e & m3 + 72 & y & & & & & & & & c & + 73 & y & & & & & & & & c & + 74 & & & 0.99 & 1.24 & 1.03 & 1.44 & 1.33 & 1.55 & b & + 75 & & & 1.37 & 1.37 & 1.38 & 1.64 & 1.35 & 1.50 & & + 76 & & & 1.58 & 1.51 & 1.61 & 1.83 & 1.78 & 1.75 & & m1 + 79 & y & & & & & & & & c & + 80 & & & 1.27 & 1.39 & 1.29 & 0.96 & 0.92 & 0.98 & b & + 81 & & & 1.59 & 1.52 & 1.66 & 1.50 & 1.28 & 1.42 & & m5 + 82 & & & 1.31 & 1.23 & 1.27 & 1.14 & 0.90 & 1.15 & e & m0 + 83 & & & & & & & & & e & + 86 & & & 1.80 & 1.61 & 1.83 & 1.28 & 0.98 & 1.23 & & m5.5 + 87 & & & & & & & & & b & m4 + 88 & & & 2.58 & 2.75 & 2.53 & 1.70 & 1.41 & 1.47 & & m0.5 + 89 & & & 1.10 & 1.21 & 1.08 & 1.62 & 1.68 & 1.71 & b & k5 + 90 & & & 1.19 & 1.19 & 1.21 & 1.54 & 1.62 & 1.67 & b & + 91 & y & & & & & & & & e & k7 + 92 & & & 1.56 & 1.60 & 1.63 & 1.92 & 1.66 & 1.97 & e & m0 + 94 & & & 1.14 & 0.85 & 0.99 & 0.46 & 0.53 & 0.54 & & + 96 & & & 1.35 & 1.23 & 1.34 & 1.53 & 1.48 & 1.41 & e & m1 + 97 & & y & & & & & & & & m2 + 98 & & y & & & & & & & & a3 + 100 & & & 1.39 & 1.47 & 1.39 & 1.90 & 1.56 & 1.89 & & + 101 & & & 1.60 & 1.55 & 1.63 & 1.95 & 1.65 & 1.92 & b & + 102 & y & & & & & & & & b & + 103 & & & 1.27 & 1.53 & 1.31 & 1.30 & 1.28 & 1.16 & & + 104 & & & 1.50 & 1.67 & 1.53 & 1.96 & 1.70 & 1.90 & & + 105 & & & 0.87 & 0.87 & 0.80 & 3.84 & 3.15 & 3.30 & & + 106 & & & 1.19 & 1.28 & 1.19 & 1.22 & 1.42 & 1.61 & & m3 + 107 & & & 1.07 & 1.04 & 1.06 & 1.24 & 1.16 & 1.37 & & + 109 & & & & & & 1.58 & 1.31 & 1.51 & b & + 111 & & & 1.24 & 1.32 & 1.29 & 0.61 & 1.11 & 1.23 & & + 113 & & & 1.10 & 1.25 & 1.12 & 1.76 & 1.82 & 1.91 & & k7 + 114 & & & 1.14 & 1.21 & 1.16 & 1.90 & 1.99 & 2.05 & b & f9 + 115 & & & 1.43 & 1.36 & 1.45 & 2.22 & 2.06 & 2.11 & & m0.5 + 116 & & & & & & 1.56 & 1.51 & 1.49 & b & + 117 & & & 1.37 & 1.38 & 1.40 & 1.87 & 1.68 & 1.81 & & k2 + 119 & & & 1.20 & 1.23 & 1.25 & 1.22 & 0.82 & 1.05 & b & k7 + 120 & & y & & & & 1.70 & 1.52 & 1.64 & b & a2 + 122 & & & 1.22 & 1.17 & 1.21 & 0.90 & 1.08 & 1.09 & b & m0 + 123 & & & 1.87 & 1.70 & 1.90 & 1.97 & 1.63 & 1.95 & b & m0 + 124 & & & 1.09 & 1.15 & 1.13 & 1.06 & 1.15 & 1.08 & b & m0 + 125 & & & 1.57 & 1.50 & 1.61 & 2.06 & 1.86 & 2.02 & b & m0 + 127 & & & & & & 1.85 & 1.71 & 2.00 & b & m2 + 129 & & & 1.19 & 1.16 & 1.20 & 1.11 & 0.95 & 1.15 & & + 130 & & & & & & 1.39 & 1.24 & 1.23 & & k6 + 131 & & y & & & & & & & & a3 + 132 & y & & & & & & & & b & + 133 & y & & & & & & & & b & + 134 & & & 0.76 & 0.88 & 0.73 & 1.25 & 1.33 & 0.89 & b & + 135 & y & & & & & & & & b & + 136 & & & 1.08 & 0.98 & 1.07 & 1.30 & 1.38 & 1.54 & b & + 137 & & & 1.27 & 1.24 & 1.28 & 1.99 & 2.06 & 2.01 & b & + 139 & & & & & & 1.88 & 1.50 & 1.76 & & a4 + 140 & y & & & & & & & & b & + 142 & & & 1.50 & 1.39 & 1.49 & 1.26 & 1.13 & 1.16 & b & m4 + 143 & & & & & & 1.42 & 1.26 & 1.47 & b & + 144 & & & 0.98 & 1.11 & 0.99 & 1.30 & 1.23 & 1.32 & b & + 145 & & & & & & 1.27 & 1.26 & 1.32 & b & g2.5 + 146 & & & 1.23 & 1.39 & 1.25 & 1.28 & 1.15 & 1.22 & a & m4 + 147 & & & 1.41 & 1.51 & 1.46 & 1.40 & 1.23 & 1.42 & b & + 148 & & & 1.81 & 1.67 & 1.83 & 1.76 & 1.51 & 1.79 & b & k7 + 149 & & & 1.90 & 1.65 & 1.89 & 1.47 & 1.31 & 1.43 & & m0 + r c c c cluster a & 0 & 1 & 1 + cluster b & 5 & 26 & 31 + group c & 9 & 10 & 19 + group d & 0 & 8 & 8 + group e & 3 & 12 & 15 + .statistics of the yso populations in serpens and taurus [ cols="<,^,^,^,^,^ " , ] [ tstats ]
here the background contamination discussed in [ sbg ] is presented in 3 categories .
the irs spectra of bright background stars , when seen through a molecular cloud , show the silicate feature at 10 @xmath0 m in absorption and not in emission as for disk sources ( see [ ssi ] ) , on top of a falling infrared spectrum .
@xcite identified candidate background stars based on optical spectra for 78 objects in this sample . after determining spectral types , extinctions and luminosities ,
those objects were placed in a hertzsprung - russell diagram ( hrd ) and compared with 2 sets of isochrones and mass tracks ( @xcite , @xcite ) .
20 out of those 78 objects proved to be much too luminous to be at the distance of serpens ( @xmath58pc , @xcite ) .
those objects were then classified as background sources , typically asymptotic giant branch ( agb ) stars with dusty shells .
interestingly , 18 of the 20 objects classified as background in @xcite have irs spectra characteristic of background sources ; the other two objects , # 81 and 86 , show silicates in emission ( just like those seen in circumstellar disks ) and were then re - classified as ysos .
besides these 18 objects , another 8 objects in this sample with falling mid - ir spectra show the 10 @xmath0 m silicate feature in absorption ( objects # 19 , 25 , 37 , 78 , 95 , 126 were not observed by @xcite ) , leading to their classification as background stars .
alternatively , those 8 objects could be highly extincted ysos , with silicate features in absorption arising from such extinction .
however , using the dense cloud dust properties , as in [ ssi_ext ] , an extinction of 10 mag would produce a @xmath59 of only 0.14 ( see also @xcite ) .
since these objects do have silicate features in absorption , it is understood that they have high extinction ( considerably higher than what is found for the cloud at their location ) .
thus , it is unlikely that they are actually highly extincted cloud members .
our working assumption is that these are background objects .
these 26 spectra are shown in figure [ irsbg ] .
irs spectra of the background stars toward serpens with the characteristic 10 @xmath0 m silicate absorption feature and falling sed . in grey
are the original spectra , while in black a binned version is overplotted .
the object # is indicated at the top right of each panel . ]
five of the sources with clear pah features show wavelength shifts in their line positions compared with galactic sources .
these features thus allow us to identify them as background galaxies in this sample .
pahs are common features in most nearby galaxies with ongoing or recent star - formation @xcite , and some contamination by background galaxies is expected @xcite . by identifying pah features from their shape and pattern ( @xcite and references therein ) , the observed wavelength ( @xmath60 )
can be related to the rest - frame wavelength ( @xmath61 ) of each feature through a redshift ( @xmath62 ) . between 3 ( # 12 ) and 5 ( # 11 and 68 ) pah features
are identified in each spectrum , and the redshift of each line is calculated .
for a given object , the @xmath63 values derived from the various lines agree within 2% and the mean of these values is taken as the redshift for that source .
these 5 redshifted galaxies can be seen , in rest - frame wavelength , in figure [ irsgal ] .
irs spectra of the background galaxies toward serpens , shown in rest - frame wavelength .
the dashed lines indicate the rest - frame pah features at 6.2 , 7.7 , 8.6 , 11.2 , 12.8 and 16.4 @xmath0 m .
the object # and @xmath63 are indicated in each panel.,scaledwidth=50.0% ] star - forming galaxies show strong correlations between the extinction - corrected paschen @xmath64 luminosity coming from hii regions and the luminosity at 24 @xmath0 m due to the emission of dust heated by hot , young stars .
@xcite obtained power - law fits to the relation between the star formation rate ( sfr ) and the observed 24 @xmath0 m flux densities ( @xmath65 ) , parameterized with respect to @xmath63 to take into account the fact that the rest - frame wavelengths sampled change with redshift ( eq . 14 in their paper )
in order to estimate the infrared luminosity ( @xmath66 ) , we use their best - fit relations between the ( rest - frame ) 24 @xmath0 m luminosity , @xmath66 and sfr ( eqs .
11 and 25 ) .
we have calculated sfrs and @xmath67(tir ) for the 5 background galaxies contaminating this sample .
these results can be found in table [ tgal ] .
@xmath66 ranges from a few times 10@xmath68 to a few times 10@xmath69 l@xmath8 , indicating that these galaxies are typical so - called luminous infrared galaxies ( lirgs , @xmath70 @xmath71 ) , while two are ultra - luminous ( ulirgs , @xmath72 @xmath71 ) .
the latter are a class of heavily obscured galaxies having enormous instantaneous sfrs that are often associated with merging galaxies @xcite .
these galaxies are very rare locally and become more common at high redshift .
their space density rises by about a factor of 10 from @xmath73 to @xmath74 @xcite , consistent with the relatively high redshifts found ( @xmath75 ) .
furthermore , due to its high sfr and the empirical correlation between the infrared and radio emission from starburst galaxies @xcite , the ulirg # 12 at @xmath76 can be identified with a previously unidentified radio source in the nrao vla sky survey ( nvss j182817 + 001604 , @xcite ) . the inferred sfrs range between 11 and 731 m@xmath8 yr@xmath77 , compared to a milky way sfr of about 35 m@xmath8 yr@xmath77 ( @xcite , and references therein ) . for further reference ,
the sfr of serpens is 5.7@xmath78 m@xmath8 yr@xmath77 @xcite .
even though the star formation observed in those galaxies is due to high mass star - forming regions such as orion , the high sfrs in the lirgs and ulirgs found are equivalent to @xmath79 molecular clouds such as serpens
! one of the objects , # 17 ( sstc2dj18282720 + 0044450 ) , in this sample presents very strong emission lines that do not seem compatible with ysos .
figure [ irspn ] shows a blow - up of the spectrum , identifying the high ionization lines at @xmath73 .
the source itself is highly reddened , being barely detected in irac1 ( 3.5 @xmath0 m ) but very prominent in mips1 ( 24 @xmath0 m ) , with a 324 @xmath0 m slope of 1.82 .
two types of galactic objects show such high excitation lines : dusty planetary nebulae and supernova remnants .
the irs spectrum of this object ( # 17 ) resembles those of shocked ejecta of supernova remnants , as in @xcite and @xcite and that of planetary nebulae such as lmc 78 , shown in @xcite .
r c c c c c c 11 & 0.1051 & 0.01800 & 16 & 1.5@xmath80 & lirg + 12 & 0.4637 & 0.00893 & 731 & 4.8@xmath81 & ulirg + 33 & 0.2742 & 0.01490 & 217 & 1.6@xmath81 & ulirg + 68 & 0.2032 & 0.00909 & 47 & 4.0@xmath80 & lirg + 128 & 0.1164 & 0.01010 & 11 & 1.1@xmath80 & lirg + | we present _ spitzer _ irs mid - infrared ( 535 @xmath0 m ) spectra of a complete flux - limited sample ( @xmath1 3 mjy at 8 @xmath0 m ) of young stellar object ( yso ) candidates selected on the basis of their infrared colors in the serpens molecular cloud .
spectra of 147 sources are presented and classified .
background stars ( with slope consistent with a reddened stellar spectrum and silicate features in absorption ) , galaxies ( with redshifted pah features ) and a planetary nebula ( with high ionization lines ) amount to 22% of contamination in this sample , leaving 115 true ysos .
sources with rising spectra and ice absorption features , classified as embedded stage i protostars , amount to 18% of the sample .
the remaining 82% ( 94 ) of the disk sources are analyzed in terms of spectral energy distribution shapes , pahs and silicate features .
the presence , strength and shape of these silicate features are used to infer disk properties for these systems .
about 8% of the disks have 30/13 @xmath0 m flux ratios consistent with cold disks with inner holes or gaps , and 3% of the disks show pah emission .
comparison with models indicates that dust grains in the surface of these disks have sizes of at least a few @xmath0 m .
the 20 @xmath0 m silicate feature is sometimes seen in absence of the 10 @xmath0 m feature , which may be indicative of very small holes in these disks .
no significant difference is found in the distribution of silicate feature shapes and strengths between sources in clusters and in the field . moreover ,
the results in serpens are compared with other well - studied samples : the c2d irs sample distributed over 5 clouds and a large sample of disks in the taurus star - forming region .
the remarkably similar distributions of silicate feature characteristics in samples with different environment and median ages if significant imply that the dust population in the disk surface results from an equilibrium between dust growth and destructive collision processes that are maintained over a few million years for any yso population irrespective of environment . |
by the definition given by j. eells and j.h .
sampson in @xcite , a map @xmath1 from a compact riemannian manifold @xmath2 to another riemannian manifold @xmath3 is harmonic if it is a critical point of the energy @xmath4 + from the first variation formula it follows that @xmath1 is harmonic if and only if its first tension field @xmath5 vanishes .
the same authors suggested a generalization of the notion of harmonicity : a map @xmath1 is biharmonic if it is a critical point of the bienergy functional @xmath6 + the first variation formula for @xmath7 , given by g.y .
jiang in @xcite and @xcite , is reobtained in @xcite and can be written as @xmath8 + where @xmath9 is a variation of @xmath1 , @xmath10 is the variational vector field along @xmath1 , and @xmath11 + here @xmath12 denotes the jacobi operator and @xmath13 , where @xmath14 represents the connection in @xmath15 .
hence the condition of biharmonicity for @xmath1 is @xmath16 + when @xmath17 is not compact , as a definition of biharmonicity one can adopt equation ( [ tensione2 ] ) instead of the definition given in terms of the integral formula of the bienergy . since any harmonic map is evidently biharmonic , we are interested in non - harmonic biharmonic maps . in general ,
the biharmonic equation is very complicated , and therefore the problem of finding examples of non - harmonic biharmonic maps between two riemannian manifolds , or examples of non - harmonic ( non - minimal ) biharmonic submanifolds of a given riemannian manifold , is difficult to solve .
still there are some results that now we mention .
* biharmonic curves on a surface in @xmath18 have been considered in @xcite . * in @xcite the authors gave the complete classification of non - minimal biharmonic submanifolds of @xmath19 .
these are certain circles , spherical helices and parallel spheres .
the general case of @xmath20 , for @xmath21 , is more interesting and has been treated in @xcite . * more recently , in @xcite , it has been shown how to construct non - harmonic biharmonic maps from @xmath17 to @xmath22 by deforming conformally the metric of @xmath17 . * when @xmath17 is a submanifold of the euclidean space @xmath23 , the biharmonicity condition seems to be very restrictive .
in fact , b.y .
chen conjectured in @xcite that any biharmonic submanifold of the euclidean space @xmath23 is minimal .
this was proved to be true in the case of submanifolds in @xmath18 ( in @xcite ) and in other special cases ( see @xcite , @xcite and @xcite ) . *
the same result was proved in @xcite for any @xmath24-dimensional riemannian manifold @xmath25 with constant negative sectional curvature @xmath26 . * other results about non - existence of non - harmonic biharmonic maps ,
when the sectional curvature or the ricci curvature of @xmath22 is non - positive , can be found in @xcite , @xcite and @xcite .
it seems then natural , as the next step , to consider biharmonic submanifolds in a @xmath24-manifold @xmath22 with non - constant sectional curvature .
we choose as @xmath22 the heisenberg group @xmath0 .
this nilpotent lie group is not - symmetric ; nevertheless it has many symmetries , in the sense that its isometry group has dimension four , the biggest possible for a 3-space of non - constant sectional curvature .
it is probably in virtue of these symmetries that some main problem turns out to be easier than expected . as , for instance , the problem of finding the equations of geodesics ( see @xcite and @xcite ) , and that of determining all surfaces with constant mean or gaussian curvature invariant with respect to some subgroup of the isometry group ( see @xcite , @xcite , @xcite and @xcite ) .
this manifold has many interesting properties .
we mention , for example , that there are no totally umbilical surfaces and therefore there are no totally geodesic surfaces in @xmath0 ( see @xcite ) . in this paper
we first write down the conditions that any non - harmonic ( non - geodesic ) biharmonic curve in @xmath0 must satisfy .
then we prove that the non - geodesic biharmonic curves in @xmath0 are helices .
a similar fact occurs in @xmath19 .
finally we deduce the explicit parametric equations of the non - geodesic biharmonic curves in @xmath0 . _ notation .
_ we shall work in @xmath27 category , i.e. manifolds , metrics , connections , maps will be assumed to be smooth .
the heisenberg group @xmath0 can be seen as the euclidean space @xmath18 endowed with the multiplication @xmath28 + and with the riemannian metric @xmath29 given by @xmath30 + the metric @xmath29 is invariant with respect to the left - translations corresponding to that multiplication .
this metric is isometric to the other , also quite standard , which is left - invariant with respect to the composition arising from the multiplication of the @xmath31 heisenberg matrices . at each point
the metric @xmath29 has an axial symmetry : the 4-dimensional group of its isometries contains the group of rotations around the @xmath32 axis ( in the classical terminology ( see @xcite ) a space with such a property is called _ systatic _ ) .
+ first of all we shall determine the levi - civita connection @xmath33 of the metric @xmath29 with respect to the left - invariant orthonormal basis @xmath34 + which is dual to the coframe + @xmath35 + we obtain @xmath36 + also , we have the well known heisenberg bracket relations @xmath37=e_3 , \quad [ e_3,e_1]=[e_2,e_3]=0.\ ] ] + we shall adopt the following notation and sign convention .
the curvature operator is given by + @xmath38}z,\ ] ] + while the riemann - christoffel tensor field and the ricci tensor field are given by @xmath39 + where @xmath40 are smooth vector fields on @xmath0 .
moreover we put , as in @xcite or @xcite , @xmath41 + where the indices @xmath42 take the values @xmath43 .
the non vanishing components of the above tensor fields are @xmath44 @xmath45 @xmath46 + respectively , and those obtained from these by means of the symmetries of @xmath47 .
thus the curvatures of @xmath0 have both positive and negative components .
to study the biharmonic curves in @xmath0 , we shall use their frenet vector fields and equations .
let @xmath48 be a differentiable curve parametrized by arc length and let @xmath49 be the orthonormal frame field tangent to @xmath0 along @xmath50 and defined as follows : by @xmath51 we denote the unit vector field @xmath52 tangent to @xmath50 , by @xmath53 the unit vector field in the direction of @xmath54 normal to @xmath50 , and we choose @xmath55 so that @xmath49 is a positive oriented orthonormal basis .
then we have the following frenet equations @xmath56 + where @xmath57 is the geodesic curvature of @xmath50 and @xmath58 its geodesic torsion . by making use of equations and of expression of the curvature tensor field ,
we obtain from the biharmonic equation for @xmath50 : @xmath59 + where @xmath60 , @xmath61 , and @xmath62 .
thus we have [ eq : theo_1 ] let @xmath48 be a differentiable curve parametrized by arc length .
then @xmath50 is a non - geodesic biharmonic curve if and only if @xmath63 by analogy with curves in @xmath18 , also following @xcite , we keep the name _ helix _ for a curve in a riemannian manifold having constant both geodesic curvature and geodesic torsion .
now , for any helix in @xmath0 , the system becomes @xmath64 + and therefore , in this case , @xmath65 must be constant , too .
thus biharmonic helices satisfy @xmath66 + we shall come back to biharmonic helices in the next section , after showing that all the biharmonic curves in @xmath0 are helices .
first we prove that for a biharmonic curve in @xmath0 the geodesic torsion must be constant .
[ gener - prop ] let @xmath48 be a non - geodesic curve parametrized by arc length .
if @xmath67 is constant and @xmath68 , then @xmath50 is not biharmonic . from @xmath69
we obtain @xmath70 ; then , if we put @xmath71 and @xmath72 , we get @xmath73 . hence we write @xmath74 + and then we use the first frenet equation , that gives @xmath75 + since @xmath76 , we have @xmath77 + now we replace @xmath78 in the above expression of @xmath79 , and we obtain + @xmath80 + as @xmath81 , we have @xmath82 then the second frenet equation gives @xmath83 + on the other hand we have @xmath84 + by comparing and we obtain @xmath85 + next we replace @xmath86 , @xmath87 and @xmath88 in , and we get @xmath89 + assume now that @xmath50 is biharmonic . then @xmath90 and we can write @xmath91 by substituting @xmath92 in and then by integrating we get @xmath93 + where @xmath94 is a constant . on the other hand , from the second equation in , we obtain @xmath95 . thus equation becomes @xmath96 where @xmath97 is a constant , and therefore also @xmath58 is constant , and we have a contradiction
. from theorem [ eq : theo_1 ] and proposition [ gener - prop ] we have , in conclusion , [ eq : biharmoniccurves ] let @xmath48 be a non - geodesic curve parametrized by arc length
. then @xmath50 is biharmonic if and only if @xmath98
now we want to determine all helices in @xmath0 that are biharmonic but non - geodesic . from theorem [ eq : biharmoniccurves ]
it is clear that , to this aim , we have to study the behaviour of @xmath92 and @xmath65 . for one thing , it follows from ( [ eq : biharmonichelix ] ) that @xmath65 must be constant .
we shall show that for a curve @xmath50 satisfying ( [ eq : biharmonichelix ] ) the constant @xmath65 can not vanish .
more precisely we prove [ 5.1 ] let @xmath48 be a non - geodesic curve parametrized by arc length .
if @xmath99 , then @xmath100 and @xmath50 is not biharmonic .
as @xmath50 is s parametrized by arc length , we can write @xmath101 where @xmath102 , @xmath103 . by using ( [ eq : conn_1 ] ) we first have @xmath104 + next we compute @xmath81 , and obtain @xmath105 assume now @xmath106 .
we exclude the case @xmath107 , that implies @xmath108 and therefore that @xmath50 is a geodesic .
thus we must have @xmath109 , and hence @xmath110 + without loss of generality , we can assume that @xmath111 ( when @xmath112 one has a geodesic ) .
then we have + @xmath113 and + @xmath114 + now we make use of the second frenet equation to obtain @xmath115 thus we have [ coroelica ] let @xmath48 be a non - geodesic biharmonic helix parametrized by arc length . then @xmath116
in this section we use the previous results to derive the explicit parametric equations of non - geodesic biharmonic curves in the heisenberg group @xmath0 .
+ we first prove the following in order to find the integral curves of the field given by ( [ tangelica ] ) that are biharmonic but non - geodesic , we must first determine the function @xmath126 .
after which , a simple integration will give the wanted parametric equations .
we shall prove [ eq : biharmonicintegralcurves ] the parametric equations of all non - geodesic biharmonic curves @xmath50 of @xmath0 are @xmath127 + where @xmath128 , @xmath129\cup[\arccos(-\frac{2\sqrt{5}}{5}),\pi)$ ] and @xmath130 .
without loss of generality , we can always assume that @xmath133 .
then we obtain + @xmath134 and @xmath135 + next we have @xmath136 + and @xmath137 + it follows that the geodesic torsion @xmath58 of @xmath50 is given by @xmath138 biharmonic curves can be obtained by intersecting the two surfaces @xmath149 and @xmath150 given by : @xmath151 and @xmath152 + the surface @xmath149 has constant non zero mean curvature ; it is the `` round cylinder '' with rulings parallel to the axis of revolution of @xmath0 at the point @xmath153 and as directrix the circle in the plane @xmath154 around this point ; this circle has constant ( geodesic ) curvature also in @xmath0 .
this cylinder has constant non zero mean curvature and zero gaussian curvature also in @xmath0 .
the surface @xmath150 is a `` helicoid '' which is minimal in the heisenberg group @xmath0 , as one can easily check by using the formulas given by bekkar in @xcite .
moreover , the biharmonic curves are geodesics of the cylinder and the cylinder is never a biharmonic surface .
( with regard to the study of biharmonic surfaces , a paper devoted to invariant non - minimal biharmonic surfaces in @xmath0 is in preparation . ) at each point @xmath157 the vectors tangent to biharmonic curves form a solid cone @xmath158 in @xmath159 . for each point @xmath160 and each vector @xmath161 ,
the only biharmonic curve @xmath50 arising from @xmath162 and such that @xmath163 is the geodesic determined by @xmath162 and @xmath164 .
thus , any @xmath165 is simultaneously tangent to a geodesic and to a non - geodesic biharmonic curve .
we visualize this fact in the following picture , where the geodesic is the curve not lying on the helicoid , while the interior edge is the biharmonic curve .
let s denote by @xmath50 the curve when @xmath166 , and by @xmath167 the curve when @xmath168 .
of course , both are helix in @xmath0 .
every biharmonic curve @xmath167 can be obtained from @xmath50 by means of a left - translation , i.e. @xmath169 , and we note that @xmath50 is a helix in @xmath18 , but @xmath167 is not .
the vector field @xmath51 tangent to a curve given by ( [ eq : biharmonichelices1 ] ) is transverse to the contact structure of @xmath0 ( for the contact geometry one can see , for example , @xcite ) determined by the @xmath170-form consider now a curve @xmath48 tangent to the contact structure ( such a curve is called a _ legendre curve _ ) , parametrized by arc lengh .
its velocity vector field @xmath174 has then the expression @xmath175 + it is not difficult to see that the vector fields @xmath51 and @xmath174 coincide for @xmath176 .
it follows that _ the legendre curves of the heisenberg group @xmath177 are biharmonic if and only if they are geodesic . _ the one - parameter subgroups @xmath178 are biharmonic if and only if they are geodesic .
in fact , if @xmath178 is not a geodesic , then @xmath67 and @xmath58 are always related by the formula ( see @xcite ) finally we note that the methods of this paper can be extended to study biharmonic curves in the cartan - vranceanu @xmath24-manifolds , namely the riemannian spaces ( @xmath180 ) , where the riemannian metrics @xmath181 are defined by @xmath182^{2 } } + \left(dz + \frac{l}{2 } \frac{ydx - xdy}{[1 + m(x^{2 } + y^{2})]}\right)^{2 } , \hspace{0.5 cm}l , m \in \r .\ ] ] + this two - parameter family of metrics reduces to the metric ( [ metricaheis ] ) for @xmath183 and @xmath184 ( see @xcite , @xcite,@xcite and @xcite for a discussion of these metrics and their properties ) . 1 . if @xmath189 , then @xmath190 , and we are in the euclidean space , where @xmath191 is biharmonic if and only if it is a line ; 2 .
if @xmath192 , the metric ( [ 5.35 ] ) has constant positive sectional curvature , and therefore the non - geodesic biharmonic curves are circles or spherical helices ( see @xcite ) .
p. baird , d. kamissoko , _ on costructing biharmonic maps and metrics _
, annals of global analysis and geometry , march 2003 , 23,1 , 65 - 75 . m. bekkar , _ exemples de surfaces minimales dans lespace de heisenberg _ , rend .
cagliari 61 ( 1991 ) , no .
2 , 123130
. l. bianchi , _ gruppi continui e finiti _
zanichelli , bologna , 1928 .
r. caddeo , s. montaldo , c. oniciuc , _ biharmonic submanifolds of @xmath19 _ , internat .
, 12 ( 2001 ) , no . 8 , 867 - 876 .
r. caddeo , s. montaldo , c. oniciuc , _ biharmonic submanifolds in spheres _ , israel j. math . , 130 ( 2002 ) , 109 - 123 .
r. caddeo , s. montaldo , p. piu , _ biharmonic curves on a surface _ , rendiconti di matematica e delle sue applicazioni , serie vii , volume 21 ( 2001 ) , 143 - 157 .
r. caddeo , s. montaldo , p. piu , _ on biharmonic maps _ , contemporary mathematics , volume 288 , ( 2001 ) , 286 - 290 .
r. caddeo , p. piu , a. ratto , _ rotational surfaces in @xmath193 with constant gauss curvature _ , bollettino u.m.i .
( 7 ) , ( 1996 ) , 9-b , 341 - 357 .
r. caddeo , p. piu , a. ratto , _
so(2)-invariant minimal and constant mean curvaturesurfaces in 3-dimensional homogeneous spaces _ , manuscripta math .
87 , ( 1995 ) , 1 - 12 .
cartan , _ leons sur la gomtrie des espaces de riemann _ , gauthier villars , paris ( 1946 ) . b.y
. chen , _ some open problems and conjectures on submanifolds of finite type _ , soochow j. math . 17 ( 1991 ) , 169 - 188 .
chen , s. ishikawa , _ biharmonic pseudo - riemannian submanifolds in pseudo - euclidean spaces _ , kyushu j. math .
52 , ( 1998 ) , 167 - 185 .
i. dimitric , _ submanifolds of @xmath194 with harmonic mean curvature vector _ , bull .
inst . math .
sinica 20 ( 1992 ) , 53 - 65 .
j. eells , l. lemaire , _ selected topics in harmonic maps _ , conf . board math .
j. eells , j.h .
sampson , _ harmonic mappings of riemannian manifolds _ , amer .
, 86 ( 1964 ) , 109 - 160 . c. figueroa , f. mercuri , r.h.l .
pedrosa , _ invariant surfaces of the heisenberg groups _ , ann .
pura e appl .
( 4 ) 177 ( 1999 ) , 173194 .
h. gluck , _ geodesics in the unit tangent bundle of a round sphere _ ,
( 2 ) 34 ( 1988 ) , no . 3 - 4 , 233246 . c. godbillon , _ gomtrie diffrentielle et mcanique analytique _ , hermann , paris ( 1969 ) .
2-harmonic isometric immersions between riemannian manifolds _ , chinese ann .
ser . a 7 ( 1986 ) , no 2 , 130 - 144 .
2-harmonic maps and their first and second variational formulas _ , chinese ann .
ser . a 7 ( 1986 ) , no 4 , 389 - 402 .
hangan , _ sur les distributions totalement godsiques du group nilpotent riemannien @xmath195 _ , rend . sem .
cagliari 55 ( 1985 ) , n. 1 , 31 - 37 .
t. hasanis , t. vlachos , _ hypersurfaces in @xmath196 with harmonic mean curvature vector field _ , math .
nachr . 172
( 1995 ) , 145 - 169 .
o. kowalski , _ additive volume invariants of riemannian manifolds _ , acta mathematica , 145 , ( 1980 ) , 205 - 225 . c. oniciuc ,
_ tangen i proprieti de armonicitate _ , tez de doctorat , univ .
i. cuza , iasi , ( 2002 ) . c. oniciuc , _ biharmonic maps between riemannian manifolds _ , an .
. i. cuza iasi , 48 ( 2002 ) , 237 - 248 .
p. piu , a. sanini , _ one - parameter subgroups and minimal surfaces in the heisenberg group _
, note di matematica 18 ( 1998 ) n.1 , 143 - 153 .
p. piu , _ sur certains types de distributions non - integrables totalement godsiques _ , thse de doctorat , univ .
de haute alsace , ( 1988 ) .
p. piu , _ sur les flots riemanniens des espaces de datri de dimension 3 _ , rend .
sem . mat .
torino 46 ( 1988 ) , 171 - 187 .
a. sanini , _ gauss map of a surface of the heisenberg group _ , bollettino u.m.i .
( 7 ) 11-b ( 1997 ) , suppl .
fasc . 2 , 79 - 93 .
tomter , _ constant mean curvature surfaces in the heisenberg group_. differential geometry : partial differential equations on manifolds ( los angeles , ca , 1990 ) , 485495 , proc . sympos .
pure math .
, 54 , part 1 , amer .
soc . , providence , ri , 1993 .
g. vranceanu , _ leons de gomtrie diffrentielle _
acad . rep .
, vol i , bucarest ( 1957 ) . | we consider the biharmonicity condition for maps between riemannian manifolds ( see @xcite ) , and study the non - geodesic biharmonic curves in the heisenberg group @xmath0 .
first we prove that all of them are helices , and then we obtain explicitly their parametric equations . |
about 100 years ago , bernstein @xcite introduced a concrete sequence of polynomials approximating a continuous function on a compact interval . that polynomials are dense in the set of continuous functions was shown by weierstrass @xcite , but bernstein was the first to give a concrete method , one that has withstood the test of time .
we refer to @xcite for a history of approximation theory , including _
inter alia _ historical references to weierstrass life and work and to the subsequent work of bernstein .
bernstein s approach was probabilistic and is nowadays included in numerous textbooks on probability theory , see , e.g. , @xcite or ( * ? ? ? * theorem 6.2 ) .
several years after bernstein s work , the nowadays known as wright - fisher stochastic model was introduced and proved to be a founding one for the area of quantitative genetics .
the work was done in the context of mendelian genetics by ronald a. fisher @xcite and sewall wright @xcite .
this paper aims to explain the relation between the wright - fisher model and the bernstein operator @xmath0 , that takes a function @xmath1 $ ] and outputs a degree-@xmath2 approximating polynomial .
bernstein s original proof was probabilistic .
it is thus natural to expect that subsequent properties of @xmath0 can also be explained via probability theory . in doing so , we shed new light to what happens when we apply the bernstein operator @xmath0 a large number of times @xmath3 to a function @xmath5 .
in fact , things become particularly interesting when @xmath3 and @xmath2 converge simultaneously to @xmath7 .
this convergence can be explained by means of the original wright - fisher model as well as a continuous - time approximation to it known as wright - fisher diffusion .
our paper was inspired by the monthly paper of abel and ivan @xcite that gives a short proof of the kelisky and rivlin theorem @xcite regarding the limit of the iterates of @xmath0 when @xmath2 is fixed .
we asked what is the underlying stochastic phenomenon that explains this convergence and found that it is the composition of independent copies of the empirical distribution function of @xmath2 i.i.d .
uniform random variables .
the composition turns out to be precisely the wright - fisher model .
being a markov chain with absorbing states , @xmath8 and @xmath9 , its distributional limit is a random variable that takes values in @xmath10 ; whence the kelisky and rivlin theorem @xcite .
composing stochastic processes is in line with the first author s current research interests @xcite . indeed
, such compositions often turn out to have interesting , nontrivial , limits @xcite .
stochastic compositions become particularly interesting when they explain some natural mathematical or physical principles .
this is what we do , in a particular case , in this paper . besides giving fresh proofs to some phenomena , stochastic compositions help find what questions to ask as well
. we will specifically provide probabilistic proofs for a number of results associated to the bernstein operator .
first , we briefly recall bernstein s probabilistic proof ( theorem [ berthm ] ) that says that @xmath11 converges uniformly to @xmath5 as the degree @xmath2 converges to infinity .
second , we look at iterates @xmath12 of @xmath0 , meaning that we compose @xmath0 @xmath3 times with itself and give a probabilistic proof of the kelisky and rivlin theorem stating that @xmath13 converges to @xmath14 as the number of iterations @xmath3 tends to infinity ( theorem [ krthm ] ) .
third , we exhibit , probabilistically , a geometric rate of convergence to the kelisky and rivlin theorem ( proposition [ krrate ] ) .
fourth , we examine the limit of @xmath13 when both @xmath2 and @xmath3 converge to infinity in a way that @xmath15 converges to a constant ( theorem [ jointlimsthm ] ) and show that probability theory gives us a way to prove and set up computation methods for the limit for `` simple '' functions @xmath5 such as polynomials ( proposition [ poly ] ) .
a crucial step is the so - called voronovskaya s theorem ( theorem [ vorthm ] ) which gives a rate of convergence to bernstein s theorem but also provides the reason why the wright - fisher model converges to the wright - fisher diffusion ; this is explained in section [ secwfdiff ] .
regarding notation , we let @xmath16 $ ] be the set of continuous functions @xmath17\to \r$ ] , and @xmath18 $ ] the set of functions having a continuous second derivative @xmath19 , including the boundary points , so @xmath20 ( respectively , @xmath21 ) is interpreted as derivative from the right ( respectively , left ) . for a bounded function @xmath22 \to \r$ ] , we denote by @xmath23 the quantity @xmath24 .
the bernstein operator @xmath0 maps any function @xmath22\to \r$ ] into the polynomial @xmath25 we are mostly interested in viewing @xmath0 as an operator on @xmath16 $ ] .
bernstein s theorem is : [ berthm ] if @xmath1 $ ] then @xmath11 converges uniformly to @xmath5 : @xmath26 the proof of this theorem is elementary if probability theory is used and goes like this .
let @xmath27 be independent bernoulli random variables with @xmath28 , @xmath29 for some @xmath30 .
if @xmath31 denotes the number of variables with value @xmath9 then @xmath31 has a binomial distribution : @xmath32 therefore @xmath33 now let @xmath34 since @xmath5 is continuous on the compact set @xmath35 $ ] it is also uniformly continuous and so @xmath36 as @xmath37 .
let @xmath38 be the event that @xmath39 and @xmath40 the indicator of @xmath38 ( a function that is @xmath9 on @xmath38 and @xmath8 on its complement ) .
we then write @xmath41 by chebyshev s inequality , @xmath42 therefore , @xmath43 letting @xmath44 the last term goes to @xmath8 and letting @xmath37 the first term vanishes too , thus establishing the theorem . a variant of bernstein s theorem due to marc kac @xcite gives better estimate if @xmath5 is lipschitz or , more generally , hlder continuous .
indeed , if @xmath5 satisfies @xmath45 for some @xmath46 then , @xmath47 where the first inequality used the hlder continuity of @xmath5 , while the second used jensen s inequality twice ; indeed , if @xmath48 is a positive random variable then @xmath49 , by the concavity of the function @xmath50 when @xmath51 .
hlder continuous functions with small @xmath52 are `` rough '' functions .
the previous remark tells us that we may not have a good rate of convergence for these functions . on the other hand ,
if @xmath5 is smooth can we expect a good rate of convergence ?
a simple calculation with @xmath53 shows that @xmath54 . excluding the trivial case @xmath55 ( the only functions @xmath5 for which @xmath56 ) , can the rate of convergence be better than @xmath57 for some smooth function @xmath5 ?
no , and this is due to voronovskaya s theorem ( theorem [ vorthm ] in section [ vsec ] ) .
[ pj ] @xmath58 + ( i ) @xmath0 is an increasing operator : if @xmath59 then @xmath60 .
( proof : @xmath11 is an expectation ; see . ) + ( ii ) if @xmath5 is a convex function then @xmath61 .
indeed , @xmath62 , by jensen s inequality , and , clearly , @xmath63 .
+ ( iii ) if @xmath5 is a convex function than @xmath11 is a also convex .
see lemma [ bconv ] in section [ isec ] for a proof .
let @xmath64 be the composition of @xmath0 with itself and , similarly , let @xmath65 ( @xmath3 times ) .
abel and ivan @xcite give a short proof of the following .
[ krthm ] for fixed @xmath66 , and any function @xmath17\to \r$ ] , @xmath67 note that this says that @xmath68 , as @xmath69 , uniformly in @xmath70 $ ] .
if @xmath5 is convex by remark [ pj](ii ) , @xmath71 . by remark [ pj](i ) , @xmath72 is increasing in @xmath3 and hence the limit of theorem [ krthm ] is actually achieved by an increasing sequence of convex functions ( remark [ pj](iii ) ) . to prepare the ground , we construct the earlier bernoulli random variables in a different way .
we take @xmath73 to be independent random variables , all uniformly distributed on the interval @xmath35 $ ] , and their _ empirical distribution function _
@xmath74 we shall think of @xmath75 as a random function .
note that , for each @xmath76 , @xmath77 has the binomial distribution .
the advantage of the current representation is that @xmath76 , instead of being a parameter of the probability distribution , is now an explicit parameter of the new random object @xmath78 .
we are allowed to ( and we will ) pick a sequence of independent copies @xmath79 of @xmath75 . for a positive integer @xmath3 let @xmath80 be the composition of the first @xmath3 random functions .
so @xmath81 is itself a random function . by using the independence and the definition of @xmath0 we have
that ( see also section a1 in the appendix ) @xmath82 for any function @xmath5 .
hence the limit over @xmath83 of the right - hand side is the expectation of the limit of the random variable @xmath84 , if this limit exists .
( we make use of the fact that is a finite sum ! ) to see that this is the case , we fix @xmath2 and @xmath76 and consider the sequence @xmath85 with values in @xmath86 we observe that this has the markov property , namely , @xmath87 is independent of @xmath88 , conditional on @xmath89 . by
, the one - step transition probability of this markov chain is @xmath90 since @xmath91 , states @xmath8 and @xmath9 are absorbing , whereas for any @xmath92 , we have @xmath93 for all @xmath94 .
define the absorption time @xmath95 elementary markov chain theory ( * ? ? ?
11 ) tells us that [ markovelem ] for all @xmath76 , @xmath96 .
therefore , with probability @xmath9 , we have that @xmath97 hence @xmath98 for all bur finitely many @xmath3 and so @xmath99 but the random variable @xmath100 takes two values : @xmath8 and @xmath9 .
notice that @xmath101 for all @xmath3 .
hence @xmath102 . but @xmath103 .
thus @xmath104 , and @xmath105 .
hence @xmath106 this proves the announced limit of theorem [ krthm ] but without the uniform convergence .
however , since all polynomials of the sequence are of degree at most @xmath2 , and @xmath2 is a fixed number , convergence for each @xmath76 implies convergence of the coefficients of the polynomials . to prove a rate of convergence in theorem [ krthm ] we need to show what was announced in remark [ pj](iii ) .
recall that @xmath107 .
[ bconv ] if @xmath5 is convex then so is @xmath11 .
let @xmath22\to \r$ ] and @xmath108 .
we shall prove that @xmath109 \ ] ] this can be done by direct computation using .
alternatively , we can give a probabilistic argument .
consider @xmath110 $ ] and compute first - order terms in @xmath111 . by
, @xmath112 is nonzero if and only if at least one of the @xmath113 s falls in the interval @xmath114 $ ] .
the probability that @xmath115 or more variables fall in this interval is @xmath116 , as @xmath37 .
hence , if @xmath117 is the event that _ exactly _ one of the variables falls in this interval , then @xmath118\,\p(s_n(x)=k , f_{\varepsilon})+o({\varepsilon}).\ ] ] if we let @xmath119 be the event that only @xmath120 is in @xmath121 , then @xmath122 is independent of @xmath123 , so @xmath124 .
so becomes @xmath125 + o({\varepsilon}),\ ] ] and , upon dividing by @xmath111 and letting @xmath37 , we obtain .
applying the same formula once more ( no further work is needed ) we obtain @xmath126.\ ] ] bring in now the assumption that @xmath5 is convex , whence @xmath127 , for all @xmath128 , and deduce that @xmath129 for all @xmath30 .
so @xmath11 is a convex function . [
cols="^,^ " , ] + _ convergence of the iterates of @xmath130 as @xmath83 + for a convex @xmath5 ( left ) and a nonconvex one ( right ) .
_ we can now exhibit a rate of convergence .
[ krrate ] for all @xmath30 , @xmath131 , @xmath132 moreover , @xmath133 and @xmath134 we have , for all positive integers @xmath135 and @xmath3 , @xmath136 \\ & = \e\left [ f(g_n^{k+\ell } { \raisebox{0.1ex}{\scriptsize $ \circ$}}\cdots{\raisebox{0.1ex}{\scriptsize $ \circ$}}g_n^{k+1}(h_n^k(x ) ) ) - f(h_n^k(x))\right ] \\ & = \sum_{y\in { \mathbb{i}}_n\setminus\{0,1\ } } \p(h_n^k(x)=y)\ , \e\left [ f(g_n^{k+\ell}{\raisebox{0.1ex}{\scriptsize $ \circ$}}\cdots{\raisebox{0.1ex}{\scriptsize $ \circ$}}g_n^{k+1}(y ) ) - f(y)\right],\end{aligned}\ ] ] whence @xmath137 letting @xmath138 and using theorem [ krthm ] yields .
to prove , we notice that @xmath139 , where @xmath140 if @xmath141 and @xmath9 otherwise , and , by , @xmath142 .
note that @xmath143 is a concave function . by ,
@xmath144 which is also concave by lemma [ bconv ] .
hence @xmath145 since , by symmetry , @xmath146 , inequality follows . for the final inequality ,
notice that @xmath147 \le n\ , \e \left[h_n^{k-1}(x ) \left(1-h_n^{k-1}(x)\right)\right ] , \ ] ] where we used the inequality @xmath148 , for all @xmath149 .
therefore , @xmath150 = : n\ , \gamma(k-1,x).\ ] ] using @xmath151 = \left(1-\frac{1}{n}\right)\ , x\ , ( 1-x),\ ] ] we obtain the recursion @xmath152 taking into account that @xmath153 we find that @xmath154 . combining and we get @xmath155 this should be compared with ( * ? ? ? * eq . ( 4 ) ) that says that @xmath156 , for some constant @xmath157 which has not been computed in @xcite , whereas we have an explicit constant @xmath158 .
now , the factor @xmath2 is probably wasteful and this comes from the fact that the inequality @xmath148 is not good when @xmath2 is large .
we only used it because of the simplicity of the right - hand side that enabled us to compute @xmath159 very easily .
we have a better inequality , namely , but to make it explicit one needs to compute @xmath160 .
we now take a closer look at the markov chain described by the sequence @xmath161 for fixed @xmath2 .
we repeat formula : @xmath162 we recognize that it describes the simplest stochastic model for reproduction in population genetics that goes as follows .
there is a population of @xmath163 individuals each of which carries 2 genes .
genes come in 2 variants , i and ii , say , thus , an individual may have 2 genes of type i both , or of type ii both , or one of each .
hence there are @xmath164 genes in total .
we observe the population at successive generations and assume that generations are non - overlapping .
suppose that the @xmath3-th generation consists of @xmath165 genes of type i and @xmath166 of type ii . in figure [ figgener ] below , type i genes are yellow , and type ii are red . to specify generation @xmath167
, we let each gene of generation @xmath167 select a single `` parent '' at random from the genes of the previous generation .
the gene adopts the type of its parent .
the parent selection is independent across genes .
the probability that a specific gene selects a parent of type i is @xmath168 . since we have @xmath2 independent trials , the probability that generation @xmath167 will contain @xmath123 genes of type i is given by the right - hand side of formula .
if we start the process at generation @xmath8 with genes of type i being chosen , independently , with probability @xmath76 each , then the number of alleles at the @xmath3-th generation has the distribution of @xmath169 .
this stochastic model we just described is known as the wright - fisher model , and is fundamental in mathematical biology for populations of fixed size .
the model is very far from reality , but has nevertheless been extensively studied and used .
early on , wright and fisher observed that performing exact computations with this model is hard .
they devised a continuous approximation observing that the probability @xmath170 as a function of @xmath76 , @xmath171 and @xmath3 can , when @xmath2 is large , be approximated by a smooth function of @xmath76 and @xmath171 .
( see , e.g. , kimura @xcite and the recent paper by tran , hofrichter , and jost @xcite . ) rather than approximating this probability , we follow modern methods of stochastic analysis in order to approximate the discrete stochastic process @xmath161 by a continuous - time continuous - space stochastic process that is nowadays known as wright - fisher diffusion .
our eventual goal is to understand what happens when we consider the limit of @xmath130 , when both @xmath3 and @xmath2 tend to infinity . from the bernstein and the kelisky - rivlin theorems
we should expect that the order at which limits over @xmath3 and @xmath2 are taken matters .
it turns out that the only way to obtain a limit is when the ratio @xmath15 tends to a constant , say , @xmath172 .
this is intimately connected to the wright - fisher diffusion that we introduce next .
we assume that the reader has some knowledge of stochastic calculus , including the it formula and stochastic differential equations driven by a brownian motion at the basic level of ksendal @xcite or at the more advanced level of bass @xcite .
we first explain why we expect that the markov chain @xmath173 , @xmath174 has a limit , in a certain sense , as @xmath44
. our explanation here will be informal .
we shall give rigorous proofs of only what we need in the following sections .
the first thing we do is to compute the expected variance of the increment of the chain , and examine whether it converges to zero and at which rate : see , theorem [ mc2diff ] , section a3 in the appendix .
the rate of convergence gives us the right time scale .
our case at hand is particularly simple because we have an exact formula : @xmath175 = \e\left[\left(g_n(y)-y\right)^2\right ] = \frac{1}{n } y ( 1-y).\ ] ] this suggests that the right time scale at which we should run the markov chain is such that the time steps are of size @xmath57 . in other words , consider the points @xmath176 and draw the random curve @xmath177}(x)\ ] ] ( where @xmath178 $ ] is the integer part of @xmath179 ) as in figure [ figcurve ] .
this is at the right time scale .
the second thing we do ( see , theorem [ mc2diff ] , section a3 ) is to compute the expected change of the markov chain . in our case , this is elementary : @xmath180 = \e\left[g_n(y)-y\right ] = 0.\ ] ] the functions @xmath181 and @xmath182 obtained in and suggest that the limit of the random curve @xmath183}(x ) , \
, t \ge 0)$ ] should be a diffusion process @xmath184 , @xmath185 , satisfying the stochastic differential equation @xmath186 with initial condition @xmath187 , where @xmath188 , @xmath189 , is a standard brownian motion .
it is actually possible to prove that @xmath183 } , t \ge 0)$ ] converges _ weakly _ to @xmath190 , but this requires an additional estimate on the size of the increments of the markov chain that is , in our case , provided by the following inequality : for any @xmath108 , @xmath191 to see this , apply hoeffding s inequality ( see section a2 in the appendix ) .
we thus have that , , are the conditions , and of theorem [ mc2diff ] , section a3 .
in addition , it can be shown that the stochastic differential equation admits a unique strong solution for any initial condition @xmath76 .
this is , e.g. , a consequence of the yamada - watanabe theorem ( * ? ? ?
* theorem 24.4 ) .
hence , by theorem [ mc2diff ] , the sequence of continuous random curves @xmath183}(x ) , t \ge 0)$ ] converges weakly to the continuous random function @xmath192 .
one particular conclusion of weak convergence is that @xmath193}(x ) ) \to \e f(x_t(x))$ ] for any @xmath1 $ ] or , equivalently , that [ jointlimsthm ] for any @xmath1 $ ] and any @xmath185 , @xmath194 } f(x ) = \e f(x_t(x)),\quad \text{uniformly in } x.\ ] ] since understanding the theorem of stroock and varadhan requires advanced machinery , we shall prove theorem [ jointlimsthm ] directly .
the proof is deferred until section [ jointsec ] . the nice thing with this theorem is that we have a way to compute the limit by means of stochastic calculus , the tools of which we shall assume as known .
let @xmath5 be a twice - continuously differentiable function . then , it s formula ( ( * ? ? ?
11 ) , ( * ? ? ?
* ch . 4 ) ) says that @xmath195 if we let @xmath196 take expectations in , and differentiate with respect to @xmath172 , we obtain @xmath197 the so - called forward equation of the diffusion .
now let , for all @xmath198 , @xmath199 noticing that @xmath200 is defined for all bounded and measurable @xmath201 and that @xmath202 .
if @xmath201 is such that @xmath203 then we can set @xmath204 in .
now , @xmath205 , because of the markov property of @xmath206 , @xmath185 , and so becomes @xmath207 letting @xmath208 , we arrive at the backward equation @xmath209 which is valid if @xmath200 is twice continuously differentiable .
the class of functions @xmath201 such that both @xmath201 and @xmath200 are in @xmath210 is nontrivial in our case .
it contains , at least polynomials .
this is what we show next .
it turns out that in order to prove theorem [ jointlimsthm ] we need to compute @xmath211 when @xmath5 is a polynomial .
[ poly ] for a positive integer @xmath212 , the following holds for the wright - fisher diffusion : @xmath213 where @xmath214 ( where , as usual , a product over an empty set equals @xmath9 ) .
write @xmath215 to save space . by it s formula applied to @xmath216 , @xmath217 since
the first integral is ( as a function of @xmath172 ) a martingale starting from @xmath8 its expectation is 0 .
thus , if we let @xmath218 we have @xmath219 thus , @xmath220 , as expected , and @xmath221 defining the laplace transform @xmath222 and using integration by parts to see that @xmath223 we have @xmath224 iterating this easy recursion yields @xmath225 where the second equality was obtained by partial fraction expansion ( and the notation is as in ) . since the inverse laplace transform of @xmath226 is @xmath227 ,
the claim follows .
formula was proved by kelisky and rivlin ( * ? ? ?
( 3.13 ) ) and karlin and ziegler ( * ? ? ?
* eq . ( 1.13 ) ) by entirely different methods .
( the latter paper contains a typo in the formula ) .
( 3.13 ) of @xcite reads : @xmath228 comparing this with and we obtain @xmath229 } = \frac{i}{r } { r\choose i}^2 \frac{\displaystyle ( -1)^{i+j } { r - i\choose j - i}^2 } { \displaystyle { 2\,j-2\choose j - i}{j+r-1 \choose r - j}},\ ] ] valid for any integers @xmath230 with @xmath231 . this equality can be verified directly by simple algebra .
an important result in the theory of approximation of continuous functions is voronovskaya s theorem @xcite .
it is the simplest example of _ saturation _ , namely that , for certain operators , convergence can not be too fast even for very smooth functions .
see devore and lorentz ( * ? ? ?
* theorem 3.1 ) .
voronovskaya s theorem gives a rate of convergence to bernstein s theorem . from a probabilistic point of view , the theorem is nothing else but the convergence of the generator of the discrete markov chain to the generator of the wright - fisher diffusion .
we shall not use anything from the theory of generators , but we shall give an independent probabilistic proof below for @xmath210 functions @xmath5 , including a slightly improved form under the assumption that @xmath19 is lipschitz . in this case ,
its lipschitz constant is @xmath232 recall that @xmath233 is defined by .
[ vorthm ] for any @xmath234 $ ] , @xmath235}| n ( b_n f(x ) - f(x ) ) - \ll f(x ) |=0.\ ] ] if moreover @xmath19 is lipschitz then , for any @xmath236 , @xmath237}| n ( b_n f(x ) - f(x ) ) -
\ll f(x ) |\le
\frac{\lip(f'')}{16 \cdot 3^{1/4 } } \ , n^{-1/2}\ ] ] using taylor s theorem with the remainder in integral form , @xmath238 since @xmath239 , we have , from , @xmath240 .
therefore , @xmath241 we estimate @xmath242 by splitting the expectation as @xmath243 + \e[j_n(x ) ; |g_n(x)-x| > \delta],\ ] ] where @xmath244 is chosen by the uniform continuity of @xmath19 : for @xmath108 let @xmath245 be such that @xmath246 whenever @xmath247 .
thus , @xmath248 and so @xmath249 on the other hand , with @xmath250 , we have @xmath251| \le 2 \|f''\| \p(|g_n(x)-x| > \delta ) \le 4 \|f''\| e^{-\delta n^2/2},\ ] ] by .
hence @xmath252 letting @xmath44 , and since @xmath253 is arbitrary , we obtain the first claim .
+ assume next that @xmath19 is lipschitz with lipschitz constant @xmath254 .
then @xmath255 we have @xmath256 , where @xmath31 is a binomial random variable
we can easily find ( or look in a textbook ) that @xmath257 since @xmath258 and , for @xmath259 , the function @xmath260 is concave in @xmath76 , it follows that @xmath261 . on the other hand , for @xmath262 , @xmath263 for all @xmath76 .
thus , the last term of is upper bounded by @xmath264 probabilistically , the operator @xmath265 , where @xmath266 is the identity operator , maps a function @xmath1 $ ] to the function @xmath267 $ ] which is the expected change of the function @xmath5 ( under the action of the chain ) per unit time , conditional on the current state being equal to @xmath171 .
since the natural time scale is counted in time units that are multiples of @xmath57 , we can interpret @xmath268 as the expected change of @xmath5 per unit of real time .
thus , its `` limit '' @xmath269 should play a similar role for the diffusion . and ,
indeed , it does , but we shall not use this here . for further information on diffusions and their generators see , e.g. , karlin and taylor @xcite .
the goal of this section is a probabilistic proof of theorem [ jointlimsthm ] .
first notice that it suffices to prove for polynomial functions . indeed ,
if @xmath5 is continuous on @xmath35 $ ] and @xmath253 , there is a polynomial @xmath270 such that @xmath271 ( by bernstein s theorem [ krthm ] ) .
but then @xmath272}_n f - f\| \le \|b_n^{[nt ] } f - b_n^{[nt ] } h\| + \|\sp_t f - \sp_t h\| + \|b_n^{[nt ] } h - \sp_t h\| \le 2{\varepsilon}+ \|b_n^{[nt ] } h - \sp_t h\|,\ ] ] where we used the fact that both @xmath273}$ ] and @xmath274 are defined via expectations , and so @xmath275 } f - b_n^{[nt ] } h\| = \|b_n^{[nt]}(f - h)\| \le \|f - h\|$ ] , and , similarly , @xmath276 .
therefore , if @xmath275 } h - \sp_t h\| \to 0 $ ] for polynomial @xmath277 then @xmath278}_n f - f\| \to 0 $ ] for any continuous @xmath5 .
equivalently , we need to prove that @xmath279}(x ) ) = \e h(x_t(x ) ) , \text { uniformly in $ x$}.\ ] ] notice that @xmath280}(x)$ ] , @xmath185 , is not a markov process .
however if @xmath281 , @xmath189 , is a standard poisson process with rate @xmath9 , starting from @xmath282 , and independent of everything else , then @xmath283 is a markov process for each @xmath2 .
moreover , for all @xmath234 $ ] , @xmath284}_n(x))\right| \to 0 , \quad \text{uniformly in $ x$}.\ ] ] to see this , let @xmath79 be i.i.d .
copies of @xmath75 , as in , and write the triangle inequality @xmath285\right| + \left|\e [ f(g_n^1(x ) ) - f(x)]\right|.\ ] ] since @xmath286\right| \le \|b_n f - f\|$ ] for all @xmath171 , and since @xmath287 is independent of @xmath288 , we have that the first term of the right side of is @xmath289 , and so @xmath290 by the same argument , for @xmath291 , @xmath292 since @xmath89 is independent of @xmath293 , we can replace the @xmath171 in the last display by @xmath89 and obtain @xmath294\right| \le ( \ell - k ) \|b_n f - f\| , \quad \text{for all } 0 \le x \le 1.\ ] ] using the fact that the poisson process @xmath295 is independent of everything else , we obtain @xmath296}(x))\ } \right| \le \e\{|\phi(nt)-[nt]|\}\,\|b_n f - f\| , \quad \text{for all } 0 \le x \le 1.\ ] ] but @xmath297|\ } \le \e\{|\phi(nt)-nt|\ } + 1 \le \sqrt{\e\{|\phi(nt)-nt|\}^2 } + 1 = \sqrt{nt}+1 $ ] , while , from voronovskaya s theorem , @xmath298 .
therefore the right - hand side of converges to @xmath8 as @xmath44 , and this proves .
therefore will follow from @xmath299 for all polynomial @xmath277 . for each @xmath198
, define a random curve @xmath300 that follows @xmath301 up to time @xmath302 and then switches to @xmath303 . more precisely , define @xmath304 see figure [ gluing ] . it is here assumed that the brownian motion @xmath305 driving the defining equation for the wright - fisher diffusion is independent of all random variables used for the construction of @xmath301 . since , for any given initial state , the solution to is unique , we may replace the initial state by a random variable independent of the brownian motion , and this is what we did in the last formula .
thus , if we prove that @xmath306 is differentiable with respect to @xmath302 , we shall have , for @xmath307 , @xmath308 to show that the last derivative exists as well as estimate it , we estimate @xmath309 and @xmath310 . since @xmath277 is a polynomial , it follows that @xmath311 is also a polynomial function of @xmath76 ( see proposition [ poly ] ) and so the forward equation holds : @xmath312 therefore , for any random variable @xmath313 with values in @xmath35 $ ] and independent of @xmath314 , we have @xmath315 taking expectations of both sides and interchanging differentiation and expectation ( by the dct ) we have @xmath316 setting @xmath317 , we have @xmath318 assume next that @xmath17\to \r$ ] is any function . using the identity @xmath319 = \e f(g_n(y ) ) - f(y ) = b_n f(y)-f(y),\ ] ] valid for any function @xmath5 and any @xmath76 , together with the fact that @xmath320 , as @xmath321 , we arrive at @xmath322 . \ ] ] setting now @xmath323 and observing that @xmath324 , we arrive at @xmath325 \label{frontb}\ ] ] combining and we have a formula for the derivative appearing in the last term of : @xmath326 where @xmath327\ ] ] assume now that @xmath328 is lipschitz with lipschitz constant @xmath329 .
then , by , @xmath330 where @xmath331 and @xmath332 , and so @xmath333 by the formula for @xmath334 when @xmath277 is a polynomial ( proposition [ poly ] ) , it follows that the @xmath335 is a finite constant .
hence has been proved .
[ cor1 ] with @xmath336 , @xmath337 } f_r(x ) = \sum_{i=1}^r b_{i , r}(t)\ , x^i,\ ] ] where the @xmath338 are given by .
this is theorem 2 in kelisky and rivlin @xcite . with @xmath339 , we have @xmath337 } f_\theta(x ) = \e e^{-\theta x_t(x ) } = : h(t , x,\theta),\ ] ] where @xmath340 satisfies @xmath341 and the pde @xmath342 the solution to this pde can be expressed in terms of modified bessel functions .
we shall not pursue this further here .
we provided a fully stochastic explanation of the phenomenon of convergence of @xmath3 iterates of bernstein operators of degree @xmath2 when @xmath2 and @xmath3 tend to infinity in different ways .
this problem has received attention in the theory of approximations of continuous functions .
we showed that the problem can be interpreted naturally via stochastic processes .
in fact , these processes , the wright - fisher model and wright - fisher diffusion are very basic in probability theory and are well - understood .
there are a number of interesting directions that open up .
the most crucial thing is that @xmath343 is the expectation of a random variable .
we can construct different operators by using different random variables .
see , e.g. , karlin and ziegler ( * ? ? ?
* eq . ( 1.5 ) ) for an operator related to a poisson random variable .
whereas karlin and ziegler study iterates of these operators , their approach is more analytical than probabilistic . by using approximations by stochastic differential equations , and taking advantage of the tools stochastic calculus , it is possible to derive convergence rates and other interesting results , including explicit formulas , such as the formula for @xmath344 } f_r(x)$ ] ( corollary [ cor1 ] and formula ) , obtained here by a simple application of the it formula .
in section [ interlude ] we explained the most standard wright - fisher model where mutations are not allowed .
if we assume that the probability that a gene of one type changing to another type also depends on the number of genes of each type , in a possibly nonlinear fashion , we obtain a more general model .
mathematically , this is captured by letting @xmath345\to [ 0,1]$ ] be an appropriate function , and by considering the markov chain obtained by iterating independent copies of the function @xmath346 , that is the markov chain @xmath347 , @xmath348 . for the case
@xmath349 with @xmath350 chosen so that @xmath351 , see ethier and norman @xcite .
by a random map @xmath353 we mean a random function from some probability space @xmath354 into a subspace of @xmath355 of functions from @xmath356 to @xmath357 . in rigorous probability theory
, this means that all sets involved are equipped with @xmath358-algebras in a way that @xmath359 is a measurable function of @xmath360 . as a concrete example , in this paper , we considered @xmath361 $ ] and @xmath362^n$ ] .
we equipped all sets with the lebesgue measure . for @xmath363 ,
the map @xmath364 is a random variable with the uniform distribution .
moreover , @xmath73 are independent .
we defined @xmath365 the function @xmath366 by taking a different @xmath354 , we are able to construct two ( or more ) random maps @xmath367 that are independent . as usual in probability , we suppress the symbol @xmath360 from the definition of @xmath368 .
when we talk about the composition @xmath369 of two random functions as above , we are talking about the composition with respect to @xmath76 .
that is , @xmath370 .
if @xmath5 is a deterministic function we define the operator @xmath371 by @xmath372 , @xmath373 .
we can then easily see that @xmath374 the point being that the order of composition outside the expectation is the reverse of the one inside .
this is why , for instance , @xmath375 in eq . .
another example of a random map is the @xmath376 , where @xmath301 is the markov chain constructed in section [ jointsec ] . for fixed @xmath302 and @xmath2 , the initial state @xmath76
is mapped into the state @xmath377 at time @xmath302 .
and yet another random map is @xmath378 , where @xmath206 , @xmath185 , is the wright - fisher diffusion .
we composed these random maps in , after assuming that they are independent .
let @xmath27 be independent random variables with zero mean with values between @xmath379 and @xmath380 for some @xmath381 .
then , for @xmath382 , @xmath383 this inequality , due to hoeffding @xcite , is very well - known and can be found in many probability theory books .
we just prove it below for completeness . since @xmath384 is a convex function of @xmath76 for all @xmath385 and
so if we write @xmath386 as a convex combination of @xmath379 and @xmath380 , @xmath387 we obtain @xmath388 hence @xmath389 here we used the inequality @xmath390 , valid for all real @xmath172 .
this implies that @xmath391 .
hence , for any @xmath385 , @xmath392 the last exponent is minimized for @xmath393 and its minimum value is @xmath394 .
hence @xmath395 .
reversing the roles of @xmath386 and @xmath396 , we have @xmath397 also .
recall that sequence of real - valued random variables @xmath398 is said to be a time - homogeneous markov chain if @xmath399 , for all @xmath3 , @xmath76 and @xmath171 . often ,
markov chains depend on a parameter which , without loss of generality , we can take to be an integer @xmath2 .
let @xmath400 , @xmath401 be such a sequence .
frequently , it makes sense to consider this process at a time scale that depends on @xmath2 .
the problem then is to choose ( if any ) a sequence @xmath402 of positive real numbers such that @xmath403 and study instead the random function @xmath404 } } , \quad t \ge 0,\ ] ] which , hopefully , may have a limit in a certain sense .
this can be made continuous by linear interpolation .
that is , consider the random function @xmath405 , @xmath185 , defined by @xmath406 } } + ( \tau_n t-[\tau_n t ] ) ( z_{n,{[\tau_n t]+1}}-z_{n,{[\tau_n t]}}).\ ] ] we seek conditions under which the sequence of random continuous functions @xmath407 converges weakly to a random continuous function @xmath408 . to define weak convergence ,
we first define the notion of convergence in @xmath409 by saying that a sequence of continuous functions @xmath410 converges to a continuous function ( write @xmath411 ) if @xmath412 , as @xmath44 , for all @xmath413 .
then we say that @xmath414 is continuous if , for all continuous functions @xmath5 , @xmath415 whenever @xmath411 .
finally , we say that the sequence of random continuous functions @xmath416 converges weakly @xcite to the random continuous function @xmath48 if @xmath417 , as @xmath44 , for all continuous @xmath414 .
we now quote , without proof , a useful theorem that enables one to deduce weak convergence to a random continuous function that satisfies a stochastic differential equation .
for the whole theory we refer to stroock and varadhan ( * ? ? ?
* chapter 11 ) .
[ mc2diff ] let , for each @xmath66 , @xmath418 , @xmath419 , be a sequence of real random variables forming a time - homogeneous markov chain .
assume there is a sequence @xmath402 , with @xmath420 , such that @xmath421 \to \sigma^2(x),\ ] ] uniformly over @xmath422 for all @xmath423 , for some continuous function @xmath424 .
for the same @xmath402 , we also assume that @xmath425 \to b(x),\ ] ] uniformly over @xmath422 for all @xmath423 , for some continuous function @xmath426 .
assume also that , for all @xmath427 , there are positive constants @xmath428 , such that @xmath429 for all @xmath108 and all @xmath422 .
finally , assume that there is @xmath430 such that @xmath431 , for all @xmath108 .
then , as @xmath44 , the sequence of random continuous functions defined as in converges weakly to the solution of the stochastic differential equation @xmath432 provided that this equation admits a unique strong solution .
ronald a. devore and george g. lorentz ( 1993 ) .
_ constructive approximation_. springer - verlag , heidelberg .
stewart n. ethier and thomas g. kurtz ( 1986 ) .
_ markov processes : characterization and convergence .
_ wiley .
marc kac ( 1937 ) .
une remarque sur les plynomes de m. s. bernstein . _ studia math . _
* 7 * , 49 - 51 .
howard m. taylor and samuel karlin ( 1981 ) .
_ a second course in stochastic processes . _ academic press , new york .
bernt ksendal ( 2003 ) .
_ stochstic differential equationsa , 6th ed .
_ springer - verlag , berlin .
allan pinkus ( 2000 ) .
weierstrass and approximation theory .
_ j. approx .
_ * 107 * , 1 - 66 .
albert n. shiryayev ( 1984 ) .
_ probability_. springer - verlag , new york .
daniel w. stroock and s.r.s .
varadhan ( 1979 ) .
_ multidimensional diffusion processes .
_ springer - verlag , berlin .
karl weierstrass ( 1885 ) .
ber die analytische darstellbarkeit sogennanter willkrlicher functionen [ _ sic _ ] einer reelen vernderlichen .
_ sitzungsberichte der kninglich preuischen akad .
der wissensch .
zu berlin _ , 663 - 639 , 789 - 805 . | this paper presents a stochastic approach to theorems concerning the behavior of iterations of the bernstein operator @xmath0 taking a continuous function @xmath1 $ ] to a degree-@xmath2 polynomial when the number of iterations @xmath3 tends to infinity and @xmath2 is kept fixed or when @xmath2 tends to infinity as well . in the first instance , the underlying stochastic process is the so - called wright - fisher model , whereas , in the second instance , the underlying stochastic process is the wright - fisher diffusion .
both processes are probably the most basic ones in mathematical genetics . by using markov chain theory and stochastic compositions ,
we explain probabilistically a theorem due to kelisky and rivlin , and by using stochastic calculus we compute a formula for the application of @xmath0 a number of times @xmath4 to a polynomial @xmath5 when @xmath6 tends to a constant . |
let @xmath0 be a smooth closed orientable connected @xmath1-dimensional manifold .
we study _ isotropic _ subgroups ( subspaces ) @xmath2 of its cohomology group ( space ) @xmath3 , where @xmath4 is a field or the ring of integers , under the cup - product @xmath5 i.e. , @xmath6 such that @xmath7 . to avoid duplication of terminology , such as `` rank ( dimension )
, '' we will refer to @xmath8 as modules over the ring @xmath4 . specifically , we study the set @xmath9 of ranks of maximal isotropic submodules of @xmath3 and the corresponding _ isotropy index _ @xmath10 the maximum rank of an isotropic submodule .
we study the structure of ( maximal ) isotropic submodules for finite connected sums and direct products of manifolds .
in particular , we show ( theorems [ theor : conn_sum ] and [ theor : dir_prod ] ) that @xmath11[r]{$h$}(m_1 { \mathrel\#}m_2;r)&=h(m_1;r)+h(m_2;r),\\ \hphantom{{{\cal h}}}\makebox[0pt][r]{$h$}(m_1 \times m_2;r)&=\max{\{\,h(m_1;r ) , h(m_2;r)\,\ } } \end{alignedat } \label{eq : intro - prod}\end{aligned}\ ] ] under certain conditions and with certain exceptions described in the corresponding theorems ( here the sum of sets is understood element - wise ) .
isotropy index bounds the co - rank @xmath12 of the fundamental group , i.e. , the maximum rank of a free homomorphic image of @xmath13 @xcite and , obviously , is bounded by the first betti number @xmath14 : @xmath15 in @xcite , for a field @xmath16 upper and lower bounds on @xmath17 were given in terms of betti numbers ; see proposition [ prop : h - bounds ] . using and , for a given @xmath4 we describe all possible sets @xmath9 ( proposition [ prop : hs ] ) and all possible values of @xmath18 with different @xmath0 in terms of @xmath19 ( theorem [ theor : h , b ] ) , as well as extend to fields of characteristic zero and show that in this case these bounds are exact ( proposition [ prop : b<h < b1 ] ) .
the notion of isotropy has been studied in the context of algebraic geometry .
for instance , isotropic subspace theorems by catanese @xcite and bauer @xcite establish relations between isotropic subspaces of @xmath20 for a smooth quasi - projective variety @xmath0 and certain irrational pencils .
these theorems have been studied in @xcite .
the isotropy index has numerous applications to the topological study of manifolds and foliations .
as we show , isotropy for manifolds has a clear geometric meaning : ( maximal ) isotropic subgroups of @xmath21 of rank @xmath22 correspond to ( maximal ) systems of @xmath22 homologically independent , homologically non - intersecting closed orientable codimension - one submanifolds , @xmath23 being the maximum number of such submanifolds ( theorem [ theor : h(m)_geometry ] ) . while this geometric meaning is defined for @xmath24 , we show that the relevant aspects of isotropy coincide for @xmath24 and @xmath25 ( lemma [ lem : h(g)=h(l ) ] ) , which enables the use of simpler , vector space - based techniques in geometric applications of isotropy .
consider a foliation defined on @xmath0 by a morse form @xmath26 , i.e. , a closed one - form that is locally the differential of a morse function .
such foliations have important applications in modern physics , for example , in supergravity @xcite .
a morse form foliation defines a decomposition of @xmath0 into a finite number @xmath27 of minimal components and a finite number @xmath28 of maximal components , i.e. , connected components of the union of compact leaves , which are cylinders over a compact leaf .
these two numbers are bounded by @xmath23 : @xmath29 where @xmath30 is the singular set , which is finite @xcite . in homological terms , for the number @xmath31 of homologically independent compact leaves it holds @xmath32 @xcite .
a sufficient condition of existence of a minimal component has been given in @xcite in terms of @xmath33 , the rank of the group of the periods : if @xmath34 then the foliation has a minimal component , i.e. , @xmath35 .
also , in case of strong inequality in the upper bound in , i.e. , if @xmath36 the foliation of a morse form in general position has a minimal component @xcite .
if the subgroup @xmath37 generated by the homology classes of all compact leaves of a morse form foliation is maximal isotropic , then the foliation has no minimal components , i.e. , @xmath38 @xcite .
subgroups of @xmath39 are related , by poincar duality , with those of @xmath21 . in particular ,
if the homology classes of some compact leaves of a morse form foliation generate a maximal isotropic subgroup , then @xmath38 @xcite .
obviously , this is the case when the foliation has @xmath23 homologically independent compact leaves .
however , if @xmath40 , then in some cases our results allow us to conclude that @xmath38 by examening only one leaf ( examples [ ex : m22xs1-foliation ] and [ ex : h(m2gxm2 g ) ] ) .
isotropic submodules of @xmath41 can be defined for arbitrary bilinear map @xmath42 , where @xmath43 are finitely generated groups or finite - dimensional vector spaces .
this gives the corresponding notions of @xmath44 and the isotropy index @xmath45 as in . in order to establish our main results and
, we study the behavior of isotropic submodules for skew - symmetric maps under operations of extension of scalars ( proposition [ prop : h_over - diff - rings ] ) and direct sum @xmath46 ( lemmas [ lem : sum ] and [ lem : prod ] ) . by poincar duality ,
@xmath47 and @xmath23 can be equivalently defined in terms of the intersection map @xmath48 instead of the cup product . for a closed one - form @xmath26 on @xmath0 , the isotropy index @xmath49
is defined by the restriction @xmath50\times \ker[{\omega}]}$ ] of the intersection map to the group @xmath51\subseteq h_1(m)$ ] , where @xmath52 $ ] is the integration map .
this notion has been extensively used to study the structure of morse form foliations on closed orientable surfaces @xmath53 of genus @xmath54 .
for example ,
@xmath55 if the foliation has no minimal components , then @xmath56 @xcite . for so - called weakly generic forms , @xmath57 , where @xmath58 is the number of singularities surrounded by a minimal component @xcite . since @xmath59
, these inequalities hold also for @xmath23 .
the paper is organized as follows . in section [ sec_def - spaces ] , we give basic facts on isotropy in finite - dimensional vector spaces and finitely generated abelian groups . in section [ sec_def_manif ] , we introduce the isotropy index for manifolds and consider its properties and geometric meaning . in section [ sec_conn - sum ] , we calculate the isotropy index of the connected sum of two manifolds . in section
[ sec_dir - prod ] , we calculate the isotropy index of the direct product of two manifolds and describe the possible sets @xmath9 . in section [ sec_realization : h , b ] , we completely characterize the relation between @xmath60 and @xmath61 . finally , in section [ sec_corank ] we consider the relations between @xmath60 and @xmath62 .
in this section , we will define the isotropy index and discuss how it changes from groups to vector fields or between vector fields with different scalars .
we will deal with finite - dimensional vector spaces and finitely generated abelian groups . to avoid duplication of terminology , such as `` any subspace or subgroup '' or `` its dimension or rank ''
, we will use terminology from @xmath4-modules , where @xmath4 will be a field or @xmath63 , correspondingly . in particular , _ submodule _ will stand for subspace or subgroup ; _ rank _ will stand for the dimension of a space or the rank of a group .
let @xmath41 , @xmath64 be finitely generated @xmath4-modules and @xmath65 a bilinear map ; @xmath4 is a field or @xmath24 .
[ def : isotropic - space ] a submodule @xmath66 is called _ isotropic _ under the map @xmath67 if @xmath68 , i.e. , @xmath69 for any @xmath70 .
if @xmath4 is a field , @xmath4-modules @xmath41 and @xmath64 are finite - dimensional vector spaces , so we deal with _ isotropic subspaces _ ; if @xmath24 , then @xmath4-modules @xmath43 are finitely generated abelian groups , so we have _
isotropic subgroups_. since @xmath41 is noetherian , every isotropic submodule is contained in some maximal isotropic submodule , not necessarily unique .
denote by @xmath71 the set of ranks of maximal isotropic submodules under the map @xmath67 : @xmath72 obviously , @xmath44 is a finite set of non - negative integers such that @xmath73 proposition [ prop : hs ] below shows that these are the only restrictions on @xmath44 .
the _ isotropy index _
@xmath74 is the maximum rank of the isotropic submodules of @xmath41 : @xmath75 consider the skew - symmetric map @xmath76 , @xmath77,l]$ ] , where @xmath78 is a fixed vector and @xmath79 $ ] is the vector product . for any vector @xmath80 , for example @xmath81 , the subspace @xmath82 , @xmath83 , is maximal isotropic , and so is @xmath84 , @xmath85 .
thus @xmath86 , and @xmath87 . for skew - symmetric maps , usually @xmath88 and thus @xmath89 : [ lem : h(l)=0 ]
let @xmath90 be skew - symmetric .
then @xmath91 , i.e. , @xmath92 , iff either * @xmath93 or * @xmath94 , @xmath95 , and @xmath96 .
let @xmath91 , then @xmath97 for any @xmath98 .
unless @xmath93 , for a skew - symmetric map this implies @xmath95 .
suppose @xmath99 .
consider independent @xmath100 ; @xmath101 .
then for @xmath102 , we have @xmath103 , a contradiction .
the _ kernel _ of a bilinear map @xmath42 is @xmath104 obviously , @xmath105 is an isotropic submodule ; moreover , any maximal isotropic submodule contains @xmath105 , so @xmath106 . generally speaking
, the isotropy index depends on the coefficients .
namely , let @xmath41 , @xmath64 be finitely generated abelian groups and @xmath107 a skew - symmetric bilinear map . for a field @xmath16 , consider the corresponding vector spaces @xmath108 and the induced skew - symmetric bilinear map @xmath109 the isotropy index depends on the field @xmath16 , and generally @xmath110 : consider @xmath111 , @xmath112 .
it has an isotropic subgroup is @xmath113 , thus @xmath114 .
similarly , for @xmath115 , we have @xmath116 , thus again @xmath114 , in accordance with below .
however , for @xmath117 , we have @xmath118 , @xmath119 , so @xmath120 , @xmath112 ; obviously , @xmath121 and thus @xmath122 : @xmath123 on the other hand , consider @xmath124 defined by the matrix @xmath125 0&k\\ \!-k&0 \end{pmatrix * } \ ! , $ ] @xmath126 .
then @xmath127 , but @xmath128 ( thus @xmath129 ) iff @xmath130 .
so for @xmath130 and @xmath131 we have : @xmath132 however , extension of scalars for vector spaces does not decrease @xmath133 : [ lem : h_ext - of - scalars ] let @xmath134 be finite - dimensional vector spaces over a filed @xmath16 and @xmath135 be a skew - symmetric bilinear map .
let @xmath136 be a field , @xmath137 , @xmath138 vector spaces obtained from @xmath139 , @xmath140 by extension of scalars , and @xmath141 the induced map : @xmath142 then @xmath143 consider an isotropic subspace @xmath144 , @xmath145 .
a basis @xmath146 can be be extended to a basis for @xmath139 ; thus @xmath147
. extension of scalars from @xmath16 to @xmath136 gives @xmath148 , so @xmath149 . by , the subspace @xmath150 is isotropic , i.e. , @xmath151 .
we obtain @xmath152 .
a stronger fact holds for groups and @xmath153 : [ lem : h(g)=h(l ) ] let @xmath41 , @xmath64 be finitely generated abelian groups and @xmath107 be a skew - symmetric bilinear map .
denote by @xmath154 the corresponding @xmath63- and @xmath153-modules .
let @xmath155 be the induced skew - symmetric bilinear map
. then 1 .
[ item : g->l ] for every maximal isotropic subgroup @xmath66 , the subspace @xmath156 is maximal isotropic ; 2 . [
item : l->g ] for every maximal isotropic subspace @xmath157 , there is a maximal isotropic subgroup @xmath66 such that @xmath158 .
in particular , @xmath159 [ item : g->l ] let @xmath66 be a maximal isotropic subgroup , @xmath160 . then @xmath156 is an isotropic subspace , @xmath161 .
consider @xmath162 such that @xmath163 , i.e. , all @xmath164 , the torsion subgroup .
then for some @xmath165 , we have @xmath166 . since @xmath2 is maximal , @xmath167 and thus @xmath168 .
therefore , @xmath169 is maximal .
[ item : l->g ] let @xmath170 be a maximal isotropic subspace , @xmath171 , a basis .
consider @xmath172 . then all @xmath173 ; thus for some @xmath165 , all @xmath174 .
we obtain @xmath175 for an isotropic subgroup @xmath176 , @xmath165 .
consider a maximal isotropic subgroup @xmath177 .
for any @xmath178 , we have @xmath179 and thus @xmath180 . since @xmath169 is maximal , @xmath181 .
we obtain @xmath182 .
lemma [ lem : h(g)=h(l ) ] allows formulating lemma [ lem : h_ext - of - scalars ] for fields or @xmath63 : [ prop : h_over - diff - rings ] let @xmath183 be finitely generated @xmath4-modules , @xmath4 being a field or @xmath63 , and @xmath184 be a skew - symmetric bilinear map .
let @xmath185 be a field , @xmath186 , @xmath187 modules obtained by extension of scalars , and @xmath188 the induced map .
then @xmath189 in particular , in addition to extension of scalars of vector spaces , holds for a group and a corresponding vector space over @xmath16 , @xmath190 , since @xmath191 .
in this section , we introduce maximal isotropic subgroups ( subspaces ) of the first cohomology group ( space ) and the isotropy index for manifolds and discuss their geometric meaning and properties .
let @xmath0 be a smooth closed orientable @xmath1-dimensional manifold .
consider the cup product @xmath192 where @xmath24 or @xmath4 is a field .
it is a skew - symmetric bilinear map , and @xmath193 are finitely generated @xmath4-modules ; in case of a field , @xmath193 are vector spaces .
a submodule @xmath6 is called _ isotropic _ if it is isotropic under @xmath194 in the sense of definition [ def : isotropic - space ] , i.e. , if the restriction of the cup - product to @xmath195 is zero : @xmath196 .
accordingly , we denote by @xmath9 the set of ranks of maximal isotropic submodules : @xmath197 proposition [ prop : hs ] below shows that is still the only restriction on @xmath9 , i.e. , that almost any set of non - negative integers is @xmath9 for some manifold @xmath0 . the _ isotropy index _
@xmath198 is the maximum rank of isotropic submodules of @xmath3 .
lemma [ lem : h(g)=h(l ) ] allows us to work interchangeably with @xmath21 and @xmath199 : [ lem : h = h_q ] for a smooth closed orientable manifold @xmath0 , there exists a maximal isotropic subgroup @xmath200 , @xmath201 , iff there exists a maximal isotropic subspace @xmath202 , @xmath203 , i.e. , @xmath204 the notions of @xmath47 and @xmath23 have a clear geometric meaning , which can be characterized as follows : [ def : isotropic - system ] an _ isotropic system _ on a manifold @xmath0 , @xmath205 , is a set of homologically non - intersecting homologically independent smooth closed orientable connected codimension - one submanifolds @xmath206 , intersecting transversely : @xmath207=0 , \label{eq:[xy]=0}\end{aligned}\ ] ] @xmath208 ; @xmath209 .
the requirement can not be simplified to @xmath210 , since on some manifolds there exist submanifolds with non - empty , but homologically trivial , intersection : [ ex : heisenberg - detailed ] on the heisenberg 3-nilmanifold , any two homologically independent smooth closed orientable 2-submanifolds have non - empty , but homologically trivial , intersection .
this will be shown as example [ ex : heisenberg ] below ; here we only give a graphical illustration .
the heisenberg 3-nilmanifold @xmath211 is a @xmath212-bundle over the circle @xmath213 , with the monodromy being a dehn twist @xmath214 , defined as the quotient space @xmath215\times t^2}{(1,x)\sim ( 0,f(x))},\quad f=\begin{pmatrix } 1 & 0 \\ 1 & 1\end{pmatrix}.\ ] ] for the basis cycles @xmath216 , @xmath217 of the torus @xmath212 , we have @xmath218 see [ fig : heisengerg](_a _ ) .
the cycle @xmath217 is homologically trivial , being realized by the boundary of a 2-submanifold ( torus without a disk ) shown in [ fig : heisengerg](_a _ ) .
however , for the two submanifolds @xmath219 shown in [ fig : heisengerg](_b _ ) , we have @xmath220=c$ ] .
the heisenberg nilmanifold @xmath211 , represented as a @xmath212-bundle over the circle @xmath213 .
the circle is shown as the vertical line @xmath221 ; @xmath212 is shown as a horizontal square with the sides @xmath216 and @xmath217 , the opposite sides of the square being identified .
the top is identified with the bottom with the dehn twist : @xmath222 , @xmath223 ; all four vertical lines are identified . (
a ) the curve realizing @xmath217 is the boundary of a 2-submanifold shown as the hatched rectangle , triangle , and another rectangle ; thus @xmath224 .
the triangle forms a disk with two holes , which are glued to the cylinder formed by the two rectangles ; the resulting figure is a torus with a disk removed , whose boundary realizes @xmath217 .
( b ) the two tori @xmath225 , @xmath226 intersect by a curve realizing @xmath217 , thus @xmath220=0 $ ] .
they can not be made non - intersecting ; see example [ ex : heisenberg ] . ]
( 0,0)(100,0 ) ( 60,103)@xmath216 ( -8,35)@xmath221 ( 87,37)@xmath217 ( 25,14 ) ( 50,-15 ) the heisenberg nilmanifold @xmath211 , represented as a @xmath212-bundle over the circle @xmath213 .
the circle is shown as the vertical line @xmath221 ; @xmath212 is shown as a horizontal square with the sides @xmath216 and @xmath217 , the opposite sides of the square being identified .
the top is identified with the bottom with the dehn twist : @xmath222 , @xmath223 ; all four vertical lines are identified . (
a ) the curve realizing @xmath217 is the boundary of a 2-submanifold shown as the hatched rectangle , triangle , and another rectangle ; thus @xmath224 .
the triangle forms a disk with two holes , which are glued to the cylinder formed by the two rectangles ; the resulting figure is a torus with a disk removed , whose boundary realizes @xmath217 .
( b ) the two tori @xmath225 , @xmath226 intersect by a curve realizing @xmath217 , thus @xmath220=0 $ ] .
they can not be made non - intersecting ; see example [ ex : heisenberg ] . ]
( 0,0)(100,0 ) ( 45,63)@xmath225 ( 103,37)@xmath226 ( 50,-15 ) an algebraic model of @xmath211 can be given as follows : consider the 3-dimensional heisenberg group over a ring @xmath4 , @xmath227 then the heisenberg nilmanifold @xmath228 is the quotient of the real heisenberg group by the discrete heisenberg subgroup .
it is a compact orientable connected 3-manifold with nil geometry .
definition [ def : isotropic - system ] implies the cardinality of an isotropic system @xmath229 , the betti number ; thus each isotropic system is contained in a maximal isotropic system .
the following theorem relies on the fact that homology classes @xmath230 can be realized by smooth closed orientable connected codimension - one submanifolds .
[ theor : h(m)_geometry ] let @xmath0 be a smooth closed orientable connected manifold , @xmath205 , and @xmath231 be a poincar duality map . then
let @xmath232 be a ( maximal ) isotropic system .
then @xmath233\,\}}$ ] form a basis of a ( maximal ) isotropic subgroup @xmath200 .
let @xmath234 be a basis of a ( maximal ) isotropic subgroup @xmath200 .
then @xmath235 can be realized by submanifolds @xmath236 that form a ( maximal ) isotropic system .
in particular , * @xmath237 ; * the isotropy index @xmath23 is the maximum cardinality of an isotropic system of submanifolds of @xmath0 .
consider an isotropic subgroup @xmath238 , @xmath201 .
since @xmath21 is torsion - free , it has a basis , @xmath239 .
the cup product @xmath240 is dual to the homology classes intersection map @xmath241 namely , @xmath242 where @xmath243 is a poincar duality map .
realize the cycles @xmath244 by suitable submanifolds @xmath245 , @xmath246 $ ] , choosing them to intersect transversely . then @xmath247=[x_i]\circ[x_j]=du_i\circ du_j = d(u_i\smile u_j),\end{aligned}\ ] ] where the sign depends on the choice of orientation in @xmath236 and @xmath248 . since @xmath2 is isotropic , all @xmath249 ; thus @xmath250=0 $ ] for any @xmath208 . since @xmath251 are independent , so are @xmath252 $ ] .
if @xmath2 is maximal then so is this system , because expanding it would , by duality , expand @xmath2 .
similarly , given a ( maximal ) system of @xmath22 such submanifolds @xmath245 , the group @xmath253\rangle\subseteq h^1(m;{{\mathbb{z}}})$ ] is a ( maximal ) isotropic subgroup , @xmath201 .
examples [ ex : heisenberg ] and [ ex : kt ] below show that the homological interpretation of the non - intersection requirement [ eq:[xy]=0 ] is important for theorem [ theor : h(m)_geometry ] : some manifolds have fewer non - intersecting submanifolds , @xmath210 , with the properties listed in definition [ def : isotropic - system ] , than homologically non - intersecting such submanifolds , @xmath250=0 $ ] . for discussion of systems with @xmath210 ,
see section [ sec_corank ] .
recall that the betti number @xmath254 ; by definition , @xmath255 . by the universal coefficient theorem ,
if @xmath4 is a field with @xmath256 , then @xmath257 . since @xmath258 for @xmath259 , the following statements apply to @xmath213 and a point @xmath260 .
[ lem : h(m)=0 ] let @xmath0 be a smooth closed orientable manifold ; @xmath24 or @xmath4 be a field .
then @xmath261 except that @xmath262 iff any of the following conditions holds : * @xmath263 , or * @xmath264
, @xmath95 , and the cup product @xmath265 . theorem [ theor : h , b ] below states that this lemma gives the only relation between @xmath60 and @xmath61 . by definition
, we have @xmath266 . for @xmath63 , by poincar duality , @xmath267 , so @xmath268 . for a field @xmath16 ,
@xmath269 , so @xmath270 .
since the cup product is skew - symmetric , is given by lemma [ lem : h(l)=0 ] .
[ ex : rp3 ] consider @xmath271 ; it is orientable .
its cohomology ring is @xmath272}{({\alpha}^4)},\end{aligned}\ ] ] where @xmath273 .
thus each @xmath274 is a free @xmath275-module with generator @xmath276 , i.e. , @xmath277 .
we have @xmath278 , and for @xmath279 it holds @xmath280 , i.e. , @xmath281 .
obviously , @xmath282 iff @xmath283 . since @xmath284 is an isotropic submodule , @xmath285 .
there are , though , better estimates : [ prop : h - bounds ] let @xmath0 be a smooth closed orientable manifold and @xmath286 .
for @xmath24 or @xmath4 being a field , with the exception specified below , we have : 1 .
it holds @xmath287 in particular , if @xmath288 , then @xmath289 2 .
if @xmath194 is surjective , then @xmath290 as an exception , if @xmath291 then @xmath292 and of , only the upper bound in holds .
if @xmath293 , for a field this has been shown in @xcite ; for @xmath63 it follows from lemma [ lem : h = h_q ] . if @xmath292 , then by lemma [ lem : h(m)=0 ] either @xmath263 or holds . in the former case , happen to hold ; in , we have @xmath294 . in the latter case , and
the lower bound in do not hold , while in we have @xmath288 and the square root does not exist .
the exception is illustrated by example [ ex : rp3 ] .
in fact , and the lower bound in would not need an exception if we took the floor function of the corresponding expressions , which in all other cases except happen to be integer anyway . for a closed orientable surface of genus @xmath54 ,
gives @xmath295 ; for @xmath1-torus , @xmath296 , gives @xmath297 .
both cases do not fall under exception , since @xmath298 and @xmath299 .
[ ex : h(t ) ] thus , @xmath300 . indeed , gives @xmath301 the following example shows a non - singleton @xmath9 : [ ex : m22xs1 ] @xmath302 .
this is seen in [ fig : m22xs1 ] , but can also be formally proved by theorem [ theor : dir_prod ] below .
for sets , we denote @xmath303 .
[ lem : sum ] let @xmath304 , @xmath305 , @xmath306 , be finite - dimensional vector spaces over a filed @xmath16 and @xmath307 be bilinear skew - symmetric maps .
denote @xmath308 and let @xmath65 be a bilinear skew - symmetric map such that @xmath309 i.e. , @xmath67 is defined as component - wise sum of @xmath310 : @xmath311 where @xmath312 are projections . then : 1 .
a subspace @xmath66 is maximal isotropic iff @xmath313 where @xmath314 are maximal isotropic under @xmath310 .
the set of dimensions of maximal isotropic subspaces @xmath315 3 .
the isotropy index @xmath316 note that these conclusions do not necessarily hold for isotropic subspaces that are not maximal .
for example , each 1-dimensional subspace @xmath317 , @xmath318 , is isotropic , but does not hold for it . by , if @xmath319 are isotropic , then @xmath320 is isotropic .
( @xmath321 ) let @xmath66 be a maximal isotropic subspace .
consider @xmath322 . by ,
@xmath323 , i.e. , both projections @xmath324 are isotropic .
let @xmath325 be maximal isotropic subspaces of @xmath304 .
then @xmath326 is isotropic , and since @xmath327 is maximal , @xmath328 .
( @xmath329 ) conversely , let @xmath330 be maximal isotropic subspaces ; then @xmath320 is isotropic .
consider @xmath331 , i.e. , say , @xmath332 .
since @xmath333 is maximal isotropic in @xmath334 , there exists @xmath335 such that @xmath336 . by , @xmath337 .
thus @xmath2 is maximal .
[ theor : conn_sum ] let @xmath338 , @xmath339 be connected closed orientable manifolds with @xmath340 , and @xmath4 be a field or @xmath24 . then for the connected sum @xmath341 : 1 .
a submodule @xmath342 is maximal isotropic iff @xmath343 where @xmath344 are maximal isotropic submodules .
the set of ranks of maximal isotropic submodules @xmath345 3 .
the isotropy index of the connected sum @xmath346 let @xmath4 be a field .
denote @xmath347 and @xmath348 , @xmath306 . since @xmath349 , the mayer - vietoris sequence gives @xmath350 .
the additive structure is given by the induced maps of the inclusions ; the cup product translates into component - wise product : @xmath351 where @xmath352 and @xmath312 are projections .
then , for fields , lemma [ lem : sum ] gives the result .
now for @xmath24 , the result follows from lemma [ lem : h(g)=h(l ) ] .
[ ex : h(m^2_g ) ] by theorem [ theor : conn_sum ] and given example [ ex : h(t ) ] , or by if @xmath353 , for a closed orientable surface of genus @xmath54 it holds @xmath354 consider @xmath355 from example [ ex : m22xs1 ] with @xmath356 ; see [ fig : m22xs1 ] . then @xmath357 and @xmath358 .
consider @xmath359 , surfaces of genus @xmath216 and @xmath221 , @xmath360 .
theorem [ theor : dir_prod ] below gives @xmath361 therefore , @xmath362 and @xmath363 .
[ lem : xy = yx ] let @xmath41 be a vector space and @xmath364 ; @xmath365
. then @xmath366 implies @xmath367 , @xmath368 for some @xmath216 .
coordinate - wise , we have @xmath369 for all @xmath370 . for those @xmath370 for which @xmath371 , this gives @xmath372 since @xmath373 does not depend on @xmath374 or @xmath375 , all @xmath376 .
we obtain @xmath377 if @xmath378 and @xmath379 if @xmath380 .
if @xmath381 , gives @xmath382 for all @xmath375 , thus @xmath383 , and similarly @xmath384 implies @xmath385 . while in lemma
[ lem : sum ] we had @xmath386 , now consider @xmath387 as large as possible : [ lem : prod ] let @xmath304 , @xmath305 , @xmath306 , be final - dimensional vector spaces over a field @xmath16 and @xmath307 be bilinear skew - symmetric maps .
denote @xmath388 where @xmath389 , and let @xmath65 be a bilinear skew - symmetric map such that @xmath390 where @xmath312 are projections . then : 1 .
a subspace @xmath66 is isotropic iff @xmath391 where @xmath314 is isotropic under @xmath310 , for @xmath392 .
[ item : x2 ] the set of dimensions of maximal isotropic subspaces @xmath393 except that @xmath394 if @xmath395 , i.e. , if either * @xmath396 or * @xmath397 , @xmath398 , and @xmath399 .
the isotropy index @xmath400 note that in contrast to lemma [ lem : sum ] , the first conclusion does not require @xmath2 to be maximal .
let @xmath2 be an isotropic subspace .
we will show that if both projections @xmath401 , then @xmath402 .
consider @xmath178 such that both projections @xmath403 .
let @xmath404 .
since @xmath2 is isotropic and the three components of are independent , we have @xmath405 by lemma [ lem : xy = yx ] , @xmath406 .
the conditions for @xmath395 in item are given by lemma [ lem : h(l)=0 ] .
[ theor : dir_prod ] let @xmath338 , @xmath339 be connected closed manifolds and @xmath4 be a field or @xmath24 .
then for the direct product @xmath407 : 1 .
a submodule @xmath408 is isotropic iff @xmath409 where @xmath344 is isotropic for @xmath410 , @xmath411 .
the set of ranks of maximal isotropic submodules @xmath412 except that @xmath413 if @xmath414 , i.e. , if either * @xmath415 , the betti number , or * @xmath416 , @xmath95 , and @xmath417 .
3 . the isotropy index of the direct product @xmath418 let @xmath4 be a field .
denote @xmath419 by the knneth formula , @xmath420 by construction , @xmath421 for @xmath306 ; @xmath422 , and holds for the cup - products in @xmath423 and @xmath410 , respectively .
lemma [ lem : prod ] gives the result for fields and lemma [ lem : h(g)=h(l ) ] for @xmath24 .
example [ ex : heisenberg - detailed ] shows that in theorem [ theor : dir_prod ] , the direct product can not be replaced by an arbitrary fiber bundle .
[ ex : h_r(t^n ) ] by lemma [ lem : h(m)=0 ] , @xmath424 , so for a torus @xmath425 , we have @xmath297 .
[ ex : s1*sn ] by lemma [ lem : h(m)=0 ] , @xmath426 , @xmath427 , and @xmath424 , so @xmath428 .
[ prop : hs ] for any non - empty finite set @xmath429 of non - negative integers and for @xmath24 or @xmath4 being a field , @xmath430 for some smooth closed orientable connected manifold @xmath0 iff @xmath431 or @xmath432 . if @xmath433 , then @xmath434 , a surface of genus @xmath54 ; see example [ ex : h(m^2_g ) ] .
let now @xmath435 , @xmath436 . by the condition , @xmath437 .
consider @xmath438 by theorems
[ theor : conn_sum ] and [ theor : dir_prod ] , we obtain @xmath439 .
another application of theorem [ theor : dir_prod ] can be found in study of the topology of foliations defined by morse forms .
it is known that if the subgroup of @xmath39 generated by the homology classes of all compact leaves of the foliation is maximal isotropic , then the foliation has no minimal components @xcite .
this condition obviously holds true when the foliation has @xmath23 homologically independent compact leaves .
however , if @xmath40 , in some cases theorem [ theor : dir_prod ] allows to conclude that the foliation has no minimal components by considering only one leaf : [ ex : m22xs1-foliation ] as has been mentioned in example [ ex : m22xs1 ] , @xmath302 ; see [ fig : m22xs1 ] .
even though @xmath440 , if a morse form foliation has the submanifold @xmath441 as a leaf , then it has no minimal components .
in contrast , nothing can be said about a form that has @xmath442 as a leaf , because the system @xmath443 is not maximal .
[ ex : h(m2gxm2 g ) ] @xmath444 , @xmath445 . now consider a cycle @xmath446 that winds around the @xmath447 and also around the @xmath448 , that is , @xmath449 , @xmath450 . if a morse form foliation has a leaf dual to @xmath446 , then it has no minimal components .
by definition of the isotropy index , @xmath451 the only relation between @xmath60 , @xmath61 , and @xmath4 is given by lemma [ lem : h(m)=0 ] ; in particular , any gap between @xmath60 and @xmath61 is possible for a given @xmath4 : [ theor : h , b ] let @xmath452 , and @xmath4 be a field or @xmath24 . there exists a connected smooth closed orientable manifold @xmath0 with @xmath453 and @xmath454 iff any of the following conditions holds : * @xmath455 , or * @xmath456 , or * @xmath457 , @xmath458 , and @xmath95 . for @xmath456 ,
consider @xmath459 .
for @xmath457 and @xmath458 with @xmath95 , consider @xmath271 ; see example [ ex : rp3 ] .
let now @xmath455 .
choose @xmath460 such that @xmath461 for large enough @xmath1 such that @xmath462 for all @xmath374 , consider an @xmath1-manifold @xmath463 by theorem [ theor : dir_prod ] , for each summand @xmath464 , we have @xmath465 , while @xmath466 .
then by theorem [ theor : conn_sum ] , @xmath467 note that the construction used in the proof requires @xmath468 , the ceiling here being the smallest possible value for @xmath469 under .
this condition is not restrictive when @xmath470 , leading to @xmath471 ; it is not very restrictive when @xmath472 , leading to @xmath473 , etc . ; for @xmath474 , @xmath475 , we need @xmath476 .
this requires high dimension when @xmath477 . for @xmath275 ,
however , @xmath478 is enough : for @xmath479 , the manifold in theorem [ theor : h , b ] can be chosen with any given @xmath480 .
let @xmath479 and @xmath481 . for @xmath456
, consider @xmath482 .
let now @xmath483 .
consider @xmath484 example [ ex : s1*sn ] shows that @xmath485 , thus theorem [ theor : conn_sum ] implies @xmath486 this trivially generalizes to @xmath487 as @xmath488 let now @xmath489 . for @xmath490 ,
we use with one summand less in , namely , @xmath491 for @xmath470 , consider @xmath492 . finally , for @xmath457 , @xmath458 , consider an enriques surface @xmath493 .
indeed , @xmath494 , where @xmath495 is an orientation - preserving fixed point - free involution ; note that a @xmath496 surface is simply connected . then @xmath497 ; thus @xmath498 . for @xmath499 , @xmath500 is a reduction ( mod 2 ) of @xmath501 , where @xmath502 is the bockstein homomorphism .
suppose @xmath503 , i.e. , for some @xmath504 we have @xmath505 with @xmath506 because @xmath507 . since @xmath501 is @xmath508-torsion , we obtain that @xmath509 is @xmath510-torsion , where @xmath511 is the quotient map , while the latter group is torsion - free .
thus @xmath512 .
in this section , we give a lower bound on @xmath60 stronger than @xmath513 from .
[ theor : b(m)_geometry ] the co - rank of the fundamental group @xmath12 is the maximum number of non - intersecting homologically independent smooth closed orientable connected codimension - one submanifolds @xmath245 : @xmath514 @xmath208 , @xmath515 . the condition for @xmath236 to be smooth closed orientable is important for deduce that @xmath516
* theorem 2.1 ) is formulated for a compact @xmath0 , possibly with @xmath517 ; @xmath236 are also compact and @xmath518
. since our @xmath0 is closed , then @xmath236 are also closed .
these @xmath236 are two - sided ; since @xmath0 is orientable , so are @xmath236 .
these @xmath236 are not necessary smooth , but any homology class can be realized by a smooth @xmath236 .
accordingly , properties of @xmath12 closely resemble those of @xmath23 .
similarly to , it holds @xcite : @xmath519 in particular , @xmath520 iff @xmath521 .
exactly as in theorems [ theor : conn_sum ] and [ theor : dir_prod ] , for the connected sum , @xmath340 , except for non - orientable surfaces , and the direct product it holds @xmath522 * for the closed orientable surface , @xmath524 @xcite and @xmath295 @xcite ; see example [ ex : h(m^2_g ) ] . * for @xmath1-torus , @xmath525 @xcite and @xmath297 @xcite
; see example [ ex : h_r(t^n ) ] . * for manifolds with quasi - khler and 1-formal fundamental group , for example , for compact khler manifolds , @xmath526 @xcite .
* for @xmath527 from theorem [ theor : h , b ] , it holds @xmath523 .
a non - trivial theorem from @xcite implies that represent the only relation between @xmath12 and @xmath528 for any given @xmath529 .
the last item in example [ ex : b=h ] shows that the construction from theorem [ theor : h , b ] gives an elementary proof of this fact for large enough @xmath529 : comparing theorems [ theor : h(m)_geometry ] and [ theor : b(m)_geometry ] gives @xmath534 together with this gives a geometric proof of lower and upper bounds on the isotropy index @xmath23 , which have been obtained indirectly in @xcite .
we extend this to fields of characteristic zero : both bounds in are exact ( see example [ ex : b=h ] and theorem [ theor : h , b ] ) ; in particular , as we have shown , in many cases @xmath12 is a very strong lower bound for @xmath538 .
however , both inequalities can also be strict : [ ex : kt ] the kodaira
thurston nilmanifold @xmath543 gives an example of @xmath544 indeed , the fundamental group @xmath13 is nilpotent , so @xmath545 ; by theorem [ theor : dir_prod ] and given example [ ex : heisenberg ] , @xmath546 ; and , obviously , @xmath547 . | a subspace or subgroup is isotropic under a bilinear map if the restriction of the map on it is trivial .
we study maximal isotropic subspaces or subgroups under skew - symmetric maps , and in particular the isotropy index the maximum dimension of an isotropic subspace or maximum rank of an isotropic subgroup . for a smooth closed orientable manifold @xmath0
, we describe the geometric meaning of the isotropic subgroups of the first cohomology group with different coefficients under the cup product .
we calculate the corresponding isotropy index , as well as the set of ranks of all maximal isotropic subgroups , for the connected sum and the direct product of manifolds .
finally , we study the relationship of the isotropy index with the first betti number and the co - rank of the fundamental group .
we also discuss applications of these results to the topology of foliations . |
to uncover formation mechanisms of diverse exoplanetary systems , the ( mis)alignment between the planetary orbital axis and the stellar spin axis , which can be measured via the rossiter - mclaughlin ( rm ) effect ( e.g. , @xcite ) or spot - crossing events ( e.g. , @xcite ) , has been recognized as an useful clue .
previous observations have revealed that about one third of hot jupiters have tilted or even retrograde orbits relative to their host star s spin ( e.g. , ) .
this indicates that not only disk - planet interaction but also few - body interaction ( planet - planet scattering and/or the kozai mechanism ) play an important role in planetary migration processes . moreover , some studies pointed out the interesting facts that spin - orbit misalignments in planetary orbits are apparently correlated with host star s temperature @xcite and age , while these correlations can be explained by the properties and evolution of the internal structure in host stars @xcite .
those correlations are important clues to understand the whole picture of planetary migration .
however , previous discussions often overlooked the possible presence of faint distant companions . in most of the rm measurements , except for some obvious cases ( e.g. , xo-2 , hd80606 )
, observers did not check whether the host star has outer companions .
thus any correlation between the spin - orbit misalignment and the existence of binary companions has not been well investigated . to solve this problem ,
we have started high - contrast direct imaging observations for known transiting planetary systems in the course of the seeds ( strategic explorations of exoplanets and disks with subaru ; @xcite ) project .
@xcite reported two candidate companions around hat - p-7 , which was known to have a retrograde hot jupiter hat - p-7b @xcite . in this letter , we present evidence that one of the two candidate companions is indeed a true companion of hat - p-7 , confirmed by the common proper motion and the distance from earth inferred from its spectral type and apparent magnitude .
( 85mm,85mm)fig1.eps ( 85mm,85mm)fig2.eps ccccccc ut dates & inst . & band & exp .
time [ s ] & mag.(east ) & mag.(west ) & ref .
+ 2009aug06 & hiciao & @xmath1 & 585@xmath2 & @xmath3 & @xmath4 & this work@xmath5 + 2009oct30 & astralux & @xmath6 & 30 & @xmath7 & non - detection & @xcite + 2009oct30 & astralux & @xmath8 & 30 & @xmath9 & non - detection & @xcite + 2009nov02 & hiciao & @xmath1 & 100@xmath10 & @xmath11 & @xmath12 & this work + 2011aug12 & ircs & @xmath13 & 560.6@xmath14 & @xmath15 & @xmath16 & this work + 2011aug12 & ircs & @xmath17 & 560.6@xmath14 & @xmath18 & @xmath19 & this work + 2011aug12 & ircs & @xmath20 & 396@xmath21 & @xmath22 & non - detection & this work + 2012jul07 & hiciao & @xmath1 & 1810@xmath23 & @xmath24 & @xmath25 & this work + ccccccc ut dates & inst . &
sa(east ) [ ] & pa(east ) [ @xmath26 & sa(west ) [ ] & pa(west ) [ @xmath26 & ref .
+ 2009aug06 & hiciao & @xmath27 & @xmath28 & @xmath29 & @xmath30 & @xcite + 2009nov02 & hiciao & @xmath31 & @xmath32 & @xmath33 & @xmath34 & this work + 2011aug12 & ircs & @xmath35 & @xmath36 & @xmath37 & @xmath38 & this work + 2012jul07 & hiciao & @xmath39 & @xmath40 & @xmath41 & @xmath42 & this work +
for high - contrast direct imaging , we employed ircs ( for @xmath13 , @xmath17 , @xmath20 bands ) and hiciao ( for @xmath1 band ) with ao188 @xcite on the subaru 8.2 m telescope .
our observation logs ( observing dates , instruments , filters , exposure times ) and properties of the candidate companions are summarized in tables 1 and 2 .
magnitudes of the bright companion are derived by relative photometry with the host star using unsaturated images , and then magnitudes of the faint companion are similarly derived by relative photometry with the bright companion using saturated images .
figure 1 shows pictures of hat - p-7 in different epochs and bands .
figure 2 plots time changes of positions of the candidate companions . based on the tycho-2 catalogue
, hat - p-7 has a proper motion of @xmath43 mas / yr in right ascension ( ra ) and @xmath44 mas / yr in declination ( dec ) . from figure 2
, we find that the motion of the western ( fainter ) candidate is consistent with a background star ( @xmath45 for a degree of freedom of 6 ) , while the motion of the eastern ( brighter ) candidate is different from a background star ( @xmath46 for a degree of freedom of 6 ) and consistent with a common proper motion with hat - p-7 .
in addition to the common proper motion , the spectral type of the eastern candidate supports a consistent distance with hat - p-7 , as follows .
the distance to hat - p-7 was estimated as @xmath47 pc by @xcite , based on the yonsei - yale stellar evolution models @xcite and sme ( spectroscopy made easy : ) analysis .
on the other hand , using the available colors ( @xmath6 , @xmath8 , @xmath13 , @xmath1 , @xmath17 ) , we confirm that the eastern companion is an m5.5v dwarf ( see figure 3 ) . based on the absolute magnitude of an m5.5v dwarf in @xmath1 band @xcite
, we estimate a distance modulus of @xmath48 , which corresponds to @xmath49 pc .
thus the spectro - photometric distance to the eastern companion is in good agreement with the distance to hat - p-7 from earth .
we thus conclude that hat - p-7 has a common proper motion stellar ( m5.5v ) companion at a projected separation of @xmath0 arcsec ( @xmath50 au ) .
note that assuming a random distribution for the companion orbit , the expected unprojected separation is @xmath51 times larger than the projected separation ( see also table 6 of @xcite for a conversion factor from the projected separation to the semi - major axis ) .
here we name the eastern companion as `` hat - p-7b '' ( hereafter , just `` b '' ) .
( 75mm,75mm)fig3.eps ( 85mm,85mm)fig4.eps
@xcite reported the possible existence of a third ( different from hat - p-7b and hat - p-7b ) companion from a long - term radial velocity ( rv ) trend observed with the keck hires between 2007 and 2009 .
we confirm the long - term trend based on rvs measured with the subaru hds in 2008 and 2010 ( the rvs are available upon request ) .
figure 4 shows the rvs taken with the subaru hds and residuals from the best - fit one - planet model with and without a long - term linear rv trend @xmath52 ( see @xcite , for the fitting procedure ) .
we find @xmath53 m s@xmath54 yr@xmath54 , which agrees well with @xcite ( @xmath55 m s@xmath54 yr@xmath54 ) .
note that `` b '' can not explain the observed long - term rv trend .
thus we conclude that there is the third companion in this system . here
we name the long - term rv companion as `` hat - p-7c '' ( hereafter , just `` c '' ) .
the trend corresponds to the mass constraint of `` c '' as , @xmath56 the orbital period of `` c '' is presumably longer than 10 yr , since the trend is still approximately linear .
thus `` c '' is likely to be more massive than jupiter .
we do not detect `` c '' in high - contrast direct images , because it is too close to the host star .
@xcite pointed out that hat - p-7b ( hereafter , just `` b '' ) can not be migrated by the kozai mechanism caused by `` b '' in the presence of `` c '' in the system .
however , they also pointed out that there is another possibility of `` sequential '' kozai migration ( see e.g. , @xcite ) in this system .
namely , an inclined outer stellar companion `` b '' causes the kozai mechanism for the outer `` c '' first , and then the inclined `` c '' induces the kozai mechanism for the inner `` b '' . such an initial inclined configuration can be formed by planet - planet scattering of `` b '' and `` c '' ( and other possible undetected massive planets )
recently , @xcite pointed out the interesting fact that the spin - orbit alignment timescale of hat - p-7 due to the tidal effect is much shorter than the star s age ( see figure 25 of @xcite ) .
namely , the retrograde orbit of hat - p-7b appears to be a strong outlier which should be aligned at the age of hat - p-7 .
however , they did not consider the existence of the outer companions `` b '' and `` c '' .
here we propose an alternative reason of the outlier : if the kozai oscillation of `` b '' caused by `` c '' had once happened , the tidal timescale for spin - orbit alignment can be much longer . based on equation ( 1 ) and the timescale of the kozai oscillation as , @xmath57 where @xmath58 is the host star s mass , @xmath59 , @xmath60 , and @xmath61 are the mass , orbital period , and eccentricity of `` c '' , and @xmath62 is the orbital period of `` b '' @xcite , we estimate that the timescale of kozai oscillation of `` b '' caused by `` c '' is comparable to the age of hat - p-7 ( @xmath63 gyr ) .
thus a possible slowly - changing orbital inclination due to the kozai oscillation might have prevented hat - p-7b from achieving a spin - orbit alignment .
we note that , however , the above discussion may be still optimistic , as the kozai oscillation is easily suppressed by other effects such as general relativity , tides , stellar distortion , or extra bodies ( see e.g. , @xcite ) . at the current position of `` b '' , the timescale of general relativity is indeed shorter than the timescale of kozai oscillation of `` b '' caused by `` c '' .
this means the kozai oscillation does not extend the timescale of spin - orbit alignment any more , although it might have once happened .
we thus note that we may still overlook other special conditions that inhibit the spin - orbit alignment in this system .
thus far , the existence of possible faint outer companions around planetary systems has not been checked and is often overlooked , even though the kozai migration models assume the presence of an outer companion . to further discuss planetary migration using the information of the rm effect / spot - crossing events as well as significant orbital eccentricities ,
it is important to incorporate information regarding the possible or known existence of binary companions .
this is also because a large fraction of the stars in the universe form binary systems @xcite .
thus it would be important to check the presence of faint binary companions by high - contrast direct imaging .
in addition , if any outer binary companion is found , it is also necessary to consider the possibility of sequential kozai migration in the system , since planet - planet scattering , if it occurs , is likely to form the initial condition of such planetary migration .
we present evidence that hat - p-7 has a common proper motion stellar companion by high - contrast direct imaging with subaru hiciao and ircs .
the companion is located at @xmath0 arcsec to the east and estimated as an m5.5v star based on its colors .
we also confirm the presence of the third companion hat - p-7c by rv measurements with the subaru hds .
our finding that hat - p-7 is a wide binary system increase the potential that the sequential kozai migration had once occured in this system .
this scenario may be favorable to explain the discrepancy between the spin - orbit alignment timescale for hat - p-7b and the stellar age noted by @xcite . in recent years
, a number of rm and spot - crossing observations have provided useful information on the spin - orbit ( mis)alignment , which is an important clue to understand the variety of planetary migration mechanisms .
the same holds for eccentric planetary systems , since the eccentricity is another important clue to discuss planetary migration by few - body interactions .
we thus propose that high - contrast direct imaging observations should be made for known planetary systems to check the presence of outer faint companions .
since the dependence of the spin - orbit ( mis)alignment or significant eccentricity on the existence of outer companions is still unclear , it is important to conduct such observations to understand the entire picture of planetary migration .
this paper is based on data collected at subaru telescope , which is operated by the national astronomical observatory of japan ( naoj ) .
this research has benefitted from the spex prism spectral libraries , maintained by adam burgasser at http://pono.ucsd.edu/ adam / browndwarfs / spexprism .
we are grateful to the referee , dr .
simon albrecht , for insightful suggestions .
n.n . acknowledges supports by naoj fellowship , nins program for cross - disciplinary study , and the japan society for promotion of science ( jsps ) grant - in - aid for research activity start - up no .
this work is partly supported by jsps fellowships for research ( dc1 : 22 - 5935 , 23 - 271 ) , a grant - in - aid for specially promoted research , no . 22000005 from the ministry of education , culture , sports , science and technology ( mext ) , grant ast-1009203 from the national science foundation , and by the mitsubishi foundation .
part of this research was carried out at the jet propulsion laboratory , california institute of technology , under a contract with the national aeronautics and space administration . | we report that hat - p-7 has a common proper motion stellar companion .
the companion is located at @xmath0 arcsec to the east and estimated as an m5.5v dwarf based on its colors .
we also confirm the presence of the third companion , which was first reported by @xcite , based on long - term radial velocity measurements .
we revisit the migration mechanism of hat - p-7b given the presence of those companions , and propose sequential kozai migration as a likely scenario in this system .
this scenario may explain the reason for an outlier in the discussion of the spin - orbit alignment timescale for hat - p-7b by @xcite . |
the @xmath6 interaction near threshold is known , since the pioneering works by dalitz and collaborators in the 1960s,@xcite to be strongly attractive as well as strongly absorptive .
the experimental data which traditionally have been used to constrain it consist of elastic , charge - exchange and @xmath8 reaction cross sections at low energies , as low as @xmath9 mev / c , and of the @xmath10 resonance shape extracted from @xmath11 final - state interaction below threshold .
in addition , one has also three accurately determined branching ratios for @xmath12 absorption from rest .
a crucial experimental datum near threshold , the ( complex ) @xmath13 scattering length , is progressively becoming accurately determined in recent years from measurements of the energy shift and width of the @xmath4 atomic @xmath3 state.@xcite the scattering length @xmath14 may be expressed in terms of the @xmath6 @xmath15 scattering lengths @xmath16 and @xmath17 , respectively , as follows@xcite : @xmath18 where @xmath19 is the @xmath20 cm momentum with respect to the @xmath4 threshold . using the nlo corrected deser formula proposed recently by meiner _
et al._@xcite @xmath21 the value of @xmath14 derived from the dear measurement@xcite @xmath22 appears inconsistent with comprehensive fits to low - energy @xmath13 scattering and reaction data , as discussed recently in refs . and as shown pictorially in fig .
[ fig : raha ] taken from ref . .
furthermore , such fits leave the @xmath23 scattering length @xmath24 poorly determined . [
fig : raha ] a precise measurement of the energy shift and width of the @xmath0 atomic @xmath3 state is called for , in order to determine @xmath25 , and this indeed is one of the prime aims of the dear / siddharta experimental collaboration ( see e.g. johann zmeskal s talk in these proceedings ) .
a theoretically meaningful extraction of @xmath24 from @xmath14 and @xmath25 would require a relatively handy and accurate multiple - scattering ( ms ) approximation to the more involved coupled - channel faddeev calculation of the @xmath0 system at low energy .
several faddeev calculations of this sort have been reported.@xcite@xcite however , although the subject of ms approximations to the @xmath0 scattering length has been discussed extensively in recent years,@xcite@xcite only _ one _ of these works produced a correct expression@xcite within the commonly used ` fixed - center ' assumption for the nucleons , with a follow up recently in ref . .
the main purpose of this work is to review the existing ms approximations and to propose a new ms approximation for @xmath25 .
the proposed version goes beyond the ` fixed - center ' assumption .
this should enable one to make a reasonable prediction for the @xmath0 atom @xmath3 level- shift and width on the basis of given values of @xmath14 and @xmath24 without having to perform new time - consuming faddeev calculations or , _ vice - versa _ , to derive fairly reliably @xmath24 from the joint measurement of the @xmath3 level- shift and width in @xmath4 and @xmath0 atoms .
it was shown in the first fully reactive @xmath1 ( @xmath26 ) faddeev calculation@xcite that @xmath25 may be approximated to a few percent by neglecting the hyperonic channels altogether , in accordance with the conjecture by schick and gibson,@xcite provided the input to the @xmath27 faddeev equations consists of @xmath6 t matrices which implicitly account for the @xmath28 coupling .
the faddeev equations for the @xmath29 3-body system , with initial @xmath30 configuration , assume the form @xmath31 where @xmath32 is the free green s function and the full @xmath33 matrix is given by its breakdown into partial @xmath33 matrices : @xmath34 .
the partial @xmath35 and @xmath36 , upon a suitable choice of channels , stand for all the ms terms that start with a @xmath4 collision ( 2-body @xmath37 matrix ) and with a @xmath5 collision ( 2-body @xmath38 matrix ) , respectively , whereas @xmath39 consists of the terms that start with @xmath40 collisions ( 2-body @xmath41 matrix ) .
the choice of @xmath0 initial state dictates that the last of eqs .
( [ eq : t ] ) does not have an inhomogeneous term @xmath41 .
no antisymmetrization is explicitly applied yet for the @xmath42 subsystem .
the @xmath33 matrix which reduces in the limit of zero kinetic energy to the @xmath0 scattering length , up to a kinematical factor , is @xmath43 .
we now outline several ms approximations , progressively by order of complexity , beginning with the brueckner formula@xcite which was motivated by @xmath44 scattering for static nucleons .
charge - exchange degrees of freedom are then introduced and , wherever charge independence holds , the brueckner formula is generalized in terms of isoscalar and isovector meson - nuclear scattering lengths .
the brueckner formula is obtained by neglecting @xmath41 and hence also @xmath39 in eqs .
( [ eq : t ] ) , @xmath45 and treating the nucleons as fixed centers .
furthermore , mostly for the sake of illustration , the charge - exchange two - body channel @xmath46 is switched off , leading to the following expression : @xmath47 for zero - range interactions , and considering the necessary kinematical factors to transform from @xmath48 matrices to scattering lengths @xmath49 , it is straightforward to obtain from eq .
( [ eq : tbru ] ) the brueckner formula : @xmath50 @xmath51 where @xmath52 , with @xmath49 standing generically for @xmath53 and for @xmath54 in the @xmath55 cm system .
the numerator in the brueckner formula consists of single- and double - scattering terms , whereas the denominator provides for the renormalization of these terms to include all the higher - order scattering terms .
the relevant expansion parameter is @xmath56 , where @xmath57 for the deuteron wavefunction .
hence , this series faces divergence once the scattering length is of order @xmath58 fm or more , as we encounter for the @xmath59 channel .
in contrast , for low - energy pions . ] @xmath60 fm and the single- and double - scattering terms ( augmented by charge - exchange scattering as done below ) provide an excellent approximation to the brueckner formula for the @xmath61 scattering length.@xcite the generalization of the brueckner formula to include the charge exchange reaction on the proton ( on - shell for @xmath62 , off - shell for @xmath12 ) is given below .
the inclusion of charge - exchange degrees of freedom for @xmath0 scattering at threshold is due to ref . :
@xmath63 where @xmath64 , with @xmath65 and @xmath66 the threshold scattering amplitudes for @xmath67 and @xmath68 respectively . the derivation of eq .
( [ eq : brux ] ) shows that it holds also for @xmath69 , provided that the replacement @xmath70 is made in the @xmath6 amplitudes .
the charge - exchange scattering contribution to @xmath69 as well as to @xmath25 is substantial . when charge independence is assumed , all four @xmath6 amplitudes in eq .
( [ eq : brux ] ) are given in terms of the isoscalar and isovector threshold amplitudes @xmath71 which are related to the isospin scattering lengths @xmath72 by thresholds.@xcite ] @xmath73 eq .
( [ eq : brux ] ) simplifies then to @xmath74 in close analogy to deloff s expression@xcite for @xmath61 scattering at threshold : @xmath75 where the isoscalar and isovector @xmath76 threshold amplitudes are now given in terms of the isospin scattering lengths @xmath77 by @xmath78 eqs .
( [ eq : iso],[eq : pid ] ) may be obtained directly from eq .
( [ eq : tbru ] ) treating it as an operator relationship in isospace . in table
[ tab : fad ] we show the quality of the ` fixed center ' ms approximation .
( [ eq : int],[eq : iso ] ) were used to compare with values for @xmath25 reported in faddeev calculations done in the isospin basis , with the further approximation of replacing @xmath79 by the tpe matrix elements @xmath80 for @xmath81 from table 1 in ref . .
the values of @xmath82 and @xmath83 listed in the table are sufficient to determine the input scattering lengths @xmath16 and @xmath17 for eq .
( [ eq : iso ] ) .
this version of ms approximation reproduces the real part of @xmath25 to better than about @xmath84 and the imaginary part to within @xmath2 . for the particle - basis faddeev calculation of ref .
( [ eq : int],[eq : brux ] ) were used .
the extra scattering lengths , beyond @xmath82 and @xmath83 , are not listed here .
the renormalized tpe value for @xmath85 was used.@xcite the resulting value of @xmath25 is sensitive to the value used for this matrix element which diverges unless renormalized ; scaling it down from @xmath86 by the ratio of @xmath87 , we get @xmath88 fm , in better agreement with the faddeev value @xmath89 fm.@xcite in table [ tab : ms ] we list two ms evaluations of @xmath25 using a deuteron wavefunction based on the paris potential and @xmath6 input which is similar to that used in the tdd@xcite faddeev calculation .
the accuracy of these ms evaluations is typically @xmath90 .
also listed is our own ms approximation ( from table [ tab : fad ] ) which , perhaps fortuitously , works to about @xmath91 . under the ` fixed center ' assumption we have suppressed the terms @xmath92 and @xmath93 in the coupled faddeev equations for @xmath35 and @xmath36 , respectively , in eq . ( [ eq : t ] ) .
these terms may be rewritten in the form @xmath94 @xmath95 where @xmath96 , so that @xmath97 is a green s function which takes full account of the @xmath42 interaction but is still free with respect to the meson - nucleon interactions .
this leads to the following coupled equations for @xmath35 and @xmath36 : @xmath98 which are in the form of the ` fixed center ' faddeev equations ( [ eq : tfc ] ) with @xmath32 replaced by @xmath97 and with _
@xmath37 and @xmath38 : @xmath99 we thus obtain the following improvement over the brueckner formula eq .
( [ eq : tbru ] ) : @xmath100 this expression for @xmath101 provides as transparent and applicable ms expansion , with @xmath102 and @xmath103 , as the brueckner formula ms expansion is .
charge exchange degrees of freedom can be introduced in a straightforward manner , as done in subsection [ subsec : bru2 ] .
it is conceivable that one can reach in this way a level of accuracy of @xmath2 for approximating fully - reactive faddeev calculations .
in this talk i have surveyed schematically the derivation of ms approximations to the few available fully - reactive faddeev calculations for @xmath25 under the generally accepted practice of treating the nucleons as fixed centers . these ` fixed center ' ms approximations , which provide an excellent approximation scheme already at the double - scattering order for @xmath69 , require the full summation of the ms series for @xmath25 , using eq .
( [ eq : brux ] ) in the particle basis or eq .
( [ eq : iso ] ) in the isospin basis .
the accuracy thus provided is in the range of @xmath104 .
it should be stressed that effective field theory approaches do not yet offer any alternative scheme at present.@xcite i have subsequently outlined a theoretical ms approach for going beyond the ` fixed center ' assumption while keeping the formal appearance of the ` fixed center ' ms formulae , at the expense of renormalizing the input @xmath6 scattering amplitude and the free green s function which for fixed centers gives rise to the @xmath79 dependence of the ms series terms .
it is conceivable that one can reach in this way a level of accuracy of @xmath2 for approximating fully - reactive faddeev calculations .
the aim of this project is to free oneself of depending on available faddeev calculations , which might become obsolete if the dear measurement@xcite of the @xmath4 atomic @xmath3 level shift and width is confirmed and the error bars further reduced , since _ all _ the published faddeev calculations of @xmath25 use considerably larger values for the @xmath14 input .
this work was supported in part by grant 757/05 of the israel science foundation , jerusalem , israel .
ivanov , m. cargnelli , m. faber , h. fuhrmann , v.a .
ivanova , j. marton , n.i .
troitskaya and j. zmeskal , _ eur .
j. a _ * 23 * , 79 ( 2005 ) ; * 25 * , 329 ( 2005 ) ; arxiv : nucl - th/0505022 , presented at the _ international conference on exotic atoms , exa05 _ , stefan meyer institute of subatomic physics , vienna , austria , february 2005 ( http://www.oeaw.ac.at/smi/exa05 ) .
s. wycech and b. loiseau , arxiv : nucl - th/0508013 , presented at the _ international conference on exotic atoms , exa05 _ , stefan meyer institute of subatomic physics , vienna , austria , february 2005 ( http://www.oeaw.ac.at/smi/exa05 ) . | multiple - scattering approximations to faddeev calculations of the @xmath0 scattering length are reviewed and compared with published @xmath1 fully reactive faddeev calculations .
a new multiple - scattering approximation which goes beyond the ` fixed - center ' assumption for the nucleons is proposed , aiming at accuracies of @xmath2 .
a precise value of the @xmath0 scattering length from the measurement of the @xmath0 @xmath3 atomic level shift and width , planned by the dear / siddharta collaboration , plus a precise value for the @xmath4 scattering length by improving the @xmath4 atom measurements , are essential for extracting the @xmath5 scattering length , for resolving persistent puzzles in low - energy @xmath6 phenomenology and for extrapolating into @xmath7-nuclear systems . |
recently the @xmath8 collaboration at jinr , carrying out searches for the @xmath0 decoupled dibaryon resonances in the reaction @xmath9 at an energy below the @xmath10 threshold @xcite , has reported @xcite the observation of such a state with a mass of about 1956 mev , which was called @xmath1 .
the conclusion for the existence of this resonance was drawn from the analysis of the energy spectrum for energetic ( @xmath11 mev ) coincident photons emitted at angles of @xmath2 in the laboratory system from the reaction @xmath9 at an energy of about 216 mev .
the measured spectrum shown in fig.1 consists of a narrow peak at a photon energy of about 24 mev and a relatively broad peak in the energy range ( 45 - 75 ) mev .
the statistical significances for the narrow and broad peaks are 5.3@xmath12 and 3.5@xmath12 , respectively .
the width of the narrow peak ( fwhm ) is about 8 mev , compatible with that of experimental resolution .
this behavior of the photon spectrum conforms to the signature of the @xmath0 decoupled dibaryon resonance @xmath1(1956 ) that is believed to be formed in the radiative process @xmath13 and then to decay via the @xmath14 mode , thereby contributing to the @xmath3 reaction .
the photon energy spectrum calculated with the help of monte carlo simulations under the assumption that the process @xmath15 is the only mechanism of the @xmath16 reaction proved to be in reasonable agreement with the experimental one .
the differential cross section of the resonance production of two photons emitted symmetrically at @xmath17 from the process @xmath18 at an energy of 216 mev was estimated to be @xmath19 9 nb / sr@xmath20 .
the total cross section for the @xmath1(1956 ) production in the @xmath21 reaction at 216 mev found as a result of multiplying this differential cross section by @xmath22 amounts to about 1.4 @xmath23 .
the existence of the @xmath1(1956 ) resonance implies , in particular , that in addition to the conventional mechanisms of photon production in @xmath0 interactions below the @xmath10 threshold there should exist a new one , the dibaryon mechanism @xmath24 .
obviously , depending on kinematical conditions of the @xmath25 and @xmath26 reactions this mechanism would make some contribution to inclusive photon energy spectra of these reactions .
such a contribution should be mainly concentrated in two photon energy regions specific for this dibaryon resonance , which are determined by energies of photons arising from the @xmath1(1956 ) formation and decay . in the center - of - mass system ,
the energy of a photon connected with the dibaryon formation ( @xmath27 ) is determined by its mass @xmath28 and the energy of colliding nucleons ( @xmath29 ) as @xmath30 , while the energy of a photon associated with the radiative dibaryon decay ( @xmath31 ) in the rest frame of this resonance is defined as @xmath32 , where @xmath33 is the invariant mass of the final @xmath0 state populated as a result of this decay .
the contribution from the dibaryon mechanism to the photon energy spectrum of the @xmath25 or @xmath26 reaction would compete with a substantial contribution from the conventional mechanisms .
moreover , in the case of @xmath34 collisions even at an energy below the @xmath10 threshold one should have in view the additional contribution from @xmath35 decays produced in @xmath0 collisions due to the nucleon fermi motion in nuclei . at the same time , since the bremsstrahlung cross section in @xmath36 scattering at energies around 200 mev is substantially larger than that for @xmath37 scattering , one can expect that the contribution of the dibaryon mechanism in @xmath36 and , hence , in @xmath34 scattering is also larger than in @xmath37 scattering . in search of further evidence for the existence of the @xmath1(1956 ) dibaryon resonance we decided to analyze the available experimental data on photon production in @xmath0 and @xmath34 collisions at energies below the @xmath10 threshold resulting in the photon energy spectra of the processes explored .
the inclusive photon energy spectrum of the reaction @xmath38 was measured at a neutron energy of 170@xmath3935 mev for high energy photons ( @xmath40 mev ) emitted at the angle @xmath41 @xcite .
this spectrum ( data were taken from ref .
@xcite and binned in 6 mev energy interval to reduce their statistical fluctuations ) is shown in fig . 2a
together with the prediction of the obe model @xcite folded with a gaussian with @xmath42 mev to allow for an energy spread of an incident beam and detector resolution . a resonance - like structure in the energy range from 45 to 75 mev
is clearly seen in this spectrum . at the same time
the calculated photon energy spectrum does not contain even a hint of the presence of any structure there . to solve the problem of this structure
the authors @xcite assumed that it was caused by the contribution from a capture process , that is , the process @xmath43 , which was not included in the calculations @xcite .
but since such a process ( see fig .
2a ) should contribute in an energy range extending beyond the structure in question , it was suggested in ref .
@xcite that the authors @xcite had incorrectly calibrated the energy response of their @xmath4 detector and so the capture peak was shifted by about 20 mev . however , if this is not the case and at least the shape of the @xmath5 spectrum @xcite is correct , then the question arises : what is the reason for this structure origin ? in fig .
2a we also show the energy spectrum of photons ( @xmath44 20 mev ) emitted at @xmath45 from the process @xmath46 at 170@xmath3935 mev calculated by the monte carlo method for the @xmath1(1956 ) production cross section of 4.3 @xmath23 .
as can be seen , the @xmath1(1956 ) contribution to the @xmath5 photon spectrum lies in the same energy range as the structure of interest and therefore the dibaryon mechanism can be considered as a likely cause for appearance of this structure . in this connection
the existing photon energy spectra of the @xmath47 @xcite reaction are of particular interest .
this is because for low and intermediate photon energies the @xmath48 spectrum is a good approximation to the free @xmath5 spectrum and thus it should reflect features inherent in the @xmath49 reaction .
the inclusive @xmath6 photon energy spectra were measured at an incident energy of 200 mev by the grenoble group @xcite and at 145 and 195 mev by the michigan state group @xcite .
the spectra measured by those groups for @xmath41 at an energy of 200 mev and 195 mev are in excellent agreement .
the experimental @xcite and calculated @xcite inclusive photon energy spectra of the @xmath6 reaction at 195 mev and @xmath41 are shown in fig .
we see again that the experimental spectrum clearly exhibits a characteristic structure in the range ( 45 - 75 ) mev , the presence of which leads to serious disagreement between the experimental and theoretical data .
this fact makes it reasonable to assume that the structure observed in the energy range ( 45 - 75 ) mev of both the @xmath5 and @xmath6 inclusive photon spectra can arise from the same underlying mechanism .
we believe that such a mechanism is the dibaryon mechanism in question . in fig .
2b we show the dibaryon mechanism contribution to the @xmath6 photon energy spectrum for the @xmath1(1956 ) production cross section of 9.4 @xmath23 calculated by the monte carlo method with the fermi motion of nucleons in the deuteron taken into account . simply adding this contribution to the theoretical spectrum @xcite we find that the agreement between the experimental and theoretical spectra becomes significantly better .
finally , the inclusive photon energy spectra were also measured in @xmath50 collisions @xcite .
obviously , the clearest manifestation of the dibaryon mechanism should be anticipated in spectra for light nuclei . in fig .
2c we show the inclusive photon energy spectrum for carbon ( lighter nucleus ) measured at an energy of 200 mev and corrected for the @xmath35 contribution @xcite .
the structure in the energy range ( 45 - 75 ) mev is also seen in this spectrum , it is very likely to be of the same origin as that observed in the @xmath5 and @xmath6 spectra . in fig . 2c
we also show the dibaryon mechanism contribution to the @xmath51 photon energy spectrum for the @xmath1(1956 ) production cross section of 56.5 @xmath23 .
we can see that this contribution is situated in the same energy region as the structure in question .
99 s. b. gerasimov and a. s. khrykin , mod .
phys . lett .
* a8 * , ( 1993)2457 .
s. b. gerasimov , s. n. ershov , a. s. khrykin , phys . at .
nucl.*58 * , ( 1995)844 .
a. s. khrykin et al . , phys .
c * 64 * , 034002(2001 ) .
f. malek et al.,phys.lett .
* b 266 * , 255(1991 ) .
pinston , et al.,phys.lett .
* b 249 * , 402(1990 ) .
j. clayton , et al .
c * 45 * , 1810(1992 ) .
f.p . brady et al.,phys.lett .
* b 349 * , 272(1995 ) .
s.f . shfer et al.,z.phys . * a 339 * , 391(1991 ) .
m. jetter and h.v .
von geramb , phys .
c * 49 * , 1832(1994 ) .
k. nakayama , phys .
c * 45 * , 2039(1992 ) .
j. clayton , et al .
c * 45 * , 1815(1992 ) . | we give reasons for the existence of the @xmath0 decoupled dibaryon resonance @xmath1(1956 ) .
strong evidence for its presence has first been found in the energy spectrum of coincident photons emitted at @xmath2 from the @xmath3 process at 216 mev measured by the dib2@xmath4 collaboration at jinr . as further experimental indications of the @xmath1(1956 )
existence we present those found in the available photon energy spectra of @xmath5 , @xmath6 , and @xmath7 reactions below the pion threshold . it is noted that serious discrepancies between the @xmath5 and @xmath6 experimental cross sections and theoretical calculations can reasonably be explained by the fact that the latter did not take into account the @xmath1 effect . |
= 0.15mg / ml . ( b ) steady - state viscosity as obtained from the strain window where the stress is constant ( see dashed lines in ( a ) ) as a function of shear rate for the two concentrations .
the dashed line gives the solvent viscosity .
error bars are due to the low torque .
( c ) schematic depiction of the counter - rotating cone - plate shear cell placed on an inverted microscope .
( d ) typical reconstructed 3d stack of confocal images , showing a small fraction of labeled actin filaments embedded in a dark background of unlabeled .
green line : example of a tracked filament .
scale bar : 10 m , tick unit : m. [ fig_mm],scaledwidth=45.0% ] over the filament length l[fig_rl ] for short filaments @xmath0 m and ( b ) long filaments @xmath1 m .
black bars : zero strain and red bars strain @xmath2 .
( c ) filament length distribution at different dimensionless strain values .
all distributions were determined for @xmath3 . [ rl_length ] , scaledwidth=45.0% ]
we subjected solutions of 0.02 and 0.15mg / ml to shear flow , increasing the shear rate from 0.075 to @xmath4 in 4 steps of 5 minutes , while keeping track of the total acquired strain , see section [ sec_methods ] .
the stress as a function of strain at 0.15mg / ml is plotted in fig.[fig_mm]a .
the stress decreases with increasing strain after start - up of the shear flow at the lowest shear rate .
this behavior is indicative for strain - softening .
after about 8 strain units the stress stays constant and the viscosity of the system can be determined . in
fig.[fig_mm]b the viscosity is plotted as a function of shear rate for both concentrations .
both systems display a clear shear - thinning behavior .
thus we observe both strain - softening and shear - thinning . the viscosity at 0.02mg /
ml is an order of magnitude less than that at 0.15mg / ml and approaches for high shear rates the buffer viscosity , exemplifying the effect of entanglement .
from now on we refer to these samples as @xmath3=0.15mg / ml and @xmath5=0.02mg / ml , where the indices refer to the viscosity of the samples . to visualize shear - induced conformational transitions of entangled individual actin filaments , we embedded trace amounts of fluorescently labeled actin filaments in the host dispersion . for imaging we used a home - built counter - rotating cone - plate cell mounted on an inverted microscope ( fig.[fig_mm]c ) that was equipped with a multi - pinhole confocal scanning head , allowing frame rates of 8 frames per second .
we obtain the full 3d contour of on average 150 filaments per analyzed dimensionless strain .
fig.[fig_mm]d shows several filaments , including one for which the contour is fitted .
the coordinate system of this fitted contour , parametrized by @xmath6 ( @xmath7 is the index of the coordinates along the contour ) , is given by the velocity direction ( along the shear direction ) , gradient direction ( along the gap direction ) and vorticity direction .
from @xmath6 we can extract information about the global filament conformation , such as the filament contour length @xmath8 and its end - to - end vector @xmath9 .
fig.[rl_length]a and b show the distribution of the ratio @xmath9@xmath10 for @xmath3 , which is a measure for the degree of stretching .
this distribution does not show a significant change with strain for the short filaments ( @xmath11 m , fig . [ rl_length]a ) . for the long filaments ( @xmath12 m , [ rl_length]b ) , the distribution widens towards small @xmath9@xmath10 while a pronounced peak is formed for @xmath9@xmath13 .
this means that the filaments become more stretched and at the same time more bent , which is indicative for the formation of hairpins .
fig.[rl_length]c displays the length distribution at different dimensionless strain values for @xmath3 .
the distribution does not shift to lower values of filament lengths with increasing strain , so that we conclude that shear - softening and shear - thinning is not due to rupture of actin filaments .
we will characterize the filament conformations by extracting the local curvature , @xmath14 , tangent vector , @xmath15 , and binormal vector , @xmath16 , along the contour , see fig.[fig_examples]a . to calculate these values we average over two neighboring points in the contour .
thus , the contour is split up in segments of an average length of = 2.36 m ( supplementary note1 ) .
fig.[fig_examples]b and c display the projections of two typical examples of filaments strained at @xmath5 ( b ) and @xmath3 ( c ) . at @xmath5 , which corresponds to the onset of entanglements , we find filaments with highly curved segments ( region i ) and segments that are stretched and orientated with the flow direction ( region ii ) . at @xmath3 ,
the filament has a hairpin conformation ( region iii ) , with two dangling ends which are stretched in the flow direction ( region iv ) .
strikingly , for @xmath3 the hairpin is confined to the flow / vorticity plane , with its binormal vectors ( blue arrows ) pointing in the gradient direction fig.[fig_examples]c .
this behavior is markedly different from the behavior observed for filaments in dilute solutions @xcite , which tumble in the gradient direction .
the behavior at @xmath5 is intermediate .
the typical configuration of a sheared filament as displayed in fig.[fig_examples]a shows a hairpin , which is characterized by two stretched tails connected by a bent part .
the formation of these hairpins suggests that the distribution of the curvature of the segments changes when entangled semi - flexible polymers are sheared .
stretching and bending of filaments both contribute to the free energy of a filament .
usually this is connected to @xmath9 which measures the stretching of a filament @xcite .
@xmath9 , however , can be small for hairpins , which can thus be incorrectly taken to suggest that the entropic contribution to the energy is low .
this is clearly a flaw because the main part of the filament is stretched .
one therefore needs to analyze the filaments at smaller length scales .
we do this by using the local curvature @xmath14 , which is related to the scaled end - to - end vector @xmath17 ( fig.[fig_examples]a ) . in this analysis
we only used filaments with length @xmath1 m to assure good statistics and sufficient flexibility ( @xmath18 ) . on average
we analyzed about 1200 segments per image stack , allowing us to obtain distributions for the curvature @xmath19 . *
( a ) @xmath5 ( b ) @xmath3 .
open symbols : zero strain @xmath20 .
solid symbols : high strain , @xmath21 .
dashed line is fit to a log - norm distribution , while the solid lines are fits to a gaussian distribution.[fig_diskappa],scaledwidth=45.0% ] the distribution of @xmath14 are plotted in fig.[fig_diskappa ] for zero shear strain and high strain ( @xmath22 ) . for both concentrations
the equilibrium curvature distribution is well - described by a gaussian distribution .
the distribution in equilibrium deviates from earlier studies of unsheared f - actin @xcite , probably due to the different sample environment and relatively high concentrations of filaments that are used in this study . for the low concentration the distribution remains unchanged when straining the system .
in contrast , at the high concentration we observe that the distribution takes the form of a log - normal distribution .
compared to the unsheared situation , the peak of the distribution shifts towards smaller values of @xmath14 , showing that the majority of segments are more stretched .
in addition the distribution also has a long tail with a power law form , showing that strain induces curvature in the system .
these observations hint that at @xmath3 energy is stored in the system , but not at @xmath5 .
this is consistent with the presence of entanglements at @xmath3 .
however , these observations do not yet explain the strain - softening behavior .
d ) with its end - to - end vector , @xmath9(black ) .
the filament is subdivided in segments @xmath7 , allowing us to determine the local orientation in terms of tangent @xmath23 ( red ) and binormal @xmath24 ( blue ) vectors , and the local curvature using the local end - to - end vector @xmath9@xmath25 .
part of the hairpin can be described a circle with curvature @xmath14 .
the relevant angles characterizing the orientation are indicated on the right .
( b , c ) typical conformations of actin filaments at @xmath5 ( b ) and @xmath3 ( c ) . left : views in the flow / vorticity plane .
right : corresponding views in the flow / gradient plane . at @xmath5 ,
the filament has a highly curved segment ( region i ) , and stretched segments that are somewhat oriented in the flow direction ( region ii ) .
it shows no confinement in any plane . for @xmath3
, the filament also has a highly curved segment ( iii ) and a stretched segment ( iv ) and is confined to the flow / vorticity plane .
tick unit : m.[fig_examples],scaledwidth=45.0% ] to quantify the orientations of the filaments , we collected all local tangent and binormal vectors along the filament contour , as plotted in fig.[fig_distribute]a , d on a unit sphere , and calculated the corresponding two - dimensional angular orientational distribution functions which are best fitted with a 2d lorentzian given by @xmath26 . here the angles @xmath27 and @xmath28 , which set up the vectors @xmath15 and @xmath16 , are defined as shown in fig.[fig_examples]a right .
as examples we show in fig.[fig_distribute ] the distribution for @xmath3 after straining the sample to @xmath29215@xmath3012 , separating the orientation of the stretched segments , using all segments with @xmath31m@xmath32 ( top ) and bent segments using all segments with @xmath33m@xmath32 ( bottom ) .
the stretched segments display a strong shear - alignment of the tangent : fig.[fig_distribute]a , b shows that the segments all point in the flow direction .
note however that the distributions are slightly biaxial , i.e. not symmetric around the maximum .
the binormal distribution has a low degree of order and is strongly biaxial , fig.[fig_distribute]c .
this can be rationalized as follows .
when the tangent is well aligned , then there is no clearly defined plane that is spanned by two sequential segments .
thus the binormal is ill - defined though per definition in the plane perpendicular to the tangent : the distribution in fig.[fig_distribute]c is very sharp with respect to @xmath34 and very broad with respect to @xmath35 . on the contrary ,
the bent segments have a pronounced uniaxial ordering of the binormal pointing in the gradient direction ( fig.[fig_distribute]d , f ) , while the distribution of the tangent is biaxial ( fig.[fig_distribute]e ) , which is to be expected for hairpins , see for example region i in fig.[fig_examples]b . in order to quantify separately the orientational behavior of the stretched and bent parts of the filaments , we use the distribution functions @xmath36 to calculate the orientational order tensors @xmath37 and similarly @xmath38 . for our purpose
the traceless diagonalized form @xmath39 is particularly useful since fig.[fig_examples ] and fig.[fig_distribute]b both suggest that the orientational distributions can be biaxial . @xmath40 may be written as @xmath41 where @xmath42 is the orientational order parameter of the main orientation axis and @xmath43 parametrizes the biaxiality of the system .
we will now discuss the behavior of these parameters for the four features that we indicated in fig.[fig_examples]b and c : segments with high curvatures of @xmath33m@xmath32 for @xmath5 ( i ) and @xmath3 ( iii ) , and segments with low curvatures of @xmath31m@xmath32 for @xmath5 ( ii ) and @xmath3 ( iv ) .
the stretched segments at @xmath3 ( iv in fig.[fig_examples ] ) clearly display a high degree of ordering of the tangent with @xmath44 ( solid symbols in fig.[fig_orient]b ) .
the ordering is uni - axial ( solid symbols in fig.[fig_orient]f ) and almost along the flow direction , as expected ( solid symbols in fig.[fig_angles]c and f ) .
@xmath45 is higher for @xmath3 than for @xmath5 ( compare fig.[fig_orient]a and b , ii and iv ) , which shows that entanglements enhance ordering when the sample is sheared .
interestingly , @xmath45 jumps immediately to a high value when strain is applied for @xmath3 . on the contrary , @xmath45 displays a moderate increase with increasing strain for @xmath5 , while the corresponding eigenvector tilts towards the flow direction ( solid symbols in fig.[fig_angles]a and e ) .
this complies with measurements on wormlike micelles @xcite .
thus , the entanglements not only cause stretching of parts of the filaments , as seen from fig.[fig_diskappa ] , but also strong shear - alignment of these stretched segments along the flow direction . whilst this shear - alignment is a well known phenomenon ,
the orientational behavior of segments with high curvature ( i and iii in fig.[fig_examples ] ) is _ a priori _ not obvious .
the binormal is a particularly informative parameter since it is oriented along the normal of the plane spanned by a curved segment . when there is a train of segments with similar curvature , which is the case in the bent part of a hairpin , then binormal vectors point in the same direction
this can be seen for @xmath3 , where the blue arrows in region iii of fig.[fig_examples]c on the right all point in the gradient direction .
the behavior of the binormal order parameter @xmath46 confirms this observation . indeed @xmath46 is higher for the highly curved than for the stretched segments ; compare open and solid symbols in fig.[fig_orient]d .
the eigenvectors belonging to @xmath46 point along the gradient direction ( fig.[fig_distribute]f and open symbols in fig . [ fig_angles]d and h ) and
therefore also the plane spanned by the highly curved segments .
the fact that the binormal is well defined for @xmath33m@xmath32 implies that the ordering of the tangent is low and biaxial .
this is indeed confirmed by fig.[fig_orient]f , where the biaxiality of the tangent @xmath47 is plotted : for @xmath3 the biaxiality of the tangent @xmath47 increases with strain when @xmath48m@xmath32 .
we therefore conclude that the plane spanned by the bent part of the hairpin , which we labeled iii in fig.[fig_examples]c , indeed is located in the flow - vorticity plane , with its normal pointing in the gradient direction .
the behavior of the binormal and biaxiality is distinctively different for @xmath5 .
no increase in @xmath46 nor in @xmath47 is observed with increasing strain ( fig.[fig_orient]c , e ) .
thus , there is no well defined orientation of hairpins at the low concentration , as is exemplified in region i of fig.[fig_examples]b . in fig.[fig_orient ] we also plot the points measured about 150s after cessation of shear flow .
we note a clear difference between the relaxation times at low and high concentration : for @xmath5 the orientation is partly lost , whereas for @xmath3 the orientations remain unchanged .
this indicates that the relaxation time at @xmath3 is indeed much longer than for @xmath5 . a time series after cessation of flow confirms these observations : the shape is mainly lost for @xmath5 ( fig.[fig_orient]g ) , while at @xmath3 the filament conserves its shape ( fig.[fig_orient]h ) .
m@xmath32 ) and the open symbols are the results for the bent segments ( @xmath33m@xmath32 ) .
the grey zone indicates data points about 150s after cessation of shear flow .
also shown is the overlay of the 2d projection of filaments ( g and h ) in four sequential time windows following cessation of shear flow ( time average @xmath49s each ) .
the dotted white lines indicate the confining tube for the first 45s following cessation of shear flow .
scale bar 5 m .
the left column is data for @xmath5 and the right column for @xmath3 .
the lines in ( b , d , f ) are guides to the eye .
i - iv refer to the regions shown in figure [ fig_examples ] .
[ fig_orient],scaledwidth=50.0% ]
by directly visualizing the full 3d conformation of individual actin filaments within entangled solutions , we can show how entanglements influence the conformation of the filaments in response to shear flow .
first , the distribution of curvature for @xmath5 , at the onset of the entangled regime , remains unchanged when applying shear flow , while the number of stretched as well as bent segments increases for @xmath3 when applying shear , see fig .
[ fig_diskappa ] .
this explains the much higher viscosity for @xmath3 .
second , we observe at @xmath3 that hairpins form in shear flow , which tilt into the flow - vorticity plane .
the behavior of entangled filaments is in marked contrast with the dynamics of filaments in dilute solutions which tumble in the gradient direction @xcite .
our observations also explain why tumbling was previously shown to be strongly reduced at concentrations three times higher than @xmath3 @xcite , although the shear rates used in this reference were @xmath50 higher .
we find that the response at @xmath5 lies between these two extremes , since there is no well defined plane into which the filaments turn . in this case
we find shear - thinning which is purely due to increased alignment of the filaments , as we conclude from fig.[fig_orient]a and e. we will now try to relate the formation of the strongly aligned hairpins with the strain - softening in entangled solutions .
hairpins are generally viewed as a signature of an entanglement .
two entangled filaments will strongly bend around the point where they are entangled when they are moved in opposite directions faster than the time they need to relax .
mechanically this leads to strain - stiffening , while pairs of hairpins with the orientations of the bent segments roughly perpendicular to each other result in a very flat distribution of the binormal of the bent segments .
we observe , however , exactly the opposite : the binormal of the bent segments is highly aligned when straining the system at @xmath3 .
moreover , there is a strong increase in the number of stretched segments and a marginal increase in the number of bent segments and we predominantly find only one hairpin per filament , see fig .
[ fig_diskappa]b .
these observations strongly hint that entanglements disappear as the system is strained .
since the hairpins and their stretched tails are strictly located in the flow - vorticity plane , they have no components in the gradient direction that cause the shear stress , which results in strain softening .
the filaments slide over each other facilitating lamellar flow .
these findings are contrary to theoretical predictions that hairpins cause strain - stiffening . however , the theory considered the contour lengths to be much longer than the persistence length @xcite , while in our experiments they are of the same order .
theory does predict an instability where shear deformation pushes out contacts between filaments causing strain - softening @xcite .
this strain - induced loss of entanglements , very similar to the convective constraint release @xcite , is exactly what we find .
m@xmath32 ) and the open symbols are the results for the bent segments ( @xmath33m@xmath32 ) .
the grey zone indicates data points 150s after cessation of shear flow .
[ fig_angles ] , scaledwidth=45.0% ] there are no predictions of the shape of the filaments at high strains after strain softening .
we find experimentally that the effect of the surrounding filaments is to confine the hairpins in the flow - vorticity plane . instead of a confining tube
as can be defined in equilibrium , there are now confining planes , without entanglements .
this can also be appreciated from the movie ( supplementary movie1 ) , where fluctuations in the vorticity directions are observed , but not in the gradient direction , in contrast to the movie of a filament at @xmath5 ( supplementary movie2 ) .
while the hairpins immediately form and can be related to the strain - softening , we also observe shear thinning , see fig .
[ fig_mm]b .
this behavior is likely to relate to the small but significant increase in the orientational ordering of the stretched segments ( solid symbols in fig [ fig_orient]b ) and decrease in the biaxiality ( solid symbols in fig [ fig_orient]f ) parameterizing the flow alignment of the segments .
these new scenarios for strain - softening and shear - thinning exclude the need for scission of the filaments as a pathway to explain non - linear rheology , which is known for living polymers such as wormlike micelles @xcite .
indeed , we observe in fig . [ fig_rl]b no change in the filament length distribution over the full measured range of strain . the mechanism for stress release by the formation of disentangled sliding hairpins could be a precursor for the formation of flow instabilities , which is often related with local reorganization of the constituting particles @xcite .
the distinct orientation of the hairpins also suggests that a normal stress builds up which could lead to flow instabilities .
polymer solutions @xcite , polymer melts @xcite and sticky carbon nanotubes @xcite are all systems that display pronounced normal stresses as well as flow instabilities .
flow instabilities have also been observed for actin dispersions at higher concentrations than we used @xcite .
we did not find any signature of such a behavior , scanning a significant part of the gap of the shear cell at a fixed position from the center of the shear cell .
this could be due to the limited strain applied to the system as well as the relatively low filament concentration as compared to ref .
@xcite . in conclusion , we believe that the mechanism of stress release we identified here may be generally valid for solutions of semi - flexible polymers , including supramolecular systems @xcite .
thus , our findings will aid to the understanding of the complex flow behavior of such systems .
likewise , this mechanism could impact the self - organization of cytoskeletal filaments in response to intracellular shear flows created by processes like cytoplasmic streaming @xcite .
g - actin was isolated from rabbit skeletal muscle @xcite , stored in g - buffer solution ( 5.0 mm tris , 0.2 mm cacl@xmath51 , 0.2 mm atp , 0.2 mm dtt , ph7.4 , 21@xmath52c ) at 4@xmath52c .
fluorescently labeled filaments were obtained by polymerizing g - actin ( 0.2mg / ml ) for 1h at 21@xmath52c by adding 10 vol% 10xf - buffer solution ( 20 mm tris , 2.0 mm cacl@xmath51 , 1.0 m kcl , 20 mm mgcl , ph8.5 , 21@xmath52c , 5.0 mm atp , 2.0 mm dtt ) in the presence of an equimolar amount of atto647n - phalloidin ( atto - tec ) .
unlabeled g - actin ( 0.2 - 0.3mg / ml ) was similarly polymerized , but in the presence of unlabeled phalloidin ( sigma / p2141 ) .
these filaments have a persistence length of @xmath53 9 - 18 m @xcite samples were prepared by diluting labeled filaments in a 1:2000 ratio with gfs - buffer solution ( 10% 10xf - buffer , 60% sucrose in g - buffer ) and mixing this solution in equal volume with unlabeled to reach final concentrations of 0.02 and 0.15mg / ml . the final buffer solution thus contained 30% sucrose , which reduced the off - rate of labeled phalloidin @xcite , improving the signal to noise ratio during image acquisition .
all measurements were done at 21@xmath52c . to produce shear flows with a well - defined linear velocity gradient , we used an adapted version of the counter - rotating cone - plane shear cell used in ref.@xcite ( supplementary fig.1 ) .
it consists of a bottom glass plate ( diameter 80 mm , thickness 170 m , menzel ) which is fixed by two teflon rings that are pressed together between a titanium plate at the bottom and a stainless steel block at the top . in this block
there is a hole where the top cone is inserted , which is also made out of stainless steel .
both the glass plate and steel cone can move independently and are in our experiments moved counter clockwise .
we used an epiplan - neofluar 50x/1.0 pol objective ( zeiss ) .
the shear cell was mounted on an inverted microscope ( zeiss / axiovert 200 m ) , equipped with a multi - pinhole - confocal system ( visitech / vt - infinity - i ) and an epiplan - neofluar 50x/1.0 oil pol objektiv ( zeiss ) .
an argon / krypton laser ( spectra physics / stabilite 2250 ) operating at 647 nm was used for excitation of the fluorescent dye .
an observation area of 151 m x 151 m was imaged onto an emccd - camera ( andor / ixon du-897 ) operated with iq software .
confocal stacks consisted of 51 frames taken at 1 m intervals at a rate of 7.2s per stack .
the difference in the geometrical and the optical path lenght , due to the different refractive indices of the immersion oil ( @xmath54 ) objective and the actin solution ( @xmath55 ) was determined by two independent methods .
first we measured the refractive indices of the immersion oil and the actin solution and calculated the correction factor @xmath56 based on the ratio of those indices .
second we filled a glass capillary of known thickness ( 10 m @xmath57 ) with a solution of fluorescent beads ( @xmath58 m , latex beads , sigma ) and measured the geometrical path length , by the use of a calibrated piezo element , calculating the correction factor @xmath56 based on the difference of the capillary and geometrical length : @xmath59 .
both methods gave equivalent results .
data were taken 30 - 80 m into the sample to reduce wall effects .
our shear protocol consists of four blocks of five minutes where a shear rate of 0.075 , 0.15 , 0.225 , 0.3s@xmath32 , respectively , is applied .
the reason for the shear protocol is that after sample loading and inserting the cone , the orientation of the filaments for @xmath3 is not well defined , while it does not relax .
this is a fact that can not be avoided .
thus quite some strain units are needed to remove this memory effect .
for @xmath5 the sample does relax before we start the experiment and indeed we can follow trends with increasing strain .
rheology data are taken with an anton paar mcr501 , using a cone - plate geometry with 30 mm diameter and 1 degree angle .
the 3d filament contours were tracked with the autodepth - function of the visualization and analysis software imaris ( bitplane/ imaris 6.1 ) , and subsequent filament position analysis was done using matlab ( version , mathworks ) . as a control for the conformation of the filaments before the shear experiment ( @xmath60 )
one frame was taken before placing the cone .
this frame contained about 50 tracer filaments , that were tracked with imaris .
for the analysis of the filament conformation at a certain strain value , 3 statistically independent stacks ( circa 30s apart ) before changing the shear rate , were analyzed , containing also about 50 tracer filaments each , corresponding to a strain of @xmath61=21@xmath30 3 , 59 @xmath30 6 , 126 @xmath30 7 and 215 @xmath30 12 .
for the last data point 3 frames , 120s , 150s and 180s after cessation , were analyzed .
g.h.k . was supported by the foundation for fundamental research on matter ( fom ) , which is part of nwo , and n.a.k .
acknowledges support by a marie curie fp7 iif fellowship .
d.g . was supported by the international helmholtz research school ( biosoft ) .
we thank f. mackintosh , r. winkler , j. dhont , m. giesen and r. merkel for fruitful discussions and y. jia for help with the data analysis . | while semi - flexible polymers and fibers are an important class of material due to their rich mechanical properties , it remains unclear how these properties relate to the microscopic conformation of the polymers .
actin filaments constitute an ideal model polymer system due to their micron - sized length and relatively high stiffness that allow imaging at the single filament level . here
we study the effect of entanglements on the conformational dynamics of actin filaments in shear flow .
we directly measure the full three - dimensional conformation of single actin filaments , using confocal microscopy in combination with a counter - rotating cone - plate shear cell .
we show that initially entangled filaments form disentangled orientationally ordered hairpins , confined in the flow - vorticity plane .
in addition , shear flow causes stretching and shear alignment of the hairpin tails , while the filament length distribution remains unchanged .
these observations explain the strain - softening and shear - thinning behavior of entangled f - actin solutions , which aids the understanding of the flow behavior of complex fluids containing semi - flexible polymers
. networks of semi - flexible polymers or fibers form the basis of many materials that we encounter in daily life .
concentrated dispersions of wormlike micelles are used to engineer the viscoelastic properties of industrial and consumer products @xcite , while polysaccharides and self - assembled supra - molecular structures are used in tissue engineering @xcite and smart gels @xcite . in biology ,
eukaryotic cells are mechanically supported by an internal cytoskeleton composed of protein filaments including filamentous ( f-)actin .
these filaments can form striking non - equilibrium patterns when they are subjected to cytoplasmic flows within plants @xcite and animal embryos @xcite generated by molecular motor activity .
the basis of understanding these phenomena is knowledge of the conformational response of the stiff polymers constituting the materials to an applied flow . here
, we directly visualize the full three - dimensional ( 3d ) contour of labeled subjected to shear flow , in order to resolve the microscopic basis of the macroscopic non - newtonian flow response .
our system lies in between the limiting cases of permanently cross - linked networks , where filaments do not relax , and of dilute non - interacting polymers , where filaments relax freely .
permanently cross - linked networks , which represent a model for cytoskeletal networks , show remarkable viscoelastic properties such as elasticity at low filament density and strong strain - stiffening behavior , where the stress increases with increasing strain @xcite .
theoretical models can capture this behavior on the basis of microscopic properties of the filaments , such as the bending rigidity , length distribution , and cross - link density @xcite .
the rheology of solutions in the absence of cross - links @xcite is comparatively poorly understood .
the complex behavior of such solutions is due to entanglements between the filaments that form at concentrations above the overlap density , where diffusion takes place by reptation within tube - like confinement zones defined by the entanglements @xcite .
the tube itself is not a rigid object , but describes rather a confining potential with a varying tube radius @xcite .
the linear viscoelastic response of such entangled solutions to small deformations that leave the tubes unaffected have been described by wormlike chain models @xcite . when starting up shear flow entangled solutions first display strain - stiffening , followed by strain - softening , where stress decreases with increasing strain @xcite .
in addition , the viscosity of entangled f - actin decreases with increasing shear rate , known as shear - thinning @xcite .
these are key ingredients for the formation of flow instabilities , which often occur in concentrated semi - flexible polymers @xcite such as dna @xcite , wormlike micelles @xcite and also @xcite .
though the connection between shear - thinning and flow instabilities is fairly well understood @xcite , the microscopic mechanism of shear - thinning is still under debate mainly because real space information for nanoscopic systems like dna and wormlike micelles is limited .
is a particularly useful semi - flexible model polymer because the micron - sized lengths and relatively high stiffness allows imaging of the conformation at the single filament level by fluorescence microscopy @xcite .
2d filament tracking in microfluidic devices for dilute solutions , where the filaments do not interact , has shown that the conformational dynamics is dominated by the interplay between the brownian motion of the filament ends and the shear flow @xcite . due to this competition , the filaments tumble in the gradient direction forming so - called hairpins , which contain highly curved segments between the stretched parts of the filaments . here , the only relevant stress component is the shear stress , while reorientational motion takes place in the flow - gradient plane .
the frequency of this tumbling motion in this direction was shown to decreases when the system is entangled @xcite .
since , however , stress can develop in all directions for entangled systems @xcite , knowledge of the 3d contour of the filament is crucial for a full understanding of the mechanical response of these systems .
we use here concentrations just below and above the concentration where a confining tube can be defined @xcite .
we achieve 3d in situ imaging by employing a counter - rotating cone - plate shear cell ( supplementary fig.1 ) in combination with a fast confocal microscope @xcite .
the advantage of this shear cell is that it induces a simple shear flow with a linear velocity gradient , while the presence of a zero - velocity plane guarantees that the filaments stay sufficiently long in the field of view to obtain the full 3d contour .
this approach allows us to test predictions on the structural origin of strain - softening of entangled semi - flexible polymers @xcite .
we quantitatively measure the effect of strain on the distribution of the local curvature . moreover ,
a detailed analysis of the orientational distributions of the filaments reveals that entanglements are lost during strain deformation while hairpins are formed that are confined in the flow - vorticity plane .
thus we identify the mechanism for strain - softening and shear - thinning of entangled solutions . |
heavy ion collisions ( hic ) at relativistic energies are a privileged tool for creating very hot and dense matter in a laboratory .
in particular , lattice chromo - dynamics ( lqcd ) predicts a cross - over toward a new state of matter called quark gluon plasma ( qgp ) at a temperature @xmath0 mev for vanishing chemical potential @xmath1 @xcite .
heavy ion collisions allow to experimentally study the properties of this new state of matter .
this experimental program started in the mid 80 s with fixed target heavy ion experiments at the ags and sps @xcite and continued with the physics program developed at the rhic collider ( bnl ) @xcite .
heavy ion collisions at the future large hadron collider ( lhc ) at cern will open new experimental insights in the study of hadronic matter at high temperature . the alice experiment will be the only experiment at the lhc devoted to heavy ion physics , whereas the atlas and the cms experiments plan to develop a heavy - ion program @xcite in parallel with their main physics goals .
the lhc collider will provide lead ( argon ) high luminosity beams @xmath2@xmath3s@xmath4 ( @xmath5@xmath3s@xmath4 ) at @xmath65.5a tev ( @xmath7a tev ) .
in addition , the lhc will deliver proton beams @xmath8@xmath3s@xmath4 at @xmath614 tev and d+pb beams @xmath9@xmath3s@xmath4 at @xmath106.2a tev , providing a solid baseline for the study of medium effects in hic @xcite . at such ultra - relativistic energies new phenomena
emerge , improving the experimental conditions for studying the hadronic matter in hic : * * initial conditions .
* the initial conditions will be under control by the gluon saturation scenario . at these energies
the initial nucleus - nucleus interaction can be viewed as _ weak _ interactions of a huge number of _ small x _ gluons which are freed in the beginning of the collision leading to the formation of a big gluonic ball @xcite . up
to 8000 gluons in the early stage of the collisions are predicted @xcite .
such a system will rapidly evolve toward a equilibrium @xcite .
most of these processes ( as secondary interaction of minijets ) leading to thermalization will be governed by hard processes ( @xmath11 ) . * * equilibrated matter . * after equilibration of the initial gluonic ball , a hot and long - lived hadronic matter will be formed ( hotter and longer - lived than at lower energy hic ) .
the increase of the beam energy will favor the creation of vanishing baryonic potential hadronic matter with a temperature around 0.5 - 1 gev , well above the critical temperature predicted by lqcd . *
* observables . *
the experience acquired during the last 20 years of heavy - ion physics in the relativistic regime has shown the necessity to measure most of the probes of the reaction dynamics , from hard processes and qgp formation observables , until the freeze - out of the expanding hadron gas .
a coherent explanation of the full set of observables will be the only way to study the properties of the ephemeral qgp .
final states probes like particle multiplicities , hadron @xmath12 distributions , particle ratios , strangeness production , azimuthal asymmetries , etc ... will shed light on the condition of the phase transition and the dynamical evolution of the expanding hadron gas .
penetrating probes , as real and virtual photon production , charmonium production , and light vector meson properties will provide informations about the qgp formation and properties .
moreover , the study of the qgp at lhc energies will be enriched by exploiting new probes which can be efficiently studied in this energy regime : 1 .
hard processes leading to very fast leading partons will provide information about the qgp through the interaction of jets with the surrounding partonic matter @xcite .
huge parton energy losses ( around 1 - 3 gev / fm ) in the qgp will modify the hadronization process of these very energetic jets , inducing a jet - quenching and a suppression of high p@xmath13 hadron production . at the lhc , jet production cross - section in the very high @xmath12 range ( p@xmath1450 gev ) will increase by more of 4 orders of magnitude with respect to rhic beam energies ; 2 .
the debye screening of bottomonium states in qgp will be studied .
suppression of the upsilon family production will be exploited due to its sensitivity to the density of color charges in the medium @xcite ; 3 .
open charm and open beauty production will be exploited as new probes of the strong interacting system of partons @xcite .
open charm ( beauty ) production cross - section will increase by a factor 10 ( 100 ) with respect to rhic energies ; 4 .
the production of massive electroweak bosons ( w@xmath15 and z ) will open the possibility to check the validity of the glauber scaling in hic .
these bosons do not interact with the surrounding medium and they are produced in hard parton collisions ; 5 .
finally , the huge particle multiplicity at the lhc will make possible to measure a large number of observables on an event - by - event basis , increasing the sensitivity to non - statistical fluctuations predicted to occur in a phase transition scenario .
heavy quarks ( charm and beauty quarks ) in hic are produced in the first stages of the collisions and then they coexist with the surrounding medium due to their long life - time .
production rates , transverse momentum ( @xmath12 ) and rapidity ( @xmath16 ) distributions , quarkonia production rates , etc ... will allow for probing the properties of the medium , as the qgp .
charm quarks will be copiously produced in hic at lhc energies : more than 100 @xmath17 pairs per central pb+pb collision and around 5 @xmath18 pairs per collision are expected @xcite .
production of heavy quarks will be dominated by prompt parton - parton scattering .
open heavy flavor production will be dominated by gluon fusion .
next leading - order ( nlo ) diagrams noticeably contribute to the heavy quark production cross - section @xcite although the lo one particle differential cross - section shapes are not appreciably modified by the nlo contribution .
it has been observed that pqcd calculations underestimate the measured differential cross - sections in @xmath19 collisions at tevatron energies , although they are compatible within uncertainties @xcite .
all those experimental data concerns high p@xmath13 production ( larger than 5 gev / c ) . at lhc energies ,
heavy quark production will explore a @xmath20 and transfered momentum @xmath21 domain which has not been well studied .
for instance in rapidities @xmath22 at lhc charm production will be sensible to @xmath23 and @xmath24 gev@xmath25 . in the case of hic
, one should expect gluon shadowing which will modify the heavy quark production based on glauber scaling .
it is expected a heavy quark yield suppression of about 20% ( 10% ) for charm ( beauty ) production in p+pb collisions at lhc energies @xcite due to gluon shadowing .
about 1% of the heavy quark pairs will end in the formation of a colorless bound state : quarkonium .
the process of direct formation of quarkonia from the initially produced @xmath26 pair is a non - perturbative process .
models like color evaporation model ( cem ) or non - relativistic qcd @xcite have been successfully tested in @xmath27 and @xmath19 collisions . in hic direct production of quarkonia
will take place inside the impinging nuclei , leading to an initial nuclear absorption of quarkonia ( normal suppression in hic ) .
the nuclear absorption increases with the enhancement of the relative rapidity between the quarkonia and the impinging nuclei . for this reason ,
nuclear absorption becomes higher at lhc energies , a factor 2 higher than at sps energies .
the nuclear absorption exhibits a relatively flat distribution as a function of the hic centrality for central and semi - central collisions and becomes maximal for central pb+pb collisions at the lhc , where the survival probability of the charmonium is @xmath28 25% ( 50% for bottomonium ) @xcite .
pre - equilibrium stage at lhc energies will be dominated by secondary mini - jet interactions which could noticeably contribute to the total production cross - section of charm quark pairs @xcite .
one could expect up to 200 @xmath17 in central pb+pb collisions taking into account the production of heavy quarks during the thermalization phase .
heavy quarks will be then embedded in a matter mainly formed by gluons and light quarks ( @xmath29 ) .
it is a open question how heavy quarks will behave in such a medium .
will heavy quarks thermalize ? will heavy quarks develop collective motion ?
some models assume that heavy quarks behave like _ brownian _ particles ( m@xmath30m@xmath31 ) @xcite and their transverse momentum ( p@xmath13 ) and rapidity distributions , collective motion ... will probe the properties of the qgp .
although pqcd heavy quark interaction cross - sections with the medium seems too small for reaching full thermalization of heavy quarks , non - perturbative effects in qgp could noticeably increase the heavy - quark cross - section , leading to their full thermalization @xcite .
other models do not address the problem of thermalization and directly assume statistical coalescence of heavy quarks @xcite .
the study of the heavy quark bound states will allow for probing the medium via the debye screening @xcite , the gluon dissociation of quarkonia @xcite , kinetic recombination of heavy quarks @xcite and/or their statistical hadronization @xcite .
it is expected that charmonium production in hic at lhc will be dominated by recombination of @xmath17 pairs in the qgp or at the hadronization . due to the huge production rate of charmed quarks at lhc , these models predict a spectacular enhancement of the charmonium yield in central pb+pb collisions , more than a factor 10 . based on these models , charmonium yield will exhibit centrality dependence proportional to @xmath32 @xcite , which will be a very clear signature of qgp . in the case of charmonium ,
a non - negligible production could occur after hadronization of the qgp due to @xmath33 annihilation @xcite .
for the bottomonium production at the lhc , the recombination mechanism will play a less important role due to the lower bottom pair production rate .
recombination mechanism would compete with the suppression mechanism due to debye screening and/or gluon dissociation .
charm measurements at rhic would help to predict what should occur with beauty at lhc since production rates per central hic are similar .
there will be a production of @xmath34 due to charmonium decay modes of the b mesons . at lhc energies the relative contribution of @xmath34 from b - meson decay should be about 25% @xcite .
upsilon family production will be increased by radiative decay of @xmath35 quarkonia .
about 60% of the total upsilon family production will originate from these resonances .
electroweak bosons ( @xmath36 , @xmath37 ) will be abundantly produced at lhc and this will allow for several precision measurements , for instance high precision measurement of the w mass @xcite
. however , proton luminosities in alice experimental region will be much lower ( @xmath8@xmath3s@xmath4 ) than in other lhc experiments , due to alice detector performances . at leading - order ,
w bosons are produced in quark - antiquark collisions : @xmath38 , @xmath39 etc ... the lo production cross - section of @xmath40 obtained from pythia @xcite amounts to @xmath2817 nb in p+p collisions at 14 tev @xcite , leading to a production of half a millions w bosons decaying in the muonic channel in one year of data taken in the alice interaction point . at these energies ,
contributions to @xmath41 production from higher order diagrams should be about 10% @xcite . in proton+proton collisions
, there will be more @xmath42 than @xmath43 valence quarks , leading to a enhancement of @xmath44 with respect @xmath45 which is more pronounced at high rapidities where the w - boson production is dominated by valence - quarks . at mid - rapidity
the production of @xmath44 and @xmath45 is almost symmetric . in hic ,
w bosons production will not be affected by the medium effects since w bosons will leave the hot and dense bulk of matter without any secondary interaction . in this respect ,
w bosons production will be a very powerful tool for checking the validity of the glauber scaling hypothesis at these energies .
since shadowing could be different in quarks and gluons , measurements of w production in d+pb ( or p+pb ) will be necessary .
w - boson production could provide a powerful normalization of very high @xmath12 ( @xmath4630 - 40 gev / c ) beauty and charm production where a suppression could be expected due to jet quenching of heavy quarks .
the asymmetry between @xmath44 and @xmath45 in hic will be given by the isospin of the nuclei and the elementary asymmetry in p+p , n+n , p+n and n+p collisions @xcite .
in the framework of the alice physics program @xcite , the goal of the muon spectrometer of alice @xcite is the study of open heavy flavor production and quarkonia production ( @xmath34 , @xmath47 and @xmath48(1@xmath49 ) , @xmath48(2@xmath49 ) and @xmath48(3@xmath49 ) ) via the muonic channel . for hic the dependence with the collision centrality and with the reaction plane will also be studied .
the study of electroweak bosons is also foreseen and preliminary studies are under progress @xcite . the main experimental requirement is to measure the quarkonia production in central pb+pb collisions at lhc energies , down to very low p@xmath13 , since low p@xmath13 quarkonia will be sensitive to medium effects like heavy - quark potential screening .
since muons are passively identified by the absorber technique , a lorentz boost is needed to be able to measure quarkonia at low p@xmath13 . on the other hand
, the muon spectrometer has to be as close as possible to the physics of the qgp which occurs in the mid - rapidity region . as a compromise
, the muon spectrometer allows for measuring muons and quarkonia production in an intermediate rapidity range @xmath50 . the acceptance plot in this rapidity range for @xmath34 and @xmath51 mesons decaying into muon pairs
is presented in fig .
[ fig : acceptance ] .
the muon spectrometer will be a unique apparatus at the lhc to measure charmonia production down to p@xmath52 and will cover a rapidity range which completes the one measured by the alice central barrel , and by the cms and atlas experiments . and @xmath51 in the rapidity range @xmath50 , via their muon pair decay and with a muon low p@xmath13 threshold equal to 1 gev / c .
[ fig : acceptance],width=377 ] the other important experimental requirement for the muon spectrometer is to be able to disentangle the different resonances of the @xmath48 family .
in particular the separation between the resonance @xmath53 and @xmath54 ( @xmath55 m @xmath28 300 mev / c@xmath25 ) imposes to the spectrometer an invariant mass resolution about 100 mev / c@xmath25 in the @xmath48 mass region . in order to reach these requirements ,
the muon spectrometer is located downstream ( side c of point 2 at lhc ) of the alice detector covering the angular range @xmath56 , ( -4.0@xmath57 - 2.5 ) consisting of 3 absorbers , a muon magnet , a trigger system and a tracking system .
a more detailed description of the muon spectrometer is given in these proceedings @xcite .
the muon physics program is focused on the measurement of heavy flavor production in p+p , p+a like and a+a collisions at lhc energies .
many studies have been undertaken @xcite mainly in the framework of the technical design report @xcite and the physics performance report of alice @xcite and other studies are still in progress .
invariant mass analysis of muon pairs will provide a direct measurement of the quarkonia production : @xmath34 , @xmath47 , @xmath51 , @xmath53 , @xmath54 .
quarkonia measurement in central pb+pb collisions will be very challenging due to the huge low - energy background @xcite and the large muon combinatorial background ( see in tab .
[ smbattable ] the quarkonia signal to background ratio and the significance from reference @xcite ) . in the case of @xmath34 , a ratio signal to background ( s / b ) in the order of @xmath28 0.15 is expected in most central ( 0 - 10% ) pb+pb collisions at 5.5a tev ( see fig .
[ fig : jpsicentral ] ) . during the first pb+pb run with nominal luminosity at the lhc @xcite ( this corresponds to a period of 10@xmath58 s and an average luminosity of @xmath59 @xmath3 s@xmath4 ) we expect to measure about @xmath60 of @xmath34 and @xmath61 @xmath51 .
those numbers would fluctuate depending on the physics of the heavy quark production and quarkonia production in heavy ion collisions : from total suppression due to color screening , to enhancement due to recombination .
measurement of the ratio of @xmath53/@xmath51 will provide a very powerful experimental observable to constraint the different models on quarkonia suppression in the qgp @xcite .
the measurement of @xmath34 elliptic flow will also be possible and the study of very high p@xmath13 @xmath34 will represent a very promising hard probe .
the increase of luminosity for the ar+ar will allow for cumulating about a factor 5 more quarkonia statistics than in pb+pb lhc run .preliminary quarkonia signal over background ( s / b ) and significance ( @xmath62 ) for quarkonium resonances measured versus centrality in the alice forward muon spectrometer @xcite .
the input cross - sections are taken from @xcite .
shadowing is taken into account . any other suppression or enhancement effects
are not included .
the numbers correspond to one month of pb - pb data taking and are extracted with a 2@xmath63 mass cut .
[ smbattable ] [ cols= " < , < , < , < , < , < , < " , ] in the case of proton - proton run at 14 tev @xcite ( this corresponds to a period of 10@xmath64 s and an average luminosity of @xmath65 @xmath3 s@xmath4 ) , statistics per lhc run will be @xmath66 @xmath34 , and @xmath67 @xmath51 . during the first 10 years of lhc running it
is planned to deliver deuterium lead collisions at @xmath68=6.2a tev and the cumulated statistic per d+pb lhc run will be @xmath69 @xmath34 s and @xmath70 @xmath51s0.12 ( depending on the deuterium or lead beam direction : d+pb or pb+d ) with respect the laboratory frame due to the asymmetry in the deuterium and lead beam energies .
this will slightly change the acceptance of the muon spectrometer and could slightly increase or decrease the reported statistics . since we are interested to measure shadowing for small @xmath20 values inside the lead nucleus , the lead beam will come from the back of the spectrometer , and the spectrometer will cover the rapidity range @xmath71 .
this effect would be more pronounced in p+pb collisions . ] the measurement of open heavy flavors will also be a priority of our physics program .
being very interesting in its own , it will allow for a normalization of the quarkonia production rates .
different alternative analysis will be applied , which exploit the muon production from open charm and beauty mesons via their semi - leptonic decay . [ [ single - muon - measurements . ] ] single muon measurements .
+ + + + + + + + + + + + + + + + + + + + + + + + + the measurement of muon p@xmath13 distributions will provide the first measurement of heavy quark production at high rapidities at lhc . for p@xmath13 larger than 5 - 8 gev / c muon production will be dominated by the semi - muonic decay of d and b mesons .
the expected integrated single muon contribution in the most central pb+pb collisions at 5.5a tev is presented in fig .
[ fig : singlemuons ] . for very high
@xmath12 a noticeable contribution to the single muon @xmath12 distribution will be originated from the muonic decay of the w - bosons .
this contribution will be maximum in the @xmath12 range between 30 and 40 gev / c . in fig .
[ fig : wsinglemuons ] the differential cross - section for the muonic channel in the muon spectrometer acceptance is represented . during a p+p run at 14 tev
it is expected to detect around 50000 muons coming from the decay of the w boson .
[ [ muon - pairs - measurements . ] ] muon pairs measurements .
+ + + + + + + + + + + + + + + + + + + + + + + + like and unlike sign muon pairs originating from the same hard scattering or same heavy quark , will present a residual correlation in the low ( 1 - 3 gev / c@xmath25 ) and high ( 4 - 8 gev / c@xmath25 ) invariant mass ( @xmath72 ) regions .
correlated unlike - sign muon pairs in high @xmath72 region will be mainly produced by semi - muonic decays of @xmath73 and @xmath74 mesons from the same hard scattering : ( 400,150 ) ( 0 , 75.0)hadron collision @xmath75 @xmath76 + @xmath77 + ... ( 150 , 105)@xmath78 ( or @xmath79 , @xmath80 , @xmath81 , @xmath82 ) + ... ( 150 , 45)@xmath83 ( or @xmath84 , @xmath85 , @xmath86 @xmath87 ) + ... ( 320 , 135)@xmath88 + ... ( 320 , 15)@xmath89 + ... ( 105 , 82)(0 , 1)24 ( 105 , 106)(1 , 0)35 ( 130 , 67)(0 , -1)20 ( 130 , 47)(1,0)15 ( 175 , 120)(0 , 1)18 ( 175 , 138)(1 , 0)135
( 175 , 40)(0 , -1)23 ( 175 , 17)(1 , 0)135 ( 325 , 127)(0,-1)35 ( 325 , 22)(0 , 1)35 ( 270 , 75)residual @xmath90 correlation @xmath73 ( 400,150 ) ( 0 , 75.0)hadron collision @xmath75 @xmath91 + @xmath92 + ... ( 150 , 105)@xmath93 ( or @xmath94 , @xmath95 , @xmath96 , ... ) + ... ( 150 , 45)@xmath97 ( or @xmath98 , @xmath99 , @xmath100 , ... ) + ... ( 320 , 135)@xmath89 + ... ( 320 , 15)@xmath88 + ... ( 105 , 84)(0 , 1)22 ( 105 , 106)(1 , 0)35 ( 130 , 67)(0 , -1)20 ( 130 , 47)(1,0)15 ( 175 , 120)(0 , 1)18 ( 175 , 138)(1 , 0)135 ( 175 , 40)(0 , -1)23 ( 175 , 17)(1 , 0)135 ( 325 , 127)(0,-1)35 ( 325 , 22)(0 , 1)35 ( 270 , 75)residual @xmath90 correlation @xmath74 the low @xmath72 region will be populated by @xmath101 or @xmath102 semi - muonic decays from the same bottom quark fragmentation . like - sign
muon pairs will be dominated by @xmath103 and , more exotic mechanism , like @xmath104 oscillations .
the study of such correlations will allow for measuring the beauty production cross - section @xcite and to study the relative contribution of nlo diagrams to the total beauty production cross - section . in hic ,
the @xmath90 residual correlations could be modified by final state effects . in this respect ,
@xmath90 residual correlations could become a probe of the heavy quark re - scattering in the qgp medium .
[ [ correlation - with - the - central - barrel . ] ] correlation with the central barrel .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + in the same way , residual correlation between unlike sign muons and electrons ( or kaons ) can be exploited to get similar physics informations . since
electrons and kaons are detected in the alice central barrel , such a correlation will be sensible to heavy quark production in an intermediate rapidity range @xmath105 . additionally , unlike - sign
@xmath106 pairs can not not be generated by any resonance , nor thermal or direct production @xcite .
[ [ multi - correlation - of - muons . ] ] multi - correlation of muons .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + residual correlation of 3 ( or more ) muons will only be sensible to beauty production .
in particular , 3 muon correlation will probe the beauty production decaying in the charmonium channel which will noticeably contribute to the @xmath34 total production cross - section .
unfortunately , the combinatorial background of 3-muon correlations in pb+pb central collisions will be huge and this technique could only be efficiently applied in p+p , d+pb and ( perhaps ) in ar+ar collisions @xcite .
in summary , the muon spectrometer supports an ambitious physics program in the alice experiment , focused on heavy quark physics for studying qgp at the lhc collider .
the muon spectrometer has entered in its construction phase and will be ready in 2007 at the interaction point 2 of the lhc for the first data taking .
i would like to thank my colleagues of alice collaboration for fruitful discussions .
in particular , i would like to thank philippe crochet , christian finck , rachid guernane , smbat grigoryan and andreas morsch for helping me in the preparation of this talk .
i would like to thank laurent aphecetche , zaida conesa del valle , philippe crochet and christian finck for carefully reading this manuscript .
99 f. karsch _ et al . _ , hep - ph/0103314 ( 2001 ) .
et al . _ , hep - lat/0408046 ( 2004 ) .
r. gavai and e. laermann in these proceedings .
satz , nucl .
a715 ( 2004 ) 3c .
hemmick , j. phys.g : nucl .
( 2004 ) s659 .
y. schutz in these proceedings .
takai , j. phys.g : nucl .
( 2004 ) s1105 and lhcc 2004 - 009/i-013 ( 2004 ) .
alice collaboration , alice physics performace report vol i , j. phys .
g30 ( 2004 ) 1517 .
a. morsch , private communication ( 2005 ) .
lerran , hep - ph/0104285 ( 2001 ) .
e. iancu and r. venugopalan , hep - ph/0303204 ( 2003 ) .
a. krasnitz and r. venugopalan , phys .
86 ( 2001 ) 1717 .
r. baier , a.h .
mueller , d. shiff and d.t .
son , hep - ph/0009237 ( 2000 ) .
wiedemann , nucl .
b588 ( 2000 ) 303 ; x .-
wang , phys .
c58 ( 1998 ) 2321 ; x .-
wang and z. huang , phys .
c55 ( 1997 ) 3047 .
t. matsui and h. satz , phys .
lett b178 ( 1986 ) 416 .
wong , phys .
c58 ( 1998 ) 2358 ; e. shuryak , phys .
c55 ( 1997 ) 961 ; z. lin , r. vogt , and x .-
wang , phys .
c57 ( 1998 ) 899 .
crochet , nucl - ex/0503008 ( 2005 ) .
et al . _ ,
phys . b327 ( 1989 ) 49 .
m. bedjidian _
et al . _ , cern yellow report hep - ph/0311048 ( 2003 ) .
cdf collaboration , phys .
d65 ( 2002 ) 052005 ; cdf collaboration , phys .
91 ( 2003 ) 241804 ; d0 collaboration , phys .
b487 ( 2000 ) 264 .
j. raufeisen , hep - ph/0403195 and j. phys .
g30 ( 2004 ) s1159 . b. muller and x.n .
wang , phys .
68 ( 1992 ) 2437 .
gossiaux , hep - ph/0411324 ( 2004 ) . h. van hees and r. rapp , nucl - ex/0412015 and phys .
c71 ( 2005 ) 034907 .
p. braun - munzinger and j. stachel , phys .
b490 ( 2000 ) 196 .
a. andronic _
_ , nucl - ex/0303036 and phys .
b571 ( 2003 ) 36 .
et al . _ , j. phys.g : nucl .
( 2004 ) s1347 .
l. grandchamp _
et al . _ ,
92 ( 2004 ) 212301 .
et al . _ ,
c63 ( 2001 ) 054905 .
et al . _ , phys .
lett . b444
( 1998 ) 237 .
p. braun - munzinger and k. redlich , eur .
j. c16 ( 2000 ) 519 .
et al . _ ,
phys . rev .
93 ( 2004 ) 072301 .
atlas collaboration , atlas detector and physics performance technical design report , http://atlas.web.cern.ch/atlas/groups/physics/tdr/access.html ( 1999 ) .
pythia http://www.thep.lu.se/~ torbjorn / pythia.html z. conesa del valle , in preparation ( 2005 ) .
s. frixione and m.l .
mangano cern - ph - th/2004 - 081 and hep - ph/0405130 ( 2004 ) .
finck in these proceedings .
alice collaboration , cern / lhcc 99 - 22 ( 1999 ) .
alice collaboration , alice physics performance report vol.2 , in preparation ( 2004 ) .
s. grigoryan private communication and alice internal note , in preparation ( 2004 ) .
gunion and r. vogt , nucl .
b492 ( 1997 ) 301 .
r. guernane _ et al .
_ , alice - int-2004 , in preparation ( 2004 ) .
e. dumonteil and ph .
crochet , alice - int-2005 - 002 ( 2005 ) .
a. zintchenko _
et al . _ , alice - int-2003 - 006 ( 2003 ) . | the main goal of the alice muon spectrometer experiment is the measurement of heavy quark production in p+p , p+a and a+a collisions at lhc energies , via the muonic channel .
physics motivations and expected performances have been presented in this talk . |
we study _ impartial _ _ subtraction _ games on two _ heaps of tokens _ @xcite .
for a background , see also @xcite .
there are two players who obey the same rules and alternate in moving .
we follow the _ normal play _ convention , meaning that a player who can not move loses and the other player wins . in general an impartial game
consists of a _ position _ , which contains information of whose turn it is and describes the given state of the game , and a _ ruleset _ , which decides what move options there are for a given position .
sometimes the term `` game '' is adapted to mean only the ruleset , other times the term `` position '' encompasses the usual meaning of a game , that is whenever the ruleset is understood . here
we study so - called _ invariant _ games @xcite , where the rules in general do not depend on the given position .
for example , in _ 2-heap nim _
@xcite the players remove tokens from precisely one of two finite heaps , at least one token and at most a whole heap .
the game is invariant in the sense that , independent of the size of a heap , a given number of tokens can be removed provided the heap contains at least this number of tokens . as for all impartial games , this game has a _ perfect winning strategy _ , here : the second player to move wins if and only if the two heaps have the same non - negative number of tokens . in general , for an impartial game @xmath12 without drawn moves ,
the _ outcome _ , of a given position , is either a _ previous player _ win ( @xmath13 ) or a _ next player _
win ( @xmath14 ) .
it belongs to @xmath15 , if none of its options is in @xmath13 .
otherwise it belongs to @xmath14 .
this gives a recursive characterization of all @xmath13-positions of a given game @xmath12 , starting with the _ terminal _ positions in @xmath16 , from which no move is possible .
let @xmath17 where @xmath18 are given non - negative integer _
game constants _ , with @xmath19 .
we let the allowed _ positions _ or _ heap - sizes _ be represented by ordered pairs of non - negative integers , bounded by ` rational slopes ' as given by the set @xmath20 in particular , for the special cases @xmath21 all combinations of heap - sizes are allowed ; we omit the index and denoted this set simply by @xmath22 also , we let @xmath23 . following @xcite , a two heap _ subtraction game _ , @xmath24 , is defined via a given set of ordered pairs of non - negative integers @xmath25 .
a legal move from @xmath26 is to @xmath27 for some @xmath28 and provided @xmath29 .
let @xmath30 . in general ,
each allowed move for a _
@xmath31-subtraction game _ is of the form @xmath32 where @xmath9 belongs to a given set of ordered pairs of non - negative integers @xmath33 , and provided @xmath34 .
our main interest in this paper is the following subset of the @xmath31-subtraction games . for a purpose that will become clear later ( in lemma [ l:5 ] )
we will alter the notation somewhat . in a _ @xmath35-subtraction game _ @xmath36
we require that @xmath9 in ( [ 3 ] ) is of the form @xmath37 and @xmath38 where @xmath39 .
hence for these games it is @xmath40 which belongs to `` the set of allowed moves '' @xmath41 ( and not @xmath9 ) .
thus , a typical move in @xmath42 is @xmath43 where @xmath44 and @xmath45 . as before ,
whenever a @xmath35-subtraction game is a subtraction game we write simply @xmath46 , that is whenever @xmath47 and the allowed heap sizes are as given by @xmath48 in ( [ b ] ) .
let us define a surjective map @xmath49 , which takes as input a position in @xmath50 and produces as output a position in @xmath48 , @xmath51 by @xmath52 .
our main result is that any @xmath35-subtraction game @xmath53 is `` @xmath49-equivalent '' to the subtraction game @xmath54 in the following sense .
[ t:1 ] given game constants @xmath55 , suppose that @xmath30 .
then @xmath56 if and only if @xmath57 .
this is equivalent to @xmath58 if and only if , for all @xmath59 , @xmath60 . by this result we make the following definition : the @xmath35-subtraction games @xmath61 and @xmath62 are _ @xmath63-equivalent _ if @xmath64 .
@xmath41-canonical game _ is the subtraction game @xmath46 . in section [ s:2 ]
we prove theorem [ t:1 ] .
a reader who wishes to study some examples before plunging into the proof of the main theorem should skip to section [ s:3 ] , where we illustrate theorem [ t:1 ] via generalizations of nim and wythoff nim .
this discussion is continued in section [ s:4 ] with some open questions where we relate a certain `` splitting behavior '' of wythoff type @xmath65-subtraction games to similar @xmath31-subtraction games .
a generic game has several terminal positions from which no move is possible .
[ l:1 ] given @xmath66 , the set @xmath67 is a subset of all _ terminal positions _ of the @xmath35-subtraction game @xmath42 . in particular
@xmath68 is the set of all terminal positions , if and only if @xmath69 .
also @xmath70 . *
proof . * since , by assumption @xmath71 , we need to show that @xmath72 for all positive integers @xmath73 and @xmath74 . hence , by definition of @xmath50 , we need to show that @xmath75 and that @xmath76 for all positive integers @xmath73 and @xmath74 , but this is clear since , by definition of @xmath77 , it holds for @xmath78 and @xmath79 , and the game constants satisfy @xmath1 .
the set @xmath77 is the complete set of terminal positions of @xmath42 if @xmath80 since @xmath81 gives @xmath82 or @xmath83 .
hence @xmath84 or @xmath85 respectively . for `` only if '' , suppose that @xmath86 , then @xmath87 .
the other case is similar . @xmath88 + the next two lemmas discuss how the positions in @xmath50 can be viewed as linear translations of those in @xmath77 .
[ l:2 ] let @xmath59 .
then @xmath89 if and only if @xmath90 . *
* what is required is to show that @xmath91 and @xmath92 if and only if both @xmath93 and @xmath94 are non - negative , assuming that @xmath71 .
but , as in the proof of lemma [ l:1 ] , the `` if''-part is immediate by @xmath1 . hence , suppose that @xmath95 . by definition
there is no move from any @xmath35-subtraction game from @xmath59 .
hence , by negativity of @xmath93 , we must have @xmath96 , which contradicts ( [ second ] ) .
the case @xmath97 is similar .
@xmath88 + [ l:3 ] if @xmath30 , then @xmath98 , for some unique @xmath99 and some unique non - negative integers @xmath93 and @xmath94
. * proof . *
suppose that @xmath100 , with @xmath101 , @xmath102 and non - negative integers @xmath103 and @xmath104 .
suppose that @xmath105 .
this gives @xmath106 which implies @xmath107 which is impossible by lemma [ l:1 ] .
hence @xmath108 and @xmath109 which gives @xmath110 which is impossible . next ,
let us find @xmath111 such that @xmath112 .
we have that @xmath113 and @xmath114 .
hence , by @xmath115 , there is a largest non - negative @xmath94 such that @xmath116 and a largest non - negative @xmath93 such that @xmath117 .
then @xmath118 defines @xmath59 by our choices of @xmath93 and @xmath94 .
@xmath88 + [ r:1 ] following lemma [ l:3 ] , we say that ( the game ) @xmath119 belongs to the _
@xmath120-class _ , where @xmath59 .
then , by @xmath121 , @xmath122 belongs to the @xmath123-class if and only if the associated restriction of @xmath49 is @xmath124 .
we get the following consequence of the above lemmas for our @xmath65-subtraction games .
[ l:5 ] let @xmath66 .
suppose that there is a move of the form @xmath125 in the subtraction game @xmath46 , then for any given game constants @xmath126 and for each @xmath59 , there is a move in the @xmath65-subtraction game @xmath42 of the form @xmath127 suppose on the other hand that there is a move in @xmath128 from @xmath30 via @xmath44 .
then there is a corresponding move @xmath125 in @xmath12 where @xmath129 for some @xmath59 .
thus , with notation as in lemma [ l:5 ] and remark [ r:1 ] , each game in the @xmath120-class ends in @xmath120 if @xmath69 .
let us restate and prove our main theorem , where the function @xmath49 is as in ( [ varphi ] )
. given game constants @xmath55 , suppose that @xmath30 .
then @xmath56 if and only if @xmath57 .
this is equivalent to @xmath58 if and only if , for all @xmath59 , @xmath60 . * proof .
* we begin with the second part .
suppose that @xmath58 .
then , none of its options is in @xmath130 .
we need to prove that none of the options in the @xmath65-subtraction game @xmath128 from @xmath131 is in @xmath132 .
we have that , for all @xmath44 such that @xmath133 and @xmath134 , @xmath135 .
then , by lemma [ l:5 ] , induction gives that @xmath136 if and only if @xmath133 and @xmath134 .
if , on the other hand @xmath137 , then there is an option @xmath138 , with @xmath139 , and so induction gives that @xmath140 .
for the first part , by lemma [ l:3 ] , if @xmath30 , then @xmath98 , for some unique @xmath99 and some unique non - negative integers @xmath93 and @xmath94 .
we plug this into the expression @xmath141 by @xmath142 and @xmath143 , since @xmath59 .
but then the first part of the proof gives the result .
our first example of a 2-heap subtraction game , 2-heap nim , was discussed briefly in the introduction . the game of _ wythoff nim _
@xcite is also played on two heaps , all the moves in nim are allowed and also the possibility of removing the same positive number of tokens from both heaps in one and the same move , a number bounded by the number of tokens in the smallest heap .
see figure [ f : nim ] for the @xmath13-positions of nim and wythoff nim ( ) respectively . for the latter game , it is known @xcite that they are of the form : @xmath144 where @xmath145 ranges over the non - negative integers and @xmath146 denotes the golden ratio .
-positions of 2-heap nim , @xmath147 to the left and wythoff nim , @xmath148 to the right.,title="fig:",scaledwidth=40.0% ] -positions of 2-heap nim , @xmath147 to the left and wythoff nim , @xmath148 to the right.,title="fig:",scaledwidth=40.0% ] for our first example of a @xmath65-subtraction game , a player may move @xmath149 for @xmath150 and any positive integer @xmath74 , provided @xmath151 .
we denote this game by _ rational nim _
( rn ) , @xmath152-rn . here
the ratios in the prefix are merely symbols and so for example @xmath153-rn and @xmath154-rn denote different games .
note that , given @xmath65 , this game may be represented simply by the set @xmath155 and that the case @xmath156 , @xmath157 is the classical game of nim on 2 heaps .
see figure [ f:1 ] for the @xmath13-positions of an rn game .
an _ extension _ of a game @xmath12 has the same set of positions as @xmath12 , contains all the moves in @xmath12 and possibly some new .
thus a trivial extension of 2-heap nim is nim itself .
a non - trivial extension is for example wythoff nim .
further , a _
@xmath65-extension _ is an extension of a @xmath65-subtraction game which is also a @xmath65-subtraction game .
thus another way of expressing this is that @xmath158 is a @xmath65-extension of both @xmath42 and @xmath159 .
note that @xmath65 is fixed
. given game constants @xmath55 , we denote by _ rational wythoff nim _ ( rw ) , _
@xmath152- _ , the following @xmath65-extension of @xmath152-rn .
in addition to the moves in ( [ 2 ] ) , the new moves are of the form @xmath160 provided that @xmath161 . hence , given @xmath65 , this game may equivalently be represented by @xmath162 and the classical game of wythoff nim is the game where @xmath156 , @xmath157 .
see figure [ f:2 ] for the initial @xmath13-positions of an rw game .
-positions for the game @xmath163-rw.,scaledwidth=50.0% ] -positions for the game @xmath163-rw.,scaledwidth=50.0% ] theorem [ t:1 ] gives that @xmath164 is periodic with period @xmath165 and @xmath166 is aperiodic similar to @xmath167 , as described by the following result .
[ c:1 ] the @xmath13-positions of rational nim are given by @xmath168 , where @xmath59 and where @xmath145 ranges over the non - negative integers .
the set @xmath166 is given by all positions of the forms @xmath169 and @xmath170 where @xmath59 and @xmath145 is a non - negative integer .
see also corollary [ c:2 ] in the next section .
is it true that adjoining new moves to an existing impartial game changes its set of @xmath13-positions ? for the particular games studied in section [ s:3 ] this is certainly true by corollary [ c:1 ] , as is also illustrated in figures @xmath171 and @xmath172 .
in fact , we have seen a particular shift of behavior inherited from the relation between nim and wythoff nim . in describing rn s and rw s asymptotic behavior we go from one accumulation point to two accumulation points . following the terminology in @xcite , we note that the particular @xmath65-extension of rn we have introduced , rw , _ splits _ the ` old ' set of @xmath13-positions of rn into two ` new ' @xmath13 sets for which the ratios of the heap - sizes _ converge _ to two distinct real numbers . we have the following simple consequence of corollary [ c:1 ] . [ c:2 ] let the game constants @xmath126 be given .
let @xmath173 , with the @xmath174s in lexicographical order .
then @xmath175 as @xmath176 .
let @xmath177 , where the _ lower _ @xmath13-positions are the @xmath178s in lexicographical order and the _ upper _ @xmath13-positions are the @xmath179s in lexicographical order for which @xmath180 then @xmath181 @xmath182 as @xmath183 for another example , take the set @xmath184 .
the first few @xmath13-positions are : @xmath185 for all non - negative integers @xmath145 , by corollary [ c:1 ] and since the only terminal position is @xmath123 , @xmath186 and thus , the two convergents of the @xmath13-positions of this game are @xmath187 and @xmath188 , whereas for @xmath189-rn the single convergent is @xmath190 .
see also figure [ f3 ] .
this game has a particular interest as being a new simple ` restriction ' of the game @xmath191gdwn in @xcite , the latter which is an extension of wythoff nim where the new moves are of the form : in one and the same move remove @xmath74 tokens from one of the heaps and @xmath192 from the other , at most a whole heap .
the game @xmath191gdwn is conjectured to split the upper ( and lower ) @xmath13-positions of wythoff nim from the single convergent @xmath187 to a pair of convergents @xmath193 and @xmath194 .
see also @xcite for another example where a further split of the @xmath13-positions of wythoff nim is obtained via a certain _ blocking maneuver _ on the regular wythoff nim moves .
-positions for the games @xmath189-rn and @xmath189-rw respectively.,title="fig:",scaledwidth=45.0% ] -positions for the games @xmath189-rn and @xmath189-rw respectively.,title="fig:",scaledwidth=45.0% ] let us finish off this paper with some extensions of our sample game in figure [ f:1 ] , that ( unlike the game in figure [ f:2 ] ) are not @xmath65-extensions , but indeed @xmath50-subtraction games .
our final three extensions of @xmath163-rn are as follows : adjoin moves of the form + game ( a ) : @xmath195 , + game ( b ) : @xmath196 , + game ( c ) : @xmath197 , + for positive integers @xmath74 , bearing in mind that the positions satisfy ( [ 2 ] ) .
the first two extensions have complicated patterns of @xmath13-positions which we do not yet understand , of which at least the first appears to exhibit a similar splitting ` behavior ' as does @xmath163-rw , but game ( c ) has the same set of @xmath13-positions as does @xmath163-rn , which can be proved by elementary methods , see also @xcite for related results .
hence the answer to the first question in this final section is negative ; see also @xcite for a discussion of this question in the context of heap games , computational complexity and algorithmic undecidability .
the purpose of this paper has been to introduce the subject of @xmath50-subtraction games and resolve the most basic question , that is of @xmath49-equivalence of @xmath65-subtraction games .
as we indicate in figures [ f:4 ] and [ f:5 ] , the most interesting `` new '' games are the generic @xmath50-subtraction games .
( for example , could a study of extensions of @xmath198 lead to new revelations of the behavior of @xmath191gdwn ? ) of course one would like to extend the notion of @xmath49-equivalence to games on several heaps , which can be done essentially by applying the same ideas as in section [ s:1 ] and [ s:2 ] .
we do not know of any literature on @xmath50-subtraction games even on two heaps .
+ * acknowledgment . *
i thank ragnar freij for an interesting discussion regarding generalizations to several heaps . | we study variations of classical combinatorial games on two finite heaps of tokens , a.k.a . _
subtraction games_. given non - negative integers @xmath0 , where @xmath1 , @xmath2 and @xmath3 , two players alternate in removing @xmath4 tokens from the respective heaps , where the allowed ordered pairs of non - negative integers are given by a certain move set @xmath5 .
there is a restriction imposed on the allowed heap sizes @xmath6 , they must satisfy @xmath7 and @xmath8 . a player who can not move loses and
the other player wins . for a certain restriction of these games , namely where each allowed move option @xmath9 is of the form @xmath10 , for some ordered pair of non - negative integers @xmath11
, we show that all games have equivalent outcomes via a certain surjective map to a canonical subtraction game . other interests in our games
are various interactions with classical combinatorial games such as _ nim and wythoff nim_. |
we consider a @xmath9 cubic crystal with a scalar displacement field @xmath10 defined on each lattice site @xmath11 where @xmath12 and @xmath13 .
the hamiltonian is taken to be of the fermi - pasta - ulam ( fpu ) form : h&= & _ + _ , [ ( x_-x_+)^2 + ( x_-x_+)^4 ] , + [ ham ] where @xmath14 denotes unit vectors in the three directions . we have set the values of all masses and harmonic spring constants to one and the anharmonicity parameter is @xmath15 .
two of the faces of the crystal , namely those at @xmath16 and @xmath17 , are coupled to white noise langevin type heat baths so that the equations of motion of the particles are given by : _ & = & - _ + & + & _ n_1,1 ( -_+^l _ ) + _ n_1,n ( -_+^r _ ) .
[ eqm ] the noise terms at different sites are uncorrelated while at a given site the noise strength is specified by @xmath18 , where @xmath19 and @xmath20 are the temperatures of the left and right baths and we have chosen units where the boltzmann constant @xmath21 . fixed boundary conditions were used for the particles connected to the baths and periodic boundary conditions were imposed in all the other directions .
we simulate these equations using a velocity - verlet algorithm @xcite and calculate the heat current and the temperature profile in the nonequilibrium steady state of the crystal . the heat current @xmath22 from the lattice site @xmath23 to @xmath24 where @xmath25 , is given by @xmath26 , with @xmath27 being the force on the particle at site @xmath24 due to the particle at site @xmath23 . in our simulations
we calculate the average current per bond given by j= _ n_1=1^n-1 _ n_2,n_3=1^w j _ .
we also calculate the average temperature across layers in the slab and this is given by @xmath28 .
in all our simulations we set @xmath29 and @xmath30 . we first address the question of the dependence of @xmath31 on the width @xmath1 of the system and the nature of the cross - over from @xmath32 behaviour , for small values of the ratio @xmath33 , to true @xmath9 behavior for @xmath34 .
the numerical results are given in fig .
( [ crossover ] ) .
we see that for any fixed length @xmath0 , the value of @xmath31 decreases as we increase @xmath1 but saturates quickly to the @xmath9 value .
the cross - over width @xmath35 is seen to increase slowly with @xmath0 .
the inset shows that as we increase @xmath0 , the cross - over from @xmath32 to @xmath9 behavior takes place at decreasing values of @xmath36 and presumably in the thermodynamic limit @xmath37 , the cross - over occurs at @xmath38 .
thus our study suggests that @xmath39 with @xmath40 .
a similar result was obtained by grassberger and yang @xcite for a @xmath41 fpu system .
next we look at the dependence of @xmath31 on @xmath0 for the @xmath9 case .
the fast cross - over from @xmath32 to @xmath9 behaviour implies that we can extrapolate the results for small @xmath36 to estimate the true value of the @xmath9 current ( at @xmath42 ) .
thus we can get results for quite large values of @xmath0 from simulations on systems with small widths . for sizes up to @xmath43 we obtained data for @xmath44 .
for the largest system size , namely @xmath45 we have data for @xmath46 .
we show our results for the @xmath0 dependence of @xmath2 in fig .
( [ jvsn ] ) .
there are three sources of error in the values of current : ( i ) numerical errors , arising from the finite time discretization value ( @xmath47 ) , and from rounding off errors ; ( ii ) statistical errors arising from averaging over a finite number of time steps ; and ( iii ) errors arising from the extrapolation of the small aspect ratio ( @xmath36 ) results to the @xmath9 case . the error from ( iii )
was taken to be the difference in current values for the two largest widths studied . for smaller system sizes we verified that the numerical error was much smaller than the statistical and extrapolation errors and
we assume that this is true also at larger system sizes .
the error - bar for each data point plotted in fig .
( [ jvsn ] ) is the larger of errors from ( ii ) and ( iii ) . versus width @xmath1 for different fixed values of the length @xmath0 .
the inset plots @xmath31 versus the aspect ratio @xmath33.,width=307 ] -versus-@xmath0 in different dimensions .
the inset shows the running slope @xmath48 as a function of @xmath0 .
the dashed line is a guide to the eyes .
, width=307 ] the slope of the @xmath2 versus @xmath0 curve is decreasing slowly with @xmath0 and a straight line fit to the last three points gives an exponent @xmath49 .
for comparison we also show in fig .
( [ jvsn ] ) the @xmath32 and @xmath41 data for the fpu system .
the @xmath41 results are from data for @xmath50 samples for systems up to @xmath51 while for larger sizes the results shown are extrapolated values from small width data . in @xmath32
we get @xmath52 @xcite while in @xmath41 we get @xmath53 . in the inset of fig .
( [ jvsn ] ) we have plotted the running slope defined as @xmath48 against system size . from this we see that while the slopes in @xmath32 and @xmath41 tend to saturate , the @xmath9 slope seems to be decreasing .
the @xmath9 slope can be fitted by the dashed line with a power law form .
this suggests that the asymptotic system size behaviour will give @xmath54 implying diffusive transport and validity of fourier s law .
one of the remarkable features of @xmath32 systems with anomalous heat transport is the form of the steady state temperature profile obtained in these systems .
typically one finds that the temperature profile is concave upwards in part of the system and concave downwards elsewhere and this is true even for small temperature differences @xcite .
this means that the temperature gradient is non - monotonic as a function of distance across the sample . in fig.([temp_pro ] ) we plot the temperature profiles for the @xmath32 , @xmath41 and @xmath9 samples .
we see that the variation of the temperature gradients are non - monotonic in both @xmath32 and @xmath41 while in @xmath9 they are monotonic .
the inset in fig .
( [ temp_pro])shows that the @xmath9 temperature profile is concave upward everywhere .
we have also confirmed that the profile becomes more linear on decreasing the temperature difference between @xmath19 and @xmath20 .
this again supports our finding based on the size - dependence of the current , that heat transport in @xmath9 is diffusive while in lower dimensions it is anomalous .
finally we look at the temperature dependence of thermal conductivity . temperature and nonlinearity
are highly correlated @xcite , and temperature dependence can be understood from the nonlinearity dependence of thermal conductivity .
we note that eq .
( [ eqm ] ) leads to the scaling relation @xmath55 , where @xmath56 , @xmath57 , and @xmath58 is an arbitrary scale factor .
taking the limit @xmath59 , this gives the scaling relation for thermal conductivity as @xmath60 .
putting @xmath61 and @xmath62 , we then get ( t , ) = ( t , 1 ) [ conductivity_scaling ] .
thus the thermal conductivity is a function of @xmath63 .
one may expect that large @xmath15 suppresses heat currents due to enhancement of phonon - phonon interactions .
hence from the scaling ( [ conductivity_scaling ] ) we expect that @xmath2 must also decrease with increasing @xmath64 . to check this , we show the dependence of the heat current on @xmath63 for a @xmath65 system with a small temperature difference @xmath66 . in fig.([temp_depe ] ) , we compared two cases : one with @xmath67 fixed and @xmath64 varied , and another with @xmath68 fixed and @xmath15 varied .
we find that current decreases as a function of @xmath63 , consistent with the scaling relation eq .
( [ conductivity_scaling ] ) .
we note that fourier s law ( [ fourier ] ) leads to @xmath69 and so the decrease of @xmath2 in the region @xmath70 $ ] with @xmath67 is consistent with the concave curve in the @xmath9 temperature profiles .
interestingly at large anharmonicity the current does not seem to go to zero but instead appears to saturate to a constant value . at low temperatures
the effect of anharmonicity becomes weaker and we expect the conductivity to increase , eventually diverging in the limit @xmath71 .
it is difficult to numerically access the low temperature regime since the mean free path becomes large and one would need much larger system sizes to see diffusive behaviour .
in summary , we have given the first numerical evidence for the validity of fourier s law of heat conduction in an anharmonic crystal in three dimensions .
this confirms the belief that in three dimensions anharmonicity is a _ sufficient _ condition for normal transport .
this is not a necessary condition since , for example , a @xmath9 pinned disordered purely harmonic crystal also shows normal transport @xcite .
our conclusion was based on three evidences .
the first is the system - size dependence of the thermal conductivity , the second is temperature profile , and the third is the consistency between temperature profile and temperature dependence of conductivity .
it has been known that the one - dimensional fpu system shows slow convergence of the thermal conductivity to it s asymptotic behavior @xcite . here
we show that this is also the case in @xmath9 .
unlike @xmath32 and @xmath41 , the running slope of the size dependence of @xmath2 in @xmath9 showed decreasing behavior even at the largest system size and this gives us a clear signature for finite @xmath2 .
the temperature profiles in @xmath9 are completely different type from the @xmath32 and @xmath41 case where nonmonotonic behavior of the gradient is robust even for small temperature differences .
we note that a recent simulation of heat conduction in the @xmath9 fpu crystals reported diverging thermal conductivity ( the reported exponent is about @xmath72 ) @xcite .
the reasons for this is probably because of the small values of anharmonicity used in those simulations and also the much smaller system sizes that were studied ( maximum size in that study was @xmath73 ) . in @xmath41
we find a divergence of the conductivity with an exponent @xmath53 which is similar to the value obtained in @xcite . for a sample of fixed length @xmath0
we find that the current density decreases on increasing its width @xmath1 and the cross - over from @xmath32 to @xmath9 behaviour takes place at a value @xmath74 with @xmath40 .
this has implications for experiments measuring thermal conductivity of nanowires @xcite .
if the cross - over width were independent of @xmath0 and the width of the nanowire larger than it , then the thermal conductivity of long nanowires could well be finite and not diverge as expected for true @xmath32 systems .
on the other hand , since the cross - over width gradually increases with increasing @xmath0 , a gradual transition from @xmath9-like to @xmath32 behavior will take place when the cross - over width is comparable to the width of nanowires .
this scenario is an interesting system size effect that may be observed in experiments on nanowires .
system , a @xmath50 @xmath41 system and a @xmath75 @xmath9 system with different aspect ratios @xmath33 .
temperature profiles for the three aspect ratios overlap with each other .
the inset shows that the @xmath9 temperature profile is concave upward everywhere.,width=288 ] ) for @xmath65 system with @xmath66 .
thermal conductivities decrease as increasing temperature @xmath64 or nonlinearity @xmath15 .
this temperature dependence of @xmath2 explains the slightly concave curve in temperature profiles in @xmath9 ( see fig.([temp_pro])).,width=288 ] | we perform nonequilibrium simulations of heat conduction in a three dimensional anharmonic lattice . by studying slabs of length @xmath0 and width @xmath1
, we examine the cross - over from one - dimensional to three dimensional behavior of the thermal conductivity @xmath2 .
we find that for large @xmath0 , the cross - over takes place at a small value of the aspect ratio @xmath3 . > from our numerical data we conclude that the three dimensional system has a finite non - diverging @xmath2 and thus provide the first verification of fourier s law in a system without pinning .
macroscopic behavior of heat transport in the linear response regime is governed by fourier s law |j & = & -| t(|x ) , [ fourier ] where @xmath4 , @xmath5 are respectively the heat current density and temperature gradient at the position @xmath6 , and @xmath2 is the thermal conductivity .
this implies diffusive behavior of heat .
what are the necessary and sufficient conditions for the validity of fourier s law ?
this question is a longstanding unsolved problem @xcite . for solids one
starts with the description in terms of a harmonic crystal where heat conduction takes place through lattice vibrations or phonons .
scattering of the phonons can occur due to phonon - phonon interactions ( _ i.e _ anharmonicity in the interactions ) or by impurities ( e.g isotopic disorder , defects ) @xcite .
for one dimensional systems , from a large number of numerical and analytical studies it is now established that these scattering mechanisms are insufficient in ensuring normal diffusive transport .
instead one finds anomalous transport @xcite , one of the main signatures of this being that the thermal conductivity @xmath2 in such systems is no longer an intrinsic material property but depends on the linear size @xmath0 of the system . a power law dependence @xmath7
is typically observed .
for two dimensional anharmonic crystals a @xmath8 divergence of the conductivity is predicted from various analytical theories @xcite and also from an exactly solved stochastic model @xcite , but the numerical evidence for this so far is inconclusive @xcite . a recent experiment has reported the breakdown of fourier s law in nanotubes @xcite while another experiment on graphene flakes @xcite also indicates a divergence of @xmath2 . for systems with pinning ( an external substrate potential ) and anharmonicity , fourier s law has been verified in simulations on one and two dimensional systems @xcite .
there is a strong belief that fourier s law should be valid in three dimensional ( @xmath9 ) systems , even without pinning .
a recent work @xcite examined heat transport in a @xmath9 disordered harmonic crystal .
analytical arguments showed that heat conduction in the system was sensitive to boundary conditions .
for generic boundary conditions a finite conductivity was predicted but this could be numerically verified only for the pinned case . in this letter
we investigate the effect of anharmonicity on heat conduction in ordered crystals . through extensive simulations of a @xmath9 anharmonic crystal
we give strong numerical evidence for normal transport and the validity of fourier s law in this system . |
an optical lattice is the result of the interference of one or several pairs of laser beams to produce a stationary wave , felt by a neutral atom as an external potential of the form @xcite : @xmath0 this expression , that depends on the laser wavelengths ( @xmath1 ) , defines a three dimensional lattice .
@xmath2,@xmath3 are the depths of the potential barriers felt by the atoms and can be modified by changing the light intensity . in a very common arrangement , two of those depths
are fixed to values high enough to produce a tight confinement in two directions .
we have then quasi - one dimensional systems well described by the continuous hamiltonian : @xmath4 + \sum_{i < j } v(r_{ij}).\ ] ] here , @xmath5 represents the interaction between the atoms @xmath6 and @xmath7 , separated a distance @xmath8 from each other , and @xmath9 is the mass of those atoms .
@xmath10 is the harmonic frequency we obtain at developing the first two terms of the right - hand side of eq .
( [ pot ] ) around @xmath11 and @xmath12 , a good approximation when @xmath13 and @xmath14 are deep enough .
when the interaction between particles is modeled by a hard spheres one , a system described by this hamiltonian at a density of one particle per potential well exhibits superfluid - mott insulator transitions @xcite for values of @xmath3 that depend on @xmath10 @xcite .
( [ hamiltonian ] ) does not consider the system to be harmonically trapped in the longitudinal direction , @xmath15 .
that would mean to include an additional term of the form @xmath16 in the hamiltonian @xcite .
disorder can be introduced on this simple arrangement .
two are the most common experimental routes : the use of two superimposed lasers of different ( and incommensurate ) wavelengths on the @xmath15 direction @xcite , or the introduction of a speckle field on the optical lattice @xcite . in both cases
, we end up having potential wells of unequal depths , with a quasi - periodic variation in the first case and a more randomly distribution of potential minima in the second .
what the phase diagrams of those arrangements have in common is the presence of at least a superfluid ( sf ) and a bose glass ( bg ) phase , depending on the particular experimental setup .
if the average number of atoms per potential well is an integer , a mott insulator ( mi ) can be also stable for some values of the lattice defining parameters @xcite . in this work , we considered the first disorder option , and solved the schrdinger equation corresponding to the continuous hamiltonian in eq .
( [ hamiltonian ] ) with an optical lattice potential of the form : @xmath17 where @xmath18 and @xmath19 are the two wavelengths that define the bichromatic lattice . here ,
e@xmath20 and e@xmath21 are the recoil energies corresponding to @xmath18 and @xmath19 ( e@xmath22= @xmath23 ) ) . since we chose to reproduce as closely as possible the experimental setup used in ref .
, we took @xmath18 = 830 nm , @xmath24 = 40 ( in units of e@xmath20 ) , and @xmath9 as the mass of a @xmath25rb atom .
there is a difference , tough .
we considered @xmath19 = 1067 nm instead of the 1076 nm originally used .
this makes the ratio @xmath26 = @xmath27 instead of the 0.77137 of the experimental data .
@xmath28 and @xmath29 where taken to be in the same range as the experimental data ( @xmath30 16 ; 0 @xmath31 2.5 , in units of their respective recoil energies ) . to use rational approximations to
the real wavelength ratio is a common practice ( see for instance ref . , in which @xmath26 = @xmath32 , or ref .
in which @xmath26 = @xmath33 ) , and allowed us to use relatively small simulation cells .
the interatomic potential was described by a hard - spheres ( hs ) interaction : @xmath34 for @xmath35 and 0 otherwise , @xmath36 being fixed to a single value , that of the scattering length of the @xmath25rb atoms @xcite .
this potential was previously used to describe bosons loaded in optical lattices @xcite . in most previous literature , the continuous hamiltonian of eq.([hamiltonian ] )
is approximated by the discrete bose - hubbard ( bh ) model @xcite : @xmath37 in which all the particles are allowed to stay only at the minima of their corresponding optical lattice potentials ( @xmath6 sites ) .
besides , only interactions of the type @xmath38 were considered , @xmath7 being one of the two nearest neighbors ( a quasi - one dimensional system translates into a pure one - dimensional hamiltonian ) to any particular @xmath6 .
@xmath39 is the bosonic creation ( annihilation ) operator at @xmath6 , and @xmath40 represents the number of neutral atoms at that same location .
@xmath41 , @xmath42 and @xmath43 are parameters that can be related meaningfully to the experimental ones ( @xmath3 , @xmath44 , and @xmath45 ) only when the optical lattice potentials are deep enough @xcite . for the kind of hamiltonians we are dealing with in this work @xcite : @xmath46 and
@xmath47 @xmath41 is the hopping matrix element between sites @xmath6 and @xmath7 , and @xmath42 models the repulsive interaction between two atoms located at the same @xmath6 site . to deduce the expression for @xmath42
, it was supposed that the full interaction potential between atoms in the original continuous hamiltonian was adequately described by a pseudopotential @xcite .
( [ j ] ) and ( [ u ] ) were deduced for monochromatic lattices , i.e. , lattices defined by a single laser of wavelength @xmath44 , and recoil energy @xmath48 .
if another laser is introduced , both expressions would have to be recalculated to include the effect of this second wavelength @xcite .
however , this is usually not done @xcite , even though a variation in @xmath41 could produce an appreciable change in the phase diagram .
@xcite the remaining parameter in eq.([hub ] ) , @xmath49 , represents the energy offset at each potential well , and for a monochromatic lattice with no longitudinal confinement is the same for each site .
this means that could be set to zero without loss of generality .
however , when disorder is introduced , this is no longer the case . for bichromatic lattices , @xmath50 $ ] @xcite , while for random disorder
, @xmath49 follows a uniform distribution in the interval @xmath51 $ ] or @xmath52 $ ] @xcite . in this work
, we calculate the phase diagram of the continuous hamiltonian defined by eq .
( [ hamiltonian ] ) with the optical lattice potential of eq .
( [ fa ] ) .
this implies that the disorder arises naturally at the introduction of the second laser , with no _ ad hoc _ change in the depths of the potential wells .
the only approximation was to use a hard spheres interparticle interaction , a very common approach @xcite .
we considered two densities , one particle per optical lattice minimum , and half a particle per minimum , and compared our results to those calculated with a bose - hubbard model for the same systems @xcite .
however , our primary goal will be to compare our results to the experimental ones , when available . since the lattice defined by @xmath18 is in most cases deeper than the defined by @xmath19 ( @xmath53 @xmath54 @xmath55 ) , to define the average density we used the number of potential minima corresponding to the lattice defined by @xmath18 .
this convention will not affect the limits of the phase diagram , since when @xmath56 , the system is always a superfluid , as we will see below .
the initial step to calculate the phase diagrams was to obtain the ground state of a set of atoms described by the hamiltonian of eq .
( [ hamiltonian ] ) .
this means that we are operating a t=0 , a reasonable approximation commonly used @xcite to describe the very low temperature regimes in which the experiments are performed @xcite . in any case
, we expect our description should be valid for temperatures lower than the corresponding to the excitations of the system , which are in the range of the khz @xcite . to do so , we used the diffusion monte carlo ( dmc ) method @xcite .
that technique needs an initial approximation to the ground state wavefunction to work , the so - called _ trial _ function .
that initial expression depends on the positions of the @xmath57 atoms in the simulation cell ( * r@xmath58*,@xmath59,*r@xmath60 * ) . in this work ,
we used : @xmath61 where @xmath62 are the coordinates of the atom @xmath6 and @xmath63 is the distance between atoms @xmath64 and @xmath9 .
we took @xmath65 to be a gaussian of variance @xmath66 , i.e. , the exact solution to the harmonic potential that traps the particles transversally .
@xmath67 was taken to be the numerical solution to an one - dimensional schrdinger equation including only the monochromatic lattice defined by @xmath18 .
the two body correlations were modeled by : @xcite : @xmath68 @xmath36 is here the scattering length of the pair of atoms .
the particular way to obtain the five constants in this equation has been detailed elsewhere @xcite . as indicated above
, we chose the @xmath28 and @xmath29 parameters in the same range that the experimental ones of ref . .
the main difference is that instead of limiting ourselves to the case @xmath69 = 16 , we considered the entire range 0 @xmath70 16 . for each @xmath69 we performed calculations in the interval 0
@xmath71 2.5 for a density of one atom per potential well and 0 @xmath71 2 when the average atom occupation was 0.5 . in most cases ,
a simulation cell including 36 potential minima (= 18 @xmath18 ) was used .
we checked that to increase that number to 54 , 72 or even 90 did not vary the results presented in this work .
a bose glass is usually defined as an insulating ( non - superfluid ) and finite - compressibility ( gapless ) phase @xcite . on the other hand ,
a mott insulator is a non - superfluid gaped state .
so , to determine whether a particular arrangement of atoms can be described as a glass , we should calculate the superfluid fraction ( to know if we are dealing with an insulator ) and the compressibility of the system ( to see if we have a mi or a bg ) .
we define the compressibility , @xmath72 , as : @xmath73 where @xmath74 is the chemical potential , @xmath75 here , @xmath76 is the total energy of the hard spheres arrangement loaded in the quasi - one dimensional optical lattice and @xmath77 is the density of the system ( @xmath78 , @xmath57 , number of atoms ; @xmath79 , length of the simulation cell ) .
this means that a finite @xmath72 at a given @xmath77 is associated with a continuous slope of the energy per atom versus the density at that @xmath77 . in fig .
[ fig1 ] , we represent @xmath80 , in units of the recoil energy of the primary lattice versus its filling fraction ( @xmath81/2 ) , subtracting from the energy the constant contribution per atom due to the transversal harmonic confinement , @xmath82 , in the same units . all the cases displayed correspond to @xmath69 = 16 and different values for @xmath29 : @xmath83 = 1.2 ( bottom ) , @xmath83 = 1.3 ( middle ) , and @xmath83 = 1.4 ( top ) . to see the influence of the size of the simulation cell in our results
, we display there the energies corresponding to simulation cells including 36 ( full symbols ) , 54 ( open symbols ) and 72 ( crosses ) potential minima . the results for bigger cells were similar and are not shown for simplicity .
the dashed and full lines are least - squares fits to the simulation results for densities below and above @xmath81/2 = 1 when only the data for 36 wells are considered .
the fact that all the symbols are basically on top of those fits , indicates that our energies are not affected by size effects .
the dotted line corresponds to the continuation of the result in the 0.85 @xmath84 @xmath81/2 @xmath84 1 range beyond that last density .
that line helps us to see that for @xmath83 = 1.3 , the slope of the energy versus the filling fraction is discontinuous at @xmath81/2 = 1 , something similar to what happens for @xmath83 = 1.2 . on the other hand , when @xmath83 = 1.4 there is no such a slope change , what indicates that for those parameters we have a gapless phase . using that information , and since we considered the smallest difference between consecutive @xmath83 values to be 0.1 , we estimated the upper limit of the mott insulator for @xmath69 = 16 to be @xmath83 = 1.3 @xmath85 0.1 .
this corresponds to the rightmost open symbol in fig .
[ fig2 ] . for larger @xmath83 values
, we have a gapless arrangement that can be either a superfluid or a bose glass . that value is in reasonable agreement with the experimental results of ref .
, in which we can see a disappearance of the mi somewhere between @xmath83 = 0.5 and @xmath83 = 1.5 .
the remaining open symbols in fig .
[ fig2 ] were obtained in the same way that the already described for @xmath69 = 16 and indicate the upper limit of stability of the mott insulator phase for each @xmath69 value . ) for @xmath28 = 16 in units of @xmath86 versus the filling fraction .
symbols correspond to the simulation results for different values of @xmath29 ( from bottom to top s@xmath87 = 1.2 , s@xmath87 = 1.3 and @xmath29 = 1.4 , all of them in units of @xmath88 ) .
full symbols correspond to simulation cells with 36 potential wells , open symbols to cells with 54 minima and crosses display the case for 72 wells .
lines correspond to least squares fits to the simulation data considering only the cell with 36 potential minima . , scaledwidth=44.0% ] and @xmath29 are the parameters defined in eq.([fa ] ) , in units of their respective recoil energies .
see further explanation in the text .
, scaledwidth=44.0% ] we calculated the superfluid fraction ( @xmath89/@xmath90 ) using a variant of the winding number technique @xcite .
the formal expression employed was : @xmath91 where @xmath92 .
@xmath93 indexes each monte carlo step of a given simulation , from the initial step after the monte carlo equilibration ( @xmath93= 0 ) to the end of the markov chain ( in our case , @xmath93 = 20000 ) .
@xmath94 is defined as : @xmath95 here , @xmath96 is the position of the center of mass of all the particles in the simulation cell at a step labeled by @xmath93 .
@xmath57 is , as before , the number of particles .
this definition means that if we represent @xmath97 as a function of @xmath93 , for large enough @xmath93 , we should get a straight line whose slope will give us the value of the superfluid fraction .
an example of that procedure is displayed in fig .
each curve represents the average of 50 independent dmc simulations .
there , @xmath83= 1.5 in all cases , while @xmath28 goes from @xmath69 = 16 ( bottom ) to @xmath69 = 5 ( top ) .
the solid line corresponds to the case of a perfect superfluid ( @xmath97 = @xmath98 ) , and it is included as a reference .
the error bars are displayed only for the case @xmath69 = 5 , are fairly typical , and correspond to an standard deviation of the data produced in those 50 independent simulations .
what we can see is that the slope of @xmath97 for large values of @xmath93 ( @xmath54 10000 ) increases from zero ( @xmath69 = 16,12 ) , to values that imply the existence of superfluids ( @xmath69 = 10,8,5 ) . .
all cases correspond to @xmath83 = 1.5 and different values of @xmath69 . from bottom to top : @xmath28 = 16,12,10,8,5 .
the solid line gives the slope corresponding to @xmath89/@xmath90 = 1 .
the error bars are only displayed in one case for simplicity .
, scaledwidth=44.0% ] from the slopes of curves similar to those in fig .
[ fig3 ] , and with the help of eq .
( [ super1 ] ) , we can deduce the superfluid fraction for the set of parameters we are interested in .
[ fig4 ] shows those results for different @xmath69 s as a function of @xmath83 .
the errors bars for each @xmath89/@xmath90 were calculated by the same method as in ref .
: least - squares fitting procedures were performed for the average values of @xmath97 as a function of @xmath93 , and for the same values after the subtraction ( or addition ) of the corresponding limits given by the error bars of that ratio .
this allows us to obtain a range of most probable slopes , and from that , a superfluid fraction interval .
what we observe is that , except for the @xmath69 = 5 case , there is a range of @xmath83 values above which the system is an insulator , and below which we have a superfluid .
we consider an arrangement to be non - superfluid when the error bar for the superfluid fraction includes the value @xmath89/@xmath90 = 0 .
so , for instance , when @xmath69 = 12 ( full squares in fig .
[ fig4 ] ) , the last insulator corresponds to @xmath83= 1.6 , since in that case @xmath89/@xmath90= 0.01 @xmath85 0.01 .
this translates into the point @xmath69= 12 , @xmath83 = 1.6 in fig .
[ fig2 ] . for larger @xmath83 s
, we have a superfluid , and when 1.3 @xmath99 , the phase is a gapless insulator .
i.e. a bose glass .
the remaining of the full symbols in fig .
[ fig2 ] have been obtained in a similar way for other @xmath69 values .
the error bars displayed there have the same origin as the ones for the mi - bg limit .
the full symbols in fig .
[ fig2 ] are related to the superfluid behavior of the system ( see below ) .
( in units of @xmath88 ) for different @xmath28 values ( in units of @xmath86 ) .
open squares , @xmath69 = 16 ; full squares , @xmath69 = 12 ; open circles , @xmath69 = 8 ; full circles , @xmath69= 5 .
the error bars are only displayed in the last two cases for simplicity , the ones not shown being of the same size ( @xmath100 0.01 ) . ,
scaledwidth=44.0% ] exactly the same calculations can be performed for an average density of half a particle per potential well .
the only difference is that in the phase diagram we will have only superfluid and bose glass phases , since a mi is only possible for a filling fraction of one .
[ fig5 ] is the counterpart of fig .
[ fig4 ] for this lower density .
there , we displayed the superfluid fraction as a function of @xmath83 for @xmath69 = 8 ( upper curve ) and @xmath69 = 10 ( lower curve ) , both sets of calculations including their corresponding error bars for each point .
we can see that for @xmath69 = 8 , the system is always a superfluid , irrespectively of the value of @xmath83 . on the other hand , for @xmath69 = 10
, an insulator ( bose glass ) appears for @xmath101 0.9 .
this last value is represented in fig .
[ fig6 ] as the leftmost full symbol .
additional calculations indicate that for this density no bose glass appears when @xmath102 , at least for @xmath83 s in the range displayed in fig .
[ fig6 ] . for larger @xmath69 s
, the limits between the superfluid and insulator phases are also shown as full symbols .
the upper dashed line is a least - squares fit of those simulation results to a function of the form @xmath103 ( @xmath104 , constants , @xmath29 and @xmath28 in units of their respective recoil energies ) , i.e. , to a function with an asymptote for @xmath69 = 8 ( vertical dotted line ) , and it is intended simply as a guide - to - the - eye . on the other hand , the open symbols in fig .
[ fig6 ] correspond to the values obtained for a bose - hubbard hamiltonian that models the same system @xcite .
for instance , the rightmost open symbol of fig .
[ fig6 ] correspond to the largest @xmath105 value considered in ref . , ( @xmath105 = 10 ) , for which the limit between the superfluid and the bose glass phases is located at @xmath106 = 4 . using eqs . [ j ] and [ u ] above
, we have that this translates into @xmath69 = 6.1 and @xmath83 = 0.4 .
a similar procedure was used to obtain the remaining open symbols shown there , the only cases whose transformations give coordinates in the @xmath83 range displayed in fig .
the dashed line on top of them is again a least - squares fit to be used only as a guide .
what we observe is that both in our simulations and those of ref .
predict a bose glass for high enough @xmath83 and @xmath69 , even though the bose - hubbard hamiltonian underestimates the values of those parameters needed to create an insulator . , but for a density of half a particle per potential well .
full squares , @xmath69 = 10 ; open squares , @xmath69 = 8 . ,
scaledwidth=44.0% ] , but for an average density of half a particle per potential well .
the full symbols correspond to our simulation results while the open symbols are from ref . .
all the lines displayed are guides - to - the - eye .
see further explanation in the text .
, scaledwidth=44.0% ]
in this work , we described the behavior of a set of neutral atoms loaded in a quasi - one dimensional bichromatic optical lattice . to do so
, we solved the schrdinger equation corresponding to the continuous hamiltonian in eq .
( [ hamiltonian ] ) , with an external potential defined by eq .
the only approximation employed , besides the use of a stochastic method to obtain the ground state , was the description of the two - body interaction by a hard - spheres potential . within this framework
, we obtained the phase diagrams of figs .
[ fig2 ] and [ fig6 ] , as functions of the real parameters at two different atom densities .
this makes straightforward the comparison between theory and experiment .
in particular , when @xmath69 = 16 , our value for the mi - bg limit was @xmath83=1.3 @xmath85 0.1 , a value within the experimental range ( 0.5 @xmath1071.5 ) @xcite . however , the standard approximation to describe an optical lattice is the use of a bose - hubbard model . in a classic case of putting the cart before the horse ,
the fact that a set of atoms loaded in an monochromatic optical lattice is a very good experimental realization of a bh hamiltonian for _ some _ values of the lattice defining parameters @xcite , has been used as an excuse to describe _ any _ optical lattice in _ any _ circumstances by this model .
the main advantage of this approach is that the bh model is relatively easy to use and its behavior is well understood ; the main problem being that it is only an approximation that breaks down even for monochromatic lattices with shallow potential wells , as it has been established both experimentally @xcite and theoretically @xcite for the description of the superfluid - mott insulator transition .
in addition , the study of a bichromatic lattice implies corrections for both the @xmath41 and @xmath42 parameters that would include the effect of the second standing wave and that there are usually not considered .
this means that even in those cases in which the primary lattice is deep enough to be well described by a bh hamiltonian , the description of the lattice is not exact .
that compounded error would be the cause of the differences between the phase diagram for half a particle per potential well calculated with a continous and a bh hamiltonian shown in fig .
unfortunately , to our knowledge , there are no experimental data to compare those diagrams to .
we can only say that the stability zone of the superfluid phase is reduced with respect to the case of a continuous hamiltonian , the same general trend found for monochromatic lattices @xcite .
the situation is similar for a density of one particle per potential well .
as indicated above , for @xmath69 = 16 , there is a reasonable agreement between the position of the mi - bg transition predicted from our calculations and the experiment .
regretably , we can not compare any bh predictions to that value , since @xmath28 = 16 translates ( using eqs .
( [ j ] ) and ( [ u ] ) , with the scattering length taken from ref .
@xcite ) into @xmath108 130 , a larger value than the ones considered in the literature @xcite . on the other hand , the upper stability limit for the mi phase in ref . for @xmath109
( larger @xmath105 used ) corresponds to @xmath110 10 .
this means @xmath111 6.1 and @xmath112 1.1 , to be compared to @xmath69 = 6 , @xmath83 = 0.6 @xmath85 0.2 , from fig .
thus , we have that a bh description produces a shrinking of the sf stability zone with respect to the case of a continuous model , as in the case of the lower density described above .
this is corroborated by the onset of the sf - mi transition for @xmath29 = 0 .
the value in this work ( @xmath69 = 5 @xmath85 1 ) is larger than the one deduced in ref .
( @xmath111 3 ) . to conclude
, we should make a last remark about the general form of the phase diagram displayed in fig .
there , we can see that for some values of @xmath69 ( @xmath113 10 ) , an increasing in @xmath83 makes the system change from a mi to a bg and further up , to a superfluid .
however , when @xmath69 is smaller , there is only a single transition , mi @xmath114 sf .
this is in apparent contradiction to some previous theoretical results @xcite for a bose - hubbard model , that indicate that between a mi and a sf phases , there should be always a bg `` sandwich '' .
at least part of the discrepancy can be explained by the fact that those works correspond to systems with random site disorder , which behave differently than arrangements with quasiperiodic bichromatic lattices . in any case , the results of fig .
[ fig2 ] indicate that if that intermediate bg exists for @xmath115 10 , the width of its stability interval should be very narrow . | we calculated the phase diagram of a continuous system of hard spheres loaded in a quasi - one dimensional bichromatic optical lattice .
the wavelengths of both lattice - defining lasers were chosen to model an incommensurate arrangement .
densities of one particle and half a particle per potential well were considered .
our results can be compared directly to those of the experimental system [ fallani et al .
prl , * 98 * 130404 ( 2007 ) ] from which our initial parameters were taken .
the phase diagrams for both densities are significatively different to those obtained by describing the same experimental setup with a bose - hubbard model . |
in this paper we consider the following statistical problem : upon observing a high - dimensional vector , one is interested in detecting the presence of a sparse , possibly structured , correlated subset of components of the vector .
such problems emerge naturally in numerous scenarios . the setting is closely related to gaussian signal detection in gaussian white noise , on which there is an extensive literature surveyed in @xcite . in image processing ,
textures are modeled via markov random fields @xcite , so that detecting a textured object hidden in gaussian white noise amounts to finding an area in the image where the pixel values are correlated .
similar situations arise in remote sensing based on a variety of hardware .
a related task is the detection of space time correlations in multivariate time series , with potential applications to finance @xcite .
we investigate the possibilities and limitations in problems of detecting correlations in a gaussian framework .
we may formulate this as a general hypothesis testing problem as follows .
an @xmath0-dimensional gaussian vector @xmath1 is observed . under the null hypothesis @xmath2 ,
the vector @xmath3 is standard normal , that is , with zero mean vector and identity covariance matrix . to describe the alternative hypothesis @xmath4 ,
let @xmath5 be a class of subsets of @xmath6 , each of size @xmath7 , indexing the possible `` contaminated '' components .
one wishes to test whether there exists an @xmath8 such that @xmath9 where @xmath10 is a given parameter .
equivalently , if @xmath11 denotes the vector of observations , then @xmath12 where @xmath13 denotes the @xmath14 identity matrix and @xmath15 we write @xmath16 for the probability under @xmath2 ( i.e. , the standard normal measure in @xmath17 ) and , for each @xmath18 , @xmath19 for the measure of @xmath20 .
the goal of this paper is to understand for what values of the parameters @xmath21 reliable testing is possible .
this , of course , depends crucially on the size and structure of the subset class @xmath5 .
we consider the following two prototypical classes : * _ @xmath7-intervals .
_ in this example , we consider the class of all intervals of size @xmath7 of the form @xmath22 modulo @xmath0for aesthetic reasons .
( we call such an interval a _ @xmath7-interval_. ) this class is the flagship of _ parametric _ classes , typical of the class of objects of interest in signal processing . * _ @xmath7-sets . _ in this example , we consider the class of all sets of size @xmath7 , that is , of the form @xmath23 where the indices are all distinct in @xmath6 .
( we call such a set a _ @xmath7-set_. ) this class is the flagship of _ nonparametric _ classes , and may arise in multiple comparison situations .
our theory , however , applies more generally to other classes , such as : * _ @xmath7-hypercubes .
_ in this example , the variables are indexed by the @xmath24-dimensional lattice , that is , @xmath25 , so that the sample size is @xmath26 , and we consider the class of all hyper - rectangles of the form @xmath27each interval modulo @xmath28of fixed size @xmath29 .
this class is the simplest model for objects to be detected in images ( mostly @xmath30 in applications ) . *
_ perfect matchings .
_ suppose @xmath0 is a perfect square with @xmath31 .
the components of the observed vector @xmath3 correspond to edges of the complete bipartite graph on @xmath32 vertices and each set in @xmath5 corresponds to the edges of a perfect matching .
thus , @xmath33 . in this example @xmath5 has a nontrivial combinatorial structure . *
_ spanning trees .
_ in another example , @xmath34 and the components of @xmath3 correspond to the edges of a complete graph @xmath35 on @xmath36 vertices and every element of @xmath5 is a spanning tree of @xmath35 . as usual , a _ test _ is a binary - valued function @xmath37 . if @xmath38 , then the test accepts the null hypothesis @xmath2 ; otherwise @xmath2 is rejected by @xmath39 .
we measure the performance of a test based on its _ worst - case risk _ over the class of interest @xmath5 , formally defined by @xmath40 we will derive upper and lower bounds on the _ minimax risk _ @xmath41 a standard way of obtaining lower bounds for the minimax risk is by putting a prior on the class @xmath5 and obtaining a lower bound on the corresponding _ bayesian risk _ , which never exceeds the worst - case risk .
because this is true for any prior , the idea is to find one that is hardest ( often called _ least favorable _ ) .
most classes we consider here are invariant under some group action : @xmath7-intervals are invariant under translation and @xmath7-sets are invariant under permutation .
invariance considerations ( @xcite , section 8.4 ) lead us to considering the uniform prior on @xmath5 , giving rise to the following _ average risk _ : @xmath42 where @xmath43 and @xmath44 is the cardinality of @xmath5 .
the advantage of considering the average risk over the worst - case risk is that we know an optimal test for the former , which , by the neyman
pearson fundamental lemma , is the likelihood ratio test , denoted @xmath45 . introducing @xmath46 for all @xmath8 , the likelihood ratio between @xmath2 and @xmath4
may be written as @xmath47 and the optimal test becomes @xmath48 note that @xmath49 .
the ( average ) risk @xmath50 of the optimal test is called the _ bayes risk _ and it satisfies @xmath51 note that , with the only exception of the case of spanning trees , in all examples mentioned above , the minimax and bayes risks coincide , that is , @xmath52 .
this is again due to invariance ( @xcite , section 8.4 ) .
( the class of spanning trees is not sufficiently symmetric for this equality to hold .
however , as we will see below , even in this case , @xmath53 and @xmath54 are of the same order of magnitude . )
we focus on the case when @xmath0 is large and formulate some of the results in an asymptotic language with @xmath55 though in all cases explicit nonasymptotic inequalities are available . of course
, such asymptotic statements only make sense if we define a sequence of integers @xmath56 and classes @xmath57 .
this dependency in @xmath0 will be left implicit . in this asymptotic setting , we say that _ reliable _ detection is possible ( resp . , impossible ) if @xmath58 ( resp . ,
@xmath59 ) as @xmath60 . in this paper
we assume that , under the alternative hypothesis , the correlation between any two variables in the `` contaminated '' set is the same .
while this model has a natural interpretation ( see lemma [ lemrepresent ] below ) , it is clearly a restrictive assumption .
this simplification is in understanding the fundamental limits of detection ( i.e. , in obtaining lower bounds on the risk ) . at the same time
, the tests we exhibit also match these lower bounds under more general correlation structures , such as @xmath61 that said , dealing with more general correlation structures remains an interesting and important challenge , relevant in the detection of textured objects in textured background , for example .
the vast majority of the literature on detection is concerned with the detection of a signal in additive ( often gaussian ) noise , which would correspond here to an alternative where @xmath62 for @xmath63 , where @xmath64 is the ( per - coordinate ) signal amplitude .
we call this the _ detection - of - means _ setting .
the literature on this problem is quite comprehensive .
indeed , the detection of @xmath7-intervals and @xmath7-hypercubes is treated extensively in a number of papers ; see , for example , @xcite . a more general framework that includes
the detection of perfect matchings and spanning trees is investigated in @xcite , and the detection of @xmath7-sets is studied in @xcite . in the literature on detection of parametric objects , the phrase `` correlation detection '' usually refers to the method of _ matched filters _ , which consists of correlating the observed signal with signals of interest .
this is not the problem we are interested in here . while the problem of _ detection - of - correlations _ considered here is mathematically more challenging than the detection - of - means setting ,
there is a close relationship between the two . the connection is established by the representation theorem of @xcite stated here for the case gaussian random variables . [ lemrepresent ]
let @xmath65 be standard normal with @xmath66 for @xmath67 .
then there are i.i.d .
standard normal random variables , denoted @xmath68 , such that @xmath69 for all @xmath70 .
thus , given @xmath71 , the problem becomes that of detecting a subset of variables with nonzero mean ( equal to @xmath72 ) and with a variance equal to @xmath73 ( instead of 1 ) .
this simple observation will be very useful to us later on .
when @xmath71 is random , the setting is similar to that of detecting a gaussian process ( here equal to @xmath72 for @xmath74 , and equal to 0 otherwise ) in additive gaussian noise .
however , the typical setting assumes that the gaussian process affects all parts of the signal @xcite . in our setting
, the signal ( the subset of correlated variables ) will be sparse .
since we only have one instance of the signal @xmath3 , the problem can not be considered from the perspective of either multivariate statistics or multivariate time series .
if indeed we had multiple copies of @xmath3 , we could draw inspiration from the literature on the estimation of sparse correlation matrices @xcite , from the literature on multivariate time series @xcite , or on other approaches @xcite ; but this is not the case as we only observe @xmath3 .
closer in spirit to our goal of detecting correlations in a single vector of observation is the paper of @xcite , which aims at testing whether a gaussian random field is i.i.d . or has some markov dependency structure .
their setting models communication networks and is not directly related to ours .
it transpires , therefore , that @xmath75 in the detection - of - correlations setting plays a role analogous to @xmath76 in the detection - of - means setting .
while this is true to a certain extent , the picture is quite a bit more subtle .
the detection - of - means problem for parametric classes such as @xmath7-intervals is well understood . in such cases , @xmath76 needs to be of order at least @xmath77 for reliable detection of @xmath7-intervals to be possible .
this remains true in the detection - of - correlations setting , and the _ generalized likelihood ratio test _ ( _ glrt _ ) is near - optimal , just as in the detection - of - means problem ; see , for example , @xcite .
our inspiration for considering @xmath7-sets comes from the line of research on the detection of sparse gaussian mixtures .
very precise results are known on @xmath78 that make detection possible @xcite and optimal tests have been developed , such as the `` higher criticism '' @xcite .
in fact , the recent paper @xcite deals with heteroscedastic instances of the detection - of - means problem where the variance of the anomalous variables may be different from 1 .
for example , it is known that , when @xmath79 [ resp . ,
@xmath80 , @xmath76 needs to be of order at least @xmath81 [ resp .
, @xmath82 for reliable detection of @xmath7-sets to be possible , and the test based on @xmath83 ( resp . , @xmath84 ) is near - optimal .
though more precise results are available when @xmath85 , these can not be translated immediately to our case via the representation theorem of lemma [ lemrepresent ] . as a bonus ,
we show that the glrt is clearly suboptimal in some regimes see theorem [ thmglrt - bad ] .
note that in the detection - of - means problem it is not known whether the glrt has any power .
this paper contains a collection of positive and negative results about the detection - of - correlation problem described above . in section [ seclower ]
we derive lower bounds for the bayes risk .
the usual route of bounding the variance of the likelihood ratio , that is very successful in the detection - of - means problem , leads essentially nowhere in our case .
instead , we develop a new approach based on lemma [ lemrepresent ] .
we establish a general lower bound for the bayes risk in terms of the moment generating function of the size of the overlap of two randomly chosen elements of the class @xmath5 .
this quantity also plays a crucial role in the detection - of - means setting and we are able to use inequalities worked out in the literature in various examples . in section [ secupper ]
we study the performance of some simple and natural tests such as the squared - sum test based on @xmath86 , the generalized likelihood ratio test ( glrt ) and a goodness - of - fit ( gof ) test , as well as some variants .
we show that , in the case of parametric classes such as @xmath7-intervals and @xmath7-hypercubes , the glrt is essentially optimal .
the squared - sum test is shown to be essentially optimal in the case of @xmath7-sets when @xmath87 is large , while the glrt is clearly suboptimal in this regime .
this is an interesting example where the glrt fails miserably .
when @xmath87 is small , detection is only possible when @xmath75 is very close to @xmath88 .
we show that a simple gof test is near - optimal in this case .
the analysis of tests such as the squared - sum test and the glrt involves handling quadratic forms in @xmath3 .
this is technically more challenging than the analogous problem for the detection - of - means setting in which only linear functions of @xmath3 appear ( which are normal random variables ) .
in this section we investigate lower bounds on the risk , which are sometimes called information bounds .
first we consider the special case when @xmath5 contains only one element as this example will serve as a benchmark for other examples .
then we consider the standard method based on bounding the variance of the likelihood ratio under the null hypothesis , and show that it leads nowhere .
we then develop a new bound based on lemma [ lemrepresent ] that has powerful implications , leading to fairly sharp bounds in a number of examples . as a warm - up , and to gain insight into the problem ,
consider first the simplest case where @xmath5 contains just one set , say @xmath90 . in this case , the alternative hypothesis is simple and the likelihood ratio ( neyman pearson ) test may be expressed by @xmath91 this follows by the fact that @xmath92 which is easy to check by straightforward calculation .
the next simple lemma helps understand the behavior of the bayes risk .
[ lemqf ] under @xmath16 , @xmath93 is distributed as @xmath94 and under the alternative @xmath19 , it has the same distribution as @xmath95 where @xmath96 and @xmath97 denote independent @xmath98 random variables with degrees of freedom @xmath88 and @xmath99 , respectively .
if @xmath100 denotes a standard normal vector , then under @xmath2 , the quadratic form @xmath93 is distributed as @xmath101 , and under the alternative , it has the distribution of @xmath102 , since @xmath3 is distributed as @xmath103 .
now , observe that for any symmetric matrix @xmath104 with eigenvalues @xmath105 , the quadratic form @xmath106 has distribution @xmath107 this follows simply by diagonalizing @xmath104 and using the rotational invariance of the standard normal distribution .
the lemma follows from this simple representation and the fact that @xmath108 has eigenvalue @xmath73 with multiplicity @xmath99 , @xmath109 with multiplicity @xmath88 , and the eigenvalue @xmath88 with multiplicity @xmath110 .
now it is straightforward to analyze the bayes risk . in particular
, we immediately have the following : [ prpsimple ] if @xmath5 is a singleton , @xmath111 if and only if @xmath112 .
similarly , @xmath113 if and only if @xmath114 .
suppose @xmath112 .
it suffices to show that there exists a threshold @xmath115 such that @xmath116 and @xmath117 .
we use lemma [ lemqf ] and the fact that , by chebyshev s inequality , @xmath118 for any sequence @xmath119 , and the fact that @xmath120 we choose @xmath121 and define @xmath122 . then under the null , @xmath123 and under the alternative , setting @xmath124 , @xmath125 we then conclude with the fact that , for @xmath7 large enough , @xmath126 . if @xmath127 is bounded , the densities of the test statistic under both hypotheses have a significant overlap and the risk can not converge to @xmath128 .
the proof of the second statement is similar .
clearly , the role of @xmath0 is immaterial in this specific example as the optimal test ignores all components whose indices are not in @xmath129 . when the class @xmath5 contains more than one element , the likelihood ratio with uniform prior on @xmath5 is given by ( [ l ] ) . a common approach for deriving a lower bound on the bayes risk is via an upper bound on the _ variance _ of @xmath130 under the null
indeed , by the cauchy
schwarz inequality , @xmath131 - 1}}{2}.\ ] ] therefore , an upper bound on @xmath132 -1 = { \operatorname{var}}_0(l(x))$ ] leads to a lower bound on @xmath53 .
let @xmath133 , which is independent of @xmath8 . by fubini s theorem
, we have @xmath134 where @xmath135 is defined in ( [ z ] ) .
we focus on terms of the double sum for which @xmath136 .
the following result is a straightforward consequence of the representation ( [ sum - chi ] ) and the well - known expression for the moment generating function of @xmath96 .
[ lemchi - eig ] suppose @xmath3 is a standard normal vector in @xmath137 and @xmath138 is an @xmath14 symmetric matrix with eigenvalues strictly less than @xmath139
. then @xmath140 if @xmath138 has an eigenvalue exceeding @xmath139 , then @xmath141 . since @xmath142 has eigenvalue @xmath143 with multiplicity @xmath7 , eigenvalue @xmath144 with multiplicity @xmath88 , and eigenvalue 0 with multiplicity @xmath145 , @xmath146 = { { \mathbb{e}}}_0 \exp(x^t { \mathbf{m}}x ) = + \infty$ ] unless @xmath147 .
the implications are rather insubstantial .
it only shows that , when @xmath148 with @xmath149 fixed , the bayes risk does not tend to zero .
as we shall see , this lower bound is grossly suboptimal , except in the case where @xmath5 is a singleton ( as in section [ secsimple ] ) or does not grow in size with @xmath0 .
a refinement of this method consists in bounding the first and second _ truncated _ moments of @xmath130 , again under the null hypothesis .
for example , this is the approach used in @xcite in the detection - of - means setting for the case of @xmath7-sets to obtain sharp bounds .
unfortunately , in our case this method only provides a useful bound when the class @xmath5 is not too large ( i.e. , has size polynomial in @xmath7 ) while it does not seem to lead anywhere in the case of @xmath7-sets .
the computations are quite involved and we do not provide details here , as we were able to obtain a more powerful general bound that applies to both @xmath7-intervals and @xmath7-sets .
this is presented in the next section . in this section
we derive a general lower bound for the bayes risk . as in the detection - of -
means problem @xcite , the relevant measure of complexity is in terms of the moment generating function of the size of the overlap of two randomly chosen elements of @xmath5 . in the detection - of - means setting ,
this is a consequence of bounding the variance of the likelihood ratio .
we saw in section [ secmoment ] that this method is useless here .
instead , we make a connection between the two problems using lemma [ lemrepresent ] .
[ thmlower ] for any class @xmath5 and any @xmath150 , @xmath151 where @xmath152 and @xmath153 , with @xmath154 drawn independently , uniformly at random from @xmath5 . in particular , taking @xmath155 , @xmath156 where @xmath157 .
the starting point of the proof is lemma [ lemrepresent ] , is as described in distribution . ] which enables us to represent the vector @xmath3 as @xmath158 where @xmath159 are independent standard normal random variables .
we consider now the alternative @xmath160 , defined as the alternative @xmath4 given .
let @xmath161 , @xmath162 , @xmath45 [ resp .
, @xmath163 , @xmath164 , @xmath165 be the risk of a test @xmath39 , the likelihood ratio , and the optimal ( likelihood ratio ) test , for @xmath2 versus @xmath4 [ resp .
, @xmath2 versus @xmath160 ] .
for any @xmath167 , @xmath168 , by the optimality of @xmath169 for @xmath2 versus @xmath160
. therefore , conditioning on @xmath71 , @xmath170 [ @xmath171 is the expectation with respect to @xmath172 . ]
using the fact that @xmath173 for all @xmath174 , we have @xmath175 } { { \mathbb{e}}}_0 @xmath176 } { { \mathbb{e}}}_0 \\ & \geq & { { \mathbb{p}}}\{|u|\le a\}\biggl(1 - \frac12 \max_{u \in[-a , a ] } \sqrt{{{\mathbb{e}}}_0
l_u^2(x ) - 1}\biggr).\end{aligned}\ ] ] since @xmath177 we get @xmath178 it is easy to check that @xmath179 which implies @xmath180 which concludes the proof .
we now apply theorem [ thmlower ] to a few examples .
the theorem converts the problem into a purely combinatorial question and @xcite offers various estimates for the moment generating function of @xmath181 which we may use for our purposes .
consider first the simplest case when @xmath5 contains @xmath182 disjoint sets of size @xmath7 .
[ cordisjoint ] let @xmath5 be the class of all sets of size @xmath7 . if @xmath183 then the bayes risk satisfies @xmath184 , and @xmath185 if @xmath186 or if @xmath187 .
clearly , the size @xmath181 of the overlap of two randomly chosen elements of @xmath5 equals zero with probability @xmath188 and @xmath7 with probability @xmath189 .
thus , @xmath190 which is bounded by @xmath88 if @xmath191 .
the first part then follows from the second part of theorem [ thmlower ] . for the second part ,
we need to find @xmath192 such that @xmath193 .
( note that in this case the upper bound above tends to zero . )
first assume that @xmath194 . in that case , @xmath195 , so it suffices to take @xmath192 slowly enough that @xmath196 .
next assume that @xmath197 . in this case , we have @xmath198 , and we simply choose @xmath192 slowly enough that @xmath199 .
consider the class of all @xmath7-intervals .
the situation is similar to that of nonoverlapping sets .
( in fact , since this class of @xmath7-intervals contains @xmath200 $ ] nonoverlapping sets of size @xmath7 , we could immediately deduce a lower bound via corollary [ cordisjoint ] . )
[ corkint ]
let @xmath5 be the class of all @xmath7-intervals . if @xmath201 then the bayes risk satisfies @xmath184 , and @xmath185 if @xmath202 or if @xmath203 . for two @xmath7-intervals chosen independently and uniformly at random , @xmath204 thus , @xmath205 and proceed as in the proof of corollary [ cordisjoint ] , using the fact that @xmath206 .
consider the class of all sets of size @xmath7 .
[ corksets ] let @xmath5 be the class of @xmath7-sets .
if @xmath207 then the bayes risk satisfies @xmath184 , and @xmath185 if either @xmath208 and @xmath209 , or @xmath210 . by @xcite ,
proposition 3.4 , which uses negative association , @xmath211 where the last expression is bounded by @xmath212 under the postulated condition , and tends to 1 if either @xmath208 and @xmath213 , or @xmath214 and @xmath215 .
first assume that @xmath208 and @xmath209 . by choosing @xmath192 slowly enough that @xmath216 we ensure that @xmath217 .
next assume that @xmath218 .
since @xmath219 , it suffices to take @xmath192 slowly enough that @xmath199 to ensure that @xmath220 .
the result then follows from theorem [ thmlower ] .
consider now the example of perfect matchings described in the . here
once again , theorem [ thmlower ] applies and implies that testing is impossible for moderate values of @xmath75 .
[ cormatch ] let @xmath5 be the class of all perfect matchings .
if @xmath222 , the bayes risk satisfies @xmath184 .
also , @xmath185 if @xmath223 .
the random variable @xmath181 for this class is considered by @xcite , who prove that @xmath224 this is bounded by @xmath212 whenever @xmath225 , which is satisfied whenever @xmath222 , and tends to 1 if @xmath226 .
we then apply theorem [ thmlower ] .
a similar argument applies for the class of all spanning trees of a complete graph with @xmath36 vertices [ and @xmath227 edges ] as described in the .
[ corsp ] let @xmath5 be the class of all spanning trees .
if @xmath228 , then the bayes risk satisfies @xmath229 .
we also have @xmath185 if @xmath223 .
it is shown in @xcite that @xmath230 which is bounded by @xmath231 whenever @xmath232 , which is satisfied whenever @xmath233 , and tends to 1 if @xmath226 .
we then apply theorem [ thmlower ] .
we already know that the likelihood ratio test is optimal in the bayesian setting .
we study here other tests for multiple reasons .
first , the likelihood ratio test seems difficult to compute in most situations .
second , the likelihood ratio test is heavily dependent on the prior we choose here , the uniform distribution on the class .
the third , and perhaps most important , reason is that it is difficult to obtain directly upper bounds for the ( worst - case ) risk of the likelihood ratio test whereas the tests considered below are easier to analyze and often yield near - optimal performance . whenever we obtain an upper bound for the risk of a test that matches the lower bounds developed in the previous section
, we have a full understanding of the limitations and possibilities of detection for the particular case considered , and this is our main goal in this paper .
we consider the squared - sum test , which corresponds to the anova test in the detection - of - means setting , the generalized likelihood ratio test ( glrt ) and a goodness - of - fit ( gof ) test , as well as some variants .
we say that a test is _ near - optimal _ for a certain setting if it achieves the information bound for that setting to first order .
one of the simplest tests is based on the observation that the magnitude of the squared - sum @xmath234 may be substantially different under the null and alternative hypotheses due to the higher correlation under the latter .
indeed , under @xmath16 , @xmath234 is distributed as @xmath235 , while for any @xmath236 with @xmath237 , under @xmath19 , @xmath234 has the same distribution as @xmath238 ; in fact , under the more general correlation model ( [ model - general ] ) , this is a ( stochastic ) lower bound .
this immediately leads to the following result .
[ prpsq ] let @xmath5 be an arbitrary class of sets of size @xmath7 and suppose that @xmath239 in ( [ model - general ] ) . if @xmath240 is such that @xmath241 but @xmath242 , then the test which rejects the null hypothesis if @xmath243 has a worst - case risk converging to zero .
however , any test based on @xmath234 is powerless if @xmath209 in ( [ model ] ) . in corollary [ corksets ] , we saw that reliable detection of @xmath7-sets is impossible if @xmath208 and @xmath209 .
here we see that , when @xmath239 , the squared - sum test is asymptotically powerful .
hence , the following statement : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the squared - sum test is near - optimal for detecting @xmath7-sets in the regime where @xmath208 .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ on the other hand , in the regime @xmath214 , the squared - sum test is powerless even if @xmath244 .
the test does not require knowledge of @xmath75 , though knowing @xmath75 allows one to choose the threshold @xmath240 in an optimal fashion ; if @xmath75 is unknown , we simply choose @xmath245 very slowly . in this section
we investigate the performance of the generalized likelihood ratio test ( glrt ) .
we show that for parametric classes such as @xmath7-intervals , the test is near - optimal .
however , for the nonparametric class of @xmath7-sets , the test performs poorly in some regimes . by definition
, the glrt rejects for large values of @xmath246 , or simply @xmath247 when all the sets in the class @xmath5 are of same size , since @xmath248 only depends on the size of @xmath249 .
hence , the glrt is of the form @xmath250 for some appropriately chosen @xmath251 .
we immediately notice that the glrt requires knowledge of @xmath75 our analysis of the glrt is based on lemma [ lemqf ] , which provides the distribution of the quadratic form @xmath93 under the null @xmath16 and under the alternative @xmath19 . under the null we need to control the maximum of such quadratic forms over @xmath8 , which we do using exponential concentration inequalities for chi - squared distributions . recalling corollary [ corkint ] , when detecting @xmath7-intervals all tests are asymptotically powerless when @xmath252 . we assume for concreteness that @xmath253 , for otherwise detecting @xmath7-intervals for very small @xmath7 has more to do with detecting @xmath7-sets .
we state a general result that applies for classes of small cardinality .
[ prpglrt - small ] consider a class @xmath5 of sets of size @xmath7 , with cardinality @xmath254 such that @xmath255 .
when @xmath256 , the generalized likelihood ratio test with threshold value @xmath257 has worst - case risk tending to zero .
we first bound the probability of type i error .
indeed , under the null , by lemma [ lemqf ] and its proof , we can decompose @xmath258 where @xmath259 and @xmath260 .
hence , @xmath261 it is well known that the maximum of @xmath182 standard normals is bounded by @xmath262 with probability tending to 1 as @xmath263 .
hence , the second term on the right - hand side is bounded by @xmath264 with high probability .
for the first term , we combine the union bound and chernoff s bound to obtain , for all @xmath265 , @xmath266\\[-8pt ] & \leq & n \exp\biggl(-\frac{(k-1)}2 ( a~- 1 - \log a)\biggr ) .
\nonumber\end{aligned}\ ] ] using the fact that @xmath267 when @xmath268 , the right - hand side tends to zero when @xmath269 .
we arrive at the conclusion that the glrt with threshold @xmath270 has probability of type i error tending to zero .
now consider the alternative under @xmath271 . by lemma [ lemqf ] and chebyshev s inequality , @xmath272 with high probability when @xmath273 .
we then conclude by the fact that the right - hand side is larger than @xmath251 when @xmath273 sufficiently slowly . comparing the performance of the glrt in proposition [ prpglrt - small ] with the lower bound for @xmath7-intervals in corollary [ corkint ] , we see that the glrt is near - optimal for detecting @xmath7-intervals .
this is actually the case for all parametric classes we know of .
consider now the example of the class of all @xmath7-sets .
compared to the previous section , the situation here is different in that @xmath182 , the size of the class @xmath5 , is much larger .
for example , for @xmath7-sets , @xmath274 , and therefore @xmath275 with @xmath60 .
the equivalent of proposition [ prpglrt - small ] for this regime is the following : [ prpglrt - large ] consider a class @xmath5 of sets of size @xmath7 , with cardinality @xmath263 such that @xmath276 . when @xmath277 , the generalized likelihood ratio test with threshold value @xmath278 has worst - case risk tending to zero .
we follow the proof of proposition [ prpglrt - small ] .
the only difference is in ( [ glrt - small - eq1 ] ) , where we now need @xmath279 and that right - hand side tends to zero when @xmath280 .
choose @xmath281 , obtaining that , with high probability , @xmath282 as before , with high probability under @xmath271 , @xmath283 so we only need to check that the threshold @xmath251 is larger than the right - hand side in ( [ glrt - large - eq1 ] ) and smaller than the right - hand side in ( [ glrt - large - eq2 ] ) , which is the case by the assumptions we made .
notice that in proposition [ prpglrt - large ] the condition on @xmath284 implies that @xmath285 , which is much stronger than what the squared - sum test requires when @xmath208 . for @xmath7-sets ,
@xmath274so that @xmath286and the requirement is that @xmath287 , which is substantially stronger than what the lower bound obtained in corollary [ corksets ] requires .
moreover , if we restrict @xmath284 to be bounded away from @xmath88 , then the glrt may be powerless .
[ thmglrt - bad ] let @xmath5 be the class of all @xmath7-sets .
if @xmath288 and @xmath289 , the glrt has a bayes risk bounded away from zero .
the proof is in the . in view of theorem
[ thmglrt - bad ] , the glrt is clearly suboptimal when in the situation stated there , and compares very poorly with the squared - sum test , which is asymptotically powerful if @xmath290 as seen in proposition [ prpsq ] .
we do not know of any other situation where the glrt fails so miserably .
while the glrt is near - optimal for detecting objects from a parametric class such as @xmath7-intervals , it needs knowledge of @xmath75 .
however , a simple modification solves this drawback .
indeed , consider the following `` local '' squared - sum test : @xmath291 for some appropriate threshold @xmath251 .
[ prplocal - sum ] consider a class @xmath5 of sets of size @xmath7 , with cardinality @xmath254 such that @xmath255 .
when @xmath292 in ( [ model - general ] ) , the local squared - sum test with threshold @xmath293 has worst - case risk tending to zero .
the proof is quite straightforward . indeed , under the null , for any @xmath249 of size @xmath7 we have @xmath294 so that @xmath295 with probability tending to 1 . under an alternative ( [ model - general ] )
, @xmath249 denoting the anomalous set of variables , we have @xmath296 when @xmath292 .
specializing this result to the case of @xmath7-intervals leads to the following statement ( which ignores logarithmic factors ) : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the localized squared - sum test is near - optimal for detecting @xmath7-intervals in the regime where @xmath297 .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ when @xmath7 is unknown .
_ we might only know that some interval is anomalous , without knowing the size of that interval . in that case , multiple testing at each @xmath7 using the local squared - sum test yields adaptivity .
computationally , this may be done effectively by computing sums in a multiscale fashion as advocated in @xcite .
in fact , here it is enough to compute the sums over all _ dyadic _ intervals since each interval @xmath249 contains a dyadic interval of length at least @xmath298and this can be done in @xmath299 flops in a recursive fashion . by now , the parametric case is essentially solved , with the local squared - sum test being not only near - optimal but also computable in polynomial time ( in @xmath0 and @xmath7 ) for the case of @xmath7-intervals , for example . in the nonparametric case , so far , the story is not complete .
we focus on the class of all @xmath7-sets . there we know that the squared - sum test is near - optimal if @xmath208 . if @xmath214 , it has no power , and we only know that the glrt works when @xmath300 , which does not match the rate obtained in corollary [ corksets ] .
worse than that , it is not clear whether computing the glrt is possible in time polynomial in @xmath301 . we now show that a simple goodness - of - fit ( gof ) test performs ( almost ) as desired . the basic idea is the following .
let @xmath302 , where @xmath303 is the standard normal distribution function . under the null ,
the @xmath304 s are i.i.d .
uniform in @xmath305 . under an alternative with anomalous set denoted by @xmath249 ,
the @xmath306 are closer together , especially since we place ourselves in the regime where @xmath285 .
more precisely , we have the following . [ lemclose ]
suppose @xmath65 are zero - mean , unit - variance random variables satisfying @xmath307 , for all @xmath308 .
let @xmath309 denote their average .
then for any @xmath310 , @xmath311 let @xmath312 .
elementary calculations show that @xmath313 = 1 -\frac1k -\frac{\lambda}{k^2 } \leq(1 -1/k)(1 - \rho ) \leq 1-\rho.\ ] ] by markov s inequality , we then have @xmath314 the statement follows from observing that @xmath315 the idea , therefore , is detecting unusually high concentrations of @xmath304 s , which is a form of gof test for the uniform distribution . under a general correlation model as in ( [ model - general ] ) , with lemma [ lemclose ] we see that the concentration will happen over an interval of length slightly larger than @xmath316 .
this is apparent from lemma [ lemrepresent ] under the simple correlation model ( [ model ] ) .
choose an integer @xmath28 such that @xmath317 and partition the interval @xmath318 $ ] into @xmath28 bins of length @xmath319 , denoted @xmath320 .
let @xmath321 be the bin counts thus , we are computing a histogram .
then consider the following gof test : @xmath322 where @xmath251 is some threshold .
[ prpgof ] consider the class @xmath5 of all @xmath7-sets in the case where @xmath323 and @xmath253 . in the gof test
above , choose @xmath28 such that @xmath324 . when @xmath325 in ( [ model - general ] ) , the resulting test with threshold @xmath326 has worst - case risk tending to zero .
bernstein s inequality , applied to the binomial distribution , gives that @xmath327.\ ] ] this and the union bound imply that , indeed , @xmath328 consider now an alternative of the form ( [ model - general ] ) , with @xmath249 denoting the anomalous set .
let @xmath329 though the set @xmath330 is random , by lemma [ lemclose ] and the fact that @xmath331 , we have that @xmath332 define the event @xmath333 for some @xmath150 .
note that , since the variance of @xmath334 is bounded by 1 , @xmath335 .
define @xmath336 .
on @xmath337 , using a simple taylor expansion , we have @xmath338 where @xmath339 denotes the standard normal density function and @xmath340 is taken sufficiently large . therefore ,
when @xmath341 and @xmath337 hold , at least @xmath342 of the anomalous @xmath304 s fall in an interval of length at most @xmath343 .
since such an interval is covered by at most @xmath344 bins , by the pigeonhole principle , there is a bin that contains @xmath345 anomalous @xmath304 s . by bernstein s inequality
, the same bin will also contain at least @xmath346 nonanomalous @xmath304 s ( with high probability ) , so in total this bin will contain @xmath347 points . by our choice of @xmath28 , @xmath348
, so it suffices to choose @xmath192 slowly enough that @xmath349 still .
then , with high probability , there is a bin with more than @xmath251 points . ignoring logarithmic factors , we are now able to state the following : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the gof test is near - optimal for detecting @xmath7-sets in the regime where @xmath214 and @xmath253 . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ when @xmath350 , things are somewhat different .
there , the gof test requires that @xmath351 , which is still close to optimal when @xmath352 , but far from optimal when @xmath7 is bounded ( e.g. , when @xmath353 , the exponent is 4 instead of 2 ) .
indeed , when @xmath350 , @xmath28 needs to be chosen larger than @xmath0 , and bernstein s inequality is not accurate .
instead , we use the simple bound @xmath354 note that bennett s inequality would also do . ( the analysis also requires some refinement showing that , with probability tending to 1 under the alternative , one cell contains at least @xmath7 points . ) note that in the remaining case , @xmath355 , the glrt is optimal up to a logarithmic factor , since it only requires that @xmath356 , as seen in section [ secglrt - nonparametric ] .
we do not know whether a comparable performance can be achieved by a test that does not have access to @xmath75 .
_ when @xmath7 is unknown .
_ in essence , we are trying to detect an interval with a higher mean in a poisson count setting .
as before , it is enough to look at dyadic intervals of all sizes , which can be done efficiently as explained earlier , following the multiscale ideas in @xcite .
the proof is divided into three steps .
the first step formalizes the fact that we want to prove that ( under @xmath4 ) , the contaminated set has no influence ( with high probability ) on the glrt statistic .
the second step exhibits a useful high probability event .
finally , in the third step we show that on this high probability event , the contaminated set has no influence on the glrt .
it can easily be seen that for every @xmath249 of size @xmath7 , @xmath357 introduce the function @xmath358 defined by @xmath359 for @xmath360 . denoting , for @xmath361 and @xmath236 , the vector of components of @xmath362 belonging to @xmath249 by @xmath363 , we may write the glrt as @xmath364 note that by the symmetry of @xmath5 and the test , @xmath365 given @xmath366 , define the coupling @xmath367 as follows : @xmath368 for @xmath369 , and @xmath370 are independent for @xmath371 .
note that @xmath372 . then ,
no matter what the threshold @xmath251 is , we have @xmath373 in the following we show that , with probability tending to @xmath88 , we have @xmath374 which then implies that the glrt is asymptotically powerless . by lemma [ lemrepresent ] , there exist @xmath68 independent standard normal such that for all @xmath371 , @xmath375 using the fact that @xmath376 with high probability , with probability tending to 1 , we have @xmath377,\ ] ] where @xmath378 and @xmath379 is any sequence such that @xmath380 .
fix @xmath381 to be determined later and define @xmath382 where @xmath172 . by the fact that @xmath383 are i.i.d .
standard normal , @xmath384 , so that @xmath385 if @xmath386 .
when @xmath387 is bounded away from @xmath88 , this is the case if @xmath388 . in conclusion
, we proved that the event @xmath389 has a probability that tends to @xmath88 if @xmath390 as long as @xmath387 is bounded away from 1 .
we specify @xmath391 .
note that , as required , @xmath387 exceeds and is bounded away from 1 .
assume that we are on the event @xmath392 .
first note that @xmath393\\[-8pt ] & = & k ( k-1 ) \zeta^2 ( 1 - \rho\gamma^2 ) , \nonumber\end{aligned}\ ] ] and the same holds for @xmath394 .
let @xmath8 be such that @xmath395 .
we want to show that there exists @xmath396 such that @xmath397 .
this entails that @xmath398 , since for @xmath399 we have @xmath400 .
first remark that we can assume that @xmath401 since otherwise by ( [ eqminordered ] ) we can simply take @xmath402 . to simplify notation , we may assume that @xmath403 . by definition of @xmath392 and the fact that @xmath249 contains at least one index in @xmath404
, there exist @xmath405 such that @xmath406 and @xmath407 do not appear in @xmath408 .
we want to show that by replacing @xmath409 by either @xmath406 or @xmath407 , in @xmath408 , one increases the value of @xmath410 .
more precisely , we want to show that @xmath411 then by induction one can show the existence of the @xmath396 described above .
note that , for @xmath412 and @xmath413 , @xmath414 consider the case where @xmath415 ( the case @xmath416 can be dealt with similarly ) . since @xmath417 , it suffices to show that @xmath418 , which follows from @xmath419 this concludes the proof .
we thank omiros papaspiliopoulos for his illuminating remarks and the anonymous referees for challenging us to obtain stronger results in the sparse setting and for pointing out a mistake in proposition [ prpgof ] . | we consider the hypothesis testing problem of deciding whether an observed high - dimensional vector has independent normal components or , alternatively , if it has a small subset of correlated components .
the correlated components may have a certain combinatorial structure known to the statistician .
we establish upper and lower bounds for the worst - case ( minimax ) risk in terms of the size of the correlated subset , the level of correlation , and the structure of the class of possibly correlated sets .
we show that some simple tests have near - optimal performance in many cases , while the generalized likelihood ratio test is suboptimal in some important cases . , . |
in order to study harmonic fields in two dimensions , very powerful techniques have been developed based on complex analysis@xcite . among these , we will use in the following conformal maps , which allow to relate the geometry we wish to study ( i.e. a semi - infinite domain limited by a rough interface ) , to a regular one as schematically illustrated in figure ( [ fig : method ] ) .
0.8 > as usual , we will identify a point in the plane @xmath3 , with the complex number @xmath4 .
we note @xmath5 the complex conjugate of @xmath6 . the two variables @xmath6 and @xmath7 can be treated as independent variables instead of @xmath8 and @xmath9 .
a mapping from the complex plane onto itself is simply defined as a complex function @xmath10 of @xmath6 and @xmath7 which transforms one point of the complex plane @xmath6 into another point @xmath11 .
for the mapping to be of physical interest it has to be bijective in a domain of interest , and thus inversible .
the mapping is _ conformal _ if the function is _ holomorphic _ , i.e. it only depends on @xmath6 and not on @xmath7 .
it can be shown that in this case , local angles are preserved in the transformation apart from singular points and hence the term `` conformal '' . moreover ,
the real part and the imaginary part of any such holomorphic function are both harmonic , i.e. @xmath12 . the latter property results from the expression of the laplacian operator in terms of the variables @xmath6 and @xmath7 : @xmath13 two obvious choices are well suited for our geometry : i ) a semi - infinite plane @xmath14 , @xmath15 ii ) the unit circle @xmath16 , @xmath17 .
these two domains can be related by the transformation @xmath18 , and thus they are basically equivalent . since the boundary we consider is periodic in the @xmath8-direction , the mapping to the unit circle is well suited . however , in the following , we will rather use the mapping to the half - plane , since it corresponds directly to the `` reference '' problem where the roughness vanishes .
the domain of interest , denoted by @xmath19 is limited by a rough interface @xmath20 which is a periodic function of @xmath21 of period @xmath22 . the conformal map , @xmath10 , is a function of @xmath6 which associate one point of the reference domain , @xmath14 or @xmath16 , to another point in @xmath19 . from now on in order to distinguish the initial and the image domain
, we will note @xmath23 a point in the image plane , and keep the notation @xmath4 for the initial plane unless otherwise mentioned . before specifying the particular form of the boundary ,
it is possible to guess an adequate form for these transformations .
let us first consider the mapping from the half - plane @xmath14 to @xmath24 . as @xmath25 tends to @xmath26 , the mapping should approach the identity , @xmath27 , since the roughness of the boundary is not expected to play any significant role at a large distance ( compared to the period @xmath22 ) from the boundary .
we introduce the function @xmath28 such that @xmath29 functions of the form @xmath30 with @xmath31 real , thus appears to be natural candidates for @xmath28 .
they are indeed periodic functions of @xmath32 , and vanish exponentially as @xmath25 goes to @xmath26 when @xmath33 . moreover , in order to satisfy the same periodicity as @xmath20 we require that @xmath34 where @xmath35 is an integer .
thus we propose as a writing of the transformation @xmath10 the following decomposition @xmath36 without loss of generality we will set @xmath37 for all the rest of this study .
the rough boundary is to be identified with the image of the @xmath8 axis , so that @xmath20 obeys the parametric equations @xmath38 the corresponding transformation from the unit disk @xmath16 to the domain @xmath19 can be obtained from the above form ( [ map ] ) and the tranformation from the disk to the semi - infinite plane @xmath39
. the resulting transformation reads @xmath40 where @xmath41 has been used . evidently , the image of the unit circle @xmath42 provides the same parametric form as ( [ boundary ] ) .
the form of the transformation being imposed , one needs to check that the transformation is bijective : a point should have a single image , and an image point a unique parent .
this condition does impose some restriction on the transformation @xmath10 .
it can be rephrased simply for the transformation ( [ map ] ) as @xmath43 for all @xmath44 . in principle , it is sufficient to impose this condition only in the strict interior of the domain .
if @xmath45 on the boundary , a kink may appear at this point . in the following , we will assume that the interface is smooth at a small scale so that poles are forbidden on the boundary .
as an example , if the transformation is simply @xmath46 then the condition ( [ inversible ] ) reduces to @xmath47 or @xmath48 . for this maximum value ,
the image of the @xmath8-axis is a cycloid , with cusp points .
figure ( [ fi : cycloid ] ) illustrates this limit case . 0.8
the above presented transformation is only useful for a particular application if the transformation @xmath10 can be computed , once the boundary @xmath20 is imposed .
this section is devoted to this problem .
the algorithm that we have developed generates the transformation very efficiently .
different numerical techniques applied to computing the map from arbitrary closed domains to the unit disk can be found in ref.@xcite .
our algorithm can be shown to be related to the jacobi method used in these studies .
we define the rough boundary @xmath20 as a single valued real function @xmath49 such that @xmath50 is the equation of the boundary . in other words ,
the boundary is given by the parametrised form @xmath51 . to comply with the framework we have chosen here
, we use in the following a 2@xmath52 -periodic @xmath49 . from the particular form of the transformation @xmath10 , we expand the real and imaginary parts on the boundary @xmath20 @xmath53 \\
h(u)= & \mathop{\im m}\left[\sum_k \omega_k e^{-ikx}\right ] \end{array } \right.\ ] ] if @xmath54 were equal to @xmath8 , the second equation would be close to a fourier transform expression of the function @xmath49 . more precisely rewriting the last equation as @xmath55\ ] ]
we see that the coefficient @xmath56 can be computed from the fourier transform of @xmath57 .
the difficulty is that @xmath58 is _ a priori _ unknown .
however , we note that if the roughness is small enough , say of order @xmath59 , @xmath54 can be written as @xmath60 .
therefore , identifying @xmath8 with @xmath54 is a zeroth order approximation . from the latter ,
the coefficient @xmath61 can be computed by the fourier transform of @xmath62 .
this provides a first order approximation of @xmath58 , from which an improved estimate of @xmath61 can be obtained , by taking the fourier transform of @xmath57 , i.e. a non - uniform sampling of the profile @xmath63 .
iterating this scheme is the basis of our algorithm .
we will omit for the time being the prerequisite on the amplitude of the roughness .
we will return to this point by considering the stability of the algorithm .
the intermediate quantities appearing at the @xmath31th iteration will be labelled with a superscript @xmath64 .
we also formulate the algorithm directly in discrete terms suited for a numerical implementation . in the remainder of this article
, all functions will be decomposed over a set of @xmath65 discrete values .
the number of fourier modes will thus be limited to @xmath65 .
we first introduce a series of sampling points @xmath66 with @xmath67 which is initially set to an arithmetic series @xmath68 .
the sampling of @xmath62 by the @xmath66 gives the array @xmath69 the discrete fourier transform of this array is the complex valued array @xmath70 for @xmath71 .
the latter is shortly written as @xmath72\ ] ] where @xmath73 denotes the fourier transform , which will be chosen as the fast fourier transform ( fft ) algorithm , thus imposing that @xmath35 is an integer power of 2 .
the intermediate mapping @xmath74 is computed from the @xmath75 as @xmath76 the latter form is obtained from the identification of eq.([eqboom]b ) and the definition of @xmath75 , taking care of the fact that one sum is over positive index , while the other extends over the interval @xmath77 $ ] .
then , one computes the series @xmath78 this linear transformation is shortly noted as @xmath79\ ] ] where @xmath80 is the above detailed transformation .
the form of @xmath80 is dictated by eq.([eqboom]a ) for positive index , and from the fact that the inverse fourier transform of @xmath81 ( see below ) is real .
the new sampling series is finally obtained from @xmath82 \label{eqalgend}\ ] ] the equations ( [ eqalgbeg]-[eqalgend ] ) define one step in the algorithm relating @xmath83 to @xmath74 .
briefly we note this step @xmath84 .
the searched function @xmath10 is clearly a fixed point of the transformation @xmath85 defined above in a discretized version .
the uniqueness of the transformation @xmath10 results from that of the harmonic field in the domain @xmath19 with an equipotential condition on the boundary and a constant gradient perpendicular to the boundary at infinite distance from it . therefore
, the only condition to consider is the stability of the fixed point .
let us assume that we have an approximate solution of the transformation @xmath86 , from which we compute the series @xmath87 .
all intermediate quantities computed from the exact solution are denoted by a superscript @xmath88 .
following one complete iteration of the algorithm , we obtain the following expressions @xmath89 \\
b = & b^*+{\cal g}[a - a^ * ] \\ ( \delta u)^{\prime}= & { \cal f}^{-1}[b - b^ * ] \end{array}\ ] ] where a taylor expansion of @xmath49 has been used to estimate the values of @xmath90 and where indices are omitted when unnecessary .
the resulting difference @xmath91 after one cycle is thus @xmath92\ ] ] let us introduce the norm @xmath93 parseval s theorem relates the above norm in real and fourier spaces according to @xmath94 in a similar fashion , the transformation @xmath80 does not affect the norm @xmath95 using the two previous results , we can estimate the norm of @xmath96 as @xmath97 therefore , if the absolute value of the slope of the objective profile satisfies , @xmath98 for all @xmath54 , then the fixed point @xmath99 is attractive for the transformation @xmath85 .
it should however be noted that the number of modes @xmath35 should be large enough so that the perturbation @xmath100 should be small enough to legitimate the taylor expansion of @xmath49 used in the stability analysis . in practice
, the convergence is very fast provided the sufficient condition ( [ eqcondcv ] ) is fulfilled .
moreover it has to be noted that one step in the algorithm requires a rather limited amount of computing time of order @xmath101 ( i.e. as for a fft operation ) . considering that this computation gives the solution of an harmonic problem in a semi - infinite domain
, this cost appears to be extremely low .
when our algorithm is applied to a simple monochromatic sine ( or cosine ) profile , @xmath102 , it turns out that as soon as the condition ( [ eqcondcv ] ) is violated ( i.e. @xmath103 ) the scheme is unstable , and a loop begins to appear around the origin where the slope exceeds 1 .
thus the sufficient condition is also a necessary condition .
the limit @xmath104 can simply be broken if one uses an under - relaxation scheme .
the optimum determination of the under - relaxation parameter , or the use of other algorithms can be found in ref.@xcite for mapping arbitrary domains on the unit disk .
the transposition of these algorithms to our problem can be worked out in details .
other ways to break this limit is to decompose the transformation @xmath105 in two ( or more ) substeps .
suppose one could map the real axis onto an intermediate profile using a first transformation @xmath106 and then the intermediate profile onto the objective one using a second transformation @xmath107 .
the combination of the two transformations @xmath108 is then the searched mapping . by breaking the problem into two steps , it is possible that each step can be handled by the above presented algorithm , while the combination of the two giving a profile having a slope larger than unity .
the difficulty here is to device a suited intermediate step .
one could consider for example filtering the initial objective profile so that the filtered profile may fulfill the slope constraint .
we did not investigate this extension any further .
we present in the following calculations of conformal transformations associated with a simple sine interface .
we will use a norm on the error similar to the one introduced in the previous subsection .
the distance @xmath109 between the objective profile and the calculated one is defined as follows : @xmath110 ^ 2 dx\ ] ] it is convenient to make this distance dimensionless , normalising it by the amplitude of the profile , @xmath111 where the objective profile has the equation @xmath102 .
it is worth noting that the problem is far from being as simple as it might appear on the surface . in real space
one single fourier mode is sufficient to entirely characterise the interface .
the transformation @xmath10 however requires much more modes .
figure ( [ fi : pssine ] ) shows the power spectrum of the @xmath112 series , for different amplitudes @xmath113 . 0.7
the convergence of the algorithm is shown at play on figure ( [ fi : cvsine ] ) , where the obtained profile obtained after the first few iterations are shown . in this particular example @xmath114 and
the number of modes is 32 .
0.7 the importance of allowing for enough fourier modes is also illustrated by considering the minimum error @xmath109 obtained as a function of @xmath35 as shown in figure ( [ fi : ersine ] ) .
( for this particular study we did not resort to a fft algorithm to handle any value of @xmath35 ) . for @xmath114
we observe that about 20 modes are necessary to reach the single precision used in the computation .
as the amplitude increases , the number of modes needed to reach a small enough error becomes larger and larger .
in the description of rough surfaces and interfaces , some recent progress has been achieved by recognising some scaling invariance properties which has been observed in a number of real surfaces , and has been shown to result naturally in a number of growth models .
recent reviews@xcite have covered this field . due to the different roles played by the directions normal and parallel to the surface , the scaling invariance when applicable
involves different scale factors depending on orientation , a property called self - affinity .
we consider here only two dimensional media so that the boundary is self - affine if it remains ( statistically ) invariant under the transformation @xmath115 for all values of @xmath116 .
the exponent @xmath117 is called the `` hurst '' or roughness exponent .
it is characteristic of the scaling invariance . from this property , we derive easily that @xmath118 where @xmath119 is a prefactor . it is noteworthy that the self - affinity property does not involve the scaling of any measure .
however , studying the scaling of the length of the curve , two regimes are revealed . for large distances , larger than a scale @xmath120 ,
the curvilinear length of the profile is simply proportional to the projected length along the @xmath8 axis , hence on can identify a trivial fractal dimension equal to 1 . on the other hand , for distances smaller than @xmath120 , the arc length scales in a non - trivial fashion with the projected length .
this allows to define a fractal dimension equal to @xmath121 .
the cross - over scale @xmath116 between these two regimes is such that the typical slope of the profile is 1 , i.e. using the notations of eq.([eqsa ] ) , @xmath122 once a roughness profile has been measured , a very convenient way @xcite to check the self - affinity is to compute the power spectral density ( psd ) of the profile . in the case of a self - affine profile of exponent @xmath117 , the psd is expected to have the following behaviour : @xmath123 it is important to stress that the approach developed in this article is not specific to self - affine boundaries .
however , being given the practical importance of such boundaries , and the expected generality of scaling results , we will essentially focus on self - affine boundaries as practical applications of the concepts developed in the framework of harmonic field in the vicinity of rough boundaries . in view of the form of the transformation @xmath10 , and of the previous scaling , eq.([eqsaps ] ) , we introduce a particular set of transformation : let us choose @xmath124 where @xmath125 are random gaussian variable with 0 mean and unit variance for the real and imagnary part independently , we can write @xmath126 = 1 + a. \mathop{\im m } \left [ \sum_k \epsilon_k k^{1/2 - \zeta } e^{-ikx } \right]\ ] ] then for a given set of @xmath125 , we can define a maximum amplitude such that the mapping is bijective : @xmath127}}}\ ] ] this method gives a short way to generate directly @xmath10 transforms which image the real axis to a self - affine interface shape .
this approach is useful to study generic properties of self - affine boundaries .
when the amplitude @xmath113 is small enough , @xmath128 , and thus the series @xmath61 is equal to the fourier transform of the profile .
the transformation @xmath10 sends the real axis onto a periodic function whose power - spectrum is of the form eq.([eqsaps ] ) .
when the amplitude increases , the first iteration of the algorithm turns out to be rather approximative . in order to show that the power spectrum of @xmath61 is not significantly altered by further steps , we show in figure ( [ fi : pssyntsa ] ) the power spectrum of @xmath61 as compared to the initial zeroth order approximation .
the result have been obtained from an average over 100 profiles with 2048 modes each .
we see that the synthetic generation of the transform does not modify the power spectrum of @xmath112 coefficients . 0.7 therefore
, we can directly generate mappings @xmath112 which will transform the real axis into a periodic boundary which is self - affine with any prescribed roughness exponent for distances smaller than the period .
such a construction may appear as artificial in the sense that the rough boundary is not imposed but on the contrary it results from the choice of the mapping .
it is however useful , as will be shown later , because it allows to study generic properties of harmonic fields close to self - affine boundaries .
the alternative way consists in using the mapping construction algorithm .
we have in a previous section analysed the convergence of the algorithm applied to the special case of a sine profile .
we now consider the case of a self - affine boundary in a similar fashion .
this interface has been calculated in the real space with 64 modes , and we have used 256 modes in the conformal transformation . the standard deviation of the height distribution is called @xmath129 .
the chosen @xmath117 exponent chosen for this example is @xmath130 . from eq.([eqampmax ] ) , we note that the maximum amplitude decreases as the number of modes @xmath35 increases .
this is natural since as the lower cut - off in the scaling regime decreases , the self - affine function will tend toward a continuous but non - differentiable curve when @xmath131 .
the distribution of local slopes is indeed expected to gets wider and wider as the number of modes increases .
quantitatively , @xmath132 .
it is to be noted that as @xmath35 increases , the standard deviation of the height @xmath129 does not increase .
it is to be noted that these conclusions are drawn under the hypothesis that the longest wavelength remains fixed , here set to @xmath133 . alternatively , if the smaller cut - off and the amplitude of the corresponding mode were kept constant while increasing the number of modes , then the maximum amplitude would remain constant . as in the previous example ( sine profile )
, we can observe on figure ( [ fi : exsyntsa ] ) an example of the conformal map obtained for @xmath134 , where the maximum standard deviation that could be handled by the algorithm without diverging is @xmath135 .
we can see on this figure that the major differences between the objective and calculated profiles occur in areas where the local slope is maximum . for roughness amplitude greater than the convergence threshold
, one can see loops appearing in these areas .
the convergence speed , the sensitivity to the number of modes allowed in the determination of @xmath10 , the evolution of the minimum error , ... behaved for these self - affine profiles in a similar qualitative way as for the simple sine profile . 0.7
in the following , we show that the knowledge of such a conformal transformation allows to solve immediatly harmonic problems .
we essentially focus here on the case where the field is assumed to be uniform far from the boundary .
this is a typical case as soon as the roughness is of small amplitude .
this can be seen as an asymptotic expansion focusing here on the small scale details of the interface , whereas the matching with the far field can be done using a field whose variation is small on the scale of the roughness amplitude .
we will now focus on two problems : perfectly conducting boundary so that the potential gradient is normal to the boundary , and perfectly insulating boundary where the potential gradient is parallel to the surface .
since we know how to taylor mappings which image the real axis on a generic self - affine boundary , this gives us a key to consider the scaling features of harmonic fields in the vicinity of self - affine boundaries .
harmonic fields are encountered very frequently in nature .
linear transport involving scalar fields @xmath136 , where the flux @xmath137 is proportional to the field gradient plus a conservation law in the absence of sources , and in steady conditions , @xmath138 , imply the harmonic nature of the field @xmath136 , @xmath139 .
heat diffusion obeying fourier s law , gives an harmonic temperature field in steady condition .
mass diffusion with fick s law is a similar example with the concentration field .
electric conduction with ohm s law , viscous flow in confined two dimensional hele - shaw cells , vorticity in stokes flow , ... constitutes a partial list of possible applications . in this part , for the sake of concreteness
, we use the case of thermal conduction .
we are interested in the temperature field @xmath140 in the region @xmath19 limited by the rough interface @xmath20 .
let us first consider the case of a perfectly conducting interface , so that @xmath141 for each point of the boundary .
we impose in the far field a homogeneous unit flux of heat .
the problem to solve is @xmath142 the knowledge of @xmath10 allows then to define @xmath143 , image field of @xmath140 in the smooth domain @xmath14 : @xmath144 .
as @xmath145 and @xmath146 in @xmath14 , the resolution of ( [ eqpblap ] ) in @xmath19 is thus equivalent to : @xmath147 the frontier @xmath148 being the real axis , we have immediately the solution in @xmath14 : @xmath149 and then the solution in @xmath19 is : @xmath150\ ] ] figure ( [ fi : isotherm ] ) shows a set of isotherm curves close to a self - affine isotherm boundary .
these lines become smoother and smoother when the distance to the electrode increases .
the morphology of these isotherm lines have some interesting features .
if @xmath151 denotes the distance to the boundary , one can observe from the form of the mapping @xmath10 that modes with a wavelength smaller than @xmath151 will be damped whereas longer wavelength modes will only be slightly decreased .
therefore , the isotherm curves will be similar to the profile up to a low pass filtering .
in the case of a self - affine boundary , the isotherms will preserve the self - affine character with the same exponent , but their lower cut - off will increase as the distance to the actual boundary , up to the distance of order of the largest wavelength . 0.7
let us now study the temperature gradient .
quite generally , we can write the gradient in the complex plane as @xmath152 from the expression of the temperature field , we have @xmath153 from the expression of the function @xmath112 , we see that at a large distance from the rough boundary , the term @xmath154 vanishes exponentially
. therefore , one recovers the imposed condition for the temperature at infinity i.e. @xmath155 . on figure ( [ fi : hgrat ] ) we have presented on the same graph the profile of a rough electrode and the modulus of the temperature gradient .
one may see quite easily that the field is very large ( small ) in the deepest ( highest ) areas .
this field depends naturally both on the local topography and on its remote environment .
the connection between the field and the local topography can be analysed through cross - correlations as will be done in a following section .
0.7 the perfectly insulating boundary is the other archetypical problem whose expression is @xmath156 the solution to this problem can simply be obtained from the previous using duality properties of the harmonic field .
the real part of the previous solution gives the answer to the problem .
@xmath157\ ] ] the temperature gradient is then simply @xmath158 we have seen previously that once we know the conformal mapping able to transplant the half complex plane @xmath159 onto the rough domain @xmath160 , we have immediatly the solution of the electrical potential near the rough electrode @xmath161 . if the roughness amplitude remains below the convergence threshold , we are now able to solve this problem for any kind of boundary . in practice , very often , one does not worry about the details of the rough interface .
as we have seen most perturbations die away from the boundary exponentially fast .
therefore , knowing the longest wavelength of the boundary gives the scale away from the boundary where the field becomes homogeneous .
this means practically , that if one is interested only in the far field , one could replace the rough interface by a straight one so that the far field is unperturbed .
the question we want to address in this section is the following : where should the `` equivalent '' straight interface be located so as to match the asymptotic far field ?
a zeroth order guess is to place it at the geometrical average of the height distribution .
this will be shown not be an exact answer , in the following , we call @xmath162 the distance between the equivalent position and the geometrical average . in order to illustrate the problem , let us imagine the following experiment .
let us consider an electrolytic bath , where the electrical field is homogeneous between two opposite electrodes @xmath113 and @xmath163 at a distance @xmath164 from each other .
the electrical resistance @xmath165 of the set - up is measured .
then , as illustrated in figure ( [ fi : elect ] ) we place in the middle of the bath and parallel to the electrodes a rough plane @xmath119 of negligible thickness which is a good conductor , so that it can be considered as an equipotential .
we measure again the electrical resistance of the set - up , which is now reduced to @xmath166 .
what is the value of @xmath167 ?
so as to answer this question , we part the system in two , @xmath168 and @xmath169 .
each of these two problems corresponds to the situation described in the introduction of this section . extrapolating the field from electrode @xmath113
, we find an offset @xmath170 .
similarly from @xmath163 we obtain a different offset @xmath171 , so that , ignoring the details of the perturbed field in the vicinity of @xmath119 , the rough electrode will appear as being equivalent to a plane electrode of thickness @xmath172 .
this `` electrical thickness '' has nothing to do with the real thickness of the plane considered here to be zero .
if the rough electrode has the shape of a sine function , of amplitude @xmath113 and wavelength @xmath116 we will argue below that @xmath173 . finally it is a simple matter to relate the resistance drop to this effective thickness through @xmath174 . 0.7 we now revert to the notation of the previous paragraph , and deal with the temperature instead of the voltage . for distances away from the rough boundary much greater than @xmath133 , ( our longest wavelength ) ,
all exponential terms die out , and hence the far - field can be written @xmath175 where @xmath176 is the constant term in the function @xmath112 .
the off - set position of the equivalent isotherm is thus @xmath177 let us first analyse the problem for a small amplitude sine boundary of amplitude @xmath113 and wavelength @xmath116 .
the offset @xmath162 in the location of the equivalent straight boundary is to be normalised by @xmath113 to obtain a dimensionless quantity .
the latter should be a function of the dimensionless ratio @xmath178 .
taylor expansion of this function provides the perturbation expansion @xmath179 a simple argument allows to simplify the latter equation .
suppose one would analyse the problem for the profile of amplitude @xmath180 .
the latter is obtained from the former by a translation along the @xmath8 axis by an amount @xmath181 .
thus @xmath162 should be unchanged .
this imposes that odd terms in the expansion should vanish , hence @xmath182 thus the dominant correction is of order @xmath183 .
it can be interpreted as the product of the amplitude @xmath113 and a typical slope @xmath184 .
this result holds in the limit of a small amplitude and long wavelength .
if the wavelength goes to zero , clearly the offset should converge to the amplitude , but the latter limit can not be obtained from the above taylor expansion in the small parameter @xmath178 . for a sine profile of small amplitude it is possible to carry out the computation of the coefficient @xmath185 .
we briefly sketch here the solution .
the potential is to be computed to second order in @xmath113 .
we revert as in the previous sections to a wavelength @xmath186 .
the solution reads @xmath187 \vspace{0.15 cm } \\ & & & \displaystyle + \mathop{\im m}\left [ i\frac{a^2}{2 } \right]+ { \cal o}(a^3 ) \end{array}\ ] ] the offset can be read from this equation as @xmath188 . reincorporating the @xmath189 dependence ,
we arrive at @xmath190 or @xmath191 this last result is of course only valid for small @xmath113 values , @xmath162 being bounded by @xmath113 . on figures ( [ fi : eqboh - a ] ) and ( [ fi : eqboh - l ] ) , we can see comparisons between this perturbative calculation and the result directly obtained by conformal transformation . we observe an excellent agreement for small amplitudes ( resp .
large wavelengths ) and then the perturbative calculation overestimates @xmath162 for larger values of the amplitude ( resp . smaller wavelength ) . in view of the upper bound on the offset and the above perturbation expansion , we propose the following form @xmath192 which fits the data accurately as can be seen on figs .
( [ fi : eqboh - a ] ) and ( [ fi : eqboh - l ] ) , and which reproduces both limiting behaviors @xmath193 and @xmath194 . 0.7 0.7 the question we ask is now how does the result translate to a rough profile ? in particular for a self - affine profile , there is no characteristic length scales apart from the cut - offs .
the product of the amplitude times the slope is a scale dependent factor .
is it possible to reach quantitative conclusions for such profiles ? in order to estimate @xmath162 for a rough boundary , we use the formalism developed for introducing the algorithm .
we expand the function @xmath195 as well as all other intermediate quantities in series of the profile amplitude . using the linearity of the transformations @xmath73 , @xmath80 and @xmath196 ,
we arrive at @xmath197=\displaystyle{\frac{\mathop{\re e}[a_0]}{2n } } = { \frac 1{2n}}\sum_jh_j^{*}\\ & \displaystyle = { \frac 1{2n}}\sum_jh^{\prime } ( u_j)\ { \cal f}^{(-1)}\circ { \cal g}\circ { \cal f}[h_j ] \end{array}\ ] ] up to third order terms in the amplitude .
we now need an asymmetric version of parseval s theorem .
let us compute the integral for two arbitrary arrays defined in real space for @xmath54 and fourier space for @xmath198 @xmath199_j= & \displaystyle{\frac 1{2n}}\sum_j\sum_ku_jv_ke^{-ikj } \\
= & \displaystyle{\frac 1{2n}}\sum_k\overline{{\cal f}[u]}_kv_k \end{array}\ ] ] the offset can now be expressed as @xmath200}_k\ { \cal g}\circ { \cal f}[h(x)]_k\ k \\ = & \displaystyle-{\frac 1{4n^2}}\left ( \sum_{k>0}\tilde h_k\overline{\tilde h_k}\ k-\sum_{k<0}\overline{\tilde h_k}\tilde h_k\ k\right ) \\
= & \displaystyle-{\frac 1{2n^2}}\sum_{k>0}|\tilde h(k)|^2\ k \end{array}\ ] ] where @xmath201 is the fourier transform of @xmath202 . figure ( [ fi : eqbosa ] ) gives the evolution of @xmath162 with the amplitude of self - affine profiles of roughness exponent @xmath203 , and 16 modes .
again , we observe that the above given expression ( [ eqhexpress ] ) is accurate for small amplitude , but shows deviations for larger amplitudes . 0.7 it is interesting to consider the scaling of @xmath162 observed from the generic transformations where @xmath204 are postulated to be @xmath205 .
the expectation value of the offset @xmath206 reads to dominant order in the amplitude @xmath207 where the extra factor of 2 comes from the expectation value of @xmath208 are independent gaussian variables of zero mean and unit variance . depending on the value of the roughness exponent @xmath209
two cases are to be distinguished .
@xmath210 for a _ `` persistent '' _ profile i.e. @xmath211 the sum in eq.([eqhsasca ] ) is dominated by the smallest @xmath212 , i.e. the longest wavelength , and thus the scaling of @xmath206 can be expressed as @xmath213 where @xmath214 is the riemann zeta function .
we have quitted momentarily the convention that the largest wavelength is @xmath215 , hence @xmath216 and @xmath217 are respectively the smallest and largest cut - off lengths in the profile . in this case , @xmath113 , the amplitude , is such that the largest wavelength mode amounts to @xmath218 .
let us introduce the standard deviation of the profile given by @xmath219 which leads ( using parseval s theorem ) to @xmath220 @xmath209similar to the scaling eq.([eqsa ] ) .
eq.([eqhsasca ] ) can then be expressed as @xmath221 the latter equation simply means that the rough profile behaves as a the simple monochromatic profile .
this conclusion is however not always valid as is shown in the following case .
@xmath210 for a _ `` anti - persistent '' _ profile i.e. @xmath222 the sum in eq.([eqhsasca ] ) is dominated by the largest @xmath212 , i.e. the shortest wavelength , in contrast to the previous persistent case .
@xmath223 therefore , we can express the scaling of @xmath162 in an intrinsic fashion as @xmath224 in contrast to the persistent case , it appears that the offset @xmath162 is dependent on the lower cut - off scale of the profile .
in fact if @xmath225 is kept fixed , the standard deviation @xmath226 grows as @xmath227 .
therefore , one sees that the upper scale cut - off disappears , so that @xmath162 only depends on @xmath216 . in order to see this more clearly
, we introduce another measure of the roughness which is sensitive to the small scale .
let @xmath228 be the norm of the derivative of @xmath49 : @xmath229 which amounts to @xmath230 from the latter norm , the offset can be written as @xmath231 which is the counterpart of eq.([eqpers ] ) for the antipersistent case . as a conclusion , the scaling of the offset @xmath162
is controlled by the shortest ( resp .
longest ) scale cut - off of the self - affine regime for anti - persistent ( resp .
persistent ) boundaries . in a preceeding section
, we have extracted the expression of the temperature gradient as a function of the transformation @xmath112 .
we now use it to investigate the correlations between the topography and the temperature gradient .
we study these correlations in the limit of a small amplitude .
the first order perturbation in the temperature gradient can be extracted from eq.([eqgrad ] ) as @xmath232 where the amplitude of the profile is asssumed to be of order @xmath59 .
we introduce the logarithm of the temperature gradient denoted @xmath233 which can be expressed as @xmath234+{\cal o}(\epsilon^2)\ ] ] from now on we will omit the @xmath235 term , keeping in mind that we focus here only on the dominant term . in order to compute the correlation between the gradient of temperature and the height
, we form the cross - product and average over @xmath54 ( or @xmath8 for convenience , since their difference is of order @xmath59 ) .
the expectation value of the product is @xmath236\mathop{\im m}[\omega(x)]\rangle = -2 \sum_k k\vert\omega_k\vert^2\ ] ] it is amazing that the same expression appeared when computing the offset of the equivalent straight boundary .
we define now the correlation coefficient @xmath237 which can be identified as the slope of a linear regression between @xmath49 and @xmath233 .
its value is @xmath238 since @xmath239 to first order in @xmath240 .
hence we have @xmath241 this expression holds for any rough boundary of small amplitude . in the particular case of a self - affine boundary ,
we assume as in the previous section that the transformation @xmath112 can be taken as the one generated artificially from its fourier decomposition .
the latter expression can thus be written as @xmath242 from the latter expression , we have to distinguish between persistent @xmath243 and anti - persistent @xmath244 profiles depending on whether the series is convergent or divergent when the number of modes increases .
@xmath210 in the case of a _ persistent _ self - affine boundary , as the number of modes increases to infinity , the value of @xmath245 converges toward an asymptotic limit @xmath246 given by a ratio of riemann zeta functions @xmath247 the divergence of the zeta function as its argument approaches 1 , leads to a divergence of @xmath246 as @xmath248 . in the more general case where @xmath217 is not set to @xmath215 , the above equation should be corrected to @xmath249 as a practical illustration of the latter property we have studied the correlations between @xmath49 and @xmath250 by averaging @xmath251 at fixed @xmath49 for 1000 profiles having the same characteristics : amplitude @xmath252 , roughness exponent @xmath203 and 64 fourier modes .
figure ( [ fi : corrstat ] ) shows the evolution of @xmath250 versus @xmath63 . from the eq.([eqriemann ] ) we estimate @xmath253 .
as shown on figure ( [ fi : corrstat ] ) , this value of @xmath246 provides an accurate fit to the data .
the evolution of this coefficient as a function of @xmath209 is shown in figure ( [ fi : corrslop ] ) .
0.7 0.7 @xmath210 the anti - persistent self - affine profile behaves differently from the previous case .
the correlation between the surface temperature gradient and the height vanishes .
mathematically , this result can be traced to the difference of behavior of the two series in eq.([eqalpha ] ) . however , as in the previous section concerning the location of the equivalent smooth interface , one can extract the asymptotic behavior of @xmath245 .
@xmath254 this latter result shed some light on the physical meaning of the previously mentioned divergence . in our presentation
, we have chosen to fix the largest wavelength ( set to @xmath255 ) and amplitude of this mode . increasing
the number of modes implies that the shortest wavelength @xmath256 decreases . for roughness exponents in the range
@xmath131 , this implies an algebraic increase of @xmath257 with @xmath35 , while @xmath258 is bounded .
this divergence of @xmath257 is of no importance for the correlation @xmath237 only if the profile is persistent .
otherwise , eq.([eqalpha_ap ] ) holds .
the perturbation method used however assumes that both @xmath259 and @xmath257 should be small .
the above analysis simply identifies which cut - off will dictate its behavior to the correlation .
the antipersistent case is more suited to the case where @xmath256 is fixed together with its amplitude , while @xmath260 varies . in this case
, the @xmath237 coefficient increases as @xmath261 as can be read from eq.([eqalpha_ap ] ) using the scaling @xmath262 ( @xmath263 being independent of @xmath260 ) . up to now , we have only considered harmonic problems with a uniform field at infinity .
this kind of boundary condition is of particular interest for problems where the scale of variation of the field in the bulk of the solid is large compared to the scale of the roughness so that an asymptotic development can be performed where the matching is to be done on the far field as one focusses on the rough boundary .
however , from the conformal mapping , one can address more complex types of boundary conditions . in order to illustrate this ,
we develop here a particular class of solutions which can be used to solve any problem .
we will consider green functions which give the field in the medium for localized flux @xmath264 injected in the medium from the surface .
let us consider the following problem : a localised flux @xmath265 is injected at point @xmath266 , on the border of the unit circle @xmath16 .
the remaining boundary is perfectly insulating .
the same flux is withdrawn at the origin where @xmath267 . the harmonic field which fulfills such boundary conditions is @xmath268\ ] ]
this potential @xmath136 is the green function for the domain @xmath16 .
considering the transformation @xmath269 maps the unit circle to the semi - plane @xmath14 . in the transformation ,
the potential @xmath136 becomes @xmath270\ ] ] which is the green function for a unit flux localised at every site @xmath271 for all integer @xmath31 . at infinity , the potential approaches @xmath272 . from this green function
it is simple to derive the one obtained for a translated array of sources . for sources at @xmath273
, we have @xmath274\ ] ] from this latter expression , the green function for a localised and periodic source on the rough profile is obtained by combining @xmath136 and @xmath10 .
the green function thus reads @xmath275 which gives the potential at point @xmath276 for a series of sources periodically spaced with the same period as the profile @xmath277 .
we have introduced here conformal mapping technique which allows to address harmonic problems in semi - infinite domains limited by a rough interface .
this mapping is accompanied by an efficient numerical technique which allows us to compute the mapping by a few iterations of a one dimensional fourier transform . moreover
, this technique provides a natural basis for discussing analytically some practical applications .
we then defined and studied the notion of an equivalent smooth boundary , whose position has been obtained exactly in the limit of a small amplitude .
this question underlined the differences between persistent and anti - persistent boundaries , in terms of sensitivity to the lower or upper scale cut - off of the self - affine character of the boundary .
we considered the question of correlations between the gradient of the harmonic field on the boundary and the height of the profile at the same point .
the correlation has been explicitly computed and shown to converge to a precise limit for persistent boundaries .
anti - persistent profiles lead to a correlation coefficient which is dependent on the self - affinity range .
extensions of the above technique are numerous .
we essentially focused here on static problems involving harmonic fields .
however , the same mapping may also be used in connection with evolution problems such as diffusion or wave propagation ( localisation ) .
thermal diffusion in the vicinity of a rough boundary has recently been shown to display an anomalous scaling behavior at early stages which could be addressed by such methods .
impedence of rough electrodes is another potential field of extension which has been studied in recent years .
it may also be one constitutive brick of a different mapping dealing with different geometries .
an example of such extensions is the stress intensity factor ( i.e. the singular behavior of the stress field ) at a crack tip . in the framework of antiplane
elasticity one can compute the local stress intensity factor at the crak tip and relate it to the far - field singular behavior .
this problem is currently being investigated .
b. b. mandelbrot , d. e. passoja and a. j. paullay , nature * 308 * , 721 ( 1984 ) ; s. r. brown , geophys . res .
* 13 * , 1430 ( 1986 ) ; r. h. dauskardt , f. haubensak and r. o. ritchie , acta metall . mater . * 38 * , 143 ( 1990 ) ; b. l. cox and j. s. y. wang , fractals * 1 * , 87 ( 1993 ) ; e. bouchaud , g. lapasset and j. planes , europhys . lett . * 13 * , 73 ( 1990 ) ; k. j. maloy , a. hansen , e. l. hinrichsen and s. roux , phys . rev
. lett . * 68 * , 213 ( 1992 ) . | the aim of this study is to analyze the properties of harmonic fields in the vicinity of rough boundaries where either a constant potential or a zero flux is imposed , while a constant field is prescribed at an infinite distance from this boundary .
we introduce a conformal mapping technique that is tailored to this problem in two dimensions .
an efficient algorithm is introduced to compute the conformal map for arbitrarily chosen boundaries .
harmonic fields can then simply be read from the conformal map .
we discuss applications to equivalent smooth interfaces .
we study the correlations between the topography and the field at the surface .
finally we apply the conformal map to the computation of inhomogeneous harmonic fields such as the derivation of green function for localized flux on the surface of a rough boundary .
defining and computing effective properties of heterogeneous media is a subject which has been studied for a long time , and for which a number of powerful techniques have been developped . in most cases however , the heterogeneities are considered to lie in the bulk of the material .
another type of inhomogeneity is due to the random geometry of the surface on which boundary conditions are applied .
this study focusses on this second type
. we will thus consider _ homogeneous media _ which are limited by a rough surface or interface .
our purpose here is to introduce a very efficient way of solving harmonic problems in two - dimensional systems for any geometry of the boundary .
the occurence of rough interfaces in nature is more the general rule than the exception .
apart from very specific cases such as mica where a careful cleavage can produce planar surfaces at the atomic scale , surfaces are rough .
even glass with a very homogeneous composition , where the surface is obtained by a slow cooling of the material , so that surface tension can act effectively to smoothen all irregularities , does display roughness in the range 5 to 50 nanometers over a window of a few micrometers width@xcite .
similarly , the so - called `` mirror '' fracture surface which is optically smooth exhibits specific topographic patterns when examined with an atomic force microscope@xcite .
the key question is thus to identify the relevant range of scales at which roughness appears . from common observations
, this question may not admit a simple clear cut answer . indeed in a variety of cases
, the amplitude of the roughness appears to be strongly dependent on the size of the examined surface .
a particular class of such scale dependent roughness , namely _ self - affine roughness_@xcite , has recently motivated a lot of activity ( see references @xcite for recent reviews ) both because of its relevance in many different instances , and of their theoretical justification which has been obtained in statistical physics for a wide class of models ranging from growth models@xcite , molecular beam epitaxy@xcite , fracture surfaces@xcite , to immiscible fluid interfaces@xcite .
although the present study is not specific to self - affine surfaces , we shall consider this particular class in order to apply our method .
the interest of this choice being that _
i ) _ the description of the roughness is realistic for a number of applications , _ ii ) _ consequences can be expressed in quite general terms as a function of few parameters directly accessible experimentally , and finally _ iii ) _
the most commonly studied roughness models are `` monochromatic '' surfaces with a single asperity pattern repeated periodically , and hence the transposition to more complex geometries may reveal wrong ( examples of such cases will be explicited in the main body of this article ) .
as previously mentioned , if most surfaces are rough , this roughness may be of small amplitude macroscopically , and thus one may feel that its role can be neglected in most cases .
fortunately , this is generally true .
taking into account precisely the surface roughness may be required in two distinct classes of problems : the first class ( i ) covers applications where the roughness can not be neglected at the scale at which the bulk field varies . for obvious reasons , there is no way to avoid the accurate description of the boundary .
we may mention the following potential applications : @xmath0 in confined geometries , such as naturally encountered in surface force study , the roughness of the surface may affect the interpretation and thus the precision of the measurements since the distance between two facing surfaces is generally estimated from indirect measurements of transport in the gap between the surfaces.@xcite @xmath0 field which are rapidly varying in space will be sensitive to fine details of the boundary geometry .
the most obvious example in this field is the reflection and scattering of a wave by a rough boundary.@xcite of particular importance is the case of surface waves , evanescent waves , rayleigh waves in elasticity , ...
@xmath0 in a similar spirit , diffusion processes may display anomalous behaviors at short times where the diffusion length is smaller or comparable to the roughness.@xcite the second class of problems ( ii ) where roughness can not be neglected is when one has to focus on the boundary , either because only this part matters for extraneous reasons or because the system is sensitive to high fields which can be induced by the roughness itself .
some examples of these two cases are listed below : @xmath0 surface phenomenon such as electrofiltration requires a proper solution of say a stokes flow field , in the immediate vicinity ( typically debye length scale ) of a rough boundary where an electric boundary layer is present and can be entrained by the fluid to give rise to an electric current in response to a fluid flow in a porous medium.@xcite @xmath0 the brittle fracture of glass is generally due to surface defects which induce locally high stresses , which reduce significantly the breaking limit of this material . in the absence of specific surface degradation the most important source of surface defect is the topography itself@xcite .
@xmath0 some growth models have a local growth rate which depends on a harmonic field locally .
the development of unstable modes which will finally induce a macroscopic roughening does require the proper analysis of the field at the surface.@xcite the relative independence of the bulk field on the small scale roughness of the boundary for a slowly varying field ( class ii problems ) can be used to explore the local field close to the boundary using an asymptotic analysis with a double scale technique .
the large scale problem consists in solving the problem at hand replacing the rough boundary by a smooth equivalent one .
the small scale problem deals with the details of the rough boundary and matches at `` infinity''with a homogeneous field .
this local problem will be considered in full details in the following .
these examples are obviously not exhaustive .
inhomogeneous boundary conditions may arise for instance in contact problems where the roughness can not be neglected.@xcite one may also consider application outside the realm of physical applications , such as the use of harmonic problems and particularly conformal map for providing a simple means of meshing a domain limited by a rough boundary . in the present article we will essentially focus on harmonic problems .
the latter arise in a variety of different domains in physics , such as electrostatics , thermal or concentration diffusion , flow in porous media , anti - plane elasticity to mention a few .
another use of conformal mappings is the resolution of bi - harmonic problems near a rough interface ; both stress field in elasticity and velocity field in low reynolds number fluid mechanics@xcite can be derived from potentials that obey such bi - laplacian equations .
we refer the reader to the companion paper which is completely devoted to this specific problem .
this paper is devoted to the study of harmonic problems in 2d semi - infinite media limited by a rough boundary . to extend the definition of the profile of the boundary to infinity
, we use periodic boundary conditions along the boundary .
although very specific , this type of geometry will be very convenient as soon as no other boundary lies close to the first one .
the distance threshold to consider in such a case is typically of the order of magnitude of the larger spatial wavelength of the profile _
i.e. _ the spatial period in the geometry we have described .
we use a conformal mapping technique .
it consists in contructing a map from the domain of interest ( in the complex plane ) onto a regular semi - plane .
the conformity of the map allows to preserve harmonicity through the map transform . in a first part of this paper
, we define the form of the conformal mapping suited to our geometry . then
, we adress the question of constructing the mapping associated with any prescribed interface .
we show that this problem can be solved with an iterative algorithm using fast fourier transforms ( fft ) .
this algorithm allows to get the conformal map in a few iterations of fft , whose computation time scales as @xmath1 where @xmath2 is the number of fourier modes used to describe the interface .
note the remarquable efficiency of such a technique , considering that the map gives the solution of a laplacian field in the entire two - dimensional problem .
this problem is very close to the so - called `` theodorsen problem '' @xcite in a circular geometry .
we also show that one can generate maps which naturally give rise to self - affine boundaries , a powerful technique to explore generic properties of such problems .
specific applications of this technique to self - affine profile are studied , which includes _
i ) _ the question of defining an equivalent smooth ( planar ) interface , and finding its height compared to the geometrical average height of the interface , _ ii ) _ the correlation between the height and the field which is computed exactly in the limit of a small roughness amplitude .
these two examples demonstrate the unexpected difference in behaviors for persistent and anti - persistent profiles .
finally , we give the expression of the green function for localized flux on a rough interface . |
the bdg hamiltonian , @xmath58 , of a superconductor has always particle - hole symmetry @xmath59 where @xmath60 is the particle - hole operator . in the basis @xmath61 , the particle - hole operator is @xmath62 where @xmath63 and @xmath64 are pauli matrices in particle - hole and spin space respectively .
the time - reversal operator has the form @xmath65 .
if the system is time - reversal invariant , we have @xmath66 . theoretically speaking , the hamiltonian for superconducting pairing potential generally has the form @xmath67 where @xmath68 is the real ( imaginary ) part of the s - wave gap , @xmath69 is the imaginary ( real ) part of the p - wave gap , @xmath70 and @xmath71 in the basis @xmath72 .
we find that the matrices @xmath73 and @xmath74 satisfy the dirac algebra @xmath75 where @xmath76 .
therefore , the pairing potential can be constructed from these eight gamma matrices . in the following
, we will prove that @xmath77 expand a three dimensional vector space and construct a three dimensional representation of a su(2 ) group whose generators correspond to the rotation in spin space .
first , we define @xmath78=\frac{1}{2}\epsilon_{ijk}\tau_0 \otimes \sigma_k=\frac{1}{2}\epsilon_{ijk}\left(\begin{array}{cc } \sigma_k & 0 \\ 0 & \sigma_k \end{array}\right)=\epsilon_{ijk}s_k,\end{aligned}\ ] ] where @xmath79 are three generators of a su(2 ) algebra .
next , we will show that the matrix @xmath80 corresponds to rotating the spin from z axis to the @xmath81 direction where @xmath81 is a three dimensional unit vector .
generally a wave function in the basis @xmath61 has the form @xmath82 where @xmath83 is the 2-component spinor wave function for the electron ( hole ) .
when we rotate the electron spin as @xmath84 the spinor wave function for the hole will be transformed as @xmath85 therefore the wave function is transformed as @xmath86 therefore the matrix @xmath80 corresponds to the rotation in the spin space . because the su(2 ) generators @xmath79 are constructed from the eq .
[ su(2)-1 ] , the three gamma matrices @xmath87 construct a three dimensional representation of a su(2 ) .
therefore @xmath87 are the spinnors of the su(2 ) with generators @xmath88 .
this is consistent to the fact the @xmath87 are the basis of spin - triplet pairing .
it is easy to see that @xmath89 represent the superconducting pairing in the same spin channel .
therefore , as long as the superconducting pairing can be written as the superposition of the basis @xmath87 , we can always find a rotation operator which can rotate a general spinor to the axis in which the electron and hole have the spin .
also it is easy to check that @xmath90=0 $ ] .
therefore @xmath91 is a one - dimensional representation of the su(2 ) spin rotation group which is consistent to the fact that @xmath91 is the basis of the spin - singlet pairing .
the similar results can be applied to @xmath74 .
the equal space spectral function at energy @xmath92 is defined as @xmath93 which is a hermitian matrix .
based on the lehmann representation , the retarded green s function is related to the spectral function as @xmath94 where @xmath35 and @xmath95 are the spatial coordinates .
because the spectral function @xmath96 is hermitian , the principle integral @xmath97 is hermitian , and @xmath98 is anti - hermitian . the spectral function can be written as @xmath99 . as we are interested in the superconducting pairing , in the following , we will focus on the off diagonal part of the spectral function which has the form @xmath100 where @xmath101 and @xmath102 are the superconducting pairing amplitudes . because @xmath103 , @xmath104 and the spectral function @xmath96 are hermitian , the coefficients @xmath101 and @xmath102 are real .
we now focus on the retarded green s function at @xmath50 because it is related to the mfs in the topological superconductors .
as the hamiltonian @xmath105 has particle - hole symmetry , the retarded green s function at @xmath50 also respects particle - hole symmetry which is shown below : @xmath106 particularly we are interested in the block off diagonal part of the retarded green s function @xmath107 based on eq .
[ gf-2 ] , the off diagonal retarded green s function satisfies @xmath108 on the other hand , particle - hole symmetry requires the anomalous green s function to satisfy @xmath109 where we use the fact that @xmath110 comparing eq . [ gf-4 ] with eq .
[ gf-5 ] , we conclude that spin - singlet pairing is completely forbidden in the anti - hermitian part of the anomalous green s function @xmath111 which indicates that the mfs can not have spin - singlet pairing .
when a normal lead is attached to a superconductor , the andreev reflection at the sc / nm interface will induce the superconducting condensate in the normal region . in the following and for a one - dimensional ( 1d ) sc / nm junction , we use scattering matrix theory to show that the induced superconducting condensate by the mf resonant andreev reflection is always spin - triplet .
we first focus on a diii class sc / nm junction . because of time - reversal symmetry ,
the scattering matrix is self dual @xcite which requires that @xmath112 it is straightforward to show that the above scattering matrix satisfies the condition @xmath113 .
it is also noted that @xmath114 are the spin - triplet scattering amplitudes and @xmath115 is the spin - singlet scattering amplitude .
we are interested in the andreev reflection at @xmath50 . in this case , according to the particle - hole symmetry of the system , we have @xmath116 therefore , we have @xmath117 is real , @xmath115 and @xmath118 are pure imaginary . at last , the matrix @xmath119 should be unitary .
this requires @xmath120 because @xmath121 we have either @xmath122 , @xmath123 or @xmath124 , @xmath125 .
these indicate that the spin - singlet andreev reflection ( the former case ) and spin - triplet andreev reflection ( the latter case ) can not coexist . for the spin - triplet andreev reflection
, we further have @xmath126 which requires either @xmath127 or @xmath122 and indicates that the normal reflection and spin - triplet andreev reflection also can not coexist .
therefore , we conclude for a diii class sc / nm junction and at @xmath50 , the spin - triplet andreev reflection must be a perfect andreev reflection with vanished spin - singlet component .
it is also straightforward to show that @xmath128 which is consistent to the condition of the nontrivial topological quantum number @xcite in terms of the scattering matrix for a diii class superconductor .
the hamiltonian of the soc - semiconductor is @xmath129 where @xmath130 is the wave vector , @xmath131 is the lattice constant , @xmath28 is the soc hopping constant in the soc - semiconductor and the effective magnetic field of the soc is along @xmath29 direction .
soc will shift the two spin - subbands to opposite direction in the momentum space . in the 1d tight - binding model with @xmath132-like soc hamiltonian at low energy limit
, the dispersion relation is @xmath133 where @xmath134 .
if there are n sites in the normal region , the electron and hole will take time @xmath135 to go through the normal region . because the effective magnetic field due to soc is @xmath136 , the spin precession angle of an electron or hole in the normal region is @xmath137 . in the continuous limit
, we have @xmath138 , @xmath139 .
therefore , the spin precession angle in the continuous limit is @xmath140 where @xmath141 is the length of the normal region . for a time - reversal invariant tsc , there is a pair of majorana zero modes with opposite spins at the boundary .
the induced spin - triplet condensates are described by @xmath142 which is even under time - reversal operation . here
we assume the majorana spins are parallel or anti - parallel to z axis .
if the spin - orbit coupling ( soc ) field direction is along y axis as we consider in the manuscript , the spins are rotated so that the spin - triplet states are changed to @xmath143 where @xmath43 and @xmath144 represent the spins lying in the x - z plane with wave functions @xmath145 and @xmath146 respectively .
therefore the projection of the condensates is @xmath147 which is independent of the soc .
therefore in this case the tsc / soc / tsc junction performs like a normal josephson junction .
however if the soc field direction is along x axis so that the majorana spin is rotated in y - z direction , the spins @xmath43 and @xmath144 correspond to @xmath148 and @xmath149 respectively .
the projection is @xmath150 which will result in a 0- and @xmath1- josephson junction transition .
2ifxundefined [ 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.69.731 [ * * , ( ) ] | we establish that majorana fermions on the boundary of topological superconductors have only spin - triplet superconducting correlations independent of whether the bulk superconducting gap is spin singlet or triplet .
this is universal for time - reversal broken ( respected ) topological superconductors ( tscs ) with an odd number of ( pairs of ) majorana fermions on the boundary .
consequently , resonant andreev reflection induced by majorana fermions only occurs in spin - triplet channels and always injects spin - triplet cooper pairs into the leads .
this spin - triplet condensate results in a spin - orbit coupling ( soc ) controlled oscillatory critical current without a @xmath0-@xmath1 transition in the tsc / soc - semiconductor / tsc josephson junction .
the observation of this unique current - phase relation can serve as the definitive signal for majorana fermions .
our study shows a technique for manipulating majorana fermions based on their spin - triplet superconducting correlations . _
introduction _ due to recent theoretical proposals of realizing topological superconductors ( tscs ) with an @xmath2-wave superconducting proximity effect in spin - orbit coupling ( soc ) systems @xcite , finding and manipulating majorana fermions @xcite have become possible in superconductor - semiconductor devices @xcite .
majorana fermions ( mfs ) in this context are symmetry protected zero - energy subgap states of tscs with equal weight electron - hole superposition .
this leads to some unique transport properties , such as quantized zero - bias conductance @xcite and fractional josephson effect @xcite , and exotic superconducting condensates such as an odd frequency @xmath2-wave and spin - triplet pairing in a time - reversal broken @xmath3-wave - superconductor / normal - metal junction @xcite .
however , so far , most studies ignore the spin states of mfs . on the other hand ,
controlling or manipulating spin states is a conceptional revolution in fundamental science as well as emerging technologies @xcite and is triggering important fundamental discoveries of spin - dependent transport phenomena such as giant magnetoresistance @xcite and spin hall effect @xcite .
therefore , it is natural to ask whether we can manipulate mfs by their spin properties . in this paper
, we show that mfs , zero - energy bound states of tscs , have only spin - triplet superconducting correlations which are model independent and guaranteed by particle - hole symmetry for time - reversal broken tscs ( d class @xcite ) and particle - hole as well as time - reversal symmetries for time - reversal respected tscs ( diii class @xcite ) .
this leads to our experimental prediction that no matter whether the bulk superconducting gap is spin singlet @xcite or triplet @xcite , the mf - induced resonant andreev reflection ( ar ) only injects spin - triplet cooper pairs into the normal lead .
we perform a symmetry analysis as well as numerical calculations of the reflection matrix at the interface of either time - reversal broken or respected tsc / normal - metal ( nm ) junctions .
we further propose to detect spin - triplet cooper pairs in a tsc / soc - semiconductor / tsc experimental setup by observing a soc controlled oscillatory critical josephson current .
it has been recently shown that soc can rotate the @xmath4 vector of the spin - triplet cooper pairs and leave spin - singlet cooper pairs unaffected in the proximity region .
therefore , the observation of our predicted soc controlled critical current can serve as direct confirmation for spin - triplet cooper pairs and the mf transport signal . _
spin - triplet superconducting correlations of mfs _ in any dimension superconductor , the mf , a self - hermitian quasiparticle , can be defined as @xmath5 , where @xmath6 is generally a vector and @xmath7 is the @xmath8th zero - energy quasiparticle wave function .
electron and hole correlations at the fermi surface show up in the block off - diagonal part of the equal space zero energy spectral function , which is a hermitian matrix and has the form @xmath9 @xcite . in the four - component spinor basis @xmath10
, the block off - diagonal part of the spectral function ( bosf ) can be written as @xmath11^{\dagger } & \hat{0 } \end{array}\right),\end{aligned}\ ] ] where @xmath12 and @xmath13 are the @xmath14 identity and pauli matrices in electron - hole and spin space respectively , and @xmath15 and are the spin - singlet and spin - triplet pairing amplitudes , respectively . in a superconductor ,
the particle - hole symmetry requires the zero energy spectral function to satisfy @xcite @xmath16 where @xmath17 is the particle - hole operator , with @xmath18 the complex conjugate operator and @xmath19 the pauli matrices in particle - hole space .
a straightforward calculation shows that @xmath20^{\dagger } \end{array}\right).\end{aligned}\ ] ] comparing eq .
( [ pair - ph ] ) with eq .
( [ a - ph ] ) , we find that @xmath15 must vanish at the fermi surface .
this implies that the particle - hole symmetry completely forbids spin - singlet pairing but allows spin - triplet pairing in the zero - energy bosf .
for the d class tsc , when there is one mf @xmath21 on the boundary , the majorana condition @xmath22 requires @xmath23 , which results in a spin - triplet bosf , @xmath24 , by choosing @xmath25 .
if there is an additional zero mode , @xmath26 , the anomalous spectral function is zero because the two mfs can emerge as a single electron or hole @xcite . for a diii class tsc , although there is always at least a pair of mfs on the boundary , the time - reversal symmetry guarantees that these two mfs belong to different spin channels and therefore manifest spin - triplet pairing @xcite .
the above analysis applies equally well to the multi - mf situation and we conclude that an odd number ( of pairs ) of mfs in the d ( diii ) class have only spin - triplet superconducting correlations . it is noted that our symmetry analysis is independent of the details of tsc models and robust against the spin - independent disorder . therefore the spin - triplet superconducting correlation is a universal and robust property of mfs .
[ cols= " < " , ] _ soc controlled critical josephson current _ so far , we have shown that the mf - induced resonant ar occurs only in spin - triplet channels .
it is known that the spin property of the ar - induced cooper pairs is the same with that of the associated ar @xcite .
therefore , the mf - induced resonant ar always injects spin - triplet cooper pairs .
recent studies @xcite show that soc can rotate the vector of the ar - induced spin - triplet cooper pairs but leave the spin - singlet pairs unaffected in the proximity region .
thus soc provides a mechanism to distinguish between spin - triplet and spin - singlet cooper pairs . to confirm our prediction
, we therefore consider a tsc / soc - semiconductor / tsc josephson junction ( fig .
[ andreev - josephson](a ) ) .
the hamiltonian of the soc - semiconductor system has the form @xmath27\tau_z\otimes \sigma_0 + 2t'_{\rm so } \sin(k ) \tau_z\otimes \sigma_y,\end{aligned}\ ] ] where @xmath28 is the soc hopping amplitude and the effective magnetic field of the soc is along the @xmath29 direction . on the boundary @xmath30 and @xmath31 [ fig .
[ tsc - diii](a ) ] of time - reversal broken tscs , based on our previous discussion , the cooper pairs can be described as @xmath32 , with spin @xmath33 along the @xmath34 direction .
when a cooper pair travels from @xmath30 to @xmath31 in the soc - semiconductor region , because the effective soc field is along the @xmath29 direction , the spins of the electron and hole start to precess in the @xmath35-@xmath34 plane with the same precession angle . as the spin splitting energy due to soc
is @xmath36 , the precession angle of the particle traveling from @xmath30 to @xmath31 satisfies @xcite @xmath37 where @xmath38 in the limit @xmath39 , and we assume there are @xmath40 sites in the normal soc region whose length is @xmath41 ( fig .
[ andreev - josephson](a ) ) .
the state of this cooper pair at @xmath31 generally can be described as @xmath42 where @xmath43 represents a spin lying in the @xmath44 plane and its spin wave function is @xmath45 .
therefore , the projection of the condensate @xmath42 to @xmath46 takes the form @xmath47 especially for @xmath48 , the projection in eq .
( [ projection ] ) is zero which means no cooper pair can travel through the josephson junction , and this indicates zero critical current . as a result , we expect the critical current will oscillate with a large amplitude and may even vanish as the soc strength is suitably tuned . to confirm our theoretical argument , we plot the andreev levels in fig .
[ andreev - josephson](b ) in the cases of @xmath49 respectively .
because the andreev levels cross @xmath50 at @xmath51 , the josephson current has the form @xmath52 where @xmath53 is the critical current and @xmath54 is the phase difference between two tscs .
it is noted that at @xmath48 , the energy - phase relation of the andreev levels is almost flat
. therefore we conclude @xmath55 in this case which is consistent with our theoretical analysis . in fig .
[ andreev - josephson](c ) , @xmath53 is plotted as a function of soc controlled precession angle @xmath56 , showing a @xmath57 period as we expected based on eq . [ projection ] . generally , time - reversal broken tscs can be considered as cooper pair filters which only allow the cooper pairs in the state @xmath32 to enter the tscs from the normal region .
however the situation for time - reversal invariant tscs is very different .
because the superconducting condensates exist in both spin - up and spin - down channels , the spin - triplet cooper pairs can always enter the time - reversal invariant tscs which is independent of their spin direction @xcite .
our prediction can also be applied to multi - band and even two - dimensional quantum wells with equal rashba and dresselhaus soc strengths , which have been realized experimentally @xcite . in this case , the spin precession angles for all transverse channels are the same @xcite and the d - vector of the triple pairs will not decay even in the presence of spin - independent disorder @xcite .
the oscillatory critical current has also been predicted and observed in the non - topological superconductor / ferromagnet / superconductor ( sfs ) josephson junction where time - reversal is broken and mfs do not play any role in contrast to our system @xcite .
the current - phase relation in the sfs junction undergoes a 0-@xmath1 transition when the critical current approaches zero .
we emphasize that the current - phase relation of our time - reversal broken josephson junction does not undergo a @xmath0-@xmath1 transition .
this makes our prediction qualitatively different from that in the sfs junction .
therefore , our prediction of a soc controlled critical current is a transport signature of mfs . in a realistic experimental setup ,
the semiconductor wire may not necessarily be perfect 1d but may have several channels . however , our prediction requires the scattering at the tsc / nm interface to be 1d .
the recent development of a surface split - gates design @xcite can make a real 1d point contact between the topological superconducting semiconductor wire and the normal lead .
therefore , our experimental proposal can be realized by an existing experimental technique .
_ acknowledgment : _
x.l . acknowledges useful discussions with alejandro m. lobos and xiaopeng li .
this work is supported by microsoft q , jqi - nsf - pfc , and lps - cmtc .
61ifxundefined [ 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/physrevlett.100.096407 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.101.160401 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.103.020401 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.104.040502 [ * * , ( ) ]
link:\doibase 10.1103/physrevlett.105.077001 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.177002 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.214509 [ * * , ( ) ] link:\doibase 10.1103/physrevb.81.125318 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.111.056402 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevx.4.021018 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevb.91.064505 [ * * , ( ) ] link:\doibase 10.1103/physrevb.61.10267 [ * * , ( ) ] http://stacks.iop.org/1063-7869/44/i=10s/a=s29 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.80.1083 [ * * , ( ) ] link:\doibase 10.1126/science.1222360 [ * * , ( ) ] link:\doibase 10.1021/nl303758w [ * * , ( ) ] , , http://dx.doi.org/10.1038/nphys2429 [ * * , ( ) ] http://dx.doi.org/10.1038/nphys2479 [ * * , ( ) ] link:\doibase 10.1126/science.1216466 [ * * , ( ) ] link:\doibase 10.1103/physrevb.87.241401 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.112.217001 [ * * , ( ) ] link:\doibase 10.1103/physrevb.63.144531 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.103.237001 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.180516 [ * * , ( ) ] link:\doibase 10.1140/epjb / e2004 - 00066 - 4 [ * * , ( ) ] link:\doibase 10.1103/physrevb.79.161408 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.113.036401 [ * * , ( ) ] link:\doibase 10.1103/physrevb.87.104513 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.76.323 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.61.2472 [ * * , ( ) ] link:\doibase 10.1103/physrevb.39.4828 [ * * , ( ) ] link:\doibase 10.1126/science.1087128 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.92.126603 [ * * , ( ) ] link:\doibase 10.1103/physrevb.78.195125 [ * * , ( ) ] link:\doibase 10.1103/physrevb.72.140503 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.98.037003 [ * * , ( ) ] @noop link:\doibase 10.1103/physrevb.83.085413 [ * * , ( ) ] @noop ( ) , link:\doibase 10.1103/physrevlett.112.037001 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.69.731 [ * * , ( ) ] link:\doibase 10.1103/physrevb.23.6851 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.113.227002 [ * * , ( ) ] link:\doibase 10.1103/physrevb.86.134522 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.110.117003 [ * * , ( ) ] link:\doibase 10.1103/physrevb.89.134517 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.98.076604 [ * * , ( ) ] link:\doibase 10.1038/nature07871 [ *
* , ( ) ] link:\doibase 10.1126/science.1195816 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.246603 [ * * , ( ) ] link:\doibase 10.1038/nphys2383 [ * * , ( ) ] @noop arxiv:1403.3518 @noop * * , ( ) link:\doibase 10.1103/physrevb.75.125307 [ * * , ( ) ] link:\doibase 10.1103/physrevb.86.174301 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.77.1321 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.77.935 [ * * , ( ) ] @noop * * , ( ) |
the author is grateful to m. birse and n. walet for valuable discussions . | functional renormalisation group approach is applied to a system of kaons at finite chemical potential .
a set of approximate flow equations for the effective couplings is derived and solved . at high scale
the system is found to be at the normal phase whereas at some critical value of the running scale it undergoes the phase transition ( pt ) to the phase with a spontaneously broken symmetry with the kaon condensate as an order parameter .
the value of the condensate turns out to be quite sensitive to the kaon - kaon scattering length .
the study of phase structure of dense hadron matter is one of the key problems of the strong interaction physics described by qcd . at low energy
the qcd can be reduced to an effective theory containing goldstone bosons as effective degrees of freedom .
the extension of this effective theory to finite density requires taking into consideration nonzero isospin / strange chemical potential .
the physical motivation for such an extension is provided by a wide range of phenomenon occurring in heavy - ion collisions and the equation of state ( eos ) of the neutron stars @xcite . at finite density
these bosons may condense .
this condensation shows up in a variety of the physical systems .
for example , there is a strong possibility that a kaon condensation may exist in a core of the neutron stars so that a realistic analysis of such possibility as well as estimates of the value of the condensate may turn out to be important for establishing correct eos .
it is rather well known that , depending on the value of the chemical potential , the following phases can be realised : normal phase with the trivial vacuum , pion condensation / kaon condensation . in this letter
we focus on the kaon condensation which will be studied in the framework of the functional renormalisation group ( frg ) .
the more general case of the mixed pion / kaon condensation will be reported elsewhere .
the frg method has been applied to a wide spectrum of the problems ranging from cold atoms to quantum gravity . in the context of pt
the approach has been used to study the phase structure of neutron matter @xcite , the phenomena of bose - einstein condensation ( bec)@xcite and both balanced @xcite and imbalanced @xcite systems of cold fermionic atoms etc .
the approach makes use of the legendre transformed effective action : @xmath0=w[j]-j\cdot \phi_c$ ] , where @xmath1 is the usual partition function in the presence of an external source @xmath2 . the action functional @xmath3 generates the 1pi green s functions and it reduces to the effective potential for homogeneous systems . in the frg
one introduces an artificial renormalisation group flow , generated by a momentum scale @xmath4 and we define the effective action by integrating over components of the fields with @xmath5 .
the rg trajectory then interpolates between the classical action of the underlying field theory ( at large @xmath4 ) when the quantum fluctuation effects are excluded , and the full effective action ( at @xmath6 ) with all quantum fluctuations taken into account .
the evolution equation for @xmath3 @xcite in the frg has a one - loop structure and can be written as @xmath7 .
\label{eq : gamevol}\ ] ] here @xmath8 is the matrix containing second functional derivatives of the effective action with respect to the boson fields and @xmath9 is a matrix containing the corresponding cut - off functions which must vanish when the running scale approaches zero .
the flow equation as written above is just a differential form of a general functional intergal and has therefore a little practical sense . in order to actually solve the flow equations one needs to choose a certain ansatz for the effective action motivated by the form of the action at the starting scale and the undelying symmetries .
the ansatz assumed in this work is @xmath10 = \int d^4x\ , \left[z_\phi\,(\partial_0 + i \mu)\phi^\dagger\,(\partial_0 - i \mu)\phi -z_m\,\partial_i\phi^\dagger\partial_i\phi - u(\phi,\phi^\dagger)\right],\ ] ] where @xmath11 and @xmath12 are the renormalisation factors depending on the running scale and @xmath13 is a complex doublet field defined as follows @xmath14 the first and second components of the doublet can be identified with the pair of ( @xmath15 ) and ( @xmath16 ) mesons correspondingly .
the effective potential depends only on the combination @xmath17 .
we expand the effective potential @xmath18 near its minima and keep terms up to order @xmath19 .
@xmath20 where first three terms correspond to the expansion near the minimum with respect to the first doublet and the rest is the expansion near the minimum with respect to the second doublet .
note that the standard mass term is included in the definition of the @xmath21 coupling .
the chemical potential is provided by the external conditions ( we assume @xmath22 0 ) .
the action is invariant under the global @xmath23 group , where @xmath24 is the isospin group and @xmath25 is related to hypercharge .
it essentually captures the mean features of the kaon condensation phenomena .
substituting the ansatz for @xmath3 into the flow equation and performing the contour integration one can get the evolution equation for the effective potential which acts as a driving term generating the flow of the couplings . at large scale
we expect symmetric state with the trivial minimum of the effective potential whereas at lower scale comparable with the value of chemical potential a formation of the condensate is expected . @xmath26 where @xmath27 and @xmath28 here @xmath29 .
@xmath30 and @xmath31 determine the pole positions of the corresponding propagator @xmath32 another two poles are those with the overall minus sign in front .
the positive and negative values of energy correspond to creation and annihilation of the exitations , respectively . for our purposes
it is sufficient to pick up only positive eigenvalues .
note that the pole position defines the corresponding dispersion relations in the general case of nonzero regulator @xmath33 0 .
taking @xmath34 , @xmath35 and @xmath36 0 one can recover the dispersion relations in the broken phase derived in refs .
@xcite .
@xmath37 another dispersion relations , also derived in refs .
@xcite in the limit of @xmath34 , @xmath35 take the form @xmath38 as one can easily be seen from the above expressions two of the dispersion relations describe goldstone bosons , i.e gapless modes when @xmath39 0 .
this is the example of the nontrivial situation when the number of the broken generators is not equal to the number of the massless modes . indeed , since the initial global @xmath23 symmetry is broken to @xmath25 one could expect three massless modes .
the problem was studied in details in @xcite where is was shown that in the system with a broken lorenz symmetry the number of gapless modes may be lesser then the number of broken generators .
the physical reason is the presence of the chemical potential which induces a mass splitting between the doublets ( @xmath15 ) and ( @xmath16 ) so that the first one acquire the effective mass @xmath40 whereas the effective mass for the second one becomes @xmath41 .
the couplings may in general depend not only on running scale but also on the magnitude of the condensate so that we define the total derivative as @xmath42 where @xmath43 .
applying this to the effective potential and neglecting higher order terms gives the set of the flow equations @xmath44 the actual flow is determined by the choice of the cut - off function @xmath45 . in this exploratory study we have chosen the cut - off in the form suggested in @xcite @xmath46 the advantage of this form of cut - off is that it simplifies some algebra , allowing part of the calculations to be carried out analytically .
one notes , however that this advantage may disappear for more complicated cases like , for example , in the mixed boson - fermion systems . to solve the system of the flow equations one needs to fix a set of initial conditions .
they were determined as follows .
one starts from the effective action in vacuum where @xmath47 , @xmath48 , @xmath49 and @xmath50 is determined by the value of kaon - kaon scattering length in vacuum .
the flow equations in vacuum can easily be obtained from the general expression for the @xmath51 by differentiating it with respect to @xmath52 and @xmath53 and putting @xmath54 , @xmath55 afterwards .
for example , the flow equation for the couplings @xmath56 and @xmath57 take the form @xmath58 = 510 mev , 550 mev and 595 mev correspondingly.,width=453 ] and @xmath59 the flow equations for the other couplings can be obtained in a similar way .
the value of @xmath60 is found to be very small at @xmath61 compared to the lower order couplings and the coupling @xmath62 does not really run in vacuum so that we put @xmath63 at starting scale .
these equations are solved using the values of couplings at @xmath6 as initial conditions . from the vacuum flow the values of the couplings at the starting scale for the general flow @xmath64
are extracted .
the value of the scale @xmath65 is chosen to be much larger then any other mass scale involved in the problem .
we have taken the value @xmath66 gev as the starting scale .
it is large enough to provide practically independent results for the couplings at the physical scale @xmath67 . at the scale @xmath68 the mass term approaches zero thus signaling the onset of spontaneous symmetry breaking ( ssb ) .
the corresponding vacuum expectation value of the field @xmath13 becomes nonzero and the system undergoes phase transition .
we show on fig.1 the behaviour of the mass term in the vicinity of the critical scale at several values of the chemical potential .
the critical value of the chemical potential for such a transition is found to be at @xmath69 as it should . at the scale @xmath70 3 gev the curves merge and follow this pattern up to starting point .
the change of the transition point as a function of the chemical potential is such that it grows with the increase of the chemical potential but , in any case the system undergoes the transition to the broken phase at the scale @xmath71 . at this scale the mass term vanishes and
the condensate develops . on the other hand the character of
the change of the mass term for the other fields in @xmath72 is such that it stays positive for any scale so that ssb never happens .
one may , therefore conclude that , whereas @xmath73 and @xmath74 mesons condense , the pair of the @xmath75 and @xmath76 mesons does not in agreement with the results from @xcite and @xcite .
one notes that the mechanism of the condensation considered in this paper and in refs .
@xcite is related to an instabilities in the colour - flavour locked phase of qcd @xcite is therefore quite different from the conventional condensation of @xmath76 mesons suggested in @xcite which is related to attractive interaction between @xmath76 mesons and nucleons .
corresponding to kaon - kaon scattering lengths from the lattice calculations @xcite(upper curve ) , lowest order chiral perturbation theory @xcite ( middle curve ) and phenomenological model @xcite ( lower curve),width=381 ] .
in fig.2 we show the behaviour of the condensate as the function of chemical potential at physical scale when @xmath77 and for three values of the @xmath78 coupling .
as one can see from fig.2 the value of the condensate grows with the increase of the chemical potential in all three cases .
this growth is relatively fast at the threshold @xmath79 and then slows down for the larger values of the chemical potential .
we note however that the results become less reliable with an increase of the chemical potential as the energy / momentum dependence of the couplings neglected in this study may become progressively more important at larger @xmath53 .
the important ingredient of the approach is the running coupling @xmath78 , which is related to the kaon - kaon scattering length . to get the results shown by the upper curve on fig.2 we used the value for the scattering length obtained in the lattice calculations @xmath80=
-0.310 @xmath81 , @xcite .
the middle curve corresponds to the use of the lowest order chiral perturbation theory result @xmath82 @xcite and the lower one is obtained using the phenomenological model from ref .
@xcite .
one can see from fig.2 that the difference between the results is quite noticeable , especially at the threshold .
it becomes less sensitive to the kaon - kaon interaction at the chemical potentials significantly larger then kaon mass but then using energy / momentum independent interaction vertices may not be sufficient . taking into account the next to leading order in chiral perturbation theory does not make this difference any smaller . on the contrary , the next - to - leading order chiral perturbation theory leads to even larger deviation from the lattice results .
in any case the chiral perturbation theory based estimates are not very reliable as @xmath83 chiral perturbation theory does not converge well and some sort of resummation techniques are probably required like those developed for the kaon - nucleon scattering @xcite .
overall one can conclude that the @xmath78 coupling generates a significant source of the theoretical uncertainty related to a poor knowledge of the kaon - kaon scattering amplitude at threshold . on the other hand
this quantity is important for establishing the correct equation of state ( e0s ) for neutron stars . in the present work
we have used the frg approach to analyse the phenomena of kaon condensation .
we found that at @xmath84 the system undergoes the transition to the broken phase with a nontrivial vacuum resulting in formation of the condensate .
however the exact value of the condensate ca nt in present be determined accurately because of lack of knowledge about the magnitude of the kaon - kaon scattering length .
there are several ways of how the approach can be extended and improved .
first , the energy / momentum dependence of the interaction vertices should be taken into account .
it may be important for estimation the effect of depletion of the condensate .
second , instead of rather toy model for effective action , used in this paper one can utilize a more realistic action based on effective lagrangian like that derived in ref .
@xcite .
the induced or effective chemical potential considered there will emerge as the result of the rg flow .
third , physically interesting case of competing chiral and kaon condensates can be considered .
apart from that it would be interesting to consider a mixed system with both kaon and pion condensates .
these condensates may influence each other via @xmath85 scattering which may have important consequences for the neutron stars eos .
the results for the mixed pion - kaon system will be reported in future publications . |
the 1990s saw four epochal advances in our understanding of the ly@xmath0 forest .
spectra of quasar pairs showed coherence over scales of a hundred kpc and more , implying large sizes and thus low densities for the absorbing structures @xcite .
keck hires spectra of unprecedented resolution and signal - to - noise demonstrated the ubiquity of weakly fluctuating ly@xmath0 absorption in the high redshift universe @xcite , and they revealed the presence of metal lines associated with low column density hydrogen absorbers @xcite .
finally , and most directly relevant to this review , a combination of numerical simulations and related analytic models led to a compelling new physical picture of ly@xmath0 forest absorption @xcite .
the basic numerical result is simple to summarize : given a cosmological scenario motivated by independent observations , 3-d simulations that incorporate gravity , gas dynamics , and photoionization by the uv background produce something very much like the observed ly@xmath0forest , an outcome that requires no _ ad hoc _ adjustments to the model .
the top three rows of figure [ fig : spec ] illustrate this point , showing , respectively , the observed ly@xmath0 forest of the @xmath6 quasar q1422 + 231 , expanded views of four selected regions of this spectrum , and simulated spectra of the same length along four randomly selected lines of sight through a cosmological simulation .
the simulation uses smoothed particle hydrodynamics ( sph ) with @xmath7 dark matter particles and @xmath7 gas particles in a periodic cube of comoving size @xmath8 ( @xmath9 at @xmath10 ) .
it assumes a @xmath11cdm model ( inflationary cold dark matter with a cosmological constant ) , with @xmath12 , @xmath13 , @xmath14 , @xmath15 , inflationary spectral tilt @xmath16 , and a power spectrum normalization that corresponds to @xmath17 at @xmath18 .
spectra are extracted from the @xmath10 simulation output using the methods of @xcite .
= 5.0truein studies of the ly@xmath0 forest have traditionally focused on `` lines '' identified by a decomposition procedure and the `` clouds '' or `` absorbers '' that produce them .
hydrodynamic simulations show that typical marginally saturated absorption features at @xmath19 arise in filamentary structures , which are analogous to ( but smaller in scale than ) today s galaxy superclusters .
however , in the simulations there is no sharp distinction between the `` lines '' and the `` background , '' and one can also characterize the ly@xmath0 forest as `` gunn - peterson '' @xcite absorption produced by a smoothly fluctuating intergalactic medium @xcite .
what makes the fluctuating igm perspective especially powerful is the tight relation between the density and temperature of low density cosmic gas , @xmath20 with @xmath21 , which emerges from the balance between photoionization heating and adiabatic cooling @xcite .
the ly@xmath0 optical depth of photoionized gas is @xmath22 , where @xmath23 accounts for the temperature dependence of the recombination rate and @xmath24 is the rate at which neutral atoms are ionized by the cosmic uv background .
the temperature - density relation then leads to the `` fluctuating gunn - peterson approximation '' ( fgpa ) , @xmath25 $ ] , which relates the continuum normalized flux @xmath26 to the local gas overdensity @xcite .
the index @xmath27 lies in the range @xmath28 depending on the thermal history of the gas , and the proportionality constant @xmath29 is itself proportional to @xmath30 .
thus , one can think of a ly@xmath0 forest spectrum as providing a 1-dimensional , non - linear map of the gas overdensity @xmath31 along the line of sight .
the map is smoothed by thermal broadening and distorted by peculiar velocities , but these effects are small enough to leave a tight correlation between ly@xmath0 optical depth and local gas density even in redshift space @xcite . furthermore , pressure gradients in the diffuse , photoionized igm are usually weak compared to gravitational forces , so the gas overdensity traces the dark matter overdensity fairly well .
the fgpa provides a valuable way of thinking about the information content of the ly@xmath0 forest , and it can be a useful calculational tool if one has a way to produce realizations of cosmic density and velocity fields .
the analytic models of the ly@xmath0 forest by @xcite essentially combine the fgpa with a log - normal or zeldovich approximation model for creating non - linear density and velocity fields .
alternatively , one can run an inexpensive , gravity - only n - body simulation and assume that the diffuse gas traces the dark matter ( or add an approximate treatment of gas pressure @xcite ) .
the dotted lines in the third row of figure [ fig : spec ] show spectra extracted from a particle - mesh ( pm ) n - body simulation , run from the same initial conditions as the sph simulation , illustrating both the effectiveness and the limitations of this approach .
the two spectra trace each other over most of their length , with the largest breakdowns occurring in regions where shock heating has pushed the gas above the @xmath32 temperature - density relation , which makes the optical depth in the sph simulation lower than the fgpa predicts .
the n - body+fgpa technique offers a convenient way to investigate the response of the ly@xmath0 forest to changes in cosmological parameters . comparing the third and fourth rows of figure [ fig : spec ]
illustrates the effect of increasing the matter fluctuation amplitude by 20% , to @xmath33 , with the same fourier phases in the initial conditions .
the increased clustering of the high amplitude model is evident in the greater incidence of saturated absorption and in the merging of multiple small scale features into single larger scale features .
the bottom row shows spectra from a model with @xmath17 and the same linear power spectrum but a lower matter density , @xmath34 .
the reduction in @xmath35 lowers the value of the hubble parameter @xmath1 at @xmath10 , and as a result features are more densely packed in redshift space , producing `` choppier '' spectra .
in fact , our @xmath8 simulation cube is only @xmath36 in length for @xmath34 , and we have `` padded '' each spectrum to @xmath9 by replicating the first @xmath37 .
reducing @xmath35 also raises the fluctuation amplitude at @xmath10 ( since @xmath38 is held fixed at @xmath18 and there is less late time growth for lower @xmath35 ) , but this effect is less important than the effect on the hubble parameter .
the relative simplicity of the underlying physics and the existence of superb data at redshifts that are only sparsely probed by other observables make the ly@xmath0 forest a potentially powerful tool for testing cosmological models .
given independent estimates of @xmath24 ( e.g. , from the quasar luminosity function or the proximity effect ) , one can use the mean opacity of the forest to obtain a lower limit to the cosmic baryon density @xcite .
however , the observational constraints on @xmath24 are loose , so for higher precision applications one must treat @xmath24 as a free parameter and adjust it to match a single observable , usually taken to be the mean opacity .
this normalization also absorbs uncertainties in @xmath39 , @xmath40 , and @xmath41 .
we have followed this practice for each of the simulations in figure [ fig : spec ] and will adopt it in all of our subsequent analyses . after normalizing to the mean opacity , the variation about the mean i.e. , the structure of the ly@xmath0 forest is a prediction of the cosmological model , driven primarily by the structure in the underlying dark matter distribution .
ideally , one would like to calculate predictions for each cosmology using high resolution hydrodynamic simulations , but for some purposes other numerical or analytic approximations may be accurate enough . the adequacy of such approximations must be tested on a case - by - case basis , and the improving quality of the observational data places ever more stringent demands on the accuracy of the theoretical calculations .
because of the tight predicted correlation between the local ly@xmath0 optical depth and the local gas density , the most natural statistics for characterizing the ly@xmath0 forest are those that treat each quasar spectrum as a continuous 1-dimensional field , rather than a collection of discrete `` lines . ''
we have reviewed this approach elsewhere @xcite , and here we provide a brief recap with different examples and emphasis .
= 3.0truein figure [ fig : fluxpk ] shows the power spectrum of continuum normalized flux , converted from 1-d to 3-d as described by @xcite , and multiplied by @xmath42 to yield the variance per ln@xmath43 .
the curves are noisy because they are based on a single @xmath8 comoving cube , but we use the same phases in each simulation , so the relative behavior should be minimally affected by noise . at scales @xmath44 , the power spectra turn over because of the combined effects of non - linearity , peculiar velocities , and thermal broadening .
the predictions in this regime are also affected by the finite resolution of the simulations . at larger scales , the sph and
n - body+fgpa methods give similar but not identical results for the same cosmological model , and the amplitude of the flux power spectrum increases with the amplitude of the matter power spectrum , as expected based on the fgpa and on figure [ fig : spec ] . on large scales ,
the shape of the 3-d flux power spectrum is similar to that of the linear matter power spectrum @xmath45 @xcite .
the close connection between the shape and amplitude of the flux power spectrum and the shape and amplitude of @xmath45 is the basis of croft et al.s @xcite method for recovering @xmath45 from ly@xmath0 forest data . applying this method to a sample of 30 keck hires spectra and 23 keck lris spectra yields a matter power spectrum in remarkably ( or , perhaps , disappointingly ) good agreement with the `` concordance '' @xmath11cdm model favored by cmb , supernova , weak lensing , and low-@xmath46 large scale structure data ( @xcite ; see @xcite for an independent analysis of the same flux power spectrum and @xcite for an independent comparison to other cosmological constraints ) .
mcdonald et al .
@xcite reach similar conclusions from a `` forward '' comparison of hydrodynamic simulation predictions to the flux power spectrum measured from eight hires spectra .
the ly@xmath0 forest power spectrum tests the @xmath11cdm model in a previously unexplored regime of redshift and lengthscale .
it confirms one of the scenario s key predictions , a linear power spectrum that bends from the primeval @xmath47 towards @xmath48 on small scales .
the implied constraints on cosmological parameter combinations complement those from the cmb and other data .
the sloan digital sky survey ( sdss ) has already obtained moderate resolution spectra of several thousand high redshift quasars , and these will soon provide measurements of the flux power spectrum with much greater precision on large scales .
= 4.5truein figure [ fig : fdf ] shows a different statistic , the probability distribution function ( pdf ) of the flux decrement @xmath49 , for the same set of simulations .
the measurements in the left hand panel are from simulated spectra smoothed with a gaussian of @xmath50 , comparable to the resolution of keck hires or vlt uves spectra .
the sph and n - body+fgpa predictions agree well for @xmath17 .
models with higher fluctuation amplitude have a broader distribution of densities @xmath51 and a correspondingly broader distribution of flux decrements @xmath52 , with more saturated and low - opacity pixels and fewer pixels of intermediate opacity .
the differences in the predicted pdfs for @xmath53 are @xmath54 not enormous , but readily measurable at high statistical significance with reasonable observational samples .
the fraction of saturated pixels ( defined here by @xmath55 ) increases from 7.2% to 8.3% to 9.1% as @xmath38 goes from 0.6 to 0.8 to 1.0 .
there are larger differences in the number of nearly transparent pixels , but continuum fitting uncertainties make it difficult to measure the pdf accurately near @xmath56
. one can investigate the scale dependence of matter clustering by smoothing the spectra ( or by observing them at lower spectral resolution ) , analogous to studying galaxy counts in cells of increasing size .
figure [ fig : fdf ] ( right ) shows results for @xmath57 smoothing .
the predicted pdfs are narrower , since smoothing drives pixel values towards the mean , but the dependence on @xmath38 is similar . the saturated pixel fraction doubles , from 1.9% to 3.8% , as @xmath38 rises from 0.6 to 1 . for cosmological models with gaussian initial conditions ,
the shape of the flux pdf depends mainly on the amplitude of mass fluctuations , once one has chosen @xmath24 to match the mean opacity , @xcite .
the effective physical scale of this amplitude measurement is determined by the spectral smoothing or , for spectra that fully resolve the observed absorption features , by a combination of thermal broadening and gas pressure effects .
models with non - gaussian initial conditions predict significantly different flux pdfs @xcite .
observational tests to date show good agreement with models that have gaussian initial conditions and a @xmath45 amplitude compatible with @xmath11cdm predictions @xcite .
the uncertainties in these tests are comparable to the model differences in figure [ fig : fdf ] .
reducing them requires careful attention to the effects of continuum fitting and noise and to the accuracy of the theoretical predictions , but it should be possible to obtain tight constraints on the normalization of the matter power spectrum , and some information on its shape by studying different smoothing scales .
consistency between results from the flux power spectrum and the flux pdf can be a sensitive diagnostic for primordial non - gaussianity .
= 4.5truein figure [ fig : dcross ] shows the threshold crossing frequency , the number of times per unit redshift that the absorption spectrum crosses a decrement threshold @xmath52 .
following @xcite , we plot @xmath58 as a function of filling factor , the fraction @xmath59 of pixels with flux decrement less than @xmath52 , which cleanly separates the information in @xmath58 from the information in the pdf and makes the model predictions nearly independent of the photoionization rate @xmath24 .
as one would expect from figure [ fig : spec ] , the threshold crossing frequency increases as @xmath35 decreases because a lower hubble parameter @xmath1 `` squeezes '' redshift separations relative to comoving distances .
the threshold crossing frequency drops slightly as the amplitude of @xmath45 increases and gravitationally driven merging `` smooths '' structure @xcite , but the effect is small over the range @xmath60 .
the crossing frequency also drops for redder matter power spectra , since these also lead to smoother structure @xcite . for high spectral resolution
, @xmath58 also depends on the gas temperature , which determines the level of thermal broadening .
the great promise of the threshold crossing statistic is its potential for constraining the hubble parameter at high redshift . in spatially flat models with a cosmological constant ,
the ratio @xmath61 is determined by @xmath35 . alternatively ,
if @xmath35 is known independently , the ratio @xmath61 can constrain the equation of state of dark energy @xcite .
unfortunately , @xmath58 also depends on other cosmological and igm parameters , and it is numerically difficult to predict with high accuracy even when the model is fully specified . the difference between the pm and sph predictions in figure [ fig : dcross ] is comparable to the @xmath35 effects themselves , and the sph result is still affected by the finite numerical resolution . on the observational side , accurate measurement of @xmath58 requires high signal - to - noise spectra .
furthermore , while moderate resolution spectra can provide useful diagnostics for the shape and amplitude of the matter power spectrum @xcite , the @xmath1 application demands high spectral resolution , since smoothing over a scale that is fixed in redshift units erases the sensitivity to @xmath1 ( see fig . [
fig : dcross ] , right ) . exploiting @xmath58 as a probe of dark energy thus represents a theoretical and observational challenge .
one can also use the observable correlations between lyman break galaxies ( lbgs ) and the ly@xmath0 forest to study the environments of high redshift galaxies .
several groups have investigated this issue theoretically using simulations @xcite , and adelberger et al . have carried out an observational study using an lbg survey in fields probed by seven quasar lines of sight ( @xcite , hereafter assp ) . on large scales cubes of comoving size @xmath62 assp find a clear correlation of galaxy overdensity with ly@xmath0 flux decrement , the expected signature of galaxy formation in overdense environments @xcite .
however , the observed ly@xmath0 decrement _ decreases _ within @xmath63 of lbgs ( comoving , redshift - space separation ) , where the simulations predict that absorption should be strongest @xcite .
= 4.5truein figure [ fig : wind ] illustrates this conflict . in the upper left panel , triangles show the predicted mean decrement in bins of @xmath64 , while filled circles show the assp data points .
the predictions come from an sph simulation of a @xmath65 cube , and we select the 40 galaxies with the highest star formation rates to approximate the magnitude limit of assp s spectroscopic survey ( see @xcite for details ) .
plausible random errors in galaxy redshifts can reduce the discrepancy at @xmath66 , but they can not explain the results at the smallest scales , especially the innermost data point at @xmath67 , @xmath68 @xcite .
the right column of figure [ fig : wind ] shows simulated spectra along 12 lines of sight that pass within @xmath69 of a target galaxy , centered on the galaxy redshift space position .
all but one of these spectra have @xmath70 at the galaxy redshift , and half have @xmath71 .
= 4.5truein the obvious conclusion is that feedback from the observed galaxies reduces neutral hydrogen in their immediate surroundings .
local photoionization by the galaxies stars or active nuclei proves insufficient ; this argument can be cast in general terms that seem difficult to escape @xcite .
the natural alternative is some form of supernova or agn - driven wind @xcite . here
( and in @xcite ) we have investigated highly idealized `` wind '' models in which we simply eliminate all neutral hydrogen in a sphere of comoving radius @xmath72 around each target galaxy .
squares , pentagons , and hexagons in figure [ fig : wind ] show results for @xmath73 , 1.0 , and @xmath74 .
stars show a model where we place winds around all 641 resolved galaxies in the simulation volume ( instead of the top 40 that constitute the `` observed '' sample ) and scale the wind volume in proportion to the galaxy baryon mass , with a normalization @xmath75 around the 40th - ranked galaxy .
in the right column , dotted curves show spectra for the @xmath76 model .
the most important lesson from figure [ fig : wind ] is that eliminating neutral gas to a _
real space _
distance @xmath72 does not eliminate absorption within a _ redshift space _
separation @xmath72 ( e.g. , to @xmath77 in the spectrum plots ) .
peculiar infall velocities allow gas at larger distances to produce absorption near the galaxy redshift , making the energetic requirements on any wind explanation of the observed ly@xmath0 deficit much more stringent .
only the mass - scaled and @xmath76 models come close to matching the assp data .
winds that fully ionize or perfectly entrain gas to such large distances are not a natural outcome of hydrodynamic simulations with stellar feedback @xcite , and even reaching @xmath74 in @xmath78gyr requires a sustained propagation speed @xmath79 . given the challenges facing the wind explanation , it is worth considering the alternative possibility that extended ly@xmath0 _ emission _ from the target galaxies is `` filling in '' the corresponding region of the absorption spectrum .
steidel et al .
@xcite observed two extended ly@xmath0 `` blobs '' apparently associated with lbgs , with angular extents @xmath80 and ab apparent magnitudes 21.02 and 21.14 in the ly@xmath0 band . cooling radiation from gas settling into massive galaxies at @xmath10 naturally produces ly@xmath0 fluxes of this order @xcite .
the three quasars that contribute to assp s innermost data point have @xmath81-band ab magnitudes of 20.1 , 21.6 , and 23.4 , so if _ all _ of a target galaxy s ly@xmath0 cooling radiation went down the slit it could potentially replace the quasar flux absorbed by the surrounding igm .
however , a @xmath82 slit at an angular separation @xmath83 from a galaxy should intercept at most @xmath84 of the galaxy s extended ly@xmath0 flux , at least on average , so this explanation seems to fail by @xmath85 orders of magnitude .
furthermore , a fourth pair involving the @xmath86 quasar q0302 - 0019 , shows no sign of absorption near the galaxy redshift , and in this case the quasar is clearly too bright for galaxy emission to compete with it .
( this pair and two others are dropped from assp s @xmath87 calculation because of possible ly@xmath27 contamination . ) at this point , the ly@xmath0 emission explanation seems unlikely ; it can be conclusively ruled out by observing more close pairs involving bright quasars or by obtaining symmetrically placed spectra away from the observed quasars to search for galaxy emission .
assuming for now that winds are the correct explanation for the observed ly@xmath0 deficit , one can ask whether they completely spoil the picture painted in 1 and 2 .
the lower left panel of figure [ fig : wind ] shows the flux power spectrum for the various `` wind '' models .
eliminating neutral hydrogen to distances @xmath73 or @xmath88 around bright galaxies has only a small impact on the flux power spectrum , because the filling factor of the `` bubbles '' is small and absorption close to the galaxies remains nearly saturated in any case .
however , these models also do not explain the assp results .
in the two more extreme models , winds suppress the flux power spectrum on large scales by @xmath89 dex , comparable to the 0.1-dex @xmath90 uncertainty that @xcite quote for the normalization of the matter power spectrum .
thus , winds of this magnitude could have systematic effects on @xmath45 determinations that are significant relative to the present observational uncertainties .
the influence of winds on the flux pdf and threshold crossing frequency is smaller , as shown in figure [ fig : windstats ] .
even the more extreme models have a negligible impact on @xmath58 , and they only slightly alter the shape of the pdf .
the most significant effect is on the fraction of saturated pixels for @xmath57 smoothing , which drops from 2.5% for the no - wind case to 1.9% for @xmath76 and 1.3% for the mass - scaled model .
from the above discussion , it is clear that one immediate challenge is to better understand the galaxy proximity effect .
figures [ fig : wind ] and [ fig : windstats ] show that winds can reach substantial distances from bright galaxies without having much impact on the global statistics of the ly@xmath0 forest ( see also @xcite ) , but even the more extreme wind models considered in 3 do not reproduce the assp results very well
. simulations with more realistic wind physics @xcite can shed further light on this problem , but major progress will have to await further observational studies , since current inferences rest crucially on a handful of galaxy - quasar pairs .
studies with keck hires and vlt uves have provided a ( still growing ) trove of high quality data on the ly@xmath0 forest at @xmath91 . the fluctuating igm perspective
, furthermore , shows that large samples of moderate resolution spectra can be a powerful resource for studying structure on large scales , since continuous flux statistics do not require resolution of individual `` lines . ''
such samples can be assembled quickly with large telescopes , and the sdss is producing an enormous sample at resolution @xmath92 in the course of its normal operations .
one of the observational frontiers is the use of correlations across multiple lines of sight .
quasar pair studies provided the first decisive evidence for large coherence scales of absorbing structures @xcite , but larger samples allow more ambitious goals , such as using the alcock - pacyznski test @xcite to measure spacetime geometry @xcite , improving measurements of the flux power spectrum with cross correlations @xcite , and mapping large scale 3-dimensional structure at high redshift @xcite .
analyses of existing data have already led to an important cosmological conclusion , namely that models with matter clustering similar to that of `` concordance '' @xmath11cdm at @xmath19 are consistent with the observed ly@xmath0 forest while models with substantially different clustering amplitudes or suppression of small scale power are not @xcite .
they have also provided constraints on the temperature of the diffuse igm @xcite and indications of helium reionization at @xmath93 @xcite .
however , to do justice to the quality and quantity of data and keep pace with the tightening observational constraints from other observables , we must play for higher stakes .
it looks possible in principle to achieve precision of @xmath5 on quantities like the matter fluctuation amplitude and @xmath1 , but even without the potential complications of galaxy feedback , inferences at this level require more extensive theoretical modeling .
many effects that are unimportant at the 25% level differences between approximate methods and full hydrodynamics , numerical resolution and box size limitations , spatial fluctuations in igm temperature , details of continuum determination may become critical at the @xmath5 level . despite these challenges ,
the ly@xmath0 forest is the most promising tool we have for precision cosmological measurements at @xmath91 .
these measurements might , in the end , simply confirm the cosmological model favored by other data , but complementary constraints have the potential to break parameter degeneracies and thereby reveal subtle quantitative discrepancies .
these in turn could yield insight into the nature of dark energy , the mechanisms of inflation , or some other fundamental aspect of our universe .
bechtold , j. , crotts , a. p. s. , duncan , r. c. , & fang , y. 1994 , apj , 437 , l83 dinshaw , n. , impey , c. d. , foltz , c. b. , weymann , r. j. , & chaffee , f. h. 1994 , apj , 437 , l87 dinshaw , n. , foltz , c. b. , impey , c. d. , weymann , r. j. , & morris , s. l. 1995 , nature , 373 , 223 crotts , a. p. s. , & fang , y. 1998 , apj , 502 , 16 hu , e.m . ,
kim , t.s . ,
cowie , l.l .
, songaila , a. , & rauch , m. 1995 , aj , 110 , 1526 songaila , a. & cowie , l.l .
1996 , aj , 112 , 335 ellison , s. l. , lewis , g. f. , pettini , m. , chaffee , f. h. , & irwin , m. j. 1999 , apj , 520 , 456 bi , h.g . , 1993 ,
apj , 405 , 479 cen , r. , miralda - escud , j. , ostriker , j.p . , & rauch , m. 1994 , apj , 437 , l9 petitjean , p. , mcket , j. p. , & kates , r. e. 1995 , a&a , 295 , l9 zhang , y. , anninos , p. , & norman , m.l .
1995 , apj , 453 , l57 hernquist l. , katz , n. , weinberg , d.h . , & miralda - escud , j. 1996 , apj , 457 , l51 bi , h.g . , & davidsen , a. 1997 , apj , 479 , 523 hui , l. , gnedin , n. , & zhang , y. 1997 , apj , 486 , 599 theuns , t. , leonard , a. , efstathiou , g. , pearce , f. r. , & thomas , p. a. 1998 , mnras , 301 , 478 gunn , j.e .
, & peterson , b.a .
1965 , apj , 142 , 1633 croft , r.a.c . ,
weinberg , d.h . , katz , n. , hernquist , l. , 1997 , apj , 488 , 532 rauch , m. , miralda - escud , j. , sargent , w. l. w. , barlow , t. a. , weinberg , d. h. , hernquist , l. , katz , n. , cen , r. , & ostriker , j. p. , 1997 ,
apj , 489 , 7 weinberg , d. h. , katz , n. , & hernquist , l. 1998 , in asp conference series 148 , origins , eds . c. e. woodward , j. m. shull , & h. thronson , ( asp : san francisco ) , 21 , astro - ph/9708213 croft , r. a. c. , weinberg , d. h. , katz , n. , & hernquist , l. 1998 , apj , 495 , 44 weinberg , d. h. , et al . 1999 , in evolution of large scale structure : from recombination to garching , eds .
banday , r. k. sheth , & l. n. da costa , ( twin press : vledder nl ) , 346 astro - ph/9810142 bi , h. , ge , j. , & fang , l .- z .
1995 , apj , 452 , 90 gnedin , n. y. , & hui , l. 1998 , mnras , 296 , 44 zaldarriaga , m. , hui , l. , & tegmark , m. 2001 , apj , 557 , 519 viel , m. , matarrese , s. , mo , h. j. , theuns , t. , & haehnelt , m. g. 2002 , mnras , 336 , 685 weinberg , d.h .
, miralda - escud , j. , hernquist , l. , & katz , n. , 1997 , apj , 490 , 564 croft , r. a. c. , weinberg , d. h. , bolte , m. , burles , s. , hernquist , l. , katz , n. , kirkman , d. , tytler , d. 2002 , apj , 581 , 20 gnedin , n. y. & hamilton , a. j. s. 2002 , mnras , 334 , 107 tegmark , m. & zaldarriaga , m. 2002 , prd , 66 , 103508 mcdonald , p. , miralda - escud , j. , rauch , m. , sargent , w. l. w. , barlow , t. a. , cen , r. , & ostriker , j. p. 2000 , apj , 543 , 1 cen , r. 1997 , apj , 479 , l85 miralda - escud j. , cen r. , ostriker , j.p . , & rauch , m. 1996 , apj , 471 , 582 kujat , j. , linn , a. m. , scherrer , r. j. , & weinberg , d. h. 2002 , apj , 572 , 1 viel , m. , matarrese , s. , theuns , t. , munshi , d. , & wang , y. 2002 , mnras , submitted , astro - ph/0212241 mcdonald , p. , miralda - escud ' e , j. , & cen , r. 2002 , apj , 580 , 42 croft , r. a. c. , hernquist , l. , springel , v. , westover , m. , white , m. 2002 , apj , 580 , 634 kollmeier , j.a . , weinberg , d.h . , dav , r. , katz , n. , 2002 , apj , submitted , astro - ph/0209563 bruscoli , m. , ferrara , a. , marri , s. , schneider , r. , maselli , a. , rollinde , e. , aracil , b. , mnras , submitted , astro - ph/0212126 adelberger , k.l . , steidel , c.c . ,
shapley , a.e . , pettini , m. 2002 , apj , in press , astro - ph/0210314 kollmeier , j.a . , weinberg , d.h . ,
dav , r. , & katz , n. , these proceedings theuns , t. , viel , m. , kay , s. , schaye , j. , carswell , r. f. , & tzanavaris , p. 2002 , apj , 578 , l5 steidel , c. c. , adelberger , k. l. , shapley , a. e. , pettini , m. , dickinson , m. , & giavalisco , m. 2000 , apj , 532 , 170 haiman , z. , spaans , m. , & quataert , e. 2000 , apj , 537 , l5 fardal , m. a. , katz , n. , gardner , j. p. , hernquist , l. , weinberg , d. h. & dav , r. 2001 , apj , 562 , 605 alcock , c. , & paczyski , b. 1979 , nature , 281 , 358 hui , l. , stebbins , a. , & burles , s. 1999 , apj , 511 , 5 mcdonald , p. & miralda - escud , j. 1999 , apj , 518 , 24 mcdonald , p. 2001 , apj , submitted , astro - ph/0108064 viel , m. , matarrese , s. , mo , h. j. , haehnelt , m. g. , & theuns , t. 2002 , mnras , 329 , 848 liske , j. , webb , j. k. , williger , g. m. , fern ' andez - soto , a. , & carswell , r. f. 2000 , mnras , 311 , 657 rollinde , e. , petitjean , p. , pichon , c. , colombi , s. , aracil , b. , dodorico , v. , & haehnelt , m. g. 2002 , mnras , submitted croft , r. a. c. , weinberg , d. h. , pettini , m. , katz , n. , & hernquist , l. 1999 , apj , 520 , 1 theuns , t. , leonard , a. , schaye , j. , & efstathiou , g. 1999 , mnras , 303 , l58 narayanan , v. k. , spergel , d. n. , dav ' e , r. , & ma , c. 2000 , apj , 543 , l103 theuns , t. , schaye , j. , & haehnelt , m. g. 2000 , mnras , 315 , 600 meiksin , a. , bryan , g. , & machacek , m. 2001 , mnras , 327 , 296 ricotti , m. , gnedin , n. y. & shull , j. m. 2000 , apj , 534 , 41 mcdonald , p. , miralda - escud , j. , rauch , m. , sargent , w. l. w. , barlow , t. a. , & cen , r. 2000 , apj , 562 , 52 schaye , j. , theuns , t. , rauch , m. , efstathiou , g. & sargent , w. l. w. 2000 , mnras , 318 , 817 theuns , t. , zaroubi , s. , kim , t. , tzanavaris , p. , & carswell , r. f. 2002 , mnras , 332 , 367 theuns , t. , bernardi , m. , frieman , j. , hewett , p. , schaye , j. , sheth , r. k. , & subbarao , m.2002 , apj , 574 , l111 bernardi , m. , et al .
2003 , aj , 125 , 32 | we review recent developments in the theory of the ly@xmath0 forest and their implications for the role of the forest as a test of cosmological models .
simulations predict a relatively tight correlation between the local ly@xmath0 optical depth and the local gas or dark matter density .
statistical properties of the transmitted flux can constrain the amplitude and shape of the matter power spectrum at high redshift , test the assumption of gaussian initial conditions , and probe the evolution of dark energy by measuring the hubble parameter @xmath1 .
simulations predict increased ly@xmath0 absorption in the vicinity of galaxies , but observations show a ly@xmath0 deficit within @xmath2 ( comoving ) .
we investigate idealized models of `` winds '' and find that they must eliminate neutral hydrogen out to comoving radii @xmath3 to marginally explain the data .
winds of this magnitude suppress the flux power spectrum by @xmath4 dex but have little effect on the distribution function or threshold crossing frequency . in light of the stringent demands on winds , we consider the alternative possibility that extended ly@xmath0 emission from target galaxies replaces absorbed flux , but
we conclude that this explanation is unlikely .
taking full advantage of the data coming from large telescopes and from the sloan digital sky survey will require more complete understanding of the galaxy proximity effect , careful attention to continuum determination , and more accurate numerical predictions , with the goal of reaching @xmath5 precision on key cosmological quantities .
address = the ohio state university , dept . of astronomy , columbus , oh 43210 address = university of arizona , dept . of astronomy , tucson , az 85721
address = university of massachusetts , dept . of physics and astronomy , amherst , ma , 91003
address = ohio state university , dept . of astronomy , columbus , oh 43210 |
many formal definitions of key concepts are postponed until section [ sec : basics ] .
we study the complexity of prefix - convex regular languages @xcite . for words @xmath6 over an alphabet @xmath1 ,
if @xmath7 , then @xmath3 is a _ prefix _ of @xmath8 .
a language @xmath9 is _ prefix - convex _ if , whenever @xmath3 and @xmath4 are in @xmath0 , then @xmath5 is also in @xmath0 .
the class of prefix - convex languages includes three well - known subclasses : right - ideal , prefix - closed , and prefix - free languages .
a language @xmath0 is a _
right ideal _ if it is non - empty and satisfies the equation @xmath10 .
right ideals play a role in pattern matching : if one is searching for all words beginning with words in some language @xmath0 in a given text ( language ) , then one is looking for words in @xmath11 .
right ideals also constitute a basic concept in semigroup theory .
a language @xmath0 is _ prefix - closed _ if , whenever @xmath8 is in @xmath0 and @xmath3 is a prefix of @xmath8 , then @xmath3 is also in @xmath0 .
the complement of every right ideal is a prefix - closed language .
the set of allowed input sequences to a digital system is a prefix - closed language .
a language @xmath0 is _ prefix - free _ if no word in @xmath0 is a prefix of another word in @xmath0 .
prefix - free languages ( other than @xmath12 , where @xmath13 is the empty word ) are prefix codes .
they play an important role in coding theory , and have many applications @xcite .
the _ alphabet of a regular language _
@xmath0 is @xmath1 ( or _ @xmath0 is a language over @xmath1 _ ) if @xmath14 and every letter of @xmath1 appears in a word of @xmath0 . the _ ( left ) quotient _ of @xmath0 by a word @xmath15 is @xmath16 .
a language is regular if and only if it has a finite number of distinct quotients .
so the number of quotients of @xmath0 is a natural measure of complexity for @xmath0 ; it is called the _ quotient complexity _ @xcite of @xmath0 and is denoted it by @xmath17 .
an equivalent concept is the _ state complexity _
@xcite of @xmath0 , which is the number of states in a complete minimal deterministic finite automaton ( dfa ) recognizing @xmath0 .
if @xmath18 is a regular language of quotient complexity @xmath19 , and @xmath20 is a unary operation , then the _ quotient / state complexity of @xmath20 _ is the maximal value of @xmath21 , expressed as a function @xmath22 , as @xmath18 ranges over all regular languages of complexity @xmath19 . if @xmath23 and @xmath18 are regular languages of quotient complexities @xmath24 and @xmath19 respectively , and @xmath20 is a binary operation , then the _ quotient / state complexity of @xmath20 _ is the maximal value of @xmath25 , expressed as a function @xmath26 , as @xmath23 and @xmath18 range over all regular languages of complexity @xmath24 and @xmath19 , respectively .
the quotient / state complexity of an operation gives a worst - case lower bound on the time and space complexities of the operation , and has been studied extensively @xcite ; we refer to quotient / state complexity simply as _
complexity_. in all the past literature on binary operations
it has always been assumed that the alphabets of the two operands are restricted to be the same .
however , it has been shown recently @xcite that this is an unnecessary restriction : larger complexity bounds can be reached in some cases if the alphabets differ . in the present paper
we examine both _ restricted complexity _ of binary operations , where the alphabets must be the same , and _ unrestricted complexity _ where they may differ . to find the complexity of a unary operation one first finds an upper bound on this complexity , and then exhibits languages that meet this bound .
since we require a language @xmath18 for each @xmath27 , we need a sequence @xmath28 ; here @xmath29 is usually a small integer because the bound may not hold for a few small values of @xmath19 .
we call such a sequence a _ stream _ of languages .
usually the languages in a stream have the same basic structure and differ only in the parameter @xmath19 .
for example , @xmath30 is a stream . for a binary operation
we require two streams .
there exists a stream @xmath31 of regular languages @xmath32 which meets the complexity bounds for reversal , ( kleene ) star , product ( concatenation ) , and all binary boolean operations @xcite .
there are reasons , however , why the complexity of languages is not a sufficiently good measure .
two languages may have the same complexity @xmath19 but the syntactic semigroup @xcite of one may have @xmath33 elements , while that of the other has @xmath34 elements @xcite .
for this reason , the size of the syntactic semigroup of a language which is the same as the size of the transition semigroup of a minimal dfa accepting the language @xcite has been added as another complexity measure . secondly , _
star - free _ languages , which constitute a very special subclass of the class of regular languages , meet the complexity bounds of regular languages for all operations except reversal , which only reaches the bound @xmath35 instead of @xmath36 @xcite . while regular languages are the smallest class containing the finite languages and closed under boolean operations , product and star , star - free languages are the smallest class containing the finite languages and closed only under boolean operations and product
. quotient / state complexity does not distinguish these two classes .
the complexities of the atoms of a regular language have been proposed as an additional measure @xcite .
atoms are defined by the following left congruence : two words @xmath3 and @xmath37 are equivalent if @xmath38 if and only if @xmath39 for all @xmath40 . thus @xmath3 and @xmath37 are equivalent if @xmath41 if and only if @xmath42 .
an equivalence class of this relation is an _ atom _ of @xmath0 @xcite .
thus an atom is a non - empty intersection of complemented and uncomplemented quotients of @xmath0 .
if @xmath43 are the quotients of @xmath0 , and @xmath44 , then atom @xmath45 is the intersection of quotients with subscripts in @xmath46 and complemented quotients with subscripts in @xmath47 . for more information about atoms
see @xcite .
a language stream that meets the restricted and unrestricted complexity bounds for all boolean operations , product , star , and reversal , and also has the largest syntactic semigroup and most complex atoms @xcite , is called a
_ most complex stream_. a most complex stream should have the smallest possible alphabet sufficient to meet all the bounds .
here we present a regular language stream similar to that of @xcite but one that is better suited to prefix - convex languages .
most complex streams are useful when one designs a system dealing with regular languages and finite automata .
if one would like to know the maximal sizes of automata the system can handle , one can use the one most complex stream to test all the operations .
in this paper we exhibit most complex language streams for right - ideal , prefix - closed , prefix - free , and proper prefix - convex languages , where a prefix - convex language is _ proper _ if it is not one of the first three types .
a _ deterministic finite automaton ( dfa ) _ is a quintuple @xmath48 , where @xmath49 is a finite non - empty set of _ states _ , @xmath1 is a finite non - empty _ alphabet _ , @xmath50 is the _ transition function _ , @xmath51 is the _ initial _ state , and @xmath52 is the set of _ final _ states .
we extend @xmath53 to a function @xmath54 as usual .
a dfa @xmath55 _ accepts _ a word @xmath56 if @xmath57 . the language accepted by @xmath55
is denoted by @xmath58 .
if @xmath59 is a state of @xmath55 , then the language @xmath60 of @xmath59 is the language accepted by the dfa @xmath61 .
a state is _ empty _ or _ dead _ or _ a sink _ if its language is empty .
two states @xmath62 and @xmath59 of @xmath55 are _ equivalent _ if @xmath63 ; otherwise they are _ distinguishable_.
a state @xmath59 is _ reachable _ if there exists @xmath15 such that @xmath64 .
a dfa is _ minimal _ if all of its states are reachable and no two states are equivalent .
usually dfas are used to establish upper bounds on the complexity of operations and also as witnesses that meet these bounds .
a _ nondeterministic finite automaton ( nfa ) _ is a quintuple @xmath65 , where @xmath49 , @xmath1 and @xmath66 are as in a dfa , @xmath67 , and @xmath68 is the _ set of initial states_. an _
@xmath13-nfa _ is an nfa in which transitions under the empty word @xmath13 are also permitted . without loss of generality
we use @xmath69 as the set of states of every dfa with @xmath19 states .
a _ transformation _ of @xmath70 is a mapping @xmath71 .
the _ image _ of @xmath72 under @xmath73 is denoted by @xmath74 . in any dfa , each letter @xmath75 induces a transformation @xmath76 of the set @xmath70 defined by @xmath77 ; we denote this by @xmath78 . by a slight abuse of notation we use the letter @xmath79 to denote the transformation it induces ; thus we write @xmath80 instead of @xmath81 .
we extend the notation to sets of states : if @xmath82 , then @xmath83 . we also write @xmath84 to mean that the image of @xmath85 under @xmath79 is @xmath86 . if @xmath87 are transformations of @xmath70 , their composition is denoted by @xmath88 and defined by @xmath89 ; the @xmath90 is usually omitted .
let @xmath91 be the set of all @xmath34 transformations of @xmath70 ; then @xmath91 is a monoid under composition . for @xmath92 ,
a transformation ( permutation ) @xmath73 of a set @xmath93 is a _ @xmath29-cycle _ if @xmath94 .
this @xmath29-cycle is denoted by the transformation @xmath95 of @xmath70 , which acts as the identity on the states outside the cycle .
a 2-cycle @xmath96 is called a _
transposition_. a transformation that sends all the states of @xmath85 to @xmath59 and acts as the identity on the remaining states is denoted by @xmath97 . if @xmath98 we write @xmath99 for @xmath100
. the identity transformation is denoted by @xmath101 .
the notation @xmath102 denotes a transformation that sends @xmath59 to @xmath103 for @xmath104 and is the identity for the remaining states , and @xmath105 is defined similarly .
let @xmath106 be a dfa .
for each word @xmath56 , the transition function induces a transformation @xmath107 of @xmath70 by @xmath8 : for all @xmath108 , @xmath109 the set @xmath110 of all such transformations by non - empty words forms a semigroup of transformations called the _ transition semigroup _ of @xmath55 @xcite .
we can use a set @xmath111 of transformations to define @xmath53 , and so the dfa @xmath55 .
the _ myhill congruence
_ @xcite @xmath112 of a language @xmath9 is defined on @xmath113 as follows : @xmath114 this congruence is also known as the _ syntactic congruence _ of @xmath0 .
the quotient set @xmath115 of equivalence classes of the relation @xmath112 is a semigroup called the _ syntactic semigroup _ of @xmath0 .
if @xmath55 is a minimal dfa of @xmath0 , then @xmath110 is isomorphic to the syntactic semigroup @xmath116 of @xmath0 @xcite , and we represent elements of @xmath116 by transformations in @xmath110 .
the size of the syntactic semigroup has been used as a measure of complexity for regular languages @xcite .
recall that binary operations require two language streams to determine the complexity of the operation . sometimes the same stream can be used for both operands , and it has been shown in @xcite that for all common binary operations on regular languages the second stream can be a `` dialect '' of the first , that is , it can `` differ only slightly '' from the first and all the bounds can still be met .
let @xmath117 be an alphabet ordered as shown ; if @xmath9 , we denote it by @xmath118 to stress its dependence on @xmath1 .
a _ dialect _ of @xmath0 is a related language obtained by replacing or deleting letters of @xmath1 in the words of @xmath0 .
more precisely , for an alphabet @xmath119 and a partial map @xmath120 , we obtain a dialect of @xmath0 by replacing each letter @xmath121 by @xmath122 in every word of @xmath0 , or deleting the word entirely if @xmath122 is undefined .
we write @xmath123 to denote the dialect of @xmath118 given by @xmath124 , and we denote undefined values of @xmath124 by `` @xmath125 '' . for example , if @xmath126 then its dialect @xmath127 is the language @xmath128 .
undefined values for letters at the end of the alphabet are omitted ; thus , for example , if @xmath129 , @xmath130 , @xmath131 , @xmath132 and @xmath133 , we write @xmath134 for @xmath135 . the language stream that meets all the complexity bounds
is referred to as the _ master _ language stream .
every master language stream we present here uses the smallest possible alphabet sufficient to meet all the complexity bounds .
individual bounds are frequently met by dialects on reduced alphabets , and we prefer to use the smallest alphabet possible for each bound . for binary operations ,
we try to minimize the size of the combined alphabet of the two dialects .
as each letter induces a transformation on the states of a dfa ( or equivalently , the quotients of a language ) we count the number of distinct transformations induced by letters of the alphabet . in any language
this number is at most the size of the alphabet , but there may be multiple letters which induce the same transformation ; this does not occur in this paper as no language has a repeated transformation . for binary operations on two dialects of the same master language ,
we count the number of distinct transformations of the master language present in either dialect . for example , suppose @xmath136 and @xmath137 are two dialects of a language @xmath138 , which we assume has four distinct transformations .
each dialect has three letters and three distinct transformations , and between them they have three letters and four distinct transformations . although a given complexity bound may be met by many dialects of the master language , we favour dialects , or pairs of dialects , that use small alphabets and few distinct transformations . in many cases the dialects we present are minimal in these respects , though we do not always justify this .
we now define a dfa stream that we use as a basic component .
it is similar to the stream defined in @xcite for the case of equal alphabets , except that there the transformation induced by @xmath139 is @xmath140 .
it is also similar to the dfa of @xcite , except that there the transformation induced by @xmath139 is @xmath140 and an additional input @xmath141 is used .
[ def : regular ] for @xmath142 , let @xmath143 , where @xmath144 , and @xmath145 is defined by the transformations @xmath146 , @xmath147 , and @xmath148 .
let @xmath149 be the language accepted by @xmath150 .
the structure of @xmath151 is shown in figure [ fig : regwit ] .
8.5pt ( 37,7)(0,3 ) ( 3dots)(22,7)@xmath152 ( n-2)(29,7)@xmath153 ( n-1)(36,7)@xmath33(n-1 ) ( 0)@xmath139 ( 0,1)@xmath154 ( 1,0)@xmath155 ( 1,2)@xmath79 ( 2)@xmath155 ( 2,3dots)@xmath79 ( 3dots , n-2)@xmath79 ( n-2)@xmath155 ( n-2,n-1)@xmath79 ( n-1,0)@xmath79 ( n-1)@xmath155 [ thm : regular ] for each @xmath142 , the dfa of definition [ def : regular ] is minimal and its language @xmath156 has complexity @xmath19 .
the stream @xmath157 with some dialect streams is most complex in the class of regular languages . in particular
, it meets all the complexity bounds below .
at least three letters are required in any witness meeting all these bounds and a total of four distinct letters is required for any two witnesses for unrestricted union and symmetric difference . in several cases
the bounds can be met with a smaller alphabet as shown below . 1 .
the syntactic semigroup of @xmath156 has cardinality @xmath34 .
each quotient of @xmath158 has complexity @xmath19 .
3 . the reverse of @xmath156 has complexity @xmath36 , and @xmath156 has @xmath36 atoms .
each atom @xmath45 of @xmath156 has maximal complexity : @xmath159 5 .
the star of @xmath160 has complexity @xmath161 .
6 . 1 .
restricted complexity : + the product @xmath162 has complexity @xmath163 .
unrestricted complexity : + the product @xmath164 has complexity @xmath165 .
7 . 1 .
restricted complexity : + the complexity of @xmath166 is @xmath167 for @xmath168 .
unrestricted complexity : + the complexity of union and symmetric difference is @xmath169 and this bound is met by @xmath170 and @xmath171 , that of difference is @xmath172 and this bound is met by @xmath170 and @xmath173 , and that of intersection is @xmath167 and this bound is met by @xmath174 and @xmath173 .
a total of four letters is required to meet the bounds for union and symmetric difference .
clearly @xmath158 has complexity @xmath19 as the dfa of definition [ def : regular ] is minimal . 1 .
* syntactic semigroup * the transformations @xmath175 , @xmath176 , and @xmath177 were used in @xcite .
it is well known that these transformations as well as @xmath79 , @xmath178 , and @xmath179 generate the semigroup of all transformations of @xmath70 .
* quotients * obvious .
* reversal * this follows from a theorem in @xcite which states that if the transition semigroup has @xmath34 elements , then the complexity of reversal is @xmath36 .
it was shown in @xcite that the number of atoms is the same as the complexity of the reverse .
* atoms * proved in @xcite .
* star * proved in @xcite .
* product * let @xmath180 and @xmath181 be minimal dfas of languages @xmath182 and @xmath0 , respectively .
we use the standard construction of the @xmath13-nfa @xmath183 for the product @xmath184 : the final states of @xmath185 becomes non - final , and an @xmath13-transition is added from each state of @xmath186 to the initial state @xmath187 of @xmath55 . + the subset construction on this nfa yields sets @xmath188 where @xmath189 and @xmath190 and sets @xmath191 where @xmath192 and @xmath193 , as well as sets @xmath190 which can only be reached by letters in @xmath194 .
hence the restricted complexity of @xmath184 is bounded by @xmath195 , and the unrestricted complexity of @xmath184 is bounded by @xmath196 .
+ 8.5pt + ( 38,15)(-3,2.8 ) ( 0)(-4,15)@xmath197(0 ) ( 1)(3,15)@xmath198 ( 3dots)(10,15)@xmath152 ( m-2)(17,15)@xmath199 ( m-2)@xmath178 ( m-1)(17,11)@xmath200 ( m-1)@xmath178 + ( 0,1)@xmath154 ( 1,0)@xmath178 ( 1,3dots)@xmath79 ( 3dots,m-2)@xmath79 ( m-2,m-1)@xmath79 ( m-1,0)@xmath79 + ( 0)(11,6)0 ( 0)@xmath178 ( 1)(18,6)1 ( 3dots)(25,6)@xmath152 ( n-2)(32,6)@xmath153 ( n-2)@xmath155 ( n-1)(39,6)@xmath33(n-1 ) ( n-1)@xmath155 ( 0,1)@xmath201 ( 1,0)@xmath155 ( 1,3dots)@xmath79 ( 3dots , n-2)@xmath79 ( n-1,0)@xmath79 ( n-2,n-1)@xmath79 ( m-1,0)@xmath13 + _ restricted complexity : _ consider @xmath174 and @xmath202 of definition [ def : regular ] ; we show that their product meets the upper bound for restricted complexity . as before
, we construct an nfa recognizing @xmath203 and then apply the subset construction to obtain a dfa .
figure [ fig : regularproduct ] shows the nfa for the unrestricted product @xmath204 ; the product @xmath203 is the same except @xmath139 is omitted .
+ the initial state is @xmath205 and each state @xmath206 for @xmath207 is reached by @xmath208 .
consider @xmath209 , where @xmath210 with @xmath211 .
if @xmath212 then @xmath213 . if @xmath214 and @xmath215 , then @xmath216 .
state @xmath217 is reached by @xmath218 .
hence for any non - empty @xmath190 , state @xmath209 is reachable from @xmath219 for some @xmath220 of size @xmath221 .
we reach @xmath188 from @xmath222 by @xmath208 , where @xmath223 denotes @xmath224 taken mod @xmath19 . by induction
, @xmath188 is always reachable and thus all @xmath163 states are reachable .
+ we check that all states are pairwise distinguishable . 1 . any two sets which differ by @xmath225 are distinguished by @xmath226 .
2 . states @xmath227 and @xmath228 with @xmath229 are distinguished by @xmath230 .
3 . states @xmath217 and @xmath231 are distinguished by @xmath232 if @xmath233 ; otherwise apply @xmath234 to simplify to this case .
4 . states @xmath235 and @xmath236 reduce to case ( c ) by @xmath237 . 5 . states
@xmath238 and @xmath239 , where @xmath240 , reduce to case ( d ) by @xmath241 since @xmath242 and @xmath241 permutes @xmath243 .
+ we can distinguish any pair of states ; hence the complexity of @xmath203 is @xmath163 for all @xmath244 .
+ _ unrestricted complexity : _ the nfa for the product of @xmath204 is illustrated in figure [ fig : regularproduct ] . the nfa is the same as the restricted case except it has the additional transformation @xmath245 .
hence the subset construction yields the @xmath246 sets of the restricted case , as well as all sets @xmath190 since @xmath46 is reachable from @xmath209 by @xmath247 .
we check that these sets are distinguishable from all previously reached sets .
any two sets which differ by @xmath225 are distinguished by @xmath226 .
state @xmath206 is distinguishable from @xmath248 by @xmath249 .
3 . states @xmath217 and @xmath250 are distinguished by @xmath251 if @xmath252 is not a multiple of @xmath19 , and by @xmath253 otherwise .
4 . states @xmath231 and @xmath250 reduce to case ( c ) by @xmath254 . 5 . states @xmath188 and @xmath46 , where @xmath240 , reduce to case ( d ) by @xmath241 since @xmath242 and @xmath241 permutes @xmath243 .
+ hence @xmath204 has complexity @xmath255 .
* boolean operations * + _ restricted complexity : _ all boolean operations have complexity at most @xmath167 @xcite .
applying the standard construction for boolean operations we consider the direct product of @xmath256 and @xmath257 which has states @xmath258 ; the final states vary depending on the operation . by
* theorem 1 ) and computation for the cases @xmath259 , the states of @xmath258 are reachable and pairwise distinguishable for each operation @xmath260 ; hence each operation has complexity @xmath167 .
+ note that two letters are required to meet these bounds : to a contradiction suppose a single letter @xmath261 is sufficient to reach @xmath262 in the direct product , where @xmath263 are not coprime .
letter @xmath261 must induce an @xmath24-element permutation on @xmath243 ; otherwise there is an unreachable state in @xmath243 or the sequence @xmath264 never returns to @xmath197 .
similarly @xmath261 must induce an @xmath19-cycle in @xmath70 .
hence @xmath261 has order @xmath265 in the direct product .
it must have order @xmath167 in the direct product , which only occurs when @xmath24 and @xmath19 are coprime .
+ _ unrestricted complexity : _ the upper bounds on the unrestricted complexity of boolean operations are derived in @xcite . to compute @xmath266 , where @xmath20 is a boolean operation ,
add an empty state @xmath267 to @xmath268 , and send all the transitions from any state of @xmath243 under @xmath141 to @xmath267 .
similarly , add an empty state @xmath248 to @xmath269 together with appropriate transitions ; now the alphabets of the resulting dfas are the same .
we consider the direct product of @xmath270 and @xmath271 which has states @xmath272 .
a dfa recognizing @xmath273 is shown in figure [ fig : regularcross ] for @xmath274 and @xmath275 .
+ ( 35,21)(0,-3 ) + ( 04,14)@xmath79 ( 14,24)@xmath79 ( 24,04)@xmath79 + ( 20,30)@xmath141 ( 21,30)@xmath141 ( 22,32)@xmath141 ( 23,33)@xmath141 ( 24,34)@xmath141 + ( 30,31)@xmath178 ( 31,32)@xmath178 ( 32,33)@xmath178 ( 33,30)@xmath178 + ( 03,04)@xmath139 ( 13,04)@xmath139 ( 23,24)@xmath139 ( 33,34)@xmath139 + as in the restricted case all the states of @xmath258 are reachable by words in @xmath276 . since @xmath79 and @xmath178 induce permutations on the direct product , it follows that every state in @xmath258 is reachable from every other .
the remaining states in @xmath277 and @xmath278 are easily reachable using @xmath139 and @xmath141 in addition to @xmath79 and @xmath178 . hence all @xmath279 states are reachable .
+ for union and symmetric difference , the states of @xmath280 are pairwise distinguishable by words in @xmath281 and they are distinguished from all other states by words in @xmath282 .
similarly the states of @xmath283 are distinguishable from each other and all other states ; hence all @xmath169 states are distinguishable .
+ for difference , the final states are @xmath284 for @xmath285 . the states of @xmath283 are all empty , and they are only reachable by @xmath141 . as the words of @xmath286
do not contain @xmath141 , the alphabet is @xmath287 ; hence we can omit @xmath141 and delete the states of @xmath283 , and be left with a dfa recognizing the same language .
we check that the remaining @xmath172 states are pairwise distinguishable .
any states @xmath288 and @xmath289 where @xmath290 and @xmath291 are distinguished by words in @xmath281 .
state @xmath292 is distinguished from @xmath293 by some @xmath294 that maps @xmath293 to @xmath295 , since @xmath8 must send @xmath296 to the final state @xmath297 . + for intersection the only final state is @xmath295 .
the alphabet of @xmath298 is @xmath299 ; hence we can omit @xmath139 and @xmath141 and delete the states of @xmath300 , and be left with a dfa recognizing the same language . the remaining @xmath167 states are pairwise distinguishable as in the restricted case .
+ note that a total of four letters between the alphabets @xmath119 of @xmath301 and @xmath1 of @xmath150 is required for union and symmetric difference . as in the restricted case ,
two letters in @xmath302 are required to reach the states of @xmath258 for general values of @xmath24 and @xmath19 .
letters in both alphabets can not be used to reach states @xmath292 and @xmath303 as the empty states in each coordinate are only reached by letters outside the corresponding alphabet .
thus two additional letters are required , one in @xmath304 and one in @xmath194 .
hence each alphabet must contain at least three letters , and @xmath305 must contain at least four .
in contrast , the bound for difference is met by @xmath170 and @xmath173 , and the bound for intersection is met by @xmath174 and @xmath173 .
the results in this section are based on @xcite ; however , the stream below is different from that of @xcite , where @xmath306 and @xmath307 .
[ def : ideal ] for @xmath308 , let @xmath309 , where @xmath310 and @xmath145 is defined by the transformations @xmath311 , @xmath147 , @xmath148 , and @xmath312 .
let @xmath313 be the language accepted by @xmath314 .
for the structure of @xmath315 see figure [ fig : ridealwit ] . 8.5pt ( 37,8)(0,-4 ) ( 0)@xmath139 ( 0,1)@xmath316 ( 1,0)@xmath155 ( 1,2)@xmath317 ( 2)@xmath155 ( 2,3dots)@xmath317 ( 3dots , n-2)@xmath317 ( n-2)@xmath155 ( n-2,n-1)@xmath141 ( n-2,0)@xmath79 ( n-1)@xmath318 [ thm : rightideals ] for each @xmath308 , the dfa of definition [ def : ideal ] is minimal and @xmath32 is a right ideal of complexity @xmath19 . the stream @xmath319 with some dialect streams is most complex in the class of right ideals . in particular , it meets all the bounds below .
at least four letters are required to meet these bounds . 1 .
the syntactic semigroup of @xmath32 has cardinality @xmath320 .
there is only one maximal transition semigroup of a minimal dfa accepting a right ideal , since it consists of all the transformations of @xmath70 that fix @xmath33 .
at least four letters are needed for this bound .
2 . the quotients of @xmath321 have complexity @xmath19 , except that @xmath322 .
the reverse of @xmath321 has complexity @xmath323 , and @xmath321 has @xmath323 atoms .
each atom @xmath45 of @xmath32 has maximal complexity : @xmath324 5 .
the star of @xmath321 has complexity @xmath325
. restricted complexity : + the product @xmath326 has complexity @xmath327 .
unrestricted complexity : + the product @xmath328 has complexity @xmath329 .
at least three letters for each language and four letters in total are required to meet this bound .
restricted complexity : + the complexity of @xmath20 is @xmath167 if @xmath330 , @xmath331 if @xmath332 , and @xmath333 if @xmath334 , and these bounds are met by @xmath335 .
at least two letters are required to meet these bounds .
unrestricted complexity : + the complexity of @xmath336 is the same as for arbitrary regular languages : @xmath169 if @xmath337 , @xmath172 if @xmath332 , and @xmath167 if @xmath338 . at least three letters in each language and four letters in total are required to meet the bounds for intersection and symmetric difference .
the bound for difference is also met by @xmath339 and the bound for intersection is met by @xmath340 .
[ thm : ideal ] dfa @xmath341 is minimal because the shortest word in @xmath342 accepted by state @xmath59 is @xmath343 , and @xmath32 is a right ideal because it has only one final state and that state accepts @xmath344 . 1 .
* semigroup * the transformations induced by @xmath79 , @xmath178 , and @xmath139 generate all transformations of @xmath345 .
also , since the transformation induced by @xmath346 is @xmath347 , the transition semigroup of @xmath348 contains the one in @xcite , which is maximal for right ideals .
hence the syntactic semigroup of @xmath32 has size @xmath320 as well .
the fact that at least four letters are needed was proved in @xcite .
* quotients * if the initial state of @xmath349 is changed to @xmath59 with @xmath350 , the new dfa accepts a quotient of @xmath18 and is still minimal ; hence the complexity of that quotient is @xmath19 .
* reversal * it was proved in @xcite that the reverse has complexity @xmath323 , and in @xcite that the number of atoms is the same as the complexity of the reverse .
* atoms * the proof in @xcite applies since the dfa has all the transformations that fix @xmath33 .
* star * if @xmath18 is a right ideal , then @xmath351 . if we add a new initial state @xmath197 to the dfa of definition [ def : ideal ] with the same transitions as those from @xmath187 and make @xmath197 final , the new dfa accepts @xmath352 and is minimal
@xmath197 does not accept @xmath79 , and so is not equivalent to @xmath33 .
* product * let @xmath353 and @xmath354 be minimal dfas of @xmath182 and @xmath0 , respectively , where @xmath182 and @xmath0 are right ideals .
we use the standard construction of the nfa for the product @xmath184 : the final state @xmath200 of @xmath185 becomes non - final , and an @xmath13-transition is added from that state to the initial state @xmath187 of @xmath55 .
we bound the complexity of the product by counting the reachable states in the subset construction on this nfa .
the @xmath252 non - final states @xmath206 of @xmath185 may be reachable , as well as @xmath355 . from @xmath355
we may reach all @xmath356 subsets of @xmath70 which contain 0 but not @xmath33 , and @xmath356 states that contain both 0 and @xmath33 ; however , the latter @xmath356 states all accept @xmath344 and are therefore equivalent .
so far , we have @xmath357 states ; these are the only reachable sets if the witnesses are restricted to the same alphabet . + for the unrestricted case , suppose that @xmath358 and @xmath359 . by applying @xmath261 to @xmath360 , @xmath361 , we may reach all @xmath35 non - empty subsets of @xmath70 , and then by applying @xmath362 we reach the empty subset .
however , the @xmath323 subsets of @xmath70 that contain @xmath33 all accept @xmath344 .
hence there are at most @xmath363 additional sets , for a total of @xmath364 reachable sets .
+ 8.5pt + ( 38,17)(-4,2.5 ) ( 0)(-4,14)@xmath197(0 ) ( 0)@xmath139 ( 1)(3,14)@xmath198 ( 2)(10,14)@xmath365 ( 2)@xmath139 ( 3dots)(17,14)@xmath152 ( m-2)(24,14)@xmath199 ( m-2)@xmath139 ( m-1)(24,10)@xmath200 + ( m-1)@xmath1 ( 0,1)@xmath317 ( 1,0)@xmath139 ( 1,2)@xmath317 ( 2,3dots)@xmath317 ( 3dots,m-2)@xmath317 ( m-2,0)@xmath79 ( m-2,m-1)@xmath141 + ( 0)(6,6)0 ( 0)@xmath139 ( 1)(13,6)1 ( 2)(20,6)2 ( 2)@xmath139 ( 3dots)(27,6)@xmath152 ( n-2)(34,6)@xmath153 ( n-2)@xmath139 ( n-1)(34,10)@xmath33(n-1 ) ( n-1)@xmath1 ( 0,1)@xmath317 ( 1,0)@xmath139 ( 1,2)@xmath317 ( 2,3dots)@xmath317 ( 3dots , n-2)@xmath317 ( n-2,0)@xmath79 ( n-2,n-1)@xmath141 ( m-1,0)@xmath13 + _ restricted complexity : _ to prove the bound is tight , consider the two dialects of the dfa of definition [ def : ideal ] shown in figure [ fig : ridealproduct1 ] .
the @xmath252 sets @xmath206 for @xmath366 are reachable in @xmath367 by words in @xmath342 , and @xmath355 is reached by @xmath368 .
the @xmath356 sets of the form @xmath360 , where @xmath361 , are reachable using words in @xmath369 as follows : to reach @xmath360 , where @xmath370 , @xmath371 , we have first @xmath372 .
state 1 will then be moved to the right by applying either @xmath141 or @xmath373 repeatedly : if @xmath374 , use @xmath141 ; otherwise use @xmath373 . repeating this process @xmath375 times
we eventually construct @xmath46 .
for example , to reach @xmath376 use @xmath377 .
the @xmath356 sets @xmath378 that contain @xmath33 all accept @xmath379 ; hence they are all equivalent .
+ the remaining states are pairwise distinguishable : states @xmath206 and @xmath380 with @xmath381 are distinguished by @xmath382 , and @xmath206 is distinguished from @xmath360 by @xmath383 .
two non - final states @xmath360 and @xmath384 with @xmath385 are distinguished by @xmath386 .
thus the product has complexity @xmath327 .
+ 8.5pt + ( 38,17)(-4,2.5 ) ( 0)(-4,14)@xmath197(0 ) ( 0)@xmath139 ( 1)(3,14)@xmath198 ( 2)(10,14)@xmath365 ( 2)@xmath139 ( 3dots)(17,14)@xmath152 ( m-2)(24,14)@xmath199 ( m-2)@xmath139 ( m-1)(24,10)@xmath200 + ( m-1)@xmath119 ( 0,1)@xmath317 ( 1,0)@xmath139 ( 1,2)@xmath317 ( 2,3dots)@xmath317 ( 3dots,m-2)@xmath317 ( m-2,0)@xmath79 ( m-2,m-1)@xmath141 + ( 0)(6,6)0 ( 0)@xmath139 ( 1)(13,6)1 ( 2)(20,6)2 ( 2)@xmath139 ( 3dots)(27,6)@xmath152 ( n-2)(34,6)@xmath153 ( n-2)@xmath139 ( n-1)(34,10)@xmath33(n-1 ) ( n-1)@xmath1 ( 0,1)@xmath387 ( 1,0)@xmath139 ( 1,2)@xmath387 ( 2,3dots)@xmath387 ( 3dots , n-2)@xmath387 ( n-2,0)@xmath178 ( n-2,n-1)@xmath141 ( m-1,0)@xmath13 + _ unrestricted complexity : _ consider two dialects of the dfa of definition [ def : ideal ] shown in figure [ fig : ridealproduct2 ] . here
@xmath388 and @xmath389 . by the restricted case ,
all states @xmath206 for @xmath366 and @xmath378 for @xmath390 are reachable by words in @xmath369 .
apply @xmath178 from @xmath205 to reach the empty subset . by applying @xmath178 to @xmath360 , @xmath361 ,
we reach all @xmath35 non - empty subsets of @xmath70 ; hence all states are reachable .
however , the @xmath323 sets @xmath190 that contain @xmath33 all accept @xmath391 and are sent to the empty state by @xmath79 ; hence they are all equivalent .
similarly , the @xmath356 sets @xmath378 that contain @xmath33 all accept @xmath391 and are sent to @xmath392 by @xmath79 ; hence they are also equivalent . + the remaining states are pairwise distinguishable .
states @xmath206 and @xmath380 with @xmath381 are distinguished by @xmath382 , and @xmath206 is distinguished from @xmath360 by @xmath383 or from @xmath46 , where @xmath393 by @xmath383 .
two states @xmath360 and @xmath384 with @xmath394 are distinguished by @xmath395 .
two states @xmath46 and @xmath396 with @xmath394 are distinguished by @xmath395 .
a state @xmath360 is distinguishable from @xmath396 where @xmath397 by @xmath398 .
thus all @xmath364 states are pairwise distinguishable .
+ at least three inputs to each dfa are required to achieve the bound in the unrestricted case : there must be a letter in @xmath1 ( like @xmath141 ) with a transition to @xmath33 to reach sets containing @xmath33 , and this letter must be in @xmath119 in order to reach the sets that contain both @xmath200 and @xmath33 .
however no single letter in @xmath302 is sufficient to reach every set of the form @xmath360 , regardless of its behaviour on @xmath70 .
for example , if the letter maps @xmath399 and @xmath400 then it is impossible to reach the state @xmath401 by repeatedly applying the letter from @xmath355 , as it can never delete @xmath402 .
hence there must be at least two letters in @xmath302 .
furthermore there must be some @xmath359 to reach the empty state , and there must be some @xmath403 to distinguish @xmath404 from @xmath405 .
thus each alphabet must contain at least three letters to meet the bound .
* boolean operations * + _ restricted complexity : _ the bounds for right ideals were derived in @xcite .
we show that the dfas @xmath406 and @xmath407 of the right ideals of definition [ def : ideal ] meet the bounds .
+ ( 35,22)(0,-2 ) + ( 00,10)@xmath141 ( 01,10)@xmath141 ( 02,12)@xmath141 ( 03,13)@xmath317 + ( 10,20)@xmath141 ( 11,20)@xmath141 ( 12,22)@xmath141 ( 13,23)@xmath317 ( 23,03)@xmath79 + ( 20,01)@xmath79 ( 20,30)@xmath141 ( 21,30)@xmath141 ( 22,32)@xmath141 ( 23,33)@xmath141 + ( 30,31)@xmath79 ( 31,32)@xmath79 ( 32,33)@xmath79 + ( 00,11)@xmath79 ( 10,21)@xmath79 ( 01,12)@xmath79 ( 11,22)@xmath79 ( 21,02)@xmath79 ( 02,13)@xmath79 ( 12,23)@xmath79 ( 22,03)@xmath79 + ( 30)@xmath141 ( 31,30)@xmath141 ( 32)@xmath141 ( 33)@xmath317 + consider the direct product of @xmath408 and @xmath409 , illustrated in figure [ fig : idealcross1 ] for @xmath410 . for @xmath366 state @xmath411 is reached by @xmath412 . since the first column of @xmath413 is reachable and @xmath414 , where @xmath415 is taken mod @xmath252 , we can reach every state in @xmath413 .
state @xmath284 is reached by @xmath416 ; hence the states of @xmath258 are reachable .
+ we now check distinguishability , which depends on the final states of the dfa .
the direct product is made to recognize @xmath335 by setting the final states to be @xmath417 .
+ for intersection and symmetric difference , all states are pairwise distinguishable .
states that differ in the first coordinate are distinguished by words in @xmath418 and states that differ in the second coordinate are distinguished by words in @xmath419 .
hence the complexity is @xmath167 .
+ for difference , the states @xmath420 are all empty , and therefore equivalent .
the remaining states are non - empty , and they are distinguished by words in @xmath342 if they differ in the first coordinate or by words in @xmath419 if they differ in the second coordinate .
hence the complexity is @xmath421 . + for union , the states @xmath422 are all final and equivalent as they accept @xmath423 .
the remaining states are distinguished by words in @xmath342 if they differ in the first coordinate or by words in @xmath281 if they differ in the second coordinate .
hence the complexity is @xmath333 . + as in regular languages , one letter in @xmath302 is not sufficient to reach all the states of @xmath424 for all values of @xmath24 and @xmath19 ; hence two letters are required to meet any of the bounds .
+ _ unrestricted complexity : _ the unrestricted bounds for right ideals are the same as those for arbitrary regular languages @xcite . we show that the dfas @xmath425 and @xmath426 of definition [ def : ideal ] meet the bounds .
+ ( 35,23)(0,-6 ) + ( 00,10)@xmath141 ( 01,10)@xmath141 ( 02,12)@xmath141 ( 03,13)@xmath141 + ( 10,20)@xmath141 ( 11,20)@xmath141 ( 12,22)@xmath141 ( 13,23)@xmath141 + ( 20,01)@xmath79 ( 20,30)@xmath141 ( 21,30)@xmath141 ( 22,32)@xmath141 ( 23,33)@xmath141 + ( 04,14)@xmath141 ( 14,24)@xmath141 ( 24,34)@xmath141 ( 34,44)@xmath178 + ( 30,31)@xmath79 ( 31,32)@xmath79 ( 32,33)@xmath79 + ( 00,11)@xmath79 ( 10,21)@xmath79 ( 01,12)@xmath79 ( 11,22)@xmath79 ( 21,02)@xmath79 ( 02,13)@xmath79 ( 12,23)@xmath79 ( 22,03)@xmath79 + ( 40,41)@xmath79 ( 41,42)@xmath79 ( 42,43)@xmath79 ( 43,44)@xmath139 + ( 03,04)@xmath139 ( 13,04)@xmath139 ( 23,24)@xmath139 ( 33,34)@xmath139 + ( 30,41)@xmath178 ( 31,42)@xmath178 ( 32,40)@xmath178 ( 33,43)@xmath178 + to compute @xmath336 , where @xmath20 is a boolean operation , add an empty state @xmath267 to @xmath425 , and send all the transitions from any state of @xmath243 under @xmath178 to @xmath267 . similarly , add an empty state @xmath248 to @xmath269 together with appropriate transitions ; now the alphabets of the resulting dfas are the same .
the direct product of @xmath427 and @xmath428 is illustrated in figure [ fig : idealcross2 ] for @xmath410 .
+ as in the restricted case , the @xmath167 states of @xmath258 are reachable by words in @xmath423 .
the remaining states @xmath292 and @xmath303 are easily seen to be reachable using @xmath178 and @xmath139 .
+ we now check distinguishability , which depends on the final states of the dfa .
the direct product is made to recognize @xmath336 by setting the final states to be @xmath429 .
+ for union and symmetric difference , all states are pairwise distinguishable : states that differ in the first coordinate are distinguished by words in @xmath430 and states that differ in the second coordinate are distinguished by words in @xmath431 .
+ for difference , the final states are @xmath284 for @xmath285 .
the alphabet of @xmath432 is @xmath433 ; hence we can omit @xmath178 and delete all states @xmath303 and be left with a dfa recognizing the same language .
the remaining states are distinguished by words in @xmath430 if they differ in the first coordinate or by words in @xmath434 if they differ in the second coordinate . + for intersection , the only final state is @xmath295 .
the alphabet of @xmath435 is @xmath299 ; hence we can omit @xmath178 and @xmath139 and delete all states @xmath292 and @xmath303 .
the remaining @xmath167 states are pairwise distinguishable as in the restricted case .
+ note that the bound for difference is met by @xmath339 , and that of intersection is met by @xmath340 .
however the bounds for union and symmetric difference all require three letters in each dialect : there must be a letter in @xmath436 to reach states of the form @xmath292 , and there must a letter in @xmath437 to reach states of the form @xmath303 .
as in regular languages , one letter in @xmath302 is not sufficient to reach all the states of @xmath258 for all values of @xmath24 and @xmath19 ; hence @xmath438 and so both @xmath119 and @xmath1 must contain at least three letters . it has been shown in @xcite that at least two letters are needed for each right ideal that meets the bounds for star or reversal . hence almost all our witnesses in theorem [ thm : rightideals ] that meet the bounds for the common operations use minimal alphabets
the complexity of operations on prefix - closed languages was studied in @xcite , but most complex prefix - closed languages were not considered . as every prefix - closed language has an empty quotient , the restricted and unrestricted complexities are the same for binary operations .
[ def : closedwit ] for @xmath308 , let @xmath439 , where @xmath310 , and @xmath145 is defined by the transformations @xmath311 , @xmath147 , @xmath440 , and @xmath441 .
let @xmath313 be the language accepted by @xmath150 .
the structure of @xmath348 is shown in figure [ fig : closedwit ] .
8.5pt ( 37,8)(0,3 ) ( 0)(1,7)0(0)(0 ) ( 1)(8,7)1(1 ) ( 2)(15,7)2(2 ) ( 3dots)(22,7)@xmath152 ( n-2)(29,7)@xmath153(n-2 ) ( n-1)(36,7)@xmath33 ( 0,1)@xmath154 ( 0,n-1)@xmath141 ( 1,0)@xmath442 ( 1,2)@xmath79 ( 2,1)@xmath141 ( 2,3dots)@xmath79 ( 3dots,2)@xmath141 ( 3dots , n-2)@xmath79 ( n-2,3dots)@xmath141 ( n-2,0)@xmath79 ( 0)@xmath139 ( 2)@xmath155 ( n-2)@xmath155 ( n-1)@xmath318 [ thm : closed ] for each @xmath308 , the dfa of definition [ def : closedwit ] is minimal and @xmath32 is a prefix - closed language of complexity @xmath19 .
the stream @xmath319 with some dialect streams is most complex in the class of prefix - closed languages .
at least four letters are required to meet the bounds below . 1
. the syntactic semigroup of @xmath32 has cardinality @xmath320 .
2 . the quotients of @xmath321 have complexity @xmath19 , except for @xmath248 , which has complexity 1 .
the reverse of @xmath321 has complexity @xmath323 , and @xmath321 has @xmath323 atoms .
each atom @xmath45 of @xmath32 has maximal complexity : @xmath443 5 . the star of @xmath444 has complexity @xmath445 .
the product @xmath446 has complexity @xmath447 .
7 . for any proper binary boolean function @xmath20 ,
the complexity of @xmath448 is maximal .
in particular , the complexity is @xmath167 if @xmath337 , @xmath449 if @xmath332 , and @xmath333 if @xmath338 .
the dfa is minimal since state @xmath62 rejects @xmath450 if and only if @xmath451 .
it is prefix - closed because all non - empty states are final . 1 . * semigroup * let @xmath452 induce the transformation @xmath453 ( this was called @xmath141 in the right - ideal section ) .
since @xmath454 , the transition semigroup of the dfa of figure [ fig : closedwit ] is the same as that of the dfa of the right ideal of figure [ fig : ridealwit ] .
* quotients * obvious .
* reversal * since reversal commutes with complementation , we consider the complement of the language accepted by the dfa of figure [ fig : closedwit ] restricted to the alphabet @xmath455 . it was proved in @xcite that the reverse of a right ideal has complexity at most @xmath323 , and in @xcite that the number of atoms is the same as the complexity of the reverse .
it remains to prove that all @xmath323 states of the dfa @xmath456 obtained by the subset construction from the nfa @xmath183 obtained by reversal of the dfa of the right ideal @xmath55 are reachable and distinguishable .
the proof is similar to that of @xcite .
subset @xmath405 is the initial state of @xmath183 , and @xmath33 appears in every reachable state of @xmath456 .
every subset @xmath457 of size @xmath29 , where @xmath458 and @xmath459 , is reached from the subset @xmath460 of size @xmath461 by @xmath462 . since only state @xmath59 , @xmath350 , accepts @xmath463 , any two subsets differing by @xmath59 are distinguishable by @xmath463 .
* atoms * let @xmath0 be a prefix - closed language with quotients @xmath43 , @xmath464 . recall that @xmath465 is a right ideal with quotients @xmath466 . for @xmath467 ,
the atom of @xmath0 corresponding to @xmath46 is @xmath468 .
this can be rewritten as @xmath469 , which is the atom of @xmath465 corresponding to @xmath470 ; hence the sets of atoms of @xmath0 and @xmath465 are the same .
the claim follows from the theorem for right ideals .
the proof in @xcite applies since the dfa that accepts the complement of the prefix - closed language of figure [ fig : closedwit ] has all the transformations that fix @xmath33 .
star * it was proved in @xcite that @xmath445 is the maximal complexity of the star of a prefix - closed language .
we now show that @xmath444 meets this bound .
since @xmath444 accepts @xmath13 , no new initial state is needed and it suffices to delete the empty state and add an @xmath13-transition from each final state to the initial state to get an nfa @xmath183 for @xmath352 . in this nfa
all @xmath356 subsets of @xmath345 containing 0 are reachable and pairwise distinguishable .
any non - empty set @xmath471 of size @xmath29 with @xmath472 is reached from @xmath473 of size @xmath461 by @xmath474 .
moreover , the empty set is reached from @xmath250 by @xmath141 , giving the required bound .
sets @xmath475 and @xmath476 with @xmath394 are distinguished by @xmath477 .
* product * it was shown in @xcite that the complexity of the product of prefix - closed languages is @xmath447 .
we now prove that our witness @xmath478 with minimal dfa @xmath479 together with the dialect @xmath480 with minimal dfa @xmath481 meets this bound .
construct the following nfa @xmath183 for the product .
start with @xmath479 , but make all of its states non - final .
delete the empty state from @xmath481 and all the transitions to the empty state , add an @xmath13-transition from each state @xmath482 to the initial state @xmath187 of @xmath481 .
we will show that @xmath483 states of the form @xmath484 , where @xmath485 , and @xmath323 states of the form @xmath486 , where @xmath487 are reachable and pairwise distinguishable .
+ the initial state of the subset automaton of @xmath183 is @xmath488 .
state @xmath489 is reachable by @xmath178 and @xmath231 for @xmath490 is reachable from @xmath489 by @xmath491 .
state @xmath492 where @xmath493 and @xmath494 is reachable from @xmath495 by @xmath496 for some @xmath497 . by induction ,
all @xmath483 states @xmath492 are reachable . from @xmath498 by @xmath499
we reach @xmath500 .
further apply @xmath501 to reach @xmath486 .
hence all @xmath323 subsets of the form @xmath486 are reachable .
+ we check that the states are pairwise distinguishable in four cases
@xmath486 and @xmath502 with @xmath503 are distinguished by @xmath504 .
2 . @xmath188 and @xmath505 with @xmath503 reduces to case ( a ) by @xmath506 .
@xmath188 and @xmath502 with @xmath507 are distinguished by @xmath508 .
4 . @xmath188 and @xmath509 with @xmath510 reduces to case ( c ) by @xmath511 .
7 . * boolean operations * it is again convenient to consider the ideal languages defined by the complements of the prefix - closed languages of figure [ fig : closedwit ] restricted to the alphabet @xmath512 and then use de morgan s laws . since every prefix - closed language has an empty quotient , it is sufficient to consider boolean operations on languages over the same alphabet .
the problems are the same as those in @xcite , except that there the transformation induced by @xmath141 is @xmath513 .
+ let @xmath348 denote the dfa for the complement of the prefix - closed language of definition [ def : closedwit ] of complexity @xmath19 and let @xmath18 be the language accepted by @xmath150 .
we consider boolean operations on right ideals @xmath23 and @xmath18 .
+ 8.5pt + ( 35,20)(0,-2 ) + ( 00,44)@xmath141 ( 01,40)@xmath141 ( 02,41)@xmath141 ( 03,42)@xmath141 + ( 10,04)@xmath141 ( 20,14)@xmath141 + ( 40,44)@xmath141 ( 04,44)@xmath141 + ( 43,42)@xmath141 ( 42,41)@xmath141 ( 41,40)@xmath141 + ( 42,43)@xmath178 ( 41,42)@xmath178 ( 40,41)@xmath178 ( 43,40)@xmath178 + ( 24,14)@xmath141 ( 14,04)@xmath141 + ( 14,24)@xmath79 ( 04,14)@xmath79 ( 24,04)@xmath79 + ( 13,23)@xmath79 ( 03,13)@xmath79 ( 23,03)@xmath79 + ( 22,23)@xmath178 ( 21,22)@xmath178 ( 20,21)@xmath178 ( 23,20)@xmath178 + the direct product is illustrated in figure [ fig : closedcross ] .
the states in @xmath424 are reachable from the initial state @xmath514 by ( * ? ? ? * theorem 1 ) . then @xmath515 is reached from @xmath516 by @xmath141 and states of the form @xmath284 , @xmath350 ,
are then reached by words in @xmath517 .
similarly , @xmath518 is reached from @xmath519 by @xmath141 and states of the form @xmath520 , @xmath521 , are then reached by words in @xmath281 . finally , @xmath522 is reached from @xmath523 by @xmath141 . hence all @xmath167 states are reachable .
+ let @xmath524 , @xmath525 , and @xmath526 .
the final states of the direct product to recognize @xmath448 are @xmath527 .
+ consider the following dfas : @xmath528 and @xmath529 @xmath530 . by
* theorem 1 ) , the states of @xmath46 are pairwise distinguishable with respect to the states @xmath531 for any @xmath260 .
one can verify that if @xmath8 distinguished two states of @xmath46 with respect to @xmath531 , then @xmath532 distinguishes them with respect to @xmath527 for each @xmath533 .
the rest of the argument depends on the operation @xmath260 .
+ @xmath534 : all @xmath167 states are pairwise distinguishable .
the states of @xmath283 are distinguished by words in @xmath342 .
the states of @xmath280 are similarly distinguishable .
the states of @xmath283 are distinguished from the states of @xmath280 by words in @xmath423 .
every state of @xmath46 is sent to a state of @xmath283 by a word in @xmath423 , and similarly to a state of @xmath280 by a word in @xmath535 ; thus the states of @xmath46 are distinguishable from the states of @xmath283 or @xmath280 .
+ @xmath536 : the states of @xmath280 are all empty , leaving @xmath537 distinguishable states .
the states of @xmath283 are distinguished by words in @xmath342 .
+ @xmath538 : the states of @xmath283 and @xmath280 are equivalent final states accepting all words , leaving @xmath539 distinguishable states .
+ by de morgan s laws we have @xmath540 , @xmath541 , @xmath542 , and @xmath543 .
thus the prefix - closed witness meets the bounds for boolean operations .
since the semigroup of a prefix - closed language is the same as that of its complement , which is a right ideal , at least four letters are required to meet all the bounds .
the complexity of operations on prefix - free languages was studied in @xcite , but most complex prefix - free languages were not considered . as every prefix - free language has an empty quotient , the restricted and unrestricted complexities are the same for binary operations .
[ def : prefix - free ] for @xmath308 , let @xmath544 and define the dfa @xmath545 @xmath546 where @xmath145 is defined by the transformations @xmath547 , @xmath548 , @xmath549 , @xmath550 , @xmath551 for @xmath552 .
the transformations induced by @xmath79 and @xmath178 coincide when @xmath275 .
let @xmath553 be the language accepted by @xmath554 .
the structure of @xmath554 is shown in figure [ fig : free ] .
( 40,16)(0,0 ) ( 0)(1,10)0(0 ) ( 1)(9,10)1 ( 2)(17,10)2 ( 3dots)(25,10)@xmath152 ( n-4)(33,10)@xmath555 ( n-3)(41,10)@xmath556 ( n-2)(17,1)@xmath153(n-2 ) ( n-1)(33,1)@xmath33 ( 0,1)@xmath154 ( 1,2)@xmath79 ( 2,3dots)@xmath79 ( 3dots , n-4)@xmath79 ( n-4,n-3)@xmath79 ( n-3,0)@xmath79 ( 0,n-2)@xmath557 ( 1,n-2)@xmath558 ( 2,n-2)@xmath559 ( n-4,n-2)@xmath560 ( n-3,n-2)@xmath561 ( n-2,n-1)@xmath562 ( 1,0)@xmath155 [ thm : prefix - free_witness ] for @xmath308 , the dfa of definition [ def : prefix - free ] is minimal and @xmath563 is a prefix - free language of complexity @xmath19 . the stream @xmath564 with dialect stream @xmath565 @xmath566 is most complex in the class of prefix - free languages .
at least @xmath567 inputs are required to meet all the bounds below . 1
. the syntactic semigroup of @xmath568 has cardinality @xmath569 .
there is only one maximal transition semigroup of minimal dfas accepting prefix - free languages . moreover , fewer than @xmath325 inputs do not suffice to meet this bound .
2 . the quotients of @xmath321 have complexity @xmath19 , except for @xmath13 and @xmath248 , which have complexity 2 and 1 , respectively .
3 . the reverse of @xmath570 has complexity @xmath445 , and @xmath570 has @xmath445 atoms .
each atom @xmath45 of @xmath571 has maximal complexity : @xmath572 5 .
the star of @xmath321 has complexity @xmath19 .
the product @xmath573 has complexity @xmath574 .
7 . for @xmath575 but @xmath576 , and for any proper binary boolean function @xmath20 ,
the complexity of @xmath577 is maximal .
in particular , these languages meet the bounds @xmath578 for union and symmetric difference , @xmath579 for intersection , and @xmath580 for difference .
since only state @xmath59 accepts @xmath386 for @xmath581 , dfa @xmath349 is minimal .
since it has only one final state and that state accepts @xmath12 , @xmath321 is prefix - free . 1 .
* semigroup * the proof that the size of the semigroup is @xmath569 is very similar to that in @xcite . inputs @xmath79 ,
@xmath178 , and @xmath139 generate all transformations of @xmath582 .
moreover , any state @xmath583 can be sent to @xmath153 by @xmath584 and to @xmath33 by @xmath585 .
hence we have all @xmath569 transformations of @xmath70 that fix @xmath33 and send @xmath153 to @xmath33 .
the maximal transition semigroup is unique , since it must contain all these transformations .
+ to prove that at least @xmath325 inputs are necessary , we see that @xmath586 is in the transition semigroup of @xmath150 .
there are two types of states in @xmath587 : those of type 1 , for which @xmath584 is a generator ( that is , the transformation of @xmath584 is induced by a single letter ) , and those of type 2 , for which it is not . if @xmath584 and @xmath588 are generators , then clearly @xmath589 .
+ if @xmath584 is not a generator , then it must be a composition , @xmath590 , where @xmath591 is in the semigroup and @xmath592 is a generator .
no state can be mapped by @xmath591 to @xmath153 because then @xmath592 would map @xmath153 to @xmath33 .
hence @xmath591 must be a permutation of @xmath582 .
if @xmath593 and @xmath584 and @xmath594 are not generators , then there exist @xmath595 and @xmath596 as above , such that @xmath590 and @xmath597 .
then we must have @xmath598 ; otherwise both @xmath59 and @xmath599 would be mapped to @xmath153 .
hence @xmath600 and all the generators of this type are distinct .
+ finally , if @xmath584 is a generator and @xmath601 is as above , then @xmath602 , for otherwise @xmath603 would be the identity and @xmath599 would be of type 1 . therefore , @xmath153 generators are required in addition to those induced by @xmath79 , @xmath178 and @xmath139 .
* quotients * this is clear from the definition .
* reversal * this was proved in @xcite .
* atoms * first we establish an upper bound on the complexity of atoms of prefix - free languages .
consider the intersection @xmath604 , where @xmath605 , and @xmath606
. clearly @xmath33 must be in @xmath470 if @xmath45 is an atom , for @xmath607 .
since a prefix - free language has only one final state and that state accepts @xmath13 , if @xmath608 , no other quotient is in @xmath46 , for then @xmath45 would not be an atom . hence if @xmath609 then @xmath610 , and @xmath611 . +
now suppose @xmath612 ; then @xmath613 .
since @xmath614 appears in every quotient of @xmath45 , there are at most @xmath323 subsets of @xmath345 that can be reached from @xmath45 together with @xmath33 .
hence @xmath615 .
+ if @xmath616 , then @xmath617 and @xmath618 .
if we reach @xmath619 , then any intersection which has @xmath620 in the complemented part is equivalent to one that has only @xmath405 , since no quotient other than @xmath621 contains @xmath13 .
hence we can reach at most @xmath622 subsets of @xmath582 , along with the intersection @xmath623 , and the empty quotient , for a total of @xmath445 states .
+ finally , consider the case where @xmath624 .
then we have from 1 to @xmath625 uncomplemented quotients @xmath626 with @xmath627 , and from 1 to @xmath628 quotients @xmath626 with @xmath627 in the complemented part ; this leads to the formula given in the theorem .
+ it remains to be proved that the atoms of @xmath571 meet these bounds .
atom @xmath629 is equal to @xmath12 and thus has two quotients as required ; assume now that @xmath630 .
we are interested in the number of distinct quotients of @xmath604 , where @xmath631 .
the quotients @xmath632 have the form @xmath633 where @xmath634 and @xmath635 .
for brevity , we write @xmath636 and @xmath637 ; this notation is in agreement with the action of @xmath8 on the states of @xmath150 corresponding to @xmath46 and @xmath470 .
+ notice @xmath638 for all @xmath639 and @xmath640 , except for the case @xmath641 in which @xmath642 .
thus it is sufficient to assume @xmath643 from now on , as @xmath644 contains every quotient of @xmath45 .
we show that whenever @xmath645 , @xmath646 , @xmath643 , and @xmath647 , there is a word @xmath648 such that @xmath636 and @xmath637 and hence @xmath649 is a quotient of @xmath45 . when @xmath650 we reach all quotients @xmath651 where @xmath652 by words in @xmath653 , we reach @xmath654 from @xmath655 by @xmath557 , and from there we reach the empty quotient by @xmath557 .
similarly , when @xmath656 , we reach @xmath657 for @xmath658 and @xmath659 by words in @xmath653 , and the remaining quotients are easily reached using @xmath557 . + it remains to show that non - empty quotients @xmath649 and @xmath660 are distinct whenever @xmath661 or @xmath662 . notice @xmath663 if either @xmath664 or @xmath665 , and @xmath666 if and only if @xmath667 .
apart from these special cases , every @xmath649 is non - empty and does not contain @xmath13 .
+ for any @xmath658 , let @xmath668 denote a word that maps @xmath669 and @xmath670 ; there is such a word in @xmath671 because @xmath653 contains @xmath672 , and then @xmath673 .
observe that @xmath674 for all @xmath675 and @xmath676 for all @xmath677 .
hence , if @xmath658 and @xmath678 , then @xmath679 and @xmath680 .
+ let @xmath681 and @xmath682 be any disjoint subsets of @xmath70 where @xmath683 and @xmath684 . if @xmath685 then either @xmath686 or @xmath687 . similarly , if @xmath688 ( and @xmath689 ) then either @xmath690 or @xmath691
thus , any two quotients @xmath649 and @xmath660 , where @xmath692 , are distinct .
+ when we established the upper bound on @xmath693 , we counted the number of reachable , potentially distinct quotients @xmath649 of each @xmath45 .
we have now shown that every reachable @xmath649 is a quotient of @xmath45 and determined all the cases when @xmath694 .
it follows that every bound is met by @xmath571 .
star * proved in @xcite . for the purpose of proving that @xmath567 inputs are required for a most complex prefix - free witness , an outline of the proof is repeated here .
+ suppose that @xmath0 is a prefix - free language with @xmath19 quotients whose syntactic semigroup is maximal , and @xmath695 has maximal complexity .
we show that @xmath0 requires an alphabet of size @xmath567 . towards a contradiction ,
let @xmath696 be a dfa for @xmath0 where @xmath697 .
assume @xmath698 are non - final , non - empty states , @xmath153 is the unique final state , and @xmath33 is the empty state . by @xcite
, @xmath55 must have this structure and @xmath699 for any @xmath700 .
+ since the syntactic semigroup of @xmath0 is maximal , each letter of @xmath1 has a specific role in @xmath55 as described in * 1 * of this theorem .
three letters @xmath701 , @xmath702 , and @xmath703 are required to induce the transformations on @xmath582 ; these letters can not map any state of @xmath582 to @xmath153 or to @xmath33 .
an additional @xmath153 letters @xmath704 are required to generate @xmath586 for each @xmath587 , where the action of @xmath584 is induced by a word in @xmath705 .
notice @xmath592 can not map any state of @xmath582 to @xmath33 , since @xmath584 does not . in summary , @xmath706 and for all @xmath707 and @xmath587 , @xmath708 .
+ an nfa for @xmath695 is produced by adding to @xmath55 a new initial state @xmath197 , which is final , adding an @xmath13-transition from @xmath153 to @xmath187 , and deleting the empty state @xmath33 .
the transitions from @xmath197 are exactly the same as the transitions from @xmath187 .
perform the subset construction on this nfa .
the @xmath33 states @xmath709 are all reachable and distinguishable by words in @xmath710 .
the only way to reach a set containing more than one state is by moving to @xmath153 and using the @xmath13-transition .
this leads to the state @xmath711 , but applying any word @xmath700 deletes @xmath153 ; thus , @xmath711 is the only reachable set with two or more states .
however , @xmath205 and @xmath711 are indistinguishable , since both are final and @xmath712 for @xmath700 . + so far , there are only @xmath33 reachable , distinguishable states in the subset construction .
the remaining state is @xmath248 , which can only be reached if there is a letter @xmath261 that moves from @xmath587 to @xmath33 in @xmath55 ; a transition from @xmath153 to @xmath33 is not sufficient to reach the empty state .
we showed that in our witness no @xmath707 has @xmath713 .
therefore , @xmath714 , a contradiction . to achieve @xmath715 , an additional letter is required .
therefore , any most complex prefix - free language stream requires @xmath567 inputs .
* product * proved in @xcite .
* boolean operations * let @xmath716 .
for @xmath717 , let @xmath718 , and for @xmath719 let @xmath720 .
these are the sets of states in the rows and columns of figure [ fig : freecross ] .
the states in @xmath46 are reachable from the initial state @xmath514 by ( * ? ? ?
* theorem 1 ) .
every other state in the direct product is reachable from some state in @xmath46 , as illustrated in figure [ fig : freecross ] .
+ 8.5pt + ( 35,25)(0,-9 ) + ( 02,32)@xmath557 ( 20,23)@xmath557 + ( 02,03)@xmath559 ( 12,13)@xmath559 ( 22,33)@xmath559 ( 20,30)@xmath559 ( 21,31)@xmath559 + ( 30,40)@xmath559 ( 31,41)@xmath559 ( 03,04)@xmath559 ( 13,14)@xmath559 + ( 30,41)@xmath178 ( 31,42)@xmath178 ( 32,43)@xmath559 ( 40,41)@xmath178 ( 41,42)@xmath178 ( 42,43)@xmath559 ( 42,40)@xmath178 + ( 03,14)@xmath79 ( 13,24)@xmath79 ( 23,34)@xmath559 ( 04,14)@xmath79 ( 14,24)@xmath79 ( 24,34)@xmath559 ( 24,04)@xmath79 + ( 33,44)@xmath1 + ( 43,44)@xmath1 + ( 34,44)@xmath1 + for @xmath260 , the direct product recognizes @xmath721 if the final states are set to be @xmath722 .
now we must determine which states are distinguishable with respect to @xmath722 for each value of @xmath20 .
consider the dfas @xmath723 and @xmath724 . by
* theorem 1 ) , the states of @xmath46 are pairwise distinguishable with respect to @xmath725 . for any pair of states in @xmath46 ,
let @xmath8 be a word that distinguishes them in @xmath725 ; one verifies that further applying @xmath726 distinguishes them with respect to @xmath722 .
the rest of the distinguishability argument depends on @xmath260 .
+ @xmath538 : states @xmath727 , @xmath728 , and @xmath729 are equivalent , since all three are final and any letter sends them to @xmath522 .
+ states of @xmath730 are distinguished by words in @xmath731 .
states of @xmath732 are distinguished by words in @xmath733 . excluding @xmath727 and @xmath728 , which are equivalent , states of @xmath730 are distinguished from states of @xmath732 by words in @xmath733 .
+ states of @xmath734 are moved to states of @xmath735 by applying @xmath726 .
excluding @xmath727 , @xmath728 , and @xmath729 , which are equivalent , every state is mapped by @xmath726 to a different state of @xmath735 ; hence they are distinguishable .
+ finally , we must show that states of @xmath46 are distinguishable from the states of @xmath735 . for any @xmath736 , there exists @xmath294 such that @xmath737 , since both @xmath738 and @xmath739 are reached from @xmath514 by words in @xmath276 , and @xmath79 and @xmath178 permute @xmath46 .
then @xmath740 and we have already shown that @xmath741 is distinguishable from all states in @xmath735 .
thus , the @xmath578 remaining states are pairwise distinguishable .
+ @xmath742 : states @xmath727 and @xmath728 are equivalent , and states @xmath729 and @xmath522 are equivalent .
the rest of the states are distinguishable by an argument similar to that of union .
+ @xmath743 : state @xmath744 is the only final state .
the remaining non - final states of @xmath745 are all empty .
clearly the states of @xmath46 are non - empty , since @xmath746 .
thus , the remaining @xmath747 states are pairwise distinguishable .
+ @xmath536 : the states of @xmath730 and @xmath729 are all equivalent .
states @xmath748 and @xmath749 are equivalent for @xmath750
. the final states ( @xmath751 ) are all equivalent . + the states of @xmath732
are distinguished by words in @xmath733 .
it remains to show that states of @xmath46 are distinguishable from the states of @xmath732 .
notice @xmath752 is distinguished from @xmath753 by @xmath726 , and from every other state of @xmath732 by @xmath754 . for any state of @xmath46
, there exists @xmath294 that sends that state to @xmath752 , and notice @xmath755 .
so all @xmath756 remaining states are pairwise distinguishable .
note that the stream @xmath757 with dialect stream @xmath758 meets the bounds for quotients , reversal , atomic complexity , star , product and boolean operations .
using some results from @xcite we define another prefix - free witness stream that meets all the bounds except those for syntactic complexity and atom complexity .
_ moreover , all the bounds are met by dialects over minimal alphabets .
_ [ def : prefix - free_small ] for @xmath308 , let @xmath759 where @xmath760 , and @xmath145 is defined by the transformations * @xmath547 , * @xmath549 . * @xmath550 , * @xmath761 , * @xmath762 , * @xmath763 . note that @xmath178 is not used , @xmath79 , @xmath139 , @xmath141 , and @xmath764 induce the same transformations as @xmath79 , @xmath139 , @xmath141 , and @xmath561 in definition [ def : prefix - free ] .
dfa @xmath765 is shown in figure [ fig : small ] .
let @xmath563 be the language accepted by @xmath765 .
8.5pt ( 40,16)(0,0 ) ( 0)(1,10)0(0 ) ( 1)(9,10)1 ( 2)(17,10)2 ( 3dots)(25,10)@xmath152 ( n-4)(33,10)@xmath555 ( n-3)(41,10)@xmath556 ( n-2)(17,1)@xmath153(n-2 ) ( n-1)(17,5.5)@xmath33 ( 0,1)@xmath766 ( 1,2)@xmath766 ( 2,3dots)@xmath766 ( 3dots , n-4)@xmath766 ( n-4,n-3)@xmath766 ( n-3,0)@xmath79 ( 0,n-2)@xmath141 ( 1,n-1)@xmath141 ( 2,n-1)@xmath141 ( n-4,n-1)@xmath141 ( n-3,n-1)@xmath141 ( n-3,n-2)@xmath767 ( n-2,n-1)@xmath1 ( 1,0)@xmath139 [ prop : small ] for @xmath308 , the dfa of definition [ def : prefix - free_small ] is minimal and @xmath563 is a prefix - free language of complexity @xmath19 .
moreover , all the witnesses for individual operations have minimal alphabets . 1 .
the quotients of @xmath768 have complexity @xmath19 , except for the quotient @xmath13 and the empty quotient , which have complexity 2 and 1 respectively .
2 . the reverse of @xmath769 has complexity @xmath445 , and @xmath444 has @xmath445 atoms .
the star of @xmath321 has complexity @xmath19 .
4 . for @xmath575
, the product of @xmath770 and @xmath771 has complexity @xmath574 .
5 . 1 . @xmath772 and @xmath773 have complexity @xmath578 .
2 . @xmath774 has complexity @xmath580 .
3 . @xmath775 has complexity @xmath579 .
the first claim is obvious .
the second and third claims were proved in theorem [ thm : prefix - free_witness ] .
( a ternary witness was also used in @xcite for the reverse , but it had more complicated transitions than our witness . ) the fourth claim is from @xcite . the results for union , symmetric difference and intersection were proved in @xcite , and that for difference in @xcite .
[ idealorempty ] let @xmath0 be a prefix - convex language over @xmath1 .
either @xmath0 is a right ideal or @xmath0 has an empty quotient .
suppose that @xmath0 is not a right ideal .
if @xmath776 , then @xmath777 is an empty quotient of @xmath0 .
if @xmath778 , we can not have @xmath779 for all @xmath780 , because then @xmath0 would be a right ideal .
hence there exists some @xmath780 such that @xmath781 .
pick any @xmath782 ; then @xmath780 , but @xmath783 .
there can not be a word @xmath784 such that @xmath785 , because then @xmath786 would be in @xmath0 by prefix convexity .
therefore , @xmath787 is an empty quotient .
[ prefixchar ] let @xmath0 be a regular language having @xmath19 quotients and @xmath29 final quotients , and let @xmath788 be a minimal dfa recognizing @xmath0 .
the following statements are equivalent : 1 .
@xmath0 is prefix - convex .
2 . for all @xmath789 ,
if @xmath62 and @xmath790 are final , @xmath59 is reachable from @xmath62 , and @xmath790 is reachable from @xmath59 , then @xmath59 is final .
every state reachable in @xmath791 from any final state is either final or empty .
( * 1 * @xmath792 * 2 * ) assume * 1 * is true .
suppose there exist @xmath793 and @xmath225 such that @xmath59 is reachable from @xmath62 and @xmath790 is reachable from @xmath59 .
let @xmath794 be such that @xmath795 , @xmath796 , and @xmath797 .
it follows that @xmath8 and @xmath798 are both in @xmath0 , and thus @xmath786 is in @xmath0 by prefix convexity .
since @xmath799 , state @xmath59 is final . + ( * 2 * @xmath792 * 3 * ) assume * 2 * is true . take any @xmath800 , @xmath225 , and @xmath801 such that @xmath802 .
if a final state @xmath790 is reachable from @xmath59 , then @xmath59 is final by * 2*. otherwise , @xmath59 is the empty state . + ( * 3 * @xmath792 * 1 * ) assume * 3 * is true .
let @xmath803 , and @xmath37 be words in @xmath344 such that @xmath780 and @xmath785 .
there are states @xmath804 , and @xmath790 in @xmath70 such that @xmath805 , @xmath799 , and @xmath806 .
state @xmath59 can not be empty , since final state @xmath790 is reachable from @xmath59 .
since @xmath59 is reachable from final state @xmath62 , it follows from * 3 * that @xmath59 is final .
thus , @xmath807 . therefore @xmath0 is prefix - convex .
[ convexclasses ] let @xmath0 be a non - empty prefix - convex language having @xmath19 quotients and @xmath29 final quotients , and let @xmath788 be a minimal dfa recognizing @xmath0 . 1 .
@xmath0 is prefix - closed if and only if @xmath808 .
2 . @xmath0 is prefix - free if and only if @xmath791 has a unique final state @xmath62 , an empty state @xmath809 , and @xmath810 for all @xmath121 .
@xmath0 is a right ideal if and only if @xmath791 has a unique final state @xmath62 and @xmath811 for all @xmath121 .
note that @xmath812 and @xmath813 since @xmath0 is non - empty . 1 .
suppose @xmath0 is prefix - closed . clearly , @xmath13 is a prefix of some word in @xmath0 , since @xmath0 is non - empty .
thus @xmath814 , and so @xmath808 .
conversely , suppose @xmath808 .
for any @xmath807 , there are states @xmath815 such that @xmath816 , and @xmath790 is final . by proposition [ prefixchar ] , since @xmath817 , @xmath59 is reachable from @xmath187 , and @xmath790 is reachable from @xmath59 , we have @xmath818 .
hence @xmath780 , and therefore @xmath0 is prefix - closed .
2 . suppose @xmath0 is prefix - free .
if @xmath225 and @xmath800 are distinct and @xmath59 is reachable from @xmath62 , then @xmath59 can not be final as that would imply @xmath819 .
in particular , for any @xmath800 and @xmath121 , @xmath820 . by proposition
[ prefixchar ] , @xmath821 must be the empty state for all @xmath121 .
thus , the transitions from all final states are identical , and hence all final states are equivalent . by minimality , @xmath791 has a unique final state @xmath62 , an empty state @xmath809 , and @xmath810 for all @xmath121 .
+ for the converse , suppose @xmath822 , @xmath823 is an empty state , and @xmath824 for all @xmath121 . then @xmath780 if and only if @xmath825 . for all @xmath780 and @xmath121
, we have @xmath826 .
thus , whenever @xmath780 and @xmath807 , we have @xmath827 .
therefore , @xmath0 is prefix - free .
suppose @xmath0 is a right ideal .
for all @xmath780 we have @xmath828 , and hence @xmath829 , meaning that @xmath779 .
hence , for any final state @xmath818 and @xmath801 , @xmath830 .
this implies that all final states are equivalent . by minimality
, there is a unique final state @xmath62 . since @xmath831 for all @xmath121
, it follows that @xmath811 for all @xmath121 . for the converse , suppose @xmath822 and @xmath811 for all @xmath121 .
then @xmath780 if and only if @xmath832 . hence ,
for all @xmath780 and @xmath801 , we have @xmath833 .
thus , @xmath834 for all @xmath780 , and so @xmath835 .
therefore , @xmath0 is a right ideal .
recall that a prefix - convex language @xmath0 is _ proper _ if it is not a right ideal and it is neither prefix - closed nor prefix - free .
moreover , it is _
@xmath29-proper _ if it has @xmath29 final states , @xmath458 .
every minimal dfa for a @xmath29-proper language with complexity @xmath19 has the same general structure : there are @xmath836 non - final , non - empty states , @xmath29 final states , and one empty state .
every letter fixes the empty state and , by proposition [ prefixchar ] , no letter sends a final state to a non - final , non - empty state .
as every proper language has an empty quotient , the restricted and unrestricted complexities are the same for binary operations .
next we define a stream of particular @xmath29-proper dfas and languages .
[ def : proper ] for @xmath837 and @xmath838 , let @xmath839 where @xmath840 , @xmath841 , and @xmath842 is defined by the transformations @xmath843 additionally define @xmath844 ; it is often useful to partition @xmath70 into @xmath845 , @xmath846 , and @xmath405 .
note that letters @xmath79 and @xmath178 have complementary behaviours on @xmath845 and @xmath846 , depending on the parities of @xmath19 and @xmath29 .
letters @xmath847 , and @xmath848 act on @xmath845 in exactly in the same way as @xmath849 , and @xmath850 act on @xmath846 .
in addition , @xmath848 and @xmath850 send states @xmath851 and @xmath153 , respectively , to state @xmath33 , and letter @xmath764 connects the two parts of the dfa . the structure of @xmath852 is shown in figures [ fig : properwit1 ] and [ fig : properwit2 ] for certain parities of @xmath836 and @xmath29 .
let @xmath853 be the language recognized by @xmath854 .
8.5pt ( 40,12)(0,-1.6 ) ( 0)(-1,7)0(0 ) ( 1)(9.5,7)1 ( 2)(18,7)2 ( 3dots)(26.5,7)@xmath152 ( n-2-k)(35,7)@xmath851 ( n-1)(41,3.5)@xmath33 ( n-1-k)(-1,0)@xmath836(n-1-k ) ( n - k)(9.5,0)@xmath855(n - k ) ( n - k+1)(18,0)@xmath856(n - k+1 ) ( 3dots1)(26.5,0)@xmath152 ( n-2)(35,0)@xmath153(n-2 ) ( 0,1)@xmath857 ( n-2-k , n-1)@xmath848 ( 0,n-1-k)@xmath764 ( 1,0)@xmath858 ( n-2-k,0)@xmath79 ( 1,2)@xmath859 ( 2,3dots)@xmath859 ( 3dots , n-2-k)@xmath859 ( n-1-k , n - k)@xmath860 ( n - k , n-1-k)@xmath861 ( n - k , n - k+1)@xmath862 ( n - k+1,3dots1)@xmath862 ( 3dots1,n-2)@xmath862 ( n-2,n-1)@xmath850 ( n-2,n - k)@xmath178 8.5pt ( 40,14)(0,-1.6 ) ( 0)(-1,7)0(0 ) ( 1)(9.5,7)1 ( 2)(18,7)2 ( 3dots)(26.5,7)@xmath152 ( n-2-k)(35,7)@xmath851 ( n-1)(41,3.5)@xmath33 ( n-1-k)(-1,0)@xmath836(n-1-k ) ( n - k)(9.5,0)@xmath855(n - k ) ( n - k+1)(18,0)@xmath856(n - k+1 ) ( 3dots1)(26.5,0)@xmath152 ( n-2)(35,0)@xmath153(n-2 ) ( 0,1)@xmath863 ( n-2-k , n-1)@xmath848 ( 0,n-1-k)@xmath764 ( 1,0)@xmath858 ( n-2-k,1)@xmath79 ( 1,2)@xmath859 ( 2,3dots)@xmath859 ( 3dots , n-2-k)@xmath859 ( n-1-k , n - k)@xmath864 ( n - k , n-1-k)@xmath861 ( n - k , n - k+1)@xmath862 ( n - k+1,3dots1)@xmath862 ( 3dots1,n-2)@xmath862 ( n-2,n-1)@xmath850 ( n-2,n-1-k)@xmath178 [ thm : propermain ] for @xmath142 and @xmath838 , the dfa @xmath852 of definition [ def : proper ] is minimal and @xmath853 is a @xmath29-proper language of complexity @xmath19 . this language is most complex in the class of @xmath29-proper prefix - convex languages ; in particular , it meets all the complexity bounds below .
at least 7 letters are required to meet these bounds . 1
. the syntactic semigroup of @xmath853 has cardinality @xmath865 ; the maximal value @xmath866 is reached only when @xmath867 . at least 7 letters are required to meet this bound .
the non - empty , non - final quotients of @xmath868 have complexity @xmath19 , the final quotients have complexity @xmath869 , and the empty quotient has complexity 1
the reverse of @xmath870 has complexity @xmath323 ; moreover , the language has @xmath323 atoms for all @xmath29 .
4 . for each atom @xmath45 of @xmath853 , write @xmath871 , where @xmath872 and @xmath873 . let @xmath874 and @xmath875 . if @xmath876 , then @xmath877 @xmath878 if @xmath879 and @xmath880 , then @xmath877 @xmath881 otherwise , @xmath882 and @xmath883 .
5 . the star of @xmath884 has complexity @xmath885 . the maximal value @xmath886 is reached only when @xmath887 . 6 .
the product of @xmath888 and @xmath889 has complexity @xmath890 . the maximal value @xmath891 is reached only when @xmath892 . 7 .
for @xmath893 , @xmath894 , and @xmath838 , let @xmath895 and @xmath896 .
for any proper binary boolean function @xmath20 , the complexity of @xmath897 is maximal .
in particular , 1 .
@xmath898 and @xmath899 have complexity @xmath167 .
2 . @xmath900 has complexity @xmath449 .
3 . @xmath901 has complexity @xmath333 .
the remainder of this paper is devoted to the proof of this main theorem .
the longer parts of the proof are separated into individual propositions and lemmas .
dfa @xmath902 is easily seen to be minimal .
language @xmath853 is @xmath29-proper by propositions [ prefixchar ] and [ convexclasses ] . 1 .
see proposition [ prop : propersyntactic ] .
if the initial state of @xmath903 is changed to @xmath904 , the new dfa accepts a quotient of @xmath905 and is still minimal ; hence the complexity of that quotient is @xmath19 .
if the initial state is changed to @xmath906 then states in @xmath845 are unreachable , but the dfa on @xmath907 is minimal ; hence the complexity of that quotient is @xmath869 .
the remaining quotient is empty , and hence has complexity 1 . by proposition [ prefixchar ] , these are maximal for @xmath29-proper languages .
see propositions [ prop : properreversebound ] and [ prop : properreverse ] for the reverse .
it was shown in @xcite that the number of atoms is equal to the complexity of the reverse .
4 . see proposition [ prop : properatoms ] .
see propositions [ prop : properstarbound ] and [ prop : properstar ] .
see propositions [ prop : properproductbound ] and [ prop : properproduct ] .
the maximal value of @xmath908 occurs only when @xmath909 , since the complexity is increasing with @xmath910 .
7 . by ( * ? ? ?
* theorem 2 ) , all boolean operations on regular languages over the same alphabet have the upper bound @xmath167 . since all proper languages have an empty quotient this gives the bound for * _ a _ * ; the bounds for * _ b _ * and * _ c _ * follow from ( * ? ? ? * theorem 5 ) .
see proposition [ prop : properboolean ] for the proof that all these bounds are tight for @xmath897 .
[ lem : properpermutations ] let @xmath911 and @xmath838 . for any permutation @xmath73 of @xmath70
such that @xmath912 , @xmath913 , and @xmath914 , there is a word @xmath294 that induces @xmath73 on @xmath915 .
only @xmath79 and @xmath178 induce permutations of @xmath70 ; every other letter induces a properly injective map .
furthermore , @xmath79 and @xmath178 permute @xmath845 and @xmath846 separately , and both fix @xmath33 .
hence every @xmath294 induces a permutation on @xmath70 such that @xmath916 , @xmath917 , and @xmath918 .
each such permutation naturally corresponds to an element of @xmath919 , where @xmath920 denotes the symmetric group on @xmath24 elements . to be consistent with the dfa ,
assume @xmath921 contains permutations of @xmath922 and @xmath923 contains permutations of @xmath924 .
let @xmath925 and @xmath926 denote the group elements corresponding to the transformations induced by @xmath79 and @xmath178 respectively .
we show that @xmath925 and @xmath926 generate @xmath919 .
it is well known that @xmath927 , and @xmath928 generate the symmetric group on @xmath929 for any @xmath930 .
note that @xmath931 and @xmath932 are also generators , since @xmath933 .
if @xmath934 and @xmath887 then @xmath919 is the trivial group . if @xmath935 and @xmath215 , then @xmath936 and @xmath926 is either @xmath937 or @xmath938 , and either pair generates the group .
there is a similar argument when @xmath939 .
assume now @xmath940 and @xmath215 . if @xmath836 is odd then @xmath941 , and hence @xmath942 .
similarly if @xmath836 is even then @xmath943 , and hence @xmath944 .
therefore @xmath945 is always generated by @xmath925 . by symmetry , @xmath946 is always generated by @xmath926 regardless of the parity of @xmath29 .
since we can isolate the transposition component of @xmath925 , we can isolate the other component as well : @xmath947 is either @xmath948 or @xmath949 . paired with @xmath946 , either element is sufficient to generate @xmath950 .
similarly , @xmath925 and @xmath926 generate @xmath951 . therefore @xmath925 and @xmath926 generate @xmath919 .
it follows that @xmath79 and @xmath178 generate all permutations @xmath73 of @xmath70 such that @xmath912 , @xmath913 , and @xmath914 .
[ prop : propersyntactic ] the syntactic semigroup of @xmath853 has cardinality @xmath865 , which is maximal for a @xmath29-proper language . furthermore , seven letters are required to meet this bound . the maximum value @xmath866 is reached only when @xmath867 .
let @xmath0 be a @xmath29-proper language of complexity @xmath19 and let @xmath55 be a minimal dfa recognizing @xmath0 .
by lemma [ idealorempty ] , @xmath55 has an empty state . by proposition [ prefixchar ] , the only states that can be reached from one of the @xmath29 final states are either final or empty .
thus , a transformation in the transition semigroup of @xmath55 may map each final state to one of @xmath869 possible states , while each non - final , non - empty state may be mapped to any of the @xmath19 states .
since the empty state can only be mapped to itself , we are left with @xmath865 possible transformations in the transition semigroup .
therefore the syntactic semigroup of any @xmath29-proper language has size at most @xmath865 .
now consider the transition semigroup of @xmath852 .
every transformation @xmath73 in the semigroup must satisfy @xmath952 and @xmath914 , since any other transformation would violate prefix - convexity .
we show that the semigroup contains every such transformation , and hence the syntactic semigroup of @xmath853 is maximal .
first , consider the transformations @xmath73 such that @xmath953 and @xmath954 for all @xmath955 . by lemma [ lem : properpermutations ] ,
@xmath79 and @xmath178 generate every permutation of @xmath845 .
when @xmath73 is not a permutation , we can use @xmath847 to combine any states @xmath62 and @xmath59 : apply a permutation on @xmath845 so that @xmath956 and @xmath957 , and then apply @xmath847 so that @xmath958 . repeat this method to combine any set of states , and further apply permutations to induce the desired transformation while leaving the states of @xmath959 in place .
the same idea applies with @xmath848 ; apply permutations and @xmath848 to send any states of @xmath845 to @xmath33 .
hence @xmath79 , @xmath178 , @xmath847 , and @xmath848 generate every transformation @xmath73 such that @xmath960 and @xmath961 for all @xmath962 .
we can make the same argument for transformations that act only on @xmath846 and fix every other state . since @xmath849 and @xmath850 act on @xmath846 exactly as @xmath847 and @xmath848 act on @xmath845 , @xmath79 , @xmath178 , @xmath847 , and @xmath848 generate every transformation @xmath73 such that @xmath952 and @xmath961 for all @xmath963 .
it follows that @xmath79 , @xmath178 , @xmath847 , @xmath849 , @xmath848 , and @xmath850 generate every transformation @xmath73 such that @xmath964 , @xmath952 , and @xmath914 . note the similarity between this dfa restricted to the states @xmath965 ( or @xmath959 ) and the most complex witness for right ideals .
the argument for the size of the syntactic semigroup of right ideals is similar to this ; see @xcite . finally , consider an arbitrary transformation @xmath73 such that @xmath952 and @xmath914 .
let @xmath966 be the number of states @xmath967 such that @xmath968 .
we show by induction on @xmath966 that @xmath73 is in the transition semigroup of @xmath55 .
if @xmath969 , then @xmath73 is generated by @xmath970 .
if @xmath971 , there exist @xmath972 such that @xmath968 and @xmath59 is not in the image of @xmath73 .
consider the transformations @xmath973 and @xmath974 defined by @xmath975 and @xmath976 for @xmath977 , and @xmath978 and @xmath979 for @xmath980 .
notice that @xmath981 and @xmath982 .
one can verify that @xmath983 . from this expression
, we see that @xmath973 is the composition of transpositions induced by words in @xmath276 and the transformation @xmath984 induced by @xmath764 , and hence @xmath973 is generated by @xmath1 .
thus , @xmath981 is in the transition semigroup . by induction on @xmath966
, it follows that the syntactic semigroup of @xmath905 is maximal .
now we show that seven letters are required to meet this bound .
two letters ( like @xmath79 and @xmath178 ) that induce permutations are required to generate permutations , since clearly one letter is not sufficient .
every other letter will induce a properly injective map .
a letter ( like @xmath847 ) that induces a properly injective map on @xmath845 and permutes @xmath846 is required .
similarly , a letter ( like @xmath849 ) that permutes @xmath845 and induces a properly injective map on @xmath846 is required .
a letter ( like @xmath848 ) that sends a state in @xmath845 to @xmath33 and permutes @xmath846 is required .
similarly , a letter ( like @xmath850 ) that sends a state in @xmath846 to @xmath33 and permutes @xmath845 is required . finally , a letter ( like @xmath764 ) that connects @xmath845 and @xmath846 is required . for a fixed @xmath19 , we may want to know which @xmath985 maximizes @xmath986 ; this corresponds to the largest syntactic semigroup of a proper prefix - convex language with @xmath19 quotients .
we show that @xmath987 is largest at @xmath988 .
consider the ratio @xmath989 .
notice this ratio is increasing with @xmath29 , and hence @xmath990 is a convex function on @xmath991 .
it follows that the maximum value of @xmath990 must occur at one the endpoints , @xmath992 and @xmath153 .
now we show that @xmath993 for all @xmath837 . we can check this explicitly for @xmath994 .
when @xmath995 , @xmath996 ; hence the largest syntactic semigroup of @xmath853 occurs only at @xmath988 for all @xmath837 .
[ prop : properreversebound ] for any regular language @xmath0 of complexity @xmath19 with an empty quotient , the reversal has complexity at most @xmath323 .
let @xmath997 be a minimal dfa recognizing @xmath0 .
construct an nfa @xmath998 recognizing the reverse of @xmath0 from @xmath55 by reversing each transition , letting the initial state @xmath187 be the unique final state , and letting the final states @xmath66 be the initial states .
notice that the empty state of @xmath55 is not reachable in @xmath998 since the transitions have been reversed ; thus we may delete it . applying the subset construction to @xmath998 yields a dfa whose states are subsets of @xmath345 , and hence has at most @xmath323 states .
[ prop : properreverse ] the reverse of @xmath999 has complexity @xmath323 for @xmath837 and @xmath838 .
let @xmath1000 denote the dfa @xmath902 of definition [ def : proper ] and let @xmath1001 .
as before , construct an nfa @xmath998 recognizing the reverse of @xmath905 .
applying the subset construction to @xmath998 yields a dfa @xmath456 whose states are subsets of @xmath345 , with initial state @xmath846 and final states @xmath1002 .
we show that @xmath456 is minimal , and hence the reverse of @xmath905 has complexity @xmath323 .
we check reachability and distinguishability .
recall from lemma [ lem : properpermutations ] that @xmath79 and @xmath178 generate all permutations of @xmath845 and @xmath846 in @xmath915 and , although the transitions are reversed in @xmath456 , they still generate all such permutations .
let @xmath1003 be such that @xmath1004 induces @xmath1005 and @xmath1006 induces @xmath1007 in @xmath456 .
consider a state @xmath1008 where @xmath1009 .
if @xmath1010 , then @xmath1011 is the initial state . when @xmath1012 , @xmath1013 . by induction ,
all such states are reachable .
now we show that any state @xmath1014 where @xmath1009 and @xmath1015 is reachable . if @xmath1016 , then @xmath1008 is reachable by the argument above . when @xmath1017 , choose @xmath1018 and see that @xmath1011 is reached from @xmath1019 by @xmath1020 .
by induction , every state is reachable . to prove distinguishability , consider distinct states @xmath1011 and @xmath1021 .
choose @xmath1022 .
if @xmath904 , then @xmath1011 and @xmath1021 are distinguished by @xmath1023 .
when @xmath906 , they are distinguished by @xmath1024 .
hence , the states are pairwise distinguishable and @xmath456 is minimal .
[ prop : properatoms ] for each atom @xmath45 of @xmath853 , write @xmath871 , where @xmath872 and @xmath873 .
let @xmath874 and @xmath875 . if @xmath876 , then @xmath877 @xmath878 if @xmath879 and @xmath880 , then @xmath877 @xmath881 otherwise , @xmath882 and @xmath883 . the atomic complexity is maximal for @xmath29-proper languages .
let @xmath0 be a @xmath29-proper language with quotients @xmath1025 where @xmath1026 are non - final quotients , @xmath1027 are final quotients , and @xmath1028 . for @xmath1029 , we have @xmath468 ; note @xmath1030 since @xmath45 must be non - empty .
the quotients are @xmath1031 .
however @xmath1032 is always another quotient @xmath1033 .
thus @xmath632 has the form @xmath1034 where @xmath634 and @xmath635 . for brevity , we write @xmath636 and @xmath637 ; this notation is in agreement with the action of @xmath8 on the states of @xmath915 corresponding to @xmath46 and @xmath470 . to establish the upper bound , we just count the number of possible distinct @xmath649 for each value of @xmath46 .
notice @xmath1035 and if @xmath1036 then @xmath649 is the empty quotient .
write @xmath871 where @xmath872 and @xmath873 , and let @xmath874 and @xmath875 . by proposition [ prefixchar ] every word @xmath8 maps @xmath1037 to a subset of @xmath70 and @xmath1038 to a subset of @xmath1039 .
similarly , @xmath8 maps @xmath1040 to a subset of @xmath70 , @xmath1041 to a subset of @xmath959 , and @xmath33 to itself .
the number of disjoint sets @xmath1042 that are reached from @xmath46 and @xmath470 respectively by some word @xmath8 is an upper bound on the number of non - empty quotients of @xmath45 . specifically , we require @xmath647 , @xmath645 , @xmath646 , @xmath1043 , @xmath1044 , and @xmath1045 .
thus we have the initial estimate @xmath1046 where @xmath1047 counts @xmath1048 , @xmath1049 counts @xmath1050 , @xmath1051 counts @xmath1052 , and @xmath1053 counts @xmath1054 . with some refinements ,
this estimate leads to the three cases in the statement .
note if @xmath240 then @xmath1055 .
also , if @xmath876 , then any non - empty quotient @xmath649 must have @xmath1056 since @xmath1038 can not be mapped to @xmath33 ; this has the effect that @xmath1049 can not be 0 in our initial estimate .
since the estimate only counts non - empty quotients , we add 1 to represent the empty quotient achieved when @xmath639 and @xmath640 intersect , yielding the expression for the case @xmath876 .
if @xmath879 and @xmath880 , then we can not have @xmath1057 since that would correspond to @xmath1058 ; the subtracted term in the statement is the value of the estimate when @xmath1059 . as
before , add 1 for the empty quotient . finally ,
if @xmath612 , then @xmath1060 and @xmath1061 with @xmath647 . there are @xmath323 possible values of @xmath640 , hence @xmath1062 .
since @xmath639 and @xmath640 can not intersect in this case , there is no need to add 1 for an empty quotient , as in the previous two cases .
thus we have the three cases in the statement .
it remains to prove that @xmath853 of definition [ def : proper ] meets this upper bound .
let the quotient @xmath1063 of @xmath905 be the language accepted by state @xmath59 in @xmath915 .
we must show that every @xmath649 can be reached from @xmath45 by some word in @xmath344 , and that every non - empty @xmath649 is distinct from @xmath660 whenever @xmath1064 . by proposition [ prop : propersyntactic ] , the syntactic semigroup is as large as possible for @xmath29-proper languages .
hence , whenever @xmath647 , @xmath645 , @xmath646 , @xmath1043 , and @xmath1044 , there is a word @xmath56 such that @xmath636 and @xmath637 .
thus each quotient @xmath649 counted by the upper bound is reachable in @xmath45 .
consider @xmath649 where @xmath1065 and @xmath647 . if @xmath1055 then there exists @xmath1066 such that @xmath1067 and @xmath1068 ; hence @xmath1069 since @xmath1070 . if @xmath1058 choose @xmath8 such that @xmath1068 ; hence @xmath1069 .
thus @xmath649 is non - empty .
now take @xmath660 where @xmath1064 , @xmath1071 and @xmath683 .
we must show that @xmath649 and @xmath660 are distinct .
note @xmath1072 where @xmath1073 and @xmath1074 . if @xmath1075 , then choose @xmath8 that maps @xmath1076 in @xmath915 ; @xmath1077 is non - empty , since @xmath1078 , and @xmath1079 since @xmath1080 .
similarly , if @xmath1081 and @xmath1082 , then choose @xmath8 that maps @xmath1083 and @xmath1084 . then @xmath1085 is non - empty and @xmath1086 . finally , if @xmath1087 and @xmath1082 , then distinguish @xmath649 and @xmath660 by a word that sends @xmath1088 and @xmath1089 .
hence , @xmath649 and @xmath660 are distinct in all cases .
therefore , the quotients of @xmath45 counted in the upper bound are pairwise distinct and @xmath905 has maximal atomic complexity .
[ prop : properstarbound ] suppose @xmath0 is a language with @xmath19 quotients including an empty quotient .
let @xmath813 be the number of final quotients
. then @xmath1090 .
this bound also holds for all prefix - convex languages when @xmath837 .
if @xmath1091 , then @xmath776 has an empty quotient , and obeys the upper bound .
assume @xmath1092 .
let @xmath1093 be a minimal dfa recognizing @xmath0 .
since @xmath0 has an empty quotient , let @xmath33 be the empty state of @xmath55 . to obtain an @xmath13-nfa for @xmath695
, we add a new initial state @xmath197 which is final and has the same transitions as @xmath187 .
we then add an @xmath13-transition from every state in @xmath66 to @xmath187 . applying the subset construction to this
@xmath13-nfa yields a dfa @xmath1094 recognizing @xmath695 , in which the states are non - empty subsets of @xmath1095 .
many of these states are unreachable or indistinguishable from other states .
first , since there is no transition in the @xmath13-nfa to @xmath197 , the only reachable state in @xmath1096 containing @xmath197 is @xmath205 .
second , because of the @xmath13-transitions , any reachable final state @xmath1097 must have @xmath1098 .
finally , for any @xmath1099 , we have @xmath1100 if and only if @xmath1101 .
furthermore , @xmath1102 for all @xmath56 .
therefore , the states @xmath1011 and @xmath1103 are equivalent in @xmath1104 .
this allows us to deal with subsets of @xmath345 instead of @xmath70 .
combined with the first two reductions , this proves that @xmath1105 is equivalent to a reduced dfa with the states @xmath1106 .
this dfa has @xmath1107 states .
thus , @xmath1090 .
this bound also applies when @xmath0 is any prefix - convex language and @xmath837 : by lemma [ idealorempty ] , @xmath0 is either a right ideal or has an empty state . if @xmath0 is a right ideal , then @xmath1108 , which is at most @xmath885 for @xmath837 .
[ prop : properstar ] the star of @xmath884 has complexity @xmath1109 for @xmath837 and @xmath838 .
let @xmath1110 denote the dfa @xmath1111 of definition [ def : proper ] and let @xmath1001 .
we apply the same construction and reduction as in proposition [ prop : properstarbound ] to obtain a dfa @xmath1112 recognizing @xmath1113 with states @xmath1114 .
it is clear that @xmath1115 .
we must prove that @xmath1116 is minimal .
to do so , we show that every state in @xmath1096 is reachable and that all distinct states are distinguishable . by lemma [ lem : properpermutations ] , @xmath79 and @xmath178 generate all permutations of @xmath845 and @xmath846 in @xmath915 .
let @xmath1003 be such that @xmath1004 induces @xmath1005 and @xmath1006 induces @xmath1007 in @xmath915 . for reachability , we consider three cases . 1
. state @xmath205 is reachable by @xmath13 .
2 . let @xmath1117 .
for any @xmath904 , we can reach @xmath1118 by @xmath1119 ; hence if @xmath1011 is reachable , then every subset of @xmath1011 is reachable .
observe that state @xmath845 is reachable by @xmath1120 , and we can reach any subset of this state . therefore , all non - final states are reachable .
3 . if @xmath1121 , then @xmath1122 where @xmath1123 and @xmath1124 and @xmath1125 .
we prove that @xmath1011 is reachable by induction on @xmath1126 .
if @xmath1127 , then @xmath1011 is reachable by * 2*. for any @xmath1125 , we can reach @xmath1011 from @xmath1128 by @xmath1129 .
therefore , all states of this form are reachable .
now we show that the states are pairwise distinguishable . 1 .
the initial state @xmath205 is distinguishable from any other final state @xmath1011 since @xmath1130 is non - final and @xmath1131 is final .
2 . suppose @xmath1011 and @xmath1021 are distinct subsets of @xmath845 .
there is some @xmath1022 .
we distinguish @xmath1011 and @xmath1021 by @xmath1132 .
3 . suppose @xmath1011 and @xmath1021 are distinct and final and neither one is @xmath205 .
there is some @xmath1022 . if @xmath904 , then @xmath1133 and @xmath1134 are distinct , non - final states as in * 2*. otherwise , @xmath906 and we distinguish @xmath1011 and @xmath1021 by @xmath1135 . therefore , @xmath1116 is minimal and @xmath1136 .
[ prop : properproductbound ] let @xmath182 and @xmath0 be proper prefix - convex languages with @xmath1137 and @xmath1092 quotients respectively .
let @xmath910 be the number of final quotients in @xmath182 and @xmath29 be the number of final quotients in @xmath0 .
the product of @xmath182 and @xmath0 has complexity at most @xmath1138 .
let @xmath185 and @xmath55 be minimal dfas for @xmath182 and @xmath0 respectively .
let @xmath70 be the state set of @xmath55 .
to avoid confusing states in the product , take @xmath185 to have the states @xmath1139 . applying the standard dfa construction for the product @xmath184
, we obtain states of the form @xmath188 for @xmath1140 and @xmath190 .
we prove the upper bound by determining reachability and distinguishability for these states . by lemma [ idealorempty ] , each of @xmath182 and @xmath0 has an empty quotient .
assume @xmath200 is the empty state in @xmath185 and @xmath33 is the empty state in @xmath55 .
take @xmath1141 to be the non - final , non - empty states and @xmath1142 to be the final states in @xmath185 . by proposition [ prefixchar ] ,
no non - final , non - empty state can be reached from a final state in @xmath185 .
therefore , every reachable state in the product is either @xmath206 for @xmath1143 , @xmath191 for @xmath1144 and @xmath1145 , or @xmath486 for @xmath1146 .
this gives the preliminary estimate of @xmath1147 states .
for @xmath1148 , @xmath188 is equivalent to @xmath1149 for any @xmath190 .
this removes @xmath1150 states from our estimate , leaving @xmath1151 states .
[ prop : properproduct ] for @xmath244 , @xmath894 , and @xmath838 , the product of @xmath1152 and @xmath1153 has complexity @xmath1154 .
let @xmath1155 and @xmath915 be the dfas of definition [ def : proper ] for languages @xmath888 and @xmath1153 respectively . as before , take @xmath1155 to have the states @xmath1139 and let @xmath1156 .
using the standard construction of the @xmath13-nfa @xmath998 for the product , we delete the empty state @xmath33 , change the final states of @xmath1155 to non - final states , and add @xmath13-transitions from each final state of @xmath1155 to the initial state of @xmath915 . the subset construction on @xmath998 yields states of the form @xmath188 , where @xmath1140 and @xmath1029 .
however , some of these sets are not reachable in the product : if @xmath1157 then we must have @xmath882 , and if @xmath1158 then @xmath1159 because of the @xmath13-transitions in @xmath998 .
thus , we have the states @xmath206 for @xmath1157 , @xmath492 for @xmath1158 and @xmath1160 , and @xmath486 for @xmath1029 .
this totals to @xmath1161 different states .
we show that they are reachable and pairwise distinguishable .
state @xmath206 is reached by @xmath1162 for all @xmath1157 .
state @xmath1163 is reached by @xmath764 .
for @xmath1164 we have @xmath1165 .
now consider states of the form @xmath1166 where @xmath1158 and @xmath1167 .
these states are reachable when @xmath1168 .
inductively assume the states are reachable when @xmath1169 for some @xmath1125 .
let @xmath1170 where @xmath1171 , and let @xmath1172 . then @xmath1173 .
notice @xmath178 induces a permutation on @xmath1155 , so for any @xmath1158 there is a state @xmath1174 such that @xmath1175 .
thus , @xmath1176 is reachable from @xmath1177 .
extend this to states of the form @xmath1178 , where @xmath1179 , @xmath1180 , and @xmath1167 .
these states are reachable when @xmath882 .
inductively assume the states are reachable when @xmath1181 for some @xmath1012 .
let @xmath1182 where @xmath1183 , and let @xmath1184 . then @xmath1185 . in the last derivation
, @xmath1186 denotes the set @xmath1187 .
state @xmath1188 is reachable from @xmath1189 by @xmath1190 , where @xmath1191 is the order of @xmath850 in @xmath915 ( i.e. @xmath1190 induces the identity transformation on @xmath915 ) . finally , state @xmath1192 is reachable from @xmath1188 : by lemma [ lem : properpermutations ] , the permutation @xmath1193 of @xmath915 is generated by some @xmath1194 , and @xmath1195 .
thus all states are reachable .
we now check distinguishability in cases . using lemma [ lem : properpermutations ] , take words @xmath1196 such that @xmath1004 induces @xmath1193 and @xmath1006 induces @xmath1197 on @xmath915 .
note @xmath1004 and @xmath1006 act on @xmath1155 as well . 1 .
let @xmath1198 and let @xmath1021 be any other state .
notice @xmath1011 is the empty state .
we show that @xmath1021 is non - empty
if @xmath1199 then by the minimality of @xmath915 there is a word @xmath8 such that @xmath1200 ; hence @xmath1201 is final .
2 . otherwise @xmath1202 for some @xmath1157 .
there is a word @xmath8 such that @xmath1203 ; hence @xmath1204 and this reduces to * _ a_*. 2 .
let @xmath1205 and @xmath1206 where @xmath1207 and @xmath451 .
then @xmath1208 and @xmath1209 is non - empty by * 1*. 3 .
let @xmath1205 and @xmath1210 where @xmath1157 , @xmath1174 , and @xmath1160 .
then @xmath1011 and @xmath1021 are distinguished by @xmath764 .
4 . let @xmath1205 and @xmath1211 where @xmath1157 and @xmath1029 .
if @xmath882 this reduces to * 1*. if @xmath1212
then @xmath1021 is final .
otherwise there is some @xmath1213 , and @xmath1214 is final .
notice @xmath1215 is non - final because @xmath1216 .
let @xmath1217 and @xmath1218 where @xmath1219 ; pick @xmath503 . without loss of generality ,
say @xmath1220 . 1 . if @xmath1221 , then @xmath1222 and @xmath1223 .
2 . if @xmath1224 , then we reduce to * _ a _ * by applying @xmath1225 .
3 . if @xmath1226 , then @xmath1227 is final and @xmath1228 is non - final .
4 . if @xmath1229 , then we reduce to * _ c _ * by applying @xmath1230 . 6 . let @xmath1231 and @xmath1218 where @xmath1158 , and @xmath1232 .
notice @xmath1233 is non - empty since @xmath809 is not mapped to @xmath200 , but @xmath1234 ; this reduces to * 1*. 7 .
let @xmath1231 and @xmath1235 where @xmath1236 , @xmath451 , and @xmath1232 .
reduce to * 6 * by applying @xmath1237 .
let @xmath1231 and @xmath1238 where @xmath1158 and @xmath1219 .
pick @xmath503 and assume without loss of generality that @xmath1220 .
1 . if @xmath1239 , then @xmath1240 fixes @xmath790 and maps @xmath809 to @xmath200 ; hence this reduces to * 5*. 2 . if @xmath1241 , then apply @xmath850 to reduce to * 5*. notice @xmath1242 and @xmath1243 are distinct since @xmath850 induces a permutation on @xmath915 .
if @xmath1244 and @xmath940 , then applying @xmath848 reduces to * _ a_*. 4 .
if @xmath1244 and @xmath1245 , then observe that @xmath79 and @xmath178 both fix @xmath992 in @xmath915 . by lemma [ lem : properpermutations ] , there is a word @xmath294 such that @xmath1246 .
since @xmath1247 , @xmath79 and @xmath178 do not alter @xmath845 .
hence @xmath1248 and @xmath1249 , so this reduces to * _ b_*. [ prop : properboolean ] for @xmath893 , @xmath894 , and @xmath838 , let @xmath895 and @xmath896 of definition [ def : proper ] .
for any proper binary boolean function @xmath20 , the complexity of @xmath897 is maximal .
in particular , 1 .
2 . @xmath1251 .
3 . @xmath1252 .
let @xmath1155 and @xmath915 be the dfas of definition [ def : proper ] for languages @xmath1253 and @xmath1254 respectively . as before , take @xmath1155 to have the states @xmath1139 .
there is a standard construction for @xmath897 for any boolean set operation @xmath20 in terms of the direct product .
the direct product of @xmath1155 and @xmath915 has states @xmath1255 , initial state @xmath514 , and transition function @xmath53 such that @xmath1256 .
if we set the final states to be @xmath1257 , it is a dfa recognizing @xmath897 . for each @xmath260
, we construct the dfa @xmath1258 to recognize @xmath897 .
all four dfas have the same states and transitions as the direct product and will only differ in the set of final states .
the dfa @xmath1259 for symmetric difference is shown in figure [ fig : properboolean ] .
we can usefully partition the states of the direct product .
let @xmath1260 , @xmath1261 , @xmath1262 , @xmath1263 , and @xmath1264 .
let @xmath525 and @xmath1265 . 8.5pt ( 35,23)(0,-6 ) ( 00,20)@xmath764 ( 00,02)@xmath847 ( 00,01)@xmath850 ( 10,11)@xmath850 ( 00,10)@xmath848 ( 01,11)@xmath848 ( 03,14)@xmath848 ( 30,41)@xmath850 ( 10,40)@xmath848 ( 11,41)@xmath848 ( 01,04)@xmath850 ( 11,14)@xmath850 ( 23,24)@xmath848 ( 33,34)@xmath848 ( 04,14)@xmath848 ( 14,44)@xmath848 ( 24,34)@xmath850 ( 34,44)@xmath850 ( 32,42)@xmath850 ( 33,43)@xmath850 ( 40,41)@xmath850 ( 41,44)@xmath850 ( 42,43)@xmath848 ( 43,44)@xmath848 ( 22,32)@xmath850 ( 23,33)@xmath850 ( 22,23)@xmath848 ( 32,33)@xmath848 ( 02,22)@xmath764 ( 03,23)@xmath764 ( 20,22)@xmath847 ( 30,32)@xmath847 we check that every state in the direct product is reachable . since @xmath1266 , @xmath1267 , @xmath1268 , and
@xmath1269 have the same structure as the direct product , this argument will apply to them as well . by lemma [ lem : properpermutations ] there exist @xmath1003 such that @xmath1004 induces @xmath1270 and @xmath1006 induces @xmath1271 in @xmath1272 .
note that @xmath1004 and @xmath1006 permute @xmath845 and @xmath846 in @xmath1273 .
similarly , there exist @xmath1274 such that @xmath1275 induces @xmath1005 and @xmath1276 induces @xmath1277 in @xmath1273 , and they permute @xmath1278 and @xmath1279 in @xmath1272 . 1 . state @xmath1280 is reachable since @xmath1281 .
2 . state @xmath1282 is reachable since @xmath1283 .
an arbitrary @xmath1284 is then reached by @xmath1285 from some @xmath1286 where @xmath1287 is chosen so that @xmath1288 in @xmath1272 .
state @xmath1289 is reachable by symmetry with * 2*. 4 .
state @xmath1290 is reachable since @xmath1291 .
state @xmath1292 is reachable since @xmath1293 , and @xmath809 is reachable in @xmath1272 .
state @xmath1294 is reachable by symmetry with * 5*. hence all states are reachable . as a tool for distinguishability
, we show that the states of @xmath46 are distinguishable with respect to @xmath300 ; that is , for any pair of distinct states in @xmath46 , we show that there is a word that sends one state to @xmath300 and leaves the other state in @xmath46 .
we check this fact in cases .
note that @xmath850 fixes the states of @xmath639 and @xmath848 fixes the states of @xmath640 .
1 . states of @xmath1295 and @xmath639 are distinguished by words in @xmath1296 .
states of @xmath1295 and @xmath640 are distinguished by words in @xmath1297 .
states of @xmath639 and @xmath640 are distinguished by words in @xmath1297 .
states of @xmath639 and @xmath1298 are distinguished by words in @xmath1296 .
states of @xmath640 and @xmath1298 are distinguished by words in @xmath1297 .
6 . to distinguish states of @xmath1295 and @xmath1298 ,
we reduce to * 5 * by a word in @xmath1299 .
any two states of @xmath1295 are distinguished by a word in @xmath1297 if they differ in the first coordinate , or by a word in @xmath1296 if they differ in the second coordinate .
any two states of @xmath1298 are distinguished by a word in @xmath1296 if they differ in the first coordinate , or by a word in @xmath1297 if they differ in the second coordinate .
9 . to distinguish two states of @xmath639 , reduce to * 4 * by a word in @xmath1299 if they differ in the first coordinate , or reduce to * 8 * by a word in @xmath1299 if the first coordinate is the same
any two states of @xmath640 are distinguishable by symmetry with * 9*. now we determine which states are pairwise distinguishable with respect to the final states of @xmath1300 for each @xmath260 . let @xmath1301 ; observe that @xmath8 maps every state of @xmath46 to a state of @xmath1298 .
@xmath538 , @xmath742 : in @xmath1302 , @xmath293 is final if @xmath1158 or @xmath906 . in @xmath1259 , @xmath293 is final if @xmath1158 and @xmath1303 or @xmath1304 and @xmath906 .
we show that all @xmath167 states are pairwise distinguishable in both cases .
the states of @xmath283 are pairwise distinguishable by the minimality of @xmath1273 .
similarly , the states of @xmath280 are pairwise distinguishable by the minimality of @xmath1272 .
the states of @xmath283 and @xmath280 are distinguishable by @xmath1305 , since @xmath1306 and @xmath1307 .
the states of @xmath280 and @xmath46 are distinguishable since @xmath1308 , and we can distinguish states of @xmath283 and @xmath280 .
the states of @xmath283 and @xmath46 are similarly distinguishable .
finally , states of @xmath46 are pairwise distinguishable because they can be distinguished with respect to @xmath300 , and we can distinguish states of @xmath46 and @xmath300 .
@xmath536 : in @xmath1309 , @xmath293 is final if @xmath1158 and @xmath1303 . the states of @xmath283 are all empty , and the remaining states are pairwise distinguishable for a total of @xmath1310 distinguishable states
. the states of @xmath280 are pairwise distinguishable by the minimality of @xmath1272 .
the states of @xmath280 and @xmath46 are distinguishable since @xmath1308 , and every state in @xmath283 is empty .
finally , states of @xmath46 are pairwise distinguishable because they can be distinguished with respect to @xmath300 , and we can distinguish states of @xmath46 and @xmath300 .
@xmath743 : in @xmath1311 the final state set is @xmath1298
. the states of @xmath300 are all empty , leaving @xmath1312 distinguishable states .
the states of @xmath46 are non - empty since @xmath1313
. we can distinguish the states of @xmath46 with respect to @xmath300 ; hence they are pairwise distinguishable .
our results are summarized in table [ tab : summary ] .
the largest bounds are shown in boldface type , and they are reached in the classes of ideal languages , closed languages , and proper languages . recall that for regular languages we have the following results : semigroup : @xmath34 ; reverse : @xmath36 ; star : @xmath161 ; restricted product : @xmath1314 ; unrestricted product : @xmath165 ; restricted @xmath538 and @xmath742 : @xmath167 ; unrestricted @xmath538 and @xmath742 : @xmath279 ; restricted @xmath536 : @xmath167 ; unrestricted @xmath536 : @xmath172 ; restricted @xmath743 : @xmath167 ; unrestricted @xmath743 : @xmath167 . in table
[ tab : summary ] , @xmath283 and @xmath1011 stand for _ restricted _ and _ unrestricted _ , respectively .
j. bell , j. brzozowski , n. moreira , and r. reis .
symmetric groups and quotient complexity of boolean operations . in j.
esparza and et al . ,
editors , _ icalp 2014 _ , volume 8573 of _ lncs _ , pages 112 .
springer , 2014 .
han , k. salomaa , and d. wood .
operational state complexity of prefix - free regular languages . in z.
sik and z flp , editors , _ automata , formal languages , and related topics _ , pages 99115 .
institute of informatics , university of szeged , hungary , 2009 .
g. jirskov and m. krausov .
complexity in prefix - free regular languages . in i.
mcquillan , g. pighizzini , and b. trost , editors , _ proceedings of the 12th international workshop on descriptional complexity of formal systems @xmath1316dcfs@xmath1317 _ , pages 236244 .
university of saskatchewan , 2010 .
m. krausov .
prefix - free regular languages : closure properties , difference , and left quotient . in z.
kotsek , j. bouda , i. cern , l. sekanina , t. vojnar , and d. antos , editors , _ memics _ , volume 7119 of _ lecture notes in computer science _ , pages 114122 .
springer , 2011 .
b. krawetz , j. lawrence , and j. shallit .
state complexity and the monoid of transformations of a finite set . in m.
domaratzki , a. okhotin , k. salomaa , and s. yu , editors , _ proceedings of the implementation and application of automata , @xmath1316ciaa@xmath1317 _ , volume 3317 of _ lncs _ , pages 213224 .
springer , 2005 . | a language @xmath0 over an alphabet @xmath1 is prefix - convex if , for any words @xmath2 , whenever @xmath3 and @xmath4 are in @xmath0 , then so is @xmath5 .
prefix - convex languages include right - ideal , prefix - closed , and prefix - free languages .
we study complexity properties of prefix - convex regular languages .
in particular , we find the quotient / state complexity of boolean operations , product ( concatenation ) , star , and reversal , the size of the syntactic semigroup , and the quotient complexity of atoms . for binary operations we use arguments with different alphabets when appropriate ; this leads to higher tight upper bounds than those obtained with equal alphabets .
we exhibit most complex prefix - convex languages that meet the complexity bounds for all the measures listed above .
* keywords : * complexity of operations , different alphabets , prefix - closed , prefix - convex , prefix - free , regular languages , right ideals , state complexity , syntactic semigroup , unrestricted |
the skein polynomial ( as called in ( * ? ? ? * chapter 8) , also known as homfly or homfly - pt polynomial ) , @xmath0 $ ] , is an invariant for oriented links . here
$ ] is the ring of laurent polynomials in two variables @xmath2 and @xmath3 , with integer coefficients .
it is defined to be the invariant of oriented links satisfying the axioms @xmath4{fig_012/1a_dd-0 } \end{pmatrix } -a\cdot p \begin{pmatrix } \includegraphics[width=.8cm]{fig_012/1b_dd-0 } \end{pmatrix } = z\cdot p \begin{pmatrix } \includegraphics[width=.8cm]{fig_012/1o_dd-0 } \end{pmatrix } ; \tag*{\rm(i ) } \\ p \begin{pmatrix } \includegraphics[width=.8cm]{fig_012/o-0 } \end{pmatrix } = 1 .
\tag*{\rm(o)}\end{gathered}\ ] ] the alexander - conway polynomial @xmath5 $ ] and the jones polynomial @xmath6 $ ] are related to the skein polynomial : @xmath7 our main result is [ thm : homfly ] the skein polynomial @xmath8 $ ] is the invariant of oriented links determined uniquely by the following four axioms . @xmath9{fig_012/1a1a_dd-0 } \end{pmatrix } + a^2\cdot p \begin{pmatrix } \includegraphics[width=.8cm]{fig_012/1b1b_dd-0 } \end{pmatrix } = ( 2+z^2 ) \cdot p \begin{pmatrix } \includegraphics[width=.8cm]{fig_012/1oo_dd-0 } \end{pmatrix } ; \tag*{\rm(ii ) } \\
a^{-1}\cdot p \begin{pmatrix } \includegraphics[width=1.2cm]{fig_3icons/1a2a1b_ddd-0 } \end{pmatrix } -a\cdot p \begin{pmatrix } \includegraphics[width=1.2cm]{fig_3icons/1a2b1b_ddd-0 } \end{pmatrix } = a^{-1}\cdot p \begin{pmatrix } \includegraphics[width=1.2cm]{fig_3icons/1b2a1a_ddd-0 } \end{pmatrix } -a\cdot p \begin{pmatrix } \includegraphics[width=1.2cm]{fig_3icons/1b2b1a_ddd-0 } \end{pmatrix } ; \tag*{\rm(iii ) } \\ p \begin{pmatrix } \includegraphics[height=1.2cm]{fig_012/io-0 } \end{pmatrix } = z^{-1}(a^{-1}-a)\cdot p \begin{pmatrix } \;\ ; \includegraphics[height=1.2cm]{fig_012/i-0 } \;\ ; \end{pmatrix } ; \tag*{\rm(io ) } \\ p \begin{pmatrix } \includegraphics[width=.8cm]{fig_012/o-0 } \end{pmatrix } = 1
. \tag*{\rm(o)}\end{gathered}\ ] ] a parallel result is for the jones polynomial . it is not a direct corollary of the above theorem , because the substitutions @xmath10 and @xmath11 do not send @xmath1 $ ] into @xmath12 $ ] . [ thm : jones ] the jones polynomial @xmath6 $ ] is the invariant of oriented links determined uniquely by the following four axioms . @xmath13{fig_012/1a1a_dd-0 } \end{pmatrix } + t^2\cdot v \begin{pmatrix } \includegraphics[width=.8cm]{fig_012/1b1b_dd-0 } \end{pmatrix } = ( t+t^{-1 } ) \cdot v \begin{pmatrix } \includegraphics[width=.8cm]{fig_012/1oo_dd-0 } \end{pmatrix } ; \tag*{\rm(ii$_v$ ) } \\
t^{-1}\cdot v \begin{pmatrix } \includegraphics[width=1.2cm]{fig_3icons/1a2a1b_ddd-0 } \end{pmatrix } -t\cdot v \begin{pmatrix } \includegraphics[width=1.2cm]{fig_3icons/1a2b1b_ddd-0 } \end{pmatrix } = t^{-1}\cdot v \begin{pmatrix } \includegraphics[width=1.2cm]{fig_3icons/1b2a1a_ddd-0 } \end{pmatrix } -t\cdot v \begin{pmatrix } \includegraphics[width=1.2cm]{fig_3icons/1b2b1a_ddd-0 } \end{pmatrix } ; \tag*{\rm(iii$_v$ ) } \\ v \begin{pmatrix } \includegraphics[height=1.2cm]{fig_012/io-0 } \end{pmatrix } = -(t^{\frac12}+t^{-\frac12})\cdot v \begin{pmatrix } \;\ ; \includegraphics[height=1.2cm]{fig_012/i-0 } \;\ ; \end{pmatrix } ; \tag*{\rm(io$_v$ ) } \\ v \begin{pmatrix } \includegraphics[width=.8cm]{fig_012/o-0 } \end{pmatrix } = 1 .
\tag*{\rm(o$_v$)}\end{gathered}\ ] ] for the alexander - conway polynomial , the result takes a slightly different form .
we switch to a ( @xmath14)-type axiom because the ( io)-type one degenerates into a consequence of ( ii ) and ( iii ) ( see corollary [ cor : stabilization_alexander - conway ] ) .
[ thm : alexander - conway ] the alexander - conway polynomial @xmath5 $ ] is the invariant of oriented links determined uniquely by the following four axioms . @xmath15{fig_012/1a1a_dd-0 } \end{pmatrix } + \delta \begin{pmatrix } \includegraphics[width=.8cm]{fig_012/1b1b_dd-0 } \end{pmatrix } = ( t+t^{-1 } ) \cdot \delta \begin{pmatrix } \includegraphics[width=.8cm]{fig_012/1oo_dd-0 } \end{pmatrix } ; \tag*{\rm(ii$_\delta$ ) } \\ \delta \begin{pmatrix } \includegraphics[width=1.2cm]{fig_3icons/1a2a1b_ddd-0 } \end{pmatrix } - \delta \begin{pmatrix } \includegraphics[width=1.2cm]{fig_3icons/1a2b1b_ddd-0 } \end{pmatrix } = \delta \begin{pmatrix } \includegraphics[width=1.2cm]{fig_3icons/1b2a1a_ddd-0 } \end{pmatrix } - \delta \begin{pmatrix } \includegraphics[width=1.2cm]{fig_3icons/1b2b1a_ddd-0 } \end{pmatrix } ; \tag*{\rm(iii$_\delta$ ) } \\ \delta \begin{pmatrix } \includegraphics[height=1.2cm]{fig_012/phi+-0 } \end{pmatrix } = ( t^{\frac12}-t^{-\frac12})\cdot \delta \begin{pmatrix } \;\ ; \includegraphics[height=1.2cm]{fig_012/i-0 } \;\
; \end{pmatrix } ; \tag*{($\phi_\delta$ ) } \\ \delta \begin{pmatrix } \includegraphics[width=.8cm]{fig_012/o-0 } \end{pmatrix } = 1 .
\tag*{\rm(o$_\delta$)}\end{gathered}\ ] ] if we restrict our attention to oriented links with a fixed number @xmath16 of components , the axiom ( io ) becomes irrelevant but we must pick a suitable normalization .
let @xmath17 denote the @xmath18-component unlink , and let @xmath19 denote the @xmath18-component oriented chain where adjacent rings have linking number @xmath20 .
( in terms of closed braids , @xmath17 is the closure of the trivial braid @xmath21 , and @xmath19 is the closure of the braid @xmath22 . )
we can use either @xmath17 or @xmath19 ( but @xmath17 is preferred ) to normalize the skein or jones polynomial , but for alexander - conway polynomial we can only use @xmath19 .
[ thm : homfly_knot ] the skein polynomial @xmath23 is the invariant of oriented @xmath18-component links determined uniquely by the following three axioms . @xmath9{fig_012/1a1a_dd-0 } \end{pmatrix } + a^2\cdot p \begin{pmatrix } \includegraphics[width=.8cm]{fig_012/1b1b_dd-0 } \end{pmatrix } = ( 2+z^2 ) \cdot p \begin{pmatrix } \includegraphics[width=.8cm]{fig_012/1oo_dd-0 } \end{pmatrix } ; \tag*{\rm(ii ) } \\
a^{-1}\cdot p \begin{pmatrix } \includegraphics[width=1.2cm]{fig_3icons/1a2a1b_ddd-0 } \end{pmatrix } -a\cdot p \begin{pmatrix } \includegraphics[width=1.2cm]{fig_3icons/1a2b1b_ddd-0 } \end{pmatrix } = a^{-1}\cdot p \begin{pmatrix } \includegraphics[width=1.2cm]{fig_3icons/1b2a1a_ddd-0 } \end{pmatrix } -a\cdot p \begin{pmatrix } \includegraphics[width=1.2cm]{fig_3icons/1b2b1a_ddd-0 } \end{pmatrix } ; \tag*{\rm(iii ) } \\[2ex ] p(u_\mu ) = ( z^{-1}(a^{-1}-a))^{\mu-1 } .
\tag*{\rm(u)}\end{gathered}\ ] ] [ thm : jones_knot ] the jones polynomial @xmath24 is the invariant of oriented @xmath18-component links determined uniquely by the following three axioms .
@xmath13{fig_012/1a1a_dd-0 } \end{pmatrix } + t^2\cdot v \begin{pmatrix } \includegraphics[width=.8cm]{fig_012/1b1b_dd-0 } \end{pmatrix } = ( t+t^{-1 } ) \cdot v \begin{pmatrix } \includegraphics[width=.8cm]{fig_012/1oo_dd-0 } \end{pmatrix } ; \tag*{\rm(ii$_v$ ) } \\
t^{-1}\cdot v \begin{pmatrix } \includegraphics[width=1.2cm]{fig_3icons/1a2a1b_ddd-0 } \end{pmatrix } -t\cdot v \begin{pmatrix } \includegraphics[width=1.2cm]{fig_3icons/1a2b1b_ddd-0 } \end{pmatrix } = t^{-1}\cdot v \begin{pmatrix } \includegraphics[width=1.2cm]{fig_3icons/1b2a1a_ddd-0 } \end{pmatrix } -t\cdot v \begin{pmatrix } \includegraphics[width=1.2cm]{fig_3icons/1b2b1a_ddd-0 } \end{pmatrix } ; \tag*{\rm(iii$_v$ ) } \\[2ex ] v(u_\mu ) = ( -(t^{\frac12}+t^{-\frac12}))^{\mu-1 } . \tag*{\rm(u$_v$)}\end{gathered}\ ] ] [ thm : alexander - conway_knot ] the alexander - conway polynomial @xmath25 is the invariant of oriented @xmath18-component links determined uniquely by the following three axioms . @xmath15{fig_012/1a1a_dd-0 } \end{pmatrix } + \delta \begin{pmatrix } \includegraphics[width=.8cm]{fig_012/1b1b_dd-0 } \end{pmatrix } = ( t+t^{-1 } ) \cdot \delta \begin{pmatrix } \includegraphics[width=.8cm]{fig_012/1oo_dd-0 } \end{pmatrix } ; \tag*{\rm(ii$_\delta$ ) } \\ \delta \begin{pmatrix } \includegraphics[width=1.2cm]{fig_3icons/1a2a1b_ddd-0 } \end{pmatrix } - \delta \begin{pmatrix } \includegraphics[width=1.2cm]{fig_3icons/1a2b1b_ddd-0 } \end{pmatrix } = \delta \begin{pmatrix } \includegraphics[width=1.2cm]{fig_3icons/1b2a1a_ddd-0 } \end{pmatrix } - \delta \begin{pmatrix } \includegraphics[width=1.2cm]{fig_3icons/1b2b1a_ddd-0 } \end{pmatrix } ; \tag*{\rm(iii$_\delta$ ) } \\
\delta(c_\mu ) = ( t^{\frac12}-t^{-\frac12})^{\mu-1 } .
\tag*{\rm(c$_\delta$)}\end{gathered}\ ] ] note that the foundational relation ( i ) can not appear in theorems [ thm : homfly_knot][thm : alexander - conway_knot ] because it involves links with different number of components .
our approach is via closed braids .
we explain the language of relators in section [ sec : relators ] and give an algebraic reduction lemma in section [ sec : key_lem ] .
this approach is adapted from the corresponding sections of @xcite on conway s potential function for colored links .
the current context of uncolored links makes the reduction argument more transparent .
section [ sec : stabilizations ] discusses closed braids with different number of strands .
the theorems are proved in the last two sections .
for braids , we use the following conventions : braids are drawn from top to bottom .
the strands of a braid are numbered at the top of the braid , from left to right .
the product @xmath26 of two @xmath27-braids is obtained by drawing @xmath28 below @xmath29 .
the set @xmath30 of all @xmath27-braids forms a group under this multiplication , with standard generators @xmath31 .
it is well known that links can be presented as closed braids .
the closure of a braid @xmath32 will be denoted @xmath33 .
two braids ( possibly with different number of strands ) have isotopic closures if and only if they can be related by a finite sequence of two types of moves : 1 .
conjugacy move : @xmath34 @xmath35 @xmath36 where @xmath37 are conjugate in a braid group @xmath30 ; 2 .
markov move : @xmath32 @xmath35 @xmath38 .
let @xmath39 be the laurent polynomial ring @xmath1 $ ] .
let @xmath40 be the group - algebra on @xmath30 with coefficients in @xmath39 .
[ defn : skein_relation ] we say that an element @xmath41 of @xmath40 is a _ skein relator _ , or equivalently , say that the corresponding formal equation ( in which @xmath42 stands for the @xmath43 of the link @xmath44 ) @xmath45 is a _
skein relation _ , if the following condition is satisfied : for any links @xmath46 that are identical except in a cylinder where they are represented by the braids @xmath47 respectively , the formal equation becomes an equality in @xmath39 .
[ exam : relator_vs_relation ] to every element @xmath48 , by taking braid closures we have a corresponding element @xmath49 the latter vanishes if the former is a skein relator . the skein relations ( i ) , ( ii ) and ( iii ) in section [ sec : intro ]
correspond to the following relators , respectively : ( the symbol @xmath50 stands for the trivial braid . )
@xmath51 & a^{-1}\cdot{\sigma_1\sigma_2\sigma_1^{-1 } } + a\cdot{\sigma_1^{-1}\sigma_2^{-1}\sigma_1 } \\ & - a^{-1}\cdot{\sigma_1^{-1}\sigma_2\sigma_1 } - a\cdot{\sigma_1\sigma_2^{-1}\sigma_1^{-1 } } .
\end{aligned}\end{gathered}\ ] ] [ prop : relator_ideal ] assume that @xmath52 is a skein relation .
then for any given braid @xmath53 , the following equations are also skein relations : @xmath54 hence skein relators form a two - sided ideal @xmath55 ( called the _ relator ideal _ ) in @xmath40 .
look at the cylinder where the links @xmath56 are represented differently by braids @xmath57 , respectively . in the upper half cylinder
they are represented by braids @xmath47 .
so the assumption implies the first equality .
similarly for the second equality . by definition [ defn : skein_relation ] , this means skein relators form a two - sided ideal .
[ defn : equivalence ] let @xmath58 be the two - sided ideal in @xmath40 generated by @xmath59 and @xmath60 .
( when @xmath61 we ignore @xmath60 . )
two elements of the algebra @xmath40 are _ equivalent modulo @xmath58 _ ( denoted by @xmath62 ) if their difference is in @xmath58 . for example , by conjugation in @xmath30 we have @xmath63 and @xmath64 , for any @xmath65 . [
lem : reduction ] modulo @xmath58 , every braid @xmath32 is equivalent to a @xmath39-linear combination of braids of the form @xmath66 with @xmath67 and @xmath68 .
a braid @xmath32 can be written as @xmath69 where @xmath70 and @xmath71 .
we allow that @xmath72 and @xmath73 be trivial , but assume other @xmath74 s are nontrivial .
the number @xmath75 will be denoted as @xmath76 .
the lemma will be proved by an induction on the double index @xmath77 .
note that the lemma is trivial when @xmath61 , or @xmath78 .
it is enough to consider the case @xmath79 , because induction on @xmath75 works beyond @xmath80 . indeed , if @xmath81 , let @xmath82 , then @xmath83 . by inductive hypothesis @xmath36
is equivalent to a linear combination of elements of the form @xmath84 , hence @xmath34 is equivalent to a linear combination of elements of the form @xmath85 .
this brings the problem back to the @xmath79 case .
henceforth we assume @xmath79 .
since the initial and terminal part of @xmath34 , namely @xmath72 and @xmath73 , do not affect the conclusion of the lemma , we can drop them .
so we assume @xmath86 , where @xmath87 . by the induction hypothesis on @xmath27 ,
@xmath87 is a linear combination of elements of the form @xmath88 .
note that @xmath89 commute with @xmath90 .
so it suffices to focus on braids of the form @xmath91 . for the sole purpose of controlling the length of displayed formulas ,
we assume @xmath92 below . the proof for a general @xmath27 can be obtained by a simple change of subscripts , replacing @xmath93 with @xmath94 and replacing @xmath95 with @xmath96 , respectively .
thus , lemma [ lem : reduction ] has been reduced to the following [ lem : key_reduction ] every @xmath97 is equivalent to a linear combination of braids of the form @xmath98 where @xmath99 is @xmath100 , @xmath101 or @xmath80 .
modulo @xmath102 , we may restrict the exponent @xmath103 to take values @xmath104
, @xmath80 and @xmath105 ( we are done if @xmath103 is @xmath100 ) .
if @xmath106 we can decrease @xmath103 by looking at @xmath107 , so it suffices to prove the case @xmath108 .
again modulo @xmath102 , we can restrict the exponents @xmath109 to the values @xmath101 and @xmath80 .
there are altogether 9 cases to verify .
_ 5 trivial cases ( braid identities ) _ : @xmath110 _ the case @xmath111 _ : multiplying @xmath60 by @xmath112 on the right and @xmath113 on the left , and taking braid identities into account , we get the relation @xmath114 then @xmath111 is equivalent to a linear combination of braids of the form @xmath115 . so the case @xmath111 is verified .
_ the case @xmath116 _ : multiplying the previous relation by @xmath112 on the right , and taking braid identities into account , we see that @xmath117 similar to the above case , this reduces @xmath116 to the verified case @xmath111 .
_ the case @xmath118 _ : multiplying @xmath60 on the right by @xmath119 , we get @xmath120 this verifies the case @xmath118 . _
the case @xmath121 _ : multiplying @xmath122 by @xmath112 on the right , we get @xmath123 the case @xmath121 is also verified .
we have verified all 9 cases .
modulo @xmath59 we can assume @xmath124 . thus lemma [ lem : key_reduction ] is proved .
the inductive proof of lemma [ lem : reduction ] is now complete .
the resulting @xmath39-linear combination of braids of the form @xmath66 with @xmath67 in the lemma is not unique , but the inductive proof gives us a recursive algorithm to find one . to compare the ideal @xmath58 with the relator ideal @xmath55 of section [ sec : relators ] , we have @xmath125 but @xmath126 .
the inclusion is easy .
indeed , @xmath127 is in the relator ideal @xmath55 , and @xmath128 so both @xmath59 and @xmath129 are in @xmath55
. therefore @xmath125 . to show they are not equal , we need the notion of homogeneity .
each @xmath27-braid @xmath34 has an _ underlying permutation _ of @xmath130 , denoted @xmath131 , where @xmath132 is the position of the @xmath65-th strand at the bottom of @xmath34 .
in this way the braid group @xmath30 projects onto the symmetric group @xmath133 .
an element of @xmath40 is called _ homogeneous _ if all its terms ( with nonzero coefficients ) have the same underlying permutation . as a @xmath39-module ,
@xmath40 splits into a direct sum according to underlying permutations of braids . under this
splitting , every element of @xmath40 decomposes into a sum of its _ homogeneous components_. since @xmath59 and @xmath129 are homogeneous , the ideal @xmath134 is generated by homogeneous elements .
then every homogeneous component of any element of @xmath58 is also in @xmath58 . now
the relator @xmath135 has a homogeneous component @xmath136 which is not a relator .
hence @xmath137 is not in @xmath58 .
thus @xmath58 is strictly smaller than @xmath55 .
suppose a braid @xmath32 is written as a word in the standard generators @xmath138 , @xmath112 , , @xmath90 . the same word @xmath34 gives a braid in @xmath139 for any @xmath140 . thus @xmath30 is standardly embedded in @xmath139 .
however , when talking about a closed braid @xmath33 , the number of strands in @xmath34 does matter .
we shall use the notation @xmath141_n$ ] to emphasize that @xmath34 is regarded as an @xmath27-braid , and use @xmath141_n^{\;{\widehat}{}}$ ] for its closure .
for example , @xmath141_{n+1}^{\;{\widehat}{}}$ ] adds a free circle to @xmath141_n^{\;{\widehat}{}}$ ] . the markov move says @xmath142_{n+1}^{\;{\widehat}{}}$ ] is isotopic to @xmath141_n^{\;{\widehat}{}}$ ] . for a braid @xmath32 and an integer @xmath140 ,
we shall use @xmath143 to denote the @xmath103-th shifted version of @xmath34 , i.e. , the braid obtained from the word @xmath34 by replacing each generator @xmath144 with @xmath145 .
its closure @xmath146_{n+k}^{\;{\widehat}{}}$ ] is isotopic to @xmath141_{n+k}^{\;{\widehat}{}}$ ] .
suppose @xmath147 and @xmath148 .
observe from the diagram defining braid closure that the closed braid @xmath149_{n+p}^{\;{\widehat}{}}$ ] is isotopic to @xmath150_{n+p-1}^{\;{\widehat}{}}$ ] ( which is in fact a connected sum of oriented links @xmath151_n^{\;{\widehat}{}}$ ] and @xmath152_p^{\;{\widehat}{}}$ ] ) . by an abuse of language , we will call this a _
markov move_. if @xmath36 brings the @xmath27-th position at its top to the same position at its bottom , then @xmath150_{n+p-1}^{\;{\widehat}{}}$ ] is isotopic to @xmath153_{n+p-1}^{\;{\widehat}{}}$ ]
. we will refer to it as a _ slide _ move ( in the connected sum , sliding @xmath152_p^{\;{\widehat}{}}$ ] down the last strand of @xmath36 ) .
[ lem : stabilization3 ] assume that @xmath8 $ ] is an invariant of oriented links that satisfies skein relations and .
then for @xmath32 and @xmath148 we have @xmath154_{n+p}^{\;{\widehat } { } } \right ) = ( a^{-2}-1)\cdot p\left ( [ \beta\sigma_{n}^2\gamma^{\triangleright n}]_{n+p}^{\;{\widehat } { } } \right ) .\ ] ] the braid form of axioms ( ii ) and ( iii ) are the relators ( ii@xmath155 ) and ( iii@xmath155 ) , respectively . multiplying ( iii@xmath155 ) by @xmath156 on the right
we get another relator @xmath157 it gives us an equality between the @xmath43 s of closed @xmath158-braids : @xmath159_{n+p+1}^{\;{\widehat } { } } \right ) + a\cdot p\left ( [ \beta(\sigma_{n+1}\sigma_{n}^{-2})\gamma^{\triangleright ( n+1)}]_{n+p+1}^{\;{\widehat } { } } \right ) \\ & - a^{-1}\cdot p\left ( [ \beta\sigma_{n+1}\gamma^{\triangleright ( n+1)}]_{n+p+1}^{\;{\widehat } { } } \right ) - a\cdot p\left ( [ \beta(\sigma_{n}^2\sigma_{n+1}^{-1}\sigma_{n}^{-2})\gamma^{\triangleright ( n+1)}]_{n+p+1}^{\;{\widehat } { } } \right ) = 0 .\end{aligned}\ ] ] these closed braids can be simplified via isotopy moves ( c = conjugacy , m = markov and s = slide ) : @xmath160_{n+p+1}^{\;{\widehat } { } } & \overset{\text{c}}\rightsquigarrow [ \beta\sigma_{n}^2\sigma_{n+1}\gamma^{\triangleright ( n+1)}\sigma_{n+1}^{-2}]_{n+p+1}^{\;{\widehat } { } } \\ & \overset{\text{s}}{\rightsquigarrow } [ \beta\sigma_{n}^2\sigma_{n+1}^{-1}\gamma^{\triangleright ( n+1)}]_{n+p+1}^{\;{\widehat } { } } \overset{\text{m}}{\rightsquigarrow } [ \beta\sigma_{n}^2\gamma^{\triangleright n}]_{n+p}^{\;{\widehat } { } } ; \\ [ \beta\sigma_{n+1}\sigma_{n}^{-2}\gamma^{\triangleright ( n+1)}]_{n+p+1}^{\;{\widehat } { } } & \overset{\text{m}}{\rightsquigarrow } [ \beta\gamma^{\triangleright n}\sigma_{n}^{-2}]_{n+p}^{\;{\widehat } { } } \overset{\text{s}}\rightsquigarrow [ \beta\sigma_{n}^{-2}\gamma^{\triangleright n}]_{n+p}^{\;{\widehat } { } } ; \\ [ \beta\sigma_{n+1}\gamma^{\triangleright ( n+1)}]_{n+p+1}^{\;{\widehat } { } } & \overset{\text{m}}{\rightsquigarrow } [ \beta\gamma^{\triangleright n}]_{n+p}^{\;{\widehat } { } } ; \\ [ \beta\sigma_{n}^2\sigma_{n+1}^{-1}\sigma_{n}^{-2}\gamma^{\triangleright ( n+1)}]_{n+p+1}^{\;{\widehat } { } } & \overset{\text{m}}{\rightsquigarrow } [ \beta\sigma_{n}^2\gamma^{\triangleright n}\sigma_{n}^{-2}]_{n+p}^{\;{\widehat } { } } \overset{\text{s}}{\rightsquigarrow } [ \beta\gamma^{\triangleright n}]_{n+p}^{\;{\widehat } { } } .\end{aligned}\ ] ] since @xmath23 is isotopy invariant , the above equality becomes @xmath161_{n+p}^{\;{\widehat } { } } \right ) + a\cdot p\left ( [ \beta\sigma_{n}^{-2}\gamma^{\triangleright n}]_{n+p}^{\;{\widehat } { } } \right ) = ( a^{-1}+a)\cdot p\left ( [ \beta\gamma^{\triangleright n}]_{n+p}]_{n+1}^{\;{\widehat } { } } \right ) .
\\ \intertext{comparing it with the equality ( from ( ii$_{\text{b}}$ ) ) } a^{-2}\cdot p\left ( [ \beta\sigma_{n}^2\gamma^{\triangleright n}]_{n+p}^{\;{\widehat } { } } \right ) + a^2\cdot p\left ( [ \beta\sigma_{n}^{-2}\gamma^{\triangleright n}]_{n+p}^{\;{\widehat } { } } \right ) = ( 2+z^2)\cdot p\left ( [ \beta\gamma^{\triangleright n}]_{n+p}^{\;{\widehat } { } } \right ) , \end{gathered}\ ] ] we get the desired conclusion . [ cor : stabilization ] under the assumption of the above lemma , the following two relations are equivalent to each other : @xmath162{fig_012/io-0 } \end{pmatrix } = z^{-1}(a^{-1}-a)\cdot p \begin{pmatrix } \;\ ; \includegraphics[height=1.2cm]{fig_012/i-0 } \;\ ; \end{pmatrix } ; \tag*{\rm(io ) } \\ p \begin{pmatrix } \includegraphics[height=1.2cm]{fig_012/phi+-0 } \end{pmatrix } = az^{-1}(1+z^2-a^2)\cdot p \begin{pmatrix } \;\ ; \includegraphics[height=1.2cm]{fig_012/i-0 } \;\ ; \end{pmatrix } .
\tag*{($\phi$)}\end{gathered}\ ] ] the braid form of these two relations are , respectively , @xmath163_{n+1}^{\;{\widehat } { } } \right ) & = z^{-1}(a^{-1}-a)\cdot p\left ( [ \beta]_{n}^{\;{\widehat } { } } \right ) & & \quad \text{for any braid } \beta\in b_n ; \tag*{(io$_{\text{b}}$ ) } \\ p\left ( [ \beta\sigma_{n}^2]_{n+1}^{\;{\widehat } { } } \right ) & = az^{-1}(1+z^2-a^2)\cdot p\left ( [ \beta]_{n}^{\;{\widehat } { } } \right ) & & \quad \text{for any braid } \beta\in b_n . \tag*{($\phi_{\text{b}}$)}\end{aligned}\ ] ] they are equivalent to each other by the above lemma with @xmath152_p:=[e]_1 $ ] .
there is a parallel statement for jones polynomial : [ cor : stabilization_jones ] assume that @xmath6 $ ] is an invariant of oriented links that satisfies skein relations and .
then the following two relations are equivalent to each other : @xmath164{fig_012/io-0 } \end{pmatrix } = -(t^{\frac12}+t^{-\frac12})\cdot v \begin{pmatrix } \;\ ; \includegraphics[height=1.2cm]{fig_012/i-0 } \;\ ; \end{pmatrix } ; \tag*{\rm(io$_v$ ) } \\ v \begin{pmatrix } \includegraphics[height=1.2cm]{fig_012/phi+-0 } \end{pmatrix } = -t^{\frac32}(t+t^{-1})\cdot v \begin{pmatrix } \;\ ; \includegraphics[height=1.2cm]{fig_012/i-0 } \;\ ; \end{pmatrix } .
\tag*{($\phi_v$)}\end{gathered}\ ] ] for the alexander - conway polynomial , we have : [ cor : stabilization_alexander - conway ] assume that @xmath5 $ ] is an invariant of oriented links that satisfies skein relations and .
then @xmath165 for any split link @xmath166 .
in particular , the following relation holds true : @xmath167{fig_012/io-0 } \end{pmatrix } = 0 .
\tag*{\rm(io$_\delta$)}\ ] ] for links @xmath168_n^{\;{\widehat}{}}$ ] and @xmath169_p^{\;{\widehat}{}}$ ] , the split link @xmath170_{n+p}^{\;{\widehat}{}}$ ] . then apply lemma [ lem : stabilization3 ] with substitutions @xmath171 and @xmath11 .
we shall focus on theorem [ thm : homfly ] , then remark on the other two .
let us forget about the original definition of the skein polynomial , and regard the symbol @xmath23 as a well - defined invariant of oriented links which satisfies the axioms ( ii ) , ( iii ) , ( io ) and ( o ) . by corollary [ cor : stabilization ]
, @xmath23 also satisfies axiom ( @xmath14 ) .
we shall show that such an invariant @xmath23 is computable , hence uniquely determined .
suppose @xmath34 is an @xmath27-braid .
by lemma [ lem : reduction ] , the braid @xmath32 is equivalent to ( in a computable way ) a @xmath39-linear combination of braids of the form @xmath66 with @xmath67 and @xmath68 . by example [ exam : relator_vs_relation ] the ( mod @xmath58 ) equivalence preserves the @xmath43 of closure of braids .
so @xmath172_n^{\;{\widehat } { } } \right)$ ] is a @xmath39-linear combination ( with computable coefficients ) of @xmath176_n^{\;{\widehat } { } } \right)$ ] s . for @xmath177 ,
respectively , we have @xmath178_n^{\;{\widehat } { } } \right ) & = p \left ( [ \alpha \gamma]_{n-1}^{\;{\widehat } { } } \right ) & & \qquad\text{by isotopy } , \\
p \left ( [ \alpha\sigma_{n-1}^0 \gamma]_n^{\;{\widehat } { } } \right ) & = z^{-1}(a^{-1}-a ) \cdot p \left ( [ \alpha \gamma]_{n-1}^{\;{\widehat } { } } \right ) & & \qquad\text{by ( io ) } , \\
p \left ( [ \alpha\sigma_{n-1}^2 \gamma]_n^{\;{\widehat } { } } \right ) & = az^{-1}(1+z^2-a^2 ) \cdot p \left ( [ \alpha \gamma]_{n-1}^{\;{\widehat } { } } \right ) & & \qquad\text{by $ ( \phi)$}.\end{aligned}\ ] ] since @xmath179_{n-1}^{\;{\widehat } { } } \right)$ ] is computable by the inductive hypothesis claim(@xmath175 ) , we see @xmath176_n^{\;{\widehat } { } } \right)$ ] is also computable .
thus claim(@xmath27 ) is proved .
a remarkable feature of this algorithm is that it never increases the number of components of links .
in fact , all the reductions in section [ sec : key_lem ] are by axioms ( ii ) and ( iii ) which respect the components , while in this section , components could get removed but never added , by axioms ( io ) and ( @xmath14 ) .
so if we start off with a knot , we shall always get knots along the way , the axioms ( io ) and ( @xmath14 ) becoming irrelevant .
this observation works even for links with any given number of components , once we set up a suitable normalization .
hence the theorem [ thm : homfly_knot ] .
the case of alexander - conway polynomial is only slightly different .
corollary [ cor : stabilization_alexander - conway ] says ( io@xmath180 ) is a consequence of axioms ( ii@xmath180 ) and ( iii@xmath180 ) , and ( @xmath181 ) is taken as an axiom .
so the proof above also works through with the substitutions @xmath171 and @xmath11 .
actually , the argument in section [ sec : stabilizations ] can be adapted to work for conway potential function of colored links , to the effect that in ( * ? ? ?
* main theorem ) , the relation ( io ) is a consequence of axioms ( ii ) and ( iii ) hence can be removed from the list of axioms .
suppose @xmath18 is a given positive integer .
regard @xmath23 as a well - defined invariant of oriented @xmath18-component links which satisfies the axioms ( ii ) , ( iii ) and ( u ) .
we shall temporarily expand the ring @xmath182 $ ] , where the invariant @xmath23 takes value , to @xmath183 $ ] , to allow fractions with denominator a power of @xmath184 .
we shall show that such an invariant @xmath23 is computable , hence uniquely determined .
it is the normalization ( u ) that brings the value @xmath23 back into the original @xmath39 .
[ lem : reduction_homfly ] suppose @xmath32 , @xmath185 , and @xmath141_{n+p}^{\;{\widehat}{}}$ ] has @xmath18 components . if @xmath186 , then @xmath172_{n+p}^{\;{\widehat } { } } \right)$ ] is computable as a @xmath187-linear combination of terms of the form @xmath188_{n'+p'}^{\;{\widehat } { } } \right)$
] , each with @xmath18 components , @xmath189 , @xmath190 , @xmath191 , and the @xmath192-exponent of the corresponding coefficient is at most @xmath193 . by lemma [ lem : reduction ] ,
the braid @xmath32 is equivalent to ( in a computable way ) a @xmath39-linear combination of braids of the form @xmath66 with @xmath67 and @xmath68 .
so @xmath172_{n+p}^{\;{\widehat } { } } \right)$ ] is a @xmath187-linear combination ( with computable coefficients ) of @xmath176_{n+p}^{\;{\widehat } { } } \right)$ ] s . for @xmath177 ,
respectively , we have @xmath178_{n+p}^{\;{\widehat } { } } \right ) & = p \left ( [ \gamma\alpha \sigma_{n-1}^{\pm1}]_{n+p}^{\;{\widehat } { } } \right ) & & \quad\text{by braid conjugation } , \\ & = p \left ( [ \gamma\alpha ] _ { ( n-1)+p}^{\;{\widehat } { } } \right ) & & \quad\text{by markov move } , \\
p \left ( [ \alpha\sigma_{n-1}^0 \gamma]_{n+p}^{\;{\widehat } { } } \right ) & = p \left ( [ \alpha \gamma]_{(n-1)+(p+1)}^{\;{\widehat } { } } \right ) & & \quad\text{obvious } , \\
( [ \alpha\sigma_{n-1}^2 \gamma]_{n+p}^{\;{\widehat } { } } \right ) & = p \left ( [ \gamma\alpha\sigma_{n-1}^2]_{n+p}^{\;{\widehat } { } } \right ) & & \quad\text{by braid conjugation } \\ &
= \frac{1+z^2-a^2}{a^{-2}-1 } \cdot p \left ( [ \gamma \alpha]_{(n-1)+(p+1)}^{\;{\widehat } { } } \right ) & & \quad\text{by lemma~\ref{lem : stabilization3}}.\end{aligned}\ ] ] the @xmath43 s on the right hand sides satisfy the required conditions .
suppose a link @xmath166 with @xmath18 components is presented as @xmath141_{n}^{\;{\widehat}{}}=[\beta]_{n+0}^{\;{\widehat}{}}$ ] .
apply lemma [ lem : reduction_homfly ] repeatedly until no such reduction is possible .
then @xmath172_{n+0}^{\;{\widehat } { } } \right)$ ] is computed as a @xmath187-linear combination of terms @xmath188_{n'+p'}^{\;{\widehat } { } } \right)$ ] , each with @xmath194 hence @xmath195_1 $ ] , such that * every @xmath196_{n'+p'}^{\;{\widehat}{}}=[e]_{1+p'}^{\;{\widehat}{}}$ ] has @xmath18 components , hence @xmath197 , and @xmath196_{n'+p'}^{\;{\widehat } { } } = [ e]_\mu^{\;{\widehat } { } } = u_\mu$ ] ; and * the @xmath192-exponent of every coefficient is at most @xmath198 . therefore , @xmath199 is computable and , by axiom ( u ) , every term @xmath188_{n'+p'}^{\;{\widehat } { } } \right ) $ ] has a factor @xmath200 that can cancel the @xmath192-exponent in its coefficient , so @xmath201 .
theorem [ thm : jones_knot ] can be proved similarly , but theorem [ thm : alexander - conway_knot ] needs modifications .
define @xmath202 whose closure @xmath203_p^{\;{\widehat}{}}$ ] is the oriented @xmath204-component chain @xmath205 .
[ lem : reduction_alexander - conway ] suppose @xmath206 , @xmath207 , and @xmath208_{n+p}^{\;{\widehat}{}}$ ] has @xmath18 components .
if @xmath209 , then @xmath210_{n+p}^{\;{\widehat } { } } \right)$ ] is computable as a @xmath211$]-linear combination of terms of the form @xmath212_{n'+p'}^{\;{\widehat } { } } \right)$ ] , each with @xmath18 components , @xmath213 , @xmath190 and @xmath191 . by lemma [ lem : reduction ] ( with substitutions @xmath171 and @xmath11 ) , the braid @xmath206 is equivalent to ( in a computable way ) a @xmath211$]-linear combination of braids of the form @xmath214 with @xmath215 and @xmath68 .
multiplication by @xmath216 makes the braid @xmath217 equivalent to a linear combination of braids of the form @xmath218 , so @xmath210_{n+p}^{\;{\widehat } { } } \right)$ ] is computed as a linear combination of the @xmath219_{n+p}^{\;{\widehat } { } } \right)$ ] s . for @xmath177 ,
respectively , we have @xmath220_{n+p}^{\;{\widehat } { } } \right ) & = \delta \left ( [ \gamma\alpha\sigma_{n}^{\pm1 } \delta_p^{\triangleright n}]_{n+p}^{\;{\widehat } { } } \right ) & & \quad\text{by braid conjugacy } , \\ & = \delta \left ( [ \gamma\alpha \delta_p^{\triangleright ( n-1)}]_{n+p-1}^{\;{\widehat } { } } \right ) & & \quad\text{by a markov move } , \\ \delta \left ( [ \alpha\sigma_{n}^0 \gamma\delta_p^{\triangleright n}]_{n+p}^{\;{\widehat } { } } \right ) & = 0 & & \quad\text{by corollary~\ref{cor : stabilization_alexander - conway } } , \\ \delta \left ( [ \alpha\sigma_{n}^2 \gamma\delta_p^{\triangleright n}]_{n+p}^{\;{\widehat } { } } \right ) & = \delta \left ( [ \gamma\alpha\sigma_{n}^2 \delta_p^{\triangleright n}]_{n+p}^{\;{\widehat } { } } \right ) & & \quad\text{by braid conjugacy } \\ & = \delta \left ( [ \gamma\alpha \delta_{p+1}^{\triangleright ( n-1)}]_{n+p}^{\;{\widehat } { } } \right ) & & \quad\text{by definition}.\end{aligned}\ ] ] the @xmath221 s on the right hand sides are in the desired form with @xmath222 and with @xmath18 components .
suppose a link @xmath166 with @xmath18 components is presented as @xmath141_{n+1}^{\;{\widehat}{}}=[\beta\delta_1^{\triangleright n}]_{n+1}^{\;{\widehat}{}}$ ] .
apply lemma [ lem : reduction_alexander - conway ] repeatedly until no such reduction is possible .
then @xmath223_{n+1}^{\;{\widehat } { } } \right)$ ] is computed as a @xmath211$]-linear combination of terms @xmath212_{n'+p'}^{\;{\widehat } { } } \right)$ ] , with @xmath224 .
hence each @xmath195_1 $ ] , @xmath225 the number of components , and each @xmath226_{n'+p'}^{\;{\widehat } { } } = [ \delta_\mu]_\mu^{\;{\widehat } { } } = c_\mu$ ] .
therefore @xmath227 is computable and moreover , by axiom ( c@xmath180 ) , divisible by @xmath228 .
8 j w alexander , topological invariants of knots and links , trans .
soc . , 30 ( 1928 )
no.2 , 275306 .
j h conway , an enumeration of knots and links , and some of their algebraic properties , pp . 329358 in book computational problems in abstract algebra " , pergamon press , new york , 1970 .
p freyd , d yetter ; j hoste ; w b r lickorish , k millett ; a ocneanu . a new polynomial invariant of knots and links , bull .
, 12 ( 1985 ) no.2 , 239 - 246 .
richard hartley , the conway potential function for links , comment .
helv . , 58 ( 1983 )
no.3 , 365378 .
boju jiang , on conway s potential function for colored links , acta math .
sinica , english series , 32 ( 2016 ) no.1 , 2539 .
v f r jones , a polynomial invariant for knots via von neumann algebras , bull .
, 12 ( 1985 ) no.1 , 103111 .
akio kawauchi , a survey of knot theory " , birkhuser , basel , 1996 .
j h przytycki , p traczyk , invariants of links of conway type , kobe j. math . , 4 ( 1987 ) , 115139 . | we give characterizations of the skein polynomial for links ( as well as jones and alexander - conway polynomials derivable from it ) , avoiding the usual `` smoothing of a crossing '' move . as by - products
we have characterizations of these polynomials for knots , and for links with any given number of components . |
ngc 4527 is a nearby spiral galaxy ( table 1 ) , and has been classified as a starburst galaxy because of it s abundant molecular gas ( @xmath3 ) and high _ star formation rate _ ( young & devereux 1991 ) .
however , young & devereux ( 1991 ) pointed out that the star formation rate per unit gas mass indicated by @xmath4 , the _ star formation efficiency _ , of ngc 4527 is very low compared with their starburst sample .
what controls the star formation properties in the central region of ngc 4527 ? in order to address the relationship between the physical conditions of molecular gas and star formation , we have performed high resolution multiple molecular emission observations in the central a few kpc region of ngc 4527 with the nobeyama millimeter array ( nma ) .
.properties of ngc 4527 .
reference ( 1)(2)hummel et al . 1987 ; ( 3 ) rubin et al . 1997 ; ( 4 ) tully 1988 ; ( 5)(6 ) young & devereux 1991 ; [ cols="<,<,<",options="header " , ]
the co(1 - 0 ) map in figure 1 shows two outstanding features .
one is the strong concentration of co emission toward the nucleus ( r @xmath5 5@xmath6 or 350 pc ) with a slight elongation along the major axis of the galaxy ( p.a .
= 64@xmath7 ) .
another feature is two offset ridges of co emission extends over r @xmath5 27@xmath6 ( 1800 pc ) along p.a . of 64@xmath7 .
the overall distribution of the co in ngc 4527 is very similar to that of barred spiral galaxies , which shows two offset ridges at the leading edges of a bar with a central concentration of gas ( e.g. , kenney et al . 1992 ) , although the optical morphology can not make clear whether the presence of the bar ( table 1 ) .
our co(1 - 0 ) map recovers about 80 % of a single dish flux ( young et al . 1995 ) .
molecular gas mass estimated from the observed co flux is about @xmath8 @xmath9 , and the peak molecular gas surface density , @xmath10 , at the center is @xmath11 @xmath9 pc@xmath12 adopting a galactic co - to - h@xmath13 conversion factor ( @xmath14@xmath15 ) .
this velocity width ( full width of zero intensity ) of 500 km s@xmath16 is the same as that of the single - dish line profiles obtained with the fcrao 14 m ( young et al . 1995 ) and nrao 12 m ( helfer & blitz 1993 ) .
the distribution of the co(2 - 1 ) emission resembles the co(1 - 0 ) map , although the extended ridges seen in co(1 - 0 ) map ( r @xmath5 27@xmath6 or 1800 pc ) do not appear in this co(2 - 1 ) map .
it is probably due to insufficient sensitivity of co(2 - 1 ) observations , and also the lack of short spacing .
the summary of co observations are listed in table 2 .
hcn(1 - 0 ) and hco@xmath0(1 - 0 ) emission concentrated toward the center have been detected .
extended components correspond to the co ridges may be also seen in the maps , although the s / n ratios are insufficient to make detail comparisons .
we assumed almost all the single dish flux is recovered in both hcn and hco@xmath0 maps , and the previous observations of hcn emission seems to support the validity of this assumption ( see helfer & blitz 1997 and kohno et al .
we summarize hcn and hco@xmath0 observations in table 2 .
the co(2 - 1 ) to co(1 - 0 ) integrated intensity ratio , @xmath17 , is a measure of dense gas fraction to the total molecular gas ( e.g. , sakamoto et al .
1994 ) , and sensitive in a density range of about 10@xmath18 - 10@xmath19 @xmath20 .
we convolved the co(1 - 0 ) map to the same beam size as that of the co(2 - 1 ) to calculate @xmath21 values .
the peak @xmath17 of about 0.6 at the center of ngc 4527 is somewhat smaller than @xmath17 values(@xmath22 ) in the central regions of spiral galaxies ( braine & combes 1992 ) , suggesting a presence of a large amount of low density molecular gas in this region .
note the observed @xmath17 could contain considerable error due to lack of single dish co(2 - 1 ) measurements , however .
a radial decrease of @xmath17 may be apparent ; the ratio is about 0.6 at the center , while the ratio decreases to 0.5 at r = 6@xmath6 ( 400 pc ) .
the hcn(1 - 0 ) to co(1 - 0 ) integrated intensity ratio , @xmath23 , is another measure of dense gas fraction to the total molecular gas ( e.g. , kohno et al .
1999 ) , and sensitive in a density range of about 10@xmath24 - 10@xmath25 @xmath20 .
the observed peak @xmath23 of 0.06 is significantly smaller than a @xmath23 values in starburst galaxies such as ngc 253 ( 0.2 @xmath26 ; e.g. , paglione et al . 1997 ; sorai 1997 ) and m82 ( @xmath27 ; shen & lo 1995 ) , and is similar to those in `` normal '' galaxies such as the galactic center ( @xmath28 ; paglione et al .
1998 ) . because stars are formed from dense cores of molecular clouds rather than their diffuse envelopes(e.g .
lada 1992 ) , study of dense molecular gas is essential to understand star formation in galaxies . in fact , good spatial coincidence between hcn and h@xmath29 emission has been reported in nearby starburst galaxies ( e.g. paglione et al .
1997 ; kohno et al 1999 ) .
both @xmath17 and @xmath23 suggest that the fraction of dense molecular clouds to the total molecular gas in the center of ngc 4527 is smaller than that of prototypical starburst galaxies such as ngc 253 and m82 .
it has been demonstrated that the fraction of dense molecular gas measured with the @xmath23 correlates with the star formation efficiency ( solomon et al .
1992 ) , which is indicated by @xmath30 . considering these results , we suggest that the small dense gas fraction is responsible for the low star formation efficiency in the center of ngc 4527 .
but the star formation rate of ngc 4527 would be enhanced existence of rich gas in central region .
aalto et al . 1995 , a&a , 300 , 369 | we have performed high resolution co(2 - 1 ) , co(1 - 0 ) , hcn(1 - 0 ) and hco@xmath0(1 - 0 ) observations of a `` low star formation efficiency starburst galaxy '' ngc 4527 with the nobeyama millimeter array .
the integrated intensity ratios , co(2 - 1)/co(1 - 0 ) and hcn(1 - 0)/co(1 - 0 ) , are found to be 0.6@xmath1 and 0.06@xmath2 , respectively , at the center .
these line ratios are smaller than those in prototypical starburst galaxies such as ngc 253 and m82 , and we suggest that the fraction of dense molecular gas to the total molecular mass in the central a few kpc region of ngc 4527 is small
. this fact may be responsible for the low star formation efficiency in the center of ngc 4527 .
-30 mm -14 mm ts [email protected] |
in this paper we investigate the @xmath0 regularity of alexandrov solutions to degenerate monge - ampre equations of the form @xmath2 where @xmath3 is a bounded convex domain in @xmath4 . in the case that @xmath5 also has a positive lower bound
, @xmath0 estimates were first obtained by de philippis and figalli ( @xcite ) .
they showed that @xmath6 is integrable for any @xmath7 .
it was subsequently shown that @xmath8 is in fact @xmath9 for some @xmath10 depending on dimension and @xmath11 ( see @xcite , @xcite ) .
these estimates are optimal in light of two - dimensional examples due to wang ( @xcite ) with the homogeneity @xmath12 these estimates fail when @xmath5 degenerates . in three and higher dimensions , it is not hard to construct solutions to ( @xmath13 ) that have a lipschitz singularity on part of a hyperplane , so the second derivatives concentrate ( see section [ highdim ] )
. however , in two dimensions , a classical result of alexandrov ( @xcite , see also @xcite ) shows that lipschitz singularities of solutions to @xmath14 propagate to the boundary .
thus , in two dimensions solutions to ( @xmath13 ) are @xmath15 and @xmath8 has no jump part . however , this leaves open the possibility that @xmath8 has nonzero cantor part .
the main result of this paper is the construction of a solution to ( @xmath13 ) in two dimensions that is not @xmath0 .
this negatively answers an open problem stated in both ( @xcite ) and ( @xcite ) , which was motivated by potential applications to the semigeostrophic equation .
we also prove that singularities that are logarithmically slower than lipschitz propagate , which generalizes the theorem of alexandrov and is optimal by example .
the @xmath0 estimates mentioned above have applications to the global existence of weak solutions to the semigeostrophic equation ( @xcite , @xcite ) . in this context
, the density @xmath5 solves a continuity equation that preserves @xmath16 bounds .
this is the only regularity property of @xmath5 that is globally preserved , due to nonlinear coupling between @xmath5 and the velocity field .
it is therefore useful to obtain estimates that depend on @xmath16 bounds for @xmath5 but not on its regularity .
to apply the results in @xcite and @xcite one must assume that @xmath5 is supported in the whole space .
however , in physically interesting cases , the initial density is compactly supported .
it is thus natural to ask what one can show about solutions to ( @xmath13 ) .
our construction shows that , even in two dimensions , one must rely more on the specific structure of the semigeostrophic equation to obtain existence results for compactly supported initial data .
the idea of our construction is to start with a one - dimensional convex function of @xmath17 in the half - space @xmath18 whose second derivative has nontrivial cantor part , and extend to a convex function on @xmath19 which lifts from these values without generating too much monge - ampre measure . to accomplish this we start with a `` building block '' @xmath20 that agrees with @xmath21 in @xmath22 for some @xmath23 , and in the cusp @xmath24 grows with the homogeneity @xmath25 by superposing vertically - translated rescalings of ( a smoothed version of ) @xmath20 in a self - similar way
, we obtain our example .
our main theorem is : [ main ] for all @xmath26 , there exist solutions to @xmath27 that are not @xmath0 .
it is obvious that solutions to ( @xmath13 ) in one dimension are @xmath28 . in our examples ,
the support of @xmath5 is irregular . in particular , in the higher - dimensional examples ,
the support of @xmath5 is a cusp revolved around an axis , and in the two - dimensional example , the support of @xmath5 has a very irregular `` fractal '' geometry . in e.g. @xcite and @xcite
the authors obtain interesting regularity results when @xmath5 degenerates in a specific way , motivated by applications to prescribed gauss curvature . our second result concerns the behavior of solutions to @xmath27 near a single line segment .
since lipschitz singularities propagate , @xmath8 can not concentrate on a line segment .
( in our two - dimensional counterexample to @xmath0 regularity , @xmath8 concentrates on a family of horizontal rays . ) on the other hand , by modifying an example in @xcite one can construct , for any @xmath29 , a solution to @xmath27 that grows like @xmath30 , with second derivatives not in @xmath31 ( see section [ propagation ] ) .
it is natural to ask whether one can take @xmath32 .
we show that this is not possible .
indeed , we construct a family of barriers that agree with @xmath33 away from arbitrarily thin cusps around the @xmath34 axis , where we can make the monge - ampre measure as large as we like . by sliding these barriers
we prove that singularities of the form @xmath33 propagate .
our second theorem is : [ logarithmicpropagation ] assume that @xmath35 is convex on @xmath19 and that @xmath14 .
then if @xmath36 and @xmath37 in a neighborhood of the origin for some @xmath38 , then @xmath35 vanishes on the @xmath34 axis .
note that we assume the growth in a neighborhood of @xmath39 . for a lipschitz singularity
it is enough to assume the growth at a point , which automatically extends to a neighborhood by convexity .
( see e.g. @xcite for a short proof that lipschitz singularities propagate . )
theorem [ logarithmicpropagation ] shows , after taking a legendre transform , that a solution to @xmath40 in two dimensions can not separate from a tangent plane more slowly than @xmath41 in any fixed direction .
this quantifies the classical result that such functions are strictly convex .
the paper is organized as follows . in section [ highdim ]
we construct simple examples of solutions to ( @xmath13 ) in the case @xmath42 which have a lipschitz singularity on a hyperplane . in section [ mainexample ]
we construct a solution to ( @xmath13 ) in two dimensions whose second derivatives have nontrivial cantor part .
this proves theorem [ main ] . in section [ propagation ]
we first construct examples showing that theorem [ logarithmicpropagation ] is optimal .
we then construct barriers related to these examples .
finally , we use the barriers to prove theorem [ logarithmicpropagation ] .
in this section we construct simple examples of solutions to ( @xmath13 ) in three and higher dimensions that have a lipschitz singularity on a hyperplane .
denote @xmath43 by @xmath44 and let @xmath45 .
more precisely , we show : [ higherdimex ] in dimension @xmath42 , for any @xmath46 there exists a solution to @xmath27 that is a positive multiple of @xmath47 in @xmath48 .
let @xmath49 .
we search for a convex function @xmath50 in @xmath51 , with @xmath23 , that glues `` nicely '' across the boundary to @xmath47 . to that end
we look for a function with the homogeneity @xmath52 so that @xmath53 is invariant under the rescaling .
let @xmath54 then @xmath55 is continuous on @xmath56 .
furthermore , @xmath57 is the line segment between @xmath58 , which has measure zero .
thus , in the alexandrov sense , @xmath59 can be computed piecewise . in
the cylindrical coordinates @xmath60 one easily computes @xmath61 for @xmath46 the right hand side is locally bounded .
the bound on @xmath62 can be understood by looking at the gradient map of @xmath35 , which takes a `` ring '' of volume like @xmath63 to a `` football '' of length of order @xmath64 and radius of order @xmath65 ( see figure [ gradmap ] ) . then impose that it decreases volume
. decreases volume if @xmath46 . ]
observe that @xmath59 grows like @xmath66 from its zero set .
this is in a sense optimal ; if @xmath67 then one can modify alexandrov s two - dimensional argument to show that the singularity has no extremal points .
in this section we prove the main theorem [ main ] .
we construct our example in several steps .
first , let @xmath69 be a smooth , convex function such that @xmath70 for @xmath71 and @xmath72 for @xmath73 , where @xmath23 . then define @xmath74 it is easy to check that @xmath20 is a @xmath28 convex function , and in the alexandrov sense , @xmath75 in particular , @xmath76 is bounded , and decays like @xmath77 for @xmath34 large .
let @xmath78 be the rescalings defined by @xmath79 observe that @xmath80 and we have @xmath81 in the following key lemma we show that any superposition of @xmath82 vertically - translated copies of @xmath78 has bounded monge - ampre measure in @xmath83 , and separates from its tangent planes when we step away from the @xmath17 axis .
[ keylemma ] let @xmath84 be fixed numbers with @xmath85 for all @xmath86 , where @xmath87 is any positive integer .
let @xmath88 then @xmath89 for some @xmath90 independent of @xmath87 and the choice of @xmath91 , and @xmath92 for some @xmath93 independent of @xmath87 and the choice of @xmath91 .
* proof of ( [ detbound ] ) : * since @xmath94 is bounded we may assume that @xmath95 .
let @xmath96 .
since @xmath97 is @xmath28 , the curves @xmath98 do nt contribute anything to @xmath99 , so we may assume that @xmath100 for any @xmath86 . then in a neighborhood of @xmath101 , a subset of @xmath102 of the translates are not linear , and all are linear if in addition @xmath103 . up to relabeling the indices and subtracting a linear function of @xmath17 , by ( [ invariance ] )
we can write @xmath104 in a neighborhood of @xmath101 . since @xmath85
, one easily computes that @xmath105 and @xmath106 if @xmath103 .
we conclude that @xmath107 where @xmath90 does not depend on @xmath87 .
* proof of ( [ separation ] ) : * since @xmath20 is monotone increasing in the @xmath108 direction , we have for @xmath109 that @xmath110 since @xmath23 , the lower bound @xmath111 is strictly positive . by ( [ invariance ] ) the same argument gives @xmath112 for @xmath109 .
finally , since @xmath85 , we have for @xmath113 that @xmath114 \geq \sum_{i = 1}^n \frac{\mu}{n } = \mu > 0,\ ] ] completing the proof .
we can now complete the construction . roughly , at stage @xmath7 we superpose @xmath115 vertical translations of @xmath116 , starting at the endpoints of the intervals removed up to the @xmath117 stage in the construction of the cantor set .
fix @xmath118 and define @xmath119 then @xmath120 is a piecewise linear function of @xmath17 outside of four equally spaced cusps in @xmath121 connected to thin strips in @xmath18 ( see figure [ cantor1 ] ) .
is a piecewise linear function of @xmath17 outside of the four equally spaced cusps between @xmath122 and @xmath123 . ]
define @xmath124 inductively by @xmath125.\ ] ] we first claim that @xmath126 are uniformly bounded ( in @xmath7 ) in @xmath83 .
indeed , each @xmath124 is a sum of @xmath115 vertical translates of @xmath116 by values in @xmath127 $ ] , so this follows from ( @xmath128 ) .
next we show that @xmath126 are uniformly bounded in @xmath19 .
note that @xmath124 are linear functions of @xmath17 in @xmath129 , so in @xmath130 , the rescaled copies of @xmath124 in the definition of @xmath131 are linear where the other is nontrivial ( the determinants `` do nt interact '' , see figure [ cantor2 ] ) . since the rescaling @xmath132 preserves hessian determinants
, we conclude that @xmath133 one easily checks that @xmath134 is bounded , so the claim follows by induction .
is obtained by superposing two rescaled copies of @xmath120 , whose hessians do nt affect each other in @xmath130 . ] since @xmath135 in @xmath136 , the functions @xmath124 are locally uniformly lipschitz and bounded and thus converge locally uniformly to some @xmath137 .
the right hand sides @xmath126 converge weakly to @xmath138 ( see @xcite ) , so @xmath139 in all of @xmath19 .
finally , let @xmath140 be the function obtained by translating @xmath137 to the right and reflecting over the @xmath17 axis .
it is clear that @xmath35 is even in @xmath34 and @xmath17 , and is a one - dimensional function @xmath141 in the strip @xmath142 .
it is easy to show that @xmath143 is the standard cantor function ( appropriately rescaled ) , so @xmath144 has a nontrivial cantor part . indeed , @xmath145 jumps by @xmath146 over each of @xmath115 intervals of length @xmath147 centered at the endpoints of the sets removed in the construction of the cantor set .
by ( [ separation ] ) we also have @xmath148 since @xmath35 is even over both axes we conclude that @xmath149 \times [ -c , c].\ ] ] by convexity , @xmath35 has bounded sublevel sets , completing the proof .
the second derivatives of a solution to @xmath27 can not concentrate on a single line segment , since lipschitz singularities propagate .
( compare to the example above , where the second derivatives concentrate on a family of horizontal rays . ) in this section we investigate more closely how the second derivatives of a solution to @xmath27 can behave near a single line segment .
we first construct , for any @xmath29 , examples that grow from the origin like @xmath150 , with @xmath8 not in @xmath151 .
we then construct a family of barriers related to these examples in the case @xmath152 .
finally , we use these barriers to prove that singularities of the form @xmath33 propagate .
[ logexamples ] for any @xmath153 there exists a solution to @xmath27 that vanishes at @xmath39 and lies above @xmath154 , and whose hessian is not in @xmath155 .
let @xmath156 for some positive even function @xmath157 to be determined .
( by @xmath158 we mean @xmath159 )
. in @xmath160 , define @xmath161 we would like to glue this to a function of @xmath17 on @xmath162 , which imposes the condition @xmath163 on the boundary .
computing , we find that @xmath164 in this way we ensure that @xmath165 glues in a @xmath15 manner across @xmath166 to some function @xmath167 in @xmath168 defined by @xmath169 the agreement of derivatives on @xmath170 gives @xmath171 which upon differentiation and using the formula for @xmath157 gives @xmath172 for @xmath173 small it follows that @xmath174 giving the non - integrability claimed ( after , say , replacing @xmath34 by @xmath175 ) .
it remains to show that @xmath176 is positive and bounded .
one computes for @xmath177 that @xmath178 completing the proof .
we now construct barriers that agree with @xmath33 except for in a very thin cusp around the @xmath34 axis where the monge - ampre measure is as large as we like .
let @xmath179 where @xmath153 is large .
let @xmath180 be a thin cusp around the positive @xmath34 axis and let @xmath181 be its complement .
our barrier is @xmath182 note that @xmath183 is convex and bounded by @xmath64 on @xmath184 , and @xmath183 is continuous across @xmath185 .
furthermore , on @xmath185 one computes ( from inside @xmath186 ) that @xmath187 so the derivatives have positive jumps across @xmath185 .
denote @xmath188 .
one computes in @xmath186 ( where @xmath189 ) that @xmath190 finally , let @xmath191 \times [ -1/4 , 1/4].$ ] we conclude that @xmath183 are convex in @xmath3 , with @xmath192 and furthermore @xmath193 we prove theorem [ logarithmicpropagation ] by sliding the barriers @xmath183 from the right . by rescaling and multiplying by a constant , we may assume that @xmath194 with @xmath36 and @xmath195 for some large @xmath196 .
choose @xmath62 so large that @xmath197 .
slide the barriers @xmath198 from the right .
since @xmath199 on @xmath200 for all @xmath201 small , it follows from the maximum principle that @xmath202 for some @xmath203 .
( indeed , if not , we can take @xmath204 small and get a contradiction ; see figure [ barrier ] )
. taking @xmath205 , we conclude that @xmath206 . on the right edge of @xmath207
, then we get a contradiction by sliding @xmath183 to the left . ] by convexity , near each point on the @xmath34 axis where @xmath35 is zero , there is a singularity of the same type as near the origin .
we can apply the above argument at all such points to complete the proof .
this work was supported by nsf grant dms-1501152 .
i would like to thank a. figalli and y. jhaveri for helpful comments .
9999 a. d. alexandrov , _ smoothness of the convex surface of bounded gaussian curvature _ , c. r. ( doklady ) acad .
urss ( n. s. ) * 36 * ( 1942 ) 195 - 199 .
l. ambrosio , m. colombo , g. de philippis and a. figalli , _ existence of eulerian solutions to the semigeostrophic equations in physical space : the 2-dimensional periodic case _ ,
partial differential equations * 37 * ( 2012 ) , no . 12 , 2209 - 2227 . l. ambrosio , m. colombo , g. de philippis and a. figalli , _ a global existence result for the semigeostrophic equations in three dimensional convex domains _ , discrete contin .
* 34 * ( 2014 ) , no . 4 , 1251 - 1268 .
p. daskalopoulos and o. savin , _ on monge - ampre equations with homogeneous right - hand sides _ ,
pure appl . math .
* 62 * ( 2009 ) , no . 5 , 639 - 676 . g. de philippis and a. figalli , _ the monge - ampre equation and its link to optimal transportation _ , bull .
( n. s. ) * 51 * ( 2014 ) , no . 4 , 527 - 580 . g. de philippis and a. figalli , _
@xmath0 regularity for solutions of the monge - ampre equation _ , invent .
* 192 * ( 2013 ) , 55 - 69 . g. de philippis , a. figalli and o. savin , _ a note on interior @xmath208 estimates for the monge - ampre equation _ , math .
* 357 * ( 2013 ) , 11 - 22 .
a. figalli , _ global existence for the semigeostrophic equations via sobolev estimates for monge - ampre _ , cime lecture notes , springer , to appear .
a. figalli and g. loeper , _
@xmath15 regularity of solutions of the monge - ampre equation for optimal transport in dimension two _ , calc .
partial differential equations * 35 * ( 2009 ) , no . 4 , 537 - 550 .
p. guan , _ regularity of a class of quasilinear degenerate elliptic equations _ , adv
* 132 * ( 1997 ) , no . 1 , 24 - 45 .
c. gutierrez , `` the monge - ampre equation , '' progress in nonlinear differential equations and their applications 44 , birkhuser boston , inc . , boston , ma , 2001 .
t. schmidt , _
@xmath208 estimates for the monge - ampre equation _ , adv . math .
* 240 * ( 2013 ) , 672 - 689 .
wang , _ some counterexamples to the regularity of monge - ampre equations _ ,
* 123 * ( 1995 ) , 841 - 845 . | we construct a counterexample to @xmath0 regularity for convex solutions to @xmath1 in two dimensions .
we also prove a result on the propagation of singularities in two dimensions that are logarithmically slower than lipschitz .
this generalizes a classical result of alexandrov and is optimal by example . |
in this article , we consider a semi - discrete model for the evolution of a driven interface subject to line tension in a random , heterogeneous , quenched environment .
we prove that if the driving force is large enough then such an interface propagates with a positive velocity even if the random environment contains obstacles of arbitrarily large strength .
let @xmath1 be a probability space and consider the following lattice differential equation for the height @xmath2 of the @xmath3 dimensional interface in an ambient space of dimension @xmath4 , @xmath5 where @xmath6 , @xmath7 , and @xmath8 , @xmath9 .
the initial condition is @xmath10 .
the operator @xmath11 denotes the discrete @xmath0-dimensional laplacian operator , namely @xmath12 , where @xmath13 denotes the discrete 1-norm .
the one - dimensional setting was discussed in @xcite , in this note we generalize our results to arbitrary dimension , albeit only for the ( semi-)discrete evolution .
we assume that @xmath14 , @xmath15 are independent and identically distributed functions such that the map @xmath16 is measurable with respect to the product of the borel-@xmath17 algebra on @xmath18 and @xmath19 and the map @xmath20 is locally lipschitz for almost all @xmath8 . furthermore , we assume that @xmath21 . disregarding infinitely fast growing unphysical solutions
, these assumptions guarantee that equation above admits a unique solution with non - negative velocity for every @xmath22 and that the solution depends measurably on @xmath23 for each @xmath24 .
the solution furthermore admits a comparison principle .
the main further assumption on the @xmath25 is that they admit a finite exponential moment .
as opposed to some other requirements , like independence , this assumption is central to our proof . under these conditions
, we can prove our main result .
[ thm : main ] assume in addition to the above requirements that there exists @xmath26 such that @xmath27 where @xmath28 denotes taking the integer ceiling of the argument .
then there exists a non - decreasing function @xmath29 which is not identically zero and which depends on @xmath30 and @xmath31 only , such that for all @xmath32 we have @xmath33 and therefore @xmath34 specifically , we can choose @xmath35 where @xmath36 denotes taking the integer floor of the argument .
the proof is split in two parts , first a discrete result arguing that there can be no discretized interface whose average velocity is small .
the second part is an application of this result to the coupled systems of odes .
taking @xmath37 large for small values of @xmath38 , one can see that @xmath39 for all @xmath40 .
furthermore , as @xmath38 becomes large , one can take @xmath37 closer to @xmath30 to see that there exists a constant @xmath41 , depending only on @xmath30 , @xmath31 , and @xmath0 , such that @xmath42 for all @xmath43 .
the main theorem also implies the following almost sure result for the point - wise velocity , excluding the existence of stationary solutions .
[ thm : limsup ] under the conditions of theorem [ thm : main ] we also have @xmath44 along any deterministic sequence of times @xmath45 as @xmath46 .
a model very similar to the one considered here was recently discussed in @xcite .
as opposed to our model , they use a fully discrete evolution , where in each time - step the system advances by one unit at every point where the total force is positive .
while some of their results are comparable to ours , they use a rigorous renormalization group approach to prove that in their model ( assuming also uniformly bounded obstacles ) , an interface is either completely blocked ( in the sense that a non - negative stationary solution exists ) or that it propagates ballistically , i.e. , there is no intermediate regime of sub - ballistic propagation . in more general models , like the one here
, this question remains open .
in particular , for the present model , we can only prove that there exist two critical values for the driving force : if the driving force is below the first value , the interface becomes stuck for all times .
if , on the other hand , the driving force is above the second value the interface propagates with finite velocity . the first result is a simple adaptation of our methods in @xcite and the second part is proved here .
generally , problems of the present form ( whether fully discrete , partially discrete , or fully continuous ) have received considerable interest in the physics community ( see for example @xcite ) .
many connections to questions arising from physics are discussed in the aforementioned article by bodineau and teixeira @xcite , as well as in @xcite , where the first rigorous result on non - existence of stationary states was derived .
the article is organized as follows . in section
[ sec : discrete ] , we show non - existence of states whose velocity is too small . in the following section [ sec : cont ] , we apply this result to prove our theorem .
we finish with some conclusions and an outlook in section [ sec : conc ] .
in this section , we prove the central lemma stating that in a fully discrete version of our model , one can with probability one not find any function whose average velocity is too small .
let thus now @xmath47 defined for all @xmath48 .
for convenience , we begin by introducing some notation .
we use the following abbreviations .
* @xmath49 , the @xmath0-dimensional cube of sites in @xmath50 of side - length @xmath51 , * @xmath52 , the @xmath17-algebra generated by the random functions in @xmath53 , * @xmath54 , any fixed number , later to be taken as the integer ceiling of an _ a priori _ bound on the maximal value the functions @xmath55 , solutions of can take at time @xmath56 , * @xmath57 , the set of admissible functions , * @xmath58 , the set of admissible functions within a cube @xmath59 , * @xmath60 , the size of the boundary layer around @xmath53 , * @xmath61 , the number of ways @xmath62 can be represented as the sum of @xmath63 ( ordered ) non - negative integers .
[ lem : discrete ] for each @xmath64 , there exists a set @xmath65 of full measure such that for any @xmath66 and any function @xmath67 we have @xmath68 and @xmath31 and @xmath30 are defined in theorem [ thm : main ] .
fix @xmath69 and consider for @xmath70 the sequence of random variables @xmath71 the basic underlying idea in this definition is the following .
we will show , using a martingale argument , that for sufficiently large @xmath38 the sequence @xmath72 almost surely vanishes exponentially in the size of the box @xmath53 .
for this decrease we can also establish a rate .
such a decrease , however , implies that as we look at larger and larger boxes around the origin , either the sum of the normal derivatives at the boundary of the box ( the first term in the exponential ) has to become large and negative quickly , or the sum of the velocities ( the second term in the exponential ) in @xmath53 has to increase with a rate related to the one with which @xmath72 vanishes .
the first option is excluded by the non - negativity of @xmath73 .
the second option yields the average velocity ( with a negative sign ) , after taking a logarithm and using the sum over all paths as an estimate for the supremum over all possible paths .
the first step in the proof is to relate the change in normal derivatives as @xmath74 increases to the addition of terms in the sum over the laplcian .
we use a discrete version of the divergence theorem , namely that @xmath75 and thus @xmath76 a calculation now yields @xmath77 where the sum in the second line is taken over all admissible extensions of @xmath73 to functions in @xmath78 .
taking now @xmath79 with the sum as above over all possible extensions , we get @xmath80 in order to estimate @xmath81 further , we need to rearrange and count the number of possible extensions . in the sum over all admissible extensions
we thus first take all extensions such that @xmath82 , calling these `` admissible extensions with velocity @xmath83 '' and then sum over all @xmath84 . in the case that there does not exist an admissible extension with velocity @xmath83 , we take the sum to be zero .
this yields @xmath85 where @xmath86 is the of the number of admissible extensions with velocity @xmath83 , depending on the realization of the random field @xmath87 and on @xmath73 from the previous step .
we also note that @xmath88 for @xmath89 is a fixed value inside the supremum , which allows us to use the assumption on the exponential moment of @xmath87 .
the idea for estimating @xmath90 now is the following : given @xmath83 , there are no more than @xmath91 possibilities to distribute these velocities on the @xmath92 sites . with all velocities fixed , for most sites in @xmath93 where the extension lives ,
the function value is determined due to the fact that @xmath23 and the velocity can be used to calculate the discrete laplacian ( if such a choice exists at all ) .
the number of sites where we still have freedom is @xmath94 .
we thus aim for an estimate of the type @xmath95 .
first notice that in dimension @xmath96 , there is no freedom to choose any additional values for the extension if the laplacian at the boundaries is given .
consider thus the case @xmath97 .
each given function value and the laplacian on the site can be used to write one independent linear equation for the function values on the extension , therefore the remaining number of sites with freedom is @xmath98 .
see the illustration in figure [ fig : extension ] for the two - dimensional case .
note that the number of choices for each of those nodes is limited to at most @xmath99 , and that @xmath100 .
since the above estimate was independent of @xmath101 and on @xmath102 , this yields @xmath103 using the estimate @xmath104 , the sum can be bounded as follows .
we have @xmath105 which , using @xmath106 yields @xmath107 we thus have @xmath108 for some constant @xmath41 .
equation together with the boundedness of @xmath81 and the almost sure finiteness of @xmath109 establishes that @xmath110 for @xmath111 is a non - negative supermartingale . from doob
s martingale convergence theorem we therefore find that @xmath112 on a set @xmath113 of full measure .
note furthermore that @xmath114 and thus @xmath115 where we have used that @xmath116 and dropped all terms that are of lower order than @xmath117 .
in particular , these are the terms in @xmath118 that are of lower order than @xmath119 as well as @xmath120 and the first sum inside the exponent in @xmath72 , which vanishes in the limit due to the boundedness of @xmath73 .
this proves the lemma .
the lemma from the above section allows us to complete the proof of the main theorem .
assume that the statement in the theorem is untrue .
then there exist @xmath121 and some @xmath122 such that @xmath123 . by our independence assumptions on the field @xmath87
, the processes @xmath124 , @xmath125 , @xmath15 are stationary and ergodic and take values in @xmath126 .
we write @xmath55 instead of @xmath124 . by birkhoff
s ergodic theorem , we have @xmath127 almost surely . however , taking @xmath88 to be @xmath55 rounded to the closest integer , we find @xmath128 by lemma [ lem : discrete ] .
the almost sure statement about the velocities can be derived by the following argument .
consider , for a fixed sequence of times @xmath45 , the random variables @xmath129 noting that @xmath130 is stationary , ergodic and bounded from above and below by @xmath38 and @xmath131 , respectively .
furthermore , we have @xmath132 , by fatou s lemma . by the non - negativity of the velocity and @xmath25
, it follows that @xmath133 for all @xmath134 and almost all @xmath23 , and therefore @xmath135 for almost all @xmath23 .
now let @xmath136 . by stationarity
, @xmath137 is constant in @xmath22 and we write @xmath138 . by the discrete divergence theorem , boundedness of @xmath130 and ergodicity of @xmath139 imply that @xmath140 and since @xmath135 for almost all @xmath23 we have @xmath141 almost surely and for all @xmath15 .
this yields that @xmath142 is a bounded , ergodic , and stationary process whose realizations are almost surely harmonic .
thus , @xmath142 is almost surely constant in @xmath22 and therefore @xmath143 is almost surely equal to its expected value .
the desired result follows .
in this note , we have extended our depinning result from @xcite to the case of arbitrary dimension in a semi - discrete model of coupled ordinary differential equations .
a careful inspection of the proof shows that one can furthermore extend our results to obstacle strengths coupled over a finite distance : if there exists @xmath144 such that sets of obstacles are independent if their distance ( in the first @xmath0-dimensions ) is above @xmath145 , one can still obtain similar estimates for the velocity .
the case of the fully continuous model on @xmath146 , however , remains open .
further unresolved issues are whether we have @xmath147 almost surely for sufficiently large @xmath38 , the relaxation of the result to obstacles with fat tails , as well as whether a regime of sub - ballistic propagation ( i.e. , vanishing velocity , but propagation of the interface to @xmath148 everywhere ) can exist in our models . as mentioned above , for a specific fully discrete evolution model this last question was answered recently @xcite . | we consider a discretized version of the quenched edwards - wilkinson model for the propagation of a driven interface through a random field of obstacles .
our model consists of a system of ordinary differential equations on a @xmath0-dimensional lattice coupled by the discrete laplacian . at each lattice point , the system is subject to a constant driving force and a random obstacle force impeding free propagation .
the obstacle force depends on the current state of the solution and thus renders the problem non - linear . for independent and identically distributed obstacle strengths with exponential moment we prove ballistic propagation of the interface if the driving force is large enough |
the study of billiards on euclidean polygonal tables has benefitted greatly from the use of the theory of riemann surfaces .
a classical unfolding process associates to each rational - angled polygonal table a riemann surface with holomorphic 1-form , see @xcite . using this
, kerkhoff - masur - smillie @xcite showed that for almost every direction on a polygonal table , the billiard flow is ergodic . soon thereafter , veech @xcite identified a setting where the only non - ergodic directions are those of saddle connections ; these examples are identified by an associated group , the veech group , being a lattice fuchsian group . in @xcite , veech asked if in general these groups can be non - finitely generated .
only recently have examples of infinitely generated veech groups been given @xcite , @xcite .
here we study the geometry associated to some of these groups .
the riemann moduli space @xmath0 is the space of smooth compact riemann surfaces of genus @xmath1 , up to biholomorphic equivalence .
the holomorphic 1-forms on a riemann surface @xmath2 of genus @xmath1 form a @xmath1-dimensional complex vector space , @xmath3 .
pairs @xmath4 with @xmath5 form a bundle @xmath6 ( of course , @xmath0 is not a manifold , but rather an orbifold , thus a fiber is @xmath7 ) .
each pair @xmath4 ( with @xmath8 non - null ) identifies a translation structure on the real surface @xmath2 : local coordinates are defined by integration of @xmath8 . invoking the real structure of @xmath9 , there is a @xmath10-action on these translation surfaces : post - composition with the local coordinate functions . to each @xmath4
, there is thus a map from @xmath10 to @xmath0 , given by sending group elements to corresponding orbit elements .
but , rotations act so as to leave the riemann surface @xmath2 fixed .
thus there is an induced a map from @xmath11 to @xmath0 . with respect to the teichmller metric on @xmath0 ,
this map turns out to be an isometric immersion of the hyperbolic plane @xmath12 .
the resulting image is called the _ teichmller complex geodesic _ determined by @xmath4 .
we will denote this by @xmath13 .
the @xmath10-stabilizer of @xmath4 induces a subgroup of the oriented isometry group of the hyperbolic plane .
this induced group is the _ veech group _
, denoted @xmath14 .
the teichmller complex geodesic is thus the image of the hyperbolic surface ( or possibly orbifold ) uniformized by the veech group ; the map between them is generically injective .
we use this to show : [ thmgeometry ] for any genus @xmath15 , there exists a complex teichmller geodesic in the riemann moduli space @xmath0 that has infinitely many cusps and infinitely many infinite ends . we remind the reader of the definition and basic classification of ends of a hyperbolic surface below . informally , an _ end _ is a direction in which geodesics can leave compact subsets .
finite volume hyperbolic surfaces can only have cusps as ends ; briefly , a _
cusp _ , or puncture , corresponds to a conformal punctured unit disk in the surface ( which can not be completed in the surface ) .
an end of a hyperbolic surface which corresponds to a conformal non - trivial annulus is called a hole , or a flare .
these two types of ends are called finite , an _
infinite end _ is an end which is not finite .
we prove our main results using the approach of translation surfaces .
in particular , in terms that we make precise below , whereas cusps always form a discrete set , we give examples such that any direction giving an infinite end is the limit of directions of inequivalent infinite ends . a translation surface @xmath16
is called a _ veech surface _ if @xmath17 is a lattice in @xmath18 ; the definition of a non - periodic connection point can be found in subsection [ sadparspecpts ] [ inffatboys ] let @xmath4 be a translation surface which is a finite ramified covering of a veech surface @xmath16 , with ramification above some non - periodic connection point and possibly the singularities of @xmath16 .
then @xmath19 has infinitely many cusps and infinitely many infinite ends .
choose a dirichlet fundamental domain for @xmath14 acting upon the hyperbolic plane @xmath20 ; this determines an infinite set @xmath21 of points in @xmath22 , each representing a distinct infinite end .
the set @xmath21 has no isolated points . that @xmath14 , for @xmath4 as in theorem [ inffatboys ] , is infinitely generated is the main result of @xcite . for the @xmath4 of theorem [ inffatboys ] ,
the complicated geometry at infinity of @xmath19 is projected to a small sublocus of the riemann moduli space .
[ thmbigcusp ] let @xmath23 be as in theorem [ inffatboys ] , and suppose that @xmath2 is of genus @xmath1 .
then , there is a two - complex dimensional sublocus @xmath24 such that all ends of the complex teichmller geodesic generated by @xmath23 are contained in finitely many ends of @xmath25 .
we prove theorem [ thmgeometry ] by restricting to the case where @xmath4 satisfies the hypotheses of theorem [ inffatboys ] an application of the riemann - hurwitz formula shows that we may assume that the corresponding @xmath16 be of genus two , we give explicit examples of such @xmath16 which admit non - periodic connection points in @xcite , it is now known @xcite , @xcite that the set of such examples is large the existence of infinitely many cusps is proven in [ manycusps ] , the existence of infinitely many infinite ends is proven in [ manyends ] , that every direction of an infinite end is the limit of distinct infinite end directions is also proven in [ manyends ] .
we prove theorem [ thmbigcusp ] in [ locend ] . in 2 we give background material on translation surfaces and certain @xmath10-actions , in 3 we discuss projections of horocycles to the moduli space @xmath0 .
the main results rest upon the elementary tools presented in 4 . that a hyperbolic surface uniformized by an infinitely generated fuchsian group of the first kind admits infinite ends can be deduced already from @xcite , theorem 16 .
thus , the examples of infinitely generated veech group given by mcmullen @xcite must also have these .
each of mcmullen s examples has its associated teichmller complex geodesics lying densely on a symmetric hilbert modular surface , all ends must lie in the finitely many ends of the hilbert modular surface .
thus , if here also there are infinitely many infinite ends , then they may well show the geometry of the type we discuss here .
it is a pleasure to thank f. dalbo , g. forni , and j. smillie for helpful discussions . in particular , we thank g. forni for his ideas which are reflected in [ projhoro ] .
we thank a. eskin and others , see @xcite , for continued interest in these matters .
in order to establish notation , we review basic notions . see say @xcite for more details ; @xcite , @xcite and @xcite use related notation and notions . see also @xcite , @xcite , @xcite for recent related developments . a _ translation surface _
is a real surface with an atlas such that all transition functions are translations . as usual , we consider maximal atlases . given a riemann surface @xmath2 and a holomorphic 1-form @xmath26 , integration of @xmath8 induces local coordinates on @xmath2 away from the zeros of @xmath8 ; regarding @xmath9 with its standard structure as @xmath27 , transition functions are easily seen to be translations . in particular , directions of flow on a translation surface are well - defined , by pull - back from the plane .
one easily checks that the translation surface associated to @xmath4 is well - defined up to isomorphism of translation surface ( such isomorphism is realized by local pure translations ) , see section 2.1 of @xcite for a more formal discussion .
the holomorphic 1-form @xmath8 induces a flat metric on @xmath2 off of its zeros ; when the metric is completed , a zero of multiplicity @xmath28 gives a conical singular point , of cone angle @xmath29 .
we consider only translation surfaces of this type ; the notation @xmath4 hence denotes both the ordered pair of riemann surface @xmath2 with holomorphic 1-form @xmath8 , and the translation surface structure , with which @xmath8 equips @xmath2 .
given a translation surface @xmath4 , and an element @xmath30 , we can post - compose the coordinate functions of the charts of the atlas of @xmath4 by @xmath31 .
it is easily seen that this gives again a translation surface , denoted by @xmath32 .
given a translation surface @xmath4 , let @xmath33 denote the surface which arises from deleting the singularities of @xmath4 .
an _ affine diffeomorphism _
@xmath34 of @xmath4 is a homeomorphism of @xmath2 that restricts to a diffeomorphism of @xmath33 whose differential at all points is a constant element of @xmath10 . in other terms
, @xmath34 induces a real - affine map in the charts of the @xmath8-atlas on @xmath2 , of constant ( jacobian ) derivative .
the orientation - preserving affine diffeomorphisms form a group , denoted @xmath35 .
the subgroup of @xmath10 which arises as the derivatives of affine diffeomorphisms is denoted by @xmath36 .
if @xmath37 have the same derivative , then the diffeomorphism @xmath38 acts by pure translations .
this diffeomorphism thus preserves the conformal structure on the underlying translation surface ; that is , @xmath38 is a conformal automorphism of the riemann surface @xmath2 ( that preserves @xmath8 ) .
the kernel of the group homomorphism from @xmath35 to @xmath36 is hence certainly finite whenever @xmath2 is of genus greater than one . by pull - back on charts
, each @xmath39 also gives a translation structure on the underlying surface .
indeed , if @xmath31 is the constant derivative of @xmath40 with respect to the @xmath8-charts , then @xmath40 is an isomorphism of the translation surface @xmath32 with @xmath23 .
thus , the @xmath10-stabilizer of @xmath4 is @xmath36 .
fix some @xmath4 , and let @xmath41 be given by @xmath42 ; let @xmath43 denote the standard projection .
note that the underlying riemann surface @xmath44 is common to the @xmath45-orbit of any @xmath46 the image of @xmath16 under a rotation is @xmath47 , where @xmath48 is the complex number which induces the rotation .
thus , @xmath49 induces a map ( from right cosets ) @xmath50 .
the image of @xmath34 is @xmath13 , the _ teichmller complex geodesic _ associated to @xmath4 .
right multiplication gives the natural right group action of @xmath51 on the cosets @xmath52 .
the center , @xmath53 , of @xmath51 is a subgroup of @xmath45 , and one finds that @xmath54 acts faithfully .
the @xmath18-stabilizer of the map @xmath34 is the _ veech group _ @xmath55 , the image in @xmath18 of @xmath56 .
this is trivially verified : @xmath57 implies that there is some rotation @xmath58 such that @xmath59 .
thus , the class @xmath60 is in the @xmath55-orbit of @xmath61 . that @xmath34 is single - valued on each such orbit
is obvious .
the map @xmath34 is generically injective ; it can only fail to be injective at some @xmath46 if @xmath62 is larger than the automorphism group of riemann surfaces corresponding to neighboring points of @xmath0 , see @xcite and @xcite .
when the veech group normalizes a true surface , then @xmath34 gives a ramified cover normalizing , in the sense of algebraic curves , @xmath13 . in @xcite ,
veech proved that @xmath55 is a discrete subgroup of @xmath18 .
that is , it is a fuchsian group . the celebrated _ veech dichotomy
_ , also proven in @xcite , states that if @xmath14 is a lattice that is , if @xmath63 has finite hyperbolic area , then the linear flow in any direction of the surface @xmath4 is either ergodic or periodic .
one thus says that a translation surface is a _ veech surface _ if @xmath14 is a lattice .
we also say that @xmath4 is an _ arithmetic surface _ if @xmath14 is arithmetic , in that it is @xmath18-commensurable to @xmath64 .
the teichmller metric on @xmath0 is found by giving the teichmller metric on the ( ramified ) cover teichmller space @xmath65 of @xmath0 ; the teichmller modular group ( equivalently , the mapping class group ) acts isometrically on @xmath65 , with quotient @xmath0 .
thus , the teichmller metric descends to @xmath0 . we avoid direct discussion of teichmller space here ,
the reader may wish to refer to , say , @xcite .
if @xmath66 is a non - zero real , then the translation surfaces @xmath4 and @xmath67 differ only by scaling .
this scaling and the @xmath10-action commute ; furthermore , the projection map @xmath68 ignores such scaling .
thus , we can restrict our discussion to @xmath69 , the collection of area one @xmath4 . in informal terms , whereas a conformal map of riemann surfaces takes infinitesimal circles to circles , a _ quasi - conformal _ map of a riemann surfaces takes infinitesimal circles to ellipses .
beltrami coefficient _ of a quasi - conformal homeomorphism @xmath70 is computed by @xmath71 , where @xmath72 is a local chart in @xmath44 , and the symbols @xmath73 and @xmath74 denote @xmath75 and @xmath76 , respectively .
now , if @xmath77 and @xmath30 , then the identity map on @xmath2 , from @xmath4 to @xmath78 has beltrami coefficient @xmath79 , where @xmath80 is the complex dilatation of the linear map given by @xmath31 on @xmath9 with its standard real structure .
but , the map sending elements @xmath81 to @xmath82 is constant on right @xmath83-cosets .
this can be shown by direct calculation , or see corollary 1 of @xcite .
indeed , @xmath84 injectively identifies @xmath85 with the complex unit disk , @xmath86 .
the teichmller metric on @xmath85 is the pull - back of the hyperbolic metric on @xmath86 .
but , as @xcite show in their lemma 4.2 , the map sending @xmath87 by @xmath88 is conformal ; it is thus an oriented isometry .
hence , the veech group is the stabilizer in the oriented isometry group of @xmath50 .
it is reasonable to identify @xmath85 with @xmath20 , as does @xcite . in this case
the stabilizer of the map @xmath89 giving the teichmller complex geodesic is again the veech group , but acting on the left by mbius transformations conjugated with the action of @xmath90 .
this conjugating action corresponds to @xmath91 , the constant derivative of a coset representative of the index two subgroup @xmath92 in the full group of affine diffeomorphisms of @xmath23 .
thus , @xmath93 _ is _ @xmath14 ; compare with proposition 3.2 of @xcite . in what follows
, we will ignore this innocuous conjugation , and consider @xmath14 acting on @xmath20 by mbius transformation .
any translation surface of genus greater than one must have singularities .
a geodesic line emanating from a singularity is called a _ separatrix _ , a separatrix connecting singularities ( with no singularity on its interior ) is a _
saddle connection_. we call a point @xmath94 a _ connection point _
if every separatrix passing through @xmath95 extends to a saddle connection .
we say @xmath96 is a _
parabolic direction _ for @xmath4
if there is a parabolic element of @xmath14 whose associated affine diffeomorphism preserves the direction @xmath96 .
( in fact , the full veech dichotomy states that the non - ergodic directions of a veech surface are parabolic . )
we also call these associated affine diffeomorphisms parabolic . expressing @xmath96 as a slope
, the point on @xmath22 associated to this slope is @xmath97 ; this is due to the fact that @xmath10 acts on @xmath4 by way of its _ linear _ action on @xmath27 , but the @xmath10-action on @xmath20 is _
fractional linear_. thus for example , an element of @xmath10 arising from an element of @xmath35 which fixes the vertical direction ( should such an element exist ) acts on the extended hyperbolic plane so as to fix the boundary point @xmath98 , see figure 1 .
a point @xmath94 of finite orbit under the group of affine diffeomorphisms is called a
_ periodic point_. a point of finite orbit under the subgroup of affine diffeomorphisms generated by the two ( primitive ) parabolic diffeomorphisms for some pair of transverse parabolic directions is called a _ rational point _ for @xmath4 . in @xcite ,
one shows that non - periodic points have dense orbits . in @xcite
, we show that every connection point is a rational point .
to date , constructions of infinitely generated veech groups rely on the following basic fact of fuchsian groups : ( 1 ) a fuchsian group of the first kind ( that is , whose limit set is all of @xmath99 ) is either a lattice group or else is infinitely generated . in @xcite , we used two further facts : ( 2 ) given a ramified covering @xmath100 of riemann surfaces , upon setting @xmath101 the translation surface @xmath4 is a _ balanced translation covering _ of the marked translation surface @xmath102 if every point of the fiber above a nonsingular branch point is ramified , in particular , @xmath14 is then commensurable to @xmath103 , where this latter is the image in @xmath17 of the affine diffeomorphisms which stabilize the branch locus of @xmath34 ; and , ( 3 ) if there is an element of this branch locus that is non - periodic ( that is , such that its @xmath104-orbit is of infinite order ) , then @xmath103 is of infinite index in @xmath17 ; in particular , it is thus not a lattice .
therefore , the commensurable @xmath14 also fails to be a lattice .
given any point of a translation surface , the directions of separatrices through the point are dense ( see , say , lemma 1 of @xcite , in turn a slight generalization of a result of @xcite ) ; the direction of any saddle connection on a veech surface is in fact a parabolic direction ; one finds that if the branch locus contains a connection point , then the parabolic directions of @xmath103 are dense in @xmath22 , and thus the limit set this group is certainly all of @xmath22 .
now , if the connection point is non - periodic , then we conclude that that @xmath103 is infinitely generated ; again , the commensurable @xmath14 is therefore also infinitely generated .
but , we show in @xcite that there exist @xmath16 that admit non - periodic connection points , and thus proved the existence of infinitely generated @xmath14 . + for an explicit example , we showed in @xcite ( see in particular proposition 7 there ) that the genus two surface given by the golden cross can be taken to play the role of @xmath16 .
this surface has exactly one singularity , say @xmath105 ; let us also choose one of its infinitely many non - periodic connection points , @xmath95 . if @xmath106 is even , then we can take @xmath107 copies of our genus two surface , with a slit from @xmath105 to @xmath95 ; identify these with the usual process after having chosen some transitive permutation on @xmath107 letters .
this gives a degree @xmath107 ramified covering , whose branch locus consists of @xmath108 , and which is totally ramified above each of these points .
the riemann - hurwitz formula then shows that the covering riemann surface does have genus @xmath1 .
one can choose the odd genus @xmath109 by way , say , of double coverings : choose @xmath110 connection points of our genus two surface ; create @xmath111 slits and glue two copies in the standard manner . for more on creating ramified coverings , see say @xcite .
a _ cylinder _ on a translation surface @xmath4 is a maximal connected set of homotopic simple closed geodesics .
if the genus of @xmath2 is greater than one , then every cylinder is bounded by saddle connections . upon cutting the cylinder along a line perpendicular to the direction of flow
, we can represent it as indicated in figure 1 : of width @xmath112 and height ( or circumference ) @xmath113 .
a cylinder has its affine dehn twist @xmath114 .
the _ modulus _ of the cylinder is @xmath115 .
if for some direction @xmath96 the surface @xmath4 decomposes into cylinders in this direction , then we say that @xmath96 is a _
( completely ) periodic direction_. veech @xcite shows that a completely periodic direction @xmath96 is a parabolic direction if and only if the various cylinders in the direction @xmath96 gave commensurable inverse moduli : all @xmath116 are rational .
the teichmller _ horocyclic flow _ is given by @xmath117 , where @xmath118 , with @xmath119 .
the orbit of @xmath4 under this flow is called a _
horocycle_. suppose that @xmath4 is completely periodic in the vertical direction , that is @xmath4 has a decomposition into vertical cylinders .
the effect of @xmath120 , with @xmath121 , restricted to a cylinder of the decomposition is to change the relative positions of the singularities on the right hand side of the cylinder with respect to those on the left .
it is easily seen that on a given vertical cylinder of height @xmath113 and width @xmath112 , that a standard affine twist on the cylinder has been performed when @xmath122 .
( thus , when @xmath105 is the multiplicative inverse of the modulus of the cylinder . )
see figure 1 .
horocyclic flow : moving a singularity .
if the various moduli of the cylinders are commensurable , then there is a least positive @xmath105 such that @xmath120 induces on each cylinder an integer power of the basic dehn twist of the cylinder .
this @xmath120 is then easily seen to belong to @xmath36 .
indeed , the element @xmath120 then generates a parabolic subgroup of @xmath14 , and the @xmath14-conjugacy class of this parabolic subgroup corresponds to a _
cusp_. teichmller _ geodesic flow _ is given by @xmath123 , where @xmath124 , with @xmath125 .
for positive @xmath126 , the effect of @xmath127 restricted to a vertical cylinder decomposition is to decrease heights and increase widths .
note that vertical saddle connection lengths tend to zero as @xmath128 .
thus , if @xmath129 has saddle connections in the vertical direction , then for @xmath126 sufficiently large , the shortest saddle connections for @xmath123 are in the vertical direction . a _ horodisk _ of index @xmath130 corresponding to a direction @xmath96 of @xmath4 is the union over @xmath131 of all the horocycles @xmath132 .
a direction @xmath96 of @xmath4 is a _ fat direction _ if there is a sufficiently large positive real @xmath130 such no two points of the horodisk of index @xmath130 corresponding to @xmath96 are @xmath36-equivalent .
an _ end _ of a hyperbolic manifold @xmath133 is determined by a cover of @xmath133 by compact sets @xmath134 , with each @xmath134 contained in the interior of @xmath135 ; an end @xmath136 is a sequence of components @xmath137 in the complement of @xmath134 such that @xmath138 .
the definition of an end is independent of choice of sequence @xmath134 , see say @xcite .
an end @xmath136 of a hyperbolic surface is called a _ cusp _ if there is a subset @xmath139 that is conformally equivalent to the complex punctured unit disk @xmath140 such that for all sufficiently large @xmath141 , @xmath142 .
an end is called a hole if there is a subset @xmath143 that is conformally equivalent to a complex annulus @xmath144 with @xmath145 such that for all sufficiently large @xmath141 , @xmath146 .
these two types of ends are called finite , an _
infinite end _ is an end which is not finite . in a terminology introduced by haas @xcite
, one differentiates between ends of the first kind and infinite ends of the second kind ; the infinite ends that we discuss are of the first kind . to be precise
, we say that @xmath136 is an
_ infinite end of the first kind _ if there is a subset @xmath147 that is conformally equivalent to the complex upper half - plane @xmath148 such that for all sufficiently large @xmath141 , @xmath149 .
note that each of the above definitions continues to be valid upon allowing @xmath133 to be a hyperbolic orbifold , although there is then also the further possibility that the set of elliptic fixed points enters arbitrarily far into an end .
+ given a fuchsian group @xmath150 , and a point @xmath151 , we say that @xmath152 _ determines an infinite end _ of @xmath153 if there is some horodisk based at @xmath152 such that any @xmath150-orbit in @xmath20 meets this horodisk in at most one point .
equivalently , there is an infinite end of @xmath153 which lifts to include a horodisk based at @xmath152 .
analogously , we say that a parabolic fixed point @xmath152 of @xmath150 _ determines a cusp_. the following two lemmas are virtually immediate
. [ niceends ] if the fuchsian group @xmath150 is of the first kind , then any end of @xmath153 is either a cusp or an infinite end as above .
[ fatptsareinfenddir ] let @xmath154 , and let @xmath155 be a direction for @xmath4 .
let @xmath151 be given by @xmath156 .
if @xmath96 is a parabolic direction , then @xmath152 determines a cusp of @xmath157 .
if @xmath96 is a fat direction , then @xmath152 determines an infinite end of @xmath157 . for ease of discussion ,
we introduce the following terms . given a ( possibly ramified ) covering of hyperbolic surfaces or orbifolds @xmath158 , we say @xmath159 is _ balanced with respect to ends _ if the image of any end of @xmath133 is an end of @xmath160 and also the pre - image of any end of @xmath160 is an end of @xmath133 .
the terms _ balanced with respect to infinite ends _ and _ balanced with respect to cusps _ are analogously defined .
the following is well - known .
[ fincoveringends ] finite degree ramified coverings of hyperbolic surfaces or orbifolds are balanced with respect to each of the following : ends , infinite ends , and cusps .
proposition 3 of @xcite can be reworded as the following .
[ ourends ] let @xmath16 be a veech translation surface .
suppose that @xmath161 is a non - periodic connection point .
let @xmath162 denote the subgroup of @xmath14 arising from the stabilizer of @xmath95 in @xmath35 .
then the ( possibly ramified ) covering @xmath163 is balanced with respect to ends .
however , this covering is not balanced with respect to either cusps or infinite ends ; more precisely : @xmath164 has infinite ends , but @xmath165 has no infinite ends .
points defining our ends are dirichlet points .
recall that a _ dirichlet point _ is a point @xmath166 which is on the boundary of some dirichlet region thus , @xmath152 admits some @xmath167 such that @xmath152 is on the boundary of the dirichlet fundamental region for @xmath150 .
one can show , see @xcite , that a dirichlet point @xmath151 is either a parabolic fixed point of @xmath150 , or is such that the @xmath150-orbit of @xmath168 meets the boundary of a horodisk based at @xmath152 whose interior contains no element of this orbit .
recall that @xmath151 is called , after sullivan @xcite , a _ garnett point _ for @xmath150 relative to @xmath168 if the largest horodisk based at @xmath152 whose interior contains no element of the @xmath150-orbit of @xmath168 , also does not meet this orbit .
a point @xmath152 is a _
rigid garnett point _ for @xmath150 if @xmath152 is garnett with respect to all @xmath169 .
now , if @xmath170 is any fuchsian group , and @xmath152 is a parabolic point for @xmath170 , then for any subgroup @xmath171 admitting a fundamental domain which contains some non - trivial horodisk based at @xmath152 , one can easily show that there is some @xmath167 such that @xmath152 is a dirichlet point for @xmath150 relative to @xmath168 .
thus , the points which determine our infinite ends are not rigid garnett points .
we do not yet know if they can in fact be garnett points .
we discuss the projection of certain horocycles of the teichmller disk of some @xmath4 to @xmath0 .
we discuss the avatars in the setting of @xmath172 of certain elements of the teichmller modular group .
suppose that @xmath4 admits a cylinder @xmath173 in the vertical direction .
consider the basic affine dehn twist on @xmath173 , expressed in standard coordinates simply as @xmath114 , where @xmath174 are respectively the height and width of @xmath173 , see figure 1 . recall that @xmath175 is defined as the _ modulus _ of the cylinder .
this dehn twist acts as the identity on the vertical boundaries of the cylinder ; we can extend by the identity map to obtain a well - defined homeomorphism , @xmath176 , of the underlying topological surface .
but , @xmath176 allows us to give a differentiable structure on the receiving surface ( such that the image of the singularities are again singularities ) ; furthermore , @xmath176 has locally constant derivative .
hence , using @xmath177 we pull - back the translation atlas from @xmath4 to give a new translation structure .
now , each of the various cylinders @xmath178 has image under @xmath176 of equal modulus ; by standard results in complex analysis , see say @xcite , @xmath176 is conformal
. therefore , the complex structure pulled - back by @xmath177 is the same as that of @xmath2 .
thus , we find a mapping of translation surfaces , @xmath179 .
( of course , @xmath180 only when @xmath173 is all of @xmath2 . )
we thus have that @xmath181 , where @xmath68 is the projection .
note that the horocyclic flow @xmath182 restricts to induce a map on @xmath173 which commutes with @xmath176 acting on @xmath183 ; modulo the action of @xmath176 , this map is parametrized by @xmath184 .
now suppose that there are @xmath185 cylinders in the vertical direction , @xmath186 , @xmath187 , @xmath188 .
let @xmath189 denote the modulus , and let @xmath190 denote the map as above , corresponding to the cylinder @xmath191 ; the various @xmath190 commute , as maps on the underlying surface @xmath2 .
similarly , we can define an @xmath192-action on the underlying surface : given @xmath193 , in cylinder @xmath178 , we twist by amount @xmath194 ; outside the collection of cylinders , the action is defined to be trivial .
since this action commutes with the action of the @xmath190 , we find that the projection to @xmath0 of @xmath195 factors through @xmath196 but , @xmath197 is isomorphic to the @xmath185-torus , @xmath198 .
if in fact @xmath23 admits a cylinder decomposition given by the various @xmath178 , then the horocycle @xmath182 lies in @xmath195 .
indeed , the projection of this horocycle to @xmath0 factors through a curve on @xmath197 . to generalize slightly , if @xmath96 denotes the direction of a cylinder decomposition of @xmath4 into @xmath185 cylinders , we have that the projection of the corresponding horocycle factors through a curve on the @xmath185-torus denoted @xmath199 .
commensurability of non - zero real numbers is an equivalence relation ; for each commensurability class of the moduli @xmath189 , there is a corresponding smallest positive value of @xmath105 which induces on each of the corresponding cylinders @xmath178 an integer power of its basic dehn twist .
let @xmath200 be the number of these classes ; for the @xmath201-th class , let @xmath202 be the least common positive integer multiple of the various @xmath203 in this class .
we then let @xmath204 be the composition of the various @xmath205 in this class , where @xmath206 is such that @xmath207 .
again , the projection of the horocycle factors through the torus corresponding to the group generated by the @xmath204 .
we find that in fact our horocycle factors through a curve on the @xmath200-torus @xmath208 . of course
, @xmath96 is a parabolic direction if and only if there is a single commensurability class , in which case : @xmath209 , and the horocycle factors through a circle ; this circle lies on ( a finite component of the sequence which defines the end which is ) the corresponding cusp of @xmath63 . in all other cases ,
there is more than one commensurability class defining @xmath210 , and just as @xmath211 is dense in @xmath212 , so is the image of the horocycle is dense on @xmath213 . for positive @xmath126 ,
the effect of @xmath124 on a vertical cylinder decomposition of @xmath4 is to preserve the number of cylinders , while decreasing heights and increasing widths . indeed ,
@xmath127 acts so as to multiply vertical cylinder moduli by @xmath214 .
therefore , if @xmath96 is a direction of a cylinder decomposition of @xmath4 as above , then @xmath215 has the same dimension as @xmath216 .
note that vertical saddle connection lengths tend to zero as @xmath128 ; indeed ( having begun with vertical saddle connections ) , for @xmath126 sufficiently large , the shortest saddle connections for @xmath123 are in the vertical direction .
but then , @xmath123 leaves every compact subset of @xmath0 .
hence , as @xmath128 , @xmath123 penetrates into an end of @xmath0 , see say @xcite .
in @xcite , we studied an invariant of the geometry at a cusp of a teichmller complex geodesic the multiplicities and relative lengths of saddle connection vectors .
we use the following elementary notion in a similar manner . if @xmath95 lies interior to the sides of a cylinder , then the _ splitting ratio _ of @xmath95 is the ratio of the distance from the left hand side of the cylinder to @xmath95 to the width of the cylinder .
lemma 4 of @xcite can be reworded as the following .
[ irrsplit ] let @xmath16 be a translation surface . if @xmath96 is a completely periodic direction and @xmath161 has irrational splitting ratio for some cylinder of the cylinder decomposition in the direction @xmath96 , then @xmath96 is a fat direction for @xmath217 .
[ splitratiofixed ] splitting ratios are @xmath10 invariant .
* proof * let @xmath4 be a translation surface , and suppose that @xmath218 splits some cylinder with ratio @xmath96 .
for ease of discussion , we may and do assume that the cylinder , say @xmath173 , is a vertical cylinder .
suppose that @xmath30 .
we develop @xmath173 to @xmath27 in the standard fashion .
the ratio @xmath96 is the ratio of the length of the horizontal line segment passing from the left hand side of @xmath173 to @xmath95 to the width of the cylinder . of course , this second length is the length of the horizontal line segment joining the sides of @xmath173 and passing through @xmath95 .
the image of the cylinder by @xmath31 is again a cylinder , say @xmath219 , with obvious development .
the horizontal line segment crossing @xmath173 upon which @xmath95 lies has as its image some segment in @xmath219 ; since @xmath31 acts linearly , the ratio of the length of the initial segment of this image from the appropriate side of the cylinder to the image of @xmath95 to the length the segment is again @xmath96 .
draw the line segment perpendicular to the sides of the cylinder which passes through the image of @xmath95 . due to the equality of the opposite angles
so defined , ratios of corresponding sides of the triangles so formed are equal . a trivial cancellation ,
see figure 2 , shows that the splitting ratio is indeed preserved .
figure 2 . preserving splitting ratios .
we have the immediate consequence .
[ affpressplitt ] affine diffeomorphisms preserve splitting ratios of cylinders .
we distinguish infinite ends by use of the following elementary well - known result in the theory of continued fractions .
we use a flat notation for continued fractions ; thus , for example @xmath220 = a+ \dfrac{1}{b+ \dfrac{1}{c + \alpha}}\,$ ] .
we label the partial quotients of a real number @xmath221 $ ] by @xmath222 , @xmath223 , @xmath187 .
these give rise to the convergents to @xmath96 : @xmath224 $ ] .
let @xmath225 denote the fractional part of a real number : @xmath226 .
[ rotateirrationals ] suppose that @xmath173 and @xmath219 are cylinders lying in tranverse directions which have non - trivial intersection .
if @xmath95 irrationally splits @xmath173 , then the orbit of @xmath95 under the affine dehn twist of @xmath173 meets the cylinder @xmath219 so as to give infinitely many distinct irrational splitting ratios .
* proof * let @xmath96 denote the splitting ratio of @xmath95 with respect to @xmath173 .
by lemma [ splitratiofixed ] , we may normalize such that @xmath173 is in the vertical direction and @xmath219 is in the horizontal direction . let @xmath227 be a rectangle in which our two cylinders meet . let us develop the cylinders such that @xmath95 lies on the @xmath228-axis , and such there are @xmath229 and @xmath230 such that @xmath227 is bounded below by the line @xmath231 and above by @xmath232 .
let us denote by @xmath113 the vertical height @xmath173 .
thus in these coordinates , the orbit of @xmath95 by the vertical dehn twist has heights @xmath233 with @xmath234 .
but , this is exactly the set of @xmath235 .
it suffices for us to show that there are infinitely many distinct ratios @xmath236 which are irrational , and lie between @xmath229 and @xmath237 . for completeness sake , we show this elementary , presumably well - known , result .
let @xmath238 .
now , if @xmath239 is a rational multiple of @xmath240 , then it is necessary that @xmath241 for @xmath242 integers , with @xmath168 , @xmath243 nonzero .
suppose now that @xmath96 is of the form @xmath244 , and there is an @xmath185 such that @xmath245 for nonzero integers @xmath66 and @xmath105 .
we find : @xmath246 since @xmath96 is irrational , it is not linearly related to a rational number .
hence , in particular , we must have that @xmath247 .
that is , @xmath248 ; but then @xmath249 .
this can only occur for a finite number of @xmath185 .
since in the worst case scenario there are only a finite number of orbit elements to avoid , and the orbit of @xmath95 is dense in the circle on the vertical cylinder , we conclude that there are always infinitely many elements of the orbit which irrationally split the horizontal cylinder .
[ connptsforever ] the affine diffeomorphisms of a translation surface preserve the set of connection points . *
proof * suppose that @xmath250 is an affine diffeomorphism and @xmath218 is a connection point .
given a separatrix passing from a singularity to @xmath251 , we apply @xmath252 to find a separatrix passing from some singularity through @xmath95 .
since @xmath95 is a connection point , this separatrix extends to a saddle connection .
but , the image by @xmath40 of this extended separatrix extends our initial separatrix passing through @xmath251 .
since @xmath40 sends saddle connections to saddle connections , we are done .
we use the invariant studied in @xcite to distinguish cusps .
[ infcusps ] let @xmath4 be a translation surface which is a finite ramified covering of a veech surface @xmath16 , with ramification above some non - periodic connection point and possibly the singularities of @xmath16 .
then the teichmller complex geodesic @xmath13 has infinitely many cusps . * proof * the translation structure of @xmath4 arises as the pull - back by a ramified holomorphic map @xmath253 of riemann surfaces , with @xmath254 .
let @xmath95 denote the unique non - periodic connection point in the branch locus .
the veech group @xmath14 is commensurable to the subgroup of @xmath17 formed by the derivatives of the stabilizer in @xmath255 of @xmath95 . by theorem 2 of @xcite , the image of the non - periodic @xmath95 under the action of the affine diffeomorphism group
is dense on @xmath46 .
let @xmath256 be the length of the shortest saddle connection of @xmath46 .
given @xmath257 , we can find an image of @xmath95 under @xmath258 , say @xmath259 , which is within @xmath260 of one of the singularities , say @xmath176 .
there is a unique separatrix of length less than @xmath260 connecting @xmath259 and @xmath176 . by lemma [ connptsforever ]
, we can extend this separatrix to a saddle connection , passing from @xmath176 to a singularity @xmath261 .
( of course , @xmath261 may equal @xmath176 . )
let @xmath40 be an affine diffeomorphism mapping @xmath259 to @xmath95 .
the image by @xmath40 of the above saddle connection is a saddle connection passing through @xmath95 .
since @xmath46 is a veech surface , the direction of any saddle connection is a parabolic direction .
since @xmath95 lies on this saddle connection , @xmath95 is preserved by the dehn twist of either of the cylinders ( in this given direction ) upon which @xmath95 lies .
thus this direction is a parabolic direction for @xmath23 .
consider a single lift to @xmath23 of the saddle connection that we have just determined . this lift is the union of two saddle connections on @xmath23 : the first joins a lift of @xmath262 to a lift of @xmath95 , the second lies on another branch of the ( ramified ) covering space @xmath2 , joining the lift of @xmath95 to a lift of @xmath263 .
( the connection vectors of these saddle connections are parallel ! ) now , the lengths of these saddle connections on @xmath23 are equal to the lengths of the corresponding segments on @xmath46 .
but , the affine diffeomorphism @xmath40 preserves _ ratios _ of segment lengths ; thus the ratio of the lengths of the saddle connections on @xmath23 equals the ratio of the length along the separatrix from @xmath176 to @xmath259 to the length along its continuation from @xmath259 to @xmath261 .
we find that this ratio is less than @xmath264 .
our construction thus allows us to find pairs of parallel connection vectors , in parabolic directions , of @xmath23 that are of arbitrarily small ratio .
however , the number of connection vectors in any given direction is finite .
furthermore , the ratios of the various connection vectors lying in a parabolic direction is invariant under the action of @xmath36 see @xcite .
we conclude that @xmath4 has infinitely many inequivalent parabolic directions .
but , each of these parabolic directions determines a distinct cusp of @xmath19 .
since the teichmller complex geodesic @xmath13 is a generically injective image of this quotient , it too has infinitely many cusps .
we now give the proof of theorem [ inffatboys ] .
thus , we continue with @xmath23 a ramified covering of @xmath46 as above .
the periodic directions on @xmath23 are exactly those given by the periodic , and hence parabolic , directions of @xmath46 .
we show that @xmath23 has many directions which are periodic , but not parabolic .
[ infends ] let @xmath4 be a translation surface which is a finite ramified covering of a veech surface @xmath16 , with ramification above some non - periodic connection point and possibly the singularities of @xmath16 .
then the teichmller complex geodesic @xmath13 has infinitely many infinite ends .
* proof * let @xmath95 be the branch point , thus @xmath95 is a non - periodic connection point of @xmath16 .
there is thus some parabolic direction of @xmath46 for which @xmath95 irrationally splits a cylinder . without loss of generality , we may normalize so that @xmath95 irrationally splits a vertical cylinder and such that the horizontal is a parabolic direction for @xmath16 .
consider the vertical cylinder @xmath173 in which @xmath95 lies , and any horizontal cylinder @xmath219 which meets @xmath173 . by lemma [ rotateirrationals ] , the set of splitting ratios of @xmath219 determined by the orbit of @xmath95 under the vertical affine dehn twist includes an infinite set of irrational numbers .
there is some finite power of this dehn twist which is induced by an element @xmath40 of the affine diffeomorphism group of @xmath16 ; the orbit of @xmath95 under the subgroup generated by this multiply - composed dehn twist still provides an infinite set of distinct irrational splitting ratios . for each image @xmath265 of this latter type
, apply the corresponding inverse @xmath266 . now , @xmath267 ; @xmath268 is a cylinder in a corresponding parabolic direction for @xmath269 ; and , @xmath95 splits @xmath270 with the ratio that @xmath259 splits @xmath219 . of course ,
given any direction , @xmath95 lies interior to at most one cylinder of the possible decomposition in that direction ; thus the splitting ratio defined by @xmath95 in that direction is well - defined .
furthermore , the splitting ratio that @xmath95 gives is constant on directions equivalent under the stabilizer of @xmath95 amongst the affine diffeomorphisms of @xmath16 .
since the @xmath104-orbit of @xmath95 gives infinitely many distinct irrational splitting ratios , we conclude that there are infinitely many fat directions that are inequivalent under the stabilizer of @xmath95 . therefore , @xmath271 has infinitely many infinite ends .
but , @xmath14 is commensurable with @xmath272 ; hence , @xmath271 and @xmath19 admit a common finite degree ( possibly ramified ) covering . since such coverings are balanced with respect to infinite ends , @xmath19 also has infinitely many infinite ends .
but , the teichmller complex geodesic @xmath13 is a generically injective image of @xmath19 , it too must have infinitely many infinite ends .
we next observe that the above proof actually shows a bit more .
[ fatlimits ] with notation as above , let @xmath96 be a fat direction of @xmath4 . then @xmath96 is the limit of @xmath14-inequivalent fat directions of @xmath4 .
* proof * from construction , @xmath96 is a parabolic direction for @xmath16 and the non - periodic connection branch point @xmath95 irrationally splits a cylinder @xmath173 of @xmath16 in this direction .
recall that some power of the dehn twist in this cylinder is induced by an element @xmath40 of the affine diffeomorphism group of @xmath16 .
choose any parabolic direction for @xmath16 that is transverse to @xmath96 .
there is some cylinder @xmath219 in this direction that meets @xmath173 .
the previous proof shows that there is a subsequence of @xmath270 giving cylinders which are split by @xmath95 with distinct irrational splitting ratios . since @xmath40 is parabolic , preserving the direction @xmath96 , the direction of the cylinder @xmath270 tends to @xmath96 as @xmath185 tends to infinity .
thus , @xmath96 is the limit of these inequivalent fat directions for @xmath217 .
finally , commensurability of groups allows us to choose a subsequence of these directions consisting of inequivalent fat directions for @xmath4 .
we now suppose @xmath273 is such that @xmath4 satisfies the hypotheses of theorem [ inffatboys ] ; thus : @xmath16 is a veech surface @xmath16 with distinguished non - periodic connection point @xmath95 , @xmath274 , and the branch locus of @xmath34 is @xmath95 and possibly singularities of @xmath16 .
the case of ramification above multiple non - periodic connection points is similar ; for clarity of exposition , we content ourselves with this main case . by construction , @xmath36 is commensurable to @xmath275 . for clarity of discussion , _ suppose further _ that the marked translation surface @xmath276 has some connection vector which is uniquely realized .
this is easily arranged by choice of @xmath95 . with this added hypothesis ,
proposition 5.6 of @xcite applies to show that @xmath277 .
[ everyaffhasbldcover ] let @xmath278 be a balanced covering of translation surfaces as above . for each @xmath279
, there is some @xmath280 with @xmath281 such that @xmath282 is a balanced translation covering .
* proof * we take @xmath283 such that @xmath40 gives an isomorphism of translation surfaces of @xmath284 with @xmath16 .
let us now excise the singularities from @xmath285 to form @xmath286 , and excise both @xmath251 and the singularities from @xmath46 to form @xmath287 .
we then find that @xmath288 is a covering of topological surfaces .
letting @xmath287 have its induced translation structure from @xmath46 , we can pull this back by @xmath289 to induce a possibly new structure on @xmath290 ; we can complete this flat metric to form some @xmath291 . by construction , @xmath289 gives a balanced translation covering by @xmath280 .
the translation structure of @xmath291 is compatible with a unique complex structure ; this complex structure is also ( after possibly extending over removable singularities ) the completion of the pull - back of the complex structure on @xmath287 , which in turn is simply the restriction of the complex structure on @xmath46 .
but , @xmath40 does not change complex structure , hence the pulled - back complex structure is exactly that of @xmath292 .
therefore , @xmath293 .
[ lineupcovs ] let @xmath278 be a balanced covering of translation surfaces as above .
then for all @xmath294 , there is some @xmath280 with @xmath295 and some @xmath296 such that @xmath282 is a balanced translation covering . * proof * for each @xmath297 , the function @xmath34 itself gives a balanced translation covering @xmath298 .
there is some @xmath299 giving an isomorphism of translation surfaces of @xmath284 with @xmath16 .
we now proceed as in the previous proof .
[ ptgivescover ] with notation as above , let @xmath300 $ ] .
there is a one - to-@xmath200 correspondence between elements of the @xmath301-orbit of @xmath95 and balanced coverings @xmath302 such that @xmath293 .
moreover , if @xmath291 and @xmath303 correspond to @xmath304 , then @xmath305 .
now , @xmath40 is unique modulo the stabilizer of @xmath95 . replacing @xmath40 by @xmath308 , with @xmath309 of derivative @xmath310 , upon proceeding as in the proof of lemma [ everyaffhasbldcover ]
, we find a balanced covering of translation surfaces @xmath311 , with @xmath312 such that @xmath313 of course , if @xmath314 , then @xmath315 reduces to the original mapping @xmath34 ; thus after the further composition with @xmath40 , we then recover the same balanced covering as with @xmath40 by itself .
however , if @xmath316 is a nontrivial coset representative for @xmath36 in @xmath275 , then the reduction to the original mapping can not be achieved : @xmath317 gives a distinct balanced covering of @xmath318 .
since @xmath40 is invertible , @xmath319 pull - backs a translation structure from @xmath320 that is distinct from that pulled - back by @xmath321 . in any case , each balanced covering @xmath322 announces a @xmath291 in the @xmath10-orbit of the translation structure pulled - back by @xmath289 .
furthermore , we certainly have that @xmath293 .
now recall that the number of balanced coverings of @xmath323 is finite and @xmath324 acts so as to permute them , see say @xcite .
but , the pulled - back complex structures in such an orbit are the same : @xmath305 .
the ( coarse ) moduli space of genus @xmath1 curves with @xmath66 marked points is denoted by @xmath325 .
the forgetful map , given by erasing the final marked point , is @xmath326 ; this is also known as the _
universal curve_. see , say , @xcite and @xcite for these notions .
if @xmath280 is of genus @xmath1 and has @xmath66 singularities , then since the @xmath10-action preserves the singularities of @xmath280 , the lift of @xmath327 to @xmath325 is a biholomorphic equivalence . given a balanced covering @xmath328 there is a relationship between @xmath13 and @xmath329 ; when @xmath330 is the union of some singularities and a singleton , then the relationship can be expressed in terms of the universal curve over @xmath331 .
one could discuss the current setting in terms of the generalized hurwitz spaces as presented in wewer s dissertation @xcite .
we give an elementary discussion here ; one may turn to @xcite for a deeper discussion of correspondences between moduli spaces induced by ramified coverings .
we suppose that the genus @xmath113 translation surface @xmath16 has @xmath66 singularities , and consider the universal curve over the lift of @xmath331 to @xmath332 ; the roving @xmath333-st point accounts for the non - singular branch point in coverings related to the initial ramified covering of @xmath44 by @xmath2 . * proof * for general @xmath30
, we can replace @xmath335 by @xmath336 , and apply proposition [ ptgivescover ] to find the @xmath259 in the @xmath337-orbit of @xmath95 that is uniquely associated to @xmath78 .
now , since @xmath338 is conjugated by @xmath31 to @xmath339 , it follows that when @xmath340 are equivalent modulo @xmath341 then the same point of the underlying surface is associated to both @xmath78 and @xmath342 .
of course , for each @xmath343 , the set of these points @xmath259 is dense on the complement to the @xmath66 singularies .
hence , we have a map from the @xmath10-orbit of @xmath4 to a dense subset , @xmath344 , of @xmath173 the universal curve restricted to @xmath331 , see figure 3 . [ endsalign ] with notation as
above , suppose that @xmath136 is an end of @xmath157 .
then the image of @xmath136 in @xmath13 lies within the image of a subset @xmath347 of @xmath344 fibering above a cusp of @xmath348 .
more precisely , if @xmath136 is a cusp , then @xmath347 is homeomorphic to @xmath349 .
if @xmath136 is an infinite end , then @xmath347 is homeomorphic to @xmath350 , with @xmath351 a dense subset of @xmath352 . *
proof * suppose first that @xmath136 is a cusp .
let @xmath353 be the direction on @xmath23 corresponding to @xmath136 as in lemma [ fatptsareinfenddir ] of [ endhypsurf ] ; normalize so that this is the vertical direction .
of course , @xmath96 is a parabolic direction : @xmath4 admits a cylinder composition in this direction , with all cylinders having commensurable inverse moduli .
but , the balanced covering @xmath354 sends cylinders to cylinders .
hence , @xmath96 is also a parabolic direction for @xmath16 , and is such that @xmath95 splits no cylinder irrationally .
that is , either @xmath95 lies on the boundary of some cylinder of the decomposition of @xmath16 in this direction , or else splits some cylinder in a rational manner .
if @xmath95 lies on the boundary of cylinders , then there is equality of sets of moduli of cylinders of the @xmath4 and of moduli of cylinders of the @xmath16-decomposition in this direction .
hence , the minimum value , say @xmath355 , of @xmath356 such that @xmath357 belongs to @xmath341 is also the minimum value such that this is an element of @xmath36 . again since @xmath95 lies on the boundary of @xmath16-cylinders , the action of each @xmath358 fixes @xmath95 as a point on the underlying topological surface . hence
, in accordance with the proof of theorem [ fiberunivcurve ] , @xmath359 with @xmath360 $ ] describes a circle in @xmath344 .
there is some @xmath361 such that the horodisk of index @xmath362 projects to the cuspidal region of @xmath136 ; similarly , there is some @xmath363 such that the horodisk of index @xmath363 for @xmath16 projects to the cuspidal region of the cusp of @xmath331 corresponding to the parabolic element @xmath364 . let @xmath365 .
we find that the image in @xmath13 of the horodisk of index @xmath130 of @xmath136 meets the image of the region @xmath366\,,\ , t > t\,\ } \subset \mathcal c^{0}$ ] .
we indeed have that @xmath347 is homeomorphic to @xmath349 .
if @xmath95 rationally splits some cylinder of @xmath16 , then it is still the case that the single commensurability class of the inverse moduli of the cylinder decompositions of @xmath4 and of @xmath16 agree .
however , the parabolic element @xmath367 of @xmath341 as above is not an element of @xmath36 ; there is some finite power of this element , thus some @xmath368 with @xmath369 minimal , that does belong to @xmath36 .
we find that the horocycle @xmath370 gives an @xmath185-fold cover of the corresponding horcycle of @xmath46 .
we again can let @xmath366\,,\ , t > t\,\ } \subset \mathcal c^{0}$ ] .
finally , suppose that @xmath136 is an infinite end .
we continue with the normalization that the vertical direction on @xmath4 corresponds to @xmath136 .
we have a decomposition into cylinders for @xmath4 , which projects to a decomposition of cylinders for @xmath217 .
there are at least two commensurability classes of inverse moduli of cylinders for the @xmath4-decomposition ( for otherwise @xmath136 would be a cusp ) for no @xmath121 is @xmath371 .
now , @xmath46 is a veech surface , hence this direction of a cylinder decomposition must be parabolic ; there is thus a sole commensurability class of inverse moduli for this cylinder decomposition , and we can define @xmath355 as above . since the balanced covering of @xmath217 has at least two such classes , @xmath95 must irrationally split some cylinder .
but , this introduces at most one new commensurability class ; we find that @xmath4 has exactly two commensurability classes of inverse moduli of cylinders for its cylinder decomposition .
of more importance , the orbit of @xmath95 under powers of @xmath367 is dense in the vertical circle of @xmath16 upon which @xmath95 lies . with @xmath130 as above , we now let @xmath372\,,\ , t > t\,\ } \subset \mathcal c^{0}$ ] .
this is is homeomorphic to @xmath350 , with @xmath351 a dense subset of @xmath352 .
theorem [ thmbigcusp ] now follows each end of @xmath13 must be covered by some end of @xmath19 , but each of these is given by one of finitely many ends : the image of the fibering over one of the finitely many ends of @xmath331 .
p. dbes , _ mthodes topologiques et analytiques en thorie inverse de galois : thorme dexistence de riemann _ , pp . 2741 , in : arithmtique des revtements algbriques - actes du colloque de saint - tienne , smin .
et congr . 5 , soc
france , paris , 2001 .
e. gutkin , p. hubert and t. a. schmidt , _ affine diffeomorphisms of translation surfaces : periodic points , fuchsian groups , and arithmeticity _ , ann . sci .
cole norm .
, 4e ser . , t. 36 ( 2003 ) , 847 - 866 .
d. sullivan , _ on the ergodic theory at infinity of an arbitrary discrete group of hyperbolic motions _ , in : riemann surfaces and related topics : proceedings of the 1978 stony brook conference ( state univ .
new york , stony brook , n.y . , 1978 ) , pp .
465496 , ann . of math . stud . , 97 , princeton univ .
press , princeton , n.j . , 1981 . | veech groups uniformize teichmller geodesic curves in riemann moduli space . recently ,
examples of infinitely generated veech groups have been given .
we show that these can even have infinitely many cusps and infinitely many infinite ends .
we further show that examples exist for which each direction of an infinite end is the limit of directions of inequivalent infinite ends . |
it is now widely accepted that acceleration in supernova - driven shock waves plays an important role in the production of the observed cosmic ray spectrum up to energies of @xmath0ev ( heavens 1984a ; ko 1995a ) , and it is plausible that acceleration in shocks near stellar - size compact objects can produce most of the cosmic radiation observed at higher energies ( jones & ellison 1991 ) . in the generic shock acceleration model , cosmic rays scatter elastically with magnetic irregularities ( mhd waves ) that are frozen into the background ( thermal ) gas ( gleeson & axford 1967 ; skilling 1975 ) . in crossing the shock , these waves experience the same compression and deceleration as the background gas , if the speed of the waves with respect to the gas ( roughly the alfvn speed ) is negligible compared with the flow velocity ( achterberg 1987 ) .
the convergence of the scattering centers in the shock creates a situation where the cosmic rays gain energy systematically each time they cross the shock .
since the cosmic rays are able to diffuse spatially , they can cross the shock many times . in this process , an exponentially small fraction of the cosmic rays experience an exponentially large increase in their momentum due to repeated shock crossings .
the characteristic spectrum resulting from this first - order fermi process is a power - law in momentum ( krymskii 1977 ; bell 1978a , b ; blandford & ostriker 1978 ) .
it was recognized early on that if the cosmic rays in the downstream region carry away a significant fraction of the momentum flux supplied by the incident ( upstream ) gas , then the dynamical effect of the cosmic - ray pressure must be included in order to obtain an accurate description of the shock structure ( axford , leer , & skadron 1977 ) . in this scenario ,
the coupled nonlinear problem of the gas dynamics and the energization of the cosmic rays must be treated in a self - consistent manner .
a great deal of attention has been focused on the `` two - fluid '' model for diffusive shock acceleration as a possible description for the self - consistent cosmic - ray modified shock problem . in this steady - state theory ,
first analyzed in detail by drury & vlk ( 1981 , hereafter dv ) , the cosmic rays and the background gas are modeled as interacting fluids with constant specific heat ratios @xmath1 and @xmath2 , respectively . the coupling between the cosmic rays and the gas is provided by mhd waves , which serve as scattering centers but are otherwise ignored .
the cosmic rays are treated as massless particles , and second - order fermi acceleration due to the stochastic wave propagation is ignored . within the context of the two - fluid model ,
dv were able to demonstrate the existence of multiple ( up to 3 ) distinct dynamical solutions for certain upstream boundary conditions .
the solutions include flows that are smooth everywhere as well as flows that contain discontinuous , gas - mediated `` subshocks . ''
subsequently , multiple solutions have also been obtained in modified two - fluid models that incorporate a source term representing the injection of seed cosmic rays ( ko , chan , & webb 1997 ; zank , webb , & donohue 1993 ) .
only one solution can be realized in a given flow , but without incorporating additional physics one can not determine which solution it will be .
the two - fluid model has been extended to incorporate a quantitative treatment of the mhd wave field by mckenzie & vlk ( 1982 ) and vlk , drury , & mckenzie ( 1984 ) using a `` three - fluid '' approach . during the intervening decades ,
a great deal of effort has been expended on analyzing the structure and the stability of cosmic - ray modified shocks ( see jones & ellison 1991 and ko 1995b for reviews ) .
much of this work has focused on the time - dependent behavior of the two - fluid model , which is known to be unstable to the development of acoustic waves ( drury & falle 1986 ; kang , jones , & ryu 1992 ) and magnetosonic waves ( zank , axford , & mckenzie 1990 ) .
ryu , kang , & jones ( 1993 ) extended the analysis of acoustic modes to include a secondary , rayleigh - taylor instability . in most cases
, it is found that the cosmic - ray pressure distribution is not substantially modified by the instabilities .
the two - fluid theory of dv suffers from a `` closure problem '' in the sense that there is not enough information to compute the adiabatic indices @xmath2 and @xmath1 self - consistently , and therefore they must be treated as free parameters ( achterberg , blandford , & periwal 1984 ; duffy , drury , & vlk 1994 ) .
this has motivated the subsequent development of more complex theories that utilize a diffusion - convection transport equation to solve for the cosmic - ray momentum distribution along with the flow structure self - consistently . in these models , `` seed '' cosmic rays are either advected into the shock region from far upstream , or injected into the gas within the shock itself .
interestingly , kang & jones ( 1990 ) , achterberg , blandford , & periwal ( 1984 ) , and malkov ( 1997a ; 1997b ) found that diffusion - convection theories can also yield multiple dynamical solutions for certain values of the upstream parameters , in general agreement with the two - fluid model .
frank , jones , & ryu ( 1995 ) have obtained numerical solutions to the time - dependent diffusion - convection equation for oblique cosmic - ray modified shocks that are in agreement with the predictions of the steady - state two - fluid model .
these studies suggest that , despite its shortcomings , the two - fluid theory remains one of the most useful tools available for analyzing the acceleration of cosmic rays in shocks waves ( ko 1995b ) . in their approach to modeling the diffusive acceleration of cosmic rays
, dv stated the upstream boundary conditions for the incident flow in terms of the total mach number , @xmath3 and the ratio of the cosmic - ray pressure to the total pressure , @xmath4 where @xmath5 , @xmath6 , @xmath7 , @xmath8 , @xmath9 , and @xmath10 denote the flow velocity of the background gas , the total sound speed , the gas density , and the total , cosmic - ray , and gas pressures , respectively .
we will use the subscripts `` 0 '' and `` 1 '' to denote quantities associated with the far upstream and downstream regions , respectively .
dv described the incident flow conditions by selecting values for @xmath11 and @xmath12 .
once these parameters have been specified , the determination of the flow structure ( and in particular the number of possible solutions ) in the simplest form of the two - fluid model requires several stages of root finding . since the analysis is inherently numerical in nature , the results are usually stated only for specific upstream conditions .
the characterization of the upstream conditions in terms of @xmath11 and @xmath12 employed by dv turns out to be an inconvenient choice from the point of view of finding exact critical relations describing the number of possible flow solutions for given upstream conditions . as an alternative approach , it is possible to work in terms of the individual gas and cosmic - ray mach numbers , defined respectively by @xmath13 where @xmath14 denote the gas and cosmic - ray sound speeds , respectively . according to equation ( 1.1 ) , @xmath15 , and
therefore the mach numbers @xmath16 and @xmath17 are related to the total mach number @xmath18 and the pressure ratio @xmath19 via @xmath20 or , equivalently , @xmath21 since these equations apply everywhere in the flow , the boundary conditions in the two - fluid model can evidently be expressed by selecting values for _ any two _ of the four upstream parameters @xmath22 . in their work ,
dv described the upstream conditions using @xmath23 , whereas ko , chan , & webb ( 1997 ) and axford , leer , & mckenzie ( 1982 ) used @xmath24 .
another alternative , which apparently has not been considered before , is to use the parameters @xmath25 .
although these choices are all equivalent from a physical point of view , we demonstrate below that the set @xmath25 is the most advantageous mathematically because it allows us to derive _ exact _ constraint curves that clearly delineate the regions of various possible behavior in the parameter space of the two - fluid model .
this approach exploits the formal symmetry between the cosmic - ray quantities and the gas quantities as they appear in the expressions describing the asymptotic states of the flow .
the remainder of the paper is organized as follows . in 2
we discuss the transport equation for the cosmic rays and derive the associated moment equation for the variation of the cosmic - ray energy density . in 3 we employ momentum and energy conservation to obtain an exact result for the critical upstream cosmic - ray mach number that determines whether smooth flow is possible for a given value of the upstream gas mach number . in
4 we derive exact critical conditions for the existence of multiple solutions containing a discontinuous , gas - mediated subshock .
the resulting curves are plotted and analyzed for various values of the adiabatic indices @xmath2 and @xmath1 . in 5 we present specific examples of multiple - solution flows that verify the predictions made using our analytical critical conditions .
we conclude in 6 with a general discussion of our results and their significance for the theory of diffusive cosmic - ray acceleration .
the two - fluid model is developed by treating the cosmic rays as a fluid with energy density comparable to that of the background gas , but possessing negligible mass . in this section
we review the basic equations relevant for the two - fluid model . for integrity and clarity of presentation
, we also include re - derivations of a few of the published results concerning the overall shock structure and the nature of the transonic flow . the diffusive acceleration of energetic cosmic rays due to the convergence of scattering centers in a one - dimensional , plane - parallel flow is described by the transport equation ( skilling 1971 ; 1975 ) @xmath26 where @xmath27 is the particle momentum , @xmath28 is the flow velocity of the background gas ( taken to be positive in the direction of increasing @xmath29 ) , @xmath30 is the spatial diffusion coefficient , and the operator @xmath31 expresses the comoving ( lagrangian ) time derivative in the frame of the gas .
equation ( 2.1 ) describes the effects of fermi acceleration , bulk advection , and spatial diffusion on the direction - integrated ( isotropic ) cosmic - ray momentum distribution @xmath32 , which is normalized so that the total number density of the cosmic rays is given by @xmath33 note that equation ( 2.1 ) neglects the second - order fermi acceleration of the cosmic rays that occurs due to stochastic wave propagation , which is valid provided the alfvn speed @xmath34 is much less than the flow velocity @xmath5 , where @xmath35 is the magnetic field strength .
furthermore , equation ( 2.1 ) does not include a particle collision term , and therefore it is not applicable to the background gas , which is assumed to have a thermal distribution .
the momentum , mass , and energy conservation equations for the gas can be written in the comoving frame as @xmath36 respectively , where @xmath37 is the internal energy density of the gas . the expression for @xmath38 in equation ( 2.4 ) implies a purely adiabatic variation of @xmath39 , and therefore it neglects any heating or cooling of the gas due to wave generation or damping .
this adiabatic equation must be replaced with the appropriate rankine - hugoniot jump conditions at a discontinuous , gas - mediated subshock , should one occur in the flow . in the case of a relativistic subshock , the momentum conservation equation for the gas must be modified to reflect the anisotropy of the pressure distribution ( e.g. , kirk & webb 1988 ) . the pressure @xmath9 and the energy density @xmath40 associated with the isotropic cosmic - ray momentum distribution @xmath41 are given by ( duffy , drury , & vlk 1994 ) @xmath42 where @xmath43 denote respectively the kinetic energy , the speed , and the lorentz factor of a cosmic ray with momentum @xmath27 and mass @xmath44 .
although the lower bound of integration is formally taken to be @xmath45 , in practice the cosmic rays are highly relativistic particles , and therefore @xmath41 vanishes for @xmath46 .
if the distribution has the power - law form @xmath47 , then we must have @xmath48 in order to avoid divergence in the integrals for @xmath9 and @xmath40 ( achterberg , blandford , & periwal 1984 ) , although this restriction can be lifted if cutoffs are imposed at high and/or low momentum ( kang & jones 1990 ) .
we can obtain a conservation equation for the cosmic - ray energy density @xmath40 by operating on the transport equation ( 2.1 ) with @xmath49 , yielding @xmath50 where the mean diffusion coefficient @xmath51 is defined by ( duffy , drury , & vlk 1994 ) @xmath52 and the cosmic - ray adiabatic index @xmath1 is defined by ( malkov & vlk 1996 ) @xmath53 note that in deriving equation ( 2.7 ) , we have dropped an extra term that arises via integration by parts because it must vanish in order to obtain finite values for @xmath9 and @xmath40 . the integral expression in equation ( 2.9 )
indicates that @xmath1 must lie in the range @xmath54 .
it also demonstrates that @xmath1 will evolve in response to changes in the shape of the momentum distribution @xmath41 .
the closure problem in the two - fluid model arises because @xmath41 is not calculated at all , and therefore @xmath1 must be imposed rather than computed self - consistently .
the conservation equations can be rewritten in standard eulerian form as @xmath55 @xmath56 @xmath57 where the fluxes of mass , momentum , and total energy are given respectively by @xmath58 @xmath59 @xmath60 the momentum and energy fluxes can be expressed in dimensionless form as @xmath61 @xmath62 where @xmath63 is the asymptotic upstream flow velocity and the dimensionless quantities @xmath64 , @xmath65 , and @xmath66 are defined respectively by @xmath67 note that the definition of @xmath64 implies that the incident flow has @xmath68 .
these relations can be used to rewrite equations ( 1.3 ) for the gas and cosmic - ray mach numbers as @xmath69 in this paper we shall adopt the two - fluid approximation in the form used by dv , and therefore we assume that the adiabatic indices @xmath2 and @xmath1 are each constant throughout the flow .
the assumption of constant @xmath2 is probably reasonable since the background gas is expected to remain thermal and nonrelativistic at all locations .
the assumption of constant @xmath1 is more problematic , since we expect the cosmic ray distribution to evolve throughout the flow in response to fermi acceleration , but it is justifiable if the `` seed '' cosmic rays are already relativistic in the far upstream region .
we also assume that a steady state prevails , so that the fluxes @xmath70 , @xmath71 , and @xmath72 are all conserved . in this case
the quantities @xmath65 and @xmath66 express the pressures of the two species relative to the upstream ram pressure of the gas @xmath73 , where @xmath74 is the asymptotic upstream mass density .
the eulerian frame in which we are working is necessarily the frame of the shock , since that is the only frame in which the flow can appear stationary ( becker 1998 ) . in a steady state ,
the adiabatic variation of @xmath65 implied by equation ( 2.4 ) indicates that along any smooth section of the flow , the gas pressure can be calculated in terms of the velocity using @xmath75 where @xmath76 and @xmath77 denote fiducial quantities measured at an arbitrary , fixed location within the section of interest . according to equation ( 2.19 ) , the associated variation of the gas mach number along the smooth section of the flow
is given by @xmath78 where @xmath79 denotes the gas mach number at the fiducial location .
substituting for @xmath65 in equation ( 2.16 ) using equation ( 2.20 ) and differentiating the result with respect to @xmath29 yields the dynamical equation ( achterberg 1987 ; ko , chan , & webb 1997 ) @xmath80 critical points occur where the numerator and denominator vanish simultaneously .
the vanishing of the denominator implies that @xmath81 at the critical point , and therefore the critical point is also a _ gas sonic point _
( axford , leer , & mckenzie 1982 ) .
the vanishing of the numerator implies that @xmath82 at the gas sonic point .
we can rewrite the dimensionless momentum flux @xmath83 by using equations ( 2.19 ) to substitute for @xmath65 and @xmath66 in equation ( 2.16 ) , yielding @xmath84 similarly , equation ( 2.17 ) for the dimensionless energy flux @xmath85 becomes @xmath86 using equation ( 2.23 ) to eliminate @xmath17 in equation ( 2.24 ) yields for the gradient of the cosmic - ray pressure @xmath87 where @xmath88 along any smooth section of the flow , @xmath89 depends only on @xmath64 by virtue of equation ( 2.21 ) , which gives @xmath16 as a function of @xmath64 . in the two - fluid model , the flow is assumed to become gradient - free asymptotically , so that @xmath90 in the far upstream and downstream regions ( dv ; ko , chan , & webb 1997 ) .
the function @xmath89 must therefore vanish as @xmath91 , and consequently we can express the values of @xmath83 and @xmath85 in terms of the upstream mach numbers @xmath92 and @xmath93 using @xmath94 and @xmath95 where we have also employed the boundary condition @xmath96 .
the critical nature of the dynamical equation ( 2.22 ) implies that @xmath97 at the gas sonic point .
hence if the flow is to pass _ smoothly _ through a gas sonic point , then @xmath89 must vanish at _ three _ locations .
it follows that one of the key questions concerning the flow structure is the determination of the number of points at which @xmath97 .
we can address this question by differentiating equation ( 2.25 ) with respect to @xmath64 , which yields @xmath98 where @xmath99 is the total mach number , and we have used the result @xmath100 implied by equation ( 2.21 ) . for the second derivative of @xmath89
we obtain @xmath101 since the cosmic rays have a higher average lorentz factor than the thermal background gas , @xmath102 ( cf .
2.9 ) , and therefore @xmath103 , implying that @xmath104 is concave down as indicated in figure 1 .
hence there are exactly _ two _ roots for @xmath64 that yield @xmath97 , one given by the upstream velocity @xmath105 and the other given by the downstream velocity , denoted by @xmath106 .
we therefore conclude that if the flow includes a gas sonic point , then the velocity at that point must be either @xmath107 or @xmath108 .
consequently the flow can not pass smoothly through a gas sonic point , as first pointed out by dv .
the flows envisioned here are decelerating , and therefore the high - velocity root @xmath109 is associated with the incident flow . in this case , the fact that @xmath104 is concave - down implies that @xmath110 in the upstream region , and therefore based on equation ( 2.29 ) we conclude that @xmath111 in the upstream region and @xmath112 in the downstream region , with @xmath113 at the peak of the curve where @xmath114 . hence the flow _ must _ contain a sonic transition with respect to the _ total _ sound speed @xmath6 . in this sense ,
the flow is a `` shock '' whether or not it contains an actual discontinuity .
for the flow to decelerate through a shock transition , the total upstream mach number must therefore satisfy the condition @xmath115 .
this constraint also implies that the upstream flow must be supersonic with respect to both the gas and cosmic - ray sound speeds ( i.e. , @xmath116 and @xmath117 ) , since @xmath118 .
furthermore , for a given value of @xmath92 , the upstream cosmic - ray mach number @xmath93 must exceed the minimum value @xmath119 corresponding to the limit @xmath120 . the requirement that @xmath116 forces us to conclude that if a gas sonic point exists in the flow , then it must be identical to the gradient - free asymptotic _ downstream
consequently the flow must either remain supersonic everywhere with respect to the gas , or it must cross a discontinuous , gas - mediated subshock . if @xmath121 everywhere , then the flow is completely smooth and the gas sonic point is `` virtual , '' meaning that it exists in the parameter space , but it does not lie along the flow trajectory . in this case , the gas pressure evolves in a purely adiabatic fashion according to equation ( 2.20 ) , although the total entropy of the combined system ( gas plus cosmic rays ) must increase as the flow crosses the shock , despite the fact that it is smooth . in
3 we derive the critical value for the upstream cosmic - ray mach number @xmath93 that determines whether or not smooth flow is possible for a given value of the upstream gas mach number @xmath92 .
the overall structure of a cosmic - ray modified shock governed by the dynamical equation ( 2.22 ) can display a variety of qualitatively different behaviors , as first pointed out by dv .
depending on the upstream parameters , up to three distinct steady - state solutions are possible , although only one of these can be realized in a given situation .
the possibilities include globally smooth solutions as well as solutions containing a discontinuous , gas - mediated subshock .
smooth flow is expected when the upstream cosmic - ray pressure @xmath122 is sufficiently large since in this case cosmic ray diffusion is able to smooth out the discontinuity . in this section
we utilize the critical nature of the dynamical equation to derive an analytic expression for the critical condition that determines when smooth flow is possible , as a function of the upstream ( incident ) mach numbers @xmath93 and @xmath92 . whether or not the flow contains a discontinuous , gas - mediated subshock
, it must be smooth in the upstream region ( preceding the subshock if one exists ) .
we can therefore apply equation ( 2.21 ) for the variation of @xmath16 in the upstream region , where it is convenient to use the incident parameters @xmath68 and @xmath92 as the fiducial quantities .
the requirement that @xmath123 at the gas sonic point implies that the velocity there is given by @xmath124 which we refer to as the `` critical velocity . ''
if a sonic point exists in the flow , then @xmath125 must correspond to the downstream root of the equation @xmath126 , i.e. , @xmath127 .
note that the value of @xmath125 depends only on @xmath92 , and consequently it is independent of @xmath93 .
the asymptotic states of the flow are assumed to be gradient - free , and the critical conditions associated with the dynamical equation ( 2.22 ) imply that @xmath128 at the gas sonic point .
the constancy of @xmath85 and @xmath83 therefore allows us to link upstream quantities to quantities at the gas sonic point by using equations ( 2.23 ) , ( 2.24 ) , ( 2.27 ) , and ( 2.28 ) to write @xmath129 and @xmath130 respectively , where @xmath131 denotes the value of the cosmic - ray mach number at the gas sonic point .
if we substitute for @xmath125 using equation ( 3.1 ) and eliminate @xmath131 by combining equations ( 3.2 ) and ( 3.3 ) , we can solve for @xmath93 to obtain an _ exact expression _ for the critical upstream cosmic - ray mach number required for the existence of a sonic point in the asymptotic downstream region , @xmath132 \gamma_c - \left[\left({r_0 ^ 2 \over 2 } - { 1 \over 2 } - { 1 \over \gamma_g-1 } \right ) m_{g0}^2 + { r_0 ^ 2 \over \gamma_g - 1 } \right ] \left(\gamma_c - 1 \right ) \over r_0 \ , ( r_0 - 1 ) \ , m_{g0}^2 } \right\}^{-1/2 } \ , , \eqno(3.4)\ ] ] where @xmath133 note that @xmath134 is an explicit function of the upstream gas mach number @xmath92 .
the interpretation is that if @xmath135 for a given value of @xmath92 , then the flow is everywhere supersonic with respect to the gas sound speed @xmath136 except in the far - downstream region , where it asymptotically approaches the gas sonic point .
surprisingly , this simple solution for @xmath134 has apparently never before appeared in the literature , probably due to the fact that the analytical form is lost when one works in terms of the alternative parameters @xmath11 and @xmath12 employed by dv .
this can be clearly demonstrated by using the expressions @xmath137 to substitute for @xmath92 and @xmath93 in equations ( 3.4 ) and ( 3.5 ) and then attempting to solve the resulting equation for either @xmath12 or @xmath11 .
it is easy to convince oneself that it is not possible to express either of these quantities explicitly in terms of the other . in figure 2
we depict the curve @xmath135 in the @xmath25 parameter space using equations ( 3.4 ) and ( 3.5 ) for various values of @xmath2 and @xmath1 .
we have determined that smooth flow into an asymptotic downstream gas sonic point is possible if @xmath135 .
however , in order to obtain a complete understanding of the significance of the critical upstream cosmic - ray mach number @xmath134 , we must determine the nature of the flow when @xmath138 .
the resulting flow structure can be analyzed by perturbing around the state @xmath135 by taking the derivative of the asymptotic downstream velocity @xmath108 with respect to @xmath93 , holding @xmath92 constant .
the fact that @xmath92 is held fixed implies that the critical velocity @xmath125 also remains unchanged by virtue of equation ( 3.1 ) . upon differentiating ,
we obtain @xmath139^{-1 } \ , , \eqno(3.7)\ ] ] which is always negative since the flow decelerates , and therefore @xmath140 .
this indicates that if @xmath93 is _ decreased _ from the value @xmath141 for fixed @xmath92 , then the downstream velocity @xmath108 _ increases _ above the critical velocity @xmath125 , and therefore the flow is everywhere supersonic with respect to the gas sound speed . in this case
there is no gas sonic point in the flow , and consequently a globally smooth solution is possible .
conversely , when @xmath142 , a gas sonic point exists in the flow , and therefore the flow can not be globally smooth because that would require smooth passage through a gas sonic point , which we have proven to be impossible . in this case
, the flow _ must _ pass through a discontinuous , gas - mediated subshock .
we conclude that globally smooth flow is possible in the region below each of the critical curves plotted in figure 2 .
note that in each case there is a critical value for @xmath92 , denoted by @xmath143 , to the right of which smooth flow is _ always _
possible for _ any _ value of @xmath93 .
this critical value is the solution to the equation @xmath144 \gamma_c - \left[\left({r_a^2 \over 2 } - { 1 \over 2 } - { 1 \over \gamma_g-1 } \right ) m_{ga}^2 + { r_a^2 \over \gamma_g - 1 } \right ] \left(\gamma_c - 1 \right ) = 0 \ , , \eqno(3.8)\ ] ] where @xmath145 corresponding to the limit @xmath146 in equation ( 3.4 ) .
we plot @xmath143 as a function of @xmath2 and @xmath1 in figure 3 .
when @xmath147 and @xmath148 , we find that @xmath149 , in agreement with the numerical results of dv and heavens ( 1984b ) .
dv discovered that two new discontinuous solutions become available in addition to either a smooth solution or another discontinuous solution when the upstream total mach number @xmath11 is sufficiently large .
subsequent authors have confirmed the existence of multiple dynamical solutions within the context of the two - fluid model ( achterberg , blandford , & periwal 1984 ; axford , leer , & mckenzie 1982 ; kang & jones 1990 ; ko , chan , & webb 1997 ; zank , webb , & donohue 1993 ) .
however , most of this work utilized numerical root - finding procedures and therefore it fails to provide much insight into the structure of the critical conditions that determine when multiple solutions are possible .
we revisit the problem in this section by recasting the upstream boundary conditions using the same approach employed in 3 .
in particular , we show that when the incident flow conditions are stated in terms of the upstream gas and cosmic - ray mach numbers @xmath92 and @xmath93 , respectively , it is possible to obtain exact , analytical formulae for the critical curves that form the border of the region of multiple solutions in the parameter space . the existence of multiple dynamical solutions is connected with the presence in the flow of a discontinuous subshock mediated by the pressure of the gas .
we can therefore derive critical criteria related to the multiple - solution phenomenon by focusing on the nature of the flow in the post - subshock region , assuming that a subshock exists in the flow .
as we demonstrate in the appendix , the energy , momentum , and particle fluxes for the cosmic rays and the background gas are independently conserved as the flow crosses the subshock .
this implies that the quantities associated with the gas satisfy the usual rankine - hugoniot jump conditions , as pointed out by dv .
far downstream from the subshock , the flow must certainly relax into a gradient - free condition if a steady state prevails .
in fact , it is possible to demonstrate that the entire post - subshock region is gradient - free , so that @xmath150constant downstream from the subshock .
this has already been shown by dv using a geometrical approach , but it can also be easily established using the following simple mathematical argument . first we combine the dynamical equation ( 2.22 ) with the definition of @xmath104 given by equation ( 2.25 ) to obtain the alternative form @xmath151 in the post - subshock gas , @xmath152 and therefore @xmath112 regardless of the value of @xmath17 , since @xmath153 .
it follows from equation ( 2.29 ) that @xmath154 downstream from the subshock . referring to figure 4
, we wish to prove that the subshock transition must take the velocity @xmath64 directly to the gradient - free asymptotic downstream root denoted by @xmath108 , so that @xmath155 on the immediate downstream side of the subshock . to develop the proof ,
let us suppose instead that @xmath156 in the post - subshock gas , corresponding to the post - subshock velocity @xmath157 in figure 4 .
in this case , equation ( 4.1 ) implies that @xmath158 in the downstream region , so that @xmath64 increases along the flow direction , evolving _
away _ from the gradient - free root @xmath108 in the post - subshock flow .
conversely , if @xmath159 in the post - subshock gas ( corresponding to the velocity @xmath160 in figure 4 ) , then @xmath161 in the downstream region and consequently @xmath64 decreases , again evolving away from the gradient - free root @xmath108 . hence
if the flow is to be stationary , then @xmath64 must jump _ directly _ to the value @xmath108 , and the entire post - subshock region must therefore be gradient - free .
this conclusion is valid within the context of the `` standard '' two - fluid model studied by dv , but zank , webb , & donohue ( 1993 ) suggest that it may be violated in models including injection .
we can derive an expression for @xmath104 suitable for use in the post - subshock region by employing equation ( 2.21 ) to eliminate @xmath16 in equation ( 2.25 ) .
this yields @xmath162 where we have adopted the immediate post - subshock values @xmath163 and @xmath164 as the fiducial quantities in the smooth section of the flow downstream from the subshock .
since we have already established that the post - subshock flow is gradient - free , we can obtain an equation satisfied by the post - subshock velocity @xmath163 by writing @xmath165 the gradient - free nature of the post - subshock flow also trivially implies @xmath166 where @xmath167 is the asymptotic downstream gas mach number associated with the downstream velocity @xmath108 .
equation ( 4.3 ) can be interpreted as a relation for the immediate _ pre_-subshock velocity @xmath168 by utilizing the standard rankine - hugoniot jump conditions ( landau & lifshitz 1975 ) @xmath169 where the pre - subshock gas mach number @xmath170 is related to @xmath168 and @xmath92 via @xmath171 which follows from equation ( 2.21 ) . by using equations ( 4.5 ) and ( 4.6 ) to eliminate @xmath163 , @xmath164 , and @xmath170 ,
we can transform equation ( 4.3 ) into a new equation for the pre - subshock velocity @xmath168 , which we write symbolically as @xmath172 where @xmath173 @xmath174 \ , \left(\gamma_g - 1 + 2 \ , m_{g0}^{-2 } \ ,
u_-^{-1-\gamma_g } \over \gamma_g + 1 \right ) - { \cal e } \ . \eqno(4.8)\ ] ] recall that the constants @xmath83 and @xmath85 are functions of @xmath93 and @xmath92 by virtue of equations ( 2.27 ) and ( 2.28 ) .
in addition to satisfying the condition @xmath175 , acceptable roots for the pre - subshock velocity @xmath168 must also exceed the critical velocity @xmath125 . this is because the flow must be supersonic with respect to the gas before crossing the subshock , if one exists .
equation ( 4.7 ) allows us to solve for the pre - subshock velocity @xmath168 as a function of the upstream cosmic - ray and gas mach numbers @xmath93 and @xmath92 , respectively . in general
, @xmath168 is a multi - valued function of @xmath93 and @xmath92 , and this results in the possibility of several distinct subshock solutions in certain regions of the parameter space .
fortunately , additional information is also available that can be utilized to calculate the shape of the critical curve in @xmath25 space bordering the region of multiple subshock solutions .
the nature of this information becomes clear when one examines the topology of the function @xmath176 as depicted in figures 5 and 6 for @xmath177 and @xmath178 .
we consider a sequence of situations with @xmath92 held fixed and @xmath93 gradually increasing from the minimum value @xmath179 given by equation ( 2.32 ) , corresponding to the limit @xmath120 .
note that since @xmath92 is held constant , the critical velocity @xmath125 also remains constant according to equation ( 3.1 ) .
two qualitatively different behaviors are observed depending on whether or not @xmath92 exceeds @xmath143 , where @xmath180 is the critical upstream gas mach number for smooth flow calculated using equation ( 3.8 ) with @xmath177 and @xmath178 . in figure 5_a
_ we plot @xmath181 as a function of @xmath168 and @xmath93 for @xmath182 , which yields for the critical velocity @xmath183 . in this case
the minimum upstream cosmic - ray mach number is @xmath184 .
note that initially , for small @xmath93 , there is one root for @xmath168 corresponding to the single crossing of the line @xmath175 .
since this root does not exceed @xmath125 , a subshock solution is impossible and instead the flow must be globally smooth as discussed in 3 .
the choice @xmath182 satisfies the condition @xmath185 , and therefore as @xmath93 increases , the root for @xmath168 eventually equals @xmath125 , which occurs when @xmath186 . in this case
the subshock is located at the asymptotic downstream limit of the flow , and therefore it is identical to the gas sonic point .
equations ( 3.1 ) and ( 4.6 ) indicate that the pre - subshock gas mach number @xmath187 as expected . as we continue to increase @xmath93 beyond the critical value @xmath134 in figure 5_a _ , the root for @xmath168
exceeds @xmath125 , and therefore the smooth solution is replaced by a subshock solution .
we refer to this solution as the `` primary '' subshock solution .
since @xmath188 in this region of the parameter space , the subshock plays a significant role in modifying the flow . if @xmath93 is increased further , the primary root for @xmath168 increases slowly , the concave - down shape changes , and a new peak begins to emerge at large @xmath168 .
the peak touches the line @xmath175 when @xmath189 , and therefore at this point a new subshock root appears with the value @xmath190 .
the development of this new root can be clearly seen in figure 5_b _ , where we replot @xmath181 at a much smaller scale .
as @xmath93 continues to increase , the peak continues to rise , and the new @xmath168 root bifurcates into two roots . since the two new roots for @xmath168 are larger than the primary root , the corresponding pre - subshock gas mach numbers are also larger , and therefore the two new subshocks are stronger than the primary subshock .
we conclude that in this region of the @xmath25 parameter space , _ three _ discontinuous subshock solutions are possible , although only one can occur in a given situation .
another example is considered in figure 6_a _ , where we plot @xmath181 as a function of @xmath168 and @xmath93 for @xmath191 , which yields @xmath192 and @xmath193 . in this case , @xmath194 , and consequently there is _ always _ a root for @xmath168 below @xmath125 , indicating that globally smooth flow is possible for all values of @xmath93 .
hence the `` primary '' subshock solution never appears in this example .
however , for sufficiently large @xmath93 , a peak develops in @xmath181 just as in figure 5_b_. this peak rises with increasing @xmath93 and eventually crosses the line @xmath195 at @xmath196 when @xmath197 , corresponding to the appearance of a new physically acceptable subshock root for @xmath168 .
this new root bifurcates into two roots as @xmath93 is increased , which can be clearly seen in figure 6_b _ , where @xmath181 is replotted on a much smaller scale .
it follows that in this region of @xmath25 space , two discontinuous solutions are possible in addition to a single globally smooth solution .
it is apparent from figures 5 and 6 that the onset of multiple solutions is connected with the vanishing of @xmath181 coupled with the additional , simultaneous condition @xmath198 which supplements equation ( 4.7 ) .
equations ( 4.7 ) and ( 4.9 ) can be manipulated to obtain explicit expressions for the critical upstream gas and cosmic - ray mach numbers corresponding to the onset of multiple subshock solutions as functions of the pre - subshock velocity @xmath168 .
these critical mach numbers are denoted by @xmath199 and @xmath200 , respectively .
the logical procedure for obtaining the relations is straightforward , although the algebra required is somewhat tedious .
the first step in the process is to solve equations ( 4.7 ) and ( 4.9 ) individually to derive two separate expressions for @xmath200 .
equation ( 4.7 ) yields @xmath201 where @xmath202 @xmath203 \ , u_-^{2 \gamma_g } \ , , \eqno(4.10{\rm c})\ ] ] @xmath204 @xmath205 @xmath206 \ , , \eqno(4.10{\rm e})\ ] ] @xmath207 \ , .
\eqno(4.10{\rm f})\ ] ] the solution to equation ( 4.9 ) is given by @xmath208 where @xmath209 @xmath210 @xmath211 @xmath212 \ , , \eqno(4.11{\rm e})\ ] ] @xmath213 \ , .
\eqno(4.11{\rm f})\ ] ] the similarity of the dependences on @xmath199 in equations ( 4.10 ) and ( 4.11 ) for @xmath200 suggests that we can derive an exact solution for @xmath199 as a function of @xmath168 by equating these two expressions .
the result obtained is @xmath214 where @xmath215 @xmath216 \ , u_- \bigg\ } \ , , \eqno(4.12{\rm b})\ ] ] @xmath217 @xmath218 \ , , \eqno(4.12{\rm d})\ ] ] @xmath219 @xmath220 @xmath221 @xmath222 equation ( 4.12 ) can be used to evaluate the critical upstream gas mach number @xmath199 as a function of the pre - subshock velocity @xmath168 .
once @xmath199 is determined , we can calculate the corresponding critical upstream cosmic - ray mach number @xmath200 using either equation ( 4.10 ) or equation ( 4.11 ) , which both yield the same result .
hence equations ( 4.10 ) , ( 4.11 ) , and ( 4.12 ) provide a direct means for calculating @xmath199 and @xmath200 as exact functions of @xmath168 . as an example , setting @xmath177 , @xmath178 , and @xmath190 yields @xmath223 and @xmath224 , in agreement with the results plotted in figure 5 . likewise , setting @xmath177 , @xmath225 , and @xmath196 yields @xmath226 and @xmath227 , in agreement with figure 6 .
although our expressions for @xmath199 and @xmath200 simplify considerably in the special case @xmath177 , @xmath178 , we have chosen to derive results valid for general ( but constant ) values of @xmath2 and @xmath1 in order to obtain a full understanding of the sensitivity of the critical mach numbers to variations in the adiabatic indices .
while admittedly somewhat complex , these equations can nonetheless be evaluated using a hand calculator , and replace the requirement of utilizing the root - finding techniques employed in most previous investigations of the multiple - solution phenomenon in the two - fluid model .
in figure 7 we plot the critical curves for the occurrence of multiple solutions using various values of @xmath2 and @xmath1 .
this is accomplished by evaluating @xmath199 and @xmath200 as parametric functions of the pre - subshock velocity @xmath168 using equation ( 4.12 ) along with either equation ( 4.10 ) or ( 4.11 ) .
the critical curves denote the boundary of the wedge - shaped _ multiple - solution region_. inside this region , two new subshock solutions are possible , along with either the primary subshock solution or the globally smooth solution . outside this region ,
the flow is either described by the primary subshock solution or else it is globally smooth .
the lower - left - hand corner of the multiple - solution region curves to the left and culminates in a sharp cusp .
the presence of the cusp implies that there is a minimum value of @xmath92 , below which multiple solutions are never possible for any value of @xmath93 .
note that the multiple - solution region becomes very narrow as the values of @xmath2 and @xmath1 approach each other , suggesting that the physically allowed solutions converge on a single solution as the thermodynamic properties of the two populations of particles ( background gas and cosmic rays ) become more similar to each other . in the limit @xmath228 , multiple solutions are not allowed at all , and the single available solution is either smooth or discontinuous depending on the values of @xmath92 and @xmath93 . in figure 8 we plot all of the various critical curves for the physically important case of an ultrarelativistic cosmic - ray distribution ( @xmath178 ) combined with a nonrelativistic background gas ( @xmath177 ) .
this is the most fully self - consistent example of the two - fluid model , since in this case we expect that the adiabatic indices will remain constant throughout the flow , as is assumed in the model ( see eq .
[ 2.9 ] ) .
the area of overlap between the multiple - solution region and the smooth - solution region implies the existence of four distinctly different domains within the @xmath25 parameter space . within domain
i , which lies outside both the multiple - solution region and the smooth - solution region , the flow must be discontinuous , with exactly one ( i.e. , the primary subshock ) solution available .
domain ii lies inside the multiple - solution region and outside the smooth - solution region , and therefore within this area of the parameter space , _ three _ distinct subshock solutions are possible , while globally smooth flow is impossible .
domain iii is formed by the intersection of the multiple - solution region and the smooth - solution region , and therefore two discontinuous subshock solutions are available as well as one globally smooth solution .
finally , domain iv lies outside the multiple - solution region and within the smooth - flow region , and therefore one globally smooth flow solution is possible .
the existence of domain iv is consistent with our expectation that smooth flow will occur for sufficiently large values of the upstream cosmic - ray pressure @xmath122 due to diffusion of the cosmic rays .
if we consider a trajectory through the @xmath25 parameter space that crosses into the multiple - solution region , then the sequence of appearance of the subshock roots for @xmath168 depends on whether the boundary is crossed through the `` top '' or `` bottom '' arcs of the wedge .
we illustrate this phenomenon for @xmath177 and @xmath178 in figure 8 , where we consider two possible paths approaching the point @xmath8 inside the multiple - solution region , starting outside at either points @xmath19 or @xmath229 .
the two paths cross the multiple - solution boundary on different sides of the wedge . in figure 9
we plot @xmath176 along the segment @xmath230 ( with @xmath231 ) , and in figure 10 we plot @xmath176 along the segment @xmath232 ( with @xmath233 ) .
note that the primary subshock root for @xmath168 already exists at points @xmath19 and @xmath229 since they both lie inside domain i. when the lower section of the multiple - solution boundary is crossed along segment @xmath230 ( @xmath234 in fig .
9 ) , two new roots for @xmath168 appear at larger values than the primary root , which is the same sequence observed in figure 5 . however , when the _ upper _ section of the boundary is crossed along segment @xmath232 ( @xmath235 in fig .
10 ) , the two new roots for @xmath168 appear at _ smaller _ values than the primary root .
hence the order of appearance of the subshock roots for @xmath168 is different along each path . despite this path dependence ,
the actual set of roots obtained at point @xmath8 is the same regardless of the approach path taken , and therefore there is no ambiguity regarding the possible subshock solutions available at any point in the @xmath25 parameter space . in figure 11 we present summary plots that include all of the various critical curves derived in 3 and 4 for several different values of @xmath2 and @xmath1 .
note that the area of overlap between the multiple - solution region and the smooth - solution region observed when @xmath177 and @xmath178 rapidly disappears when the difference between @xmath2 and @xmath1 is reduced , due to the increasing similarity between the thermodynamic properties of the cosmic rays and the background gas .
ko , chan , & webb ( 1997 ) and bulanov & sokolov ( 1984 ) have obtained similar parameter space plots depicting the critical curves for smooth flow and for the onset of multiple solutions .
the parameters used to describe the incident flow conditions are @xmath24 in the case of ko , chan , & webb ( 1997 ) and @xmath23 for bulanov & sokolov ( 1984 ) , where @xmath12 and @xmath11 refer to the incident pressure ratio and total mach number , respectively ( see eqs . [ 1.1 ] and [ 1.2 ] ) .
when these parameters are employed directly instead of the quantities @xmath25 utilized in our work , the critical curves must be determined by numerical root - finding .
however , if desired , equations ( 1.5 ) can be used to transform our exact solutions for the critical curves from the @xmath25 space to the @xmath24 and @xmath23 spaces .
this procedure was used to obtain the results depicted in figure 12 , which are identical to those presented by ko , chan , & webb ( 1997 ) and bulanov & sokolov ( 1984 ) .
our interpretation of the various critical mach numbers derived in
3 and 4 can be verified by performing specific calculations of the flow dynamics for a few different values of the upstream mach numbers @xmath92 and @xmath93 .
we shall focus here on cases with @xmath236 and @xmath178 , but this is not essential .
along smooth sections of the flow , we can calculate the velocity profile by integrating the dynamical equation obtained by combining equations ( 2.21 ) and ( 4.1 ) , which yields @xmath237^{-1 } \ , \eqno(5.1)\ ] ] where @xmath238 and we have introduced the new spatial variable @xmath239 the quantities @xmath79 and @xmath77 are fiducial parameters measured at an arbitrary location along the smooth section of interest , and the constants @xmath83 and @xmath85 are functions of the incident mach numbers @xmath92 and @xmath93 via equations ( 2.27 ) and ( 2.28 ) . as discussed in 3 , if @xmath240 , then the flow is _ everywhere _ supersonic with respect to the gas , and therefore we can use equation ( 5.1 ) to describe the structure of the _ entire _ flow by setting the fiducial parameters @xmath79 and @xmath77 equal to the asymptotic upstream quantities @xmath92 and @xmath68 , respectively .
conversely , if @xmath142 , then the flow must contain a discontinuous , gas - mediated subshock . in this case
the upstream region is governed by equation ( 5.1 ) with @xmath241 and @xmath242 , and the downstream region is governed by equation ( 5.1 ) with @xmath243 and @xmath244 .
the pre - subshock and post - subshock regions are linked by the jump conditions given by equations ( 4.5 ) , which determine @xmath164 and @xmath163 and yield @xmath97 in the entire downstream region as expressed by equation ( 4.3 ) . in order to illustrate the various possible behaviors predicted by our critical conditions , we shall present one example from each of the four domains discussed in iv ( see fig . 8) calculated using @xmath177 and @xmath178 . in figure 13
we plot the velocity profile @xmath245 obtained by integrating the dynamical equation ( 5.1 ) with @xmath246 and @xmath247 . according to figure 8 , this point lies within domain i , and therefore we expect to find a single subshock solution , and no globally smooth solution .
equation ( 2.25 ) confirms that in this case the downstream root for @xmath104 is @xmath248 , and therefore smooth flow is impossible since this does not exceed the critical velocity @xmath249 .
analysis of equation ( 4.7 ) yields a single acceptable value for the pre - subshock velocity root @xmath250 , with an associated pre - subshock gas mach number @xmath251 .
figure 13 also includes plots of the gas mach number @xmath252 obtained using equation ( 2.21 ) , the dimensionless gas pressure @xmath253 obtained using equation ( 2.20 ) , and the dimensionless cosmic - ray pressure @xmath254 obtained using equation ( 2.16 ) .
the first three quantities exhibit a jump at the subshock , whereas the cosmic - ray pressure is continuous since a jump in this quantity would imply an infinite energy flux .
the overall compression ratio is @xmath255 and the pressures increase by the factors @xmath256 and @xmath257 between the asymptotic upstream and downstream regions .
recall that @xmath65 and @xmath258 express the pressures of the two species divided by the upstream ram pressure of the gas . hence in this example the two species each absorb comparable fractions of the upstream ram pressure .
note that the increase in the cosmic - ray pressure is entirely due to the smooth part of the transition upstream from the discontinuous subshock , while the gas pressure experiences most of its increase in crossing the subshock .
in figure 14 we set @xmath259 and @xmath260 , corresponding to domain ii in figure 8 .
we therefore expect to find three distinct , physically acceptable subshock solutions associated with these parameter values .
analysis of equation ( 2.25 ) verifies that no smooth solutions are possible in this case because the downstream root @xmath261 is less than the critical velocity @xmath262 .
equation ( 4.7 ) indicates that the three acceptable roots for the pre - subshock velocity are @xmath263 with associated pre - subshock gas mach numbers @xmath264 respectively .
the gas and cosmic - ray pressures increase by the factors @xmath265 and @xmath266 and the overall compression ratios are @xmath267 .
note that the solution with the _ largest _ cosmic - ray pressure has the _ largest _ overall compression ratio and the _ weakest _ subshock .
this is due to the fact that the cosmic - ray pressure is amplified in the smooth - flow region upstream from the subshock , and this region is most extended when the flow does not encounter a subshock until the smallest possible value of @xmath170 .
conversely , the solution with the largest gas pressure is the one with the strongest subshock and the smallest overall compression ratio . in figure 15
we use the upstream parameters @xmath268 and @xmath269 .
this point lies within domain iii in figure 8 , and therefore we expect to find two distinct subshock solutions along with one globally smooth solution .
equation ( 2.25 ) confirms that in this case a smooth solution is possible since the downstream root @xmath270 exceeds the critical velocity @xmath271 .
analysis of equation ( 4.7 ) yields two acceptable values for the pre - subshock velocity given by @xmath272 .
the corresponding pre - subshock gas mach numbers are @xmath273 respectively . in the discontinuous ( subshock )
solutions the gas and cosmic - ray pressures increase by the factors @xmath274 and @xmath275 and the overall compression ratios are @xmath276 .
the subshocks in this example are strong , and therefore in the solutions containing a subshock most of the deceleration is due to the buildup of the pressure of the background gas , rather than the cosmic - ray pressure .
hence both of the subshock solutions are gas - dominated .
in the globally smooth solution , which is cosmic - ray dominated , the pressures increase by the factors @xmath277 and @xmath278 , and the compression ratio is @xmath279 . in this case
the cosmic rays absorb almost all of the ram pressure of the upstream gas .
finally , in figure 16 we set @xmath280 and @xmath281 .
according to figure 8 , this point lies within domain iv , and therefore we expect that only a single , globally - smooth solution exists for these upstream parameters .
this prediction is verified by equation ( 4.7 ) , which confirms that no acceptable subshock roots for @xmath168 exists .
furthermore , equation ( 2.25 ) indicates that smooth flow is possible since the downstream root @xmath282 exceeds the critical velocity @xmath283 .
hence the only acceptable solution is globally smooth , with the pressure increases given by @xmath284 and @xmath285 .
the overall compression ratio is @xmath286 and the flow is cosmic - ray dominated .
in this paper we have obtained a number of new analytical results describing the critical behavior of the two - fluid model for cosmic - ray modified shocks .
it is well known that in this model , up to three distinct solutions are possible for a given set of upstream boundary conditions .
the behaviors of the various solutions can be quite diverse , including flows that are smooth everywhere as well as flows that contain a discontinuous , gas - mediated subshock .
the traditional approach to the problem of determining the types of possible solutions , employed by ko , chan , & webb ( 1997 ) and bulanov & sokolov ( 1984 ) , is based on stating the upstream boundary conditions in terms of the incident total mach number @xmath11 and the incident pressure ratio @xmath12 ( see eqs . [ 1.1 ] and [ 1.2 ] ) . in this approach
the determination of the available solution types requires several steps of root - finding , and there is no possibility of obtaining analytical expressions for the critical relationships .
the analysis presented here utilizes a fresh approach based upon a new parameterization of the boundary conditions in terms of the upstream gas and cosmic - ray mach numbers @xmath92 and @xmath93 , respectively .
the analytical results we have obtained in 3 and 4 for the critical upstream mach numbers expressed by equations ( 3.4 ) , ( 4.11 ) , and ( 4.12 )
provide for the first time a systematic classification of the entire parameter space for the two - fluid model , which remains one of the most powerful and practical means available for studying the problem of cosmic - ray modified shocks .
these expressions eliminate the need for complex root - finding procedures in order to understand the possible flow dynamics for a given set of upstream boundary conditions , and are made possible by the symmetry between the gas and cosmic - ray parameters as they appear in the expressions describing the asymptotic upstream and downstream states of the flow .
we have compared our quantitative results with those of ko , chan , & webb ( 1997 ) and bulanov & sokolov ( 1984 ) , and they are found to be consistent .
the results are valid for arbitrary ( but constant ) values of the gas and cosmic - ray adiabatic indices @xmath2 and @xmath1 , respectively . in 5
we have presented numerical examples of flow structures obtained in each of the four parameter space domains defined in figure 8 .
these examples verify the predictions made using our new expressions for the critical mach numbers , and confirm that the largest overall compression ratios are obtained in the globally - smooth , cosmic - ray dominated cases .
the existence of multiple distinct solutions for a single set of upstream boundary conditions demands that we include additional physics in order to determine which solution is actually realized in a given situation .
this question has been addressed by numerous authors using various forms of stability analysis as well as fully time - dependent calculations .
dv speculated that when three distinct solutions are allowed ( in domains ii and iii of the parameter space plotted in fig .
8) , the solution with the intermediate value of the cosmic - ray pressure @xmath66 will be unstable .
the argument is based on the idea that if the cosmic - ray pressure were to increase slightly due to a small perturbation , then the gas would suffer additional deceleration , leading to a further increase in the cosmic - ray pressure .
this nonlinear process would drive the flow towards the steady - state solution with the largest value of @xmath66 .
conversely , a small decrease in the cosmic - ray pressure would decrease the deceleration , leading to a smaller value for the cosmic - ray pressure . in this case
the flow would be driven towards the steady - state solution with the smallest value of @xmath66 .
recently , mond & drury ( 1998 ) have suggested that this type of behavior may be realized as a consequence of a corrugational instability .
other authors ( e.g. , drury & falle 1986 ; kang , jones , & ryu 1992 ; zank , axford , & mckenzie 1990 ; ryu , kang , & jones 1993 ) have argued that the globally smooth , cosmic - ray dominated solutions are unstable to the evolution of mhd waves in certain situations .
jones & ellison ( 1991 ) suggest that even when formally stable , the smooth solutions may not be realizable in nature . on the other hand , donohue , zank , & webb ( 1994 ) report time - dependent simulations which seem to indicate that the smooth
, cosmic - ray dominated solution is indeed the preferred steady - state solution in certain regions of the parameter space .
hence , despite the fact that much effort has been expended in analyzing the stability properties of cosmic - ray modified shocks , there is still no clear consensus regarding which of the steady - state solutions ( if any ) is stable and therefore physically observable for an arbitrary set of upstream conditions . in light of the rather contradictory state of affairs regarding the stability properties of the various possible dynamical solutions , we propose a new form of _ entropy - based _ stability analysis . in this method ,
the entropy of the cosmic - ray distribution is calculated by first solving the transport equation ( 2.1 ) for the cosmic - ray distribution @xmath41 and then integrating to obtain the boltzmann entropy per cosmic ray , @xmath287 where @xmath288 is boltzmann s constant , @xmath289 is the cosmic - ray number density , @xmath290 is planck s constant , @xmath291 is the system volume , and @xmath292 . according to equation ( 2.3 ) , @xmath293 gives the probability that a randomly selected cosmic ray at location @xmath29 has momentum in the range between @xmath27 and @xmath294 .
the cosmic - ray entropy density @xmath295 is computed using @xmath296 where the final two terms stem from the fundamental indistinguishability of the cosmic ray particles ( of like composition ) , and is necessary in order to avoid the gibbs paradox ( reif 1965 ) . the inconvenient reference to the system volume @xmath291
can be removed by combining equations ( 6.1 ) and ( 6.2 ) to obtain the equivalent expression @xmath297 the total entropy per particle @xmath298 for the combined gas - particle system is calculated using @xmath299 where @xmath300 and @xmath301 denote the entropy density and the number density of the background gas , respectively .
one may reasonably hypothesize that the state with the largest value for @xmath298 will be the preferred state in nature .
this criterion may prove useful for identifying the most stable solution when multiple solutions are available .
we plan to pursue this line of inquiry in future work .
the results may shed new light on the structure of cosmic - ray modified shocks .
the authors acknowledge several stimulating conversations with frank jones and truong le .
the authors are also grateful to the anonymous referee for several useful comments .
pab also acknowledges support and hospitality from nasa via the nasa / asee summer faculty fellowship program at goddard space flight center . in this section
we demonstrate that the cosmic - ray energy flux must be continuous across a velocity discontinuity ( subshock ) , if one is present in the flow .
this is turn implies that the subshock must be mediated entirely by the pressure of the background gas , and therefore the discontinuity is governed by the classical rankine - hugoniot jump conditions .
the cosmic - ray energy flux in the @xmath302 direction is given by @xmath303 in a steady - state , the cosmic - ray energy equation ( 2.7 ) reduces to @xmath304 which can be combined with equation ( a1 ) to express the derivative of @xmath305 as @xmath306 integration in the vicinity of the subshock located at @xmath307 yields @xmath308 in order to avoid unphysical divergence of @xmath305 at the subshock , @xmath40 must be continuous , and therefore the integrand on the right - hand side of equation ( a4 ) is no more singular than a step function .
this implies that in the limit @xmath309 , the right - hand side vanishes , leaving @xmath310 hence @xmath305 remains constant across the subshock .
the energy , momentum , and particle fluxes of the gas and the cosmic rays are therefore independently conserved across the discontinuity .
this allows us to use the standard rankine - hugoniot jump conditions to describe the subshock transition in equations ( 4.5 ) .
topology of the function @xmath104 given by eq .
the quantities @xmath107 and @xmath108 respectively denote the upstream and downstream roots for the velocity in a globally smooth solution .
critical upstream cosmic - ray mach number @xmath134 ( eq . [ 3.4 ] ) for smooth flow plotted as a function of the upstream gas mach number @xmath92 in the @xmath25 parameter space for ( _ a _ ) @xmath177 and various values of @xmath1 as indicated ; ( _ b _ ) @xmath178 and various values of @xmath2 as indicated .
smooth flow is not possible in the region above each curve .
critical upstream gas mach number for smooth flow @xmath143 ( eq .
[ 3.8 ] ) plotted as a function of the gas and cosmic - ray adiabatic indices @xmath2 and @xmath1 , respectively . when @xmath194 , smooth flow is possible for any value of @xmath93 .
schematic depiction of the function @xmath104 ( eq . [ 2.25 ] ) . if the flow contains a discontinuous , gas - mediated subshock , then the velocity must jump _ directly _ to the final asymptotic value @xmath108 in crossing the shock .
otherwise the flow is unstable ( see the discussion in the text ) .
function @xmath176 ( eq . [ 4.8 ] ) is plotted for the parameters ( _ a _ ) @xmath182 , @xmath177 , @xmath178 .
the value of @xmath93 is indicated for each curve . in this example , @xmath311 , and therefore the primary subshock solution appears when the low - velocity root @xmath312 , which occurs when @xmath313 .
the same function is plotted on a smaller scale in ( _ b _ ) , where we see that two new subshock roots for @xmath168 appear when @xmath314 .
fig . 6 . function @xmath176 ( eq . [ 4.8 ] ) is plotted for the parameters ( _ a _ ) @xmath191 , @xmath177 , @xmath178 .
the value of @xmath93 is indicated for each curve . in this example , @xmath315 , and therefore the primary subshock solution never appears .
hence smooth flow is possible for all values of @xmath93 .
the same function is plotted on a smaller scale in ( _ b _ ) , where we see that two new subshock roots for @xmath168 appear when @xmath316 . fig .
critical mach numbers @xmath199 and @xmath200 for the onset of multiple solutions ( eqs .
[ 4.11 ] and [ 4.12 ] ) are plotted as parametric functions of the pre - subshock velocity @xmath168 in the @xmath25 parameter space for ( _ a _ ) @xmath177 and the indicated values of @xmath317 ; ( _ b _ ) @xmath178 and the indicated values of @xmath2 .
the interior of each wedge is the multiple - solution region for the associated parameters .
fig . 8 . critical upstream mach numbers for the occurrence of multiple solutions ( eqs .
[ 4.11 ] and [ 4.12 ] ; _ solid line _ ) and for smooth flow ( eq . [ 3.4 ] ; _ dashed line _ ) are plotted together in the @xmath25 parameter space for the case @xmath177 and @xmath178 . the minimum upstream cosmic - ray mach number required for decelerating flow
is also shown ( eq .
[ 2.32 ] ; _ dotted line _ ) .
there are four distinct domains in the parameter space as discussed in the text .
function @xmath176 ( eq . [ 4.8 ] ) is plotted for the parameters ( _ a _ ) @xmath231 , @xmath177 , @xmath178 along the segment @xmath230 in fig .
the values of the upstream cosmic - ray mach number are @xmath233 ( _ solid line _ ) , @xmath318 ( _ dashed line _ ) , @xmath319 ( _ dotted line _ ) .
when @xmath320 , there are three distinct subshock solutions available . in panel
( _ b _ ) the same function is plotted on a smaller scale .
function @xmath176 ( eq . [ 4.8 ] ) is plotted for the parameters ( _ a _ ) @xmath233 , @xmath177 , @xmath178 along the segment @xmath232 in fig .
. the values of the upstream gas mach number are @xmath231 ( _ solid line _ ) , @xmath321 ( _ dashed line _ ) , @xmath322 ( _ dotted line _ ) . when @xmath323 , there are three distinct subshock solutions available . in panel
( _ b _ ) the same function is plotted on a smaller scale .
critical upstream mach numbers for the occurrence of multiple solutions ( eqs .
[ 4.11 ] and [ 4.12 ] ; _ solid line _ ) and for smooth flow ( eq . [ 3.4 ] ; _ dashed line _ ) are plotted together in the @xmath25 parameter space for ( _ a _ ) @xmath177 , @xmath178 ; ( _ b _ ) @xmath177 , @xmath324 ; ( _ c _ ) @xmath325 , @xmath178 ; ( _ d _ ) @xmath325 , @xmath324
. also indicated is the minimum value of @xmath93 required for decelerating flow ( eq . [ 2.32 ] ; _ dotted line _ ) .
our analytical results for the critical curves generated using equations ( 3.4 ) , ( 4.11 ) , and ( 4.12 ) are combined with equations ( 1.5 ) to create corresponding curves in the alternative parameter spaces @xmath23 and @xmath24 employed by bulanov & sokolov ( 1984 ) and ko , chan , & webb ( 1997 ) , respectively .
panel ( _ a _ ) , with @xmath177 and @xmath178 , is identical to fig . 4 of bulanov & sokolov ( 1984 ) .
panel ( _ b _ ) , with @xmath326 and @xmath178 , is identical to fig .
1(_a _ ) of ko , chan , & webb ( 1997 ) .
note that these authors generated their curves using root - finding procedures .
the interpretation of the line styles is the same as in fig .
fig . 13 . numerical solutions for ( _ a _ ) @xmath64 , ( _ b _ ) @xmath16 , ( _ c _ ) @xmath65 , and ( _ d _ ) @xmath66 are plotted as functions of @xmath327 ( see eq . [ 5.3 ] ) .
the solutions were obtained by integrating the dynamical eq . [ 5.1 ] with @xmath246 and @xmath247 , which corresponds to domain i in fig .
8 . in this case
one discontinuous solution is possible , and smooth flow is impossible .
14 . same as fig . 13 , except @xmath259 and @xmath260 , which corresponds to domain ii in fig . 8 .
in this case three distinct discontinuous solutions are possible , and smooth flow is impossible .
the values of the pre - subshock gas mach number are @xmath328 ( _ solid line _ ) , @xmath329 ( _ dashed line _ ) , @xmath330 ( _ dotted line _ ) .
15 . same as fig . 13 , except @xmath268 and @xmath331 , which corresponds to domain iii in fig . 8 .
in this case two distinct discontinuous solutions are possible in addition to one globally smooth solution ( _ solid line _ ) .
the values of the pre - subshock gas mach number are @xmath332 ( _ dashed line _ ) , @xmath333 ( _ dotted line _ ) .
same as fig . 13 , except @xmath280 and @xmath281 , which corresponds to domain iv in fig . 8
. in this case
one globally smooth solution solution is possible , and discontinuous flow is impossible . | the acceleration of relativistic particles due to repeated scattering across a shock wave remains the most attractive model for the production of energetic cosmic rays .
this process has been analyzed extensively during the past two decades using the `` two - fluid '' model of diffusive shock acceleration .
it is well known that 1 , 2 , or 3 distinct solutions for the flow structure can be found depending on the upstream parameters .
interestingly , in certain cases both smooth and discontinuous transitions exist for the same values of the upstream parameters .
however , despite the fact that such multiple solutions to the shock structure were known to exist , the precise nature of the critical conditions delineating the number and character of shock transitions has remained unclear , mainly due to the inappropriate choice of parameters used in the determination of the upstream boundary conditions . in this paper
we derive the exact critical conditions by reformulating the upstream boundary conditions in terms of two individual mach numbers defined with respect to the cosmic - ray and gas sound speeds , respectively .
the gas and cosmic - ray adiabatic indices are assumed to remain constant throughout the flow , although they may have arbitrary , independent values .
our results provide for the first time a complete , analytical classification of the parameter space of shock transitions in the two - fluid model .
we use our formalism to analyze the possible shock structures for various values of the cosmic - ray and gas adiabatic indices . when multiple solutions are possible , we propose using the associated entropy distributions as a means for indentifying the most stable configuration . |
_ ab initio _ molecular dynamics based on density functional theory ( dft ) is becoming an important tool for studying the plethora of structural and dynamical processes in a broad range of systems in material science , chemistry , biology and physics.@xcite the application of this approach to very large systems is still limited by the computational scaling of the electronic structure portion of the calculation , regardless of whether one uses a lagrangian - based or born - oppenheimer - based methods .
this is because of the cubic scaling involved in solving the kohn - sham equations coupled with the need to iterate to self - consistency or to propagate the kohn - sham ( ks ) orbitals , as both of these options further increases the computational times by an order of magnitude .
significant advances in these respects have been made along two major directions .
one primary direction is based on a lagrangian formulation of density functional theory @xcite and circumvents the need for scf iterations by propagation of the ks orbitals .
this venue does not eliminate the cubic scaling and is therefore limited to relatively small systems .
another approach is based on linear - scaling techniques @xcite , that reduces the algorithmic complexity by finding the density matrix directly , relying on its asymptotic sparseness in real - space .
however , sparsity sets in only for very large systems , limiting the applicability sparse - matrix methods , especially in 3d . in a recent set of papers
we have introduced the stochastic dft ( sdft ) method @xcite which scales linearly ( or even sublinearly ) with the system size and does not rely on the sparsity of the density matrix .
sdft is a general approach to electronic structure based on a stochastic process and is applicable to extended ordered as well as disordered materials .
the success of sdft in reducing the scaling comes at a price of introducing a stochastic error in all its predictions , including forces , and that precludes application to _ ab initio _ molecular dynamics . in this paper
we show that sdft can be used to study equilibrium structural properties of large ncs , despite the statistical fluctuations in the force estimates .
for this , we invoke the langevin equation following the work of attaccalite and sorella,@xcite and generate a sequence of configurations distributed according to the canonical ensemble .
these configurations can be used in a variety of applications for studying the structural , electronic and optical properties of ncs . here
we demonstrate their use for studying the structural properties of silicon nanocrystals ( ncs ) with a diameter of up to @xmath0 nm , and @xmath1 electrons .
the structure of the paper is as follows : section [ sec : methods ] includes an overview of the sdft method , a variance reducing method based on embedded fragments sdft,@xcite applicable for covalently bonded systems ( described in detail also in appendix [ sec : the - embedded - fragments ] ) .
the section also includes a discussion of langevin dynamics propagators , the parameters , and the validation of the approach .
section [ sec : results ] includes the application of the method for studying structural properties of large si ncs from cryogenic to room temperatures .
section [ sec : conclusions ] concludes .
kohn - sham density functional theory @xcite maps a system of @xmath2 interacting electrons in an external electron - nucleus potential @xmath3 , where @xmath4 ( @xmath5 ) are the nuclei positions and @xmath6 are their charge ( @xmath7 is the electron charge ) , onto a system of _ non - interacting _ electrons ( the ks system ) , having the same ground - state density @xmath8 .
this mapping is performed by solving the ks equations @xcite @xmath9 where the ks hamiltonian is : @xmath10 and the ks potential @xmath11 is the sum of the external electron - nuclear potential @xmath12 , the density - dependent hartree potential @xmath13 , and the exchange - correlation potential @xmath14 : @xmath15 in the ks system , the density is expressed in terms of the normalized single electron ks eigenstates @xmath16 and eigenvalues @xmath17 @xmath18 where @xmath19 is the heaviside function and @xmath20 is the chemical potential chosen so that @xmath21 .
( [ eq : ks - equations])-([eq : n(r ) ] ) must be solved self - consistently , since @xmath22 depends on the density .
while the entire scheme is a significant simplification over the original many - electron problem , it remains a challenge for large systems since the computational effort scales as @xmath23 . an important step towards reducing the computational scaling of ks - dft
was recently proposed by baer _ _ , _ _ neuhauser , and rabani ( bnr),@xcite where the density of eq .
( [ eq : n(r ) ] ) was expressed as a trace over the projected density operator:@xcite @xmath24.\label{eq : trace}\ ] ] the problem now shifts into calculating self - consistently the trace in eq .
( [ eq : trace ] ) ( since @xmath22 depends on @xmath8 ) rather than solving the ks equations by brute - force diagonalization .
when the trace is performed using the ks eigenstates , the computational cost remains @xmath23 similar to the traditional approach .
however , since the trace is invariant to the basis , alternative schemes that potentially lead to improved scaling can be used .
one such scheme is based on the concept of a stochastic trace formula , which reduces the scaling of the trace operation by introducing a controlled statistical error.@xcite using the stochastic trace formula , the density can be estimated as a symmetrized stochastic trace formula , given by:@xcite @xmath25 where @xmath26 denotes an average over @xmath27 stochastic orbitals @xmath28 , defined as : @xmath29 for each grid point @xmath30 , the parameter @xmath31 ( not to be confused with the ks hamiltonian @xmath32 operator ) is the grid spacing , and @xmath33 are statistically independent random variables in the range @xmath34 $ ] ( @xmath35 ) . the density @xmath8 is , strictly speaking , given by the limit @xmath36 and we approximate it with a finite @xmath27 .
the heaviside function in eq .
( [ eq : stoch - dens ] ) is smoothed by the function @xmath37 $ ] , where @xmath38 is a large constant satisfying @xmath39 , where @xmath40 is the ks - dft fundamental gap .
the action of @xmath41 on @xmath42 is evaluated by a chebyshev expansion in powers of the sparse ks hamiltonian , @xmath43.@xcite the stochastic trace evaluation ( eq .
[ eq : stoch - dens ] ) reduces the computational scaling of ks - dft to @xmath44 and for certain properties even to a sub - linear scaling.@xcite importantly , it does not rely on the sparsity of the density matrix often utilized by other linear - scaling methods for dft @xcite and is therefore suitable for studying the electronic properties of large 3d structures with open boundary conditions .
a converged self - consistent solution of eq . ( [ eq : stoch - dens ] ) provides an estimate of the electron density and in addition can be used to generate other quantities , such as the density of states ( dos ) , the total energy per electron , and the forces acting on the nuclei .
all estimates contain a statistical error that can be controlled by increasing the number of stochastic orbitals ( @xmath27 ) used to evaluate the trace in eq .
( [ eq : stoch - dens ] ) . of particular relevance to this work
are the cartesian forces exerted by the electrons on @xmath45 nuclei ( @xmath46 ) , which can be evaluated through the hellman - feynman theorem:@xcite @xmath47 these sdft forces can be expressed as : @xmath48 where @xmath49 is the deterministic ( generally unknown ) force , @xmath50 is the pure fluctuating term , and @xmath51 is the bias expected to be proportional to @xmath52 in leading order . the choice of @xmath27 should be large enough to reduce @xmath51 to negligible values and the only source of error in the procedure is then the statistical fluctuations proportional to @xmath53 with vanishing mean ( @xmath54 ) .
the reduction of the scaling in sdft is achieved by replacing the deterministic , numerically exact , trace evaluation with a stochastic sampling of the density . in return , this leads to statistical errors in the computed observables . to reduce the size of the statistical fluctuations ,
an embedded fragment method was introduced ( see ref . ) .
the system is divided into @xmath55 small fragments that are possibly overlapping .
the division to fragments is flexible , and any desired physically motivated fragmentation can be used .
the density is then a sum of the fragment density and a small correction term : @xmath56 where @xmath57 is the density generated by the individual fragments obtained from a deterministic ks - dft calculation for each fragment and @xmath58 is a correction term evaluated using stochastic orbitals . here
, @xmath59 is given by eq .
( [ eq : stoch - dens ] ) and @xmath60 is a sum over a _
stochastic _ estimate of the fragments density . in the limit @xmath61 , eqs .
( [ eq : stoch - dens ] ) and ( [ eq : frag - corr - dens - normal - form ] ) are identical and equal to the deterministic density . for finite values of @xmath27 , the size of the statistical fluctuations of the two approaches
are quite different .
since the deterministic fragmented density , @xmath62 , provides a reasonable approximation for the full density @xmath8 , the correction term , @xmath63 , which is evaluated stochastically , is rather small , leading to a reduced variance in the relevant observables ( forces , dos , total energy per electron , etc . )
compared to the direct stochastic approach of eq .
( [ eq : stoch - dens ] ) .
an equivalent viewpoint is that the fragmentation is a device for reducing the variance in the stochastic evaluation of the density .
this is evident by rewriting eq .
( [ eq : frag - corr - dens - normal - form ] ) in the following form @xmath64 and the implementation of this form is described in appendix ( [ sec : the - embedded - fragments ] ) . to assess the accuracy of the embedded fragmented sdft ( efsdft )
, we calculated the standard deviations ( stds ) and mean absolute deviations with respect the deterministic dft ( mads ) of the atomic forces in a @xmath65 nc using hydrogen passivated @xmath66 fragments .
the results are shown in fig .
[ fig : std of frags ] .
, troullier - martins norm - conserving pseudopotentials @xcite within the kleinman - bylander approximation.@xcite fast fourier transforms were used for applying the kinetic energy operator and for determining the hartree potentials and the method of ref .
was used for treating the long range coulomb interactions in a finite simulation cell with periodic boundary conditions .
dft calculations were performed under the local density approximation ( lda ) .
] the stds and mads decrease as @xmath67 , indicating that the bias in the force estimation is negligible .
the standard deviations in the sdft forces are larger by a factor of @xmath68 compared to those of efsdft .
this implies that the required number of stochastic orbitals in efsdft is nearly an order of magnitude smaller than in sdft for similar stds .
the stds can be further reduced by using larger fragments as discussed below ( cf . ,
[ fig : std0-frags ] ) .
the standard approach to generate canonically distributed configurations _ using ab initio _ techniques is based on molecular dynamics , which requires as input accurate force estimates for each atomic degree of freedom .
since the forces generated by efsdft contain a stochastic component , we use langevin dynamics ( ld ) instead of molecular dynamics to sample configurations according to the boltzmann distribution .
a ld trajectory @xcite is a sequence of configurations @xmath69 at discrete `` times ''
@xmath70 , where @xmath71 is the time step , and @xmath72 and @xmath73 are the cartesian coordinates and conjugate momenta , respectively , for the @xmath45 atoms , the trajectory is a solution of the langevin equation ( le ) of motion:@xcite @xmath74 where @xmath75 is the mass of the atom @xmath76 , @xmath77 is its friction constant , and @xmath78 is the total efsdft force acting on it , including deterministic and fluctuating parts ( see eq .
[ eq : force - deviations ] ) .
the bias is assumed negligible , so that @xmath79 . in eq .
( [ eq : langevin eq ] ) @xmath80 is an additional uncorrelated white - noise force introduced so as to satisfy the fluctuation - dissipation ( fd ) relation .
we require that the total random fluctuation on each atom obey : @xmath81 and @xmath82 where @xmath83 are atom indices , @xmath84 designates average over the atomic force distribution , @xmath85 is the @xmath86 unit matrix , and @xmath87 is the atomic force std of atom @xmath76 , which is taken to satisfy the fluctuation - dissipation relation : @xmath88 we use the verlet - like algorithm @xcite for numerically integrating the le of motion at a fixed temperature @xmath89 and a predefined time - step @xmath71 .
the positions and momenta in time step @xmath90 depend on the positions and momenta in time step @xmath91 as well as on the forces in time step @xmath91 and the additional white noise @xmath92 is sampled from a gaussian distribution such that the discretized version of eq .
( [ eq : fluc - correlation ] ) holds : @xmath93 : @xmath94 where @xmath95 and @xmath96 .
the algorithm allows for stable and accurate solutions of the le with time step comparable to that used in molecular dynamics simulations for similar systems .
it treats the additional white noise component @xmath80 of the force differently from the force @xmath78 that result from the efsdft calculation containing deterministic and fluctuating components that can not be separated . the verlet - like @xcite temperature in @xmath97 evaluated with efsdft using : transient kinetic energy ( dotted line ) , running average kinetic energy for si ( red ) and the running average virial ( black ) using @xmath98 stochastic orbitals , a time step of @xmath99 and the friction coefficients are @xmath100 .
] in fig .
[ fig : temperature ] we plot for @xmath97 the running average of the transient temperature , @xmath101 , calculated from the kinetic energy @xmath102 and from the virial estimator , @xmath103 in the above , @xmath104 is the time average of the coordinate of atom @xmath76 .
the initial positions of the si atoms were taken from the bulk values .
all surface si atoms with more than two dangling bonds were removed and the remaining surface si atoms were passivated using one or two h atoms placed in a tetrahedral geometry at the si - h distance of @xmath105. the momenta were sampled from a boltzmann distribution at @xmath106 .
this non - equilibrium initial configuration relaxes towards equilibrium .
the agreement in fig .
[ fig : temperature ] between the two temperature estimators is consistent with a proper sampling of the canonical distribution of both positions and velocities .
the small discrepancies at the longest averaging time are due to the large fluctuations of the transient temperature , particularly when using the virial estimator .
we have also calculated the fluctuations in the kinetic energy and found very good agreement with the corresponding analytical value ( not shown ) . the normalized velocity autocorrelation function ( top panel ) and the mean nearest - neighbor si - si distance ( bottom panel ) in @xmath97 as a function of time for a ld trajectory at @xmath106 with time step @xmath107 calculated using a ddft based ld for different values of @xmath108 .
the dashed curve corresponds to a efsdft based ld calculation with @xmath109 and @xmath98 stochastic orbitals .
the simulation started intentionally from an inflated comfiguration in order to to measure the relaxation time . ] the si - si pair distribution function @xmath110 for @xmath111 calculated using ddft ( dotted lines ) and efsdft based ld ( solid lines , see table [ tab : params ] for parameters ) at @xmath112 , ( blue curves ) and @xmath113 ( red curves ) .
inset : details of the first ( nearest neighbor ) peak .
the dotted lines are for different friction coefficients @xmath108 in the range @xmath114 . ] the effect of @xmath115 on the configurational relaxation and on the velocity autocorrelation decay is illustrated in fig .
[ fig : thermalization - times ] for @xmath116 . in order to decrease the number of unknown parameters we set the value of @xmath117 and @xmath118 to be identical for all atoms of the same type ( i.e. si or h in the systems studied here ) .
the value of the added white noise is then determined by eq .
( [ eq : fdt ] ) , ensuring that the fluctuation - dissipation relation holds . before this step can be taken an estimate of the magnitude of the sdft force fluctuation @xmath119 ( @xmath120 ) appearing in eq .
( [ eq : fluc - correlation])must be obtained .
this quantity is estimated by a separate set of runs on the initial nc configuration using several independent sets of stochastic orbitals .
note , that we have tested that the magnitude of the sdft force fluctuation @xmath119 is not sensitive to the particular configuration used .
.[tab : params]value of various parameters for the ld based on efsdft calculations : the friction coefficients @xmath117 , number of stochastic orbitals @xmath27 , time - step @xmath71 and the wall time per single scf iteration @xmath121 . [ cols="^,^,^,^,^,^,^ " , ] the relaxation transient is studied in greater detail in fig.[fig : si - si - dist ] , where the average nearest - neighbor bond lengths are shown for @xmath122 ( spherical shells a - b ) and @xmath123 ( shells a - e ) ; see table [ tab : shells ] for the definition and properties of the shells . in @xmath123
the deep layer shells ( a - d ) relax slower than those near the surface showing that relaxation progresses from the surface inwards .
the difference between the relaxation times of the two systems is correlated with the smaller frequency , @xmath124 , of the breathing mode of the larger nc . in the limit of an over damped motion ( as is the case here since @xmath125 ) , the relaxation is dominated by two timescales proportional to @xmath126 and @xmath127 . the former leads to a fast relaxation while the latter is slower and depends on the value of @xmath128 .
the ratio of the breathing mode frequency for the two particles is @xmath129 ( l / s for large / small ) assuming that the breathing mode frequency scales linearly with the nc diameter.@xcite this is similar to the ratio of the relaxation times ( @xmath130 ) for the lower temperature . at the higher temperature
, one needs to consider anharmonic effects which are more pronounced in the large nc with lower acoustic phonons .
another noticeable feature in fig .
[ fig : si - si - dist ] is that the si - si bonds seem slightly shorter in @xmath122 than in @xmath131 .
this results from the difference in the bond distance of atoms in the outer shell , while the inner shell atoms have similar bond distances .
in this paper we developed an _ ab initio _ langevin dynamics approach based on a new embedded fragment stochastic dft method .
we showed how the noisy forces resulting from the efsdft calculation are used to generate a set of configurations that are distributed canonically at cryogenic and room temperatures . by proper choice of the friction coefficients and the number of stochastic orbitals
, thermalization is reached within @xmath132 time steps for these materials . using the calculations we determined the structural properties of silicon nanocrystals of size up to @xmath133 in diameter containing more than @xmath134 electrons .
e.r . and d.n .
acknowledge support by the nsf , grants che-1465064 and dmr / bsf-1611382 .
r.b . acknowledges the us - israel binational science foundation support under the bsf - nsf program , grant 2015687 .
here , we provide the technical details for the embedded fragment method described in section [ sec : methods ] . the method corrects the stochastic estimate @xmath135 for the expectation value @xmath136 of a one - body operator @xmath137 , using calculations performed on @xmath55 separate fragments ( see eq .
( [ eq : frag - corr - dens - roi - form ] ) ) : @xmath138 where the stochastic correction due to fragment @xmath139 is @xmath140 and the deterministic and the stochastic estimates , @xmath141 and @xmath142 are calculated directly on the fragment itself .
previous implementations of embedded fragment sdft were applied to systems of many weakly interacting molecules where the selection of fragments or clusters of such molecules was natural.@xcite we now describe a new method for defining and carrying calculations with fragments which can break up covalently bonded systems such as silicon ncs .
the large system is divided into @xmath55 small fragments composed of one or more bonded atoms each .
the surface dangling bonds of the fragment are passivated using a h atom placed in @xmath143 from the si atom , in the direction of the neighboring atom which is not included in the fragment .
this forms a saturated fragment . for a saturated fragment @xmath139 ,
the deterministic ks - dft method is applied to determine the ks eigenvalues @xmath144 and eigenfunctions @xmath145 .
further , occupation numbers @xmath146 are introduced for determining the saturated fragment density @xmath147 . the fragment density @xmath148 is `` carved out '' of @xmath149 using a carving function @xmath150 .
thus : @xmath151 where , inspired by hirshfeld partitioning , @xcite the carving function is defined as : @xmath152 where @xmath153 is the spherical density of neutral atom @xmath154 .
the temperature parameter @xmath38 in the definition of the population @xmath155 is chosen be the same value as that of the sdft calculation , while the chemical potential @xmath156 of each fragment is determined by the condition of neutrality of the fragment : @xmath157 defining non - orthogonal functions @xmath158 , the fragment density of eq .
( [ eq : carved - density ] ) becomes @xmath159 , so the chemical potential is determined from the condition : @xmath160 after determining @xmath156 and in order to construct the reduced density matrix ( rdm ) , we orthogonalize the functions @xmath161 by diagonalizing the overlap matrix @xmath162 , obtaining the unitary matrix @xmath163 of eigenvectors and the eigenvalues @xmath164 ( so that @xmath165 $ ] ) .
the orthogonal wavefunctions are : @xmath166 and the norm is @xmath167 . using the new wave functions , the unsaturated fragment density is given by : @xmath168 and the rdm by @xmath169 using the rdm we express the unsaturated fragment expectation value appearing in eq .
( [ eq : fragment - correction - a ] ) as : @xmath170=tr\left[\sqrt{\hat{\theta}_{f}}\hat{a}\sqrt{\hat{\theta}_{f}}\right],\ ] ] where @xmath171 by choosing the fragment grid - points to be a subset of the full system grid , each stochastic orbital @xmath172 ( @xmath173 ) of the full system appears as a stochastic orbital on the fragment grid and can be used to perform the stochastic estimate appearing in eq .
( [ eq : fragment - correction - a ] ) as : @xmath174 where the subscript @xmath139 on the left denotes integration over the fragment grid . the difference @xmath175 in eq .
( [ eq : fragment - correction - a ] ) can now be written in a unified form as : @xmath176 where : @xmath177 hence , by calculating the matrix @xmath178 all types of expectation value corrections can be obtained from eq .
( [ eq : deltaa - i ] ) .
the efficacy of embedded fragments in sdft force calculations is achieved through a reduction of the std @xmath179 of a force component .
the std @xmath179 is proportional to @xmath67 , where @xmath27 is the number of stochastic orbitals and the proportionality constant , denoted @xmath180 , is called the _ inherent _ std .
this quantity depends on the nc characteristics but not on the number of stochastic orbitals . in fig .
[ fig : std0-frags ] we plot the inherent std on each atom for @xmath97 and @xmath123 as a function of fragment size .
even the use of the smallest fragments reduces the inherent force std by a significant factor , @xmath181 ( for @xmath123 ) to @xmath182 ( for @xmath97 ) .
using larger fragments reduces the std by an additional factor of @xmath183 , with increasing effect for larger systems , since the electron density in the larger fragments is similar to that of the full system .
it is interesting to see that for the forces there is no noticeable sublinear scaling : the inherent std for both systems is similar , with the larger system having a slightly ( @xmath184 ) std . in summary ,
the embedded fragment sdft method serves as a way to expedite the sdft calculation by a judicious choice of fragment size and composition .
as the fragment size grows , the numerical effort invested in sdft decreases ( due to reduction of std ) while in ddft it increases .
for example , consider fig .
[ fig : std0-frags ] where we showed that increasing the fragment size by a factor of @xmath185 reduces the std by a factor of @xmath186 and therefore the sdft cpu time by a factor of @xmath187 . on the other hand since the fragments are ten - fold larger , the amount of ddft work on them increases ( cubically ) by a factor of more than@xmath188 . clearly then , the optimal fragment size is system dependent . embedded fragments have the additional benefit of providing an initial density for the scf calculation , significantly reducing the number of scf cycles .
34ifxundefined [ 1 ] ifx#1 ifnum
[ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) http://prb.aps.org/abstract/prb/v47/i1/p558_1 [ * * , ( ) ] @noop * * , ( ) `` , '' in @noop _ _ , , vol . ,
( , , ) p. @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.111.106402 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , , ) @noop * * , ( ) @noop _ _ ( , , ) pp .
@noop _ _ , vol .
( , ) @noop _ _ ( , ) @noop _ _ , ed .
( , , ) pp . @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) http://link.springer.com/article/10.1007%2fbf00549096 [ * * , ( ) ] | an _ ab initio _ langevin dynamics approach is developed based on stochastic density functional theory ( sdft ) within a new embedded fragment formalism .
the forces on the nuclei generated by sdft contain a random component natural to langevin dynamics and its standard deviation is used to estimate the friction term on each atom by satisfying the fluctuationdissipation relation .
the overall approach scales linearly with system size even if the density matrix is not local and is thus applicable to ordered as well as disordered extended systems .
we implement the approach for a series of silicon nanocrystals ( ncs ) of varying size with a diameter of up to @xmath0 nm corresponding to @xmath1 electrons and generate a set of configurations that are distributed canonically at a fixed temperature , ranging from cryogenic to room temperature .
we also analyze the structure properties of the ncs and discuss the reconstruction of the surface geometry . |
cmes and eruptive flares are believed to result from a sudden , explosive release of the free magnetic energy stored in the previously quasi - equilibrium , twisted / sheared coronal magnetic field ( see e.g. reviews by * ? ? ?
* ; * ? ? ?
many cme models have considered a magnetic flux rope containing helical field lines twisting about each other in the corona as the basic underlying magnetic field structure for cme precursors ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
recent observational studies of coronal prominence cavities support the picture of twisted coronal flux ropes as cme precursors ( e.g. * ? ? ?
theoretical studies have shown that a coronal flux rope confined by an external potential magnetic field can become unstable to the torus instability and erupt when it reaches a critical height where the ambient potential field decreases with height too steeply @xcite .
mhd simulations have been carried out to study the critical condition and the non - linear evolution of the torus instability of various 3d coronal flux rope structures ( e.g. * ? ? ?
* ; * ? ? ?
? * ; * ? ? ?
* ) . through a sequence of 3d mhd simulations of the evolution of the coronal magnetic field resulting from the emergence of a twisted magnetic flux rope into a pre - existing coronal arcade field ,
@xcite studied the conditions that lead to a dynamic eruption of the resulting coronal flux rope .
it is found that the critical condition for the onset of eruption is for the center of the flux rope to reach a critical height at which the corresponding potential field declines with height at a sufficiently steep rate , consistent with the onset of the torus instability of the flux rope .
the simulations show that after the flux emergence is stopped , the coronal flux rope first settles into a quasi - static rise phase with an underlying , central sigmoid - shaped current layer developing .
reconnections in the current layer is found to effectively add twisted flux to the flux rope ( as quantified by eq .
( 14 ) in * ? ? ?
* ) , allowing it to rise quasi - statically , even as the magnetic energy is declining .
we thus call these reconnections tether - cutting " reconnections in the sense that they are effectively reducing the anchoring of the flux rope by dissipating the underlying , central current layer whose current attracts the volume current in the flux rope @xcite . when the flux rope rises quasi - statically to the critical height for the onset of the torus instability , then dynamic eruption ensues .
to identify the thermal signatures that may result from the formation of the current layer and the associated reconnections , we carry out an mhd simulation with a similar setup as those in @xcite , but with a more sophisticated treatment of the thermodynamics . instead of assuming a simple isothermal evolution , we solve the total energy equation in conservative form for a perfect gas with an adiabatic index @xmath0 . in this way the non - adiabatic heating due to the dissipation of kinetic and magnetic energies by the numerical diffusions
are implicitly incorporated in the thermal energy . also thermal conduction along the magnetic field lines
is included in the energy equation .
we find from this new simulation that the main consequence of the tether - cutting reconnections in the current layers during the quasi - static phase is the formation and enlargement of a hot , low density channel , containing reconnected , twisted flux threading under the axis of the flux rope . when viewed above the limb along a line of sight that is roughly aligned with the channel , the thermal signature would manifest in sdo / aia observations as a hot central void enclosed in a `` u '' shaped dense shell on top of a relatively dense structure inside the flux rope .
this signature inside the flux rope is consistent with the recent observations of the thermal sub - structures inside coronal prominence cavities by @xcite .
we organize the remainder of the paper as follows . in section [ sec :
model ] , we describe the mhd numerical model with the more sophisticated treatment of the thermodynamics compared to the previous simulations in @xcite . in section [ sec : result ] we describe the simulation result , focusing on the thermal signature that develops inside the flux rope and compare it with observations .
we summarize and discuss the conclusions in section [ sec : conc ] .
for the simulation carried out in this study , we solve the following magneto - hydrodynamic equations in a spherical domain : @xmath1 @xmath2 @xmath3 @xmath4 @xmath5 - \rho { \bf v } \cdot \frac{g m_{\odot}}{r^2 } \hat{\bf r } - \nabla \cdot { \bf q } , \label{eqetot}\ ] ] @xmath6 where @xmath7 @xmath8 @xmath9 in the above @xmath10 is the velocity field , @xmath11 is the magnetic field , @xmath12 , @xmath13 , and @xmath14 are respectively the plasma density , pressure and temperature , @xmath15 , @xmath16 @xmath17 , @xmath18 , and @xmath19 , are respectively the internal energy density , the total energy density ( internal+kinetic+magnetic ) , the gas constant , the mean molecular weight , and the adiabatic index of the perfect gas , @xmath20 is the heat flux due to field aligned thermal conduction , with the field direction denoted by the unit vector @xmath21 and with @xmath22 , and @xmath23 and @xmath24 denote the gravitational constant and the mass of the sun .
we have assumed a perfect gas with a low value of the adiabatic index : @xmath0 , which allows the coronal plasma to better maintain its ( high ) temperature during adiabatic expansion .
viscosity and magnetic diffusion are included in the momentum and the induction equations .
however numerical dissipations are present at regions of sharp gradient , and since we are solving the total energy equation of the plasma in conservative form ( eq . [ [ eqetot ] ] ) , the non - adiabatic heating due numerical dissipation of kinetic , and magnetic energies is implicitly put back into the internal energy .
we also explicitly include the non - adiabatic effect due to the field - aligned thermal conduction ( eqs .
[ [ eqetot ] ] and [ [ conductflux ] ] ) in the energy equation , but in a limited way .
given the temperature and density in the solar corona , the field aligned thermal conduction would pose a severely limit on the time step of numerical integration .
here we artificially limit the conductive heat flux by limiting @xmath25 to a maximum value of @xmath26 , where @xmath27 is the proton mass , @xmath28 is the boltzman constant , @xmath29 is the smallest dimension of the grid cell , and @xmath30 is the numerical time step calculated from the courant condition based on all the other dynamical terms .
note , this limiting is a purely numerical consideration to avoid the extremely stringent time step that would be required by the courant condition set by the thermal diffusion at large @xmath31 where density decreases rapidly .
this limiting is mainly altering thermal conduction in the region above @xmath32 , to reduce it such that the time step set by the courant condition for the thermal diffusion does not go below that set by the fastest dynamical speed .
since we are focusing on the thermal signatures that develop during the quasi - static phase over a time scale long compared to the dynamic time scale and inside the flux rope below @xmath33 where the thermal conduction is unaltered , we do not expect this limiting of thermal conduction to significantly alter the qualitative properties of the thermal signatures .
thus with the present treatment of the thermodynamics , we aim to _ qualitatively _ identify regions of significant heating due to the formation and dissipation of current layers and the resulting regions of enhanced temperature with the redistribution of heat by the field aligned thermal conduction .
we solve the above compressible mhd equations using a numerical code which has its origin from the zeus3d code of stone and norman ( 1992a , b ) , but has been substantially modified and rewritten , in particular in the schemes used to solve the continuity , momentum and the energy equations , and in the scheme for time stepping @xcite . for ease of future reference , we call this code `` mfe '' standing for `` magnetic flux eruption '' , since it has been mainly used for modeling flux emergence from the interior into the solar atmosphere and corona in cartesian geometry @xcite , and for modeling cme initiation in the corona in spherical geometry @xcite . here
we summarize the numerical schemes used by the mfe code for solving the above mhd equations .
the dependent field variables are discretized using a staggered mesh in the spherical domain with @xmath31 , @xmath34 , @xmath35 coordinates @xcite .
a modified , second order accurate lax - friedrichs scheme described in is applied for evaluating the fluxes in the continuity and the energy equations .
the standard second order lax - friedrichs scheme is used for evaluating the fluxes in the momentum equation .
a method of characteristics that is upwind in the alfvn waves @xcite is used for evaluating the @xmath36 term in the induction equation , together with the constrained transport scheme to ensure @xmath37 to the machine precision . for temporal discretization ,
the code uses a second - order accurate predictor - corrector time stepping .
the thermal conduction term is treated via an operator splitting procedure as follows .
all the other equations and the total energy equation ( eq . [ [ eqetot ] ] ) excluding the thermal conduction term are first advanced one full time step ( with the second - order predictor - corrector time stepping ) . then a separate step ( with a second - order predictor - corrector time stepping )
is carried out to further update the internal energy for the thermal conduction term and adjust the total energy accordingly .
the initial set up of the simulation is the same as that of @xcite .
the spherical domain ( see figure [ fig_init ] ) representing the corona is given by @xmath38 $ ] , @xmath39 $ ] , @xmath40 $ ] , where @xmath41 is the solar radius .
it is resolved by a grid of @xmath42 , which is uniform in @xmath34 and @xmath35 , and non - uniform in @xmath31 : in the range from @xmath43 to @xmath44 , the grid size is @xmath45 , and @xmath46 increases gradually for @xmath47 , reaching about @xmath48 at the outer boundary .
initially , the domain is assumed to be in hydrostatic equilibrium at a uniform temperature of @xmath49 mk with the density and pressure given by : @xmath50 @xmath51 where @xmath52 g @xmath53 is the initial density at the coronal base , and @xmath54 denotes the initial pressure scale height , which is about @xmath55 mm .
the initial atmosphere contains a pre - existing potential arcade field , whose normal field @xmath56 at the lower boundary ( see the gray scale image on the sphere in figure [ fig_init ] ) is given by @xmath57 where , @xmath58 \right ] \right ] , & \frac{\pi}{2 } - \theta_t - \theta_a < \theta < \frac{\pi}{2 } - \theta_t , \\ - \frac{2 \theta_a}{\pi } \sin \theta_p b_0 r_{\odot}^2 , & \frac{\pi}{2 } - \theta_t
< \theta < \frac{\pi}{2 } + \theta_t , \\ - \frac{\theta_a}{\pi } \sin \theta_p b_0 r_{\odot}^2 \left [ 1 + \cos \left [ \frac{\pi}{\theta_a } \left [ \theta - \left ( \frac{\pi}{2 } + \theta_t \right ) \right ] \right ] \right ] , & \frac{\pi}{2 } + \theta_t < \theta
< \frac{\pi}{2 } + \theta_t + \theta_a , \\ 0 , & \frac{\pi}{2 } + \theta_t + \theta_a < \theta < \frac{7}{12 } \pi , \end{array } \right .\ ] ] in which @xmath59 , @xmath60 , @xmath61 , and @xmath62 is the peak field strength in the arcade field .
thus the peak alfvn speed at the foot point of the arcade field is @xmath63 km @xmath64 which is more than a factor of 10 greater than the initial sound speed of @xmath65 km @xmath64 . as was in @xcite ,
we impose ( kinematically ) at the lower boundary of the domain ( at @xmath66 ) the emergence of a twisted magnetic torus @xmath67 , by specifying a time dependent transverse electric field @xmath68 that corresponds to the upward advection of the flux tube at a velocity @xmath69 : @xmath70 . \label{eq_emf}\ ] ] here the imposed velocity field on the lower boundary is a constant @xmath69 in the area where the emerging tube intersects the lower boundary and zero in the rest of the area .
the field structure @xmath67 used for specifying @xmath68 is an axisymmetric torus defined in its own local spherical polar coordinate system ( @xmath71 , @xmath72 , @xmath73 ) whose origin is located at @xmath74 of the sun s spherical coordinate system and whose polar axis is parallel to the polar axis of the sun s spherical coordinate system : @xmath75 where @xmath76 @xmath77 where @xmath78 is the minor radius of the torus , @xmath79 is the distance to the curved axis of the torus , with @xmath80 being the major radius of the torus , @xmath81 denotes the angle ( in rad ) of field line rotation about the axis over a distance @xmath78 along the axis , @xmath82 gives the field strength at the curved axis of the torus . here
we have @xmath83 , @xmath84 , @xmath85 rad @xmath86 , @xmath87 .
the magnetic field @xmath67 is truncated to zero outside of the flux surface whose distance to the torus axis is @xmath88 . for specifying the flux emergence via @xmath68 given by ( [ eq_emf ] )
, it is assumed that the torus center is initially located at @xmath89 ( thus the torus is initially entirely below the surface ) , and it moves bodily towards the lower boundary at a constant velocity @xmath90 , with @xmath91 , until a time @xmath92 , when the emergence is stopped and @xmath68 is set to zero . during flux emergence , we assume that the density inside the emerging flux tube is uniformly @xmath93 ( same as the initial density at the bottom of the domain ) .
thus as the tube is being transported into the domain via the electric field at the lower boundary , there is also an inflow of mass flux @xmath94 through the lower boundary in the area where the emerging tube intersects the boundary .
when the emergence is stopped , both the electric field and the velocity at the lower boundary are set to zero , with no inflows or outflows and with field line foot points rigidly anchored .
we assume perfectly conducting walls for the side boundaries of the simulation domain .
for the outer boundary , we use a simple outward extrapolating boundary condition that allows plasma and magnetic field to flow through .
these boundary conditions described above are the same as those used in @xcite for the present simulation , we drive the emergence of the torus via the electric field at the lower boundary until @xmath95 , the same as case e4 in @xcite .
however the difference between this simulation and those in @xcite is that here we no longer assume the simple isothermal evolution , but instead we solve the total energy equation for a perfect gas with an adiabatic index @xmath0 for the coronal plasma .
thus we take into account non - adiabatic effects due to heating produced by ( numerical ) dissipation of kinetic and magnetic energies .
we also incorporate thermal conduction along the magnetic field lines , as is described earlier in [ sec : model ] .
thus temperature can vary due to adiabatic expansion / compression , as well as due to the non - adiabatic effects : heating and thermal conduction .
this enables us to identify sites of significant heating due to current sheet formation and study the resultant temperature distribution ( qualitatively ) during the evolution of the 3d flux rope in the corona .
figure [ fig_3devol ] shows snapshots of the 3d evolution of the coronal magnetic field after the imposed flux emergence at the lower boundary has stopped at @xmath96 hour .
figure [ fig_vr_em ] shows the temporal evolution of the rise speed at the apex of the tracked axis of the emerged flux rope ( upper panel ) and the total magnetic energy in the coronal domain ( lower panel ) after @xmath92 .
the way we track the axis of the emerged flux rope is the same as that described in ( * ? ? ?
* see discussion under figure 3 in that paper ) .
we find that the flux rope first settles into a phase of quasi - static rise , and then begins to accelerate and erupt when a critical height of roughly @xmath97 is reached by the tube axis . during the quasi - static phase
after the emergence is stopped , there is no more input of poynting flux from the lower boundary and the magnetic energy in the domain is slowly decreasing ( see figure [ fig_vr_em ] ) due to tether cutting reconnections in the current sheets that develop , as we will discussed later .
compared to the corresponding case e4 in @xcite with the same imposed lower boundary conditions , the present case with a different treatment of the thermodynamics is found to undergo a longer quasi - static rise phase before the flux rope reaches roughly the same critical height for the onset of the torus instability . as a result , at the time of the onset of eruption , the ( free ) magnetic energy is lower and the alfvn speed at the rope axis is also lower ( by about 15% ) in the present case . thus it is found that the eruption is slightly less energetic with the flux rope accelerating to a lower peak speed of about @xmath98 for the eruption , compared to the peak speed of about @xmath99 in the corresponding e4 case in @xcite .
here we focus on studying the development of the thermal signatures in the flux rope as a result of current sheet formation and tether - cutting reconnections during the quasi - static rise phase .
the tether - cutting reconnections after the emergence is stopped continue to build up the twisted flux of the flux rope ( as computed in equations [ 14 ] in @xcite ) , even no more poynting flux is coming through the lower boundary and the magnetic energy is declining .
they enable the flux rope to reach the critical height and eventually erupt .
therefore the thermal structures that develop due to the reconnections may be important signatures for indicating the readiness of the flux rope to erupt , and identifying such thermal signatures in the observables is useful .
figure [ fig_merievol ] shows snapshots of the density ( in log scale ) , current density normalized by magnetic field strength ( @xmath100 ) , and temperature in the central meridional cross - section of the flux rope shown in the corresponding snapshots in figure [ fig_3devol ] . in the density cross - section during the quasi - static rise phase ( panels in the left column )
, the outermost low density layer corresponds to the pre - existing arcade fields , and the low density is produced by the stretching and expansion of the pre - existing arcade fields immediately surrounding the emerged flux rope fields . inside
the outermost low density layer is a layer of relatively high density and high @xmath101 ( indicated by the yellow arrows in figure [ fig_merievol ] ) , which corresponds to the outer boundary of the emerged flux rope . inside the flux rope
, there is a central low density void ( indicated by the green arrows in figure [ fig_merievol ] ) which grows in size during the quasi - static phase .
the central void is closed at the bottom with a sharp `` u '' shaped lower boundary , and is enclosed in a high density shell .
the sharp lower boundary of the void corresponds to a `` u '' shaped current concentration of high @xmath101 which extends from a strong central vertical current layer below ( see the middle and bottom snapshots in the 2nd column of figure [ fig_merievol ] ) . in other words ,
the central void is shaped like a `` lollipop '' on top of a vertical stick " corresponding to the central vertical current layer , and the current layer extends upward to surround the two sides of the void .
the middle and lower images in the right column of figure [ fig_merievol ] show that the central growing void is the hottest part in the central meridional cross - section , hotter than the plasma in the strongest vertical current layer below it and the current layer surrounds it .
furthermore , this hot void is growing below the apex of the emerged flux rope axis marked with the yellow `` x '' point in the cross - sections shown in figure [ fig_merievol ] .
figure [ fig_3dfdlcurr ] shows the 3d coronal field lines and the isosurface of @xmath101 at a value of @xmath102 where @xmath46 is the smallest grid size in @xmath31 .
we find that at @xmath103 hour , corresponding to the top row of figure [ fig_merievol ] , the strongest portions of the current layers are two `` j '' shaped layers curving about the two legs of the flux rope at the boundaries between the flux rope legs and the arcade with the opposite polarity foot points .
this is why the central meridional cross - section of @xmath101 ( the top panel in the middle column of figure [ fig_merievol ] ) shows a double legged structure , with each leg belonging to one of the `` j''-layers .
when the `` lollipop '' shaped central hot void has developed in the central meridional cross - section ( as is shown in the middle and bottom rows of figure [ fig_merievol ] for @xmath104 hour and @xmath105 hour ) , the strongest portion of the current layer has become a central inverse s - shaped vertical layer underlying the flux rope ( see the middle and right panels of figure [ fig_3dfdlcurr ] , and the associated movie in the electronic version ) .
the inverse - s shape is consistent with the left - handed twist of the flux rope ( e.g. * ? ? ?
* ) . a forward - s shape for the current layer would result for a right - hand twisted flux rope
we find that these current layers in the simulation ( as shown in figure [ fig_3dfdlcurr ] ) develop along topological structures identified as quasi separatrix layers ( qsls ) , which are regions of the magnetic volume where the field line connectivity experiences drastic variations @xcite .
they are a generalization of the concept of separatrices at which the field line linkage is discontinuous .
similar to a separatrix , a qsl divides the coronal domain into quasi - connectivity domains , and due to the drastic change of the field line connectivity , it is a site along which thin and intense current sheets tend to build up because of the distinct dynamic behaviors and evolutions of the field lines in each of these subdomains in the line - tied corona ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
qsls are located by estimating the so - called squashing degree , @xmath106 , which is defined as the square of the norm of the jacobian matrix of the field line mapping from one foot point to the other , divided by the absolute value of the determinant of the jacobian matrix @xcite .
qsls correspond to regions of very large @xmath106 , larger than the averaged values of the rest of the domain by many orders of magnitude ( e.g. * ? ? ?
* ; * ? ? ?
we have evaluated the squashing degree @xmath106 in our simulated coronal magnetic field using the formulation described in ( * ? ? ? *
method 3 in the paper ) .
the left panels in figure [ fig_qfac ] show the computed @xmath106 in respectively the central meridional cross - section ( upper panel ) and the cross - section at the height of @xmath107 ( lower panel ) in the 3d magnetic field shown in the right panel of figure [ fig_3dfdlcurr ] ) .
compared to the corresponding 2d cuts of @xmath101 shown in the right panels of figure [ fig_qfac ] , we can see that the thin , most intense current layers indeed form along the qsls with the largest @xmath106 values , which are orders of magnitude greater than the averaged @xmath106 in the rest of the domain .
the qsls with the highest @xmath106 values shown in the central meridional cross - section ( top left panel of figure [ fig_qfac ] ) correspond to the mid cross - section of the so - called hyperbolic flux tube ( hft ) , a generalization of the x - line configuration , which divides the magnetic volume into 4 distinct domains of magnetic field connectivity ( e.g. * ? ? ? * ; * ? ? ?
* ; * ? ? ?
the main central vertical current layer shown in the right panel of figure [ fig_3dfdlcurr ] , ( whose cross - sections are shown in the 2d cuts in figure [ fig_qfac ] ) , forms along the hft and is likely a thin current sheet that can lead to significant reconnection even under the realistic high lundquist number condition of the solar corona . to understand the 3-dimensional structure of the growing central hot void on top of the central vertical current layer seen in the cross - sections in figure [ fig_merievol ] , we show in the top panels of figure [ fig_hotchannel_t2 ] the horizontal cross - sections of density and temperature at @xmath108 for @xmath105 hour ( corresponding to the height indicated by the green arrows in the bottom row of figure [ fig_merievol ] ) .
we see a hot channel of inverse - s shape , with the main middle segment of the channel tilted away from the east - west direction ( or the direction of the emerging flux rope axis ) clock - wise by roughly @xmath109 .
there is a good anti - correlation between the temperature and the density of the channel , i.e. it is a hot , low - density channel . in the lower panels of figure [ fig_hotchannel_t2 ] ,
we also show the 3d morphology of the hot channel as outlined by the pink isosurface of temperature ( at 1.2 mk ) in relation to the current layers ( orange isosurfaces of @xmath101 ) and the magnetic field lines .
the hot channel is on top of the central vertical current layer , and the current layer extends upward to surround the two sides of the hot channel in the middle portion ( see the movie associated with figure [ fig_hotchannel_t2 ] in the electronic version which show the rotating view of the 3d structures ) .
the hot channel also threads under the axial field line of the flux rope ( as indicate in the bottom panels of figure [ fig_merievol ] ) .
if this flux rope is viewed above the limb , with a line of sight that is aligned with the main middle segment of the hot channel to a suitable angle , then one would be able to observe the central void of high temperature inside the flux rope .
figure [ fig_limbview_40 ] shows the modeled images above the limb of the emission measure ( panel b ) , line - of - sight averaged temperature ( panel c ) , and sdo / aia intensities for the @xmath110 ( panel d ) , @xmath111 ( panel e ) , and @xmath112 ( f ) filters , by integrating through the simulation domain along the line - of - sight that is tilted @xmath109 clock - wise from the east - west direction ( or the direction of the axis of the flux rope ) .
also shown in the figure is the view above the limb of the 3d field lines and the iso - surface of @xmath101 ( at a value of @xmath113 with @xmath46 being the smallest radial grid size ) along the same line of sight ( figure [ fig_limbview_40]a ) . for computing the modeled emission above the limb at the aia @xmath110 , @xmath111 , and @xmath112 wavelength channels
, we use the temperature and density from the simulation and carry out the following integration along the line - of - sight @xmath114 through the domain : @xmath115 where the subscript @xmath116 denotes one of the aia wavelength channels , @xmath117 denotes the intensity at each pixel in units of dn / s / pixel , @xmath118 is the electron number density , and the function @xmath119 takes into account the atomic physics and the properties of the aia filters .
we have ignored the weak dependence of @xmath120 on @xmath118 , and obtained the temperature dependent function @xmath121 using the solarsoft routine get_aia_response.pro . the resulting modeled emission measure and the aia intensity images in figure [ fig_limbview_40 ] show a central tear shaped void with a relatively sharp `` u '' shaped lower boundary , surrounded by a relatively dense shell .
the central tear shaped void is the central low density , hot channel viewed along its main part of the length ( see figure [ fig_hotchannel_t2 ] ) , and the `` u '' shaped dense shell enclosing the void is produced by the upward extensions of the vertical current layer ( see the orange isosurfaces in figures [ fig_limbview_40]a ) which has formed along the hft . note that the tear - shaped void enclosed by a `` u '' shaped dense shell on top of a vertical dense structure is inside the larger flux rope structure .
such a central density sub - structure inside the flux rope has been observed in some of the quiescent prominence cavities at the limb ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
figure [ fig_obsexample ] shows a prominence cavity observed by sdo / aia on 12 june 2012 at about 23:01 ut on the north - west limb .
this is also the case studied by .
the aia 171 , 193 , 211 images all show an ellipse shaped central void with a sharp `` u '' shaped lower boundary situated on top of the vertical prominence .
the void is enclosed by a `` u '' shaped dense shell extending upward from the prominence .
these qualitative features are similar to those seen in the modeled aia images shown in figure [ fig_limbview_40 ] , provided that the observed prominence material is co - spatial with the vertical current layer in the model ( figure [ fig_limbview_40]a ) .
prominence condensation may develop inside the denser current layer due to radiative cooling , which is not included in our numerical simulation .
the `` u '' shaped dense shell extending upward from the prominence seen most prominently in the aia 171 observations is also called `` horns '' @xcite and is found to be common features of quiescent prominences observed with the aia .
our model suggests that this observed density feature of a dense prominence column with upward extending `` horns '' corresponds to the current sheet that forms along the hft , when viewed from a suitable line of sight angle , as illustrated by figure [ fig_horns ] , which can be compared qualitatively with the the prominence and `` horn '' morphology shown in figure 3 of @xcite . in other words , the development of a dense prominence column with upward extending `` horns '' enclosing a central void ( such as the cases shown in @xcite and @xcite ) signifies the development of the hft topology ( and the associated tether cutting reconnections ) in the cavity flux rope .
figure [ fig_horns ] also shows two field lines drawn from two points in close proximity on opposite sides of one of the upward extensions of the central current layer .
it illustrates the drastic difference in the connectivities of these two field lines which come to close proximity on opposite sides of the current layer extension ( which is a qsl ) with very different field line orientations .
a 3d rotating view of the current layer and the field lines is available as a movie associated with figure [ fig_horns ] in the online version .
the anti - correlation between the emission measure and the temperature for the central void structure ( figures [ fig_limbview_40](b)(c ) ) is also consistent with the recent study of internal thermal properties of coronal cavities by @xcite .
they found a central core with lower density but higher temperature inside the larger coronal cavity .
this central hot core is also consistent with the x - ray emitting core sometimes observed in coronal cavities .
our model suggests that this sub - structure of the hot , low density channel inside the larger flux rope corresponds to the reconnected , twisted flux being added to the flux rope via tether - cutting reconnections in the current layers which have formed along the hft during the quasi - static stage ( see orange surfaces in figure [ fig_3dfdlcurr ] ) .
figure [ fig_reconnect](b ) illustrates such a reconnected field line through the central hot void ( see the red field line ) , which at @xmath105 hour has become a field line that twists about the tracked axis ( black field line ) of the flux rope . yet
, this red field line has its foot points rooted in the arcade regions . at an earlier time @xmath122 hour ( see figure [ fig_reconnect](a ) ) , the field lines drawn from these foot points are two simple arcade field lines ( the red field lines in figure [ fig_reconnect](a ) ) .
thus the two simple arcade field lines at @xmath122 hour have undergone multiple reconnections with the flux rope fields , and by time @xmath105 hour have turn into a field line rooted in the arcade foot points but twists about the axis of the flux rope .
such addition of the twisted flux to the flux rope , threading under the axis , produces the expanding , low density hot channel in the flux rope .
when viewed in a line of sight that is roughly aligned with the channel , the resulting central void is found to enlarge and rise up during the quasi - static stage ( see figure [ fig_limbview_evol ] and the associated movie in the online version of the paper ) .
as the critical height is reached , the central void accelerates rapidly upward and is ejected , leaving behind a cusp shaped brightening ( see the last row in figure [ fig_limbview_evol ] ) .
the development of the long bright current sheet trailing the ejected central void ( see the movie ) , and the rise of the cusp shaped brightening are more prominent in the 193 and 211 channels due to their sensitivity to higher temperatures .
this evolution sequence can be compared with the observed evolution of the prominence cavity studied in @xcite , which begins to erupt at about 03:24 ut on june 13 2010 . a composite movie of the aia observation in the 171
, 193 and 211 channels of the evolution of this cavity on the north - west limb during the period from 18:00 ut 12 june 2010 to 21:30 ut 13 june 2010 can be viewed at http://sdowww.lmsal.com / suntoday / index.html?suntoday_date=2010 - 06 - 12#. the observed evolution show many qualitative similarities with the modeled one , including the morphology of the central void enclosed in the u - shaped dense shell on top of a dense column during the quasi - static phase , and the ejection of the void and the post - eruption brightening .
we have carried out an mhd simulation of the quasi - static rise and the onset of eruption of an anchored coronal flux rope .
earlier in the simulation , the emergence of a twisted flux rope is imposed at the lower boundary into a pre - existing coronal potential arcade field .
then the emergence is stopped at the lower boundary and the coronal magnetic field is allowed to settle into a quasi - static equilibrium .
it is found that the flux rope first settles into a quasi - static rise phase with an underlying sigmoid - shaped current layer developing along the hft , which is argued to be a likely site for the formation of thin current sheets and magnetic reconnections ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
tether cutting reconnections in the current layer effectively add twisted flux to the flux rope and allow it to rise quasi - statically ( even as the free magnetic energy is decreasing ) and the flux rope undergoes dynamic eruption when it reaches the critical height for the onset of the torus instability @xcite . in the present simulation
we study the thermal signatures that result from the development of the hft topology and the associated tether cutting reconnections in the the current sheets that form along the hft .
we solve the total energy equation in conservative form for a perfect gas with an adiabatic index @xmath123 , and thus incorporate the non - adiabatic heating resulting from the dissipation of kinetic and magnetic energies due to the formation of intense current layers .
furthermore , thermal conduction along magnetic field lines is also included in the energy equation so as to incoporate the redistribution of heat along the field lines . in this way
, we are able to identify qualitatively the thermal signatures in the flux rope due to heating produced by the tether - cutting reconnections in the current layers that form during the quasi - static rise phase of the flux rope intense current layers are found to develop during the quasi - static phase along qsls , which are topological boundaries where the magnetic field connectivity undergoes drastic variations and are likely sites for the formation of thin current sheets and significant magnetic reconnections @xcite .
this has been shown also in previous mhd simulations of coronal flux rope evolution as well as nonlinear force - free field models ( nlfff ) of observed sigmoid active regions ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the strongest portions of the current layers that form are initially two `` j '' shaped layers curving about the two legs of the flux rope at the boundaries between the flux rope legs and the arcade with the opposite polarity foot points ( see the left panel of figure [ fig_3dfdlcurr ] ) .
later the strongest portion of the current layer becomes a central inverse s - shaped vertical layer that forms along the hft ( see right panel of figure [ fig_3dfdlcurr ] and the 2d cuts shown in figure [ fig_qfac ] ) .
reconnections in the current layers are found to add to the flux rope twisted flux threading under the axis , creating an expanding , low density hot channel in the flux rope above the central vertical current layer .
when viewed along a line of sight that is roughly aligned with the channel , it manifests as a central hot void on top of a relatively dense structure corresponding to the central warped , vertical current layer , which also extends upward on two sides of the void ( see the @xmath101 image at @xmath104 hour in figure [ fig_merievol ] ) , producing the `` u '' shaped dense shell ( or `` horns '' ) enclosing the void in the modeled euv images ( see e.g. the @xmath124 hour and @xmath125 hour panels of the modeled aia 171 images in figures [ fig_limbview_evol ] ) . the dense vertical column and the upward
extended horns are essentially the manifestation of the hft ( see the 2d cut in the top left panel of figure [ fig_qfac ] ) .
such thermal structures inside the flux rope are consistent with several aia euv observations of coronal prominence cavities where it is found that an elevated , ellipse or tear - shaped central void is situated on top of the vertical column of dense prominence , and the void is enclosed by a `` u '' shaped dense shell ( sometimes called `` horns '' ) extending upward from the prominence column ( e.g. * ? ? ?
* ; * ? ? ?
the central hot channel produced by the tether cutting reconnections can also explain the observational result of a central hot core with low emission measure inside a coronal cavity by @xcite .
although our numerical simulation does not include radiative cooling and thus can not model the process of prominence formation , it is likely that the prominence plasma is co - spatial with the dense current layer ( e.g. * ? ? ? * ; * ? ? ?
our mhd simulation also does not model explicitly the background coronal heating but instead simply assume a high temperature coronal plasma with an adiabatic index of @xmath123 , without explicitly incorporating the coronal heating and radiative cooling processes .
however our model does take into account localized heating due to the formation and dissipation of current sheets in the corona and the redistribution of the heat by field aligned conduction in the corona .
the lower boundary of our simulation domain is placed at the coronal base instead of the chromosphere , and after the emergence is stopped , the velocity field is set to zero on the lower boundary and no more mass flux is allowed through it into the coronal domain .
this precludes the important effect of chromospheric evaporation due to conductive heat flux coming down from the corona , which may significantly affect the density of the heated coronal flux tubes .
thus there are great uncertainties in the density and temperature obtained due to our still highly simplified treatment of the thermodynamics , especially for the heated , twisted field lines in the hot channel that form as a result of the tether cutting reconnections .
one main question would be whether the heated field in the hot channel would remain a void ( i.e. of lower density ) in contrast to the surrounding as is in our model when the effect of chromospheric evaporation produced by the downward heat conduction is taken into account . in regard to this question ,
one consideration is that the reconnected field lines in the hot channel are generally long field lines ( see e.g. the red field lines in panel [ fig_reconnect](b ) and figure [ fig_horns ] ) .
the scaling law derived based on the static , thermal equilibrium coronal loop model of @xcite gives @xmath126 , where @xmath127 , @xmath13 and @xmath128 denote respectively the maximum loop temperature , the loop pressure , and the loop length .
this would mean that the plasma density of the loop @xmath129 .
thus even though the field lines in the hot core are heated to a mildly increased @xmath127 : to roughly 1.2 mk from the 1 mk original coronal temperature , these field lines are on the other hand considerably longer in @xmath128 compared to the field lines immediately outside the hot core : for example , in figure [ fig_horns ] , the red field line inside the central void is about 1.9 times the blue field line immediately outside of the horn .
therefore the above scaling law of static thermal equilibrium coronal loops would indicate that the effect of the longer length outweighs the effect of the high temperature , and the resulting density inside the central hot channel remains smaller compared to the surrounding , even with the effect of the chromospheric evaporation included .
our 3d mhd simulations with the simplified thermodynamics is an initial step to understand _ qualitiatively _ the immediate thermal effect of the formation and dissipation of current sheets along the hft in the corona .
our results suggest that the observed feature of prominence with upward extending `` horns '' enclosing an elevated central void seen in aia 171 observations of coronal cavities is a signature of the development of the hft topology and the associated tether - cutting reconnections in the current sheets that form along the hft .
it is likely that the current sheet forming along the hft remains the region of the highest density because of the significant local compression by the lorentz force even when more realistic treatment of the thermodynamics are included , and our results about the thermal sigature , i.e. relatively hot , low density central void on top of and surrounded by the higher density current sheet formed along the hft , remain qualitatively true . of course 3d
mhd simulations with a full treatment of the thermal dynamics that incorporates explicit coronal heating , radiative cooling , as well as field aligned thermal conduction , and with the lower boundary set at the chromosphere , are needed to confirm the findings here and provide quantitative determination of the density and temperature of the coronal plasma .
as has been found in 1d hydrodynamic simulations with a full treatment of the thermodynamics along long dipped coronal loops ( e.g. * ? ? ?
* ) , complex , thermal nonequilibrium formation and dynamic evolution of prominence condensations can develop along the coronal loops depending on the spatial and temporal properties of the heating profiles prescribed .
here we have presented the mhd simulation of the quasi - static evolution and onset of eruption of a particular coronal flux rope configuration .
however the results presented here of the development of the hft topology , and the associated current sheet formation and tether cutting reconnections along the hft are quite general results for the evolution of line - tied coronal flux ropes as has been demonstrated by several previous mhd simulations as well as nlfff models of sigmoid active regions based on observations ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
these models suggest that the development and the elevation of the hft may be an indication of the extent to which the tether - cutting reconnections have progressed and hence the readiness of the flux rope to erupt .
i thank b.c .
low for reading and helpful comments on the manuscript .
i thank the anonymous referee for critical review of the paper .
ncar is sponsored by the national science foundation .
this work is supported by the nasa lws tr&t grant nnx09aj89 g to ncar .
the numerical simulations were carried out on the pleiades supercomputer at the nasa advanced supercomputing division under project gid s0969 . , g. and pariat , e. and dmoulin , p. 2005
, , 444 , 961 aulanier , g. , trk , t. , demoulin , p. , and deluca , e. 2010 , , 708 , 314 .
bateman , g. , _ mhd instabilities _ , 1978 , mit press , cambridge , ma , and london , england , p. 84 - 85
berger , t. 2012 , in _
2nd atst - east workshop in solar physics : magnetic fields from the photosphere to the corona _ , asp conf .
series , editors : thomas rimmele et al .
, in press .
chen , p. f. 2011 , living rev .
solar phys .
, 8 , ( 2011 ) , 1 .
url ( cited on may 2011 ) : http://www.livingreviews.org/lrsp-2011-1 cottaar , m. , and fan , y. 2009 , , 704 , 576 , p. and henoux , j. c. and priest , e. r. and mandrini , c. h. 1996 , , 308 , 643 , p. and priest , e. r. and lonie , d. p. 1996
, , 101 , 7631 dmoulin , p. , and aulanier , g. 2010 , , 718 , 1388 fan , y. , 2009 , , 697 , 1529 fan , y. , 2010 , , 719 , 728 fan , y. , 2011 , , 740 , 68 fan , y. , & gibson , s. e. , 2007 , , 668 , 1232 fan , y. , & gibson , s. e. , 2004 , , 609 , 1123 forbes , t. g. , and 14 other authors , 2006 , space sciences reviews , 123 , 251 gibson , s. e. , and fan , y. 2006 , jgr , 111 , a12103 hudson , h. s. , acton , l. w. , harvey , k. l. , and mckenzie , d. e. 1999 , , 513 , l83 isenberg , philip a. , and forbes , terry g. 2007 , , 670 , 1453 karpen , j. t. and antiochos , s. k. 2008 , 676 , 658 kliem , b. , and trk , t. 2006 , phys .
lett . , 96 , 255002 low , b. c. , 2001 , , 10625141 low , b. c. , and zhang , m. 2002 , , 564 , l53 low , b. c. , berger , t. , casini , r. , and liu , w. 2012 , , in press , http://adsabs.harvard.edu/abs/2012arxiv1203.1056l , e. and dmoulin , p. 2012
, , 541 , a78 rempel , m , schssler , m. , and knlker , m. , 2009 , , 691 , 640 reeves , katharine k. , gibson , sarah e. , kucera , therese a. , hudson , hugh s. , and kano , ryouhei , 2012 , , 746 , 146 rgnier , s. , walsh , r. w. , and alexander , c. e. 2011 , a&a , 533 , l1 rosner , r. , tucker , w. h. , and vaiana , g. s. 1978 , , 220 , 643 , a. and van ballegooijen , a. a. and deluca , e. e. 2012 , , 744 , 78 , a. and pariat , e. and van ballegooijen , a. and aulanier , g. and deluca , e. 2012 , , 750 , 15 stone , j.m . , and norman m.l . , 1992 , , 80 , 753 stone , j.m . , and norman m.l . , 1992 , , 80 , 791 , v. s. and hornig , g. and dmoulin , p. 2002
, , 107 , 1164 , v. s. 2007 , , 660 , 863 titov , v. s. , and demoulin p. , 1999 ,
a&a , 351 , 707 trk t. & kliem b. 2007 , astron .
nachr . , 328 , 743 van ballegooijen , a. a , and cranmer , s. r. 2010 , , 711 , 164 wang , y .- m . , and stenborg , g. 2010 : , 719 , 181 | using a 3d mhd simulation , we model the quasi - static evolution and the onset of eruption of a coronal flux rope .
the simulation begins with a twisted flux rope emerging at the lower boundary and pushing into a pre - existing coronal potential arcade field . at a chosen time
the emergence is stopped with the lower boundary taken to be rigid .
then the coronal flux rope settles into a quasi - static rise phase during which an underlying , central sigmoid - shaped current layer forms along the so called hyperbolic flux tube ( hft ) , a generalization of the x - line configuration .
reconnections in the dissipating current layer effectively add twisted flux to the flux rope and thus allow it to rise quasi - statically , even though the magnetic energy is decreasing as the system relaxes .
we examine the thermal features produced by the current layer formation and the associated `` tether - cutting '' reconnections as a result of heating and field aligned thermal conduction .
it is found that a central hot , low - density channel containing reconnected , twisted flux threading under the flux rope axis forms on top of the central current layer . when viewed in the line of sight roughly aligned with the hot channel ( which is roughly along the neutral line )
, the central current layer appears as a high - density vertical column with upward extensions as a `` u '' shaped dense shell enclosing a central hot , low - density void .
such thermal features have been observed within coronal prominence cavities .
our mhd simulation suggests that they are the signatures of the development of the hft topology and the associated tether - cutting reconnections , and that the central void grows and rises with the reconnections , until the flux rope reaches the critical height for the onset of the torus instability and dynamic eruption ensues . |
the cdf collaboration has recently announced the observation of an excess of events containing a lepton ( electron or muon ) , missing transverse energy , and two jets @xcite .
this can be interpreted as the production of a new state with a mass of @xmath3150 - 160 gev which decays to quarks or gluons , along with a standard model @xmath4 decaying to an electron or muon along with a neutrino .
this excess of events , which was observed in 7.3 fb@xmath5 of cdf data , has been reported to be statistically significant at a level of 4.1@xmath6 .
although the d0 collaboration does not find evidence of such a sizable excess in their analysis of 4.3 fb@xmath5 of data @xcite , a signal roughly half as large as reported by cdf might be consistent with both experimental results .
although several explanations have been proposed for the cdf excess @xcite , perhaps the simplest is the introduction of a new gauge boson @xcite . in order to evade constraints from lep ii and the tevatron ,
however , such a state must have couplings to leptons ( or more specifically , to electrons and muons ) which are suppressed by at least a factor of 4 - 5 relative to those to quarks @xcite .
a gauge boson with these characteristics is commonly known as a leptophobic @xmath1 .
new @xmath7 gauge groups and the @xmath1 bosons which accompany them are found in many phenomenologically viable and well motivated extensions of the standard model @xcite .
such extensions introduce anomalies , however , whose cancellation requires the addition of new fermions . the sole exception to
this rule is a @xmath1 which couples to baryon minus lepton number .
thus any anomaly - free , flavor universal model that contains a leptophobic @xmath1 must invariably contain new fermions .
additional @xmath7 gauge groups are a feature common to many grand unified theories ( guts ) , including those based on the groups @xmath8 and @xmath0 ( the @xmath9 group is of the same rank as the standard model , and thus does not contain any additional @xmath7 s ) .
the @xmath1 which arises in @xmath8 models , however , is not leptophobic .
as the @xmath0 group is rank 6 ( two higher than the gauge group of the standard model ) , guts based on this group can include two additional @xmath7 s , leading to a variety of possible charge assignments and interactions @xcite .
for this reason , and as @xmath0 is well motivated within the context of @xmath10 string theory @xcite , we will focus on supersymmetric guts built upon this gauge group .
models are of course possible . however , for this initial study , we choose this more constrained scenario . ] for a review of @xmath0 models or of guts in general , see refs .
@xcite and @xcite , respectively .
@xmath0 can be broken such that @xmath11 , of which the @xmath9 contains the gauge group of the standard model .
the charges of any @xmath12 and the corresponding @xmath1 boson that appears as part of the low - energy theory can be written as a linear combination of the charges of the two additional @xmath7 s in @xmath0 : @xmath13 although for no choice of @xmath14 is @xmath15 completely leptophobic , for the well motivated value of @xmath16 ( known as the @xmath17-model of @xmath0 @xcite , such that @xmath18 ) , kinetic mixing can lead to a @xmath1 with highly suppressed couplings to leptons @xcite . in particular , although the @xmath19 and @xmath20 gauge groups undergo no kinetic mixing at the @xmath0 breaking scale , if the effective particle content at some lower energy scale does not contain full representations of @xmath0 , kinetic mixing will be generated by renormalization group ( rg ) evolution @xcite .
the full higgs sector is expected to be split in mass , as some of the scalar fields will have gut - scale vacuum expectation values ( vevs ) .
thus , during the rg running down to the weak scale , the particle content is such that a non - zero mixing parameter will evolve . [
cols="<,^,^,^,^,^,^,^,^,^,^",options="header " , ] after accounting for the effects of kinetic mixing , the effective charges of the fermions to the @xmath1 are given by @xmath21 , where the values of @xmath22 , @xmath23 and @xmath24 , are given in table i for all of the fermions in the fundamental 27-dimensional representation of @xmath0 .
for the choice of @xmath25 , the @xmath1 is completely leptophobic ( @xmath26 ) .
the size of @xmath27 is highly dependent on the composition and mass scale of the higgs sector ; while the minimal fields required to give masses to the matter * 27*s are not sufficient to generate @xmath28 , we would not expect these to be the only fields of relevance all the way up to the unification scale . in ref .
@xcite , for example , a reasonable higgs sector was described which naturally leads to @xmath29 . for purposes of explaining the cdf @xmath30dijets
excess , we do not necessarily require complete leptophobia , and values in the range of @xmath31 are acceptable .
, one may also consider models in which baryon number and lepton number are gauged @xcite . alternatively
, a @xmath1 could acquire couplings to standard model quarks through higher dimensional effective operators @xcite . ] throughout the remainder of this paper , we carry out our calculations using the choice of @xmath25 . with mixing between @xmath32 and @xmath17
, the photon , @xmath33 and @xmath1 will be linear combinations of the neutral gauge bosons of the @xmath19 , @xmath34 , and @xmath35 gauge groups .
this induces corrections to the electroweak precision observables , which place limits on leptophobic @xmath0 models from lep and other precision measurements .
although constrained , some regions of mixing parameter space remain allowed for a leptophobic @xmath1 of mass @xmath3150 gev @xcite .
in this section , we look to the matter content of @xmath0 guts for a viable candidate for the dark matter of our universe . in @xmath0 ,
the fermions of the standard model come as part of three copies ( one per generation ) of the 27-dimensional fundamental representation .
each * 27 * of @xmath0 decomposes into a * 16 * , * 10 * and * 1 * of @xmath8 . the fermions of the standard model , along with a singlet @xmath36 , fill the * 16 * , and are further decomposed into a * 10 * ( @xmath37 , @xmath38 , @xmath39 ) , @xmath40 ( @xmath41 , @xmath42 ) , and * 1 * ( @xmath36 ) of @xmath9 . in this context
, @xmath36 may be viewed as a right handed neutrino .
in addition , the * 27 * contains `` exotic '' fermions occupying the * 10 * and * 1 * of @xmath0 .
the 10 dimensional representation decomposes into a * 5 * ( @xmath43 , @xmath44 ) and @xmath45 ( @xmath46 , @xmath47 ) of @xmath9 . the state @xmath48 is a singlet under both @xmath8 and @xmath9 .
thus the exotics of the * 27 * consist of a vector - like , color singlet , weak isodoublet with electric charges of @xmath49 ( @xmath46 , @xmath43 ) , a vector - like color triplet , weak isosinglet with @xmath50 ( @xmath44 , @xmath47 ) , and two neutral singlets ( @xmath36 and @xmath48 ) .
see table i for a decomposition of the * 27 * along with their gauge charges .
the matter fields in the * 27 * get mass terms through interactions with higgs fields appearing in additional representations of @xmath0 . in the breaking scheme of the @xmath0 @xmath17-model ,
at least four higgs fields are required @xcite .
two of these are @xmath35 isodoublets , corresponding to the up- and down - type higgses of the mssm , with hypercharges of @xmath51 and @xmath52 . additionally , these two fields must have @xmath24 ( @xmath53 ) charges of @xmath54 and @xmath55 , respectively .
the remaining two higgs fields serve to give masses to the exotics in the * 27 * s ; these scalars must be isosinglets with @xmath56 and have @xmath17 charges @xmath57 and @xmath58 , giving @xmath53 charges of + 1 and + 2 .
it is the vevs of these higgs that can serve to break @xmath53 and give mass to the leptophobic @xmath1 . from the mass of the @xmath1
, we expect the new higgs vevs to be on the order of the weak scale . among these exotics ,
both @xmath36 and @xmath48 are not colored , are electrically neutral , and thus are potentially viable candidates for dark matter .
furthermore , they each carry large charges of the leptophobic gauge group ( @xmath59 ) , potentially providing an important role for the @xmath1 in cosmology as well as in collider physics .
charges as normalized in table i , absorbing the factor of @xmath60 into the definition of the gauge coupling constant . ] we take our dark matter candidate to be the scalar superpartner of the lightest right - handed neutrino @xmath36 , which will be stable by the virtue of @xmath61-parity ( this dark matter candidate has previously been considered in ref .
we could also consider the superpartner of the @xmath48 , which is identical for the purposes of direct detection , but is disfavored by relic abundance considerations as described in sec .
[ ann ] . to retain a degree of agnosticism toward this choice ( or a linear combination of these choices ) , we will refer to the scalar dark matter particle simply as @xmath62 .
in the @xmath0 @xmath17-model with @xmath29 , the light @xmath1 couples to standard model quarks with charges approximately given by : @xmath63 , @xmath64 , @xmath65 , @xmath66 , but with little or no couplings to any standard model leptons . writing the effective couplings as the product of these charges and a new gauge coupling , @xmath67 , the rate observed at cdf
favors a value of roughly @xmath68 @xcite . to bring this into better compatibility with the null results of d0 @xcite , however , we will adopt a somewhat smaller value of @xmath69 . through @xmath1 exchange ,
the dark matter can scatter elastically with nuclei @xcite .
the cross section for spin - independent scattering of our scalar dark matter candidate with @xmath70 is given by : @xmath71 ^ 2,\end{aligned}\ ] ] where @xmath72=0 , \\ f_n&=&\frac{g^2_{z'}}{m^2_{z'}}\bigg[\frac{1}{2}(q'_{u_l}+q'_{u_r})+ ( q'_{d_l}+q'_{d_r})\bigg]=\frac{-g^2_{z'}}{2 m^2_{z'}}.\nonumber \end{aligned}\ ] ] in other words , the couplings to up and down quarks cancel for the proton , leaving only neutrons with which the dark matter can scatter . with this substitution
, we arrive at @xmath73 to compare with the results of direct detection experiments , we calculate the spin - independent elastic scattering cross section of our dark matter candidate with neutrons ( not to be confused with nucleons , as is often presented ) : @xmath74 in fig .
[ direct ] , we compare this result to the cross section and mass required to explain the signals reported by the cogent @xcite and dama / libra @xcite collaborations ( see also refs .
@xcite ) . for dark matter with a mass of approximately 6 - 8 gev ( and for a velocity distribution described by @xmath75 km / s and @xmath76 km / s ) ,
the signals reported by these two experiments can be accommodated by the @xmath0 gut model discussed here .
we also show for comparison the constraints as presented by the cdms @xcite ( dashed ) and xenon100 @xcite ( dotted ) collaborations .
although the results from cdms appear to be in some degree of tension with a dark matter interpretation of cogent and dama / libra , it is not necessarily implausible that systematic uncertainties could open a region of compatibility @xcite .
the constraints placed on light dark matter particles by xenon100 ( and other liquid xenon detectors ) depend critically on the scintillation efficiency of liquid xenon at the lowest measured energies , and may well be less stringent than have been reported @xcite .
in this section , we turn our attention to the annihilation of the dark matter candidate appearing in the @xmath17-model of @xmath0 .
first , we consider annihilations through the @xmath77-channel exchange of the @xmath1 , which leads to a cross section given by : @xmath78,\end{aligned}\ ] ] where @xmath79 is the relative velocity between two dark matter particles and the sum is performed over all favors of left and right - handed quarks lighter than @xmath80 .
note that the annihilation cross section is velocity suppressed ( i.e. , @xmath81 ) . to determine the thermal relic abundance of dark matter
, we consider the thermally weighted cross section at the temperature of freeze - out : @xmath82 where @xmath83 is the ratio of @xmath80 to the freeze - out temperature .
the dark matter s annihilation cross section is related to its thermal relic abundance according to @xmath84 annihilations of dark matter through the @xmath1 thus provide only a small fraction ( @xmath35% ) of the cross section required to yield our universe s observed dark matter density .
therefore processes other than @xmath1 exchange must contribute to the dark matter s annihilation in the early universe .
the most general renormalizable superpotential describing the interactions within the * 27 * that is consistent with the symmetries of the standard model is given by : @xmath85 depending on the lepton and baryon number assignments of the exotic particle content of the * 27 * , some of these terms will need to be suppressed in order to maintain approximate low - scale lepton and baryon number conservation . among these 11 terms
, there are no interactions which directly connect the @xmath48 ( or its superpartner ) to the fields of the standard model .
for this reason , we will focus our attention on the case in which the dark matter is the superpartner of @xmath36 .
the interactions described by the terms @xmath86 and @xmath87 allow the dark matter to annihilate into down - type quarks through the @xmath88-channel exchange of a @xmath44 or to leptons through the @xmath88-channel exchange of a @xmath89 or @xmath90 .
as null results from the lhc force states with qcd color such as @xmath44 to be quite heavy ( above several hundred gev ) , we focus on the case of annihilation to charged and neutral leptons through a @xmath89 or @xmath90 , respectively .
assuming for simplicity that @xmath91 , @xmath92 and @xmath93 have similar values for each generation ( in each * 27 * ) , the dark matter s annihilation cross section to charged leptons and neutrinos is approximately given by @xcite : @xmath94 , \label{tch}\end{aligned}\ ] ] where the first term proceeds to @xmath95 while the terms in the sum proceeds to a mixture of charged leptons and neutrinos .
in the low - velocity limit , the annihilation cross section to @xmath95 given by : @xmath96 thus for the purposes of indirect detection , the dark matter will annihilate primarily to @xmath95 , possibly with an order @xmath310% contribution to @xmath97-quarks through @xmath88-channel @xmath44 exchange .
we note that in order for this annihilation process to proceed , either ( but not both of ) @xmath36 or @xmath43 must carry lepton number . after thermally averaging at the temperature of freeze - out , the @xmath98 terms of the cross section given in eq .
[ tch ] becomes : @xmath99 combining this with the results from eqs . 7 and 11 , we find a total value which is very similar to that required to thermally generate the observed density of dark matter .
the leptophilic annihilations of the dark matter is of particular interest within the context of gamma ray observations of the galactic center . in fig .
[ indirect ] , we compare the spectrum of gamma rays from dark matter annihilations in the galactic center predicted in our model to the excess emission observed by the fermi gamma ray space telescope @xcite ( assuming an extrapolation of the spectrum from the milky way s supermassive black hole ; see also ref .
@xcite ) . clearly , very good agreement is found .
dark matter in the mass range described here and which annihilates primarily to charged leptons has also been previously shown to be able to produce the observed synchrotron emission known as the wmap haze @xcite ( see also ref .
lastly we note the constraints from searches for gev - scale neutrinos from the sun as conducted by super - kamiokande @xcite are safely evaded in this model by virtue of the dark matter s suppressed couplings to protons .
in summary , motivated by the @xmath2+dijets excess recently reported by the cdf collboration , we have considered grand unified models based on the @xmath0 group . such models ( in particular , the so - called @xmath17-model of @xmath0 ) , can contain a leptophobic @xmath1 boson capable of producing the observed excess . within the fundamental representation of @xmath0 is an attractive dark matter candidate in the form of a right - handed sneutrino . from the sizable couplings of this dark matter candidate to the @xmath1 , a large elastic scattering cross section with neutrons is predicted , consistent with the direct detection signals reported by the cogent and dama / libra collaborations .
furthermore , to avoid being overproduced in the early universe , the dark matter must annihilate to leptons through the exchange of other particles contained within the fundamental representation of @xmath0 , leading to a gamma ray spectrum and flux consistent with that observed from the galactic center by the fermi gamma ray space telescope .
x. p. wang , y. k. wang , b. xiao , j. xu , s. h. zhu , [ arxiv:1104.1161 [ hep - ph ] ] ; k. cheung , j. song , [ arxiv:1104.1375 [ hep - ph ] ] ; x. p. wang , y. k. wang , b. xiao , j. xu and s. h. zhu , arxiv:1104.1917 [ hep - ph ] ; l. a. anchordoqui , h. goldberg , x. huang , d. lust , t. r. taylor , [ arxiv:1104.2302 [ hep - ph ] ] ; a. e. nelson , t. okui , t. s. roy , [ arxiv:1104.2030 [ hep - ph ] ] .
e. j. eichten , k. lane , a. martin , [ arxiv:1104.0976 [ hep - ph ] ] ; c. kilic , s. thomas , [ arxiv:1104.1002 [ hep - ph ] ] ; b. a. dobrescu and g. z. krnjaic , arxiv:1104.2893 [ hep - ph ] ; r. sato , s. shirai , k. yonekura , [ arxiv:1104.2014 [ hep - ph ] ] .
m. b. green , j. h. schwarz , phys .
* b149 * , 117 - 122 ( 1984 ) ; d. j. gross , j. a. harvey , e. j. martinec , r. rohm , phys . rev .
lett . * 54 * , 502 - 505 ( 1985 ) ; d. j. gross , j. a. harvey , e. j. martinec , r. rohm , nucl .
* b256 * , 253 ( 1985 ) ; p. candelas , g. t. horowitz , a. strominger , e. witten , nucl .
* b258 * , 46 - 74 ( 1985 ) .
k. s. babu , c. f. kolda and j. march - russell , phys .
d * 54 * , 4635 ( 1996 ) [ arxiv : hep - ph/9603212 ] . c. f. kolda , nucl .
suppl . * 52a * , 120 - 126 ( 1997 ) .
[ hep - ph/9606396 ] ; j. l. rosner , phys .
lett . * b387 * , 113 - 117 ( 1996 ) .
[ hep - ph/9607207 ] ; v. d. barger , n. g. deshpande , k. whisnant , phys .
lett . * 56 * , 30 ( 1986 ) ; v. d. barger , n. g. deshpande , j. l. rosner , k. whisnant , phys .
* d35 * , 2893 ( 1987 ) .
m. buckley , p. fileviez perez , d. hooper , e. neil , [ arxiv:1104.3145 [ hep - ph ] ] ; p. fileviez prez and m. b. wise , phys .
d * 82 * ( 2010 ) 011901 [ erratum - ibid .
d * 82 * ( 2010 ) 079901 ] [ arxiv:1002.1754 [ hep - ph ] ] ; t. r. dulaney , p. fileviez perez , m. b. wise , phys .
* d83 * ( 2011 ) 023520 .
[ arxiv:1005.0617 [ hep - ph ] ] .
a. l. fitzpatrick , d. hooper , k. m. zurek , phys .
rev . * d81 * , 115005 ( 2010 ) .
[ arxiv:1003.0014 [ hep - ph ] ] ; m. beltran , d. hooper , e. w. kolb , z. c. krusberg , phys .
* d80 * , 043509 ( 2009 ) .
[ arxiv:0808.3384 [ hep - ph ] ] ; c. boehm , p. fayet , nucl .
* b683 * , 219 - 263 ( 2004 ) .
[ hep - ph/0305261 ] .
d. hooper , j. i. collar , j. hall , d. mckinsey , phys .
rev . * d82 * , 123509 ( 2010 ) .
[ arxiv:1007.1005 [ hep - ph ] ] ; d. hooper , c. kelso , [ arxiv:1106.1066 [ hep - ph ] ] ; c. kelso , d. hooper , jcap * 1102 * , 002 ( 2011 ) . [ arxiv:1011.3076 [ hep - ph ] ] .
g. dobler , d. p. finkbeiner , astrophys .
j. * 680 * , 1222 - 1234 ( 2008 ) [ arxiv:0712.1038 [ astro - ph ] ] ; d. p. finkbeiner , arxiv : astro - ph/0409027 ; d. hooper , d. p. finkbeiner and g. dobler , phys .
d * 76 * , 083012 ( 2007 ) [ arxiv:0705.3655 [ astro - ph ] ] . | we explore the phenomenology of grand unified models based on the @xmath0 group , focusing on the @xmath1 with suppressed couplings to leptons that can appear in such models .
we find that this @xmath1 can accommodate the @xmath2+dijets anomaly reported by the cdf collaboration .
furthermore , a viable dark matter candidate in the form of a right - handed sneutrino is also present within the fundamental 27-dimensional representation of @xmath0 . through its sizable couplings to the @xmath1 , the dark matter is predicted to possess an elastic scattering cross section with neutrons which can generate the signals reported by the cogent and dama / libra collaborations . to avoid being overproduced in the early universe ,
the dark matter must annihilate to leptons through the exchange of charged or neutral fermions which appear in the * 27 * of @xmath0 , providing an excellent fit to the gamma ray spectrum observed from the galactic center by the fermi gamma ray space telescope . |
we begin by quoting the introduction from the groundbreaking paper @xcite of eisenbud and harris : * `` one of the most potent methods in the theory of ( complex projective algebraic ) curves and their linear systems since the work of castelnuovo has been degeneration to singular curves ... most problems of interest about curves are , or can be , formulated in terms of ( families of ) linear series .
thus , in order to use degenerations to reducible curves for studying smooth curves , it is necessary to understand what happens to linear series in the course of such a degeneration , and in particular , to understand what structure on a reducible curve plays the part of a linear series . ''
this has been very successful in practice and has led to important advances in the study of algebraic curves ( see @xcite for an overview ) ; we mention for example brill - noether theory @xcite , the geometry of the moduli space of curves @xcite , and weierstrass points and their monodromy @xcite . in view of the deligne - mumford compactification of @xmath11 and the stable reduction theorem @xcite
, it is particularly natural to consider degeneration of smooth curves to stable ( or more generally semistable ) curves .
eisenbud and harris were able to settle a number of longstanding open problems with their theory of limit linear series on such curves , and the theory has undergone numerous developments and improvements in the last 25 years
. however , the theory of limit linear series only applies , for the most part , to a rather restricted class of reducible curves , namely those of _ compact type _
( i.e. , nodal curves whose dual graph is a tree ) . in @xcite
, the second author introduced a new framework for degenerating linear series on curves , degenerating linear series on a regular semistable family of curves to a linear series on the _ dual graph _ of the special fiber .
this theory , which is closely related to tropical geometry and also to the theory of berkovich analytic spaces , is more or less orthogonal to the eisenbud - harris theory , in that it works best for special fibers which are _ maximally degenerate _ , meaning that the dual graph has first betti number equal to the genus of the generic fiber .
specializing linear series to the dual graph provides no information whatsoever when the special fiber is of compact type . intriguingly , both the eisenbud - harris theory and the second author s theory from @xcite lead to simple proofs of the celebrated brill - noether theorem of griffiths - harris ( see @xcite ) .
the aim of the present paper is to introduce a theoretical framework suitable for generalizing both the eisenbud - harris theory of limit linear series and the second author s theory of specialization from curves to graphs .
the main new object of study is what we call a _
metrized complex of algebraic curves _ over an algebraically closed field @xmath0 ; this is , roughly speaking , a finite metric graph @xmath1 together with a collection of marked complete nonsingular algebraic curves @xmath2 over @xmath0 , one for each vertex @xmath3 ( with a marked point on @xmath2 for each edge of @xmath1 incident to @xmath3 ) .
we define divisors , linear equivalence of divisors , and rank of divisors on metrized complexes of curves and establish a riemann - roch theorem for metrized complexes of curves which combines the classical riemann - roch theorem over @xmath0 with its graph - theoretic and tropical analogues proved in @xcite . for a curve @xmath4 defined over a non - archimedean field @xmath5 , together with a strongly semistable model @xmath6 for @xmath4 over the valuation ring @xmath7 of @xmath5 ,
we define a corresponding metrized complex @xmath8 of curves over the residue field @xmath0 of @xmath5 and a specialization map @xmath9 from divisors on @xmath4 to divisors on @xmath8 preserving degrees and linear equivalence .
we then establish generalizations of the specialization lemma from @xcite and its weighted graph analogue from @xcite , showing that the rank of a divisor can not go down under specialization from @xmath4 to @xmath8 . as an application
, we establish a concrete link between specialization of divisors from curves to metrized complexes and the eisenbud - harris theory of limit linear series . using this link
, we formulate a generalization of the notion of limit linear series to curves which are not necessarily of compact type and prove , among other things , that any degeneration of a @xmath10 in a regular family of semistable curves is a limit @xmath10 on the special fiber .
as mentioned above , the specialization theory from @xcite works best in the case where the family is maximally degenerate ; in other cases , the dual graph of the family forgets too much information .
this loss of information was partially remedied in @xcite by looking at vertex - weighted graphs , where the weight attached to a vertex is the genus of the corresponding irreducible component of the special fiber .
the theory of metrized complexes of curves offers a richer solution in which one keeps track of the ( normalization of the ) irreducible components @xmath2 of the special fiber themselves , and not just their genera .
from the point of view of non - archimedean geometry , this corresponds to retracting divisors on @xmath4 to the skeleton @xmath12 , a metric graph canonically embedded in the berkovich analytic space @xmath13 , but also keeping track , for points retracting to a vertex of @xmath12 , of the tangent direction in which the vertex is approached .
this is quite natural from the valuation - theoretic point of view : the metric graph @xmath12 associated to a strongly semistable model @xmath6 is a canonical subset of the berkovich analytic space @xmath13 , and the curve @xmath2 over @xmath0 is naturally identified ( for each @xmath14 ) with the fiber over @xmath3 of the canonical retraction map @xmath15 ( which sends a continuous valuation of arbitrary rank in the huber adic space @xmath16 to its canonical rank-1 generalization ) , c.f .
* remark 2.6 ) . though we could possibly ( with quite a bit of additional effort ) have formulated many of our theorems and proofs without mentioning berkovich spaces , it would have resulted in a significant loss of elegance and clarity , and in any case berkovich s theory seems ideally suited for the point of view taken here .
we do not , however , require any non - trivial facts from huber s theory of adic spaces , so we will not mention them again even though they are certainly lurking in the background .
there are a number of connections between the ideas in the present paper and tropical geometry ; these will be explored in more detail in future work .
for example , the theory of morphisms between metrized complexes of curves studied in @xcite sheds interesting new light on the question of which morphisms between tropical curves are liftable .
the theory developed in @xcite shows that there is a close connection between metrized complexes of curves and `` exploded tropicalizations '' in the sense of ( * ? ? ?
* definition 2.9 ) ; we plan to say more about this in the future as well .
as noted by payne in @xcite , exploded tropicalizations can be thought of as an algebraic analogue of the `` exploded torus fibrations '' studied by parker from a symplectic viewpoint @xcite .
it could be interesting to explore connections between _ loc .
cit . _ and the present work .
the ideas in the present paper also have diophantine applications to the study of rational points on curves over number fields ( specifically , to the method of coleman - chabauty ) .
indeed , eric katz and david zureick - brown have recently proved a result similar to theorem [ thm : mcclifford ] below ( clifford s theorem for metrized complexes ) , and they use this result to answer a question of m. stoll .
a special case of the main result in @xcite is the following .
let @xmath4 be a smooth projective geometrically irreducible curve of genus @xmath17 over @xmath18 and assume that the mordell - weil rank @xmath19 of the jacobian of @xmath4 is less than @xmath20 . fix a prime number @xmath21 and let @xmath6 be a proper ( not necessarily semistable ) regular model for @xmath4 over @xmath22 .
then ( letting @xmath23 denote the smooth locus of @xmath24 ) @xmath25 we discuss the theorem of katz and zureick - brown in more detail in section [ chabautysection ] below , explaining how limit linear series on metrized complexes of curves can be used to illuminate their proof and put it into a broader context .
we set the following notation , which will be used throughout this paper . a finite edge - weighted graph with vertex set @xmath26 and edge set @xmath27 a compact metric graph ( geometric realization of an edge - weighted graph @xmath28 ) a complete and algebraically closed non - trivially valued non - archimedean field the valuation ring of @xmath5 the residue field of @xmath5 a smooth , proper , connected algebraic curve defined over @xmath5 the berkovich analytification of @xmath29 a proper and flat semistable @xmath7-model for @xmath4 , usually assumed to be strongly semistable a smooth , proper , connected algebraic curve over @xmath30 ( one for each vertex @xmath31 ) a metrized complex of algebraic curves over @xmath30 the geometric realization of @xmath32 the metrized complex associated to a strongly semistable @xmath7-model @xmath33 of @xmath4 a nodal curve defined over @xmath30 an irreducible component of @xmath34 the regularization of the nodal curve @xmath34 the group of divisors on the metrized complex @xmath32 a divisor on @xmath32 an effective divisor on @xmath32 a rational function on @xmath32 a divisor on @xmath2 an effective divisor on @xmath2 a rational function on @xmath1 , usually the @xmath1-part of a rational function @xmath35 on @xmath32 a rational function on @xmath2 , usually the @xmath2-part of a rational function @xmath35 on @xmath32 the outgoing slope of @xmath36 along an edge ( or tangent direction ) @xmath37 the rank function on @xmath38 the restricted rank function with respect to a collection @xmath39 of subspaces of @xmath40 the retraction map from @xmath41 to the skeleton @xmath42 of @xmath33 the specialization map from @xmath43 to @xmath44 the marked points of @xmath2 in the metrized complex @xmath32 the point of @xmath45 corresponding to the edge ( or tangent direction ) @xmath37 the sum in @xmath46 of the points of @xmath45 we now discuss the contents of this paper in more detail .
an _ edge - weighted graph _
@xmath28 is a connected multigraph , possibly with loop edges , having vertex set @xmath26 and edge set @xmath27 , and endowed with a weight ( or length ) function @xmath47 .
metric graph _
@xmath1 is the geometric realization of an edge - weighted graph @xmath28 in which each edge @xmath37 of @xmath28 is identified with a line segment of length @xmath48 .
we call @xmath28 a _ model _ for @xmath1 . subdividing an edge of @xmath28 in a length - preserving fashion
changes the model but not the underlying metric graph .
let @xmath0 be an algebraically closed field .. ] a _ metrized complex @xmath32 of @xmath0-curves _ consists of the following data : * a connected finite graph @xmath28 with vertex set @xmath26 and edge set @xmath27 .
* a metric graph @xmath1 having @xmath28 as a model ( or equivalently , a length function @xmath49 ) . * for each vertex @xmath3 of @xmath28 , a complete , nonsingular , irreducible curve @xmath2 over @xmath0 .
* for each vertex @xmath3 of @xmath28 , a bijection @xmath50 between the edges of @xmath28 incident to @xmath3 ( with loop edges counted twice ) and a subset @xmath51 of @xmath52 .
the _ geometric realization _
@xmath53 of @xmath32 is defined to be the union of the edges of @xmath28 and the collection of curves @xmath2 , with each endpoint @xmath3 of an edge @xmath37 identified with the corresponding marked point @xmath54 .
see figure [ fig : mcsmall ] . when we think of @xmath53 as a set , we identify it with the disjoint union of @xmath55 and @xmath56
thus , when we write @xmath57 , we mean that @xmath58 is either a non - vertex point of @xmath1 ( a _ graphical point _ of @xmath32 ) or a point of @xmath52 for some @xmath14 ( a _ geometric point _ of @xmath32 ) .
the _ genus _ of a metrized complex of curves @xmath32 , denoted @xmath59 , is by definition @xmath60 , where @xmath61 is the genus of @xmath2 and @xmath62 is the first betti number of @xmath1 . given a metrized complex @xmath32 of @xmath0-curves , there is an associated semistable curve @xmath34 over @xmath0 obtained by gluing the curves @xmath2 along the points @xmath54 ( one intersection for each edge @xmath37 of @xmath28 ) and forgetting the metric structure on @xmath1 .
conversely , given a semistable curve @xmath34 over @xmath0 together with a positive real number for each node ( which we call a `` length function '' ) , one obtains an associated metrized complex of @xmath0-curves by letting @xmath28 be the dual graph of @xmath34 , @xmath1 the metric graph associated to @xmath28 and the given length function , @xmath2 the normalization of the irreducible component @xmath63 of @xmath34 corresponding to @xmath3 , and @xmath45 the preimage in @xmath2 of the set of nodes of @xmath34 belonging to @xmath63 .
let @xmath5 be a complete and algebraically closed non - archimedean field with residue field @xmath0 and let @xmath29 be a smooth , proper , connected algebraic curve .
there is a metrized complex @xmath64 canonically associated to any strongly semistable model @xmath6 of @xmath4 over the valuation ring @xmath7 of @xmath5 : the special fiber @xmath24 of @xmath6 is a semistable curve over @xmath0 , and one defines the length of an edge @xmath37 of the dual graph of @xmath24 to be the _ modulus _ of the open annulus @xmath65 , where @xmath66 is the singular point of @xmath24 corresponding to @xmath37 and @xmath67 is the canonical reduction map .
equivalently , there is a local analytic equation for @xmath66 of the form
@xmath68 with @xmath69 in the maximal ideal of @xmath7 , and the length of @xmath37 is @xmath70 .
the corresponding metric graph @xmath71 is called the _ skeleton _ of @xmath6 , and there is a canonical retraction map @xmath72 , where @xmath13 denotes the berkovich analytification of @xmath4 . by restricting to @xmath73 and then extending by linearity , we obtain a specialization map @xmath74 .
if @xmath7 is a discrete valuation ring and the fibered surface @xmath75 is regular , then one can use intersection - theoretic methods to define and study specialization of divisors from curves to graphs .
for example , in @xcite , using intersection theory , a specialization homomorphism @xmath76 was defined , where @xmath28 is the dual graph of @xmath24 , with the property that @xmath77 and @xmath78 are linearly equivalent as divisors on @xmath12 for all @xmath79 . in the present paper ,
we focus exclusively on @xmath80 and its generalization to metrized complexes of curves , though in principle all of our results can be translated ( when @xmath7 is a discrete valuation ring and @xmath6 is regular ) into results about this alternative specialization map . a _ divisor _ on a metrized complex of curves
@xmath32 is an element @xmath81 of the free abelian group on @xmath53 .
thus a divisor on @xmath32 can be written uniquely as @xmath82 where @xmath83 , all but finitely many of the @xmath84 are zero , and the sum is over all points of @xmath55 as well as @xmath52 for @xmath14 .
the _ degree _ of @xmath81 is defined to be @xmath85 . to a divisor on @xmath32 , we can naturally associate a divisor @xmath86 on @xmath1 , called the _
@xmath1-part _ of @xmath87 , as well as , for each @xmath14 , a divisor @xmath88 on @xmath2 ( called the _
@xmath2-part _ of @xmath87 ) .
the divisor @xmath88 is simply the restriction of @xmath87 to @xmath2 , i.e. @xmath89 , and @xmath86 is defined as @xmath90 where @xmath91 denotes the coefficient of @xmath58 in @xmath81 . in particular , the degree of @xmath81 equals the degree of @xmath86 .
one could equivalently define a divisor on @xmath32 to be an element of the form @xmath92 of @xmath93 such that @xmath94 for all @xmath3 in @xmath26 .
the specialization map @xmath80 can be enhanced in a canonical way to a map from divisors on @xmath4 to divisors on @xmath8 .
the fiber @xmath95 of the retraction map @xmath96 over a vertex @xmath14 can be canonically identified with @xmath97 , the fiber of the reduction map @xmath98 over the smooth locus @xmath99 of @xmath2 .
we define @xmath100 by setting @xmath101 and we extend this by linearity to a map @xmath102 . intuitively speaking , for points @xmath103 of @xmath73 which reduce to smooth points of @xmath24 , the map @xmath104 keeps track of the reduced point @xmath105 , while if @xmath103 reduces to a singular point of @xmath24 then @xmath104 instead keeps track of the retraction of @xmath103 to the skeleton of the open annulus @xmath106 , which is canonically identified with the relative interior of the corresponding edge of @xmath12 . a _ nonzero rational function _
@xmath35 on a metrized complex of curves @xmath32 is the data of a rational function @xmath36 on @xmath1 and nonzero rational functions @xmath107 on @xmath2 for each @xmath31 .
( we do _ not _ impose any compatibility conditions on the rational functions @xmath36 and @xmath107 . )
we call @xmath36 the _ @xmath1-part _ of @xmath35 and @xmath107 the _ @xmath2-part _ of @xmath35 . the _ divisor _ of a nonzero rational function @xmath35 on @xmath32
is defined to be @xmath108 where @xmath109 is defined as follows : * if @xmath110 , then @xmath111 , where @xmath112 is the sum of the slopes of @xmath36 in all tangent directions emanating from @xmath58 . * if @xmath113 , then @xmath114 . * if @xmath115 , then @xmath116 , where @xmath117 is the outgoing slope of @xmath36 at @xmath3 in the direction of @xmath37 . if @xmath86 ( resp .
@xmath88 ) denotes the @xmath1-part ( resp .
the @xmath2-part ) of @xmath118 , then @xmath119 @xmath120 divisors of the form @xmath118 are called _ principal _ , and the principal divisors form a subgroup of @xmath121 , the group of divisors of degree zero on @xmath32 .
two divisors in @xmath38 are called _ linearly equivalent _ if they differ by a principal divisor .
linear equivalence of divisors on @xmath32 can be understood quite intuitively in terms of `` chip - firing moves '' . to state this formally , for each vertex @xmath3 of @xmath28 , let @xmath122 be the sum of the @xmath123 points in @xmath45 and
let @xmath124 be the minimum length of an edge @xmath37 incident to @xmath3 .
also , for an edge @xmath37 having @xmath3 as an endpoint and a positive real number @xmath125 , let @xmath126 be the unique point of @xmath37 at distance @xmath127 from @xmath3 .
then it is not hard to show that linear equivalence of divisors on @xmath32 is the equivalence relation generated by the following three kinds of `` moves '' : 1 .
( firing on @xmath2 ) choose a vertex @xmath3 and replace the divisor @xmath88 with a linearly equivalent divisor @xmath128 on @xmath2 .
2 . ( firing a vertex )
choose a vertex @xmath3 and a positive real number @xmath129 and replace @xmath88 with @xmath130 and @xmath86 with @xmath131 .
3 . ( firing a non - vertex )
choose a non - vertex @xmath132 and a positive real number @xmath127 less than the distance from @xmath133 to the nearest vertex and replace @xmath86 by the divisor @xmath134 , where @xmath135 are the two points at distance @xmath127 from @xmath133 .
the motivation for our definitions of @xmath9 and @xmath118 come from the following fundamental relation , a consequence of the non - archimedean poincar - lelong formula due to thuillier ( see 5 of @xcite ) .
let @xmath136 be a nonzero rational function on @xmath4 and let @xmath35 be the corresponding nonzero rational function on @xmath8 , where @xmath137 is the restriction to @xmath1 of the piecewise linear function @xmath138 on @xmath13 and @xmath139 for @xmath14 is the normalized reduction of @xmath136 to @xmath2 ( cf .
[ section : reductionofrationalfunctions ] ) .
then @xmath140 in particular , we have @xmath141 .
a divisor @xmath142 on @xmath32 is called _ effective _ if @xmath143 for all @xmath58 ( or , equivalently , if the @xmath1-part @xmath144 and the @xmath2-parts @xmath145 of @xmath146 are effective for every @xmath31 )
. the _ rank _
@xmath147 of a divisor @xmath148 is defined to be the largest integer @xmath149 such that @xmath150 is linearly equivalent to an effective divisor for all effective divisors @xmath146 of degree @xmath149 on @xmath32 ( so in particular @xmath151 if and only if @xmath81 is linearly equivalent to an effective divisor , and otherwise @xmath152 ) .
if @xmath153 denotes the usual rank function @xmath154 on @xmath43 , there is an important semicontinuity result relating @xmath155 to @xmath156 : [ thm : introspecialization ] for all @xmath79 , we have @xmath157 since @xmath158 , theorem [ thm : introspecialization ] is a strengthening of the analogous specialization result from @xcite . in conjunction with a simple combinatorial argument , theorem
[ thm : introspecialization ] also refines the specialization lemma for vertex - weighted graphs from @xcite .
we also state and prove a version of theorem [ thm : introspecialization ] in which one has _
equality _ rather than just an inequality .
one can naturally associate to a rank @xmath19 divisor @xmath159 on @xmath4 not only a divisor @xmath160 on @xmath8 , but also a collection @xmath161 of @xmath162-dimensional subspaces of @xmath40 , where @xmath163 is the reduction to @xmath2 of all functions in @xmath164 ( c.f .
section [ section : reductionofrationalfunctions ] for a discussion of reduction of rational functions ) . if @xmath165 , where @xmath166 is any @xmath0-subspace of the function field @xmath167 , then for @xmath168 we define the _ @xmath169-restricted rank _ of @xmath87 , denoted @xmath170 , to be the largest integer @xmath149 such that for any effective divisor @xmath171 of degree @xmath149 on @xmath32 , there is a rational function @xmath35 on @xmath32 whose @xmath2-parts @xmath107 belong to @xmath166 for all @xmath14 , and such that @xmath172 .
( in terms of the `` chip - firing moves '' described above , one has to be able to make @xmath173 effective using only type ( 1 ) moves corresponding to functions belonging to one of the spaces @xmath166 , together with moves of type ( 2 ) or ( 3 ) . ) the following variant of theorem [ thm : introspecialization ] will be proved as theorem [ thm : limitseriesgeneral ] below : with notation as above , the @xmath174-restricted rank of the specialization of @xmath159 is equal to the rank of @xmath159 , i.e. , @xmath175 .
the theory of divisors and linear equivalence on metrized complexes of curves generalizes both the classical theory for algebraic curves and the corresponding theory for metric graphs and tropical curves found in @xcite .
the former corresponds to the case where @xmath28 consists of a single vertex @xmath3 and no edges and @xmath176 is an arbitrary smooth curve .
the latter corresponds to the case where the curves @xmath2 have genus zero for all @xmath14 . since any two points on a curve of genus zero are linearly equivalent , it is easy to see that the divisor theories and rank functions on @xmath32 and @xmath1 are essentially equivalent .
more precisely , ( i ) two divisors @xmath81 and @xmath177 on @xmath32 are linearly equivalent if and only if their @xmath1-parts @xmath86 and @xmath178 are linearly equivalent on @xmath1 , and ( ii ) for every @xmath179 we have @xmath180 . in the presence of higher genus curves among the @xmath2 , the divisor theories on @xmath1 and @xmath32 can be very different .
in addition , different choices of @xmath45 can drastically change both the linear equivalence relation and the rank function . as with the corresponding specialization theory from @xcite , the main utility of theorem [ thm : introspecialization ]
is that the rank function @xmath181 on a metrized complex of curves is surprisingly well - behaved ; for example , it satisfies an analogue of the riemann - roch formula . in order to state this result
, we need to define the canonical class .
a _ canonical divisor _ on @xmath32 , denoted @xmath182 , is defined to be any divisor linearly equivalent to @xmath183 , where @xmath184 is a canonical divisor on @xmath2 and @xmath122 is the sum of the @xmath123 points in @xmath45 .
the @xmath1-part of @xmath182 is @xmath185 , and the @xmath2-part of @xmath182 is @xmath186 .
the following result generalizes both the classical riemann - roch theorem for algebraic curves and the riemann - roch theorem for metric graphs : [ thm : introrr - metrizedcomplexes ] let @xmath32 be a metrized complex of algebraic curves over @xmath0 and @xmath182 a divisor in the canonical class of @xmath32 . for any divisor @xmath148 ,
we have @xmath187 the proof of theorem [ thm : introrr - metrizedcomplexes ] makes use of a suitable notion of _ reduced divisors _ for metrized complexes of curves . taken together , theorems [ thm : introspecialization ] and [ thm : introrr - metrizedcomplexes ] have some interesting consequences .
for example , the specialization theorem implies that the specialization to @xmath8 of any canonical divisor on @xmath4 has degree @xmath188 and rank at least @xmath189 ( where @xmath190 ) , while the riemann - roch theorem easily implies that any such divisor on @xmath8 belongs to the canonical divisor class .
thus we have : [ cor : introcanonspec ] for any canonical divisor @xmath191 on @xmath4 , its specialization @xmath192 belongs to the canonical divisor class on @xmath8 .
in particular , corollary [ cor : introcanonspec ] implies that @xmath193 is a canonical divisor on @xmath1 .
this fact was proved in @xcite , for discretely valued @xmath7 , by a completely different argument based on the adjunction formula for arithmetic surfaces .
it does not seem easy to prove the general ( not necessarily discretely valued ) case using arithmetic intersection theory , since there does not seem to be a satisfactory theory of relative dualizing sheaves and adjunction in the non - noetherian setting .
our theory of linear series on metrized complexes of curves has close connections with the eisenbud - harris theory of limit linear series for strongly semistable curves of compact type , and suggests a way to generalize the eisenbud - harris theory to more general semistable curves .
a proper nodal curve @xmath34 over @xmath0 is of _ compact type _ if its dual graph @xmath28 is a tree . for such curves , eisenbud and harris
define a notion of ( crude ) _
limit @xmath10 _ , which we discuss in detail in section [ sec : limitcompact ] below .
a _ crude limit _
@xmath194 on @xmath34 is the data of a ( not necessarily complete ) degree @xmath195 and rank @xmath19 linear series @xmath196 on @xmath63 for each vertex @xmath14 such that if two components @xmath197 and @xmath63 of @xmath34 meet at a node @xmath133 , then for any @xmath198 @xmath199 where @xmath200 denotes the @xmath201 term in the vanishing sequence of a linear series @xmath194 at @xmath133 .
a crude limit series is _ refined _ if all the inequalities in the above definition are equalities .
for simplicity , all limit linear series in the remainder of this introduction will be crude
. we can canonically associate to a proper strongly curve @xmath34 a metrized complex @xmath202 of @xmath0-curves , called the _ regularization _ of @xmath34 , by assigning a length of @xmath203 to each edge of @xmath28 , and we write @xmath63 for the irreducible component of @xmath34 corresponding to a vertex @xmath14 .
( this is the metrized complex associated to any regular smoothing @xmath6 of @xmath34 over any discrete valuation ring @xmath7 with residue field @xmath0 . )
[ thm : introlls ] let @xmath202 be the metrized complex of curves associated to a proper strongly semistable curve @xmath204 of compact type .
then there is a bijective correspondence between the following : * crude limit @xmath10 s on @xmath34 . *
equivalence classes of pairs @xmath205 , where @xmath206 , @xmath163 is an @xmath162-dimensional subspace of @xmath207 for each @xmath14 , and @xmath87 is a divisor of degree @xmath195 supported on the vertices of @xmath202 with @xmath208 .
here we say that @xmath209 if there is a rational function @xmath35 on @xmath202 such that @xmath210 and @xmath211 for all @xmath14 , where @xmath107 denotes the @xmath2-part of @xmath35 .
theorem [ thm : introlls ] , combined with our riemann - roch theorem for metrized complexes of curves , provides an arguably more conceptual proof of the fact ( originally established in @xcite ) that limit linear series satisfy analogues of the classical theorems of riemann and clifford .
the point is that @xmath212 for all @xmath213 and therefore upper bounds on @xmath214 which follow from riemann - roch imply corresponding upper bounds on the restricted rank @xmath215 . motivated by theorem [ thm : introlls ] , we propose the following definition .
let @xmath34 be a proper strongly semistable ( but not necessarily compact type ) curve over @xmath0 with regularization @xmath202 .
limit @xmath10 _ on @xmath34 is an equivalence class of pairs @xmath216 as above , where @xmath163 is an @xmath162-dimensional subspace of @xmath207 for each @xmath14 , and @xmath87 is a degree @xmath195 divisor on @xmath202 with @xmath208 . as a partial justification for this definition
, we prove that if @xmath7 is a discrete valuation ring with residue field @xmath0 and @xmath75 is a regular arithmetic surface whose generic fiber @xmath4 is smooth and whose special fiber @xmath34 is strongly semistable , then for any divisor @xmath159 on @xmath4 with @xmath217 and @xmath218 , our specialization machine gives rise in a natural way to a limit @xmath10 on @xmath34 ( see theorem [ thm : regularlimitseries ] below for a precise statement , and theorem [ thm : limitseriesgeneral ] for a more general statement ) .
let @xmath0 be an algebraically closed field of arbitrary characteristic and let @xmath32 be a metrized complex of @xmath0-curves , as defined in section [ section : overview ] .
if @xmath219 is a subgroup of @xmath220 containing the lengths of all edges of @xmath28 , we write @xmath221 for the subgroup of @xmath222 consisting of all divisors @xmath81 such that the support of the @xmath1-part of @xmath87 consists entirely of @xmath219-_rational points _ of @xmath1 ( i.e. , points of @xmath1 whose distances to the vertices of @xmath28 are in @xmath223 ) .
recall from section [ section : overview ] that the _ rank _ @xmath147 of a divisor @xmath81 on @xmath32 is the largest integer @xmath149 such that @xmath150 is linearly equivalent to an effective divisor for all effective divisors @xmath146 of degree @xmath149 on @xmath32 .
if @xmath219 is any subgroup of @xmath220 containing the lengths of all edges of @xmath28 , one can define an analogous rank function @xmath224 for divisors in @xmath225 by restricting the effective divisors @xmath146 in the above definition to lie in @xmath225 . by corollary [ cor : f - widthg ] , @xmath226 for any divisor @xmath227 , so the restriction in the definition of @xmath224 does not affect the rank of divisors .
let @xmath34 be a ( reduced ) connected projective nodal curve over @xmath30 with irreducible components @xmath228 .
we can associate to @xmath34 a metrized complex @xmath202 , called the _ regularization _ of @xmath34 , as follows . the underlying metric graph @xmath229 has model the dual graph @xmath230 of @xmath34 . recall that the vertex set @xmath231 of @xmath232 consists of vertices @xmath233 ( in bijection with the irreducible components of @xmath34 ) and the edges of @xmath232 are in bijection with singular points of @xmath34 .
( note that @xmath232 might have loop edges . )
the length of all the edges in @xmath229 are equal to one .
the @xmath30-curve @xmath234 in @xmath202 is the normalization of @xmath235 for @xmath236 .
the collection @xmath237 is the set of all @xmath30-points of @xmath234 which lie over a singular point of @xmath34 in @xmath238 . by the definition of the dual graph ,
these points are in bijection with the edges adjacent to @xmath239 in @xmath232 . if @xmath34 is _ strongly semistable _ , so that @xmath232 does not have any loop edge and @xmath240 , the rank function @xmath241 on @xmath242 can be reformulated as follows .
let @xmath243 be the picard group of @xmath34 , and consider the restriction map @xmath244 .
for any line bundle @xmath245 on @xmath34 , and @xmath246 , denote by @xmath247 the restriction of @xmath245 to @xmath63 .
two line bundles @xmath245 and @xmath248 in @xmath243 are said to be _ combinatorially equivalent _ if there exists a function @xmath249 such that @xmath250 in @xmath251 for any vertex @xmath3 of @xmath252 ( in a regular smoothing of @xmath34 , c.f .
section [ sec : limitseries ] for the definition , the role of @xmath136 is to specify a particular twist of @xmath245 by a divisor supported on the irreducible components of @xmath34 . ) in particular , two line bundles @xmath245 and @xmath248 in @xmath243 with @xmath253 are combinatorially equivalent . in the above definition
, @xmath254 is defined in analogy with ( [ eq : divf2 ] ) as follows : @xmath255 where @xmath256 is the point of @xmath2 labelled with the edge @xmath257 .
note that the function @xmath249 in the definition of combinatorially equivalent line bundles above is unique up to an additive constant .
for such an @xmath136 , we denote by @xmath258 any line bundle @xmath248 in @xmath243 such that @xmath250 for all @xmath246 . for a line bundle @xmath245 on @xmath34 , define the _ combinatorial rank _ of @xmath245 , denoted @xmath259 , to be the maximum integer @xmath19 such that for any effective divisor @xmath260 on @xmath232 of degree @xmath19 , there is a line bundle @xmath248 combinatorially equivalent to @xmath245 with @xmath261 for all @xmath14 . in particular , if no non - negative integer @xmath19 with the above property exists , the combinatorial rank of @xmath245 will be equal to @xmath262 .
the combinatorial rank of @xmath245 clearly depends only on the combinatorial equivalence class of @xmath245 .
let @xmath81 be a divisor in @xmath242 with @xmath63-part @xmath88 .
we denote by @xmath263 any line bundle on @xmath34 whose restriction to @xmath63 is @xmath264 for each vertex @xmath3 of @xmath232 .
[ prop : combinrank ] let @xmath34 be a strongly semistable curve over @xmath30 and let @xmath202 be the corresponding metrized complex .
then : 1 .
two divisors @xmath81 and @xmath177 in @xmath242 are linearly equivalent iff @xmath263 and @xmath265 are combinatorially equivalent .
2 . for any divisor @xmath266 ,
the combinatorial rank of @xmath263 is equal to @xmath267 the first part follows by definition . to see the second part ,
let @xmath19 be the combinatorial rank of @xmath263 .
it follows from corollary [ cor : f - widthz ] that @xmath268 .
thus , it will be enough to show that @xmath269 .
we first prove that @xmath270 .
let @xmath146 be an effective divisor in @xmath242 with @xmath1-part @xmath144 and @xmath63-part @xmath145 .
by the definition of the combinatorial rank , there exists a function @xmath271 such that @xmath272 since @xmath273 , this implies the existence of a global section @xmath107 of @xmath274 such that @xmath107 has order of vanishing at least @xmath275 at any point @xmath58 of @xmath63 .
let @xmath36 be the rational function defined by @xmath136 on @xmath1 .
( recall that @xmath276 and @xmath136 is integer affine with slope @xmath277 on each edge @xmath257 . ) for the rational function @xmath35 on @xmath202 defined by @xmath36 and @xmath278 , one has @xmath279 .
this shows that @xmath280 .
we now prove that @xmath281 .
let @xmath282 be a line bundle on @xmath34 defined by @xmath81 .
let @xmath27 be an effective divisor of degree @xmath283 on @xmath232 , and let @xmath146 be any effective divisor in @xmath242 with @xmath1-part equal to @xmath27 .
then there exists a rational function @xmath35 with @xmath284 such that @xmath285 .
in particular , the @xmath1-part @xmath36 of @xmath35 is linear on each edge of @xmath232 and @xmath286 , which shows that up to an additive constant , the set @xmath287 of all functions @xmath36 with these properties is finite .
let @xmath288 denote the finite set @xmath289 .
let @xmath290 , and for @xmath291 , let @xmath292 be the subset of @xmath293 defined by all collections of effective divisors @xmath145 of degree @xmath294 such that there exists a rational function @xmath35 on @xmath202 with @xmath1-component @xmath36 satisfying @xmath295 .
( here @xmath296 denotes the @xmath297 symmetric product of a curve @xmath4 . )
now let @xmath298 be the union over all @xmath291 of @xmath292 .
then one sees easily that : * each @xmath292 is zariski - closed in @xmath299 .
( use the fact that @xmath292 is the projection of a zariski closed subset of a product of proper curves . ) * @xmath300 .
therefore there exists @xmath291 such that @xmath301 .
in other words , we can suppose that the @xmath1-part @xmath36 of all the rational functions @xmath35 giving @xmath302 , when the effective divisors @xmath145 vary , is constant ( and we have @xmath303 ) .
if we fix a corresponding twist @xmath304 , then @xmath305 since @xmath27 was an arbitrary effective divisor of degree @xmath283 on @xmath232 , we obtain the desired inequality .
our aim in this section is to formulate and prove a riemann - roch theorem for metrized complexes of @xmath0-curves .
let @xmath32 be a metrized complex of algebraic curves over @xmath30 . for each vertex @xmath14 ,
let @xmath61 be the genus of @xmath2 and let @xmath306 be the genus of @xmath32 . for each curve @xmath2 ,
let @xmath184 denote a divisor of degree @xmath307 in the canonical class of @xmath2 . denote by @xmath122 the divisor in @xmath2 consisting of the sum of the @xmath123 points in @xmath45 .
canonical class of @xmath32 _ is defined to be the linear equivalence class of the divisor @xmath308 [ rem : canonicalsemi - stable ] let @xmath34 be a strongly semistable curve .
the definition of the canonical class @xmath182 in @xmath202 is compatible in a natural sense with the definition of the dualizing sheaf ( or sheaf of logarithmic @xmath203-forms ) @xmath309 .
indeed , if @xmath310 denotes the normalization of @xmath34 , and @xmath311,@xmath312 denote the points of @xmath4 above the singular point @xmath313 of @xmath34 for @xmath314 , then sections of the sheaf @xmath309 on an open set @xmath315 consist of all the meromorphic 1-forms @xmath316 on @xmath4 which are regular everywhere on @xmath315 except for possible simple poles at @xmath317 , with @xmath318 for all @xmath37 . from this , it is easy to see that the restriction of @xmath309 to each component @xmath63 of @xmath34 is the invertible sheaf @xmath319 corresponding to @xmath320 for any canonical divisor @xmath184 on @xmath63 . for a divisor @xmath81 on @xmath32 , let @xmath321 denote the rank of @xmath81 as defined in section [ section : overview ] .
[ thm : rr - metrizedcomplexes ] let @xmath32 be a metrized complex of @xmath0-curves and @xmath182 a divisor in the canonical class of @xmath32 . for every divisor @xmath148 , we have @xmath187 let @xmath34 be a strongly semistable curve of genus @xmath20 over @xmath0
. for any @xmath322 , let @xmath259 be the combinatorial rank of @xmath245 as defined in section [ sec : basics ] . as a corollary of theorem [ thm : rr - metrizedcomplexes ] , remark [ rem : canonicalsemi - stable ] , and proposition [ prop : combinrank ] , we have : for any @xmath323 , @xmath324 the riemann - roch theorem has a number of well - known formal consequences which can be transported to our situation .
for example , we obtain the following analogue of clifford s bound for the rank of a special divisor .
( a divisor @xmath81 on a metrized complex @xmath32 is called _ special _ if @xmath325 is linearly equivalent to an effective divisor . )
[ thm : mcclifford ] for any special divisor @xmath81 on a metrized complex @xmath32 , @xmath326 .
our riemann - roch theorem seems to be rather different from the classical riemann - roch theorem on semistable curves .
we refer to @xcite for a discussion of riemann s and clifford s theorems for linear series on semistable curves , and for examples showing the failure of clifford s theorem for @xmath327 with @xmath322 ( even for line bundles in the compactified picard scheme of @xmath34 ) .
the rest of this section is devoted to the proof of theorem [ thm : rr - metrizedcomplexes ] .
our proof follows and extends the original arguments of @xcite in the proof of riemann - roch theorem for graphs .
in particular , we are going to first extend the notion of reduced divisors to the context of metrized complexes of algebraic curves , and then study the minimal non - special divisors .
let @xmath328 be a fixed base point of @xmath1 .
we introduce the notion of @xmath328-reduced divisors and show that each equivalence class of divisors on @xmath32 contains a quasi - unique @xmath328-reduced divisor , in a precise sense to be defined immediately preceding theorem [ thm : reduced.divisors ] below . for a closed connected subset @xmath298 of @xmath1 and a point @xmath329 ( the topological boundary of @xmath298 ) , the number of `` outgoing '' tangent directions at @xmath3 ( i.e.
, tangent directions emanating outward from @xmath298 ) is denoted by @xmath330 ; this is also the maximum size of a collection of internally disjoint segments in @xmath331 with one end at @xmath3 .
if in addition we have @xmath14 , we denote by @xmath332 the divisor in @xmath2 associated to the outgoing edges at @xmath3 ; this is by definition the sum of all the points @xmath54 of @xmath2 indexed by the edges @xmath37 leaving @xmath298 at @xmath3 . in what follows
we refer to a closed connected subset @xmath298 of @xmath1 as a _ cut _ in @xmath1 .
let @xmath81 be a divisor on @xmath32 with @xmath1-part @xmath86 and @xmath2-part @xmath88 , and let @xmath298 be a cut in @xmath1 .
a boundary point @xmath333 is called _ saturated _ with respect to @xmath81 and @xmath298 if * @xmath334 and @xmath335 ; or * @xmath336 for some @xmath31 and @xmath337 is equivalent to an effective divisor on @xmath2 . otherwise , @xmath338 is called _ non - saturated_. a cut @xmath298 is said to be saturated if all its boundary points are saturated .
( when talking about saturated and non - saturated points , we will sometimes omit the divisor @xmath87 or the set @xmath298 if they are clear from the context . )
the divisor @xmath81 is said to be _
@xmath328-reduced _ if the following three properties are satisfied : * for all points @xmath339 of @xmath1 , @xmath340 , i.e. , all the coefficients of @xmath86 are non - negative except possibly at the base point @xmath328 .
* for all points @xmath341 ( @xmath342 if @xmath328 does not belong to @xmath26 ) , there exists an effective divisor @xmath145 linearly equivalent to @xmath88 on @xmath2 . * for every cut @xmath298 of @xmath1 which does not contain @xmath328 , there exists a non - saturated point @xmath333 .
[ rmk : burningalg ] there is an efficient `` burning algorithm '' for checking whether or not a given divisor @xmath87 on @xmath32 which satisfies ( i ) is @xmath328-reduced .
the algorithm can be described informally as follows ( compare with @xcite or @xcite ) .
imagine that @xmath1 is made of a flammable material , and light a fire at @xmath328 .
the fire begins spreading across @xmath1 in a continuous manner and can only be blocked at a point @xmath343 if one of the following holds : * @xmath334 and the number of burnt directions coming into @xmath58 exceeds @xmath344 .
( one should imagine @xmath344 firefighters standing at @xmath58 , each of whom can block fire in one incoming direction . ) * @xmath336 for some @xmath31 and @xmath88 minus the sum of the marked points corresponding to burnt directions is not equivalent to any effective divisor on @xmath2 .
it is straightforward to check , following the same ideas as the proof of ( * ? ? ?
* algorithm 2.5 ) , that @xmath87 is @xmath328-reduced if and only if the fire eventually burns through all of @xmath1 .
we omit the details since we will not need this result in the sequel .
we now show that every divisor @xmath81 on @xmath32 is linearly equivalent to a quasi - unique @xmath328-reduced divisor @xmath345 .
the quasi - uniqueness is understood in the following sense : the @xmath1-part @xmath346 of @xmath345 is unique , and for all @xmath14 , the divisor class @xmath347 $ ] on @xmath2 defined by the @xmath3-part of @xmath345 is unique .
[ thm : reduced.divisors ] let @xmath32 be a metrized complex of @xmath30-curves and @xmath328 a base point of @xmath1 . for every divisor @xmath148
, there exists a quasi - unique @xmath328-reduced divisor @xmath345 such that @xmath348 .
let @xmath349 be the set of all divisors @xmath177 linearly equivalent to @xmath81 such that : * all the coefficients of the @xmath1-part @xmath178 of @xmath177 are non - negative at every point of @xmath1 except possibly at @xmath328 .
* for each point @xmath341 , the @xmath3-part @xmath128 of @xmath177 has non - negative rank on @xmath2 . *
the coefficient of @xmath328 in @xmath178 is maximal with respect to the two properties above .
one shows as in ( * ? ? ?
* proof of theorem 2 ) or ( * ? ? ? * proof of proposition 3.1 ) that there is a divisor @xmath177 linearly equivalent to @xmath81 on @xmath32 with the property that the coefficient of every point @xmath350 is non - negative and the coefficient of @xmath178 at each point @xmath14 ( which coincides with the degree of @xmath128 ) is at least @xmath61 .
( the intuitive idea is to repeatedly fire @xmath328 and then use further chip - firing moves to spread chips wherever needed on @xmath1 , assuring that each @xmath14 gets at least @xmath61 chips and that no @xmath132 except for @xmath328 remains in debt . ) by the ( classical ) riemann - roch theorem , each @xmath128 has non - negative rank on @xmath2 , which shows that @xmath349 is non - empty .
define @xmath351 as the set of all @xmath352 for which there exists @xmath353 whose @xmath1-part is @xmath178 .
each divisor @xmath354 has degree @xmath355 , and all divisors in @xmath351 have the same coefficient at @xmath328 .
thus the number @xmath356 is independent of the choice of a divisor @xmath178 in @xmath351 . as a consequence ,
@xmath351 inherits a natural topology from the topology of @xmath1 , since it can be embedded as a subset of @xmath357 , and is compact .
let @xmath358 be the subset of @xmath359 defined by @xmath360 equipped with the total order defined by the lexicographical rule @xmath361 iff @xmath362 and @xmath363 .
one considers a map @xmath364 defined as follows . for each divisor @xmath365 , consider the multiset @xmath366 of points in @xmath367 where each point @xmath368 appears in this multiset exactly @xmath369 times .
define @xmath370 to be the point of @xmath358 defined by the multiset of distances @xmath371 for @xmath372 , ordered in an increasing way .
it is straightforward to verify that the map @xmath287 is continuous .
since @xmath351 is compact and @xmath287 is continuous , there exists a divisor @xmath373 in @xmath351 such that @xmath287 takes its minimum value at @xmath373 , i.e. , @xmath374 .
let @xmath345 be an element of @xmath349 whose @xmath1-part is @xmath373 .
* claim : * the divisor @xmath375 is @xmath328-reduced . properties ( i ) and ( ii ) above are clearly satisfied .
the only point one needs to check is that every cut @xmath298 which does not contain @xmath328 has a non - saturated point on its boundary . for the sake of contradiction , suppose this is not the case , and let @xmath298 be a closed connected set violating this condition .
this means that the following hold : * for all @xmath376 , @xmath377 . * for each @xmath378 , @xmath379 has non - negative rank on @xmath2 , i.e. , there exists a rational function @xmath107 on @xmath2 such that @xmath380 . by the definition of @xmath381 , there exists an @xmath382 such that for each vertex @xmath333 , there are closed segments @xmath383 emanating from @xmath58 with the following properties : * for @xmath333 and @xmath384 , the half - open segments @xmath385 are disjoint from @xmath298 and from each other and do not contain @xmath328 . *
each segment @xmath386 has length @xmath387 , for @xmath333 and @xmath384 .
these data give rise to a rational function @xmath388 which is identically zero on @xmath298 , is linear of slope @xmath262 on each interval @xmath386 for @xmath338 and @xmath389 , and takes the constant value @xmath390 at all points of @xmath391 .
consider the divisor @xmath392 on @xmath32 , where @xmath35 is the rational function on @xmath32 consisting of the rational function @xmath36 on @xmath1 and the rational functions @xmath107 on @xmath2 .
letting @xmath393 be the @xmath1-part of @xmath394 , one verifies that @xmath395 and @xmath393 lie in @xmath349 and @xmath351 , respectively , and that @xmath396 , contradicting the choice of @xmath373 .
this proves the claim and hence the existence part of the theorem .
it remains to prove the quasi - uniqueness .
assume for the sake of contradiction that there are linearly equivalent @xmath328-reduced divisors @xmath81 and @xmath177 on @xmath32 such that either @xmath397 or @xmath88 and @xmath128 are not linearly equivalent on @xmath2 for some @xmath14 .
then there exists a non - constant rational function @xmath35 on @xmath32 such that @xmath398 .
if @xmath36 is constant then we obtain an immediate contradiction , so we may assume that @xmath36 is non - constant .
without loss of generality , we may assume that @xmath36 does not take its maximum at @xmath328 ( otherwise , we can interchange the role of @xmath81 and @xmath177 ) .
let @xmath298 be a connected component of the set of all points where @xmath36 takes its maximum .
note that @xmath399 . for all points @xmath338
, the slope of @xmath36 at any outgoing segment @xmath386 emanating from @xmath3 is at most @xmath262 .
since @xmath81 is @xmath328-reduced , there exists a point @xmath338 such that either @xmath334 and @xmath400 , or @xmath401 for some @xmath31 and the divisor @xmath402 has negative rank . in the first case , @xmath403 , contradicting the assumption that the coefficient of @xmath178 is non - negative at @xmath339 . in the second case , @xmath404 , which implies that @xmath128 has negative rank , a contradiction .
one can presumably give an `` iterated dhar algorithm '' for algorithmically finding a @xmath328-reduced divisor equivalent to a given divisor @xmath87 by combining the ideas in the proof of theorem [ thm : reduced.divisors ] with the ideas behind ( * ? ? ?
* algorithm 2.12 ) .
let @xmath32 be a metrized complex of algebraic curves .
for each @xmath14 , let @xmath405 be the set of all _ minimal non - special _
divisors on @xmath2 : @xmath406 recall from ( * ? ? ?
* iv , example 1.3.4 ) that a divisor @xmath159 on a smooth projective curve @xmath407 of genus @xmath20 over @xmath30 is called _ special _ if @xmath408 . by riemann - roch ,
every divisor of degree at most @xmath409 on @xmath407 is special , and if @xmath410 then @xmath159 is special if and only if @xmath411 .
moreover , every non - special divisor dominates a minimal non - special divisor .
this explains the term `` minimal non - special divisor '' .
similarly , define the set of minimal non - special divisors on @xmath32 as @xmath412 in this section we provide an explicit description of the minimal non - special divisors on metrized complexes of algebraic curves , generalizing the corresponding description of minimal non - special divisors on graphs and metric graphs from @xcite .
an _ acyclic orientation _ on @xmath1 is an acyclic orientation on some model of @xmath1 , i.e. , a decomposition of @xmath1 into closed directed edges with disjoint interiors such that no directed cycles are formed . given an acyclic orientation @xmath413 on @xmath1 , we denote by @xmath414 the number of tangent directions emanating from @xmath58 which are compatible with @xmath413 .
note that for all but finitely many points of @xmath1 , @xmath415 . for @xmath14 , we denote by @xmath416 the set of edges incident to @xmath3 which ( locally near @xmath3 ) are oriented outward from @xmath3 .
a point @xmath58 with @xmath417 is called a _
sink_. it is well known and easy to prove that every acyclic orientation contains at least one sink ( start at any point and follow the orientation until you get stuck , which must eventually happen by acyclicity ) . given a collection of minimal non - special divisors @xmath418 for each @xmath31 , together with an acyclic orientation @xmath413 of @xmath1 , define a corresponding divisor @xmath419 by the formula @xmath420 where @xmath421 is the sum of all points @xmath422 for which the @xmath413-orientation on @xmath37 points away from @xmath3 .
the @xmath1-part of @xmath423 is @xmath424 and the @xmath2-part of @xmath423 is @xmath425 . if @xmath426 is a ( loopless ) model of @xmath1 such that the orientation @xmath413 of @xmath1 is induced by an acyclic orientation @xmath427 on the edges of @xmath428 , then the degree of @xmath429 is given by @xmath430 + \sum_{v\in v } g_v = g(\gamma)-1+\sum_v g_v = g({\mathfrak{c } } ) -1.\end{aligned}\ ] ] a divisor @xmath431 of the form @xmath432 is called a _
moderator _ on @xmath32 .
( this terminology comes from @xcite . )
given a moderator @xmath433 , the _ dual moderator _
@xmath434 is defined to be @xmath435 , where @xmath436 is obtained from @xmath413 by reversing the orientation of every oriented segment .
it is easy to see that @xmath437 belongs to the canonical class on @xmath32 .
the following two lemmas are essentially what we need for the proof of theorem [ thm : rr - metrizedcomplexes ] .
[ lem : non - special1 ] for any acyclic orientation @xmath413 of @xmath1 and any collection @xmath438 of minimal non - special divisors on @xmath2 , the moderator @xmath439 is a minimal non - special divisor on @xmath32 .
suppose for the sake of contradiction that there is a rational function @xmath35 on @xmath32 such that @xmath440 .
let @xmath426 be a ( loopless ) model of @xmath1 such that the orientation @xmath413 of @xmath1 is induced by an acyclic orientation @xmath427 on the edges of @xmath428 .
let @xmath298 be the closed subset of @xmath1 consisting of all the points of @xmath1 where the @xmath1-part @xmath36 of @xmath35 takes its maximum value .
it is easy to see that @xmath441 , since at a hypothetical point @xmath442 we would necessarily have @xmath443 , contradicting the assumption on @xmath35 . note that since @xmath444 , for any edge @xmath37 of @xmath428 either @xmath298 contains @xmath37 entirely or @xmath298 does not intersect the interior of @xmath37 .
consider now the restriction of the orientation @xmath413 to the induced subgraph @xmath445 $ ] .
the resulting directed graph @xmath446 $ ] is acyclic and therefore contains a sink vertex @xmath447 .
* claim : * @xmath448 .
it suffices to prove that no sink vertex of the directed graph @xmath427 belongs to @xmath298 , so assume for the sake of contradiction that @xmath449 is a sink vertex of @xmath427 .
there are two cases to consider : * if @xmath450 , then @xmath451 ( since @xmath36 achieves its maximum value at @xmath452 and we supposed that @xmath440 ) , contradicting the assumption that @xmath453 is a minimal non - special divisor on @xmath454 . *
if @xmath455 , then by definition the coefficient of @xmath452 in @xmath456 is at most @xmath262 , from which it follows that the coefficient of @xmath452 in @xmath457 is negative , a contradiction .
this proves the claim . by our choice of @xmath458 and the definition of @xmath298
, @xmath36 has strictly negative slope along all outgoing edges at @xmath447 , and it has slope zero along all other edges incident to @xmath447 .
this shows in particular that @xmath459 .
there are now two different cases to consider , both of which will lead to a contradiction ( and hence complete our proof of the lemma ) : * if @xmath460 , then @xmath461 , contradicting the choice of the rational function @xmath35 on @xmath32 . *
if @xmath462 , then for each edge @xmath37 incident to @xmath447 we have @xmath463 , with strict inequality if @xmath37 is an outgoing edge at @xmath447 in @xmath427 ( because @xmath447 is a sink of the oriented graph @xmath464 $ ] ) . by the definition of @xmath465 , we infer that @xmath466 , which shows that @xmath467 is not effective , since @xmath468 is minimal non - special .
this contradicts our assumptions on @xmath35 . for @xmath469
, we denote by @xmath470 the set of all acyclic orientations of @xmath1 with a unique sink at @xmath328 , i.e. , such that @xmath328 has out - degree zero and all other points of @xmath1 have out - degree at least one .
[ lem : reduced ] let @xmath81 be a @xmath328-reduced divisor on @xmath32
. then @xmath471 if and only if the following two conditions hold : * the coefficient @xmath472 of @xmath86 at @xmath328 is non - negative . * if @xmath473 , the divisor @xmath474 on @xmath475 has non - negative rank .
more precisely , if the above conditions do not both hold , then there exists a moderator @xmath476 with @xmath477 such that @xmath478 .
the intuitive idea is that one constructs the moderator @xmath476 using the burning algorithm described in remark [ rmk : burningalg ] .
beginning with a @xmath328-reduced divisor @xmath81 , burn through @xmath1 following the algorithm and keep track of the direction in which the fire spreads ; reversing all the arrows defines an acyclic orientation @xmath477 . if @xmath479 is the corresponding moderator , one checks that @xmath478 outside @xmath328 , and if @xmath81 is not effective then @xmath478 everywhere .
a more rigorous version of this argument is as follows .
first suppose that @xmath480 , and in the case @xmath473 that there exists a rational function @xmath481 on @xmath475 such that @xmath482 .
we infer that @xmath483 , and since each @xmath88 for @xmath484 is of non - negative rank , there exists a rational function @xmath107 on @xmath2 such that @xmath485 .
let @xmath35 be a rational function on @xmath32 consisting of a constant rational function on @xmath1 and @xmath107 on @xmath2 for @xmath14 .
we obviously have @xmath486 , which implies that @xmath487 . for the other direction ,
assume that @xmath488 if @xmath489 , and that @xmath474 has negative rank if @xmath473 . to show that @xmath490 , by lemma [ lem : non - special1 ] it will be enough to show the existence of an acyclic orientation @xmath477 and a set of minimal non - special divisors @xmath491 such that @xmath492 .
let @xmath493 be the union of @xmath26 and all the points in the support of @xmath86 .
note that in both cases above , @xmath494 .
let @xmath495 be the corresponding model of @xmath1 .
we are going to recursively define an orientation @xmath496 of @xmath428 and the collection @xmath497 by handling at each step the orientation of all the edges incident to a vertex @xmath498 and the minimal non - special divisor @xmath499 in the case @xmath14 .
we start by considering the vertex @xmath494 .
the orientations of all the edges incident to @xmath328 are defined so that @xmath500 ; in other words , all these oriented edges are incoming at @xmath328 .
note that in the case @xmath489 , we have @xmath501 . in the case @xmath473 , since @xmath502 it follows from the riemann - roch theorem for @xmath475 that there exists a minimal non - special divisor @xmath503 such that @xmath504 .
suppose that the orientation of all edges adjacent to vertices @xmath505 has been defined , and that for all @xmath506 with @xmath507 , a minimal non - special divisor @xmath508 on @xmath509 has been given .
let @xmath510 be a connected component of the induced subgraph @xmath511 $ ] .
since @xmath81 is @xmath328-reduced and @xmath510 is a cut not containing @xmath328 , there exists a point @xmath512 on the boundary of @xmath510 which is non - saturated .
( note that @xmath512 also lies in @xmath493 . )
this means that either : * @xmath513 and @xmath514 ; or * @xmath515 and @xmath516 has negative rank .
all outgoing edges from @xmath510 adjacent to @xmath512 have already been oriented ( and are outgoing from @xmath512 by the definition of the orientation @xmath413 ) .
orient all other edges incident to @xmath512 in such a way that they are all incoming at @xmath512 .
note that @xmath517 and so in case ( 1 ) , @xmath518 in case ( 2 ) , since @xmath519 , it follows from the riemann - roch theorem for @xmath2 that there exists a minimal non - special divisor @xmath499 such that @xmath520 this shows that @xmath521 . by the definition of the orientation @xmath413 we must have @xmath522 .
therefore , by the definition of the orientation @xmath413 at @xmath512 and the choice of @xmath523 , we have @xmath524 .
let @xmath413 be the orientation of @xmath1 just constructed , and let @xmath525 be the collection of minimal non - special divisors on @xmath2 defined above . by the definition of @xmath413 ,
the only vertex of @xmath493 with out - degree zero is @xmath328 and thus @xmath413 belongs to @xmath470 .
we clearly have @xmath526 as well , which completes the proof . as a corollary of the above lemma
, we obtain : [ cor : non - special2 ] any minimal non - special divisor @xmath81 on @xmath32 is linearly equivalent to a moderator .
fix @xmath527 .
let @xmath81 be a minimal non - special divisor on @xmath32 and let @xmath345 be the @xmath328-reduced divisor linearly equivalent to @xmath81 .
by lemma [ lem : reduced ] , there exists a moderator @xmath479 such that @xmath528 .
the effective divisor @xmath529 has degree zero , and so must be equal to zero .
the proof of theorem [ thm : rr - metrizedcomplexes ] can now be completed using ( a slight simplification of ) the idea behind the proof of riemann - roch theorem for graphs given in @xcite .
the key is the following formula , which provides a useful description of the rank of @xmath81 in terms of minimal non - special divisors .
we denote by @xmath530 ( resp .
@xmath531 ) the sum of non - negative ( resp .
non - positive ) coefficients in @xmath81 .
( note that @xmath532 ) .
[ prop : magicformula ] for any divisor @xmath81 on @xmath32 , we have @xmath533 fix @xmath534 , and let @xmath146 be the non - negative part of @xmath535 , which is of degree @xmath536
. then @xmath537 , since @xmath538 and @xmath539 imply that @xmath540 .
this shows that @xmath541 . to prove the opposite inequality ,
let @xmath146 be an effective divisor of degree @xmath542 such that @xmath543 .
by lemma [ lem : reduced ] , there exists @xmath544 such that @xmath545 , or equivalently , @xmath546 .
in particular , @xmath547 , which proves the proposition .
to finish the proof of theorem [ thm : rr - metrizedcomplexes ] , note that @xmath548 and observe that @xmath549 . if @xmath550 is linearly equivalent to the moderator @xmath551 , c.f .
corollary [ cor : non - special2 ] , then @xmath552 is linearly equivalent to the dual moderator @xmath553 , and so belongs to @xmath554 .
thus , by proposition [ prop : magicformula ] , we have @xmath555
let @xmath5 be a complete and algebraically closed non - archimedean field with non - trivial absolute value @xmath556 , @xmath7 the valuation ring of @xmath557 , and @xmath30 its ( algebraically closed ) residue field .
let @xmath558 be the value group of @xmath557 .
let @xmath4 be a smooth , proper , connected curve over @xmath557 and let @xmath41 be the berkovich analytic space associated to @xmath4 .
( we assume the reader is familiar with the theory of berkovich analytic curves , see e.g. ( * ? ? ?
* section 5 ) which contains everything we need . )
our first goal will be to define a metrized complex @xmath8 associated to a strongly semistable @xmath7-model @xmath6 for @xmath4 .
we then define a specialization homomorphism from @xmath43 to @xmath44 and prove a specialization inequality which refines lemma 2.8 and corollary 2.11 from @xcite . finally , we give some applications to specialization of canonical divisors , brill - noether theory , and weierstrass points as in _ loc .
_ recall that a connected reduced algebraic curve over @xmath0 is called _ semistable _ if all of its singularities are ordinary double points , and is called _ strongly semistable _ if in addition its irreducible components are all smooth .
a _ ( strongly ) semistable model _ for @xmath4 is a flat and integral proper relative curve @xmath6 over @xmath7 whose generic fiber is isomorphic to @xmath4 and whose special fiber @xmath24 is a ( strongly ) semistable curve . given a semistable model @xmath6 for @xmath4 , there is a canonical associated reduction map @xmath67 which is defined using the natural bijection between @xmath73 and @xmath559 .
this extends naturally to a map @xmath560 .
we define a metrized complex associated to a strongly semistable model @xmath6 for @xmath4 as follows .
let @xmath28 be the dual graph of @xmath24 , so that vertices of @xmath28 correspond to irreducible components of @xmath24 and edges of @xmath28 correspond to intersections between irreducible components . if @xmath313 is the ordinary double point of @xmath24 corresponding to an edge @xmath37 of @xmath28 , the _ formal fiber _
@xmath561 is isomorphic to an open annulus @xmath562 .
we define the _ length of the edge @xmath37 _ to be the length of the skeleton of @xmath562 , i.e. , the _ modulus _ @xmath563 of @xmath564 .
( the modulus of an open annulus is well - defined independent of the choice of such an analytic isomorphism . ) in this way , we have defined a metric graph @xmath71 associated to @xmath6 together with a model @xmath28 .
the irreducible components @xmath2 of @xmath24 correspond bijectively to the vertices @xmath3 of @xmath28 , and we let @xmath565 be the finite set of double points of @xmath24 contained in @xmath2 , so that there is a natural bijection between @xmath566 and the edges of @xmath28 incident to @xmath3 . in this way we have defined a metrized complex @xmath8 canonically associated to @xmath6 .
one can show that essentially every metrized complex of curves comes from this construction .
the following result is proved in ( * ? ? ?
* theorem 3.24 ) .
[ thm : graph.to.curve ] let @xmath32 be a metrized complex of @xmath0-curves whose edge lengths are contained in the value group of @xmath5 .
there exists a smooth , proper , connected curve @xmath4 over @xmath5 and a semistable model @xmath6 for @xmath4 such that @xmath567 .
it is sometimes useful to define skeleta and metrized complexes in terms of semistable vertex sets rather than semistable models .
( by ( * ? ? ?
* theorem 5.38 ) , there is a bijective correspondence between the latter two objects . )
a _ semistable vertex set _ for @xmath13 is a finite set @xmath26 of type 2 points of @xmath13 such that the complement of @xmath26 in @xmath13 is isomorphic ( as a @xmath5-analytic space ) to the disjoint union of a finite number of open annuli and an infinite number of open balls .
( such a disjoint union is called the _ semistable decomposition _ of @xmath13 associated to @xmath26 . )
any finite set of type 2 points of @xmath13 is contained in a semistable vertex set ( ( * ? ? ?
* proposition 5.27 ) ) .
the _ skeleton _
@xmath568 of @xmath13 with respect to a semistable vertex set @xmath26 is the union ( inside @xmath13 ) of @xmath26 and the skeletons of each of the open annuli in the semistable decomposition associated to @xmath26 . using the canonical metric on the skeletons of these open annuli ,
@xmath1 can be naturally viewed as a ( finite ) metric graph contained in @xmath13 ( see ( * ? ? ?
* definition 5.17 ) for more details ) .
the skeleton @xmath1 comes equipped with a natural model @xmath28 whose vertices are the points of @xmath26 and whose edges correspond bijectively to the open annuli in the semistable decomposition associated to @xmath26 .
a semistable vertex set @xmath26 is called _ strongly semistable _ if the graph @xmath28 has no loop edges . every semistable vertex set is contained in a strongly semistable vertex set . given a strongly semistable vertex set @xmath26 for @xmath13
, there is a canonical corresponding metrized complex @xmath569 of @xmath0-curves .
indeed , we have already defined a metric graph @xmath568 corresponding to @xmath26 , together with a corresponding model @xmath28 .
for @xmath14 , let @xmath2 be the unique smooth projective curve over @xmath0 with function field @xmath570 .
( for @xmath571 of type 2 , recall that the residue field @xmath572 of the completed residue field @xmath573 of @xmath58 has transcendence degree one over @xmath0 .
) it remains to specify , for each @xmath14 , a bijection @xmath574 from the edges of @xmath28 incident to @xmath3 to a subset @xmath45 of @xmath52 .
given such an edge @xmath37 , we define @xmath575 to be the point of @xmath52 corresponding to the tangent direction at @xmath3 defined by @xmath37 .
( recall from ( * ? ? ?
* paragraph 5.67 ) that if @xmath58 is of type 2 , there is a canonical bijection between @xmath576 and the set of discrete valuations on @xmath577 which are trivial on @xmath0 . )
[ rmk : modelcompatibility ] passing to a larger semistable vertex set is compatible with linear equivalence of divisors and does not change the rank of divisors .
one can thus associate a canonical group @xmath578 to @xmath13 , defined as @xmath579 for any semistable vertex set @xmath26 , together with a canonical rank function @xmath580 .
one can define in a similar way semistable vertex sets and skeleta for an affine curve @xmath581 ( see ( * ? ? ? * definition 5.19 ) ) . in this case
, one must also allow a finite number of punctured open balls in the semistable decomposition and the skeleton is a topologically finite but not necessarily finite length metric graph it will contain a finite number of infinite rays corresponding to the points of @xmath582 , where @xmath4 is the projective completion of @xmath581 .
there is a canonical embedding of @xmath12 in the berkovich analytification @xmath13 of @xmath4 , as well as a canonical retraction map @xmath583 .
in addition , there is a canonical reduction map @xmath560 sending @xmath73 surjectively onto the closed points of @xmath24 .
the retraction map @xmath584 induces by linearity a specialization map @xmath74 which is studied in @xcite .
we can promote this to a map @xmath585 whose target is the larger group @xmath44 as follows . if @xmath586 satisfies @xmath587
then either @xmath103 is the unique point of @xmath13 with @xmath588 equal to the generic point of @xmath2 , or else @xmath588 is a nonsingular closed point of @xmath2 .
the map @xmath589 is obtained by linearly extending the map @xmath590 defined by @xmath591 in this section , we show that if @xmath81 is a principal divisor on @xmath4 , then @xmath592 is a principal divisor on @xmath8 .
in fact , we show a more precise result ( theorem [ thm : plspecialization ] below ) .
let @xmath571 be a point of type 2 .
given a nonzero rational function @xmath136 on @xmath4 , choose @xmath593 such that @xmath594 . define @xmath595 to be the image of @xmath596 in @xmath597 .
although @xmath598 is only well - defined up to multiplication by an element of @xmath599 , its divisor @xmath600 is canonical and the resulting map @xmath601 is a homomorphism .
we call @xmath598 the _ normalized reduction _ of @xmath136 .
if @xmath602 is a @xmath5-linear subspace of @xmath603 , the collection of all possible reductions of nonzero elements of @xmath602 , together with @xmath604 , forms a @xmath30-vector space @xmath605 . for later use
we note the following elementary lemma , which says that @xmath606 : [ lem : dimensionreduction ] let @xmath4 be a smooth proper curve over @xmath5 , and @xmath607 a point of type 2 .
the @xmath30-vector space @xmath605 defined by the reduction to @xmath572 of an @xmath162-dimensional @xmath5-subspace @xmath608 has dimension @xmath609 .
the inequality @xmath610 follows from the observation that if @xmath611 are linearly dependent in @xmath602 , with @xmath612 for all @xmath613 , then the reductions @xmath614 are linearly dependent in @xmath605 . to prove @xmath615 , choose a @xmath5-basis @xmath616 for @xmath602 consisting of elements of norm @xmath203 and let @xmath617 in @xmath605 be the corresponding reductions
it is not true that @xmath617 necessarily form a basis of @xmath605 : for example , if @xmath136 and @xmath20 are linearly independent functions of norm @xmath203 and @xmath618 is any scalar with @xmath619 , then @xmath136 and @xmath620 are linearly independent but they have the same reduction .
however , by the following procedure one can construct another basis @xmath621 of @xmath602 with @xmath622 for all @xmath613 such that the reductions @xmath623 are linearly independent over @xmath30 .
define @xmath624 , and recursively define @xmath625 as follows : among all the rational functions of the form @xmath626 with @xmath627 for all @xmath628 , choose one , call it @xmath629 , whose norm is minimal .
note that @xmath630 so such an element exists by compactness .
let @xmath631 be a scalar with @xmath632 and define @xmath633 .
it is clear that the @xmath634 form a basis of @xmath602 .
we claim that the reductions @xmath623 form a basis of @xmath605 . indeed , if not then there exists a minimal index @xmath613 and scalars @xmath635 with @xmath636 for @xmath637 such that @xmath638 . but
this implies that @xmath639 , contradicting the choice of @xmath629 .
we recall the following useful result from ( * ? ? ?
* theorem 5.69 ) .
[ thm : pl ]
let @xmath136 be a nonzero rational function on @xmath4 , let @xmath581 be an open affine subset of @xmath4 on which @xmath136 has no zeros or poles , and let @xmath640 .
let @xmath26 be a semistable vertex set for @xmath581 , and let @xmath641 be the corresponding skeleton .
then : 1 .
@xmath642 where @xmath643 is the retraction .
2 . @xmath287 is piecewise linear with integer slopes , and @xmath287 is linear on each edge of @xmath644 .
if @xmath58 is a type-@xmath645 point of @xmath13 and @xmath646 corresponds to the discrete valuation @xmath647 on @xmath648 , then @xmath649 .
@xmath287 is harmonic at all @xmath650 .
let @xmath651 , let @xmath37 be the ray in @xmath644 whose closure in @xmath13 contains @xmath58 , let @xmath652 be the other endpoint of @xmath37 , and let @xmath653 be the tangent direction represented by @xmath37
. then @xmath654 .
to each nonzero rational function @xmath136 on @xmath4 and each strongly semistable model @xmath6 for @xmath4 , one can associate a rational function @xmath35 on @xmath8 whose @xmath1-part @xmath137 is the restriction to @xmath655 of the piecewise linear function @xmath656 on @xmath13 and whose @xmath2-part is the normalized reduction @xmath107 ( which is well - defined up to multiplication by a non - zero scalar ) . instead of @xmath657 because our convention is that @xmath658 is the sum of the outgoing slopes of @xmath287 at @xmath447 .
one could equally well take the opposite convention , defining @xmath658 to be _ minus _ the sum of the outgoing slopes of @xmath287 at @xmath447 , and then defining @xmath287 to be @xmath657 .
such a modification would also necessitate a change of sign in our definition of @xmath659 . ] as an application of theorem [ thm : pl ] , we obtain the following important formula : [ thm : plspecialization ] for every nonzero rational function @xmath136 on @xmath4 , @xmath660 let @xmath159 be the support of @xmath661 in @xmath4 .
we apply theorem [ thm : pl ] to @xmath662 , the invertible rational function @xmath136 on @xmath581 , and a semistable vertex set @xmath663 for @xmath581 which contains @xmath26 .
let @xmath664 and @xmath665 .
since @xmath26 is a semistable vertex set for @xmath4 , the closure of @xmath666 in @xmath644 is a disjoint union of metric trees @xmath667 . for each tree @xmath668 ,
denote by @xmath669 all the points in @xmath670 which are in the closure of rays in @xmath668 , and let @xmath671 .
denote by @xmath672 the unique point of @xmath673 .
in addition , if @xmath674 , let @xmath675 . for each @xmath676 , the restriction of @xmath677 to @xmath668 is harmonic at every point of @xmath678 by ( 4 ) . by ( 5 )
, @xmath138 is linear in a sufficiently small neighborhood of @xmath679 in @xmath668 , and has ( constant ) slope @xmath680 along the unique tangent direction in this interval toward the point @xmath672 .
we infer that the order of @xmath681 at @xmath672 is equal to @xmath682 . by ( 1 )
, @xmath287 is constant on any connected component of @xmath683 . by ( 4 )
, the restriction @xmath36 of @xmath287 to @xmath1 satisfies @xmath684 .
since @xmath685 is linear on each edge of @xmath1 by ( 2 ) , it follows that @xmath686 .
fix @xmath676 and suppose that @xmath674 .
for any tangent direction @xmath687 , by ( 3 ) we have @xmath688 .
since @xmath287 is locally constant on @xmath683 , we infer that @xmath689 for any @xmath690 not corresponding to a direction in @xmath644 .
this shows that @xmath691 is supported on @xmath692 . by ( 3 ) , for any point @xmath693 in @xmath694 , corresponding to a tangent direction @xmath695 , we have @xmath696 .
this shows that the coefficient of @xmath693 in @xmath118 is zero .
for a point @xmath697 , let @xmath698 be the closure in @xmath644 of the connected component of @xmath699 which contains all the points @xmath679 with @xmath700 . by ( 1 ) and ( 4 )
, @xmath287 is harmonic at any point of @xmath701 , and is linear of slope @xmath702 in a sufficiently small neighborhood of @xmath703 in @xmath704 along the unique tangent direction in this interval toward the point @xmath672 . by ( 3 ) , we infer that @xmath705 . we conclude that @xmath706 .
in particular , it follows that @xmath141 .
fix a strongly semistable @xmath7-model @xmath6 for @xmath4 and let @xmath8 be the metrized complex associated to @xmath6 .
the following is a variant of lemma 2.8 and corollary 2.11 from @xcite : [ lem : bakspeclem ] for every divisor @xmath79 , @xmath707 let @xmath708 .
it suffices to prove that if @xmath709 , then @xmath710 as well .
the base case @xmath711 is obvious .
now suppose @xmath712 , so that @xmath713
. then there exists an effective divisor @xmath714 with @xmath715 . since @xmath9 is a homomorphism and takes principal ( resp .
effective ) divisors on @xmath4 to principal ( resp .
effective ) divisors on @xmath716 , we have @xmath717 , so that @xmath718 as well .
we may therefore assume that @xmath719 . for each @xmath31 ,
let @xmath720 be a subset of size @xmath721 and denote by @xmath722 the union of the @xmath723 . by theorem [ thm :
f - width ] , @xmath722 is a rank - determining set in @xmath32 .
let @xmath146 be an effective divisor of degree @xmath149 with support in @xmath722 . for each @xmath724
, there exists a point @xmath725 whose reduction is @xmath58 , so there is an effective divisor @xmath27 of degree @xmath149 on @xmath4 with @xmath726 . by our assumption on the rank of @xmath159 , @xmath727 , and so @xmath728 . since this is true for any effective divisor of degree @xmath149 with support in @xmath722 , and since @xmath722 is rank - determining , we infer that @xmath729 as desired .
there are many examples where the inequality in theorem [ lem : bakspeclem ] can be strict .
for example , if @xmath730 for all @xmath14 then @xmath731 and thus the examples of strict inequality from @xcite apply . at the other extreme ,
if @xmath732 is smooth ( so that @xmath28 is a point ) then the specialization inequality becomes the well - known semicontinuity statement @xmath733 , where @xmath734 is the reduction of @xmath159 , and it is clear that such an inequality can be strict ; for example , take @xmath735 where @xmath4 has genus at least @xmath203 and @xmath736 are distinct points with the same reduction in @xmath737 .
in this section we use the specialization lemma for divisors on metrized complexes to provide a framework for obtaining stronger versions of the weighted specialization lemma from @xcite . our key technical tool for this will be the `` @xmath738-function '' on a metrized complex . to help motivate the definition , we first study an analogous @xmath738-function which controls the rank of divisors on connected sums .
suppose the underlying graph of a metrized complex @xmath32 has a bridge edge @xmath739 , and denote by @xmath740 and @xmath741 the two metrized complexes obtained by removing the edge @xmath37 from @xmath32 ( @xmath239 is a vertex of @xmath742 ) .
denote by @xmath743 and @xmath744 the points of @xmath740 and @xmath741 , respectively , corresponding to the edge @xmath37 .
we say that @xmath32 is a _ connected sum _ of @xmath740 and @xmath741 .
there is an addition map @xmath745 which associates to any pair of divisors @xmath746 the divisor @xmath747 on @xmath32 whose restriction to each @xmath742 is @xmath748 .
the following proposition provides a precise description of the @xmath32-rank of @xmath749 in terms of the two rank functions @xmath750 and @xmath751 . in order to state the result
, we introduce a function @xmath752 defined by the condition that @xmath753 is the smallest integer @xmath754 such that @xmath755 .
( by the riemann - roch theorem for @xmath741 , @xmath738 is well - defined . )
[ prop : connectedsumrankformetrizedcomplexes ] with notation as above , we have @xmath756 we prove the equality of the two sides of equation by showing that the inequalities @xmath757 and @xmath758 both hold . by theorem [ thm : f - width ] , we can assume that all the effective divisors we consider below are supported on geometric points ( i.e. , on points of the curves @xmath2 ) .
@xmath759 let @xmath19 be the right - hand term in and let @xmath146 be an effective divisor of degree @xmath19 on @xmath32 supported on geometric points .
we need to show the existence of a rational function @xmath35 on @xmath32 such that @xmath760 .
there is a ( unique ) decomposition of @xmath146 as @xmath761 with @xmath762 and @xmath763 effective divisors of degree @xmath764 and @xmath149 , respectively , having support on geometric points . by the definition of @xmath738
, there exists a rational function @xmath765 on @xmath741 such that @xmath766 . by the choice of @xmath19 as the minimum in equation
, we have @xmath767 , so there exists a rational function @xmath768 on @xmath740 such that @xmath769 .
we can in addition suppose that the @xmath770-part of @xmath768 takes value zero at @xmath771 .
up to adding a constant to the @xmath770-part of @xmath768 , we can define a rational function @xmath35 which restricts to @xmath768 and @xmath765 on @xmath740 and @xmath741 , respectively , and which is linear of slope @xmath772 on the oriented edge @xmath773 in @xmath1 . for this rational function
, we clearly have @xmath774 @xmath775 it will be enough to show that for each @xmath776 , the inequality @xmath777 holds .
let @xmath778 .
we can obviously restrict to the case @xmath779 .
for each effective divisor @xmath763 of degree @xmath149 supported on geometric points of @xmath741 , let @xmath780 be a rational function on @xmath741 with the property that the coefficients of @xmath781 outside @xmath782 are all non - negative , and such that in addition , the coefficient of @xmath744 in @xmath781 is maximized among all the rational functions with this property .
denote by @xmath783 the coefficient of @xmath744 in @xmath781 and let @xmath784 be an effective divisor of degree @xmath149 on @xmath741 , supported on geometric points of @xmath741 , for which @xmath785 is minimal , i.e. , @xmath786 note that by the choice of @xmath754 , for an effective divisor @xmath763 of degree @xmath149 supported on geometric points of @xmath741 and the rational function @xmath780 , we have @xmath787 , which shows that @xmath788 . by the definition of @xmath738 , it follows that @xmath789 . by the definition of @xmath790 , for any rational function @xmath765 on @xmath741 with the property that @xmath791 has non - negative coefficients outside @xmath744 ,
the coefficient of @xmath744 in @xmath792 is at most @xmath793 .
let @xmath762 be an effective divisor of degree @xmath794 on @xmath740 supported on geometric points .
there exists a rational function @xmath35 on @xmath32 such that @xmath795 .
let @xmath768 and @xmath765 be the restrictions of @xmath35 to @xmath740 and @xmath741 , respectively . by the preceding paragraph ,
the coefficient of @xmath744 in @xmath791 is at most @xmath772 .
since the coefficient of @xmath744 in @xmath796 is non - negative , the @xmath1-part of @xmath35 has slope at least @xmath753 along the tangent direction emanating from @xmath797 which corresponds to @xmath744 .
since the coefficients of @xmath118 at the interior points of the edge @xmath798 are all non - negative , it follows that the slope of the @xmath1-part of @xmath35 along the tangent direction emanating from @xmath771 on the edge @xmath798 is at most @xmath772 , which by the assumption @xmath795 implies that the coefficient of @xmath743 in @xmath799 must be at least @xmath753 .
it follows that @xmath800 .
since @xmath762 was an arbitrary effective divisor of degree @xmath801 supported on geometric points of @xmath740 , it follows that @xmath802 , and the claim follows . in the same spirit ,
consider two metric graphs @xmath770 and @xmath803 , and suppose that two distinguished points @xmath804 and @xmath805 are given .
the _ wedge _ or _ direct sum _ of @xmath806 and @xmath807 , denoted @xmath808 , is the metric graph obtained by identifying the points @xmath771 and @xmath797 in the disjoint union of @xmath770 and @xmath803 .
denote by @xmath809 the image of @xmath771 and @xmath797 in @xmath810 .
( by abuse of notation , we will use @xmath3 to denote both @xmath771 in @xmath770 and @xmath797 in @xmath803 . )
we refer to @xmath811 as a _ cut - vertex _ and to @xmath812 as the decomposition corresponding to the cut - vertex @xmath3 .
note that there is an addition map @xmath813 which associates to any pair of divisors @xmath814 and @xmath815 in @xmath816 and @xmath817 the divisor , @xmath818 on @xmath810 defined by pointwise addition of coefficients in @xmath814 and @xmath815 .
let @xmath819 , @xmath820 , and @xmath821 be the rank functions in @xmath822 , and @xmath823 , respectively . for any non - negative integer @xmath149 , denote by @xmath824 , or simply @xmath753 , the smallest integer @xmath754 such that @xmath825 .
then for any divisor @xmath814 in @xmath816 , we have ( by an entirely similar argument ) @xmath826 the above formalism is very handy for studying metric graphs such as the chain of cycles used in @xcite . in particular , the lingering lattice paths studied in @xcite in connection with a tropical proof of the brill - noether theorem can be understood in a natural way as values of @xmath738-functions .
the @xmath738-function formalism is also useful for studying limit linear series , see for example the proof of theorem [ thm : limitseries ] below .
we now define a partial analogue of the @xmath738-function introduced above for a general metrized complex @xmath32 and provide an analogue of proposition [ prop : connectedsumrankformetrizedcomplexes ] .
consider a divisor @xmath81 on @xmath32 with @xmath2-part @xmath88 for @xmath14 .
for every vertex @xmath3 of @xmath28 , define the function @xmath827 as follows . for any @xmath828
, @xmath829 is the smallest integer @xmath830 such that there exists a divisor @xmath831 of degree @xmath832 supported on the points of @xmath833 such that @xmath834 . for any effective divisor @xmath835 on @xmath28 ,
define @xmath836 note that the definition of @xmath738 depends on the data of the divisors @xmath837 .
[ prop : wrank3 ] let @xmath32 be a metrized complex of algebraic curves over @xmath0 . for any divisor @xmath81 on @xmath32 with @xmath1-part @xmath86 , we have @xmath838 where the minimum is over all effective divisors @xmath27 on @xmath28 .
fix an effective divisor @xmath27 supported on the vertices of @xmath28 .
we need to prove that @xmath839 for the sake of contradiction , suppose this is not the case , i.e. , that @xmath840 . since @xmath841 , luo s theorem ( which is a special case of theorem [ thm : f - width ] ) implies that there exists a divisor @xmath842 of degree @xmath843 supported on @xmath844 such that @xmath845 consider all the effective divisors on @xmath32 of the form @xmath846 , where each @xmath145 is an effective divisor of degree @xmath847 on @xmath2 .
note that @xmath846 has degree @xmath848 , so that by the definition of the rank function , for any choice of effective divisors @xmath145 on @xmath2 there exists a rational function @xmath35 on @xmath32 such that @xmath849 .
denote by @xmath36 and @xmath107 the different parts of @xmath35 , so in particular we have @xmath850 one can assume as in the proof of proposition [ prop : combinrank ] that the slope of @xmath36 along the edges incident to each vertex of @xmath28 is fixed .
this shows the existence of the divisor @xmath851 supported on the points @xmath833 such that @xmath852 has rank at least @xmath853 on @xmath2 . by the definition of @xmath854
, @xmath831 has degree at least @xmath855 . since the degree of @xmath831 is the sum of the slopes of @xmath36 along the edges adjacent to @xmath3 , which is by definition @xmath856 , for each @xmath857 we obtain @xmath858 , i.e. , @xmath859 .
in addition , for each non - vertex point @xmath860 we must have @xmath861 .
this implies that @xmath862 , a contradiction .
let @xmath557 be a complete and algebraically closed non - archimedean field with non - trivial absolute value @xmath556 , @xmath7 the valuation ring of @xmath557 , and and @xmath0 its residue field .
let @xmath558 be the value group of @xmath557 .
let @xmath4 be a smooth , proper , and connected @xmath557-curve and fix a semistable @xmath7-model @xmath6 . let @xmath863 denote the skeleton of @xmath6 together with the weight function @xmath864 defined by @xmath865 .
let @xmath866 .
following @xcite , we define a metric graph @xmath867 by attaching @xmath61 loops of arbitrary positive length at each point @xmath868 ( see @xcite for a discussion of the intuition behind this construction . ) any divisor @xmath159 in @xmath1 defines a divisor on @xmath867 with the same support and coefficients ( by viewing @xmath1 as a subgraph of @xmath867 ) . by an abuse of notation , we will also denote this divisor by @xmath159 . however , we distinguish the ranks of @xmath159 in @xmath1 and @xmath867 by using the notation @xmath869 for the former and @xmath870 for the latter . following @xcite ,
the _ weighted rank _
@xmath871 of @xmath159 is defined by @xmath872 . applying the formalism of the @xmath738-function from the previous subsection ,
we obtain the following useful formula for @xmath871 : [ prop : wrank ] for every @xmath873 , we have @xmath874 proceeding by induction , we are reduced to the case @xmath875 .
this case follows from ( [ eq : wrank2 ] ) applied to @xmath876 , @xmath803 a circle , @xmath877 , @xmath878 , and @xmath879 , since in this case one has @xmath880 , @xmath881 , and @xmath882 for @xmath883 .
the following result is a generalization to metric graphs @xmath1 , and to smooth curves @xmath4 over not necessarily discrete non - archimedean fields , of the corresponding statement for graphs which was obtained by different methods in @xcite .
[ thm : weightedmetricspecialization ] for every divisor @xmath79 , we have @xmath884 .
let @xmath885 and note that the @xmath1-part @xmath86 of @xmath81 is equal to @xmath78 .
applying proposition [ prop : wrank3 ] , we obtain @xmath886 for each vertex @xmath3 of @xmath28 , define the function @xmath887 by @xmath888 for @xmath889 and @xmath890 for @xmath891 . if @xmath27 is an effective divisor supported on vertices of @xmath28 , define @xmath892 . by the definition of @xmath854 and clifford s theorem for the curve @xmath2 , for any integer @xmath828
we have @xmath893 . combining this with
, we get the inequality @xmath894 note that for @xmath895 , we have @xmath896 and so , whatever the other coefficients of @xmath27 are , we have @xmath897 in other words , in taking the minimum in , one can restrict to effective divisors @xmath27 with @xmath898 for all @xmath3 , i.e. , @xmath899 .
finally , applying , theorem [ lem : bakspeclem ] , and corollary [ prop : wrank ] , we obtain the inequality @xmath900 we give some direct applications of the results of this section .
let @xmath191 be a canonical divisor on the curve @xmath29 . by theorem [ lem : bakspeclem ]
, we have @xmath901 . by the riemann - roch theorem for metrized complexes ( theorem [ thm : rr - metrizedcomplexes ] ) ,
the divisor @xmath902 has degree zero and non - negative rank , and thus @xmath903 .
in particular , we have @xmath904 .
as mentioned in the introduction , for discretely valued @xmath7 the fact that @xmath904 follows from the adjunction formula for arithmetic surfaces .
but it does not seem straightforward to generalize that argument to the not necessarily discretely valued case .
the following result is an immediate consequence of theorem [ thm : graph.to.curve ] combined with theorem [ lem : bakspeclem ] .
[ cor : bnexistence ] let @xmath30 be a field and let @xmath905 be nonnegative integers .
if the brill - noether number @xmath906 is nonnegative , then for every metrized complex @xmath32 of @xmath30-curves with @xmath907 , there exists a divisor @xmath908 such that @xmath909 . in particular , combining the original specialization lemma from @xcite with theorem [ thm : graph.to.curve ] yields a new proof of ( * ? ? ?
* theorem 3.20 ) , which says that if @xmath910 then for every metrized graph of genus @xmath20 there exists @xmath911 such that @xmath912 .
the present proof is more direct in that one does not need the approximation arguments given in lemma 3.17 and corollary 3.18 of _ loc .
_ the notion of brill - noether rank on metric graphs introduced in @xcite and the corresponding specialization lemma extend to the context of metrized complexes as well .
let @xmath32 be a metrized complex of @xmath30-curves , and let @xmath913 be the subset of all divisors of degree @xmath195 and rank at least @xmath19 .
define @xmath914 to be the largest integer @xmath754 such that for every effective divisor @xmath171 of degree @xmath915 on @xmath32 , there exists @xmath916 such that @xmath917 . if @xmath918 , define @xmath919 .
if @xmath6 is a strongly semistable @xmath7-model of a proper smooth curve @xmath29 such that @xmath920 , and @xmath921 denotes the brill - noether locus of @xmath4 , then we have the inequality @xmath922 .
let @xmath32 be a metrized complex of @xmath30-curves .
we say that a point @xmath58 in @xmath32 is a _ weierstrass point _ if @xmath923 . by theorem [ lem : bakspeclem ] , the specialization of a weierstrass point on @xmath4 is a weierstrass point on @xmath32 .
and by theorem [ thm : graph.to.curve ] combined with theorem [ lem : bakspeclem ] , any metrized complex of @xmath30-curves of genus at least two has at least one weierstrass point .
our aim in this section is to compare our divisor theory on metrized complexes with the eisenbud - harris theory of limit linear series for curves of compact type @xcite .
as we will see , our comparison results lead naturally to a generalization of the notion of limit linear series to nodal curves which are not necessarily of compact type .
note that since eisenbud and harris work over a discrete valuation ring , in order to compare our theory to theirs we will sometimes restrict to this setting ; we note , however , that most of our results hold over the valuation ring of an arbitrary non - trivially valued non - archimedean field .
we start by reviewing some basic facts and definitions concerning the degeneration of linear series in families .
let @xmath924 be a smooth projective curve over @xmath30 . recall that a linear series @xmath194 of ( projective ) dimension @xmath19 and degree @xmath195 , or simply a @xmath10 , over @xmath924 consists of a pair @xmath925 with @xmath245 an invertible sheaf of degree @xmath195 on @xmath924 and @xmath602 a subspace of @xmath926 of @xmath0-dimension @xmath609 .
let @xmath927 be a regular smoothing of a strongly semistable curve @xmath34 over a discrete valuation ring @xmath7 with fraction field @xmath928 .
this means that @xmath929 is proper , @xmath930 is regular , the generic fiber @xmath931 of @xmath929 is smooth , and the special fiber of @xmath929 is @xmath34 .
denote by @xmath932 the dual graph of @xmath34 and let @xmath933 be the set of irreducible components of @xmath34 .
any line bundle @xmath934 on @xmath931 extends to @xmath930 .
the restriction of any two extensions @xmath935 and @xmath936 of @xmath934 to the special fiber @xmath34 are combinatorially equivalent ( c.f . section [ sec : mcnodal ] ) .
the first part is a well - known consequence of the regularity of @xmath930 ( see for example ( * ? ? ?
* chapter ii , 6.5(a ) and 6.11 ) ) . for the second part ,
note that any two extensions @xmath935 and @xmath937 of @xmath938 differ by a divisor supported on the special fiber @xmath34 .
in other words , there exist integers @xmath939 for each vertex @xmath3 of the dual graph of @xmath34 such that @xmath940 .
it is now easy to see that , in the notation of section [ sec : mcnodal ] , @xmath941 and @xmath942 differ by the divisor of the function @xmath943 which sends @xmath3 to @xmath939 .
consider now a @xmath10 @xmath944 on @xmath931 and choose an extension @xmath945 of @xmath938 to @xmath930 . since cohomology commutes with flat base change
, there is a natural isomorphism @xmath946 and @xmath947 corresponds to a free @xmath7-submodule h of @xmath948 such that @xmath949 .
the restriction @xmath950 of @xmath602 to the special fiber @xmath34 is an @xmath162-dimensional subspace of @xmath951 .
this shows the existence of the limit of a @xmath10 once the extension @xmath245 of @xmath938 is fixed .
however , since the extension is not unique , the limit pair @xmath952 is not unique .
if one were to restrict to a single such extension , there would inevitably be a loss of important information ; this can be seen , for example , in the study of limits of ramification points and in the study of limits of smooth hyperelliptic , and more generally , smooth @xmath195-gonal curves , see for example @xcite .
some of the difficulties in formulating a good notion of limits for linear series are already manifest in the atypical behavior of linear series on reducible curves , since ( as mentioned earlier ) they do not satisfy many of the classical theorems which govern the behavior of ( global sections of ) linear series on irreducible curves . the eisenbud - harris theory of limit linear series on curves of compact type provides a way to keep track of the geometry in the limit by choosing an _ aspect _ of the limit for each irreducible component , with some extra conditions relating these different aspects at nodes ( see below for a precise definition ) .
we refer to @xcite for recent refinements of the eisenbud - harris theory , and to @xcite for generalizations of this theory to certain curves not of compact type .
we recall the definition of limit linear series for curves of compact type , following @xcite .
we first need to recall the sequence of orders of vanishing of a @xmath10 @xmath953 at a point on a smooth projective curve @xmath924 over @xmath0 . for any closed point @xmath133 on @xmath924 and any section @xmath954 , denote by @xmath955 the order of vanishing of @xmath136 at @xmath133 . the orders of vanishing at @xmath133 of all the sections of @xmath945 in @xmath602 define a sequence of integers @xmath956 .
( this sequence is obtained by induction : let @xmath957 be the section in @xmath602 with the highest order of vanishing , and define @xmath958 inductively as a section in @xmath602 with the highest possible order of vanishing at @xmath133 among all sections which are linearly independent of the first @xmath613 sections @xmath959 .
then @xmath960 . )
let @xmath34 be a strongly semistable curve over @xmath30 , let @xmath932 be the dual graph of @xmath34 , and for any @xmath31 , let @xmath63 be the corresponding irreducible component of @xmath34 .
when @xmath28 is a tree , the curve @xmath34 is called of _ compact type _ : in this case , @xmath961 is compact .
( recall that in general , we have an exact sequence of algebraic groups @xmath962 let @xmath34 be a curve of compact type .
a _ crude limit _
@xmath10 @xmath194 over @xmath34 is by definition the data of a @xmath10 @xmath963 over @xmath63 for each vertex @xmath14 , called the @xmath63-_aspect _ of @xmath194 , such that the following property holds : if the two components @xmath197 and @xmath63 of @xmath34 meet at a node @xmath133 corresponding to an edge @xmath964 in @xmath27 , then @xmath965 for all @xmath966 .
a crude limit linear series is a _ refined limit linear series _ if all the inequalities in are equalities .
we are now going to associate to @xmath194 a degree @xmath195 divisor on the regularization of @xmath34 ( c.f .
section [ sec : mcnodal ] for the definition ) . roughly speaking ,
this amounts to choosing a divisor of degree @xmath195 on each @xmath63 in the divisor class of @xmath196 .
however , the global divisor on @xmath202 defined by such a collection of divisors does not have degree @xmath195 . to fix this problem
, we need to choose a root @xmath967 for the tree @xmath28 and modify the local divisors using the root .
this is done as follows .
let @xmath1 be the metric graph associated to @xmath28 and let @xmath202 be the regularization of @xmath34 . for each vertex @xmath3 of @xmath28
, let @xmath968 be a divisor of degree @xmath195 on @xmath63 in the divisor class defining @xmath969 ( i.e. , @xmath970 ) .
the subspace @xmath971 can be naturally regarded as a subspace of @xmath972 .
let @xmath967 be a fixed vertex of @xmath28 and let @xmath973 denote @xmath28 considered as a rooted tree with root @xmath967 . for any vertex @xmath974 in @xmath28 , consider the unique path from @xmath447 to @xmath967 in @xmath28 and let @xmath975 be the edge adjacent to @xmath447 along this path .
let @xmath976 be the node of @xmath197 corresponding to the edge @xmath975 .
consider the divisor @xmath81 of degree @xmath195 on @xmath202 defined by @xmath977 we will see below that @xmath81 has rank at least @xmath19 in @xmath202 . in order to derive a more precise characterization of crude limits in terms of a suitable rank function , we need to introduce a refined notion of rank on a metrized complex which takes into account the spaces @xmath163 .
let @xmath32 be a metrized complex of algebraic curves , @xmath1 the underlying metric graph with model @xmath28 , and @xmath978 the corresponding family of smooth projective curves over @xmath30 .
suppose we are given , for each @xmath14 , a non - empty @xmath30-linear subspace @xmath166 of @xmath40 .
we denote by @xmath979 the collection of all @xmath166 .
define the _
@xmath979-rank _ of a divisor @xmath81 on @xmath32 , denoted @xmath980 , to be the maximum integer @xmath19 such that for any effective divisor @xmath146 of degree @xmath19 , there exists a nonzero rational function @xmath35 on @xmath32 , with @xmath2-parts @xmath981 for all @xmath14 , such that @xmath982 .
the following proposition provides an upper bound for @xmath983 in terms of the maximal dimension of @xmath166 for @xmath857 .
[ prop : basic ] let @xmath984 then for any divisor @xmath81 on @xmath32 , we have @xmath985 .
if @xmath986 for some @xmath3 , then either @xmath987 ( if @xmath988 ) or @xmath989 ( otherwise ) , and the proposition holds . for the sake of contradiction , suppose now that @xmath990 .
let @xmath3 be a vertex with @xmath991 .
for any effective divisor @xmath145 of degree @xmath992 on @xmath2 , there exists a rational function @xmath35 on @xmath32 with @xmath981 such that @xmath993 .
taking @xmath1-parts shows that @xmath994 . for a generic choice of @xmath145 , we can assume as in the proof of proposition [ prop : combinrank ] that the slopes of @xmath36 on the edges adjacent to @xmath3 are fixed .
for the divisor @xmath995 , this shows that for any choice of @xmath145 on @xmath2 , there exists an element @xmath981 with @xmath996 .
this contradicts the assumption that @xmath997 , since @xmath166 can define a linear system of projective dimension at most @xmath998 on @xmath2 .
in section [ sec : rank - csum ] , we provided a formula for the rank of a divisor on a metrized complex @xmath32 which is a connected sum of two metrized complexes @xmath740 and @xmath741 .
we now state a straightforward generalization of proposition [ prop : connectedsumrankformetrizedcomplexes ] to restricted ranks , which will be used in the next section .
suppose that the underlying graph of @xmath32 has a bridge edge @xmath999 and let @xmath740 and @xmath741 be the two metrized complexes obtained by removing the edge @xmath37 .
let @xmath770 , @xmath803 , and @xmath1 be the underlying metric graphs of @xmath740 , @xmath741 , and @xmath32 respectively ( with @xmath239 a vertex of @xmath1000 ) .
denote by @xmath743 and @xmath744 the points of @xmath740 and @xmath741 , respectively , corresponding to the edge @xmath37 .
let @xmath979 be a family of spaces of rational functions @xmath166 for vertices @xmath3 of @xmath1 , and denote by @xmath1001 ( resp .
@xmath1002 ) the collection of those @xmath166 for @xmath3 a vertex of @xmath770 ( resp .
@xmath803 ) .
[ prop : connectedsumrankformetrizedcomplexes2 ] let @xmath1003 be the sum of divisors @xmath1004 and @xmath1005 on @xmath740 and @xmath741 , respectively . for a non - negative integer @xmath149 , define @xmath753 to be the smallest integer @xmath754 such that @xmath1006 .
then @xmath1007 we retain the terminology from the previous sections .
let @xmath34 be a curve of compact type , @xmath932 the dual graph , @xmath202 the regularization of @xmath34 , and let @xmath194 be a collection of @xmath10 s @xmath1008 on @xmath63 , one for each @xmath14 .
fix a root vertex @xmath1009 and divisors @xmath1010 such that @xmath1011 , and let @xmath81 be the divisor defined by . denote by @xmath1012 the family of all @xmath1013 .
[ thm : limitseries ] the following two assertions are equivalent : * @xmath194 is a crude limit @xmath10 on @xmath34 . *
@xmath1014 .
denote by @xmath1015 all the children of @xmath967 in the rooted tree @xmath973 .
for each @xmath1016 , denote by @xmath1017 and @xmath672 the @xmath30-points of @xmath1018 and @xmath1019 , respectively , which correspond to the edge @xmath1020 .
proof of @xmath1021 : let @xmath194 be a crude limit @xmath10 and let @xmath1022 be a non - negative integer .
let @xmath1023 be a subspace of dimension @xmath992 , and denote by @xmath979 the collection of @xmath1024 and @xmath163 for @xmath1025 in @xmath26 .
we will prove that @xmath1026 .
the result then follows by taking @xmath1027 . by proposition [ prop : basic ]
, we have @xmath1028 , so it will be enough to prove the reverse inequality @xmath1029 .
the proof proceeds by induction on @xmath1030 , and by applying proposition [ prop : connectedsumrankformetrizedcomplexes2 ] . for the base case of our induction
, we have @xmath1031 , and thus @xmath1032 and @xmath1033 .
now suppose that the claim holds for all metrized complexes on at most @xmath1034 vertices .
for each @xmath1016 , denote by @xmath1035 the rooted tree at @xmath1036 obtained by removing the edge @xmath1037 from @xmath28 , and let @xmath1000 be the metric graph obtained from @xmath1035 by taking all edge lengths to be @xmath203 .
let @xmath742 be the sub - metrized complex of @xmath202 with underlying metric graph @xmath1000 and curves @xmath63 for @xmath3 a vertex of @xmath1035 , and denote by @xmath748 the restriction of @xmath81 to @xmath742 , i.e. , @xmath1038 for all vertices @xmath3 of @xmath1035 .
denote by @xmath1039 the family of all @xmath163 for @xmath3 a vertex of @xmath1035 .
let @xmath1040 be the @xmath738-function associated to the bridge edge @xmath1037 , so that @xmath1041 is the smallest integer @xmath754 such that @xmath1042 .
repeatedly applying proposition [ prop : connectedsumrankformetrizedcomplexes2 ] reduces the calculation of @xmath1043 to the calculation of the restricted ranks of certain divisors on the metrized complex @xmath1019 ( with a single vertex @xmath967 and a unique curve @xmath1019 ) with respect to the subspace @xmath1044 ; more precisely , @xmath1045 we are thus led to prove that each term @xmath1046 in the above formula , for non - negative integers @xmath1047 , is bounded below by @xmath998 .
we may clearly assume that @xmath1048 . for each @xmath1016 , @xmath1049 .
_ proof of the claim .
_ by definition , the subspace @xmath1050 of @xmath1051 consisting of all global sections of @xmath1052 having an order of vanishing at least @xmath1053 at @xmath1054 has dimension @xmath1055 .
let @xmath1056 be the divisor defined on @xmath742 by the data of the linear series @xmath1057 , for @xmath3 a vertex of @xmath1035 , and the root @xmath1036 , as in .
note that @xmath1058 .
let @xmath1059 be the collection of @xmath1050 and @xmath163 , for @xmath1060 a vertex of @xmath1035 . by the induction hypothesis applied to the subcurve of @xmath34 consisting of all curves
@xmath63 for @xmath3 a vertex of @xmath1035 and the crude limit linear series @xmath1061 , we have @xmath1062 . on the other hand , by the definition of @xmath1050
, we have @xmath1063 and thus @xmath1064 by the definition of @xmath1065 , we infer that @xmath1049 , which proves the claim .
to finish the proof of @xmath1021 , note that since @xmath194 is a crude limit @xmath10 , we have @xmath1066 this shows that the space of all global sections of @xmath1067 with an order of vanishing at least @xmath1068 at @xmath1069 has codimension at most @xmath1047 .
the intersection of all these spaces with @xmath1024 has dimension at least @xmath1070 , which shows that @xmath1071 .
proof of @xmath1072 : by corollary [ cor : f - widthz ] , in calculating @xmath1073 we can restrict to effective divisors in @xmath242 . in this case , since @xmath28 is a tree , we can suppose that all the rational functions on @xmath1 which appear in the calculation of @xmath1074 arise via linear interpolation from integral functions @xmath1075 . since @xmath28 is a tree , the data of an integral function @xmath1076 on @xmath26 is equivalent , up to an additive constant , to an assignment of labels @xmath1077 to each oriented edge @xmath1078 of @xmath28 such that @xmath1079 for any @xmath1080 .
( set @xmath1081 . ) in what follows , we denote by @xmath1082 the unique parent of a vertex @xmath1083 in the rooted tree @xmath973 .
suppose now that @xmath1084 .
we first observe that , for any other choice of a root vertex @xmath1085 , if @xmath177 denotes the corresponding divisor on @xmath202 defined as in then the two divisors @xmath81 and @xmath177 differ by the principal divisor @xmath118 , where @xmath35 is a rational function with @xmath1-part @xmath1076 and constant @xmath63-part for all @xmath14 .
thus @xmath1086 , so it will be enough to prove the validity of condition for the root @xmath967 and a child @xmath447 of @xmath967 among @xmath1087 .
denote by @xmath976 and @xmath1088 the @xmath30-points of @xmath197 and @xmath1019 , respectively , corresponding to the edge @xmath1089 , and let @xmath1090 . for any effective divisor @xmath1091 , where @xmath1092 and @xmath1093 are effective divisors of degree @xmath1094 and @xmath613 on @xmath1019 and @xmath197 , respectively
, there must exist a rational function @xmath35 with @xmath1095 for all @xmath452 such that @xmath1096 .
for any two adjacent vertices @xmath452 and @xmath66 in @xmath28 , let @xmath1097 as above . since for any vertex @xmath1098 , @xmath1099 has degree zero , a simple induction starting from the leaves and going towards the root shows that @xmath1100 .
this in particular implies that @xmath1101 and @xmath1102 . for a generic choice of @xmath1103 and @xmath1093
, we can assume as in the proof of proposition [ prop : combinrank ] that @xmath1104 is a constant .
in other words , there exists an integer @xmath1105 such that the sublinear system of @xmath1106 ( resp . ,
@xmath1107 ) which consists of those sections with an order of vanishing at least @xmath1108 ( resp .
@xmath1109 ) at @xmath1088 ( resp .
@xmath976 ) has projective dimension at least @xmath1094 ( resp .
@xmath613 ) . by the definition of the sequences @xmath1110 and @xmath1111 , this simply means that @xmath1112 and @xmath1113 .
it follows that @xmath1114 which is condition for the two irreducible components @xmath1019 and @xmath197 of @xmath34 .
let @xmath30 be of characteristic zero .
then @xmath194 is a refined limit series iff @xmath1115 and all the ramification points of @xmath196 are smooth points of @xmath63 for all @xmath857 .
this is a direct consequence of theorem [ thm : limitseries ] above and the plcker formula , c.f .
* proposition 1.1 ) .
we note that the _ a priori _ dependence of the family @xmath1012 on the @xmath63-aspects @xmath196 can be removed , and recovered from the condition on the rank , in the following sense . for any crude limit @xmath10 @xmath194 on @xmath34 with @xmath63-aspect @xmath963 , choose a divisor @xmath1116 with @xmath1117 .
two divisors @xmath87 and @xmath1118 in @xmath242 are called _ combinatorially equivalent _ if they differ by the divisor of a rational function @xmath35 on @xmath202 with all @xmath107 constant .
the following result is essentially a reformulation of theorem [ thm : introlls ] from the introduction and easily implies the result stated there .
let @xmath34 be a curve of compact type and @xmath202 the regularization of @xmath34 .
then there is a bijective correspondence between the following : * pairs @xmath1119 consisting of a crude limit @xmath10 @xmath194 on @xmath34 and a collection of divisors @xmath88 on @xmath63 with @xmath1120 for any irreducible component @xmath63 of @xmath34 ; and * pairs @xmath1121)$ ] , where @xmath206 , @xmath163 is an @xmath162-dimensional subspace of @xmath207 for each @xmath14 , and @xmath1122 $ ] is the combinatorial linear equivalence class of a divisor @xmath1123 of degree @xmath195 on @xmath202 with @xmath208 .
in particular , the data of a crude limit @xmath10 on @xmath34 is equivalent to the data of a pair @xmath1124)$ ] with @xmath1125 .
we already know by theorem [ thm : limitseries ] that any crude limit @xmath10 gives a pair @xmath1124)$ ] with @xmath1125 .
let @xmath1012 be a collection of @xmath162-dimensional subspaces @xmath1126 for @xmath14 , and let @xmath81 be a divisor of degree @xmath195 in @xmath242 with @xmath1127 we show the existence of a @xmath1128 @xmath1129 on @xmath63 for each @xmath14 such that the collection @xmath1130 defines a crude limit @xmath10 on @xmath34 , and such that @xmath81 is ( combinatorially ) linearly equivalent to the divisor @xmath1131 associated to @xmath194 by .
let @xmath86 ( resp .
@xmath88 ) denote the @xmath1-part ( resp .
@xmath63-part ) of @xmath87 . since @xmath1 has genus zero , for any vertex @xmath3 in @xmath1132 there exists a rational function @xmath137 on @xmath1 such that @xmath1133 .
define @xmath1134 , where @xmath1135 is defined as in ( [ eq : divf2 ] ) .
we claim that @xmath1136 ; this amounts to showing that @xmath1137 for any @xmath1138 .
if we denote by @xmath1139 the subspace of @xmath163 consisting of all rational functions in @xmath163 with this property , it will be enough show that @xmath1140 .
first , we note that for any rational function @xmath1141 on @xmath1 with @xmath1142 , we have @xmath1143 .
indeed , @xmath1142 implies that for any edge @xmath37 adjacent to @xmath3 , the slope of @xmath137 in the tangent direction corresponding to @xmath37 is bounded above by the sum of the coefficients @xmath1144 of the points @xmath452 lying on the connected component of @xmath1145 which contains the edge @xmath37 .
this sum is precisely the slope of @xmath137 in the tangent direction corresponding to @xmath37 . for any effective divisor @xmath27 of degree @xmath19 on @xmath63 , let @xmath146 denote the corresponding divisor on @xmath202 .
since @xmath1125 , there exists a rational function @xmath1146 such that @xmath1147 . since @xmath1148 , the preceding remark implies that @xmath1149 . by the definition of @xmath118
, we conclude that @xmath1150 , and in particular @xmath1151 belongs to @xmath1139 .
since this holds for any @xmath27 , we infer that the linear series on @xmath63 defined by @xmath1139 has ( projective ) dimension at least @xmath19 and so @xmath1139 has dimension at least @xmath609 , thus @xmath1152 .
we conclude that @xmath963 is a @xmath10 on @xmath63 , and thus the collection of @xmath196 for @xmath14 defines a crude limit @xmath10 @xmath194 on @xmath34 . if @xmath1131 is the divisor associated to @xmath194 by , it is straightforward to check that @xmath1131 and @xmath81 differ by the divisor of a rational function @xmath35 on @xmath202 whose @xmath63-parts are all constant and whose @xmath1-part is the linear interpolation of an integer - valued function @xmath1153
. let @xmath32 be a metrized complex of algebraic curves over @xmath30 , @xmath1 the underlying metric graph with model @xmath932 , and @xmath978 the corresponding collection of smooth projective curves over @xmath30 .
we define a _ ( crude ) limit @xmath10 _ on @xmath32 to be an equivalence class of pairs @xmath1154 consisting of a divisor @xmath81 of degree @xmath195 on @xmath32 and a collection @xmath1012 of @xmath162-dimensional subspaces @xmath1155 , for @xmath1156 , such that @xmath1157 .
two pairs @xmath205 and @xmath1158 are considered equivalent if there is a rational function @xmath35 on @xmath32 such that @xmath210 and @xmath211 for all @xmath14 , where @xmath107 denotes the @xmath2-part of @xmath35 . since @xmath1159 for all @xmath1012 as above and all @xmath168
, it follows immediately from theorem [ thm : rr - metrizedcomplexes ] that for any limit @xmath10 on @xmath32 we have if @xmath1160 or @xmath1161 , then @xmath1162 ( riemann s inequality ) , and if @xmath87 is _ special _ then @xmath1163 ( clifford s inequality ) . for curves of compact type ,
these inequalities are established ( by a completely different proof ) in @xcite .
[ thm : limitseriesgeneral ] let @xmath4 be a smooth proper curve over @xmath5 , @xmath6 a strongly semistable model for @xmath4 , and @xmath8 the metrized complex associated to @xmath6 .
let @xmath159 be a divisor on @xmath4 and let @xmath1164 , for @xmath1165 , be a @xmath10 on @xmath4 .
for any vertex @xmath14 , define @xmath163 as the @xmath30-vector space defined by the reduction to @xmath40 of all the rational functions in @xmath947 ( c.f .
section [ section : reductionofrationalfunctions ] ) , and let @xmath1166 .
then the pair @xmath1167 defines a limit @xmath10 on @xmath8 . by lemma [ lem : dimensionreduction ] ,
all the subspaces @xmath163 have dimension @xmath609 over @xmath30 . by proposition [ prop : basic ]
, we only have to show that @xmath1168 . by theorem [ thm : f - width ] , it suffices to show that for any effective divisor @xmath1169 of degree @xmath19 supported on a subset @xmath1170 of @xmath8 with @xmath1171 of size @xmath721 , there exists a rational function @xmath35 on @xmath8 such that @xmath1172 for all @xmath14 and @xmath1173 . for any such @xmath146 , there exists an effective divisor of degree @xmath19 on @xmath4 such that @xmath726 . since @xmath1174 is a @xmath10 on @xmath4 , there exists a rational function @xmath1175 such that @xmath1176 .
let @xmath35 be the corresponding rational function on @xmath8 ( as in the paragraph preceding theorem [ thm : plspecialization ] ) .
we conclude that @xmath1177 .
since @xmath1178 , this shows that @xmath1179 as desired . for the regularization @xmath202 of a strongly semistable curve @xmath34 over @xmath30 , theorem [ thm : limitseriesgeneral ] can be reformulated as follows .
let @xmath1180 be a regular smoothing of @xmath34 over a discrete valuation ring @xmath7 and let @xmath1181 be a @xmath10 over the generic fiber @xmath931 of @xmath929 .
consider an extension @xmath245 of @xmath1174 to @xmath930 , and let @xmath1182 be the restriction of @xmath245 to @xmath34 .
let @xmath81 be a divisor on @xmath202 in the linear equivalence class of @xmath1183 , where @xmath1184 is as in section [ sec : basics ] .
there exists a family @xmath1185 of @xmath162-dimensional subspaces @xmath1126 such that @xmath1186 .
in addition , @xmath163 can be explicitly constructed as follows . for any vertex @xmath3 , choose a rational function @xmath1187 on @xmath202 such that the divisor @xmath1188 is @xmath3-reduced and let @xmath1189 be the @xmath1-part of @xmath1190 . consider the extension @xmath1191 of @xmath934 to @xmath930 and let @xmath1192 be the closure of @xmath947 in @xmath1193 .
define @xmath1194 and @xmath1195 . finally , for any @xmath14 let @xmath1196 be the @xmath3-reduced divisor linearly equivalent to @xmath81 , and define @xmath1197 .
[ thm : regularlimitseries ] let @xmath34 be a strongly semistable curve over @xmath30 and @xmath202 the regularization of @xmath34 . for any regular smoothing @xmath1198 of @xmath34 over a discrete valuation ring @xmath7 , @xmath1199
as defined above is a limit @xmath10 on @xmath34 .
in addition , for each vertex @xmath857 we have @xmath1200 , so that @xmath1201 defines a @xmath1202 on the irreducible component @xmath63 of @xmath34 .
we conclude this section with the following strengthening of theorem [ thm : limitseriesgeneral ] which will be used in [ chabautysection ] .
[ thm : limitseriesgeneral2 ] let @xmath4 be a smooth proper curve over @xmath5 , @xmath6 a strongly semistable model for @xmath4 , and let @xmath8 be the metrized complex associated to @xmath6 .
let @xmath159 be a divisor on @xmath4 and set @xmath1203 .
let @xmath1164 be a @xmath10 on @xmath4 and let @xmath1166 be as in the statement of theorem [ thm : limitseriesgeneral ] .
let @xmath1204 be an effective divisor of degree @xmath37 supported on the smooth locus of the special fiber of @xmath6 , and define @xmath1205 , where @xmath1206 is the restriction of @xmath1204 to @xmath2 .
suppose that @xmath1207 for all @xmath3 .
then the pair @xmath1208 is a limit @xmath1209 on @xmath8 . in the view of theorem [ thm : f - width ] , the proof is similar to the proof of theorem [ thm : limitseriesgeneral ] , by requiring each subset @xmath720 to be disjoint from the support of @xmath1206 . for any effective divisor @xmath171 of degree @xmath19 supported on @xmath1210 ,
the proof of theorem [ thm : limitseriesgeneral ] shows that there exists a rational function @xmath35 such that @xmath1211 .
we claim that in fact @xmath1212 , which implies that @xmath1213 is a limit @xmath1209 on @xmath8 as desired . to see this ,
let @xmath1214 be the @xmath1-part of @xmath1215 and note that since @xmath1216 and @xmath1217 by hypothesis , it suffices to prove that the @xmath2-part of @xmath118 is at least the @xmath2-part of @xmath1218 for all @xmath14 , i.e. , @xmath1219 if @xmath1220 , then since @xmath1221 is disjoint from @xmath1222 we have @xmath1223 , and since @xmath1221 is disjoint from @xmath1224 we have @xmath1225 .
thus ( [ eq : limitseriesgeneral2 ] ) holds in this case since @xmath1226 implies that @xmath1227 . on the other hand , if @xmath1228 , then @xmath1229 and thus ( [ eq : limitseriesgeneral2 ] ) holds because by assumption we have @xmath1230 and in particular @xmath1231 .
thus ( [ eq : limitseriesgeneral2 ] ) is valid for all @xmath1232 as desired . given a divisor @xmath87 of degree @xmath195 and rank @xmath19 on @xmath32 ,
one may ask if there exists a collection @xmath1233 , with @xmath1234 an @xmath162-dimensional subspace of @xmath1235 for all @xmath613 , such that @xmath1236 . ( recall that we always have @xmath1237 . )
if so , we say that @xmath87 can be _ completed to a limit @xmath10 _ on @xmath32 .
moreover , theorem [ thm : limitseriesgeneral ] shows that if @xmath4 is a smooth proper curve over @xmath5 , @xmath1238 is a @xmath10 on @xmath4 , and @xmath8 is the metrized complex associated to a strongly semistable model @xmath6 for @xmath4 , then the associated pair @xmath1167 ( which we call the _ specialization _ of @xmath1239 ) is a limit @xmath10 on @xmath8 .
we say that a limit @xmath10 @xmath1239 on a metrized complex @xmath32 over @xmath0 is _ smoothable _ if there exists a smooth proper curve @xmath4 over @xmath5 such that @xmath1240 for some strongly semistable model @xmath6 of @xmath4 and @xmath1239 arises the specialization of a limit @xmath10 on @xmath4 .
note that by ( * ? ? ?
* theorem 3.24 ) , every @xmath1241-rational metrized complex @xmath32 over @xmath0 is of the form @xmath1240 for some smooth proper curve @xmath4 over @xmath5 and some strongly semistable model @xmath6 of @xmath4 , where _ @xmath1241-rational _ means that the edge lengths in @xmath28 all belong to the value group of @xmath5 . in this section ,
we show by example that not every divisor @xmath87 of degree @xmath195 and rank @xmath19 on a metrized complex @xmath32 can be completed to a limit @xmath10 , and not every limit @xmath10 on a @xmath1241-rational metrized complex @xmath32 is smoothable .
[ ex : notcompletable ] consider the weighted graph @xmath1 depicted in figure [ fig : nog12 ] ( with arbitrary edge lengths ) , and let @xmath32 be a metrized complex whose underlying weighted metric graph is @xmath1 ( so each @xmath234 has genus @xmath203 and @xmath2 has genus @xmath1242 ) . for any point @xmath133 in @xmath1243 , consider the degree @xmath645 divisor @xmath1244 .
we claim that @xmath81 has rank @xmath203 on @xmath32 but can not be completed to a limit @xmath1245 .
the fact that @xmath81 has rank @xmath203 is straightforward and left to the reader . assume for the sake of contradiction that there exists a limit @xmath1245 @xmath1246 on @xmath32 .
let @xmath1247 ( resp .
@xmath1248 ) be the points of @xmath1249 ( resp .
@xmath1250 ) corresponding to the three edges of @xmath1 .
fix @xmath1251 , choose a point @xmath1252 in @xmath1253 and let @xmath1254 . for any rational function @xmath35 on @xmath32 such that @xmath1096 , the fact that @xmath1253 has genus one implies that @xmath1255 ,
since there is no rational function on @xmath1253 whose divisor is @xmath1256 .
therefore @xmath1257 .
restricting @xmath35 to @xmath2 gives a rational function @xmath1258 on @xmath2 with @xmath1259 .
since @xmath1246 is by assumption a limit @xmath1245 , we must have @xmath1260 .
but a simple calculation shows that the @xmath1258 are linearly independent , implying that @xmath1261 , a contradiction .
suppose the residue field @xmath0 of @xmath5 has characteristic different from @xmath645 .
then there exists a metrized complex @xmath32 over @xmath0 ( with all curves @xmath2 isomorphic to @xmath1262 ) and a divisor @xmath87 of degree 2 and rank 1 on @xmath32 which can not be completed to a limit @xmath1245 .
indeed , any hyperelliptic metric graph which can not be realized as the skeleton of a hyperelliptic curve over @xmath5 provides an example of this phenomenon ( and one can construct many examples of such metric graphs , using the characterization of those hyperelliptic metric graphs which can be lifted to a hyperelliptic curve over @xmath5 given in @xcite and @xcite . ) here is a sketch of the argument .
it is not difficult to show that a limit @xmath1245 on a metrized complex @xmath32 gives rise , in a natural way , to a finite harmonic morphism of degree two ( in the terminology of ( * ? ? ?
* definition 2.22 ) ) from @xmath32 to a metrized complex of genus zero . by the lifting result ( * ? ? ?
* corollary 1.5 ) , such a finite harmonic morphism can be lifted to a degree two map from a smooth proper curve @xmath29 with associated metrized complex @xmath32 to @xmath1262 , which implies that the underlying metric graph of @xmath32 is the skeleton of a hyperelliptic curve over @xmath5 .
[ ex : notsmoothable ] this phenomenon occurs already in the compact type case : the original paper of eisenbud - harris on limit linear series @xcite provides an example ( example 3.2 of _ loc .
_ ) of a limit @xmath1263 which can not be smoothed ; we recall their example here .
the metrized complex @xmath32 in question is associated to a semistable curve @xmath34 with two components meeting transversely at a point @xmath133 ; one of the components is a hyperelliptic curve @xmath924 of genus at least @xmath1264 and the other component @xmath1265 has genus zero .
the point @xmath133 is a ramification point of the @xmath1245 on @xmath924 , and without loss of generality we may assume that @xmath133 corresponds to the point @xmath1242 on @xmath1266 .
the dual graph @xmath28 of @xmath34 has two vertices @xmath1267 and @xmath1268 connected by a single edge , which may take to be of length one in @xmath1 .
now define a divisor @xmath87 on @xmath32 by setting @xmath1269 for the point @xmath1270 .
if we define @xmath1271 , where @xmath1272 and @xmath1273 is the span of @xmath1274 ( with @xmath1275 a local parameter at @xmath1242 on @xmath1276 ) , then by ( * ? ? ?
* example 3.2 ) and theorem [ thm : limitseries ] above , the pair @xmath1239 defines a limit @xmath1263 on @xmath32 .
( it is also not hard to show this directly using our definitions . )
this limit @xmath1263 can not be smoothed . in this section , we explain how limit linear series on metrized complexes of curves can be used to illuminate the proof of a recent theorem due to e. katz and d. zureick - brown @xcite and put it into a broader context .
let @xmath928 be a number field and suppose @xmath4 is a smooth , proper , geometrically integral curve over @xmath928 of genus @xmath17 .
let @xmath1277 be the jacobian of @xmath4 , which is an abelian variety of dimension @xmath20 defined over @xmath928 .
if the mordell - weil rank @xmath19 of @xmath1278 is less than @xmath20 , coleman @xcite adapted an old method of chabauty to prove that if @xmath1279 is a prime which is unramified in @xmath928 and @xmath1280 is a prime of good reduction for @xmath4 lying over @xmath133 , then @xmath1281 . here
@xmath1282 denotes the special fiber of a smooth proper model for @xmath4 over the completion @xmath1283 of @xmath1284 at @xmath1280 and @xmath1285 .
stoll @xcite improved this bound by replacing @xmath188 with @xmath1286 .
lorenzini and tucker @xcite ( see also @xcite ) proved the same bound as coleman without assuming that @xmath4 has good reduction at @xmath1280 ; in their bound , @xmath1287 is replaced by @xmath1288 where @xmath6 is a proper regular model for @xmath4 over @xmath1283 and @xmath1289 is the smooth locus of the special fiber of @xmath6 .
katz and zureick - brown combine the improvements of stoll and lorenzini - tucker by proving : [ thm : katz - zb ] let @xmath928 be a number field and suppose @xmath4 is a smooth , proper , geometrically integral curve over @xmath928 of genus @xmath17 .
suppose the mordell - weil rank @xmath19 of @xmath1278 is less than @xmath20 , and that @xmath1279 is a prime which is unramified in @xmath928 .
let @xmath1280 be a prime of @xmath1284 lying over @xmath133 and let @xmath6 be a proper regular model for @xmath4 over @xmath1283
. then @xmath1290 in order to explain the main new idea in the paper of katz and zureick - brown , we first quickly recall the basic arguments used by coleman , stoll , and lorenzini - tucker .
( see @xcite for a highly readable and more detailed overview . )
assume first that we are in the setting of coleman s paper , so that @xmath1291 , @xmath1279 is a prime which is unramified in @xmath928 , and @xmath4 has good reduction at the prime @xmath1292 lying over @xmath133 . fix a rational point @xmath1293 ( if there is no such point , we are already done ! ) .
coleman associates to each regular differential @xmath1294 on @xmath4 over @xmath1295 ( the @xmath1292-adic completion of @xmath928 ) a `` definite @xmath133-adic integral '' @xmath1296 .
if @xmath1297 denotes the vector space of all @xmath1294 such that @xmath1298 for all @xmath1299 , coleman shows that @xmath1300 .
locally , @xmath133-adic integrals are obtained by formally integrating a power series expansion for @xmath1294 with respect to a local parameter . using this observation and an elementary newton polygon argument
, coleman proves that @xmath1301 where @xmath1302 is the minimum over all nonzero @xmath1294 in @xmath1297 of @xmath1303 ; here @xmath1304 denotes the reduction of a suitable rescaling @xmath1305 of @xmath1294 to @xmath1282 , where the scaling factor is chosen so that @xmath1305 is regular and non - vanishing along the special fiber @xmath1282 .
if we choose any nonzero @xmath1306 , then the fact that the canonical divisor class on @xmath1282 has degree @xmath188 gives @xmath1307 which yields coleman s bound .
stoll observed that one could do better than this by adapting the differential @xmath1294 to the point @xmath1308 rather than using the same differential @xmath1294 on all residue classes .
define the _ chabauty divisor _
@xmath1309 then @xmath1310 and @xmath1311 are both equivalent to effective divisors , so by clifford s inequality ( applied to the smooth proper curve @xmath1282 ) we have @xmath1312 . on the other hand , by the semicontinuity of @xmath1313 under specialization we have @xmath1314 .
combining these inequalities gives @xmath1315 which leads to stoll s refinement of coleman s bound .
lorenzini and tucker observed that one can generalize coleman s bound to the case of bad reduction as follows .
let @xmath6 be a proper regular model for @xmath4 over @xmath1283 and note that points of @xmath1316 specialize to @xmath1288 .
one obtains by a similar argument the bound @xmath1317 where @xmath1302 is the minimum over all nonzero @xmath1294 in @xmath1297 of @xmath1303 ; here @xmath1304 denotes the reduction of ( a suitable rescaling of ) @xmath1294 to the unique irreducible component of the special fiber of @xmath6 containing @xmath1308 and @xmath1300 as before . choosing a nonzero @xmath1306 as in coleman s bound , the fact that the relative dualizing sheaf for @xmath6 has degree @xmath188 gives the lorenzini - tucker bound . in order to combine the bounds of stoll and lorenzini - tucker , we see that it is natural to form the chabauty divisor @xmath1318 and try to prove , using some version of semicontinuity of @xmath1313 and clifford s inequality , that its degree is at most @xmath1286 .
this is the main technical innovation of katz and zureick - brown , so we state it as a theorem : [ thm : katz - zb_degree_bound ] the degree of @xmath1310 is at most @xmath1286 . combining theorem [ thm : katz - zb_degree_bound ] with yields theorem [ thm : katz - zb ] . as noted by katz and zureick - brown ,
if one makes a base change from @xmath1295 to an extension field @xmath1319 over which there is a regular semistable model @xmath1320 for @xmath4 dominating the base change of @xmath6 , then the corresponding chabauty divisors satisfy @xmath1321 .
( here @xmath1322 is defined relative to the @xmath1319-vector space @xmath1323 ; one does not want to look at the mordell - weil group of @xmath1277 over extensions of @xmath928 . ) in order to prove theorem [ thm : katz - zb_degree_bound ] , we may therefore assume that @xmath33 is a regular semistable model for @xmath4 ( and also that the residue field of @xmath1319 is algebraically closed ) .
let @xmath1324 .
we now explain how to prove that @xmath1325 when @xmath33 is a semistable regular model using limit linear series on metrized complexes of curves .
our proof is different from ( and arguably streamlined than ) the one given in @xcite .
let @xmath1326 .
we can identify @xmath1327 with an @xmath1328-dimensional space @xmath1329 of rational functions on @xmath4 in the usual way by identifying @xmath1330 with @xmath1331 for a canonical divisor @xmath191 on @xmath4 .
the divisor @xmath1310 on @xmath1289 defines in a natural way a divisor @xmath1332 of degree @xmath195 on the metrized complex @xmath1333 associated to @xmath33 , where @xmath88 is the restriction of @xmath1310 to @xmath2 .
we promote the divisor @xmath1334 to a limit linear series @xmath1335 on @xmath1333 by defining @xmath163 to be the reduction of @xmath1329 to @xmath2 for each @xmath1156 . by the definition of @xmath1310
, each element of @xmath163 vanishes to order at least @xmath1302 at each point @xmath1308 in @xmath1336 . by corollary [ cor : introcanonspec ] and theorem [ thm : limitseriesgeneral2 ] ,
@xmath945 is a limit @xmath1337 on @xmath1333 . in particular
, we must have @xmath1338 . on the other hand ,
clifford s inequality for metrized complexes ( theorem [ thm : mcclifford ] ) implies that @xmath1339 . combining these inequalities gives @xmath1325 as desired
we retain the terminology from section [ sec : basics ] .
let @xmath32 be a metrized complex of algebraic curves , @xmath1 the underlying metric graph , @xmath932 a model of @xmath1 and @xmath1340 the collection of smooth projective curves over @xmath30 corresponding to @xmath32 .
in this section , we generalize some basic results concerning rank - determining sets @xcite from metric graphs to metrized complexes by following and providing complements to the arguments of @xcite ( to which we refer for a more detailed exposition ) .
let @xmath722 be a set of geometric points of @xmath32 ( i.e. , a subset of @xmath1341 ) .
the set @xmath722 is called _ rank - determining _ if for any divisor @xmath81 on @xmath32 , @xmath321 coincides with @xmath1342 , defined as the largest integer @xmath149 such that @xmath150 is linearly equivalent to an effective divisor for all degree @xmath149 effective divisors @xmath146 on @xmath32 with support in @xmath722 . in other words , @xmath722 is rank - determining if in the definition of rank given in section [ sec : basics ] , one can restrict to effective divisors @xmath146 with support in @xmath722 .
the following theorem is a common generalization of ( a ) luo s theorem @xcite ( see also @xcite ) that @xmath26 is a rank - determining set for any loopless model @xmath932 of a metric graph @xmath1 and ( b ) the classical fact ( also reproved in @xcite ) that for any smooth projective curve @xmath407 of genus @xmath20 over @xmath30 , every subset of @xmath1343 of size @xmath1344 is rank - determining .
[ thm : f - width ] let @xmath32 be a metrized complex of algebraic curves , and suppose that the given model @xmath28 of @xmath1 is loopless . let @xmath1345 be a subset of size @xmath721 and let @xmath1346 . then @xmath722 is a rank - determining subset of @xmath32 .
let @xmath81 be a divisor on @xmath32 .
for any point @xmath1347 , let @xmath1348 be the quasi - unique @xmath103-reduced divisor on @xmath32 linearly equivalent to @xmath81 , and denote by @xmath1349 ( resp .
@xmath1350 ) the @xmath1-part ( resp .
@xmath2-part ) of @xmath1348 .
the condition @xmath1353 is equivalent to requiring that @xmath1354 for every effective divisor @xmath146 of degree @xmath203 on @xmath32 . for @xmath1355 , the divisor
@xmath1356 has non - negative rank if and only if @xmath1351 ( by lemma [ lem : reduced ] ) .
similarly , for @xmath14 and @xmath1357 , the divisor @xmath1358 has non - negative rank in @xmath32 if and only if @xmath1359 and @xmath1360 has non - negative rank on @xmath2 ( by lemma [ lem : reduced ] ) . these are clearly equivalent to ( 1 ) and ( 2 ) .
[ lem : rank - determ ] a subset @xmath1361 which has non - empty intersection with each @xmath1362 is rank - determining if and only if for every divisor @xmath81 of non - negative rank on @xmath32 , the following two assertions are equivalent : in view of lemma [ lem : rankone ] , for a rank - determining set the two conditions ( i ) and ( ii ) are equivalent .
suppose now that ( i ) and ( ii ) are equivalent for any divisor @xmath81 on @xmath32 .
by induction on @xmath19 , we prove that @xmath1367 if and only if for every effective divisor @xmath146 of degree @xmath19 with support in @xmath722 , @xmath1368 .
this will prove that @xmath722 is rank - determining .
let @xmath81 be a divisor with the property that @xmath1371 for every effective divisor @xmath146 of degree @xmath609 with support in @xmath722 .
fix an effective divisor @xmath146 of degree @xmath19 with support in @xmath722 . by the base case @xmath1369
, the divisor @xmath150 has rank at least @xmath203 on @xmath32 because @xmath1372 for any @xmath1373 .
thus @xmath1374 for any point of @xmath53 .
this holds for any effective divisor @xmath146 of degree @xmath19 with support in @xmath722 , and so from the inductive hypothesis we infer that @xmath1375 has rank at least @xmath19 on @xmath32 . since this holds for any @xmath57
, we conclude that @xmath81 has rank at least @xmath609 .
let @xmath81 be a divisor of degree @xmath195 and non - negative rank on @xmath32 , and let @xmath86 and @xmath88 be the @xmath1 and @xmath2-parts of @xmath87 , respectively .
define @xmath1376 note that @xmath1377 is a non - empty subset of the symmetric product @xmath1378 of @xmath195 copies of @xmath1 .
consider the _ reduced divisor _
map @xmath1379 which sends a point @xmath1380 to @xmath1349 , the @xmath1-part of the @xmath103-reduced divisor @xmath1381 .
the following theorem extends ( * ? ? ?
* theorem 3 ) to divisors on metrized complexes .
this is based on an explicit description of the reduced divisor map @xmath1382 in a small neighborhood around any point of @xmath1 , similar to the description provided in ( * ? ? ?
* theorem 3 ) in the context of metric graphs .
we merely give the description by providing appropriate modifications to ( * ? ? ?
* theorem 3 ) , referring to _ loc .
_ for more details .
let @xmath103 be a point of @xmath1 and let @xmath1383 be a ( unit ) tangent direction in @xmath1 emanating from @xmath103 . for @xmath382
sufficiently small , we denote by @xmath1384 the point of @xmath1 at distance @xmath387 from @xmath103 in the direction of @xmath1383 .
we will describe the restriction of @xmath1382 to the segment @xmath1385 $ ] for sufficiently small @xmath382 .
one of the two following cases can happen : we note that there is a maximum saturated cut @xmath298 ( i.e. , containing any other saturated cut ) with the property described in ( 2 ) ( see the proof of ( * ? ? ?
* theorem 3 ) for details ) . in the following @xmath298
denotes the maximum saturated cut with property ( 2 ) . in this case
, there exists an @xmath1388 such that for any @xmath1392 , the reduced divisor @xmath1393 has the following description ( the proof mimics that of ( * ? ? ?
* theorem 3 ) and is omitted ) .
let @xmath1394 be all the different tangent vectors in @xmath1 ( based at the boundary points @xmath1395 , respectively ) which are outgoing from @xmath298 .
( it might be the case that @xmath1396 for two different indices @xmath613 and @xmath628 ) .
let @xmath1397 be small enough so that for any point @xmath333 and any tangent vector @xmath1398 to @xmath1 at @xmath58 which is outgoing from @xmath298 , the entire segment @xmath1399 $ ] lies outside @xmath298 and does not contain any point of the support of @xmath86 . for any @xmath1400 and any positive integer @xmath1401
, we will define below a rational function @xmath1402 on @xmath1 .
appropriate choices of @xmath1403 and @xmath1401 will then give the ( @xmath1384)-reduced divisor @xmath1404 , for any @xmath1405 , where @xmath1406 is the rational function on @xmath32 given by @xmath1407 on @xmath1 and @xmath1408 on each @xmath2 .
* @xmath1411 takes value zero at any point of @xmath298 ; * on any outgoing interval @xmath1412 $ ] from @xmath298 , @xmath1411 is linear of slope @xmath262 ; * the restriction of @xmath1411 to the interval @xmath1413 $ ] is linear of slope @xmath1414 ; * @xmath1411 takes value @xmath1415 at any other point of @xmath1416 it remains to determine the values of @xmath1401 and @xmath1419 .
once the value of @xmath1401 is determined , @xmath1419 will be defined as @xmath1420 so that the point @xmath1421 coincides with the point @xmath1384 .
we consider the following two cases , depending on whether or not @xmath103 is a vertex of @xmath28 : * if @xmath1422 , then @xmath1423 .
( note that since @xmath298 is saturated with respect to @xmath1424 , we have @xmath1425 and thus @xmath1426 . ) * if @xmath1427 for a vertex @xmath857 , let @xmath1428 be the outgoing edges at @xmath3 with respect to @xmath298 , and consider the points @xmath1429 in @xmath1362 indexed by these edges .
suppose in addition that @xmath1430 is the edge which corresponds to the tangent direction @xmath1431 . since @xmath298 is a saturated cut with respect to @xmath1432 , the divisor @xmath1433 has non - negative rank in @xmath2 .
define @xmath1401 to be the largest integer @xmath1434 such that @xmath1435 has non - negative rank .
now for any @xmath1436 , the divisor @xmath1393 is @xmath1437-reduced .
( the argument is similar to ( * ? ? ? * proof of theorem 3 ) . )
it follows immediately that the reduced divisor map is continuous on the interval @xmath1438 , and the result follows .
so it suffices to prove that @xmath1440 and @xmath1441 .
since cardinality of @xmath723 is @xmath721 , @xmath723 is rank - determining in @xmath2 .
therefore , ( ii ) implies @xmath1441 .
we now show that @xmath1441 implies @xmath1442 .
let @xmath229 be the set of all @xmath1443 such that @xmath1444 .
by the continuity of the map @xmath1382 , @xmath229 is a closed subset of @xmath1 .
in addition , since @xmath1352 has rank at least one on @xmath2 and @xmath1445 for every vertex @xmath1446 , we have @xmath1447 .
this shows that @xmath1448 is a disjoint union of open segments contained in edges of @xmath28 .
suppose for the sake of contradiction that @xmath1449 , and let @xmath1450 be a non - empty segment contained in the edge @xmath257 of @xmath28 such that @xmath1451 . to see this , note that for any point @xmath1453 $ ] and any tangent direction @xmath1383 for which @xmath1454 $ ] for all sufficiently small @xmath382 , we are always in case ( 1 ) in the description of @xmath1382 .
otherwise , there would be an integer @xmath1455 such that @xmath1456 for all sufficiently small @xmath1391 .
in particular , this would imply ( by the definition of @xmath1457 ) that @xmath1458 , which implies that @xmath1459 , a contradiction .
this proves the claim .
a case analysis ( depending on whether @xmath103 and @xmath1460 are vertices or not ) shows that for a point @xmath1461 , the cut @xmath1462 is saturated for @xmath1463 .
since @xmath1464 , and @xmath298 does not contain @xmath1265 , this contradicts the assumption that @xmath1465 is @xmath1265-reduced . fix a rank - determining set @xmath1468 as in theorem [ thm : f - width ] . since @xmath722 is rank - determining and any effective divisor @xmath146 with support in @xmath722 obviously belongs to @xmath225 , to prove the equality of @xmath1469 and @xmath321 it will be enough to show that the two statements @xmath1470 and @xmath487 are equivalent .
obviously , the former implies the latter , so we only need to show that if @xmath487 then @xmath1470 .
let @xmath3 be a vertex of @xmath28 and @xmath1471 the @xmath3-reduced divisor linearly equivalent to @xmath81 .
by lemma [ lem : reduced ] , @xmath487 is equivalent to @xmath1472 .
now let @xmath81 be an element of @xmath225 with @xmath1473 .
since @xmath31 and @xmath223 contains all the edge - lengths in @xmath28 , it is easy to see that @xmath81 and @xmath1196 differ by the divisor of a rational function @xmath35 with support in @xmath1474 .
in other words , @xmath1475 in @xmath225 . since @xmath1196 is linearly equivalent to an effective divisor in @xmath225 ( with constant rational function on @xmath1 ) , we conclude that @xmath1476 .
[ cor : f - widthz ] let @xmath202 be the regularization of a strongly semistable curve @xmath34 over @xmath30 .
let @xmath945 be a line bundle on @xmath34 corresponding to a divisor @xmath148 .
then @xmath1477 .
e. katz and d. zureick - brown , _ the chabauty - coleman bound at a prime of bad reduction and clifford bounds for geometric rank functions_. to appear in compositio math .
preprint available at arxiv.org/1204.3335 . w. mccallum and b. poonen , _ the method of chabauty and coleman _ , june 14 , 2010 .
preprint , http://www-math.mit.edu/~poonen/papers/chabauty.pdf , to appear in panoramas et synthses , socit math .
de france .
g. mikhalkin and i. zharkov , _ tropical curves , their jacobians and theta functions_. contemporary mathematics 465 : proceedings of the international conference on curves and abelian varieties in honor of roy smith s 65th birthday . | a _ metrized complex of algebraic curves _ over an algebraically closed field @xmath0 is , roughly speaking , a finite metric graph @xmath1 together with a collection of marked complete nonsingular algebraic curves @xmath2 over @xmath0 , one for each vertex @xmath3 of @xmath1 ; the marked points on @xmath2 are in bijection with the edges of @xmath1 incident to @xmath3 .
we define linear equivalence of divisors and establish a riemann - roch theorem for metrized complexes of curves which combines the classical riemann - roch theorem over @xmath0 with its graph - theoretic and tropical analogues from @xcite , providing a common generalization of all of these results . for a complete nonsingular curve @xmath4 defined over a non - archimedean field @xmath5 , together with a strongly semistable model @xmath6 for @xmath4 over the valuation ring @xmath7 of @xmath5 ,
we define a corresponding metrized complex @xmath8 of curves over the residue field @xmath0 of @xmath5 and a canonical specialization map @xmath9 from divisors on @xmath4 to divisors on @xmath8 which preserves degrees and linear equivalence .
we then establish generalizations of the specialization lemma from @xcite and its weighted graph analogue from @xcite , showing that the rank of a divisor can not go down under specialization from @xmath4 to @xmath8 . as an application
, we establish a concrete link between specialization of divisors from curves to metrized complexes and the eisenbud - harris theory @xcite of limit linear series . using this link
, we formulate a generalization of the notion of limit linear series to curves which are not necessarily of compact type and prove , among other things , that any degeneration of a @xmath10 in a regular family of semistable curves is a limit @xmath10 on the special fiber . |
inclusions placed in an elastic medium usually deform it , and experience mediated interactions as the total elastic free energy depends on their separation .
this occurs in crystals and in soft matter systems , such as capillary interfaces , bilayer membranes , liquid crystals , and binary mixtures @xcite . in systems exhibiting near - critical fluctuations ,
a casimir - like effect adds up , which is due to the coupling between the boundary conditions and the fluctuations of the medium @xcite . while these effects are rather well understood , the _ dynamics _ of elastically mediated interactions remains a challenging subject .
mobile inclusions couple to the dynamics of the elastic medium , which produces non trivial effects @xcite , as in the dynamics of fluctuation - induced forces @xcite here , we investigate the dynamics of the force exchanged between two fixed but active inclusions , that are simultaneously triggered . for @xmath4 ,
the inclusions impose no deformation to the medium , either they are outside the medium , or they are in a state in which they do not couple with its elastic field . at @xmath1 , they are triggered , i.e. they actively change state or insert into the medium , and they start to impose an elastic deformation .
we consider two limiting cases : ( i ) soft , linear inclusions , and ( ii ) hard , non - linear inclusions . in the former case , the inclusions couple linearly to the elastic field with a _
constant force_. it is then the competition with the medium that sets the amplitude of the deformation . in the latter case ,
the inclusions couple non - linearly with the elastic field in such a way that they impose a _
constant deformation_. in both cases , we assume that the time scale associated with the switching of the inclusions is much shorter than the response time of the system ; in other words the switching of the inclusions is assumed to be instantaneous .
these are realistic situations : membrane proteins , for instance , may switch upon binding of atp from a non - curving cylindrical shape to a curving conical one @xcite .
either the protein opens - up with a finite force , or it abruptly changes its shape and imposes a local membrane curvature . for membranes proteins ,
the latter case is more realistic , as proteins are much stiffer that the surrounding membrane .
assuming an instantaneous shape change is also a very good approximation , since the dynamics of the membrane is much slower than that of the protein at distances comparable with inter - protein distances @xcite .
for the sake of simplicity , we consider a one dimensional elastic medium described by a scalar gaussian field @xmath5 , and a purely dissipative dynamics of the model a type .
the hamiltonian @xmath6 of the system is equal to the elastic medium s hamiltonian : @xmath7 ^ 2\right\},\ ] ] plus the hamiltonian of the inclusions . for soft , linear inclusions ,
triggered at @xmath1 , we take for the latter : @xmath8,\ ] ] the inclusions being placed at @xmath9 and @xmath10 . here
, @xmath11 is the heaviside step function , @xmath12 is the dirac distribution , @xmath13 is the strength of the inclusions , and @xmath14 . assuming a purely relaxational dynamics , the time evolution of the field is given by @xcite @xmath15 where @xmath16 is a thermal gaussian noise satisfying @xmath17 , with @xmath18 the temperature and @xmath19 boltzmann s constant .
let us rescale all lengths by the field s correlation length @xmath20 , all times by @xmath21 , and all energies by @xmath22 , so that everything becomes dimensionless .
the dynamical equation and the noise correlation function become : @xmath23 note that @xmath13 has been rescaled by @xmath24 and @xmath18 by @xmath25 .
the interaction between the inclusions can be computed from the stress - tensor associated to the elastic medium @xcite ( see also the discussions in refs .
@xcite ) . with the normalized hamiltonian density @xmath26 ,
the stress tensor is given by @xmath27 ^ 2\,.\ ] ] hence , the ensemble average of force acting on the inclusion at @xmath9 is given by @xmath28 ^ 2\rangle -\langle\frac{1}{2}[\nabla\phi(0^+,t)]^2\rangle\,.\ ] ] note that the @xmath29 terms have canceled out because of the continuity of @xmath30 in @xmath9 .
let us decompose the field as its ensemble average plus its stochastic component : @xmath31 since @xmath32 , the force @xmath0 exchanged between the inclusions can be written as @xmath33 , with @xmath34 its the mean - field component and @xmath35 it s casimir , fluctuation - induced , component : @xmath36 ^ 2 -\frac{1}{2}\left[\nabla\phi(0^+,t)\right]^2 , \\
\label{fc } f_\mathrm{c}(t)=\langle\frac{1}{2}[\nabla\tilde\phi(0 ^ -,t)]^2\rangle -\langle\frac{1}{2}[\nabla\tilde\phi(0^+,t)]^2\rangle.\end{aligned}\ ] ] let us first deal with the casimir component @xmath37 . because the theory is gaussian , and because @xmath38 contains only terms that are linear in @xmath30 , the correlation function @xmath39 of the stochastic part @xmath40 of the field is the same as in the case @xmath41 . using the martin siggia
rose formalism @xcite , one can show that it obeys the equation @xmath42 at all points of space and time .
as the physics is translationally invariant for @xmath43 ( in the soft , linear case ) , the quantity @xmath44 ^ 2\rangle$ ] is space independent , therefore continuous , and @xmath45 there is no fluctuation - induced component of the force in the case ( i ) of soft , linear inclusions . before studying the mean - field component @xmath34 ,
let us study the time evolution of @xmath46 .
taking the ensemble average of equation ( [ eqdyn ] ) then its laplace transform , with @xmath47 , we obtain @xmath48.\ ] ] the solution is @xmath49\right\ } & for $ 0\le x\le l$\ , , \cr c_1\exp\!\left[-(x - l)\sqrt{1+s}\right ] & for $ x\ge l$\ , , } \ ] ] with @xmath50 and @xmath51 .
the asymptotic deformation set by the inclusions is @xmath52=\frac{1}{2}b[1+\exp(-l)]$ ] .
we choose @xmath53 $ ] in order to normalize the field in such a way that @xmath54 .
we thus obtain @xmath55s\sqrt{1+s}$ ] , which completely defines @xmath56 . figure [ temporel]@xmath57 shows the profiles of @xmath58 ( in the soft , linear case ) , obtained by numerically inverting the laplace transform using durbin s method @xcite . at short times , the inclusions are not aware of each other , and the deformation is approximatively symmetrical around each inclusion , displaying a characteristic conical shape with fixed angle .
indeed , the term @xmath59 in the dynamical equation implies @xmath60 ( by symmetry ) , which yields a constant slope @xmath61 . taking the limit @xmath62 ,
as the inclusions ignore each other at short times , the deformation obeys @xmath63 for @xmath64 , which implies @xmath65 . with the constant slope
, this implies that the deformation spreads as @xmath66 , as expected for purely relaxational dynamics .
we therefore expect that the inclusions will start interacting after a time @xmath67
. created by two inclusions placed in @xmath9 and @xmath10 , and triggered at @xmath1 . ( a )
soft , linear inclusions imposing a constant force .
( b ) hard , non - linear inclusions imposing a constant deformation . from bottom
to top the time increases logarithmically , with values @xmath68 , @xmath69 , @xmath70 , @xmath71 , @xmath72 , @xmath73 , @xmath74 , @xmath75 and @xmath76 . ]
let us now calculate the force @xmath34 by using equation ( [ fcm ] ) .
the quantities @xmath77 contributing to @xmath78 can be calculated analytically by computing the inverse laplace transform of @xmath79/[s(1+\exp(-l)]$ ] .
we thus obtain for the mean - field force : @xmath80,\ ] ] where @xmath81^{-1}\ ] ] is the equilibrium force .
both quantities being positive , the interaction between the inclusions is attractive . note that @xmath82 can easily be obtained by solving the static problem .
it is obviously the common asymptotic limit of @xmath34 both in the case ( i ) of soft , linear inclusions and in the case ( ii ) of hard , non - linear inclusions , as the asymptotic profiles are identical .
the spreading of the deformation at short times as @xmath66 suggests to plot @xmath83 as a function of @xmath84 .
figure [ forces]@xmath57 shows that this scaling is reasonably good , although there is no true scale invariance in this problem ( due to the existence of a characteristic length and a characteristic time , both equal to unity here ) . between the inclusions as a function of the elapsed time @xmath85 scaled by their squared separation @xmath86 .
( a ) soft , linear inclusions , ( b ) hard , non - linear inclusions .
plain ( red ) curves : @xmath87 .
dashed ( green ) curves : @xmath88 .
dash - dotted ( blue ) curves : @xmath89 . ]
we now set @xmath90 ( i.e. @xmath41 ) and we impose the following hard " boundary conditions : @xmath91 , @xmath92 .
note that for @xmath4 the field fluctuates freely with no boundary conditions . because @xmath93 is unchanged , and because we still use the purely relaxational model a dynamics , the time evolution of @xmath5 is still given by ( [ eqdyn ] ) , with @xmath41 , and its ensemble average @xmath56 by ( [ eqdyn2 ] ) , also with @xmath41 . with the hard boundary conditions in laplace form , this yields @xmath94 the solution is still given by ( [ sol ] ) , but the boundary conditions now yield @xmath95 and @xmath96s$ ] .
figure [ temporel]@xmath97 shows the profiles of @xmath58 , obtained by numerically inverting the laplace transform using durbin s method .
again , at short times , the inclusions are not aware of each other , and the deformation around each inclusion is symmetrical .
the slope of the deformation profile is given by @xmath98 as @xmath64 .
hence , at short times @xmath99 , and since @xmath100 , the deformation spreads again diffusively as @xmath66 .
the force exchanged between the inclusions is still given by @xmath33 , with @xmath34 given by ( [ fcm ] ) and @xmath101 given by ( [ fc ] ) , as the elastic hamiltonian is unchanged .
let us first discuss the mean - field force @xmath102 .
contrary to the case ( i ) of soft , linear inclusions , there is no analytical solution , hence we used durbin s numerical method to compute the inverse laplace transforms of @xmath103 .
figure [ forces]@xmath97 shows the time evolution of @xmath104 for the same set of separations
@xmath3 as in the soft case .
the striking feature is the existence of a bump , which becomes very large when the inclusions are at a distance shorter than the field s correlation length ( i.e. @xmath105 ) .
as shown in figure [ last]@xmath57 , the maximum of the force actually diverges as @xmath106 indeed , when the deformations produced by the inclusions start to merge , the gradient of @xmath107 is of order @xmath108 , and asymmetric ( see figure [ temporel]@xmath97 ) , which by ( [ force ] ) implies the scaling of the force .
this reasoning holds only if the deformations have not relaxed spatially when they merge , which implies the disappearance of the maximum when @xmath3 reaches unity ( figure [ last]@xmath57 ) .
, in the case of hard , non - linear inclusions , as a function of the inclusions separation @xmath3 .
the dotted ( blue ) line has a slope of @xmath109 .
( b ) appearance time @xmath110 ( lower red curves ) , equilibration time @xmath111 ( upper blue curves ) , and maximum time @xmath112 ( middle black curve ) of the force @xmath34 , for a convergence criterion of @xmath113 , as a function of the inclusions separation @xmath3 . dashed curves : soft , linear inclusions ; plain curves : hard , non - linear inclusions . ]
the casimir component @xmath114 is given by ( [ fc ] ) , and it can be derived from the correlation function .
using again the martin siggia
rose formalism , we find that the correlation function @xmath39 is still given by ( [ msr ] ) , as in the soft case ( i ) , but now with boundary conditions such that @xmath115 must vanish if either @xmath116 or @xmath117 is equal to @xmath118 or @xmath3 at positive times .
we thus have dirichlet conditions in @xmath9 and @xmath10 for the fluctuating field @xmath43 at positive times , while there are no boundary conditions at negative times .
there is no need , actually , to calculate @xmath115 . since the model is gaussian , the fourier modes @xmath119 are independent .
they are thermally equilibrated at @xmath4 , then , at @xmath1 all the modes that do not satisfy the dirichlet boundary conditions are suddenly removed while the other remain thermally equilibrated .
it follows that the casimir interaction jumps abruptly from 0 to its equilibrium value : @xmath120 note that this situation differs from that discussed in @xcite , where the dynamics of the casimir force is calculated for a system prepared in the frozen state @xmath121 at time @xmath1 . to calculate the equilibrium value of the casimir force , we have used the method of li and kardar @xcite :
the static correlation function of the field free of boundary conditions is @xmath122 , which yields the casimir energy @xmath123 $ ] and the casimir force @xmath124 by differentiation with respect to @xmath3 .
note that @xmath114 always overcomes the mean - field contribution @xmath102 at small enough separations , but their detailed comparison depend on @xmath18 .
to complete the dynamical study of the mean - field force @xmath83 , let us define the _ appearance _
time @xmath110 , and the _ equilibration _ time @xmath111 , by the conditions @xmath125 with @xmath126 .
the behaviors of @xmath127 and @xmath128 are shown in figure [ last]@xmath97 for @xmath113 ( i.e. a few percent ) .
let us first discuss the case ( i ) of soft , linear inclusions ( dashed curves in figure [ last]@xmath97 ) . for @xmath129 , we expect scaling laws , as the characteristic lengths and times are irrelevant .
indeed , in this regime the force is well approximated by @xmath130 , which can be deduced from ( [ forcemou ] ) by setting @xmath131 and @xmath132 .
it follows that @xmath133 with @xmath134 ^ 2 , \quad d_+=\left[2\,\mathrm{erfc}^{-1}(1-p)\right]^2\simeq\pi p^2.\ ] ] for @xmath113 one obtains @xmath135 and @xmath136 .
note that in figure [ last]@xmath97 the regime @xmath137 is apparent , while the regime @xmath138 occurs at values of @xmath3 smaller than those displayed . in the opposite regime @xmath139 , where there are no scaling laws ,
the force at large times is well approximated by the first term of ( [ forcemou ] ) , i.e. @xmath140 $ ] .
it follows that @xmath141 where @xmath142 ^ 2\,,\ ] ] for @xmath113 one obtains @xmath143 .
the establishment of the force therefore lasts @xmath144 in this regime .
let us now turn to the case ( ii ) of hard , non - linear inclusions .
first of all , for @xmath139 , we notice on figure [ last]@xmath97 that the curves @xmath145 are superimposed on the dashed ones of case ( i ) .
they are therefore given by ( [ tapeqlgrand ] ) .
indeed , whatever the inclusion type , the deformation profiles are already well established in this regime when their tails start to overlap . in the range of values @xmath129 that we have explored , @xmath127 is well fitted numerically by a power law @xmath146 ; we found @xmath147 ( instead of @xmath148 ) , which seems to be independent of @xmath149 in the range @xmath150 $ ] .
a more detailed numerical study would be required , however , to discuss seriously this scaling law .
conversely , the time @xmath112 at which the force maximum occurs is found numerically to follow an analytic power - law : @xmath151 with @xmath152 . as for the equilibration time , we find numerically @xmath128 of order unity for @xmath105 , with @xmath153 as @xmath154 .
in other words , the equilibration time is set by the field s own timescale , even when the inclusions are very close to each other .
another characteristic feature of hard , non - linear inclusions is the existence of a discontinuity in @xmath128 , occuring at @xmath155 for @xmath113 ( see figure [ last]@xmath97 ) .
this discontinuity comes from the maximum of the force @xmath83 : if the latter is larger than @xmath156 then @xmath128 lies after the maximum , otherwise it lies before the maximum .
overall , for @xmath105 , the force in the hard inclusions case appears sooner and equilibrates later than in the soft inclusions case .
we showed that the mechanism by which active , switchable inclusions set their deformations in an elastic medium has important consequences on the dynamics of the force they exchange . in the example studied here , two inclusions instantaneously triggering their deformations may experience a transient mean - field force that can be hundreds of times larger than the equilibrium one .
inclusions setting their deformations through an instantaneous force do not exhibit such a behavior .
the timescales associated with the dynamics of the force are also largely affected by their switching mechanism .
the existence of a high transient force for inclusions setting their deformation instantaneously ( i.e. much quicker than the deformation of the medium ) should be a generic feature of elastically mediated interactions .
it might have important consequences .
for instance , two active inclusions tied to the medium by a link withstanding the equilibrium force could either break free , or remain attached , depending on the nature of their switching mechanism .
we thank f. van wijland for helpful interactions and critical reading of our manuscript . | in a one - dimensional elastic medium with finite correlation length and purely relaxational dynamics , we calculate the time dependence of the elastic force @xmath0 exchanged between two active inclusions that trigger an elastic deformation at @xmath1 .
we consider ( i ) linear inclusions coupling to the field with a finite force , and ( ii ) non - linear inclusions imposing a finite deformation . in the non - linear case ,
the force exhibits a transient maximum much larger than the equilibrium force , diverging as @xmath2 at separations @xmath3 shorter than the field s correlation length .
both the mean - field and the casimir component of the interaction are calculated .
we also discuss the typical appearance time and equilibration time of the force , comparing the linear and the non - linear cases .
the existence of a high transient force in the non - linear case should be a generic feature of elastically - mediated interactions .
_ keywords _ : statistical mechanics of complex materials : stochastic processes ( theory ) , non - equilibrium processes : exacts results , interface biology - physics : active membranes |
the modern paradigm for cosmological inflation is generically termed ` chaotic ' , referring primarily to its choice of initial conditions @xcite . in its simplest form , a single scalar field the inflaton @xmath2
evolves slowly along its potential @xmath3 towards its minimum at the origin .
the essential challenge is to embed consistently this potential in an underlying particle physics model , within a natural range of variation of the parameters involved and respecting the ` slow - roll conditions ' @xcite .
these are upper limits on the normalized slope and curvature of the potential : @xmath4 here @xmath5 is the normalised planck mass and the potential determines the hubble parameter during inflation as , @xmath6 .
inflation ends ( i.e. @xmath7 , the acceleration of the cosmological scale factor , changes sign from negative to positive ) when @xmath8 and/or @xmath9 become of @xmath10 .
it is easy to see that this challenge is _ not _ met by simple models of the potential .
consider the quadratic form @xcite @xmath11 the adiabatic scalar density perturbation generated through quantum fluctuations of the inflaton is @xcite @xmath12 where the subscript @xmath13 denotes the epoch at which a fluctuation of wavenumber @xmath14 crosses the hubble radius @xmath15 during inflation , i.e. when @xmath16 .
( we normalise @xmath17 at the present epoch , when the hubble expansion rate is @xmath18 kms@xmath19mpc@xmath19 , with @xmath20 . )
the cobe observations @xcite of anisotropy in the cosmic microwave background on large angular - scales require @xcite @xmath21 on the scale of the observable universe ( @xmath22 mpc ) .
in addition , the cobe data fix the spectral index , @xmath23 , on this scale : @xmath24 the slow - roll condition ( [ slowroll ] ) together with the cobe normalisation ( [ cobe ] ) then implies @xmath25 at the epoch when the fluctuations observed by cobe were generated .
similarly if we consider the quartic potential @xmath26 @xcite , eqs.([slowroll ] ) and ( [ cobe ] ) imply @xmath27 thus we see that the slow - roll condition ( especially on @xmath28 ) generically requires @xmath29 which is difficult to motivate in a particle physics model . in models incorporating gravity ,
non - renormalisable corrections of the form @xmath30 are unsuppressed at large @xmath31 for any @xmath32 and do not have the monomial form usually assumed in chaotic inflationary models .
furthermore the cobe measurement of the scalar perturbation amplitude implies extreme fine tuning because the only natural scale for @xmath33 is the planck scale @xmath34 , in conflict with eq.([vev ] ) .
alternatively an extremely small value is required for the self - coupling as in eq.([lambda ] ) , which also represents unnatural fine tuning .
given the difficulty of constructing realistic particle physics models which can drive inflation for large @xmath31 , it is natural to ask whether there are viable models which work for small @xmath31 , i.e. ` new inflation ' @xcite in which the potential has a maximum at the origin and the inflaton evolves away from it .
in this case low powers of @xmath35 will dominate the potential during the era of observable inflation .
starting from @xmath36 and assuming that the symmetry properties of the model forbid a linear term , the quadratic term will dominate , giving a potential of the form @xmath37 where the constant vacuum energy @xmath38 is now the leading term in the potential . from eq.([slowroll ] ) we have now the constraint @xmath39 and this is again much smaller than the natural value for the mass . to solve this problem
we are driven to consider theories in which a symmetry prevents @xmath33 from being large . the only known symmetry capable of achieving this is supersymmetry which can guarantee that @xmath33 vanishes in the limit that supersymmetry is unbroken .
however the non - vanishing potential of eq.([newinflation ] ) driving inflation _ breaks _ ( global ) supersymmetry and so , even in supersymmetric models , @xmath33 will be non - zero in general . in the extreme case that the inflaton has vanishing non - gravitational couplings , gravitational effects
will typically induce a mass of order @xmath40 for any scalar field @xcite , in particular the inflaton @xcite .
nevertheless this is a big improvement over the non - supersymmetric case , for the fine tuning problem now simply becomes one of requiring @xmath41 with @xmath42 to obtain successful inflation . in this letter
we summarise the results of a general analysis of supergravity inflationary models in which a quadratic term dominates the evolution of the potential for small values of the inflaton .
as discussed in section [ eta ] , one can find models for which @xmath43 is naturally small , consistent with the slow - roll requirements .
however , even in these models it is difficult to make the mass negligibly small so that the quadratic term is likely to be dominant at small inflaton field values .
for this reason we consider models of inflation driven by a quadratic term to be the most _ natural _ supersymmetric inflationary models .
however such models still leave unanswered several important questions concerning the initial conditions .
the most important question is why the universe should initially be homogeneous enough for slow - roll inflation to begin @xcite . in chaotic models
@xcite , inflation begins when the scale of the potential energy is of order the planck scale and the horizon ( the scale over which the universe must be homogeneous for inflation to start ) is also of order the planck scale .
by contrast in the models discussed here , inflation starts much later when the horizon contains many such planck scale horizons and , in this case , it is difficult to understand how the necessary level of homogeneity can be realised .
however this argument is not really a criticism of the possibility that there be a late stage of inflation but rather a statement that this can not be all there is .
thus we are implicitly assuming that there was some other process which ensured the necessary homogeneity at the begining of quadratic inflation , as in the original chaotic inflation scenario @xcite or in the context of quantum cosmology @xcite .
such models lead to a situation in which a homogeneous universe emerges at the planck era and potential energy is released reheating the universe and possibly setting the conditions for further periods of inflation to occur .
this leads to the second initial condition question , namely why should such a universe subsequently undergo inflation at a lower energy scale ? in the case of interest here
this amounts to asking why the inflaton field has initially such a small value that slow - roll inflation from the neighbourhood of the origin is possible ?
remarkably , within the context of supersymmetric inflationary models there is a straightforward answer because quite generally supersymmetry is broken by finite temperature effects i.e. the inflaton potential at high temperatures is not likely to be the same as that at zero temperature @xmath44 as we discuss later , thermal effects may readily drive the inflaton to the origin where the symmetry is enhanced , thus offering an elegant solution to this initial condition problem .
of course this requires that the system be initially in thermal equilibrium and this is not normally the case in slow - roll inflationary models , particularly since the inflaton should be very weakly coupled in order not to spoil the required flatness of its potential .
however quadratic inflation is special in as much as the value of the potential during inflation is not strongly constrained by the need to generate the correct magnitude of density fluctuations . for the case that the potential energy driving inflation is low , we will show that the processes leading to thermal equilibrium have time to work before the inflationary era starts .
a further possible advantage of lowering the inflationary scale is that it may be then readily identified with a scale already present in particle physics .
for example a low @xmath45 may be identified with the susy breaking scale in the hidden sector , @xmath46 . as is well known @xcite this yields a gravitino mass , @xmath47 of order the electroweak scale , as is needed to avoid the hierarchy problem of grand unified theories .
recently an even more radical solution to the hierarchy problem has been suggested , namely that the only fundamental scale is the electroweak breaking scale itself and that the ( four dimensional ) planck scale is a derived quantity @xcite . as we shall see , slow - roll quadratic inflation is possible even at the electroweak scale . in the next section
we discuss the expectation for the inflaton potential following from supergravity .
readers primarily interested in the cosmological implications may wish to skip this and go straight to section [ para ] .
in @xmath48 supersymmetric theories with a single susy generator , complex scalar fields are the lowest components , @xmath49 , of chiral superfields , @xmath50 , which contain chiral fermions , @xmath51 , as their other components . in what follows we will take @xmath50 to be left - handed chiral superfields so that @xmath51 are left - handed massless fermions .
masses for fields will be generated by spontaneous symmetry breakdown so that the only fundamental mass scale is the normalised planck scale .
this is aesthetically attractive and is also what follows if the underlying theory generating the effective low - energy supergravity theory emerges from the superstring .
the @xmath48 supergravity theory describing the interaction of the chiral superfields is specified by the khler potential @xcite , @xmath52 here @xmath53 and @xmath54 ( the superpotential ) are two functions which need to be specified ; they must be chosen to be invariant under the symmetries of the theory .
the dimension of @xmath53 is 2 and that of @xmath54 is 3 , so terms bilinear ( trilinear ) in the superfields appear without any mass factors in @xmath53 ( @xmath54 ) .
the scalar potential following from eq.([g ] ) is given by @xcite @xmath55 + { \rm d - terms } , \label{v}\ ] ] where @xmath56 at any point in the space of scalar fields @xmath57 we can make a combination of a khler transformation and a holomorphic field redefinition such that @xmath58 at that point and the khler potential takes the form @xmath59 .
in this form , the scalar kinetic terms are canonical at @xmath58 and from eq.([v ] ) , neglecting d - terms and simplifying to the case of a single scalar field , the scalar potential writes @xmath60 = v_{0 } + |\phi|^2 \frac{v_0}{m^2 } + \dots \label{sugra}\ ] ] where @xmath61 .
since inflation is driven by the non - zero value of @xmath62 , we see that the resultant breaking of supersymmetry gives all scalar fields a contribution to their mass - squared of @xmath63 , in conflict with the slow - roll condition ( [ slowroll ] ) on @xmath28 .
this is the essential problem one must solve if one is to implement inflation in a supergravity theory .
however , as stressed earlier , the problem is relatively mild when compared to the non - supersymmetric case because the suppression for @xmath28 need only be by a factor of 10 or so .
there have been several proposals for dealing with this problem .
one widely explored possibility is d - term inflation @xcite .
in particular one may consider an anomalous d - term in eq.([v ] ) of the form @xmath64 where @xmath65 is a constant and @xmath66 are scalar fields charged under the anomalous @xmath67 with charge @xmath68 . such a term , for vanishing @xmath66 , gives a constant term in the potential but does not contribute in leading order to the masses of uncharged scalar fields .
the latter occur in radiative order only giving a contribution to their mass - squared equal to @xmath69 where @xmath70 , and typically @xmath71 for an effective coupling @xmath72 between the charged and neutral fields . the other suggested ways out of the @xmath28 problem do not require non - zero d - terms during inflation but rather provide reasons why the quadratic term in eq.([sugra ] ) should be anomalously small .
one possibility follows from the fact that the mass term of @xmath73 coming from eq.([sugra ] ) actually applies at the planck scale .
since we are considering inflation for field values near the origin the inflaton mass must be run down to low scales . as shown in ref.@xcite ,
in a wide class of models in which the gauge couplings become large at the planck scale the low energy supersymmetry breaking soft masses are driven much smaller at low scales by radiative corrections .
the typical effect is to reduce the mass by a factor @xmath74 where @xmath75 is a gauge coupling evaluated at the scale @xmath76 .
while radiative corrections can cause a significant change in the coupling , the effect is limited and becomes smaller as the gauge coupling becomes small .
for this reason the effective mass at low scales can not be arbitrarily small and typically @xmath77 .
it is also possible to construct models in which the contribution to the scalar mass exhibited in eq.([sugra ] ) is cancelled by further contributions coming from the expansion of @xmath54 in eq.([v ] ) .
one interesting example which arises in specific superstring theories was discussed in ref.@xcite .
another occurs in supergravity models of the ` no - scale ' type @xcite . in these examples , while the scalar mass is absent in leading order , it typically arises in radiative order and so again there is an expectation that the effective mass can not be arbitrarily small .
to summarise , in all these cases for small field values the effective supergravity potential can be conveniently parameterised as a constant term plus a quadratic term involving a real inflaton field @xmath78 ( the real part of @xmath2 ) i.e. just the form of eq.([newinflation ] ) .
however this is the form at tree level only . in general radiative corrections
cause the effective mass , @xmath79 to depend logarithmically on @xmath78 .
thus the full potential has the form @xmath80 . \label{pot}\ ] ] note that in eq.([pot ] ) we have not included a term linear in @xmath81 .
such a term can be forbidden if the theory has a symmetry under which @xmath78 transforms non - trivially and we assume that this is indeed the case .
the form of eq.([pot ] ) provides a useful parameterisation of the inflationary potential in the neighbourhood of the origin .
however when @xmath82 is large higher order terms may be significant and so it is important to determine their structure .
such terms may occur through higher order terms in @xmath53 or @xmath54 in eq.([g ] ) or in higher order terms in the expansion of eq.([sugra ] ) .
typically these terms are expected be associated with new physics at the planck scale , for example a term in the superpotential of the form @xmath83 will give rise to terms of the form @xmath84 in the scalar potential .
however it may happen that such higher order terms arise as a result of integrating out heavy fields in the theory , in which case the mass scale in the denominator can be much smaller than @xmath34 . for example the superpotential @xmath85 describes a field @xmath86 with mass @xmath87 . at scales below @xmath87
the field @xmath86 may be integrated out to give the effective superpotential @xmath88 in which the scale of the higher dimension operator is set by @xmath87 .
this mass may be associated with any of the scales in the theory , e.g. the inflationary scale or the supersymmetry breaking scale , and can thus be much smaller than the planck scale . with this preamble
we now consider the structure of the terms responsible for ending inflation consider first the inflationary models driven by a non - zero f - term .
this is conveniently parameterised by the superpotential @xmath89 which gives @xmath90 as required .
radiative corrections then lead to the form of eq.([pot ] ) .
now consider the form of higher order corrections to this superpotential .
the terms allowed must be consistent with any gauge or discrete symmetries of the models . for example the above form linear in @xmath91 follows if @xmath91 carries non - zero @xmath92-symmetry charge under an unbroken @xmath92-symmetry .
using such symmetries it is straightforward to construct a general potential displaying the possibilities for ending inflation .
we suppose that @xmath2 is a singlet under the @xmath92-symmetry but carries a charge under a discrete @xmath93 symmetry
. then the superpotential has the form @xmath94 where we have suppressed the coefficients of @xmath10 of each term .
this gives rise to the potential @xmath95 plus terms involving @xmath91 which we drop as they do not contribute to the vacuum energy ( since @xmath91 does not acquire a vacuum expectation value ) .
this has the desired structure for ending inflation because the potential vanishes for @xmath96 for @xmath97 . for our analysis we use a slightly simplified form of eq.([einf ] ) ( c.f .
eq.([einfp ] ) ) keeping only the leading @xmath98 term and setting @xmath99 to take account of the possibility discussed above that the scale associated with the higher dimension operators may be below the planck scale .
the situation is similar for the case of @xmath100-term inflation . from eq.([dterm ] ) we see there is a contribution of the form of eq.([einf ] ) .
if @xmath101 is a heavy field it may be integrated out in a similar manner to that discussed above so that the term @xmath102 may give rise to a higher dimension term of the form @xmath103 thus the parameterisation of eq.([einfp ] ) may apply to this case as well .
we wish to study locally supersymmetric slow - roll inflationary models in which the inflaton field evolves from small to large field values . near the origin
the term involving the lowest power of the inflaton field is the most important in the potential , hence a simple parameterisation is adequate . as discussed above
, we may take the inflaton field , @xmath78 , to be real with a quadratic potential of the form ( [ pot ] ) .
it is convenient to work with dimensionless quantities @xmath104 in terms of which the potential and its derivatives are given by @xmath105 during slow - roll inflation with @xmath106 , one has @xmath107 hence at @xmath108 , the spectrum of the scalar perturbations ( [ deltah ] ) can be written as @xmath109 where the spectral index is @xmath110 the @xmath111 sign in eq.([delta ] ) takes into account that @xmath112 is positive ( negative ) at @xmath108 corresponding to the inflaton rolling towards smaller ( bigger ) values of @xmath113 .
we focus on the latter case as it offers a simple way of ending inflation as the field value increases and higher order terms in @xmath113 become important . the number of e - folds before the end of inflation when the fluctuations observed by cobe were generated is @xmath114 assuming instant reheating ; the numerical value would be smaller if reheating is inefficient or if there are later episodes of ` thermal inflation ' @xcite . in our model
the end of inflation occurs due to terms of higher order in @xmath2 of the form given in eq.([einf ] ) . for the moment we parameterise their effect through the inclusion of a further parameter @xmath115 , the value of the field at the end of inflation .
the number of e - folds from @xmath116 onwards should equal or exceed @xmath117 , giving @xmath118 in addition the slow - roll conditions during inflation require @xmath119 where @xmath120 is of @xmath10 ( and can be calculated exactly in the slow - roll formalism ) .
if eqs.([delta ] , [ si ] , [ efolds ] ) satisfy the observational constraints ( [ cobe],[ncobe ] , [ ncobe ] ) the inflationary era will lead to a satisfactory cosmology . for @xmath121 the term proportional to @xmath122 in eq.([pot ] )
is negligible .
we shall concentrate on this limit ( although acceptable inflation is also possible if this is not the case ) .
then eq.([si ] ) constrains the quantity @xmath123 $ ] .
further , eq.([delta ] ) constrains a combination of @xmath124 , @xmath125 , and @xmath126 .
given that we have four parameters to satisfy these equations and that eq.([vpp ] ) can be satisfied for a range of parameters , we see that acceptable inflation follows for a range of @xmath45 .
this opens up the possibility of having a much lower value for the inflationary energy scale than is usually considered .
the value of @xmath45 is very sensitive to the the value of @xmath127 and so it is necessary to consider in some detail what determines the latter .
we will discuss this shortly but first we consider how a reduction in the scale of inflation , @xmath45 , may allow thermal effects to naturally set the initial conditions for inflation .
we assume that some process at the planck scale , presumably quantum cosmological in nature @xcite , creates a homogeneous patch of space - time and releases energy thermalising the universe .
it is important to note that the thermalisation temperature can not be close to the planck scale , regardless of the amount of energy released , essentially because particle interactions are asymptotically free .
for example , an explicit calculation of the @xmath128 annihilation rate into gluons finds that equilibrium can only be attained below a temperature @xmath129 gev @xcite .
let us consider the requirements on the inflaton field for it to be localised at the origin through its couplings to particles in the thermal bath . on dimensional grounds ,
the @xmath130 scattering / annihilation cross - sections at energies higher than the masses of the particles involved are expected to decrease with increasing temperature as @xmath131 , where @xmath132 is the coupling .
thus if the scattering / annihilation rate , @xmath133 is to exceed the hubble expansion rate @xmath134 in the radiation - dominated plasma , then we have a limit on the thermalisation temperature @xmath135 , where @xmath136 counts the relativistic degrees of freedom ( = 915/4 in the minimal supersymmetric standard model ( mssm ) at high temperatures ) . now a yukawa coupling @xmath137 of the inflaton to mssm fields @xmath138 will generate a confining potential at high temperatures , @xmath139 i.e. an effective mass for the inflaton of @xmath140 .
this will rapidly drive the inflaton field to the origin in a time of @xmath141 as the universe cools , the potential energy @xmath142 of the inflaton will begin to dominate over the thermal energy at a temperature @xmath143 . at this epoch the inflaton field will be localised to a region @xmath144 in the neighbourhood of the origin . thus to provide natural initial conditions for new inflation we require that @xmath145 i.e. that the scale of inflation be sufficiently low .
the discussion above yields the constraint @xmath146 , taking @xmath147 .
as mentioned earlier , the end of inflation will be determined by terms which are _ not _ included in the expansion of the inflationary potential keeping only terms up to quadratic order . in general such terms
are quite model dependent as all orders in @xmath2 may contribute significantly for large @xmath2 .
however it is necessary to discuss the expectation for @xmath115 for reasonable forms of the potential since the nature of the inflationary era is strongly dependent on it . in section [ eta ]
we have discussed how higher order terms arise in supergravity theories and how they are restricted by the symmetries of the theory but here we will simply write down the structure of these terms motivated by such considerations .
although this parameterisation involves only a single inflaton field in the present context , it actually covers a wide range of models including hybrid inflationary models in which the end of inflation is triggered by a second field .
the full inflationary potential then has the form : @xmath148 where @xmath149 is a coupling which for simplicity we set to unity ( its most natural value ? ) in our calculations .
the end of slow - roll occurs at @xmath150 using this we may now determine the range of parameters for which acceptable inflation obtains by identifying @xmath126 with the field value at which the fluctuations observed by cobe were generated . in figure [ fig1 ] we show the required inflationary scale for various values of @xmath151 , setting @xmath152 and adopting @xmath153 .
the essential observation is that the quadratic potential generates acceptable inflation for a wide range of values of @xmath45 .
this is important because it opens up the possibility of identifying the inflationary scale with a mass scale already required in the particle physics model .
further if @xmath154 , thermal effects will be important in setting the initial conditions for inflation . for the cases shown in figure [ fig1 ] , inflation typically ends at @xmath155 and the fluctuations observed by cobe exited the hubble radius at @xmath156 , about 50 e - folds of expansion earlier .
we have checked that in all cases @xmath157 exceeds @xmath158 so that there will be sufficient inflation starting with thermal initial conditions .
we have studied in very general terms , an inflationary era driven by a quadratic potential .
the essential requirement for successful inflation is that the inflaton mass be reduced below the hubble parameter .
we have discussed mechanisms which do this , concentrating on the situation where the inflaton mass - squared is negative at the origin .
this has the advantage that thermal initial conditions naturally place the inflaton at the origin , initiating the inflationary era .
whether such inflation occurs is a model dependent question . in compactified string theories the multiplet structure
is completely determined and typically there are many scalar fields with the properties of the inflaton discussed here . in such models inflation driven by
a quadratic potential is a very natural possibility . in order for the above mechanism to work
, the inflaton must couple to the fields in the thermal bath and the scale of its potential must be sufficiently low .
in fact the quadratic parametrization of the inflationary potential allows any value for the inflationary scale @xmath45 .
this is essentially because the end of inflation occurs not due to the violation of the slow - roll conditions by the quadratic potential , but due to higher order non - renormalisable terms which become important as the inflaton evolves towards large field values .
two values for the inflationary scale are of particular interest and will be discussed in detail elsewhere @xcite .
one has @xmath159 , the supersymmetry breaking scale in the hidden sector .
the other with @xmath160 implements inflation at the electroweak scale and could be relevant in the context of theories with submillimeter dimensions . | we study inflation driven by a slow - rolling inflaton field , characterised by a quadratic potential , and incorporating radiative corrections within the context of supergravity . in this model
the energy scale of inflation is not overly constrained by the requirement of generating the observed level of density fluctuations and can have a physically interesting value , e.g. the supersymmetry breaking scale of @xmath0 gev or the electroweak scale of @xmath1 gev . in this mass range
the inflaton is light enough to be confined at the origin by thermal effects , naturally generating the initial conditions for a ( last ) stage of inflation of the new inflationary type . |
over the past two decades , orbiting x - ray observatories from _ einstein _ to have shown that both hot and cool stars are moderately strong sources of soft x - rays at energies betweenm 0.1 and 2 kev . for cooler , late - type stars , the scaling of x - ray luminosity with , e.g. , stellar rotation ,
suggests a solar - type origin , with surface magnetic loops confining hot , x - ray - emitting material within nearly static circumstellar coronae . for hotter , early - type stars ,
the x - ray luminosity has no discernible dependence on rotation , but instead scales nearly linearly with bolometric luminosity , @xmath8 @xcite , suggesting a different emission mechanism , perhaps related to the high - density , radiatively driven stellar wind observed from such stars .
nonetheless , an ongoing question has remained whether such hot - star x - ray emission originates from coronal - like processes occuring in nearly static regions near the stellar surface @xcite , or rather in the highly supersonic wind outflow , perhaps from shocks arising from intrinsic instabilities in the radiative driving mechanism @xcite .
based on the observed strong winds in hot stars ( e.g. , * ? ? ?
* ) , a key discriminant has been the apparent general lack of soft x - ray absorption , which , along with optical and uv spectral constraints , implies limits to the extent , temperature , and fractional contribution of the total x - ray output from a base corona @xcite .
this limited inferred absorption is thus generally seen as evidence for emission arising from lower - density regions at large radii , in general agreement with a wind - shock model . in the context of extreme ultraviolet spectroscopy , @xcite presented models of emission line profiles , from both a coronal source and from an expanding shell , affected by transfer through a stellar wind .
these authors demonstrated that emission line profiles are a key discriminant of the source location .
however , due to the small throughput and modest resolution of the _ extreme ultraviolet explorer _
( ) spectrometers , combined with the extreme opacity of the interstellar medium in this bandpass , emission lines from only one hot star ( @xmath9 cma ) were observed with .
the only high signal - to - noise line seen in this euv spectrum was the ly@xmath10 line , which was observed to be broadened by several hundred km s@xmath11 , lending credence to models that postulate a wind - origin for the hot plasma in ob stars @xcite .
it is only recently , however , that orbiting x - ray observatories like and have been able to provide the first detections of spectrally resolved x - ray emission lines from hot , luminous stars @xcite . through the broadened profiles of these resolved x - ray emission lines , these data provide a new , key diagnostic for inferring whether the x - ray emission originates from a nearly static corona , or from an expanding stellar wind outflow .
a key goal of the present paper is to provide a firm basis for interpreting such x - ray emission line spectra .
specifically , we describe below the method ( 2 ) and results ( 3 ) for computations of synthetic line profiles arising from parameterized forms of x - ray emission in an expanding stellar wind . in comparison to recent work by @xcite , who derived analytic forms for x - ray emission profiles for the case of constant expansion velocity ,
our analysis takes into account a variation of velocity @xmath12 in radius @xmath13 , as parameterized by the so - called ` beta - law ' form @xmath0 , where @xmath14 is the stellar radius and @xmath4 is the wind terminal speed . as in @xcite , the overall level of attenuation
is characterized by an integral optical depth @xmath3 , and the x - ray emission filling factor is assumed to decline outward as a power law in radius , @xmath2 .
however , we also allow for this emission to begin at a specified onset radius @xmath1 that can be set at or above the surface radius @xmath14 .
we first present scaled emission profiles for various combinations of the four parameters @xmath5 , @xmath3 , @xmath6 , and @xmath15 ( 3.1 ) . for selected cases intended to represent roughly the competing models of a `` base coronal '' emission vs. instability - generated `` wind - shock '' emission ,
we next compute synthetic profiles that are broadened by the estimated instrumental resolution for emission lines observed by and ( 3.2 ) .
finally , we discuss ( 4 ) how various features of the synthetic profiles are influenced by properties of the emitting material , and outline prospects for future work in interpreting hot - star x - ray spectra .
extended monitoring observations by x - ray missions like @xcite have shown that for most early - type stars the x - ray emission is remarkably constant , with little or no evidence of rotational modulation , and with overall variations at or below the 1% level .
this suggests that the hot , x - ray emitting material must be well - distributed within the circumstellar volume , whether in a narrow , near - surface corona , or as numerous ( @xmath16 ) , small elements of shock - heated gas embedded in the much cooler , uv - line - driven , ambient stellar wind .
here we model any such complex structure in terms of a spherically symmetric , smooth distribution of x - ray line emission , with emissivity @xmath17 at an observer s wavelength @xmath18 along direction cosine @xmath19 from a radius @xmath13 . the resulting x - ray luminosity spectrum @xmath20 is given by integrals of the emission over direction and radius , attenuated by continuum absorption within the medium , @xmath21 } .
\label{llamdef}\ ] ] the absorption optical depth @xmath22 $ ] can be most easily evaluated by converting to ray coordinates @xmath23 and @xmath24 , and then integrating over distance @xmath25 for each ray with impact parameter @xmath26 , @xmath27 = \int_z^\infty \kappa \rho [ \rp ] \ , d \zp , \label{taudef}\ ] ] where @xmath28 $ ] is the mass density at radius @xmath29 .
we assume here that the x - ray absorption cross - section - per - unit - mass , @xmath30 , is a fixed constant that is independent of wavelength @xmath18 within the line and radius @xmath13 within the wind .
for emission at a given @xmath13 along direction cosine @xmath19 , the optical depth attenuation in eqn .
( [ llamdef ] ) can then be obtained through the simple substitution , @xmath31 = t \left [ \sqrt{1-\mu^2 } \ , r , \ , \mu r \right ]
. \label{t2tau}\ ] ] for non - integer @xmath5 , the optical depth integral ( [ taupz ] ) must generally be done by numerical quadrature .
but for general integer values , e.g. @xmath41 0 , 1 , 2 , or 3 , it is still possible to obtain closed - form analytic expressions for this optical depth , though the expressions can be quite cumbersome to write down explicitly ( see also @xcite ) . in the present work ,
these analytic forms are derived and evaluated using the mathematical analysis software , _
mathematica_. to provide a specific example , we cite here the analytic result for the canonical case of @xmath42 , for which the integral in eqn .
( [ taupz ] ) becomes 2m1 @xmath43 \over \tau_\ast } = % { \pi/2 + % % \arctan ( 1/\sqp2m1 ) - \arctan ( z/(\sqp2m1 \sqrt{z^2+p^2 } ) - \arctan ( z/\sqp2m1 ) % \arctan \left ( { 1 \over \sqp2m1 } \right ) - % \arctan \left ( { z \over \sqp2m1 \sqrt{z^2+p^2 } } \right ) - % \arctan \left ( { z \over \sqp2m1 } \right ) % \over \sqp2m1 } .
\left [ { \arctan \left ( { z'/r ' \over \sqp2m1 }
\right ) + \arctan \left ( { z \over \sqp2m1 } \right ) \over \sqp2m1 } \right ] ^{z'\rightarrow \infty}_{z'=z } . \label{taupzb1}\ ] ] this expression actually applies under the full general conditions cited in eqn .
( [ taupz ] ) , though some care must be taken in its evaluation for the cases with @xmath44 .
let us next consider the line emissivity , which we take here to be of the form @xmath45 , \label{etadef}\ ] ] where @xmath46 is an emission normalization constant , @xmath47 is the rest wavelength for the line - emission , and @xmath48 represents a volume filling - factor for x - ray emission .
note that the radial dependence of this filling factor may reflect variations in the temperature distribution of the plasma , as well as variations in the total emission measure .
the form ( [ etadef ] ) simply assumes the emission is an arbitrarily sharp function centered on the local , doppler - shifted , comoving - frame wavelength @xmath49 , and so ignores all broadening ( e.g. , natural , stark , thermal , turbulent ) . for doppler shifts associated with either thermal or small - scale `` micro - turbulent '' motions ,
the delta - function should be replaced with a finite - width emission profile .
but generally , such microscopic motions can be expected to be no greater than the gas sound speed , which for x - ray emitting plasma of temperature @xmath50 k is of order a few times 100 km @xmath51 . as such
, the associated broadening should be relatively small compared to that expected from the directed motion of the wind outflow , which has a much larger characteristic speed @xmath52 km @xmath51 .
moreover , as discussed in 3.2 , for larger - scale , `` macro - turbulent '' motions , the effects can be taken into account _ a postori _ by incorporating them with the general smearing of the observed line profiles expected from the limited spectral resolution of the x - ray detector . for convenience ,
let us define a scaled wavelength measured from line center in units of the wind terminal speed @xmath4 , i.e. @xmath53 . in terms of this scaled wavelength ,
the integrals ( [ llamdef ] ) for the luminosity spectrum then become @xmath54 } \ , \delta ( x+\mu w(r ) ) \ , , \label{lxdef}\ ] ] where we have used the convention @xmath55 . integrating the dirac @xmath56-function over direction cosine @xmath19
, we then find that the wavelength - dependent x - ray luminosity can be evaluated from the single radial integral , @xmath57 \ , { f(r ) \over r^2 w^3(r ) } \ , \exp^{-\tau \left [ -x / w(r),r \right ] } \ , , \label{lxrint}\ ] ] where @xmath58 is the heaviside ( a.k.a .
unit - step ) function , with @xmath59 if @xmath60 , and zero otherwise .
figure [ fig1 ] plots the optical depth @xmath61 $ ] versus scaled wavelength @xmath62 at selected radii @xmath13 in models with various @xmath5 .
for each radius , there is a critical , positive wavelength @xmath62 where the optical depth increases steeply , representing the strong attenuation of the red - wing emission by the dense inner wind and stellar core .
such core occultation makes the integrand in eqn .
( [ lxq ] ) effectively zero for all @xmath63 where @xmath64 . for blue - wing wavelengths @xmath65
, the termination of the curves indicates the maximum blue - shift allowed by the velocity at the relevant radius .
@xcite has recently derived completely analytic results for the x - ray profiles assuming a power - law emission factor @xmath2 and constant outflow velocity , @xmath36 ( implying @xmath66 ) . as noted above , for the somewhat more general case of nonzero , but _ integer _
values of @xmath5 , the optical depth integral ( [ taupz ] ) can still be done analytically ; but in general the radial integral ( [ lxrint ] ) must be evaluated by numerical quadrature . for this numerical integration , it is convenient ( following @xcite ) to define an inverse radius coordinate @xmath67 , yielding @xmath68 } \
, , \label{lxq}\ ] ] where @xmath69 .
for any chosen model for the spatial variation of emission @xmath48 , eqn .
( [ lxq ] ) can be readily evaluated through a single numerical integration . as a specific illustration , we now present results for the example case of an emission factor with the power - law form @xmath70 for radii @xmath71 , and zero otherwise .
this cutoff implies a slight redefinition to the upper integral bound in eqn.([lxq ] ) , @xmath72 \ , .
\label{uxdef}\ ] ] to focus on the general _ form _ of x - ray line emission , let us define a normalized emission profile , @xmath73 } , \ ] ] which simply sets the profile maximum to have a value of unity .
this effectively removes the dependence on the several free parameters ( e.g. @xmath74 , @xmath75 , @xmath4 , etc . ) that affect the overall emission normalization .
the form of the resulting x - ray line - profile @xmath76 versus scaled wavelength @xmath62 thus depends on four free parameters , viz .
@xmath5 , @xmath3 , @xmath6 , and @xmath77 .
the effects of wind attenuation and other model parameters on the overall x - ray emission level have been addressed in a broad - band context by @xcite , but in focussing here on line profile shapes , we defer to future work any analysis of the overall emission line strength .
figure [ fig2 ] presents synthetic emission profiles for various combinations of these parameters , all plotted vs. the scaled wavelength @xmath62 over the range @xmath78 . for comparison ,
figure [ fig3 ] shows the flat - topped profiles for corresponding cases with optically thin emission , @xmath79 .
some notable general properties and trends are as follows : * the profiles seem most sensitive to @xmath3 and @xmath1 , with @xmath5 and @xmath6 having a comparatively modest influence . *
the emission peaks are always blueward of line center ( i.e. , at @xmath65 ) , with increasing blueshift for increasing @xmath3 and/or increasing @xmath1 . *
the overall line - width likewise increases with increasing @xmath3 and/or @xmath1 .
the narrowest lines occur when both @xmath3 and @xmath1 are near unity .
large @xmath3 with low @xmath1 tend to give blue - shifted `` peaked '' forms , while large @xmath1 tend to be blue - tilted `` flat - top '' forms . *
the blue - to - red slope near line center is always negative . for low @xmath3
it varies from very steep for low @xmath1 to nearly flat for large @xmath1 . for large @xmath3
it has a more intermediate steepness for all @xmath1 . for this simple parameterized model ,
the profiles presented above represent the expected intrinsic line emission , with the only broadening associated with the directed , large - scale wind outflow . in practice , observed line profiles can be further broadened , both by the limited spectral resolution of the detector , as well as by doppler shifting from additional , random motions of the emitting gas .
as already noted in 2.2 , for small - scale thermal or `` micro - turbulent '' motions , the associated broadening can be expected to be no larger than the gas sound speed , which for @xmath50 k plasma is order a few 100 km @xmath51 .
since this is much less than the @xmath80 km @xmath51 characteristic of the directed wind outflow , it should have comparitatively minor effect on line - profiles arising from the outer wind . for the inner wind or a more static coronal " models , larger - scale turbulent broadening , which might in principle have a larger characeristic speed , can still be accounted for _ a postori _ , together with _
instrumental broadening _ arising from the limited spectral resolution of the detectors used in the and satellites . for simplicity
, we assume here that both macro - turbulent and instrumental broadening can be characterized by convolving a simple gaussian kernel function over the intrinsic line - profile . as such
, a single parameterized width @xmath7 , defined here in units of the wind terminal speed @xmath4 , represents the combined effects of both any macroturbulent motion and intrumental broadening , with @xmath81 .
let us now consider examples of such broadened line - profiles . to keep the parameter range tractable , we identify specific parameter subsets to represent general conditions in the two distinct scenarios that have often been invoked in modeling hot - star x - ray emission , namely the coronal " ( e.g. , * ? ? ?
* ) vs. wind - shock " scenarios ( e.g. , * ? ? ?
the coronal scenario envisions x - ray emission as coming mostly from a relatively narrow , nearly static layer near the stellar surface ; we represent this here with the parameter values @xmath82 , @xmath83 , and @xmath84 .
the wind - shock scenario generally envisions x - ray production arising from instability - generated shocks that form in the acceleration region @xmath85 of the star s line - driven wind , with eventual slow decay of the shocks in the outer wind ; we represent this here with the parameter values @xmath86 , @xmath42 , and @xmath87 .
for both scenarios we again consider the effect of various optical depth parameters , @xmath3 .
figure [ fig4 ] compares results for the coronal ( left ) vs. wind - shock ( right ) scenario . in characterizing results
, we shall refer to profiles with small vs. large @xmath7 as representing good / high vs. coarse / low resolution . here
some notable general properties and trends include : * at the best resolution ( @xmath88 ) , the coronal models are , as expected , notably narrower than the wind models . *
the wind - shock models are generally more blue - shifted than the coronal models .
* at the coarsest resolution , ( @xmath89 ) , all profiles have similar width , but still generally differ in overall blue - shift .
* for the optically thin case , the emission is nearly symmetric about line center .
for other cases , the emission peaks become more blue - shifted with increasing optical depth @xmath3 .
( the coronal models with larger @xmath3 may not be relevant , however , since wind absorption would greatly reduce the overall strength of any observable emission . )
the typical spectral resolution for is around @xmath90 .
for characteristic wind terminal speed of @xmath91 , this implies a typical broadening parameter of @xmath92 .
for , the spectral broadening is comparable , though slightly larger . clearly , at such high resolutions , it should be possible to place strong constraints on the site of x - ray emission and the degree of wind attenuation , as well as on the characteristic speeds of directed vs. turbulent motions of the emitting plasma .
the parameterized emission models here provide a good basis for interpreting x - ray line - profile observations made possible by the new generation of orbiting x - ray telescopes .
while detailed , quantitative analyses must await the general availability of specific datasets , it is already clear from the few hot - star x - ray spectra published thus far that there are some noteworthy challenges for modeling the associated x - ray emission .
for example , temperature - sensitive line - ratio diagnostics provide constraints on the magnitude of shocks in dynamical models .
generally instability generated wind - shock models tend to be dominated by moderate shock - strengths @xmath93 km @xmath51 that can only heat material up to a maximum of @xmath94 k @xcite . in those stars with much higher inferred temperatures , for example @xmath95 ori c with @xmath96 k @xcite , there is a need to produce stronger shocks .
one possibilty is for wind outflow to be channeled by magnetic loops @xcite . in this case ,
wind material accelerated up opposite footpoints of a closed loop can be forced to collide near the loop top , in principle yielding shocks with velocity jumps of order @xmath97 km @xmath51 , a substantial fraction of the wind terminal speed .
another key diagnostic stems from the forbidden - to - intercombination ( f / i ) ratios of helium - like emission lines , which can provide a strong constraint on the location of the x - ray emission . in the case of the solar corona ,
such ratios are nominally interpreted as density diagnostics @xcite .
but because the strong uv radiation of hot - stars can generally dominate over electron collisions in the destruction of forbidden line upper levels , in such stars the f / i provides instead a constraint on the strength of the uv intensity @xcite . with appropriate atmospheric modeling of the relevant part of uv continuum
, any inferred dilution of the radiation thus provides a diagnostic of the proximity of the x - ray emission to the stellar surface . for observations of @xmath98 pup @xcite , initial interpretations suggest formation well away from the stellar surface ( @xmath99 ) , consistent with a wind - shock origin , with moreover a general trend toward larger radii for lower - energy lines , as would be expected from the larger optical depths and thus larger unit - optical - depth radii for such lines . for chandra observations of @xmath98 ori @xcite ,
the inferred formation radii are also generally away from the star ( @xmath100 ) , but with one stage ( ) suggesting a formation very close to the surface ( @xmath101 ) . however , in this latter case , the relevant uv flux for destruction of the forbidden upper levels lies at energies above the lyman - edge , and so any error in atmospheric modeling of the lyman jump could still alter this inferred formation radius .
chandra observations of @xmath98 pup @xcite show similar results , with most inferred formation radii in the range @xmath102 , with trends again consistent with inferred unit - optical - depth radii , but again with one stage ( in this case ) implying formation quite close to the surface ( @xmath103 ) .
of course , such fir constraints on source location must eventually be considered in conjunction with additional constraints , such as those provided by analysis of line profiles .
indeed , in the context of this paper , the most notable new diagnostic provided by recent x - ray observations lies in the form of the line profiles .
for some initial results , these have some quite puzzling characteristics .
for example , in the spectrum of @xmath98 pup @xcite , the lines all exhibit a similar , significant broadening , but with half - widths @xmath104 km s@xmath11 that are still somewhat lower than the inferred wind terminal speed @xmath105 km s@xmath11 .
the typical profiles have a rounded , gaussian form that is quite distinct from , e.g. , the flat - topped profiles expected for optically thin emission from a nearly constant velocity expansion . moreover , the profiles show only modest asymmetry , with the peak intensity only slightly blue - shifted , by @xmath106 km s@xmath11 relative to line - center , and the red - wing only slightly flatter and more extended than the blue - wing .
recent somewhat higher spectral resolution observations by chandra do show the greater shift and asymmetry expected from optically thick wind emission for @xmath98 pup @xcite .
but in chandra observations of the o9.7 ib star @xmath98 ori , the profiles are again very broad ( half - widths of @xmath107 1000 km @xmath51 ) , but nearly symmetric @xcite .
such profiles with substantial broadening , but little asymmetry and blueshift are difficult to reconcile with a simple wind - outflow model that includes attenuation by the wind and stellar core , both of which should preferentially attenuate x - rays from the back - side hemisphere , and thus in a wind or a corona with some outflow , preferentially reduce the _ red_-side emission .
in fact , from detailed modeling constrained by observations , @xcite have argued that that the wind of should be optically thick to 1 kev x - rays at 10 , with the radius of optical depth unity being even greater for softer x - rays ( 20 at 0.6 kev ; and still 2 at 2 kev ) . even for a base corona with some overall outflow velocity and _ no _ overlying wind attenuation , occultation by the stellar core should still produce line profiles that are skewed toward the blue . a coronal model with little or no net outflow
could explain the observed profile symmetry , but matching the observed line - widths would then require the coronal material to have extremely large `` turbulent '' motions .
for example , line widths in the chandra observations of @xmath98 ori @xcite require velocity dispersions of order 1000 km @xmath51 , and some of the lines observed by in @xmath95 ori c @xcite are even broader than this , yet apparently still nearly symmetric . if the lack of red - wing attenuation is interpreted to mean that red - shifted emission does not arise from wind outflow from the backside hemisphere , then this suggest that material _ in front of the star _ must be flowing _
downward _ at velocities of 1000 km @xmath51 .
since this is greater than the surface escape speed , it can not be the result of gravitational infall , but instead would imply that some mechanism , e.g. , perhaps magnetic reconnection @xcite , must actively accelerate roughly equally strong upflow and downflow with speeds of 1000 km @xmath51 ! moreover , such random motions should inevitably lead to high - speed gas collisions and their associated shock heating . for shocks with velocity jumps of order 1000 km @xmath51
the associated post - shock temperature should be of order @xmath108 k , implying generally harder x - rays and higher ionization stages than are typically observed . in this context
, these broad but symmetric profiles would indeed seem to represent a strong challenge for theoretical models .
on the other hand , our synthesized line profiles from _ wind _ models do seem capable of producing relatively broad , yet symmetric and unshifted profiles , but only in the nearly complete absence of wind attenuation , and with emission originating far from the occultation of the stellar core . to test the viabilty of this alternative ,
observed line - profiles should be interpreted in conjunction with other diagnostics , like the f / i ratio , that can constrain the radii of x - ray emission , and with an accurate account of wind absorption , as computed through opacity calculations for the observationally inferred wind mass loss rate . any wind shock model with significant wind attenuation
is expected to be strongly blueshifted and quite asymmetric ( see figure [ fig4 ] ) .
of all the line profiles discussed in the first group of and papers , the observations of @xmath98 pup provide the best evidence for the blue - shifts and asymmetry expected for optically thick wind emission @xcite . for example , figure 3 of cassinelli et al .
( 2001 ) shows 5 out of 6 lines with blueward shifts and asymmetry similar to profiles calculated here for wind - shock models with substantial optical depth , e.g. @xmath109 .
at least for this canonical o - type star , these profiles seems to provide strong evidence of a wind origin for most of the observed x - rays . in summary ,
systematic fitting of the line - profiles will be necessary to place more quantitative constraints on the spatial distribution of hot plasma and the degree of wind attenuation in the observed o stars ; but some initial qualitative inferences are already possible .
for some cases , the implied minimal wind attenuation combined with some significant bulk motion are puzzling , both in the context of coronal x - ray emission and in the context of the strong radiation - driven winds that are known to exist on these stars from uv , optical , ir , and radio observations .
one possible solution to the wind attenuation problem might be found in clumping .
a `` porous '' wind might allow many of the x - rays to stream through unattenuated , yet also lead to significant uv absorption if the interclump wind medium still has sufficient column density to produce optically thick uv lines .
( the uv line transition probabilities are significantly larger than the bound - free x - ray cross sections . )
an alternative is that the mass - loss rates of o stars may have been overestimated .
it does not seem likely , however , that the winds are dense and smooth but transparent to x - rays , as that would require levels of ionization in the wind that would not allow for efficient radiation - driving . and
, if the winds are indeed optically thin , then to discriminate between coronal and wind models will require a careful analysis of the shifts and asymmetries observed in the x - ray line profiles . finally ,
although we developed our analysis with hot stars in mind , it could be extended to other x - ray sources in which lines are emitted within a nearly symmetric expanding medium that may also have substantial continuum opacity .
this could include nova winds , planetary nebulae , active galactic nuclei , and x - ray binaries .
we thank r. ignace for providing a preprint of his paper on analytic x - ray profiles and also for his detailed comments on this manuscript .
we also thank k. gayley and a. ud - doula for helpful comments on our analysis , and j. cassinelli and j. macfarlane for many useful insights regarding the x - ray spectra of o stars .
finally , we thank the referee , m. corcoran , for helpful comments and suggestions on the submitted manuscript .
this research was supported in part by nasa grant nag5 - 3530 to the bartol research institute at the university of delaware . | motivated by recent detections by the and satellites of x - ray line emission from hot , luminous stars , we present synthetic line profiles for x - rays emitted within parameterized models of a hot - star wind .
the x - ray line emission is taken to occur at a sharply defined co - moving - frame resonance wavelength , which is doppler - shifted by a stellar wind outflow parameterized by a ` beta ' velocity law , @xmath0 . above some initial onset radius @xmath1 for x - ray emission , the radial variation of the emission filling factor is assumed to decline as a power - law in radius , @xmath2
. the computed emission profiles also account for continuum absorption within the wind , with the overall strength characterized by a cumulative optical depth @xmath3 . in terms of a wavelength shift from line - center
scaled in units of the wind terminal speed @xmath4 , we present normalized x - ray line profiles for various combinations of the parameters @xmath5 , @xmath3 , @xmath6 and @xmath1 , and including also the effect of instrumental and/or macroturbulent broadening as characterized by a gaussian with a parameterized width @xmath7 .
we discuss the implications for interpreting observed hot - star x - ray spectra , with emphasis on signatures for discriminating between `` coronal '' and `` wind - shock '' scenarios .
in particular , we note that in profiles observed so far the substantial amount of emission longward of line center will be difficult to reconcile with the expected attenuation by the wind and stellar core in either a wind - shock or coronal model . * = " 002a = # 1to 0pt#1 |
classical cepheid variable stars are of key importance in the extragalactic distance scale and in studies of galactic stellar and chemical evolution .
but the very property that makes them useful as distance indicators pulsation also makes cepheids difficult to model accurately . for several decades ,
a major problem in the understanding of cepheids has been the discrepancy between masses estimated from stellar evolution theory and those estimated from pulsation theory @xcite . while the problem was partly addressed by revision of radiative opacities and changes in evolutionary tracks @xcite , the discrepancy persists at the 1020% level @xcite .
measurements of several cepheid masses are now available from binary systems @xcite , but the comparison with evolutionary and pulsational masses has not yet produced a definitive result .
there are a number of possible solutions to the ` cepheid mass discrepancy , ' summarized by @xcite .
recent studies have come to differing conclusions on which might be responsible : @xcite suggested that mass loss can account for the discrepancy , while @xcite concluded that additional mixing in cepheid main sequence progenitors was responsible .
mass loss remains one of the less - understood parameters in stellar evolutionary theory @xcite , and its relevance to cepheid masses is still not completely clear . in the last several years , several lines of investigation have suggested that mass loss may be important in cepheids .
these include the discovery of extended near - infrared emission from a number of galactic cepheids @xcite , possibly from circumstellar shells , as well as theoretical modeling of pulsation - driven mass loss @xcite and its application to magellanic cloud cepheids @xcite .
both the presence of shells around nearby cepheids and the evidence for infrared excess in lmc cepheids have implications for the stars derived diameters and therefore for the distance scale .
systematic effects on distances are possible , if cepheid mass - loss is substantial enough to affect spectral energy distributions , and if it works differently in nearby and distant galaxies ( e.g. , is a function of metallicity ) .
@xcite concluded that mass loss in cepheids `` has an important effect on the structure of the ir p - l [ infrared period - luminosity ] relations '' although the extent of the effect is not well characterized .
attempts to estimate cepheid mass - loss rates through infrared observations began about 25 years ago .
@xcite used infrared excesses inferred from the iras point source catalog to estimate mass loss rates of @xmath8 m@xmath9 yr@xmath5 , while @xcite combined ultraviolet and infrared observations to estimate upper limits of @xmath10 m@xmath9 yr@xmath5 .
@xcite used radio observations to place similar upper limits on the mass loss in ionized gas .
in general these rates are too low to account for the mass discrepancy .
recently , the _
spitzer space telescope _ has been used in a number of studies of cepheids in the mid - infrared , including the period - luminosity relation @xcite .
@xcite combined ground - based infrared imaging and interferometric measurements with _
data ( including data from the program described here ) to detect compact circumstellar envelopes around both @xmath1 car and rs pup ; the latter was previously known to be associated with a reflection nebula . to investigate the possibility of infrared excesses due to mass loss , we have obtained
_ spitzer _
@xcite observations of a sample of 29 classical cepheids .
our results on the mid - infrared period - luminosity and period - color relations for cepheids are presented in ( * ? ? ?
* paper i ) ; a detailed analysis of extended emission near is presented by @xcite . in paper
i we determined that there was no evidence from irac colors for warm circumstellar dust . the present paper investigates the presence of cooler dust around this sample of cepheids via a search for extended mid - infrared emission .
an important consideration is that extended emission may be the result of recent star - forming activity near a cepheid rather than mass loss from the cepheid itself .
however , distinguishing those two different causes is not straightforward , so a search for extended emission places an upper limit on the amount of mass loss . except where mentioned otherwise , in all of the following analysis we use the stellar distances given by @xcite .
the sample selection and first - epoch observations are described in detail by @xcite . briefly , 29 galactic cepheid variables and 3 non - variable supergiants were observed with both irac @xcite and mips @xcite as part of _ spitzer _ space telescope general observer , program i d 30666 .
the irac observations described by @xcite were primarily made in subarray mode .
the 6 brightest stars were observed in full - array mode so that fluxes could be measured by fitting the outer part of the point spread function ( psf ) ; 4 stars were observed in both full - array and subarray modes .
the mips observations used the photometry astronomical observing template with the shortest possible observation time ( 48.2 s ) at 24 and varying times ( 37.71300.2 s ) at 70 .
three stars ( s nor , v cen , u sgr ) were observed in two adjacent fields of view at 70 to better cover the surrounding clusters . while our target sz tau ( see [ sec : sztau ] ) may also be in the cluster ngc 1647 @xcite , after inspecting its 24 image , we changed the planned 70 observation to use only one field of view centered on the star .
this allowed us to increase the observation time and , as a bonus , also provided a second epoch of 24 observations .
we performed simple aperture photometry on post - bcd mosaics primarily from the pipeline version s16.1 output .
for the three cluster stars , the 70 data were remosaiced using mopex to combine the data for the two fields of view . for two stars ( rs pup , gh lup )
the 70 mosaics were remade to remove negative sidelobes . to follow - up earlier results , we obtained some additional _ spitzer _ data in cycle 5 , on the dates 2008 october 56 .
all stars were observed in full - array mode with irac as part of guaranteed time program 50350 ( pi g. fazio ) .
this increased the field of view for stars which had previously been observed only in subarray mode , and also provided deeper observations ( 2-second frames at each of 36 dither positions ) for all stars .
each target fainter than @xmath11 was observed with the irac full frame 12 second high dynamic range ( hdr ) mode , using a 12 point reuleaux small scale dither pattern for a total integration time of 144 sec in each band .
hdr 12 second frames were used to better reveal the area closer to the saturated cores of the targets , which also helped the psf subtraction process .
for the six brightest sources , in order to avoid excess saturation , targets were observed with 2 second frames using irac full frame non - hdr mode and the 36 point reuleaux small scale dither pattern , for a total integration time of 72 s in each band .
the reuleaux dither pattern was used to achieve the required total exposure time while obtaining the best possible spatial sampling of the targets in anticipation of the psf subtraction process .
the basic calibrated data ( bcd ) for all observations were reduced using the custom post - bcd software iracproc @xcite , which generated mosaics for each source with the final exposure depth and outliers removed .
the final images had a pixel scale of 0.8627 arcsec pix@xmath5 ( @xmath12 the original irac pixel scale ) which provides a better sampling of the sources psf for a more accurate psf subtraction .
the psf subtraction was also performed using a routine part of iracproc , which has been extensively tested in other _ spitzer _
programs @xcite to search for faint structures and point sources around bright stars .
although the mips observations used in this paper are the same as those in @xcite , a major purpose of the present work is to examine extended emission around a subset of the stars in our sample .
accordingly , we have performed point source fitting and subtraction on the 24 images of some targets to separate the contributions of the star and extended emission to the total flux density .
the data were reprocessed with the mips instrument team data analysis tool using standard procedures .
the psf used for the 24 @xmath13 m subtraction was built from the observations of the a4 v star , @xmath14 eri @xcite , which matches well with the theoretical psf computed from stinytim program @xcite after proper smoothing .
the subtraction was done by iteratively shifting the psf to match the centroid of the target in sub - pixel accuracy , adjusting its scaling , and subtracting it from the target image until the residual inside @xmath15 ( @xmath16 to the first bright airy ring of the psf ) was minimized .
the residual flux is generally less than 1% of the total flux of a typical point source on a clean background .
to estimate the non - point source emission near stars showing extended emission in the mips bands , we used the results of point source fitting and subtraction .
we performed aperture photometry of the point and extended emission on images which had the field ( but not the target ) stars subtracted , using the largest apertures which could be contained in the image . from this
we subtracted the flux measured by fitting and subtracting the target stars to derive the total non - point source emission .
this procedure depends on accurate background subtraction for the large - aperture photometry , particularly difficult in the 70 images , and the resulting uncertainties are large . an alternative approach which can constrain the physical conditions within regions of extended emission
is used by @xcite : measuring the surface brightness of the extended emission in small areas . because the stars are heavily saturated in the irac bands , this is the only method available to investigate the extended emission visible in the irac images . in both irac and mips images we measured the surface brightness in several small regions near stars with extended emission , and subtracted backgrounds measured in regions located off the extended emission .
the regions used are shown in the images of the relevant stars ; the measured surface brightnesses are reported in table [ tab : surfphot ] .
infrared colors were used in paper i to demonstrate a lack of warm circumstellar dust in our sample of cepheids . however
, dust mass lost from cepheids and then transported to larger distances would be expected to be cooler , and possibly visible as extended infrared emission .
one example of this phenomenon is the infrared nebula visible around @xmath0 cep @xcite : here we complete the survey of our cepheid and comparison star sample .
none of the stars showed any extended emission in the 3.6 @xmath13 m or 4.5 @xmath13 m bands , so those bands are not discussed further here .
distinguishing between extended emission associated with a specific star and that due to nearby galactic cirrus and/or star formation , especially at low galactic latitudes , is not always straightforward .
we expect that emission associated with an individual star should be centered on or near the star , with its surface brightness decreasing with increasing distance from the star .
unrelated infrared emission will not in general have these characteristics , although of course it could hide or confuse real low level emission . for brevity
, we refer to unrelated infrared emission as ` cirrus ' below , although we recognize that it may not be true galactic cirrus emission . of our sample of 29 cepheids , 16 showed clear signs of cirrus in the irac 8 @xmath13 m images : , , , , , , , , , , , , , , , and .
cirrus is also visible in the mips images of gh lup , v cen , vy car and u sgr .
as expected , the presence of cirrus is related to galactic latitude : of the cepheids in our sample with @xmath17 , 15 of 23 appear to be surrounded by cirrus emission . the remaining star with apparent unrelated emission is dt cyg , at @xmath18 .
the list of cirrus sources include three of the four stars we believed as cluster members ( s nor , v cen , u sgr ) ; this classification is confirmed by @xcite , who also found that vy car is a member of the car ob 2 association .
the extended emission near these stars could be associated with either their clusters or the general interstellar medium .
figure [ fig - cirrus ] shows some examples of stars surrounded by cirrus . as a quantitative test for excess infrared emission near the stars
, we computed the average surface brightness around each star and compared it to a background measurement .
the irac 8 images were used for this measurement as they have a larger uniform - exposure field of view than the mips images , and much better sensitivity to extended emission than the 5.8 images .
the average surface brightnesses were measured in 1 arcminute radius apertures surrounding each star , with the psf subtraction residuals and bright star artifacts masked .
the background measurement was made in an annulus with inner and outer radii of 74 and 98 arcseconds .
point sources near the target stars , imperfect psf subtraction , and unrelated cirrus emission all complicate the comparison between low - surface - brightness emission near the stars and in their vicinities . to be conservative ,
we consider as having possible extended infrared emission only those stars for which the average surface brightness in the measurement aperture exceeds that in the background aperture by 20 times the quadrature sum of the surface brightness standard deviations .
since our measurement errors are are influenced by the factors listed above and are unlikely to be gaussian , we do not consider these detections to have true @xmath19 significance
the value chosen is intended as a starting point to guide the analysis .
we found no evidence for extended infrared emission around two stars which have been previously found to have circumstellar envelopes : @xcite and @xcite .
other stars with neither cirrus or extended infrared emission are , , , , , and .
five stars showed both cirrus and extended emission centered on or near the star : gh lup , rs pup , s mus , t mon , and x cyg .
three further stars had little detectable cirrus but did have extended infrared emission centered on the star : @xmath0 cep , @xmath1 car , and sz tau . the extended emission around @xmath0 cep is discussed in detail by @xcite and briefly summarized here .
the emission consists of a large - scale ( 0.1 pc ) arched structure visible at 70 , with diffuse , filamentary emission at 5.8 , 8.0 and 24 contained within the arch .
the brightest region of this diffuse emission is between @xmath0 cep and its hot companion .
the symmetry axis of the 70 emission near @xmath0 cep is aligned with the star s velocity vector relative to the local ism , suggesting a bow - shock at least partly swept - up from the ism .
the emission near the stars is hypothesized to be due to the interaction of their stellar winds .
the mass loss rates required for the creation of these structures are estimated to be in the range @xmath20 m@xmath21 yr@xmath5 .
the remaining stars with extended infrared emission are discussed individually below .
images of are shown in figure [ fig - ghlup ] .
the most obvious component of extended emission is an arc - like structure to the east of the star , visible at 8 , 24 , and 70 .
the structure is not centered on the star itself but roughly 20 to the south ( @xmath22 au at the star s distance ) with a radius of approximately 32 ( @xmath23 au ) .
spatially coincident with the arc is a point source visible at 5.8 and 8.0 only , which we can identify with the @xmath24 object .
because the region is crowded , with copious diffuse emission , it is unclear whether the arc near gh lup is related to the cepheid , the nearby star , or neither .
the stellar crowding also makes defining regions for surface photometry of the diffuse emission difficult ; we chose the regions marked in figure [ fig - ghlup ] to avoid stars in the 8 image and capture at least some of the extended emission .
the resulting measurements are given in table [ tab : surfphot ] , with the colors of the extended emission plotted in figure [ fig : sb_ratios ] .
using a modified blackbody spectrum to assign color temperatures from the @xmath25 ratios yields @xmath26 k , depending on the emissivity index @xmath27 for the modified blackbody .
the @xmath28 ratios for gh lup tend to be higher than those found for the other stars with extended emission .
together with the structure of the emission and the extensive nearby cirrus , this leads us to speculate that the emission near gh lup is only moderately likely to be associated with the star itself ( see [ sec : disc ] ) .
images of are shown in figure [ fig - l_car ] .
these data were also analyzed by @xcite , who failed to detect any nebulosity in the _ spitzer _ images , but found evidence of small scale ( @xmath29 ) extended emission in near- and mid - ir interferometric data , attributed to hot circumstellar dust .
our psf - subtracted irac images show low surface brightness emission , at 5.8 and 8.0 , surrounding the star with a radius of @xmath30 , or @xmath31 au at a distance of 498 pc @xcite .
we do not detect any significant extended emission at 70 , and only a tentative excess at 24 , northeast of the nearly saturated core of the psf - subtracted star .
the 8.0 emission appears brighter northwest of the star , where an arc - like structure is detected at a distance of @xmath32 ( @xmath33 au ) . following a suggestion by the referee , we computed the components of the star s galactic peculiar space motion and their projection onto the sky , using the procedure described by @xcite .
the star s position and proper motion were taken from the hipparcos catalog @xcite , the distance from @xcite , and heliocentric radial velocity of 3.6 km s@xmath5 from @xcite .
we find a space velocity of 24.9 km s@xmath5 along a position angle of 315 ( east of north ) , with the direction shown as an arrow in the second panel of figure [ fig - l_car ] .
the vector is roughly aligned along the direction from the star to the arc - like structure , suggesting a possible connection between the two ( e.g. , a bow shock , reflection nebula , or light echo ) .
image artifacts from the psf subtraction and the star make the connection difficult to secure , however .
artifacts also make it impossible to accurately measure the surface brightness of the extended emission , and thus estimate the mass loss rates that could be responsible for the observed nebulosity . higher spatial - resolution imaging covering a wider area
would clearly be useful in further study of @xmath1 car .
is embedded in a well - known reflection nebula @xcite whose infrared emission has been noted by numerous authors ( e.g. , * ? ? ?
* ; * ? ? ?
psf - subtracted images of rs pup are shown in figure [ fig - rs_pup ] .
the morphology of the extended infrared emission varies with wavelength but the brightest region is to the northeast of the star .
most of the emission is within about 80 of the star , a linear distance of 0.77 pc at the 1.992 kpc distance of rs pup @xcite , although the small field of view of the 70 image makes this difficult to pin down .
the mips data , but not the full - frame irac data , were also analyzed by @xcite , so here we present the first image of the resolved 8.0 emission around rs pup .
@xcite attributed the @xmath34 resolved emission , as well as the infrared excess in the iras data , to an extended envelope with a radius of roughly 1 pc and a temperature of @xmath35 k. @xcite also detected resolved emission at much smaller angular distances to the star , and attributed this to a `` warm and compact component '' at a distance of 3003000 au and a temperature of a few hundred k. they proposed that the warm component comes from rs pup s mass loss , and the cold component is interstellar medium gas compressed by the star s stellar wind . in @xcite a similar origin for the extended infrared emission around @xmath0 cep
was proposed .
we add to the analysis of rs pup s extended emission by examining the colors of the extended emission . using the surface brightness measurements in table [ tab : surfphot ] , figure [ fig : sb_ratios ] plots the ratios of the extended emission around rs pup and s mus , with the values derived for @xmath0 cep by @xcite for comparison . using a modified blackbody spectrum to assign color temperatures from the @xmath25 ratios yields @xmath36 k , depending on the emissivity index @xmath27 for the modified blackbody .
these temperatures are a little lower than the values derived for @xmath0 cep ; however , compared to @xmath0 cep , the rs pup nebulosity is more distant from the central star .
like the @xmath0 cep nebulosity , the rs pup extended emission has @xmath37 , suggesting that some polycyclic aromatic hydrocarbons ( pahs ) must be present ( pure dust at @xmath36 k would imply a much lower ratio ; * ? ? ?
we can also compute an average dust temperature for the rs pup nebulosity using the total flux densities of the extended emission . at 70
we used the largest aperture that will fit on the image ( 85 radius ) to derive a total flux density of 9.7 jy . at 24
we measured the flux density in a large aperture ( 72 radius ) and subtracted the psf - fitting flux of the central source to derive an extended emission flux density of ( @xmath38 ) 290 mjy .
using a modified blackbody spectrum to assign color temperatures from the ratio of the two bands yields @xmath39 k , depending on @xmath27 .
these temperatures are a little higher than the estimate of 40 k given by @xcite , who extrapolated the 70 surface brightness profile to derive a larger total flux density .
the dust mass in the rs pup nebula is best estimated with far - infrared photometry , since as figure [ fig - rs_pup ] shows , the stellar emission is insignificant at wavelengths of 70 and beyond . because our 70 mips images have a rather small field of view , the difficulty of accurate background subtraction makes any attempt to estimate the total flux density ( including that of * ? ? ?
* ) problematic . however , a rough estimate of total dust mass is possible and provides a comparison point for previous results .
we use the formula given in eq . 2 of @xcite , assuming that all of the 70 emission is from optically thin dust with @xmath40 k and emissivity @xmath41 in @xmath42 g@xmath5 .
for comparison with @xcite , we use @xmath43 , scaled from the 100 value the @xmath44 ` dirty silicate ' dust model in @xcite via a @xmath45 wavelength dependence .
the derived total dust masses for rs pup are in the range @xmath46 m@xmath21 .
( the range of values here accounts for the change in the planck function , but not any temperature dependence of the dust emissivity , over the given temperature range . )
the uncertainty due to temperature is compounded by the uncertainty in dust emissivity : for example , using the wavelength - dependent emissivity of amorphous fayalite given by @xcite would result in a value of @xmath47 nearly 15 times larger ( and thus masses 15 times smaller ) than given above .
the 100 flux density from the iras small scale structure catalog @xcite should provide a better estimate of the total flux density of the rs pup nebula . using this measurement , @xcite estimated a total dust mass of 0.028 m@xmath21 ( for their adopted distance of 1.778 kpc ) , or 0.035 m@xmath21 for the 1.992 kpc distance we use .
the mass derived from the 70 flux density is the same order of magnitude , likely smaller because our observations do not capture the full extent of the rs pup nebula . as @xcite point out , the large size of the rs pup nebula makes estimating a mass - loss rate from the nebular mass `` extremely difficult '' . with a gas - to - dust ratio of 100 , the @xcite dust mass converts to a total mass of 3 m@xmath21 . an argument in favor of the idea that some of the nebula is swept - up interstellar medium ( ism ) results from a simple comparison of the nebula mass with the helium - burning timescale of rs pup @xcite . taking the mass estimate for rs pup provided by @xcite , the helium burning time inside the instability strip is of the order of 1.2 myr .
if all of the nebula consists of mass lost by rs pup during its helium - burning phase , the implied mass - loss rate would be @xmath48 m@xmath21 yr@xmath5 .
this would imply a total mass loss at least of the order of 10% . as @xcite note
, this would be sufficient to solve the pulsational / evolutionary mass difference for rs pup .
however , as noted by @xcite , the observational mass loss rate for rs pup disagrees with their theoretical predictions ( @xmath49 m@xmath21 yr@xmath5 ) by several orders of magnitude .
these predictions do appear to fit the observational data for magellanic cloud cepheids @xcite , which makes it more believable that at least some of the mass in the rs pup nebula did not originally belong to the star .
the observed @xmath50 ratio is too high for pure dust emission , but lower than typical for the typical pah content of the ism .
since cepheid atmospheres are oxygen - rich and not expected to form pahs , this suggests that both swept - up ism ( or pre - existing material containing pahs ) and stellar wind must comprise the nebula .
images of are shown in figure [ fig - smus ] .
there are slightly extended emission features to both the north and south of the star visible at both 8 and 24 , with radial extents 3545 ( @xmath51 au at a distance of 820 pc ) .
there is also extended emission visible at 70 , to a radius of about 55 ( @xmath52 au ) , although the morphology is different at that wavelength .
the surface brightness of this emission is faint compared to that around rs pup , and the orientation of the psf subtraction residuals makes computing @xmath50 and @xmath25 at the same location difficult .
the s mus point shown in figure [ fig : sb_ratios ] is for the box marked ` 3 ' in the irac images and ` 2 ' in the mips images .
although uncertain , the colors imply an even lower pah content that in rs pup and @xmath0 cep , and a somewhat higher color temperature ( 70120 k ) .
this star has a high - temperature companion @xcite and the extended emission could be associated with that star rather than s mus itself , or the companion could be responsible for heating the circumstellar material ( as hypothesized for @xmath0 cep ; * ? ? ?
the photometry of the extended emission around s mus can be used to estimate the mass lost by the star .
large - aperture ( 35 radius ) photometry of the 70 emission yields a total flux density of @xmath53 mjy ; subtracting the point - source flux density ( measured in a 16 radius aperture as for paper i ) of @xmath54 mjy yields a nebular flux density of @xmath55 mjy .
we use the same dust mass estimation procedure as for rs pup , a temperatures of 70120 k , and a stellar distance of 0.88 kpc to derive an implied dust mass @xmath56 m@xmath21 . with a gas - to - dust ratio of 100
, the total implied mass would be approximately @xmath57 m@xmath21 .
if we instead use @xmath58 as in @xcite , the total mass agrees quite well with that derived @xmath0 cep .
given the photometric uncertainties , the agreement may well be coincidental ; however the order of magnitude comparison is intriguing .
a mass loss timescale can be estimated by dividing the distance to the emission ( 35 is 0.15 pc at the distance of s mus ) by the escape velocity ( 100 km s@xmath5 ) to yield @xmath59 yr , so the resulting mass loss rate would be @xmath60 m@xmath21 yr@xmath5 . as for @xmath0 cep ,
this estimate is consistent with the predictions of @xcite but well above those of @xcite .
there are two reasons why our value for the mass loss rate is likely to be an upper limit , and therefore larger than the @xcite predictions .
( 1 ) we assumed that all of the dust responsible for the extended emission comes from wind material as opposed to having been swept - up from the ism ; as for rs pup , the pah content implies that ism sweeping is likely relevant .
( 2 ) using the escape velocity to estimate the nebula s age results in the minimum possible age , and hence a maximum for the mass loss rate . we can estimate the total mass lost for s mus through its mass and crossing time for the instability strip , as above for rs pup .
we estimated pulsational and evolutionary masses through the relation provided by ( * ? ? ?
* tables 25 ) ; using the near - infrared photometry from ( * ? ? ?
* transformed to the 2mass system using @xcite ) and optical photometry from @xcite , corrected for extinction using the reddening given by @xcite and the reddening law of @xcite .
we used an intrinsic distance based on the near - infrared photometry and period - wesenheit relations provided by @xcite .
this procedure results in a mean mass @xmath61 m@xmath21 , and according to @xcite the corresponding crossing time of the instability strip is of the order of 0.14 myr .
given the mass loss rate above , this results in a small total mass loss of 0.006 m@xmath21 .
using the recent estimate of the radius of s mus by @xcite and the period - mass - radius relation given by @xcite , the pulsation mass is lower , @xmath62 m@xmath21 , and the instability strip crossing time much longer , of order 1 myr ; however the total mass lost would still be only 0.04 m@xmath21 , too small to significantly affect the mass discrepancy .
was the only cepheid found in paper i to have a statistically significant @xmath63- [ 70]$ ] color excess .
images of this star are shown in figure [ fig - sz_tau ] .
there is an extended object about @xmath64 clearly visible in the psf - subtracted 8.0 @xmath13 m image , about 7 to the north - east of the star . at the 512 pc distance to the star
, @xmath64 corresponds to linear scales of @xmath65 au .
the object is not clearly visible in the 5.8 @xmath13 m band , due to psf subtraction residuals , however it is visible at 24 and appears to be the source of most of the 70 luminosity .
the object is resolved at 24 and has a similar appearance in two images in that band taken about a year apart , suggesting that it is not an instrumental artifact or a solar system object .
@xcite detected a `` ridge or filament '' of similar shape but much larger angular size in iras imaging of betelgeuse ; they suggested that this structure is emission from the interstellar medium , unrelated to the star .
@xcite concluded that sz tau is a first - overtone pulsator and noted several other unusual characteristics of this star ( e.g. , period changes ) , but it is unclear how these features could be related to the presence of extended emission .
the elongated shape of the emission near sz tau suggests a background galaxy .
the infrared colors should yield further information on the object s nature , although photometry of the sz tau extended emission is complicated by the star itself . at 70
, the extended emission appears to be the dominant source of emission .
when we measure flux density in a 35 radius aperture centered on the 70 peak and a 3965 background annulus , the resulting flux density ( which may include some contribution from the star ) is 74 mjy .
we measured the flux density of the sz tau extended emission at 24 by doing a large - aperture ( 35 radius ) measurement on the extended emission+star and subtracting the psf - fitting star flux , to give an extended emission flux density of ( @xmath66 ) 31 mjy .
measuring the 8 @xmath67 flux density is the most difficult since there is significant contamination from the psf subtraction residuals .
measuring the emission on an image with the residuals masked in 35 radius aperture with 3543 background annulus gives a flux density of 10 mjy ; applying the irac extended source photometric correction changes this to 7.4 mjy .
different approaches to residual masking and photometric aperture change the resulting flux density by @xmath68% , so the value is highly uncertain .
this is hardly surprising since the extended emission is more than 100 times fainter than the star at 8 .
consistent with this result is the classification of the star s mid - infrared spectrum as normal , with no dust , based on data from the pht - s instrument on the infrared space observatory @xcite . we can ask whether the the galaxy - like appearance of the extended emission near sz tau is consistent with its mid - infrared flux densities and colors .
the object is rather bright for a galaxy : number counts at 24 @xcite show that there are only 3.54 galaxies per square degree brighter than about 15 mjy . taken together ,
all of our 24 cepheid images cover about 0.5 square degree , so we might expect 1 or 2 such galaxies in our survey .
figure [ fig : colors ] compares the colors of the sz tau extended emission with some comparison samples .
the color @xmath69 is consistent with the colors of both nearby and distant galaxies @xcite .
however , the flux ratio @xmath70 , while inconsistent with an unobscured stellar source , is rather high for a galaxy detected at 70 ; such galaxies typically have @xmath71 @xcite .
the flux ratios and colors are more consistent with those of candidate young stellar objects detected in the cepheus flare @xcite , although ysos do not typically have lenticular morphologies .
planetary nebulae can have asymmetric shapes , but the sz tau emission has quite different colors ( see figure [ fig : colors ] ) .
the best candidates to explain the extended emission near sz tau are either an unusually - shaped galactic object ( possibly a yso ) or a background galaxy with very weak cold dust emission .
this mysterious object warrants further investigation . a more definitive conclusion on its nature
could be reached with higher spatial resolution infrared observations , in which the flux from the star and extended emission were more clearly separated .
these stars show extended emission in the irac 8.0 band with only tentative detections in the mips bands ; images can be seen in figures [ fig - tmon ] and [ fig - xcyg ] . in the image of , there appears to be cirrus across the entire field of view .
however , there is brighter nebulosity closer to the star , specifically to the north and south in an hourglass - like shape extending to about 45 radius ( @xmath72 au , or 0.3 pc , at a distance of 1389 pc ) .
the emission is centered around the star , suggesting that it may be locally illuminated .
the extended emission near x cyg has the second - lowest significance ( after gh lup ) of the stars with emission above background levels .
it is brightest to the northwest of the star and concentrated within a radius of 30 ( @xmath73 au , or 0.16 pc , at a distance of 1163 kpc ) .
for both stars , the 24 images show hints of enhanced brightness at similar positions to the 8 emission , but these positions are too close to the psf subtraction residuals to make definitive conclusions , or to measure a surface brightness .
there is no corresponding detectable emission at 70 . with
both stars having extended emission observed at only one wavelength , the detections must be regarded as tentative .
observations with the james webb space telescope , particularly its coronographs , would provide an avenue to confirm or refute them .
if the detections are confirmed , the properties of these stars may provide a challenge in relating them to any mass loss .
@xcite predicted that mass loss rates , while showing large scatter with period , are generally larger for long - period stars .
t mon is a long period star with a companion @xcite , while x cyg has a period of 16.4 d and no known companion hotter than type a5 @xcite .
however , a low - mass companion with an orbit in the plane of the sky is not ruled out for x cyg and might be important in forming and/or shaping an extended nebula .
_ spitzer _ mid - infrared observations of cepheids have produced no strong evidence for large amounts of warm circumstellar dust @xcite .
here we use the same observations to place limits on cooler dust . of our sample of 29 cepheids , two stars ( rs pup and @xmath0 cep )
have well - detected extended infrared emission , two ( gh lup and s mus ) have possible extended emission at 24 and 70 , and a further three ( @xmath1 car , t mon and x cyg ) show tentative evidence for extended infrared emission at shorter infrared wavelengths only .
is there anything particularly special about the stars in which we do have ( tentative ) evidence for mass loss ?
our observations of the nebulosity around @xmath0 cep and s mus suggest that the detectability of a dustless wind might be enhanced when interstellar matter around the cepheid is swept up into a bow shock , and a hot companion heats the material .
compared to the emission around rs pup and gh lup , the @xmath25 ratio ( a proxy for dust temperature ) is higher near s mus and @xmath0 cep . both of these stars have known hot companions , while rs pup and gh lup do not ( although low - temperature companions can not be excluded using current data ) .
a counter example might be provided by t mon , which also has a hot companion , but does not have detectable extended emission at 24 or 70 or an infrared excess ; however , t mon s companion is cooler ( type a0 ) than that of the other two stars ( b3 and b7 - 8 ) .
the surface brightness ratios @xmath50 are a rather crude tool for deriving detailed physical conditions , but can give some insight into the pah content of the ism near these stars .
the low pah content indicated by the @xmath50 ratios for rs pup , s mus and @xmath0 cep much lower than in the ngc 7023 reflection nebula , or the ism in general suggests that the ism around these stars has been diluted to some extent by the stellar winds .
the higher pah content near gh lup would imply less dilution , suggesting that the emission there may be unrelated to the star .
we caution that , while the amount of dilution might be indicated by the @xmath50 ratio , high spatial - resolution spectroscopy in which the stellar and extended emission are separated would provide much greater physical insight into the conditions near these stars .
the irac - only extended emission near @xmath1 car , t mon , and x cyg also suggests that these stars are somehow unusual .
however , we have been unable to find any clear distinction between these stars and the three which also show extended mips emission , or between all seven stars and our full cepheid sample . like rs pup ,
@xmath1 car and t mon have long periods ( @xmath74 d ) and low temperatures ; gh lup , s mus and @xmath0 cep have periods @xmath75 d , and x cyg is intermediate with a period of 16.4 d. six of the seven stars have super - solar metallicity @xcite but the difference in mean metallicity between these stars and the other stars in the sample is barely statistically significant ( @xmath76 ) .
@xcite placed strong limits on the presence of hot companions in cepheids : of the stars in our sample , 3 of the 7 with possible extended emission have such companions ( s mus , @xmath0 cep , t mon ) , as do 5 of the 22 stars without extended emission ( an additional 3/23 are radial - velocity binaries ) .
with such small numbers it is not yet possible to say whether the presence of hot companions enhances the detection of extended infrared emission .
rs pup is clearly different from all of the other cepheids in our sample in terms of the extent and brightness of its infrared emission .
one possible explanation is that this star has swept up more material , either by having been on the instability strip for longer ( unlikely , since long - period stars evolve faster ) or by being located in a region of higher density in the interstellar medium .
an alternative explanation is suggested by visual examination of a schmidt plate taken by one of us ( dw ) in the region of rs pup .
the image reveals a number of near - identical dust clumps , with similar angular sizes , in the vicinity of the rs pup nebula . if these clumps are at the same physical distance and are physically related , but only one of them has a cepheid in it , it could be argued that they are all likely pre - existing nebulae that is , not the result of cepheid mass loss .
the counter - argument would be that the ring - like structures observed in the surrounding nebula appear to be centered on rs pup @xcite : this would be an extreme coincidence if the nebula were unrelated to the star . before our _ spitzer _ observations , only two cepheids had unambiguous evidence for circumstellar nebulae : @xmath1 car and rs pup .
we have now doubled that number with the addition of s mus and @xmath0 cep ; if extended emission around gh lup , t mon and x cyg is confirmed , this would increase the sample size even more .
a rough estimate of the fraction of cepheids with substantial dusty mass loss can be made as 4 to 7 of 29 , or @xmath77% .
the true fraction of cepheids with mass loss could be below this limit if the observed nebulosity is related to the gas and dust from which the star formed .
this is however unlikely , because even for small space velocities ( @xmath78 km s@xmath5 , an order of magnitude less than the observed velocity of @xmath0 cep and @xmath1 car ) , in 50 myr ( the time it takes for an intermediate mass star to become a cepheid ) these stars will have moved away from their birthplace by several tens of parsecs .
this is much larger than the size of the observed nebulosity around the stars in our sample . addressing the initial motivation for this study
, it appears that mass loss , at least its dusty component visible in the infrared , can only partially account for the cepheid mass loss discrepancy ( e.g. up to 1/5th of a 10% discrepancy in the case of @xmath0 cep , see * ? ? ?
our estimates of cepheid mass loss ( in the range @xmath3 to @xmath2 m@xmath21 yr@xmath5 ) are comparable to previous studies based on the detection of infrared and uv excess ( e.g. * ? ? ?
* ; * ? ? ?
* ) , but the higher spatial resolution of our images helps to separate local emission from background contamination .
the detection of shocked structures , as in the case of @xmath0 cep and possibly @xmath1 car , is a compelling indication of current mass loss in those stars .
while the total mass lost over the cepheid phase is small , the confirmation that these processes are active needs to be taken into account for a proper modeling of cepheid evolution .
the presence of significant cepheid winds can affect hydrodynamic models of cepheid pulsations , and the way shock waves propagate in the cepheid atmospheres .
finally , the observed level of circumstellar dust needs to be further explored , as it may have a significant effect on inferred cepheid distances as shown by @xcite .
the authors thank h. neilson for useful conversations and l.d .
matthews for assistance with computing the projected space velocity of @xmath1 car .
we thank the referee for helpful comments .
this work is based on observations made with the _
spitzer space telescope _ , which is operated by the jet propulsion laboratory , california institute of technology under nasa contract 1407 .
support for this work was provided by nasa through an award issued by jpl / caltech .
pb and dlw acknowledge research support through discovery grants from the natural sciences and engineering research council of canada ; pb also acknowledges support from an ontario early researcher award .
nre acknowledges support from the chandra x - ray center grant nas8 - 03060 .
dh was supported in part by the national science foundation reu and department of defense assure programs under grant no .
0754568 and by the smithsonian institution .
lllll gh lup 1 & @xmath79 & @xmath80 & @xmath81 & @xmath82 + gh lup 2 & @xmath83 & @xmath84 & @xmath85 & @xmath86 + gh lup 3 & @xmath87 & @xmath88 & @xmath89 & @xmath90 + gh lup 4 & @xmath91 & @xmath92 & @xmath93 & @xmath94 + rs pup 1 & @xmath95 & @xmath96 & @xmath97 & @xmath98 + rs pup 2 & @xmath99 & @xmath100 & @xmath101 & @xmath102 + rs pup 3 & @xmath80 & @xmath103 & @xmath104 & @xmath105 + rs pup 4 & @xmath106 & @xmath107 & @xmath108 & @xmath109 + rs pup 5 & @xmath110 & @xmath111 & @xmath112 & @xmath113 + s mus 1 & @xmath114 & @xmath91 & @xmath115 & @xmath116 + s mus 2 & @xmath117 & @xmath118 & @xmath119 & @xmath120 + s mus 3 & @xmath121 & @xmath122 & @xmath123 & @xmath124 + s mus 4 & @xmath79 & @xmath125 & & + m and 8.0 @xmath13 m irac bands , and the 24 @xmath13 m and 70 @xmath13 m ( not psf - subtracted ; the white x marks the position of the star ) mips bands .
the field of view is a 4 arcminute square .
numbered boxes are regions used for surface photometry ( see [ sec : mipsphot ] ) . , width=680 ]
cep ( circles ) and ngc 7023 ( star ) shown for comparison .
increasing values on the horizontal axis correspond roughly to increasing dust temperature , while increasing values on the vertical axis correspond to increasing pah content .
, width=453 ] car in the ( left to right ) 5.8 @xmath13 m and 8.0 @xmath13 m irac bands , and the 24 @xmath13 m and 70 @xmath13 m ( not psf - subtracted ) mips bands .
the field of view is a 2 arcminute square .
the white arrow in the second panel shows the direction of the star s proper motion relative to the local ism .
, width=680 ] m and 8.0 @xmath13 m irac bands , and the 24 @xmath13 m and 70 @xmath13 m mips bands ( not psf - subtracted ; the white x marks the position of the star ) .
the field of view is a 4 arcminute square .
numbered boxes are regions used for surface photometry ( see [ sec : mipsphot ] ) . , width=680 ] m and 8.0 @xmath13 m irac bands , and the 24 @xmath13 m and 70 @xmath13 m ( not psf - subtracted ) mips bands ; the white x marks the position of the star .
the field of view is a 3 arcminute square .
numbered boxes are regions used for surface photometry ( see [ sec : mipsphot ] ) . , width=680 ]
m and 8.0 @xmath13 m irac bands , and the 24 @xmath13 m and 70 @xmath13 m ( not psf - subtracted ) mips bands ; the white x marks the position of the star .
the field of view is a 2 arcminute square . of particular interest
is the elongated object to the northeast of the star . , | a deep and detailed examination of 29 classical cepheids with the _ spitzer _ space telescope has revealed three stars with strong nearby extended emission detected in multiple bands which appears to be physically associated with the stars .
rs pup was already known to possess extended infrared emission , while the extended emission around the other two stars ( s mus and @xmath0 cep ) is newly discovered in our observations .
four other stars ( gh lup , @xmath1 car , t mon and x cyg ) show tentative evidence for extended infrared emission
. an unusual elongated extended object next to sz tau appears to be a background or foreground object in a chance alignment with the cepheid .
the inferred mass loss rates upper limits for s mus and @xmath0 cep are in the range from @xmath2 to @xmath3 m@xmath4 yr@xmath5 , with the upper limit for rs pup as high as @xmath6 m@xmath4 yr@xmath5 .
mass loss during post - main - sequence evolution has been proposed as a resolution to the discrepancy between pulsational and dynamical masses of cepheid variable stars : dust in the lost material would make itself known by the presence of an infrared bright nebula , or unresolved infrared excess .
the observed frequency of infrared circumstellar emission ( @xmath7% ) and the mass loss rate we estimate for our sources shows that dusty mass loss can only account for part of the cepheid mass loss discrepancy .
nevertheless , our direct evidence that mass loss is active during the cepheid phase is an important confirmation that these processes need to be included in evolutionary and pulsation models of these stars , and should be taken into account in the calibration of the cepheid distance scale . |
modern x - ray observations show complex structures in both the spatial and spectral domains of various astrophysical sources . nonetheless , active galalactic nuclei ( agn ) including quasars nuclei remain spatially unresolved even with the highest - resolution x - ray telescopes .
most of their energy is released within the unresolved core , and only spectral and timing information is available to study the nature of the x - ray emission .
generally speaking , emission and absorption lines constitute an important part of the x - ray spectrum in that they can provide information as to the state of plasma .
one of the goals of x - ray data analysis is to understand the components present in the spectrum , and to obtain information about the emission and absorption features , as well as their locations and relation to the primary quasar emission .
the detection of weak lines in noisy spectra is the main statistical problem in such analyses : is a bump observed in the spectrum related to a real emission line or is it simply an artifact of the poissonian noise ?
although quasars x - ray spectra are usually featureless as expected based on the comptonization process ( see for example * ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , an important x - ray emission feature identified in agn and quasars spectra is the iron k emission line ( see recent review by * ? ? ?
determining the origin and the nature of this line is one of main issues in agn and quasar research .
this line is thought to come directly from illuminated accretion flow as a fluorescent process @xcite .
the location of the line in the spectrum indicates the ionization state of iron in the emitting plasma , while the width of the line tells us the velocity of the plasma @xcite .
the iron line provides a direct probe of the innermost regions of accretion flow and matter in close vicinity of a black hole .
absorption features associated with the outflowing matter ( warm wind , partial covering absorber ) have also been observed in recent x - ray observations @xcite . although the location and width of absorption lines provide information as to the velocity of the absorber and its distance from the quasar , this article focuses on statistical issues in fitting the spectral location of narrow emission lines , i.e. , identifying the ionization state .
there are two parts to the fe - k - alpha emission line observed in agn @xcite : one is a broad component thought to be a signature of a relativistic motion in the innermost regions of an accretion flow ; the other is a narrow component that is a result of a reflection off the material at larger distances from the central black hole .
a detection of the broad component is challenging as it requires a spectral coverage over a large energy range , so the continuum emission is well determined and the broad line can be separated @xcite .
the relativistic line profile is broad and skewed , and two strong peaks of the emission line that originates in a relativistic disk can be prominent and narrow . while the full profile of the broad line may not be easily separable from the continuum , these two peaks may provide a signature for this line in the x - ray spectrum .
the broad fe - line gives an important diagnostic of the gas motion and can be used to determine the spin of a black hole @xcite ; see also an alternative model for the red wing " component by @xcite .
the narrow component of the line gives diagnostics of the matter outside the accretion disk and conditions at larger distances from the black hole ; see fe - line baldwin effect discussion in @xcite .
both line components are variable and the line may `` disappear '' from the spectrum @xcite .
the spectral resolution of x - ray ccd detectors ( for example 100 - 200 ev in acis on _ chandra _ or epic on xmm-_newton _ ) is relatively low with respect to the predicted width of narrow ( @xmath0 km s@xmath1 ) emission or absorption lines in agn and quasars .
observations with grating instruments ( rgs or heg ) can provide high resolution x - ray spectra , but the effective area of the present x - ray telescopes is too low for efficient agn detections , and only a handful of bright low redshift sources have been observed with gratings to date @xcite
. therefore mainly the x - ray ccd spectra of lower resolution are used to study large samples of agn and quasars ( see for example * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . using these relatively low resolution x - ray detectors
, the fe - k - alpha emission line can be narrow enough to be contained entirely in a single detector bin . in some cases ( for example in _
chandra _ ) the line may occupy a few bins . in this article
we focus on the statistical problem of fitting the spectral location of an emission line or a set of emission lines that are narrow .
this is a common objective in high - energy analyses , but as we shall discuss fitting these relatively narrow features poses significant statistical challenges .
in particular we find evidence that using line profiles that are narrower than we actually expect the emission line to be can improve the statistical properties of the fitted emission line location .
x - ray spectra , such as those available with the _ chandra x - ray observatory _ carry much information as to the quasar s physics . taking advantage of the spectral capacity of such instruments , however , requires careful statistical analysis .
for example , the resolution of such instruments corresponds to a fine discretization of the energy spectrum . as a result
, we expect a low number of counts in each bin of the x - ray spectrum .
such low - count data make the gaussian assumptions that are inherent in traditional minimum @xmath2 fitting inappropriate .
a better strategy , which we employ , explicitly models photon arrivals as an inhomogeneous poisson process @xcite . in addition
, data are subject to a number of processes that significantly degrade the source counts , e.g. , the absorption , non - constant effective area , blurring of photons energy , background contamination , and photon pile - up .
thus , we employ statistical models that directly account for these aspects of data collection .
in particular , we design a highly structured multilevel spectral model with components for both the data collection processes and the complex spectral structures of the sources themselves . in this highly structured spectral model
, a bayesian perspective renders straightforward methods that can handle the complexity of _ chandra _ data @xcite .
as we shall illustrate , these methods allow us to use low - count data , to search for the location of a narrow spectral line , to investigate its location s uncertainty , and to construct statistical tests that measure the evidence in the data for including the spectral line in the source model .
the energy spectrum can be separated into two basic parts : a set of continuum terms and a set of several emission lines .
we begin with a standard spectral model that accounts for a single continuum term along with several spectral lines . throughout this paper
, we use @xmath3 as a general representation of model parameters in the spectral model .
the components of @xmath4 represent the collection of parameters for the continuum , ( emission ) lines , absorption , and background contamination , respectively .
( notice that the roman letters in the superscripts serve as a mnemonic for these four processes . ) because the x - ray emission is measured by counting the arriving photons , we model the expected poisson counts in energy bin @xmath5 , where @xmath6 is the set of energy bins , as @xmath7 where @xmath8 and @xmath9 are the width and mean energy of bin @xmath10 , @xmath11 is the expected counts per unit energy due to the continuum term at energy @xmath9 , @xmath12 is the set of free parameters in the continuum model , @xmath13 is the number of emission lines , @xmath14 is the expected counts due to the emission line @xmath15 , and @xmath16 is the proportion of an emission line centered at energy @xmath17 and with width @xmath18 that falls into bin @xmath10 .
there are a number of smooth parametric forms to describe the continuum in some bounded energy range ; in this article we parameterize the continuum term @xmath19 as a power law , i.e. , @xmath20 where @xmath21 and @xmath22 represent the normalization and photon index , respectively .
the emission lines can be modeled via the proportions @xmath16 using narrow gaussian distributions , lorentzian distributions , or delta functions ; the counts due to the emission line are distributed among the bins according to these proportions .
while the gaussian or lorentzian function parameterizes an emission line in terms of center and width , the center is the only free parameter with a delta function ; the width of the delta function is effectively the width of the energy bin in which it resides . while the model in equation [ park
: eq : ideal ] is of primary scientific interest , a more complex statistical model is needed to address the data collection processes mentioned in [ park : sec : hea ] .
we use the term _ statistical model _ to refer to the model that combines the _ source _ or _
astrophysical model _ with a model for the stochastic processes involved in data collection and recording .
thus , in addition to the source model , the statistical model describes such processes as instrument response and background contamination . specifically , to account for the data collection processes , equation [ park : eq : ideal ] is modified via @xmath23 where @xmath24 is the expected observed poisson counts in detector channel @xmath25 , @xmath26 is the set of detector channels , @xmath27 is the probability that a photon that arrives with energy corresponding to bin @xmath10 is recorded in detector channel @xmath28 ( i.e. , @xmath29 is the so - called redistribution matrix or rmf commonly used in x - ray analysis ) , @xmath30 is the effective area ( i.e. , arf , a calibration file associated with the x - ray observation ) of bin @xmath10 , @xmath31 is the probability that a photon with energy @xmath9 is _ not _ absorbed , @xmath32 is the collection of parameters for absorption , and @xmath33 is a poisson intensity of the background counts in channel @xmath28 . while the scatter probability @xmath27 and the effective area @xmath30 are presumed known from calibration , the absorption probability is parameterized using a smooth function ; see @xcite for details . to quantify background contamination ,
a second data set is collected that is assumed to consist only of background counts ; the background photon arrivals are also modeled as an inhomogeneous poisson process .
unfortunately , the statistical methods and algorithms developed in @xcite can not be directly applied to fitting _
narrow _ emission lines .
there are three obstacles that must be overcome in order to extend bayesian highly structured models to spectra containing narrow lines .
in particular , we must develop ( 1 ) new computational algorithms , ( 2 ) statistical summaries and methods for inference under highly multimodal posterior distributions , and ( 3 ) statistical tests that allow us to quantify the statistical support in the data for including an emission line or lines in the model .
our main objective in this paper is to extend the methods of @xcite in these three directions , and to evaluate and illustrate our proposals .
here we discuss each of these challenges in detail .
[ [ challenge-1-statistical - computation . ] ] challenge 1 : statistical computation .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + fitting the location of narrow lines requires new and more sophisticated computational techniques than those developed by @xcite .
indeed , the algorithms that we develop require a new theoretical framework for statistical computation : they are not examples of any existing algorithm with known properties .
although the details of this generalization are well beyond the scope of this article , we can offer a heuristic description ; a more detailed description is given in appendix [ ap : alg ] .
readers who are interested in the necessary theoretical development of the statistical computation techniques are directed to @xcite and @xcite . the algorithms used by @xcite to fit the structured bayesian model described in
[ park : sec : model ] are based on the probabilistic properties of the statistical models .
for example , the parameters of a gaussian line profile can be fit by iteratively attributing a subset of the observed photons to the line profile and using the mean and variance of these photon energies to update the center and width of the line profile .
the updated parameters of the line profile are used to again attribute a subset of the photons to the line , i.e. , to stochastically select a subset of the photons that are likely to have arisen out of the physical processes at the source corresponding to the emission line .
these algorithms are typically very stable . for example , they only return statistically meaningful parameters because the algorithms themselves mimic the probabilistic characteristics of the statistical model .
the family of expectation / maximization ( em ) algorithms @xcite and markov chain monte carlo ( mcmc ) methods such as the gibbs sampler @xcite are the examples of statistical algorithms of this sort .
a drawback of these algorithms is that in some situations they can be slow to converge . when fitting the location of a gaussian emission line , for example , the location is updated more slowly if the line profile is narrower
this is because only photons with energies very close to the current value of the line location can be attributed to the line .
updating the line location with the mean of the energies of these photons can not result in a large change in the emission line location .
the situation becomes chronic when a delta function is used to model the line profile : the line location parameter sticks at its starting value throughout the iteration .
it is to circumvent this difficulty that we develop both new em - type algorithms @xcite and new mcmc samplers specially tailored for fitting narrow lines .
our new samplers are motivated by the gibbs sampler , but constitute a non - trivial generalization of gibbs sampling known as partially collapsed gibbs sampling @xcite ; see appendix [ ap : alg ] .
our updated versions of both classes of algorithms are able to fit narrow lines by avoiding the attribution of photons to the emission line during the iteration .
such algorithms tend to require fewer iterations to converge regardless of the width of the emission line .
because they involve additional evaluation of quantities evolving the large dimensional redistribution matrix , @xmath34 , however , each iteration of these algorithms can be significantly more costly in terms of computing time .
a full investigation of the relative merit of the algorithms and a description of how the computational trade - offs can be played to derive optimal algorithms are beyond the scope of this paper . except in appendix
[ ap : alg ] , we do not discuss the details of the algorithms further in this article ; interested readers are directed to @xcite and @xcite . [
[ challenge-2-multimodal - likelihoods . ] ] challenge 2 : multimodal likelihoods .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the likelihood function for the emission line location(s ) is highly multimodal .
each mode corresponds to a different relatively likely location for an emission line or a set of emission lines .
standard statistical techniques such as computing the estimates of the line locations with their associated error bars or confidence intervals implicitly assume that the likelihood function is unimodal and bell shaped .
because this assumption is clearly and dramatically violated , these standard summary statistics are unreliable and inadequate . unfortunately , there are no readily available and generally applicable simple statistical summaries to handle highly multimodal likelihoods
. instead we must develop summaries that are tailored to the specific scientific goals in a given analysis . because general strategies for dealing with multimodal likelihood functions are little known to astronomers and specific strategies for dealing with multimodal likelihood functions for
the location of narrow spectral lines do not exist , one of the primary goals of this article is to develop and illustrate these methods .
a fully bayesian analysis of our spectral model with narrow emission lines is computationally demanding , even with our new algorithms .
thus , we develop techniques that are much quicker and give similar results for the location of emission lines .
these methods based on the so - called _ profile posterior distribution _ do not stand on as firm of a theoretical footing as a fully bayesian analysis , but are much quicker and thus better suited for _ exploratory data analysis_. the profile posterior distribution along with our exploratory methods are fully described and compared with the more sophisticated bayesian analysis . [ [ challenge-3-testing - for - the - presence - of - narrow - lines . ] ] challenge 3 : testing for the presence of narrow lines .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + in addition to fitting the location of one or more emission lines , we often would like to perform a formal test for the inclusion of the emission lines in the statistical model .
that is , we would like to quantify the evidence in a potentially sparse data set for a particular emission line in the source .
testing for a spectral line is an example of a notoriously difficult statistical problem in which the standard theory does not apply .
there are two basic technical problems .
first , the simpler model that does not include a particular emission line is on the boundary of the larger model that does include the line .
that is , the intensity parameter of an emission line is zero under the simpler model and can not be negative under the larger model .
an even more fundamental problem occurs if either the line location or width is fit , because these parameters have _ no value _ under the simpler model .
the behavior ( i.e. , sampling distribution ) of the likelihood ratio test statistic under the simpler model is not well understood and can not be assumed to follow the standard @xmath2 distribution , even asymptotically .
@xcite propose a monte - carlo - based solution to this problem based on the method of posterior predictive p - values @xcite . in this article
we extend the application of protassov _ et al_. s solution to the case when we fit the location of a narrow emission line , a situation that was avoided in @xcite .
the remainder of the article is organized into four sections .
[ park : sec : model - based ] reviews bayesian inference and monte carlo methods with an emphasis on multimodal distributions , outlines our computation methods , proposes new summaries of multimodal distributions , and describes exploratory statistical methods in this setting .
we introduce illustrative examples in [ park : sec : model - based ] , but detailed spectral analysis is postponed in order to allow us to focus on our proposed methods . in [ park : sec : simul ] , a simulation study is performed to investigate the statistical properties of our proposed methods , with some emphasis placed on the potential benefits of model misspecification .
[ park : sec : quasar ] presents the analysis of the high redshift quasar pg1634 + 706 , and how to test for the inclusion of the line in the spectral model .
concluding remarks appear in [ park : sec : conclusion ] .
an appendix outlines the computational methods we developed specifically for fitting the location of narrow emission lines .
using a poisson model for the photon counts , the likelihood function of the parameter in the spectral model described in
[ park : sec : model ] is given by @xmath35 $ ] where @xmath36 denotes the observed photon counts .
with likelihood - based methods , the parameter value that maximizes the probability of the observed data is generally chosen as an estimate of @xmath3 ; this estimate is called the maximum likelihood estimate ( mle ) . in bayesian methods , prior knowledge for @xmath3 can be combined with the information in the observed data .
prior distribution _ can be used to quantify information from other sources or to impose structure on a set of parameters .
the prior distribution is combined with the likelihood to form a _
posterior distribution_. the prior distribution is denoted by @xmath37 and the posterior distribution by @xmath38 .
bayesian inferences for @xmath3 are based on the posterior distribution .
using bayes theorem , the prior distribution and the likelihood function are combined to form the posterior distribution via @xmath39 where the last proportionality holds because @xmath40 does not depend on @xmath3 and , given the observed data , is considered a constant .
bayesian statistical inferences are made in terms of probability statements , which are quantified using various numerical summaries of the posterior distribution . to illustrate this
, we consider a stylized right - skewed distribution ; see figure [ park : fig : skewed ] .
this distribution is similar to the posterior distribution of the expected counts due to the emission line @xmath15 , @xmath14 , because this parameter is necessarily non - negative .
although from a bayesian perspective the posterior distribution of a model parameter is the complete summary of statistical inference for that parameter , it is often useful to summarize the posterior distribution using point estimates or intervals .
commonly used bayesian point estimates of a parameter are the mean , median , and mode(s ) of the posterior distribution .
error bars for the point estimates can be computed based on the variation of the posterior distribution .
the equal - tail interval and the highest posterior density ( hpd ) interval are both commonly used summaries of uncertainty .
for example , a 68% equal - tail posterior interval is the central interval of the posterior distribution and corresponds to the range of values of the parameter above and below which lies exactly 16% of the posterior probability .
a 68% hpd interval , on the other hand , is the interval of values that contains 68% of posterior probability and within which the density is never lower than that outside the interval .
the 68% hpd interval is the shortest possible interval that accounts for 68% of the posterior probability .
this is illustrated in figure [ park : fig : skewed](a ) and ( b ) .
the equal - tail interval achieves the same probability as the hpd interval by excluding a more likely region and by including a less likely region .
when the posterior distribution is unimodal and symmetric , the equal - tail interval and the hpd interval are identical .
in addition to computing parameter estimates and their error bars , it is often important to check if model assumptions are supported by the data .
one way to do this is to generate simulated data under the model and compare the simulated data with the observed data ; refer to @xcite where this strategy is used to determine whether emission line profiles should be included in the spectral model for a gamma - ray burst .
if the simulated data vary systematically for the observed data , it is an indication that the model used to simulate the data may not be adequate . in a bayesian analysis , we might generate such simulated data from the _ posterior predictive distribution _ , denoted by @xmath41 , i.e. , @xmath42 where @xmath43 represents an unknown future observation and the last equation follows because @xmath43 and @xmath44 are conditionally independent given @xmath3 . in words , the posterior predictive distribution averages the likelihood function over the posterior distribution of @xmath3 .
data are simulated from the posterior predictive distribution and then used to make predictive inferences ; see [ park : sec : model - check ] for details .
posterior simulation plays a central part in applied bayesian analysis because of the usefulness of a simulation that can often be relatively easily generated from a posterior distribution . in performing simulations , given a large enough sample , a histogram of a monte carlo simulation can provide practically complete information about an actual posterior distribution .
figure [ park : fig : skewed](c ) shows that the histogram of the monte carlo simulation carries the same information as the posterior distribution itself .
thus , once a monte carlo simulation is obtained , it can be used to compute the mean , variance , percentiles , and other summaries of the posterior distribution . in particular , with a random simulation of size @xmath45 , the posterior mean can be approximated as @xmath46 where @xmath47 is a simulation from the posterior distribution , @xmath38 .
a 68% equal - tail posterior interval is computed by generating a monte carlo simulation of size @xmath45 from the posterior distribution , sorting the simulated values into increasing order , and choosing the @xmath48$]th and the @xmath49$]th values in the list . with the monte carlo simulation , a 68% hpd region
is computed by segmenting the range of possible parameter values into bins , approximating the posterior probability of each bin as the proportion of the simulated values in that bin , and computing a region by beginning with the bin with the largest posterior probability and adding additional bins in the order of their posterior probabilities until the resulting region contains at least 68% of the posterior probability .
our search for a narrow emission line begins by finding modes of its posterior distribution that correspond to plausible locations of an emission line . to find the modes ( and to compte the profile posterior distribution ) , we use an algorithm optimized for this problem ( i.e. , the rotation(9 ) em - type algorithm , see appendix [ ap : alg ] and @xcite for details ) . because the posterior distribution of the line location is highly multimodal ( see [ park : sec : mm ] ) , the algorithm is run using multiple starting values selected across the entire energy range of possible line locations , e.g. , 50 starting values for the line location equally spaced between 1.0 kev and 6.0 kev .
using multiple starting values enables us to identify the important local modes of the posterior distribution .
it is important to use enough starting values to ensure all of the important modes are identified .
this is a standard strategy , long advocated in texts on bayesian data analysis @xcite and is closely related to the computation of the profile posterior distribution described in
[ park : sec : advise ] .
the profile posterior distribution is computed by fixing the line location at each value of a fine grid , finding the posterior modes of the other model parameters , and plotting the resulting maximum posterior probability as a function of the line location .
this procedure corresponds to the projected delta - chi - square method in the chi - square setting , see @xcite and @xcite .
mode finding also begins with a fine grid of starting values , but we run the mode finder allowing all parameters including the line location to be fit .
after the modes are found , monte carlo simulation techniques optimized to this problem can be run to further investigate the uncertainty of the possible line locations .
we employ state - of - the - art mcmc samplers , i.e. , partially collapsed gibbs ( pcg ) i for the delta function emission line and pcg ii for the gaussian emission line to obtain the posterior distribution of line location ; see appendix [ ap : alg ] , @xcite , and @xcite for details . to ensure the convergence of a markov chain constructed by the mcmc samplers , we run multiple chains with overdispersed starting values ( e.g. , 1 kev , 3 kev , and 5 kev for the line location parameter ) and monitor the convergence qualitatively and quantitatively .
for example , we compute the estimate of the potential scale reduction @xcite , denoted by @xmath50 , for all parameters of interest . if @xmath50 is near 1 ( e.g. , below 1.2 ) for each of the parameters , we collect the second halves of the chains together and use these monte carlo draws for inference ; see @xcite for theoretical justification and discussion .
there are of course many strategies that one might employ to construct efficient monte carlo samplers .
methods based on annealing or tempering , or that use explicitly parallel methods are often useful for exploring multimodal posterior distributions . given the low autocorrelation of the simulated values produced by our method , we have not pursued these strategies .
fitting a narrow emission line often tends to yield a highly multimodal posterior distribution of the line location , as shown in figure [ park : fig : hpd ] .
thus , we are interested in the two types of intervals for the posterior distributions : an equal - tail posterior interval and an hpd region .
figure [ park : fig : hpd ] illustrates a 68% equal - tail posterior interval and a 68% hpd region for the line location when its posterior distribution is highly multimodal .
this is only an example of a highly multimodal posterior distribution .
we come to the details of our analysis of a line location in [ park : sec : simul ] .
the 68% equal - tail posterior interval is a central interval , so that it includes segments with nearly zero posterior probability here and there within the interval . because the posterior distribution of the line location in figure [ park : fig : hpd ] is multimodal , the 68% hpd region consists of the four shaded disjoint intervals ; notice that the height of each of the histogram bars outside the hpd region is less than that of those within the region .
in such a multimodal posterior distribution , the hpd region not only is shorter in length but also conveys more information as to likely locations of the line than does the equal - tail interval . in figure [ park :
fig : hpd ] , we use a coarse binning to illustrate the distinction between the equal - tail posterior interval and hpd region .
this results in a posterior region that is relatively imprecise ; in our spectral analysis , we use a finer binning to construct a more precise region from a monte carlo simulation . when the posterior distribution of the line location is plotted with the same fine binning as the _ chandra _ energy spectrum
, however , it may not be smooth due to monte carlo errors , as shown in the top panel of figure [ park : fig : hpd - illust ] .
an hpd region computed from the unsmoothed posterior distribution may result in a combination of too many posterior intervals . to avoid such fragmentation of hpd regions , we use gaussian kernel smoothing to smooth the posterior distribution before we compute hpd regions .
the middle panel of figure [ park : fig : hpd - illust ] presents the smoothed posterior distribution resulting from applying gaussian kernel smoothing with standard deviation equal to the bin size of the _ chandra _ energy spectrum , i.e. , 0.01 kev .
the smoothed posterior distribution smooths out lower posterior probabilities but does not flatten higher posterior probabilities too much .
we propose a new graphical summary to better describe the hpd regions of a ( smoothed ) multimodal posterior distribution .
an _ hpd graph _ is constructed by plotting a series of hpd regions against their corresponding hpd levels .
for example , for the data set used to compute the posterior distribution in figure [ park : fig : hpd - illust ] , we compute 100 hpd regions , one for each of 100 levels , 1% , 2% , @xmath51 , and 100% . each of these regions is a union of possible values of the line location .
we can plot the line location on the horizontal axis and the level on the vertical axis .
each of the 100 hpd regions can then be plotted as a union of horizonal line intervals at the appropriate level of the hpd region on the vertical axis .
the resulting hpd graph lets us visualize many hpd regions with varying levels , so that all the important modes of a multimodal distribution can be effectively summarized with their relative posterior probabilities . as an illustration
, we computed the three hpd regions with levels 50% , 68% , and 95% , and plotted them in the middle panel of figure [ park : fig : hpd - illust ] along with the smoothed posterior distribution .
the solid lines indicate the disjoint intervals that compose each hpd region , and the dotted lines the intervals outside the hpd region .
the bottom panel in figure [ park : fig : hpd - illust ] shows the hpd graph in grey with the three hpd intervals from the middle plot superimposed .
, where each marginal standard deviation corresponds to the size of an energy bin in the _ chandra _ spectrum .
based on the smoothed posterior distribution , we compute ten 2-d hpd regions with levels ranging from 10% to 100% .
these regions are plotted with different shades of grey in the 2-d hpd graph shown in the right panel.,width=576 ] when there are two parameters of interest , the 1-d hpd graph can be extended to a 2-d hpd graph .
that is , a joint posterior distribution is computed from a monte carlo simulation by using bivariate gaussian kernel smoothing and used to construct 2-d hpd regions with various levels .
these hpd regions can then be plotted with different shades of grey .
for example , figure [ park : fig : twod - hpd - illust ] shows the 2-d hpd graph computed for the joint posterior distribution of two possible line locations ; see [ park : sec : simul2 ] for further discussion on the scatter plot .
the left panel of figure [ park : fig : twod - hpd - illust ] is an unsmoothed joint posterior distribution .
after applying bivariate gaussian kernel smoothing , 10 hpd regions ( one for each of 10 levels , 10% , 20% , ... , and 100% ) are computed and plotted with different shades of grey from black ( 10% ) to white ( 100% ) , as shown in the right panel of figure [ park : fig : twod - hpd - illust ] .
hpd regions with lower levels contain pixels with higher posterior probabilities , so that darker pixels indicate more probable regions of the two line locations .
multimodal likelihoods and posterior distributions pose another challenge for statistical analysis .
the calibration of many standard statistical procedures is based on the asymptotic gaussian nature of the likelihood and posterior distribution .
thus , standard methods for computing error bars and confidence intervals rely on the likelihood being at least approximately gaussian .
when the likelihood function exhibits the multimodal features that we see in figures [ park : fig : hpd - illust ] and [ park : fig : twod - hpd - illust ] , these standard results simply do not apply .
one consequence of this is that the nominal level of an interval may not match its frequency coverage .
that is , if we were able to repeat our observation many times , we would find that percentage of intervals that contain the true line location might differ from the nominal level of the interval . in [ park : sec : simul ] , we find that the posterior distribution is highly multimodal when the true emission line is weak or there are several emission lines in a spectrum . in this case , we also find that by _ misspecifying _ the model we are able to improve the frequency properties of our proposed method .
in particular , we find that using a delta function line profile improves the properties of our procedure even when the true line has appreciable width . as discussed in [ park : sec : mm ] , the posterior distribution of a narrow emission line location tends to be highly multimodal . in this case , the profile posterior distribution can be used as a handy and quick - to - compute summary of the posterior distribution to explore the possible locations of a spectral line .
the profile posterior distribution is the posterior distribution of a parameter evaluated at the values of the other parameters that maximize the posterior distribution .
the profile posterior distribution is a bayesian analogue to the profile likelihood , a standard likelihood method for dealing with nuisance parameters ; see @xcite or @xcite for applications of the profile likelihood . in the context of parameter estimation , the distribution of the minimum @xmath2 statistic described by @xcite is closely related to the profile posterior distribution .
generally we do not advocate using the profile posterior distribution as a substitute for the marginal posterior distribution because interval or region estimates computed with the profile posterior distribution have rather unpredictable statistical properties .
the marginal posterior distribution is obtained by integrating out ( i.e. , averaging over ) the other parameters and computing the marginal posterior distribution requires sophisticated numerical integration methods such as mcmc especially when the dimension of the nuisance parameters is large . however
, the profile posterior distribution can be computed without mcmc and , as an analogue to the marginal posterior distribution , can be used to roughly examine the posterior distribution of a model parameter .
thus , we believe the profile posterior distribution is well suited for the initial exploration of the data because it gives a clear and reliable set of potential locations for spectral lines . the profile posterior distribution can be computed on a fine grid of the possible values of the line location by running an optimizer to maximize over the other model parameters for each value of the line location on the grid ; we recommend using stable optimizers such as em - type algorithms @xcite .
the profile posterior distribution of the line location computed in this way is computationally less demanding and cheaper in terms of cpu time than the marginal posterior distribution produced by monte carlo methods such as mcmc .
our simulation study is conducted to assess the validity of the highly structured multilevel spectral model discussed in [ park : sec : model ] , to illustrate the connection between the multimodal posterior distribution of a single narrow emission line and evidence for multiple lines , to illustrate the possible advantage of the misspecification of an emission line width , and to illustrate the relationship among gaussian line parameters .
we consider the following six cases that we believe are representative of the cases that are of general interest : case 1 : : : there is no emission line in the spectrum .
case 2 : : : there is a narrow gaussian emission line at 2.85 kev with @xmath52 kev and equivalent width$ ] .
] , @xmath53 kev in the spectrum .
case 3 : : : there is a moderate gaussian emission line at 2.85 kev with @xmath54 kev and @xmath55 kev in the spectrum . case 4 : : : there is a narrow gaussian emission line at 2.85 kev with @xmath56 kev and @xmath57 kev in the spectrum .
case 5 : : : there are two narrow gaussian emission lines with @xmath56 kev , one at 1.20 kev with @xmath58 kev and the other at 2.85 kev with @xmath59 kev in the spectrum .
case 6 : : : there are two narrow gaussian emission lines with @xmath56 kev , one at 1.20 kev with @xmath58 kev and the other at 2.85 kev with @xmath57 kev in the spectrum .
that is , each spectrum has either 0 , 1 , or 2 lines .
data are recorded in an energy grid with bin width 0.01 kev .
thus , the narrow emission line ( i.e. , cases 2 , 4 , 5 , and 6 ) corresponds to about 17 energy bins and the moderate one ( i.e. , case 3 ) about 85 energy bins , using @xmath60 standard deviations .
as compared to the _ chandra _ resolution , the delta function emission line profile corresponds to 1 energy bin and thus it does not correctly specify the width of the gaussian emission lines in this simulation . through the simulation study , however , we illustrate possible advantage of this model misspecification in producing valid and efficient estimates and associated uncertainties for the line location .
we show that using delta functions emission lines in the model is a useful strategy even when the true line occupies multiple bins . for each of the six spectra ,
we generate twenty test data sets ( 120 data sets in total ) each with about 1500 counts similar to the observed number of counts in the _ chandra _ x - ray spectrum of pg1634 + 706 analyzed in [ park : sec : quasar ] , mimicking the real data situation .
each spectrum has a power law continuum with @xmath61@xmath625 and @xmath63 .
our simulation is done with sherpa software @xcite in ciao , assuming the _ chandra _ responses ( effective area and instrument response function ) and no background contamination . .
the vertical dashed lines represent the locations of the true emission lines.,width=576 ] for each of the test data sets , we run state - of - the - art mcmc samplers to fit a spectral model with a single delta function emission line .
based on the monte carlo draws collected from the multiple chains of the mcmc samplers , the top two rows of figure [ park : fig : simul ] present the marginal posterior distribution of the delta function line location for one simulation under each of the six cases ; the vertical dashed lines represent the true line locations .
the marginal posterior density is smoothed using gaussian kernel smoothing with standard deviation 0.01 kev , as described in
[ park : sec : mm ] . when there is no emission line in the spectrum ( i.e. , case 1 ) , the posterior distribution of the delta function line location is highly multimodal .
in the case of a weak narrow gaussian emission ( i.e. , case 2 ) , the marginal posterior distribution often remains highly multimodal , but one mode typically identifies the true line location . in practice , the local mode(s ) of such a highly multimodal posterior distribution may suggest plausible line locations and show evidence for multiple lines ; see [ park : sec : simul2 ] .
even with a moderate line ( i.e. , case 3 ) , the true line location appears well estimated with the marginal posterior distribution of the delta function line location .
as we shall see in table [ park : tbl : simulhpd ] , however , in this case the resulting posterior region under - covers the true values because the true line is 85 times wider than the specified model . with the strong narrow gaussian line ( i.e. , case 4 ) , the posterior distribution of the line location tends to be unimodal , and the posterior mode correctly identifies the true line location . the posterior distribution for case 5 in figure [ park : fig : simul ] is bimodal , with the modes corresponding to the two true line locations .
when multiple lines are present in a spectrum , the posterior distribution of the single line location can be multimodal , as shown in the case 5 of figure [ park : fig : simul ] .
thus the multiple modes may be indicative of multiple lines ; see [ park : sec : simul2 ] for details .
when one of two narrow gaussian emission lines is much stronger ( i.e. , case 6 ) , the single delta function line model tends to identify only one of the two true line locations . to visualize the uncertainty of the fitted delta function line location(s ) ,
the bottom two rows of figure [ park : fig : simul ] show the hpd graphs constructed with 100 hpd regions as described in
[ park : sec : mm ] .
the multimodality in the marginal posterior distribution of a single line location may indicate the existence of multiple lines in a spectrum , provided the lines are well separated .
when a model is fitted with one emission line , modes in the likelihood function of the line location correspond to ranges of energy with excess emission relative to the continuum .
multiple modes in the likelihood indicate that there are multiple ranges of energy with such excess emission .
the height of the mode is indicative of the degree of excess .
thus , if there are several emission lines , we might expect to see several corresponding modes in the likelihood . if there is one energy range that dominates in terms of excess emission , however , it corresponds to the dominate mode of the likelihood .
thus , if there are lines of very different intensities , only the strongest ones may show up as a mode of the likelihood .
this can be seen by comparing case 5 and case 6 in figure [ park : fig : simul ] .
if there is evidence for multiple lines in a spectrum or if we suspect multiple lines a priori , we can fit a model with two or more lines .
we illustrate this using simulated data under case 2 ( one narrow line ) and case 5 ( two narrow lines ) . beginning with case 2 ,
the actual spectrum has only one line , but we investigate what happens when we fit two lines to this data .
a scatterplot of the two fitted line locations identified when fitting two emission lines to one of the data sets generated under case 2 is presented in the top left panel of figure [ park : fig : simul - twolines ] .
there is a label switching problem between the two fitted line locations because of the symmetry of the emission lines in the model .
we can remove the symmetry by imposing a constraint on line locations . to do this
, we first fit the model with a single delta function line profile and compute the posterior mode of its line location . returning to the model with two fitted delta functions , we separate the two fitted line locations by setting the first " line location to be the one closest to the posterior mode . in case 2 ,
the posterior mode for the single line location is 2.815 kev , so that the first line location is the line location closest to 2.815 kev and the second line is the other location .
the two panels in the top right corner of figure [ park : fig : simul - twolines ] show the resulting marginal posterior distributions of the two fitted line locations .
as shown in the figure , the marginal posterior distribution of the first line location correctly identifies the true line location .
the spectrum used to generate the data under case 2 has no second line , so that the marginal posterior distribution of the second line is highly multimodal . in practice
, we may take the local mode(s ) in the second marginal posterior distribution as candidates for another line location .
however , the resulting hpd regions for the second line are wide , indicating that either there is no second line or if there is , it can not be well identified .
when two emission lines are present in a spectrum , we follow the same procedure as illustrated using the data generated under case 5 .
the bottom left panel of figure [ park : fig : simul - twolines ] shows the scatterplot of two line locations identified in the spectrum of case 5 .
label - switching is handled as above using the posterior mode for the single line location , which is computed as 2.855 kev .
the fitted marginal posterior distributions of the first and second line locations are given in the bottom right corner of figure [ park : fig : simul - twolines ] .
when there are two emission lines in a spectrum , the two true line locations are precisely specified by the two marginal posterior distributions .
this is a verification of what is suggested by the multiple modes in the marginal posterior distribution of the single line location shown in figure [ park : fig : simul ] .
the possibility of model misspecification when using a delta function to model an emission line depends on both the width of the true line and the resolution of the detector .
misspecification only occurs when the line is not contained in one energy bin , and we shall illustrate that such misspecification only has statistical consequences for the fitted line location if it is very severe . indeed there can be a possible statistical advantage of using a delta function rather than a gaussian line if we know the spectral line is not too wide .
as a toy example , consider a simple gaussian model with known standard deviation : when @xmath64 , a 95% confidence interval for @xmath65 is given by @xmath66 .
if we misspecify the modal as @xmath67 with @xmath68 , the resulting interval for @xmath65 is shorter and has lower coverage .
we similarly underrepresent the error bars of an emission line location when we use a delta function for a line that is not contained in one energy bin .
we expect this to reduce both the length and coverage of the confidence regions .
the advantage or disadvantage of this strategy is not immediately clear , however , since the nominal coverage of the intervals is based on an asymptotic gaussian approximation to the posterior distribution which clearly does not apply in this setting .
nonetheless , our simulation study illustrates that the use of a delta function line profile can result in a shorter and more informative hpd region while maintaining good coverage .
we now turn to the computation of hpd regions and the possible statistical advantage of using delta functions in place of narrow gaussian emission lines .
we fit a spectral model that includes a single delta function line or a single narrow gaussian line to the twenty simulated data sets generated under each of the 6 cases . after smoothing the marginal posterior distribution using gaussian kernel smoothing ,
we construct 95% hpd regions for the line location , as shown in figure [ park : fig : simul-10datasets ] . for visual clarity , we present results only for the first ten simulated data sets in figure [ park : fig : simul-10datasets ] ; results from all twenty simulated data sets are discussed in table [ park : tbl : simulhpd ] . because there is no emission line in the spectrum used to simulate data under case 1 , the 95% hpd regions for the line location are very wide and show large uncertainties for the fitted line location .
when there is at least one strong emission line ( i.e. , cases 3 , 4 , and 6 ) , both line models produce comparable hpd regions , although those computed under the gaussian line model appear somewhat wider .
the tradeoff between the two line models becomes evident when there is no strong emission line in the spectrum ( i.e. , cases 2 and 5 ) . in case 2 , the 95% hpd regions for a single gaussian line location are somewhat wider . with the same nominal level , the delta function line model yields more compact and informative hpd regions .
an added advantage of the delta function line model occurs in case 5 when the 95% hpd regions for a single delta function line location consist of two disjoint hpd intervals which simultaneously contain the two true line locations ; this behavior is more often observed with the delta function line model ( 2 times out of 20 ) than the gaussian line model ( 1 time out of 20 ) .
.summary of 95% hpd regions for the line location in the simulation study . [ cols="^,^,^,^,^,^,^ " , ] for example , the 95% hpd regions of the delta function line location are presented in table [ park : tbl : hpds ] along with local modes of the posterior distribution associated with each interval .
each of the 95% hpd regions is composed of a number of disjoint intervals .
only the intervals that have posterior probabilities greater than 5% are presented in table [ park : tbl : hpds ] , so that the probabilities may sum to less than 95% .
for example , the two intervals of obs - id 47 presented in table [ park : tbl : hpds ] have a combined posterior probability of 80.96% and the other eleven intervals not shown in the table have a posterior probability of about 14.04% , for a total of 95% .
the posterior modes of the delta function line location that are located near 2.74 kev are indicated in bold face in table [ park : tbl : hpds ] . and the right panel over a range near 2.85 kev.,width=576 ] the six observations of pg1634 + 706
were independently observed with _
chandra_. thus , under the flat prior distribution on @xmath65 , @xmath69 , the posterior distribution of the line location given all six observations is given by @xmath70 where @xmath71 denotes the six observations , @xmath65 denotes the delta function line location parameter , @xmath72 denotes the set of model parameters other than @xmath65 for the six observations , and @xmath73 represents a likelihood function of @xmath74 given @xmath44 .
( here we allow @xmath75 to vary among the six observations ; i.e. , we do not exclude the possibility that the six observations have somewhat different power law normalizations and photon indexes . ) the values of the posterior distribution given one of the individual data sets is sometimes indistinguishable from zero because of numerical inaccuracies .
thus we add 1/15000 to the posterior probability of each energy bin and renormalize each of the posterior distributions .
this allows the product given in equation [ park : eq : combined ] to be computed for each energy bin and is somewhat conservative as it increases the posterior uncertainty corresponding to each of the individual data sets .
figure [ park : fig : combined ] presents the marginal posterior distribution of the delta function line location given all six observations computed in this way ; the left panel examines the whole range of the line location while the right panel focuses on the range near 2.74 kev . as shown in figure [ park : fig : combined ] , the posterior distribution given all six observations is fairly unimodal and symmetric , except the little local mode near 2.655 kev . thus ,
using a nominal 95% hpd region , the most probable delta function line location is summarized as @xmath76 kev with posterior probability of 95.3% .
posterior predictive methods @xcite can be employed to check the specification of the spectral model .
this methods aim to check the self - consistency of a model , i.e. , the ability of the fitted model to predict the data to which the model is fit . to evaluate and quantify evidence for the inclusion of an emission line in the spectrum , we extend the method of posterior predictive p - values proposed by @xcite and @xcite .
with the _ chandra _ observations of pg1634 + 706 , we consider the same spectral model discussed in [ park : sec : model ] except that we compare three models for the emission line : model 0 : : : there is no emission line in the spectrum .
model 1 : : : there is a delta function emission line with location fixed at 2.74 kev but unknown intensity in the spectrum .
model 2 : : : there is a delta function emission line with unknown location and intensity in the spectrum .
we could equally well consider a gaussian line profile in models 1 and 2 ; either line profile model results in a valid test .
we consider a delta function line profile simply because we are looking for evidence of a narrow emission line .
we use ppp - values to compare the three models and quantify the evidence in the data for the delta function emission line ; see @xcite for details of this method and its advantages over the standard f - test , the standard cash statistic @xcite ( or likelihood ratio test statistic ) , and bayes factors . in the posterior predictive checks , model 1 fixes the delta function line location at 2.74 kev using prior information as to the location of the fe - k - alpha emission line . in order to combine the evidence for the line from all six observations with different exposure areas and exposure times , we base our comparisons on the test statistic that is the sum of the six loglikelihood ratio statistics for comparing model @xmath77 and model 0 , i.e. , @xmath78 where @xmath79 , @xmath80 , and @xmath81 represent the parameter spaces under models 0 , 1 , and 2 , respectively , and @xmath82 denotes the collection of six data sets simulated under model 0 .
specifically , we generate 1000 replications of each data set from the posterior predictive distribution under model 0 and compute @xmath83 for @xmath84 .
histograms of @xmath85 and @xmath86 appear in figure [ park : fig : pp - check ] . comparing the histogram of the simulated test statistics with the observed value of the test statistic yields the ppp - values @xcite shown in figure [ park : fig : pp - check ] .
the ppp - value is the proportion of the simulated test statistics that are as extreme as or more extreme than the observed test statistic .
smaller ppp - values give stronger evidence for the alternative model , i.e. , model 1 or model 2 , thereby supporting the inclusion of the line in the spectrum in our case . as shown in figure [ park : fig : pp - check ] , there is evidence for the presence of the spectral line given all six observations .
the comparison between models 0 and 1 shows stronger evidence for the line location because we are using extra a priori information about the plausible line location .
this article presents methods to detect , identify , and locate narrow emission lines in x - ray spectra via a highly structured multilevel spectral model that includes a delta function line profile .
modeling narrow emission lines with a delta function causes the em algorithms and mcmc samplers developed in @xcite to break down and thus requires more sophisticated statistical methods and algorithms .
the marginal posterior distribution of the delta function emission line location tends to be highly multimodal when the emission line is weak or multiple emission lines are present in the spectrum .
because basic summary statistics are not appropriate to summarize such a multimodal distribution , we instead develop and use hpd graphs along with a list of posterior modes .
testing for an emission line in the spectrum is a notoriously challenging problem because the value of the line intensity parameter is on the boundary of the parameter space , i.e. , zero , under a model that does not include an emission line .
thus , we extend the posterior predictive methods proposed by @xcite to test for the evidence of a delta function emission line with unknown location in the spectrum . using the simulation study in [ park : sec : simul ] ,
we demonstrate the potential advantage of model misspecification using a delta function line profile in place of a gaussian line profile .
we show that the delta function line profile may provide more precise and meaningful summaries for line locations if the true emission line is narrow .
when multiple lines are present in the spectrum , the marginal posterior distribution of a single delta function line location may indicate multiple lines in the spectrum .
our methods are applied to the six different _ chandra _ observations of pg1634 + 706 in order to identify a narrow emission line in the x - ray spectrum .
given all the six observations , the most probable delta function line is identified at @xmath76 kev in the observed frame .
the corresponding rest frame energy for the line is @xmath87 kev , which may suggest the high ionization of iron in the emitting plasma .
there is some recent evidence that high ionization iron line can be variable on short timescales ( see for example mkn766 in * ? ? ?
such variability would explain no detection of the emission line in one of the six _ chandra _ observations .
the authors gratefully acknowledge funding for this project partially provided by nsf grant dms-04 - 06085 and by nasa contract nas8 - 39073 and nas8 - 03060 ( cxc ) . this work is a product of joint work with the california - harvard astrostatistics collaboration ( chasc )
whose members include j. chiang , a. connors , d. van dyk , v. l. kashyap , x .-
meng , j. scargle , a. siemiginowska , e. sourlas , t. park , a. young , y. yu , and a. zezas . the authors also thank the referee for many helpful and constructive comments .
in this section we give an overview of the mode finding and posterior simulation methods used to fit the spectral model with a narrow emission line .
our summary is brief , and thus readers who are interested in a more detailed description should refer to @xcite and @xcite . in order to illustrate our computational strategy ,
consider a simplified example of an _ ideal instrument _ that is not subject to the data contamination processes .
in particular , the redistribution matrix is an identity matrix , the effective area is constant , there are no absorption features , and there is no background contamination .
in addition , we assume that the continuum is specified with no unknown parameters and that there is a single gaussian emission line that has a known width @xmath88 and a known intensity @xmath89 .
thus , the line location is the only unknown parameter , the source model given in equation [ park : eq : ideal ] simplifies to @xmath90 and the counts are modeled as @xmath91 .
this model can be fit using the method of data augmentation @xcite by setting @xmath92 , where @xmath93 and @xmath94 are the counts due to the continuum and the emission line in bin @xmath10 , respectively .
in particular , the em iteratively split the counts into continuum counts and emission line counts .
given the current iterate of the line location , @xmath95 , the e - step updates the line counts via e - step : : : compute @xmath96 $ ] for each bin @xmath10 .
that is , @xmath97 = y_j \frac{{\lambda}\pi_j(\mu^{(t)},\nu_0 ) } { \delta_j f(e_j)+{\lambda}\pi_j(\mu^{(t)},\nu_0)}. \label{park : eq : toy - estep } \end{aligned}\ ] ] which is the weighted average of the bin energies and uses the emission line counts as weights .
although this em algorithm is simple , it breaks down when fitting the location of a _ narrow _ emission line , i.e. , when @xmath88 is small relative to the size of the bins . in the extreme ,
the gaussian line profile becomes a delta function , so that @xmath100 is zero for all bins except the bin containing @xmath95 .
this results in an e - step that computes zero line counts in all bins except the bin containing @xmath95 and finally an m - step that computes @xmath101 .
this means that em will return the same line location at each iteration and that the algorithm will not converge to a mode . in this simplified example , this difficulty can be avoided by directly maximizing the posterior distribution @xmath102 because of the binning of the data , we can treat possible line locations within each bin as indistinguishable and compute the mode by evaluating equation [ park : eq : toy - obs - like ] on the fine grid that corresponds to the binning of the data .
the situation is more complicated in the full spectral model described in
[ park : sec : model ] .
the method of data augmentation can be used to construct efficient algorithms that both fit the parameters in the continuum and the lines and account for instrument response and background contamination . in the case of the narrow emission line ,
however , we must implement a strategy that uses less data augmentation when updating the line location / width than when updating the other model parameters .
the expectation / conditional maximization either ( ecme ) algorithm @xcite allows us to use no data augmentation when updating the line location / width , but the resulting m - step is time consuming owing to the multiple evaluations of the conditional posterior distribution of the line location / width which involve the large dimensional redistribution matrix , @xmath103 .
an intermediate strategy uses the standard data augmentation scheme to adjust for instrument response and background contamination but does not separate continuum and line photons when updating the line location / width .
this strategy is an instance of the alternating expectation / conditional maximization ( aecm ) algorithm @xcite and each iteration is much quicker than with ecme but more iterations are required for convergence .
the algorithms we use for mode finding aim to combine the advantages of ecme and aecm by running one ecme iteration followed by @xmath77 aecm iterations and repeating until convergence . @xcite call this a _
rotation(@xmath77 ) _ algorithm and illustrate the computational advantage of the strategy . returning to the simplified example of appendix [ ap : toyem ]
, we can formulate a gibbs sampler using the same data augmentation scheme .
given the current iterate , @xmath95 , step 1 simulates the line counts in bin @xmath10 via when this algorithm is applied with a narrow emission line , it breaks down just as the em algorithm does .
when a delta function is used for the line profile , the simulation in equation [ park : eq : step1 ] results in no line counts in any bin except the one containing the line , and again @xmath65 does not move from its starting value .
when a narrow gaussian line profile is used , the situation is less extreme , but the sampler exhibits very high autocorrelations and typically can not jump among the posterior modes . just as with em
, the difficulty can be avoided by computing the posterior distribution of @xmath65 on a find grid and directly simulating @xmath112 when the method of data augmentation is used to account for data contamination processes described in
[ park : sec : model ] , however , this approach should be modified in a manner analogous to the ecme and aecm algorithms .
this leads to the strategy of using conditional distributions from different data augmentation schemes . in this case , however , the resulting set of conditional distributions used to construct the gibbs sampler may be _ incompatible _ and there may be no joint distribution that corresponds to this set of conditional distributions .
although such a sampler may result in efficient computation , care must be taken to be sure the sampler delivers simulations from the target posterior distribution .
this is formalized through the partially collapsed gibbs ( pcg ) sampler of @xcite which outlines the steps that should be taken to ensure proper convergence ; refer to @xcite for the applications and illustrations of pcg samplers .
the pcg samplers can be viewed as the stochastic version of ecme and aecm , thereby allowing us to sample the line location with no data augmentation ( as in ecme ) or partial data augmentation ( as in aecm ) .
thus , the pcg sampler differs from the gibbs sampler developed in @xcite in sampling the line location ( and line width ) , and in the order of sampling steps .
@xcite design two pcg samplers to fit the spectral model with a delta function emission line or a narrow gaussian emission line , which are called pcg i and pcg ii , respectively .
as compared to pcg i , pcg
ii requires one additional sampling step for the line width , so that fitting the narrow gaussian emission line is computationally more demanding than fitting the delta function emission line .
yaqoob , t. , george , i. m. , nandra , k. , turner , t. j. , serlemitsos , p. j. , & mushotzky , r. f. 2001 , , 546 , 759 yaqoob , t. , padmanabhan , u. , dotani , t. , george , i. m. , nandra , k. , tanaka , y. , turner , t. j. , & weaver , k. a. 2001 , x - ray emission from accretion onto black holes , yaqoob , t. 2007 , astronomical society of the pacific conference series , 373 , 109 | the detection and quantification of narrow emission lines in x - ray spectra is a challenging statistical task .
the poisson nature of the photon counts leads to local random fluctuations in the observed spectrum that often results in excess emission in a narrow band of energy resembling a weak narrow line . from a formal statistical perspective
, this leads to a ( sometimes highly ) multimodal likelihood .
many standard statistical procedures are based on ( asymptotic ) gaussian approximations to the likelihood and simply can not be used in such settings .
bayesian methods offer a more direct paradigm for accounting for such complicated likelihood functions but even here multimodal likelihoods pose significant computational challenges .
the new markov chain monte carlo ( mcmc ) methods developed in @xcite and @xcite , however , are able to fully explore the complex posterior distribution of the location of a narrow line , and thus provide valid statistical inference . even with these computational tools ,
standard statistical quantities such as means and standard deviations can not adequately summarize inference and standard testing procedures can not be used to test for emission lines . in this paper
, we use new efficient mcmc algorithms to fit the location of narrow emission lines , we develop new statistical strategies for summarizing highly multimodal distributions and quantifying valid statistical inference , and we extend the method of posterior predictive p - values proposed by @xcite to test for the presence of narrow emission lines in x - ray spectra .
we illustrate and validate our methods using simulation studies and apply them to the _ chandra _ observations of the high redshift quasar pg1634 + 706 . |
when the supersonic solar wind plasma from the sun encounters an obstacle , a bow shock is formed .
the incoming solar wind particles ( electrons and ions ) upstream from the bow shock have no information about the obstacle , except in a region magnetically connected to it .
this region is known as the foreshock . at the bow shock ,
a small fraction of the solar wind particles are reflected in the sunward direction .
these backstreaming particles are subjected to the solar wind s * * e**@xmath3*b * drift , where * * e**@xmath4**v**@xmath5**b**@xmath6 , is the solar wind s convective electric field , * b * is the interplanetary magnetic field ( imf ) , @xmath7 is the solar wind velocity and @xmath8 is the speed of light .
the * * e**@xmath3*b * drift velocity is the same for all backstreaming particles , and perpendicular to the imf . as a result ,
the guiding centers of all backstreaming particles move within the @xmath7-*b * plane , gradually drifting away from the field line tangent to the bow shock toward the inner part of the foreshock and being segregated according to their parallel velocities .
electrons , because of their much smaller inertia , are much less affected by this drift and their presence can be detected right next to the field line tangent to the bow shock .
the combination of solar wind and energetic backstreaming electrons results in the production of langmuir waves at the electron plasma frequency @xcite .
backstreaming ions on the other hand , can drive a number of plasma instabilities @xcite , leading to the generation of waves .
the ion foreshock is then characterized not just by the presence of a small fraction of backstreaming ions , but also by the generation and propagation of plasma waves around the local ion cyclotron frequency .
the understanding of planetary foreshocks is far from complete .
the most studied case is the earth s foreshock , which is reasonably well understood thanks to single and multi - spacecraft measurements @xcite .
the first observations from the earth s foreshock were made by the dual spacecraft isee @xcite which identified different types of backstreaming ion distributions : reflected ( now called field - aligned beams ) , intermediate , and diffuse @xcite .
the classification of backstreaming ion populations into these three types was made on the basis of two - dimensional velocity distribution functions and energy - time spectrograms .
further results from isee demonstrated the existence of gyrophase - bunched and gyrotropic backstreaming ion distributions in the foreshock @xcite .
the field - aligned distributions are typically observed at and near the leading edge of the ion foreshock without the presence of waves @xcite . behind the field - aligned beam region , gyrophase - bunched distributions
are detected , while diffuse distributions are found even farther away ( i.e. downstream ) from the ion foreshock boundary .
the association of different linear and non - linear waves to different ion distributions was first studied by @xcite .
gyrophase - bunched and diffuse distributions are in fact observed in the presence of ultra - low frequency ( ulf ) waves . in particular , gyrophase - bunched distributions coexist with ulf quasi - monochromatic waves with substantial amplitudes ( @xmath9 ) @xcite .
the production of gyrophase - bunched ions and associated ulf waves has been studied numerically @xcite , theoretically @xcite and observationally , using data from wind spacecraft @xcite . on the other hand , non - linear ,
steepened waves have been found to be associated with diffuse ion distributions @xcite .
evidence of foreshocks has also been found in other planets as well . in the case of mars ,
@xcite found electrostatic waves which are polarized along the interplanetary magnetic field , and their peak intensity occurs at or near the local solar wind plasma frequency . @xcite identified the ulf wave foreshock boundary of venus through observations made by the pvo ( pioneer venus orbiter ) magnetometer .
plasma and magnetic field observations from the voyager 2 spacecraft reveal ulf waves in the solar wind , which are associated with neptune s , jupiter s and uranus foreshocks ( see for example * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
plasma waves have also been detected upstream from mercury s bow shock using mariner 10 measurements @xcite .
a brief review on mercury s foreshock was made by @xcite .
the region of ulf wave activity is embedded in the ion foreshock , and its boundary is known as the ulf wave foreshock boundary . in the case of the earth , @xcite located this boundary in a statistical way by using combined magnetic field and plasma data from heos 1 .
@xcite , using isse 1 data introduced the so called solar foreshock coordinates , and determined the position of the ulf wave foreshock boundary . for different imf cone angles
( they split their study into @xmath10 - @xmath11 and @xmath12 - @xmath13 data sets ) , @xcite obtained different orientations of the ulf wave foreshock boundary .
other studies have shown that for @xmath14 cone angles , there is a well defined upstream boundary , and this boundary intersects the bow shock at @xmath15 @xcite . using multi - spacecraft data obtained by cluster , @xcite presented two case studies on the onset of ulf foreshock waves , relating their appearance to changes in the orientation of the imf .
the increased wave activity is associated with the change in @xmath1 to quasi - parallel geometries ( @xmath16 ) .
these findings are consistent with previous single spacecraft studies .
although the observations at earth represent a vital element in the study of the physical processes occurring on planetary foreshocks at large , these phenomena necessarily occur at the particular parameter values relevant for our planet .
to explore how these processes change in parameter space , it is just as important to make in situ observations around other planets .
for instance , the solar wind properties vary with heliocentric distance . the parker angle between the stream lines and the radial direction to the sun at 1 au
is predicted to be about 45@xmath17 for a constant solar wind velocity of 429 km@xmath18s
@xcite . at saturn
s distance ( approx .
9 au ) , the parker angle is predicted to be about 85@xmath17 .
@xcite examined in detail the hourly averaged imf data provided by cassini from 13 august 2003 to 14 november 2004 , and found an average parker angle equal to [email protected]@xmath17 for a solar wind speed of 500 km@xmath18s .
the location and shape of planetary bow shocks are determined by the properties of the solar wind flow and by the size and shape of the obstacle to the flow .
saturn s bow shock was first observed by pioneer 11 in 1979 @xcite .
the first crossings made by cassini during saturn orbit insertion ( soi ) were analyzed and discussed by @xcite .
they presented evidence of magnetospheric compression during cassini s first immersion into the magnetosphere and the properties of the solar wind upstream from saturn s bow shock for the first six bow shock crossings observed by cassini .
the magnetic signatures of these bow shock crossings showed a clearly defined overshoot and foot regions associated with the quasi - perpendicular geometry ( @xmath20 ) . using magnetic field and plasma observations made by cassini between june 2004 and august 2005
, @xcite presented a static model of saturn s bow shock .
the model was obtained by fitting a conic section to the first 206 crossings observed by cassini . based on observations from pioneer 11 ,
voyager 1 , voyager 2 and cassini @xcite derive a small eccentricity for saturn s bow shock and found variations in the shock subsolar distance associated to variations in the solar wind dynamic pressure . using the magnetospheric imaging instrument ( mimi ) on board cassini
, @xcite showed measurements of the energetic ion population upstream from both the dusk and dawn sides of the kronian magnetosphere . during the approach phase and first orbits of cassini , the observations revealed the presence of a series of distinct upstream bursts of energetic hydrogen and oxygen ions up to distances of 120 saturn s radii .
they concluded that these oxygen upstream events must be particles leaking from saturn s magnetosphere under favorable imf conditions .
however , @xcite studied 45 hours of mass - resolved observations of cassini plasma spectrometer ( caps ) , which were performed upstream from saturn s bow shock .
the observations show supra - thermal ions composed of h@xmath21 and ions with @xmath22 , presumably solar wind he@xmath23 , with no detectable contribution from magnetospheric water group ions .
@xcite reported the first evidence of upstream low - frequency waves in saturn . during these observations ,
the spacecraft was magnetically connected to saturn s bow shock .
their results suggest that these waves were associated to the planet s foreshock .
these waves displayed a period of 550 seconds in the spacecraft frame and a relative amplitude of 0.3 .
also in the spacecraft frame , the waves are left and right - hand elliptically polarized , and propagate at about 30@xmath17 with respect to the ambient magnetic field . during the first three orbits of cassini spacecraft , @xcite present a characterization of low - frequency waves associated with saturn s foreshock based on cassini magnetometer @xcite . as a result of their survey , they identified two distinct types of waves .
they found a large majority of waves with spacecraft - frame frequencies below the local proton cyclotron frequency ( @xmath24 ) .
these waves are phase - steepened and display a left - hand elliptical polarization as seen by the spacecraft .
@xcite interpreted these waves as fast magnetosonic waves .
this kind of waves is the same presented by @xcite . in a second group , they found waves with frequencies above @xmath0 , quasi - monochromatic and steepened with a right - hand circular polarization , propagating at small angles with respect to the ambient field . according to @xcite these waves could be alfvn waves similar to those observed at earth by @xcite . in this paper ,
we have used cassini magnetometer data and single - spacecraft techniques to study the morphology of saturn s ulf wave foreshock . in situ observations made by the vector helium magnetometer ( vhm ) on board cassini
are presented in section [ obs ] .
a brief study of ulf waves in the magnetic field associated with saturn s foreshock is presented in section [ ulf - waves ] . in section
[ ulf - bound ] we show , for the first time , the determination of the outer boundary of saturn s ulf wave foreshock . for this purpose
we identified a large number of crossings of cassini to ( or from ) the wave region .
the selection criterion for these crossing is described in section [ dtc ] . in sections [ res ] and [ theta ] we introduced the solar foreshock
coordinates developed by @xcite , and present our main results .
discussion and our conclusions are summarized in section [ dis ] and [ con ] .
the observations used for the present study consist of the three components of magnetic field measured by cassini mag obtained upstream from the kronian shock during the first fifteen months of cassini s orbital data , i.e. from soi in june 2004 through august 2005 .
cassini s trajectory in kronian solar orbital ( kso ) coordinates for that period is shown in figure [ fig : orbita ] . in the kso
coordinate system centered on saturn , the @xmath25 axis points to the sun , the @xmath26 axis is anti - parallel to saturn s orbital velocity and the @xmath27 axis points towards the north pole of the ecliptic . during this phase of the orbital tour , the spacecraft explored the dawn side of saturn s magnetosphere at low kronographic latitudes .
the cassini dual magnetometer investigation consists of a vector helium magnetometer ( vhm ) and a fluxgate magnetometer ( fgm ) which provide redundant , fast vector measurements of the ambient magnetic field over a wide range .
the vhm provides accurate vector measurements with a resolution of 2 s@xmath28 over a range of @xmath29 nt , whereas fgm samples the magnetic field over a larger range ( @xmath30 nt ) and at higher frequency ( 32 s@xmath28 ) .
this dual technique is extensively described in @xcite . for the purpose of this paper
, we only consider the data provided by the vhm .
unfortunately , caps could not be used in this study due to the absence of periods with adequate pointing . -@xmath31 and the @xmath32-@xmath33 plane in kronocentric solar orbital coordinates .
the average bow shock fit is the gray - solid line.,scaledwidth=100.0% ] in the absence of multi - spacecraft observations , the reconstruction of the foreshock s geometry requires a model of the kronian bow shock . in this work , we used a static model based on a fit of the crossings using a conic section .
the functional form for the fit was introduced by @xcite , assuming that the bow shock is axially symmetric about the solar wind flow direction .
the general equation of a conic section is given by @xmath34 where @xmath35 is the distance from the planet to a point on the shock surface , @xmath36 is the corresponding polar coordinate angle with respect to the symmetry axis , @xmath37 is the semilatus rectum ( size parameter ) and @xmath38 is the eccentricity .
an empirical model has been presented by @xcite using measurements from cassini mag , caps electron spectrometer sensor ( els ) and density measurements from the radio plasma wave system ( rpws ) . assuming a constant solar wind speed ( @xmath39 km / s ) they found an eccentricity of @xmath40 ( hyperboloid ) and a size parameter for the average bow shock location of @xmath41 , which implies that the average standoff distance is @xmath42 . throughout their study
, they did not find any dependence of the global configuration of the bow shock model with the imf cone angle .
during cassini s excursions into the solar wind , we looked for intervals with low - frequency fluctuations in the magnetic field components .
we found intervals with and without the presence of waves .
figure [ fig : waves_m ] and figure [ fig : waves_nl ] show two examples of the types of wave events found in this paper . for all the wave events
, we made a characterization ( polarization and frequency ) in the spacecraft frame .
we calculated their frequencies @xmath43 ( in the spacecraft frame ) . as a result of this study
, we find two distinct types of oscillations with different properties , depending on whether their frequencies are below or above the local proton cyclotron frequency ( @xmath24 ) .
we also studied their polarization and propagation with respect to the ambient magnetic field using the minimum variance analysis ( mva ) @xcite .
the mva consists in building the variance matrix in terms of the measured magnetic field components for a given time interval and finding the three eigenvalues and corresponding eigenvectors .
the eigenvector corresponding to the smallest eigenvalue @xmath44 ( the maximum and intermediate eigenvalues are respectively @xmath45 and @xmath46 ) is used as an estimate of the direction of propagation of a plane wave . note that the eigenvector set of the variance matrix provides a convenient natural coordinate system in which to display and analyse the data .
.,scaledwidth=100.0% ] .,scaledwidth=100.0% ] figure [ fig : waves_m ] shows an example of waves between 08:31:08 ut and 10:35:00 ut on 17 march ( day 076 ) 2005 .
we obtained that these waves are quasi - monochromatic and steepened with frequencies above @xmath0 . in the spacecraft frame , these waves have periods of the order of 60 s. according to the ambient magnetic field magnitude ( @xmath470.3 nt ) , the period is significantly lower than the local proton cyclotron period t@xmath48 s. figure [ fig : mva1 ] shows the mva results for a quasi - monochromatic wave packet shown in figure [ fig : waves_m ] ( interval 08:31:12 - 08:34:04 ut ) .
figure [ fig : mva1]a shows the components of the magnetic field along the maximum , intermediate and minimum variance direction . figure [ fig : mva1]b and [ fig : mva1]c show the projection of the wave magnetic field ( hodograms ) on the maximum - intermediate and the minimum - intermediate variance planes , respectively .
the circle and the asterisk indicate the beginning and the end of the hodogram , respectively .
a large ratio between the intermediate and minimum eigenvalues of the variance matrix ( @xmath49 ) shows a clear minimum variance direction .
the hodogram in [ fig : mva1]c shows that the polarization on the minimum variance plane is circular right - handed with respect to the mean magnetic field ( b@xmath50 ) .
the angle between the mean magnetic field and the minimum variance eigenvector is @xmath51 , revealing that these waves propagate in a slightly oblique direction with respect to the ambient magnetic field .
figure [ fig : waves_nl ] shows an example of a wave train seen by cassini vhm between 01:00:00 ut and 03:00:00 ut on 27 december ( day 362 ) 2004 .
we obtained that this kind of waves are phase steepened with frequencies smaller than @xmath0 .
these waves have periods of the order of 5 to 10 min in the spacecraft frame , and they were the most frequently observed . according to the ambient magnetic field , 2 to 3 times the local proton cyclotron period for an imf magnitude between 0.35 nt and 0.5 nt . in several cases , we saw a steepening front located at the right of waves with a higher frequency wave packet attached to it .
we note a decrease in amplitude and in the period of these oscillations with increasing distance from the steepening front .
figure [ fig : mva2 ] shows the results of the mva applied on one period of these waves between 01:30:36 ut and 01:33:00 ut . from left to right , it can be seen how an early linear polarization is followed by a circular polarization toward the end of the interval .
the angle between the minimum variance vector and the mean magnetic field suggests that the propagation of these waves is quasi - parallel ( @xmath52 ) to the ambient magnetic field .
the hodogram in figure [ fig : mva2 ] shows that the magnetic field rotation around the minimum variance direction is left - handed with respect to the ambient magnetic field ( b@xmath530 ) .
the circle and the asterisk indicate the beginning and the end of the hodogram , respectively .
it is worth noticing that we identified the same two categories than @xcite had found , for a much bigger cassini mag s data set .
to identify the ulf wave foreshock boundary using the data set described in section [ obs ] , we looked for beginnings ( or endings ) of intervals in which the magnetometer detected ulf waves .
due to the lack of measurements of other physical variables ( solar wind density , velocity , dynamic pressure and composition ) , it is essential to establish a criterion to differentiate a ulf wave foreshock boundary crossing from other possible phenomena , such as discontinuities propagating in the solar wind . ) and in the zone without waves ( @xmath54 ) are in dashed - gray line .
the solid - gray lines correspond to ( @xmath55).,scaledwidth=100.0% ] for a correct determination of the ulf wave foreshock boundary , we considered those crossings in which cassini was entering or leaving the wave region under steady imf conditions . using the whole orbital data set of cassini s vhm
, we identified a total of 59 beginnings or endings of intervals in which the magnetometer detected ulf waves .
for this determination , we made no distinction regarding the kind of the field oscillations , requiring only that the transition from or to the wave region would be clearly apparent .
we defined the following selection criterion for each component of the magnetic field ( @xmath56 ) : if the difference between the average values in the wave zone ( @xmath57 ) and in the zone with no waves ( @xmath54 ) is smaller than the standard deviation in the wave zone ( @xmath58 ) , we consider that cassini crossed a stationary ulf wave foreshock boundary ( figure [ fig : good_one ] ) . ) and in the zone without waves ( @xmath54 ) are in dashed - gray line .
the solid - gray lines correspond to ( @xmath55).,scaledwidth=100.0% ] if any of the three components did not satisfy this condition , the event discarded from our analysis ( figure [ fig : bad_one ] ) .
the average values for each component of the magnetic field in the wave zone and in the zone with no waves are the mean values of the observations over two hours before ( and after ) each apparent crossing .
we found that this particular time interval is sufficiently long to obtain values representative of the mean magnetic field , because the longest wave periods found are 10 minutes long . following this selective criterion
, we reduced the original data set of 59 crossings to only 21 stationary crossings . in a second part of the study , for each component of the magnetic field we changed this 1.@xmath58 condition , requesting the difference between @xmath57 and @xmath54 to be smaller than 1/2@xmath59 ( more restrictive criterion ) and 3/2@xmath59 ( less restrictive criterion ) .
in order to identify the ulf wave foreshock boundary independently from the changes in the imf and the corresponding location of the bow shock , we analyze our data in a particular set of coordinates . in the case of the earth
, @xcite introduced the so - called _ solar foreshock coordinates_. in this coordinate system , the @xmath60 axis points toward the sun ( it is parallel to @xmath25 ) , and the @xmath60-@xmath61 plane is the @xmath62-@xmath63 plane which passes through the location of the spacecraft at a particular stationary crossing .
for each crossing , in or out of the ulf wave foreshock under steady imf conditions , we use the bow shock hyperboloidal fit from @xcite .
we used the eccentricity @xmath64 obtained by @xcite and kept it constant throughout our study . in each crossing , we estimate the value of @xmath37 using equation and the location of the nearest bow shock crossing . for this purpose , we used the list of times and locations of cassini s bow shock crossings between 27 june ( day 027 ) 2004 and 12 august ( day 224 ) 2005 published by @xcite .
the error in the determination of the parameter size was @xmath65 .
assuming @xmath66 , we calculate the foreshock coordinates in the @xmath62-@xmath63 plane or @xmath60-@xmath61 plane ( see figure [ fig : gb ] ) as : @xmath67 where @xmath68 is the imf cone angle , ( @xmath69 ) is the intersection point between the tangent imf line and the bow shock fit , while ( @xmath70 ) is the observation point ( cassini s crossing location ) .
it is worth mentioning that the @xmath62-@xmath63 plane in general does not contain the @xmath71 axis ( except for the particular @xmath62-@xmath63 plane which passes through the center of the planet ) .
in fact , there should be a third coordinate to measure the distance of any @xmath62-@xmath63 plane to the center of the planet .
@xcite have deliberately ignored that third coordinate , and showed a remarkable correlation in the @xmath72-@xmath73 plane .
as shown in figure [ fig : gb ] , the non - orthogonal coordinate set ( @xmath74 ) has its origin in the tangent point ( @xmath69 ) , @xmath72 being the distance along the tangent imf line and @xmath73 being the horizontal distance from that particular field line to cassini s location
. -@xmath63 plane , i.e. the @xmath60-@xmath61 plane .
the coordinate @xmath72 is the distance along the tangent magnetic field line between the tangent point and the observation point .
the coordinate @xmath73 is the distance along the @xmath71 direction between the tangent magnetic field line and the observation point.,scaledwidth=100.0% ] for the 21 stationary crossings considered in section [ dtc ] , we have calculated their locations in terms of the coordinate set ( @xmath74 ) . at earth ,
the location of the ulf wave foreshock boundary was found to depend on the imf cone angle ( @xcite , @xcite ) .
for this reason , we analyzed two separate sets of data , one with @xmath75 and the other with @xmath76 , i.e. for small and large cone angles .
we performed a scatter plot of the 21 ulf wave foreshock boundary crossings considered .
figure [ fig : sigma1 ] shows the cases with @xmath77 in black circles and those with @xmath78 in gray circles , and the straight line is our best linear fit ( black line ) for all our crossings .
the small cone angle cases ( gray circles ) correspond to tangent lines close to the asymptote of the hyperbola ( @xmath79 ) , and therefore have relatively large error bars . however , our best fit considering only the @xmath80 cases , yields @xmath81 ( @xmath82 ; @xmath83 ) , which is indistinguishable from the result displayed in figure [ fig : sigma1 ] .
( dark circles ) and @xmath78 ( gray squares ) .
if we only consider the @xmath77 cases , our best fit yields @xmath81 ( @xmath82 ; @xmath84 ) , which is indistinguishable from the best fit displayed in this figure.,scaledwidth=100.0% ] we also studied the variability of this result as our selection criterion become more stringent ( see sub - section [ dtc ] ) . in figure
[ fig : sigma1 ] we only consider those crossings in which the three components of the magnetic field ( @xmath56 ) satisfy that the difference between the average values in the wave zone and in the zone without waves is smaller than the standard deviation in the wave zone ( 1.@xmath58 ) . as we mentioned in sub - section [ dtc ] , if any of the three components did not satisfy this condition , the crossing was discarded from our analysis . in a more restrictive criterion , for each component of the magnetic field , we changed the 1.@xmath58 condition to 1/2@xmath59 . in this way we reduced the original data set of 21 stationary crossings to only 16 crossings . considering only these 16 cases , our best fit yields @xmath81 ( @xmath85 ; @xmath86 ) , which is almost identical to the result displayed in figure [ fig : sigma1 ] .
we also tried with a less restrictive criterion , changing the original 1@xmath59 to 3/2@xmath59 . in this case our original data set of 21 crossings , increase to 29 crossings , and our best fit yields @xmath81 ( @xmath87 ; @xmath88 ) , which is again , within the errors , the same result displayed in figure [ fig : sigma1 ] .
condition ) ; the crossings in gray are added when the 1@xmath59 criterion is considered , while the crossings indicated by white are added when the less restrictive criterion of 3/2@xmath59 is considered.,scaledwidth=100.0% ] in figure [ fig : nu_mu_all ] we show the results for the three criteria .
black circles are those crossings satisfying the most stringent criterion ( 1/2.@xmath58 ) , gray circles correspond to crossings between 1/2@xmath59 and 1@xmath59 , while white circles are those between 1@xmath59 and 3/2@xmath59 . in all these cases ,
the uncertainties in the determination of the three magnetic field components were dominated by the statistical error given by the corresponding standard deviation . as mentioned , for the semilatus rectum ( size parameter ) @xmath37 and the eccentricity @xmath38 from the hyperboloidal model , we consider the uncertainties presented in section [ ulf - waves ] .
for each crossing studied , we extend the straight magnetic field line and check for connection to the bow shock fit . as mentioned in the previous section , we checked for magnetic field line connectivity to the bow shock by using the @xcite fit , with fixed eccentricity @xmath40 .
we varied @xmath37 so as to make the bow shock fit to coincide with the closest bow shock crossing made by cassini .
if we find that cassini is connected to the shock by straight magnetic field lines , we identify the intersection point on the bow shock fit .
once this point has been identified , the angle between the magnetic field and the shock normal at the shock surface , @xmath1 , can be calculated . in all the wave events that we identified we found that the magnetic field line intersects the bow shock fit ,
i.e. cassini was in the region magnetically connected to the bow shock .
in particular , we find that the increment in wave activity is associated to foreshock regions for which @xmath16 .
in fact , 12 out of the 21 crossings have @xmath90 , while the remaining 9 crossings have @xmath1 within the range between 35@xmath17 and 55@xmath17 . if we consider the 21 crossings , the average value is @xmath91 and the standard deviation is @xmath92 .
( dot - dashed line ) and @xmath93 ( dashed line ) . in the @xmath62-@xmath63 plane @xmath94 , the average bow shock fit ( in units of @xmath37 ) , the tangent field line for @xmath95 ( point line ) , and the magnetic field line corresponding to @xmath89 ( solid line).,title="fig:",scaledwidth=100.0% ] it is important to emphasize that @xmath89 constitutes a natural and conventional distinction between quasi - parallel and quasi perpendicular shocks . for a given magnetic field orientation and parameter size , there are two lines on the hyperboloidal shock surface , for which @xmath89 . in figure
[ fig : curves ] we show a a schematic view of the bow shock ( in units of @xmath37 ) and the lines corresponding to @xmath89 ( dot - dashed lines ) and @xmath93 ( dashed lines ) , where the field lines are tangent to the hyperboloidal shock .
the * x*-*y * plane in figure [ fig : curves ] is parallel to the magnetic field lines and contains the planet in @xmath96 . within this context
, we claim that the set of magnetic field lines that cross the shock on the @xmath89 curves will determine the ulf wave foreshock boundary . on other hand , a perhaps more abstract identification of the ulf wave foreshock boundary is given by the best fit of crossings on the @xmath72-@xmath73 plane @xcite . , @xmath73 ) pair for the location of points satisfying @xmath89 at different ( parallel ) @xmath62-@xmath63 planes for @xmath95 and parameter size @xmath97 51 @xmath98 .
the distance of each plane to the planet is labeled in units of @xmath37.,scaledwidth=100.0% ] therefore , we decided to study the consistency of our linear best fit in the @xmath72-@xmath73 plane and the curve corresponding to @xmath89 . assuming a constant @xmath99 ( the magnetic field lines lay in the @xmath100-@xmath101 plane ) and parameter size @xmath97 51 @xmath98 , we computed the ( @xmath72,@xmath73 ) pair for the location of the point that satisfies @xmath89 at different ( parallel ) @xmath62-@xmath63 planes ( see figure [ fig : trian ] )
the corresponding distance of each plane ( in units of @xmath37 ) to the planet is labeled .
we can clearly see in figure [ fig : trian ] an approximate proportionality between the @xmath72-@xmath73 coordinates at any given plane .
this implies that if cassini crossed right at the intersection between the bow shock and the @xmath89 field line , we should expect a straight line passing through the origin in the @xmath72-@xmath73 plane ( see figure [ fig : sigma1 ] ) .
the fact that cassini crosses that very field line necessarily at a distance from that intersection toward the upstream direction , causes a systematic increase in the value of @xmath72 , leaving @xmath73 essentially unchanged .
the distance from the bow shock measured along field lines range between @xmath102 and @xmath103 for the 21 crossings considered . on the one hand
, this effect will reduce the quality of the linear fit , and on the other hand it will shift the line on the @xmath72-@xmath73 plane to the right ( negative intercept ) , exactly as observed in figure [ fig : sigma1 ] .
all the wave events detected by cassini s magnetometer show evidence of magnetic connectivity to saturn s bow shock , which is a clear indication that these waves are associated with saturn s foreshock . in absence of a reliable physical model of the bow shock
, we used the empirical model of @xcite , which is based on extensive observations from cassini made during the same time span of our work ( 15 months from june 2004 through august 2005 ) . assuming that the bow shock is axially symmetric about the solar wind flow direction , the average bow shock location is obtained by fitting a conic section to the bow shock crossings using a nonlinear least squares technique .
this fit provides an adequate description of the average shape , given by an eccentricity of @xmath40 , and also that of its average location , corresponding to @xmath104 . in our study
, we assume that the shape of saturn s bow shock remains constant throughout our observational time span , while its location varies in response to time variations of the solar wind s ram pressure .
in other words , for each wave event or ulf wave foreshock boundary crossing , we assumed @xmath38 to remain constant .
meanwhile , the semilatus rectum @xmath37 , which determines the distance of the bow shock to the planet , was allowed to change to account for this pressure effect .
we found values of @xmath37 between @xmath105 and @xmath106 .
we also calculated the parameter size using an average between the two closest bow shock crossing locations for a given event ( wave or crossing ) .
for the sake of comparison , we also made our calculation assuming that the parameter size remains constant ( @xmath107 , which is the average value ) for all the wave events .
we found , that the bow shock fit was not representative of many of the upstream events studied , since they fell in the downstream region of this static fit .
therefore , we concluded that the bow shock fit calculated using equation with a varying @xmath37 parameter gives a more accurate empirical representation of the real kronian bow shock . for all the wave events observed , we performed a preliminary characterization based on their frequency and polarization as measured in the spacecraft reference frame .
according to their frequencies , we observe two different types of oscillations .
waves with frequencies below @xmath0 are the most frequently observed , they are phase steepened with periods of the order of 5 to 10 minutes in the spacecraft frame . as we mentioned in section [ ulf - waves ] , we saw steepening fronts located at the right of waves with a higher frequency wave packets attached to them .
we also note a decrease in amplitude and in the period of the oscillations within each wave packet with increasing distance from the steepening front .
the angle between the minimum variance vector and the mean magnetic field suggests that the propagation of these waves is quasi - parallel .
the magnetic field rotation around the minimum variance direction is left - handed with respect to the imf direction . as was discussed in @xcite
, these signatures suggest that these waves are ion / ion resonant right - hand ( fast magnetosonic ) mode waves which steepen during the nonlinear regime and emit a dispersive whistler to stop the steepening .
we also observe waves with frequencies above @xmath0 , which are either quasi - monochromatic or steepened , with periods of the order of 60 s ( significantly lower than the local proton cyclotron period t@xmath108 s ) .
for the quasi - monochromatic events , their corresponding hodograms show that the polarization is circular right - handed with respect to the mean magnetic field .
the angle between the ambient magnetic field and the minimum variance direction reveals that these waves propagate in a slightly oblique direction with respect to the imf orientation .
this kind of wave packets is an exception rather than the rule ( we found ten events only ) . at the earth
s foreshock , there is a particular type of ulf waves , the so - called 30 s ( period ) waves which are quasi - monochromatic ( e.g. * ? ? ?
this kind of waves are found always near to the ulf wave foreshock boundary . in our study
we did not find quasi - monochromatic waves next to the saturnian ulf wave foreshock boundary .
a more detailed statistical study would shed some light about the nature of this second group of waves and whether they are the saturnian equivalent of the 30 s modes found in the earth s foreshock .
when cassini is magnetically connected to the bow shock fit by straight magnetic field lines ( since we are assuming a uniform external magnetic field ) , we identify the intersection point between the field line and the bow shock fit as well as the angle @xmath1 . in all the crossings that we observed ,
we find that the presence of magnetic fluctuations is associated with values of @xmath1 corresponding to quasi - parallel geometries ( @xmath16 ) .
this result is fully consistent with the one obtained by @xcite at earth .
using isee data , they examined the presence of ulf waves in the region immediately upstream from the earth s bow shock .
their statistical study shows a well - defined boundary on the @xmath7-*b * plane , separating the disturbed from the undisturbed magnetic field at @xmath109 .
our result for saturn s foreshock is consistent with this scenario , regardless of whether these waves are immediately adjacent to the bow shock or deep into the foreshock region .
@xcite developed a theoretical framework for studying the general trajectories of ions reflected or leaked upstream from the earth s bow shock and subjected only to the lorenz force in a steady imf . in the case known as specular reflection , the ions bounce off the shock potential reversing their velocity component along the shock normal @xmath110 , while maintaining their parallel component . in the de
hoffman - teller frame @xcite , assuming that the bow shock is infinitesimally thin and has no overshoot , @xcite found that after interacting with the shock , the guiding center motion of incident solar wind ions is reflected to the upstream region with @xmath16 .
for @xmath20 , specularly reflected ions always re - encounter the bow shock with sufficient energy to penetrate into the downstream region .
it is worth mentioning that these specularly reflected ions oriented to the upstream region for quasi - parallel geometries would have no conserving magnetic moment , in contrast with the field - aligned reflected beams , for which there is a magnetic moment conservation @xcite . in the absence of cassini plasma observations that could confirm the presence of backstreaming ions giving rise to an ion foreshock region , we can not observationally associate the presence of waves with specific ion distributions .
however , if the mechanism of generation of these waves is local , i.e. the wave growth rate @xmath111 is sufficiently large , we could expect specularly reflected ions to be responsible for these waves since they are the only population whose spatial distribution depends on @xmath1 . in this case , for quasi - parallel geometries , the reflected ions can go upstream from the bow shock and be detectes through ulf wave activity .
nevertheless , a full discussion on the wave generation caused by particle distributions as a function of @xmath1 , is beyond the purpose of this study . for a correct determination of saturn s ulf wave foreshock boundary , we only considered crossings in which cassini was entering or leaving the wave zone under steady imf conditions , i.e. stationary crossings .
we defined a criterion based on the variability of the magnetic field in order to make a proper identification of these stationary crossings . as a result , we reduced our original list of crossings to only 21 stationary ones . at earth , using single spacecraft techniques , @xcite obtained different orientations of ulf wave foreshock boundary for different cone angles ( more specifically , they considered two ranges : @xmath112 - @xmath113 and @xmath12 - @xmath13 ) . in agreement with parker s angle prediction for saturn @xcite , our study shows that 83% of the crossings correspond to cone angles larger than 55@xmath17 .
therefore , the range of @xmath99 considered is different to that at earth .
nevertheless , we analyzed the correlation between @xmath72-@xmath73 coordinates for two ranges of @xmath99 : one set of data with @xmath114 and the other with @xmath115 .
a comparison of the best linear fit for the two data set indicates no significant difference between them ( see figure [ fig : sigma1 ] ) .
this result , however , does not rule out a dependence on @xmath116 since the errors in the determination of @xmath72-@xmath73 become larger with decreasing @xmath99 .
more specifically , the main source of error propagation in the determination of each ( @xmath74 ) pair is the intersection point between the tangent imf line and the bow shock fit .
this intersection point depends critically on the imf cone angle , which is therefore the main source of error . in particular , small cone angles correspond to almost tangent magnetic field lines close to the asymptotes of the hyperbola . on the other hand ,
the appearance of waves is observed to occur at @xmath1 values of @xmath117 .
the surface determined by the condition @xmath89 is mapped on the @xmath72-@xmath73 plane as straight line , as illustrated by figure [ fig : trian ] . as a result
, we find that there is an equivalence between the ulf wave foreshock boundary determined by @xmath118 and the best fit in the @xmath72-@xmath73 plane citepgb1986 .
they are in fact two alternative ways to determine the same spatial region within the foreshock .
finally , we studied the robustness of the @xmath72-@xmath73 correlation as our selection criterion becomes more stringent .
for instance , we change the 1.@xmath58 condition to a more stringent value of 1/2.@xmath58 for the three components of magnetic field , or change it into the less restrictive value of 3/2@xmath59 for each component .
we find that regardless of whether we choose a more or less restrictive criterion , our best fit remains the same , within the fit errors .
we also find that for these three different criteria , the boundary does not seem to depend on whether the cone angle is large or small .
using cassini s data from soi in june 2004 through august 2005 , we conducted a detailed survey and analysis of ulf waves upstream from saturn s bow shock .
all the wave events show evidence of magnetic connectivity to saturn s bow shock , therefore we conclude that these wave events are undoubtedly associated with saturn s foreshock . as for their frequencies , we identify two distinct types of wave populations .
the most frequently observed have frequencies below @xmath0 and are phase steepened , with periods of the order of 5 to 10 minutes in the spacecraft frame .
these waves are often accompanied by precursor whistler wave trains . in agreement with previous works
, we suggest that these waves are ion / ion resonant right - hand ( fast magnetosonic ) mode waves which steepen during the nonlinear regime and emit a dispersive whistler to stop the steepening .
we also observe waves with frequencies above @xmath0 , which appear as either quasi - monochromatic or steepened , with periods of @xmath47 1 minute .
the quasi - monochromatic events have a circular right - handed polarization with respect to the mean magnetic field and a propagation slightly oblique with respect to the imf .
a more detailed study will be needed in order to conclude whether the latter are the saturnian equivalent of the 30 s modes found at earth .
it is worth noticing that we have identified the same two categories than @xcite have found , although for a much bigger cassini mag s data set .
we identified 21 stationary crossings inbound and outbound from the ulf wave foreshock region .
we calculated their solar foreshock coordinates in the @xmath62-@xmath63 plane and we have identified for the first time saturn s ulf wave foreshock boundary . in the @xmath72-@xmath73 plane
we do not find a clear dependence between the foreshock boundary and the imf cone angle .
we also found that the presence of waves is associated with the change in @xmath1 to quasi - parallel geometries .
moreover , we find that our determination of the ulf wave foreshock boundary as the surface given by @xmath89 , is indeed consistent with the linear fit in the @xmath72-@xmath73 plane , first proposed by @xcite for the earth s ulf wave foreshock . in this regard , we speculate that the specular reflection is the candidate process for the reflected ions ( gyrophase - bunched and diffuse distributions ) , since it is the only process to produce a signature in @xmath1 . finally , we studied the robustness of the @xmath72-@xmath73 correlation for a more restrictive criterion .
we found that regardless of the criterion that we use , we obtain the same ulf wave foreshock boundary in the @xmath72-@xmath73 plane . | even though the solar wind is highly supersonic , intense ultra - low frequency ( ulf ) wave activity has been detected in regions just upstream of the bow shocks of magnetized planets .
this feature was first observed ahead of the earth s bow shock , and the corresponding region was called the ulf wave foreshock , which is embedded within the planet s foreshock .
the properties as well as the spatial distribution of ulf waves within the earth s foreshock have been extensively studied over the last three decades and have been explained as a result of plasma instabilities triggered by solar wind ions backstreaming from the bow shock . since july 2004 , the cassini spacecraft has characterized the saturnian plasma environment including its upstream region . since cassini s saturn orbit insertion ( soi ) in june 2004 through august 2005
, we conducted a detailed survey and analysis of observations made by the vector helium magnetometer ( vhm ) .
the purpose of the present study is to characterize the properties of waves observed in saturn s ulf wave foreshock and identify its boundary using single spacecraft techniques .
the amplitude of these waves is usually comparable to the mean magnetic field intensity , while their frequencies in the spacecraft frame yields two clearly differentiated types of waves : one with frequencies below the local proton cyclotron frequency ( @xmath0 ) and another with frequencies above @xmath0 .
all the wave crossings described here , clearly show that these waves are associated to saturn s foreshock .
in particular , the presence of waves is associated with the change in @xmath1 to quasi - parallel geometries .
our results show the existence of a clear boundary for saturn s ulf wave foreshock , compatible with @xmath2 surfaces .
foreshock , ulf waves , saturn , cassini |
recently , hoava proposed a renormalizable gravity theory in four dimensions which reduces to einstein gravity with a non - vanishing cosmological constant in ir but with improved uv behaviors @xcite .
the latter theory admits a lifshitz scale - invariance in time and space , @xmath0 and @xmath1 ( in particular , @xmath2 for the present case ) , exhibiting a broken lorentz symmetry at short scales , while at large distances higher derivative terms do not contribute , and the theory reduces to standard general relativity ( gr ) .
since one of the most important aspects of the theory is a lifshitz - type anisotropic scaling , it is often called hoava - lifshitz gravity . the hoava theory has received a great deal of attention and since its formulation various properties and characteristics have been extensively analyzed , ranging from formal developments @xcite , cosmology @xcite , dark energy @xcite and dark matter @xcite , and spherically symmetric or rotating solutions @xcite .
the natural setting of hoava - lifshitz gravity is the adm formalism , where the 4-dim metric is parameterized by the following @xmath3 @xmath4 is the lapse function and @xmath5 the shift function , respectively .
the ir - modified hoava action is given by @xmath6 , \label{haction}\end{aligned}\ ] ] where @xmath7 , @xmath8 , @xmath9 , @xmath10 , @xmath11 and @xmath12 are constant parameters .
@xmath13 is the three - dimensional curvature scalar for @xmath14 ; @xmath15 is the extrinsic curvature given by @xmath16 and @xmath17 is the cotton - york tensor , defined as @xmath18 the fundamental constants of the speed of light @xmath19 , newton s constant @xmath20 , and the cosmological constant @xmath21 are provided by the parameters of the theory and are defined as @xmath22 respectively .
there are basically four versions of hoava gravity , namely , those with or without the `` detailed balance condition '' , and with or without the `` projectability condition '' .
the `` detailed balance condition '' restricts the form of the potential in the 4dim lorentzian action to a specific form in terms of a 3dim euclidean theory . in a cosmological context
, this condition leads to obstacles , and thus must be abandoned . in this context , the last term in eq .
( [ haction ] ) represents a ` soft ' violation of the ` detailed balance ' condition , which modifies the ir behavior .
this ir modification term , @xmath23 , with an arbitrary cosmological constant , represent the analogs of the standard schwarzschild-(a)ds solutions , which were absent in the original hoava model .
the `` projectability condition '' essentially stems from the fundamental symmetry of the theory , i.e. , the foliation - preserving diffeomorphism invariance , and must be respected .
foliation - preserving diffeomorphism consists of a 3dim spatial and the space - independent time reparametrization .
note that the lapse function is essentially the gauge degree of freedom associated with the time reparametrization , so that it is natural to restrict it to be space - independent , i.e. , @xmath24 .
ir - modified hoava gravity seems to be consistent with the current observational data , but in order to test its viability more observational constraints are necessary @xcite .
thus , it is the purpose of the present paper to consider the classical tests ( perihelion precession , light bending , and the radar echo delay , respectively ) of general relativity for static gravitational fields in the framework of hoava - lifshitz gravity . to do this
we shall adopt for the geometry outside a compact , stellar type object ( the sun ) , the static and spherically symmetric metric obtained by kehagias and sfetsos @xcite .
this work is essentially based on the paper @xcite .
consider a static and spherically symmetric solution , with the metric ansatz : @xmath25 by substituting the metric into the action , the resulting reduced lagrangian , after angular integration , is given by & = & e^(+)/2 . the above reduced lagrangian yields the following equations of motions @xmath26- 3 { { \lambda}}_w^2 r^2 & = & 0\ , , \\ \frac{(\nu'+\lambda')}{2 } \left[(\lambda_g-1)(\lambda ' e^{-\lambda})-2\lambda_g \frac{e^{-\lambda}-1}{r}+2({{\omega}}-{{\lambda}}_w ) r\right ] + ( \lambda_g-1 ) \left[(-\lambda''+(\lambda)')e^{-\lambda } -\frac{2(e^{-\lambda}-1)}{r^2}\right]&=&0\,,\end{aligned}\ ] ] by varying the functions @xmath27 and @xmath28 , respectively .
now , imposing @xmath29 , which reduces to the einstein - hilbert action in the ir limit , one obtains the following solution of the vacuum field equations in hoava gravity , @xmath30 } , \label{gensolution}\ ] ] where @xmath31 is an integration constant . throughout this work
, we consider the kehagias - sfetsos ( ks ) asymptotically flat solution @xcite , i.e. , @xmath32 and @xmath33 : @xmath34 if the limit @xmath35 , from eq .
( [ kssolution ] ) we obtain the standard schwarzschild metric of general relativity , @xmath36 .
there are two event horizons at @xmath37 $ ] . to avoid a naked singularity at the origin , impose the condition @xmath38 .
note that in the gr regime , i.e. , @xmath39 , the outer horizon approaches the schwarzschild horizon , @xmath40 , and the inner horizon approaches the central singularity , @xmath41 .
there are three fundamental tests which provide observational evidence for gr and its generalizations , namely , the precession of the perihelion of mercury , the deflection of light by the sun , and the radar echo delay observations . the latter have been used to successfully test the schwarzschild solution of general relativity and some of its generalizations . in this section ,
we consider these standard solar system tests of general relativity in the case of the ks asymptotically flat solution @xcite of hoava - lifshitz gravity . throughout this section
we use the natural system of units with @xmath42 .
the analysis outlined below is essentially based on the paper @xcite , and a similar analysis has been carried out in the context of branewold models @xcite .
the line element , given by eq . ( [ sssmetric ] ) , provides the following equation of motion for @xmath43 @xmath44 where @xmath45 and @xmath46 .
the analysis below is simplified by applying a change of variable @xmath47 , and using @xmath48 .
we further formally represent @xmath49 , so that the equation of motion ( [ inter1 ] ) takes the following form @xmath50 where @xmath51 , and @xmath52 is defined as @xmath53 a circular orbit @xmath54 is given by the root of the equation @xmath55 .
any deviation @xmath56 from a circular orbit must satisfy the equation @xmath57 \delta = o\left ( \delta ^{2}\right ) , \ ] ] which is obtained by substituting @xmath58 into eq .
( [ inter2 ] ) . therefore , in the first order in @xmath59 , the trajectory is given by @xmath60 where @xmath61 and @xmath62 are constants of integration .
the variation of the orbital angle from one perihelion to the next is @xmath63 , where @xmath64 is the perihelion advance , which represents the rate of advance of the perihelion .
@xmath64 is given by @xmath65 , or for small @xmath66 , by @xmath67 .
as the planet advances @xmath68 radians in its orbit , its perihelion advances @xmath69 radians . for a complete rotation
we have @xmath70 , and the advance of the perihelion is @xmath71 . for the ks solution ,
the perihelion precession is given by : @xmath72\right\}}{x^4_0 \left(4 x^3_0+\omega _ 0 \right)^{3/2}}.\ ] ] with the dimensionless parameters defined as @xmath73 , and @xmath74 .
in the `` post - newtonian '' limit @xmath75 , we obtain the gr result @xmath76 , where @xmath77 . the variation of the perihelion precession angle as a function of @xmath78
is represented in fig .
[ fig2 ] . as a function of @xmath78.,width=292 ]
the observed value of the perihelion precession of the planet mercury is @xmath79 arcsec per century @xcite .
therefore the range of variation of the perihelion precession is @xmath80 .
this range of observational values fixes the range of @xmath78 as @xmath81 the deflection angle of a light ray by the gravitational field of a massive body in a spherically symmetric geometry is derived in @xcite , and is given by @xmath82 , with @xmath83 where @xmath84 is the distance of the closest approach .
here we have taken into account that in the absence of a gravitational field a light ray propagates along a straight line . in the case of the sun ,
the deflection angle of alight ray is @xmath85 .
the deflection angle of light rays passing nearby the sun in the ks geometry is given by @xmath86 ^{-1/2}}{\sqrt{\left [ 1+\omega _ { 0}x_{0}^{2}-\sqrt{x_{0}(\omega _ { 0}^{2}x_{0}^{3}+4\omega _ { 0})}\right ] x^{2}/\left [ 1+\omega _ { 0}x_0 ^ 2x^{2}-\sqrt{x_{0}x(\omega _ { 0}^{2}x_{0}^{3}x^{3}+4\omega _ { 0})}\right ] -1}}\frac{dx}{x},\ ] ] where @xmath87 , @xmath88 , and considering @xmath89 . for the sun ,
by taking @xmath90 cm , we find for @xmath91 the value @xmath92 .
the variation of the deflection angle @xmath93 is represented , as a function of @xmath94 , in fig .
[ fig3 ] .
( in arcseconds ) as a function of the parameter @xmath78.,width=292 ] the best available data come from long baseline radio interferometry @xcite , which gives @xmath95 , with @xmath96 , where @xmath97 arcsec .
thus , the observational constraints of light deflection restricts the value of @xmath78 to @xmath98 the aim of this test is to measure the time required for radar signals to travel to an inner planet or satellite in two circumstances , namely , ( i ) when the signal passes very near the sun and , ( ii ) when the ray does not go near the sun .
the time of travel of light @xmath99 between two planets , situated far away from the sun , is given by @xmath100 , where @xmath101 and @xmath102 are the distances of the planets to the sun , respectively .
if the light travels close to the sun , the time travel is @xmath103 /2}dy,\ ] ] and the time difference is given by @xmath104 /2}-1\right\ } dy=\int_{-l_{1}}^{l_{2}}\left\ { e^{\left [ \lambda \left ( \sqrt { y^{2}+r_{\odot } ^{2}}\right ) -\nu \left ( \sqrt{y^{2}+r_{\odot } ^{2}}\right ) \right ] /2}-1\right\ } dy,\ ] ] with @xmath105 , and @xmath106 is the radius of the sun .
recent measurements of the frequency shift of radio photons to and from the cassini spacecraft as they passed near the sun have greatly improved the observational constraints on the radio echo delay . for the time delay of the signals emitted on earth , and which graze the sun
, one obtains @xmath107 , with @xmath108 @xcite .
the standard gr radar echo delay value is @xmath109 . for the ks solution , by introducing a new variable @xmath110 defined as @xmath111 , and by representing @xmath11 as @xmath112 , we obtain for the time delay the following expression @xmath113 } { 1 - 4\omega _ { 0}\left ( \xi ^{2}+a^{2}\right ) \left [ \sqrt{1+\left ( 1/2\omega _ { 0}\right ) \left ( \xi ^{2}+a^{2}\right ) ^{-3/2}}-1\right ] } d\xi , \ ] ] where @xmath114 , @xmath115 , and @xmath116 , respectively .
the variation of the time delay as a function of @xmath78 is represented in fig .
[ fig4 ] . as a function of @xmath78.,width=292 ]
the observational values of the radar echo delay are consistent with the ks black hole solution in hoava - lifshitz gravity if @xmath117
in the present work , we have considered the observational and experimental possibilities for testing , at the level of the solar system , the kehagias and sfetsos solution of the vacuum field equations in hoava - lifshitz gravity . in this context , the parameter @xmath11 , having the physical dimensions of @xmath118 , is constrained by the perihelion precession of mercury to a value of @xmath119 @xmath120 .
the deflection angle of the light rays by the sun can be fully explained in hoava - lifshitz gravity with the parameter @xmath11 having the value @xmath121 , while the radar echo delay experiment suggests a value of @xmath122 @xmath120 . from these values
we can give an estimate of @xmath11 as @xmath123 this requires a very precise fine tuning of this constant at the level of the solar system .
thus , the study of the classical tests of general relativity provide a very powerful method for constraining the allowed parameter space of hoava - lifshitz gravity solutions , and to provide a deeper insight into the physical nature and properties of the corresponding spacetime metrics .
p. hoava , jhep * 0903 * , 020 ( 2009 ) .
p. hoava , phys .
d * 79 * , 084008 ( 2009 ) .
t. p. sotiriou , m. visser and s. weinfurtner , phys .
lett . * 102 * , 251601 ( 2009 ) ; m. visser , phys .
* d80 * , 025011 ( 2009 ) ; p. horava , phys . rev .
* 102 * , 161301 ( 2009 ) ; r. g. cai , y. liu and y. w. sun , jhep * 0906 * , 010 ( 2009 ) ; b. chen and q. g. huang , arxiv:0904.4565 [ hep - th ] ; d. orlando and s. reffert , class .
* 26 * , 155021 ( 2009 ) ; r. g. cai , b. hu and h. b. zhang , phys . rev . * d80 * , 041501 ( 2009 ) ; t. nishioka , class .
* 26 * , 242001 ( 2009 ) ; m. li and y. pang , jhep * 0908 * , 015 ( 2009 ) ; c. charmousis , g. niz , a. padilla and p. m. saffin , jhep * 0908 * , 070 ( 2009 ) ; t. p.
sotiriou , m. visser and s. weinfurtner , jhep * 0910 * , 033 ( 2009 ) ; g. calcagni , arxiv:0905.3740 [ hep - th ] ; d. blas , o. pujolas and s. sibiryakov , jhep * 0910 * , 029 ( 2009 ) ; r. iengo , j. g. russo and m. serone , jhep * 0911 * , 020 ( 2009 ) ; c. germani , a. kehagias and k. sfetsos , arxiv:0906.1201 [ hep - th ] ; s. mukohyama , jcap * 0909 * , 005 ( 2009 ) ; a. kobakhidze , arxiv:0906.5401 [ hep - th ] ; c. appignani , r. casadio and s. shankaranarayanan , arxiv:0907.3121 [ hep - th ] ; j. kluson , arxiv:0905.1483 [ hep - th ] ; j. kluson , arxiv:0907.3566 [ hep - th ] ; n. afshordi , phys . rev . * d80 * , 081502 ( 2009 ) ; a. kocharyan , phys . rev . * d 80 * , 024026 ( 2009 ) ; d. orlando and s. reffert , phys . lett .
b * 683 * , 62 ( 2010 ) ;
d. blas , o. pujolas and s. sibiryakov , arxiv:0909.3525 [ hep - th ] ; m. i. park , arxiv:0910.1917 [ hep - th ] ; k. koyama and f. arroja , arxiv:0910.1998 [ hep - th ] ; m. i. park , arxiv:0910.5117 [ hep - th ] ; a. papazoglou and t. p. sotiriou , arxiv:0911.1299 [ hep - th ] ; m. henneaux , a. kleinschmidt and g. l. gomez , arxiv:0912.0399 [ hep - th ] ; m. visser , arxiv:0912.4757 [ hep - th ] ; t. takahashi and j. soda , phys . rev .
lett . * 102 * , 231301 ( 2009 ) ; g. calcagni , jhep * 09 * , 112 ( 2009 ) ; e. kiritsis and g. kofinas , nucl . phys .
* b821 * , 467 ( 2009 ) ; s. mukohyama , jcap * 0906 * , 001 ( 2009 ) ; r. brandenberger , arxiv:0904.2835 [ hep - th ] ; y. s. piao , arxiv:0904.4117 [ hep - th ] ; x. gao , arxiv:0904.4187 [ hep - th ] ; s. mukohyama , k. nakayama , f. takahashi and s. yokoyama , arxiv:0905.0055 [ hep - th ] ; t. ha , y. huang , q. ma , k. d. pechan , t. j. renner , z. wu and a. wang , arxiv:0905.0396 [ physics.pop-ph ] ; s. kalyana rama , phys . rev . * d79 * , 124031 ( 2009 ) ; b. chen , s. pi and j. z. tang , arxiv:0905.2300 [ hep - th ] ; x. gao , y. wang , r. brandenberger and a. riotto , arxiv:0905.3821 [ hep - th ] ; m. minamitsuji , arxiv:0905.3892 [ astro-ph.co ] ; a. wang and y. wu , jcap * 0907 * , 012 ( 2009 ) ; s. nojiri and s. d. odintsov , arxiv:0905.4213 [ hep - th ] ; y. f. cai and e. n. saridakis , jcap * 0910 * , 020 ( 2009 ) ; a. wang and r. maartens , arxiv:0907.1748 [ hep - th ] ; t. kobayashi , y. urakawa and m. yamaguchi , jcap * 0911 * , 015 ( 2009 ) ; p. wu and h. w. yu , arxiv:0909.2821 [ gr - qc ] ; c. g. boehmer and f. s. n. lobo , arxiv:0909.3986 [ gr - qc ] ; t. kobayashi , y. urakawa and m. yamaguchi , jcap * 0911 * , 015 ( 2009 ) [ arxiv:0908.1005 [ astro-ph.co ] ] ; s. maeda , s. mukohyama and t. shiromizu , phys . rev .
d * 80 * , 123538 ( 2009 ) ; s. carloni , e. elizalde and p. j. silva , arxiv:0909.2219 [ hep - th ] ; g. leon and e. n. saridakis , jcap * 0911 * , 006 ( 2009 ) ; a. wang , d. wands and r. maartens , arxiv:0909.5167 [ hep - th ] ; b. chen , s. pi and j. z. tang , arxiv:0910.0338 [ hep - th ] ; d. momeni , arxiv:0910.0594 [ gr - qc ] ; r. g. cai and n. ohta , arxiv:0910.2307 [ hep - th ] ; r. a. porto and a. zee , arxiv:0910.3716 [ hep - th ] ; s. dutta and e. n. saridakis , jcap * 1001 * , 013 ( 2010 ) [ arxiv:0911.1435 [ hep - th ] ] ; i. bakas , f. bourliot , d. lust and m. petropoulos , arxiv:0911.2665 [ hep - th ] ; x. gao , y. wang , w. xue and r. brandenberger , arxiv:0911.3196 [ hep - th ] ; e. czuchry , arxiv:0911.3891 [ hep - th ] ; r. garattini , arxiv:0912.0136 [ gr - qc ] ; m. wang , j. jing , c. ding and s. chen , arxiv:0912.4832 [ gr - qc ] ; r. g. cai and a. wang , arxiv:1001.0155 [ hep - th ] ; m. r. setare and m. jamil , arxiv:1001.1251 [ hep - th ] ; c. g. boehmer , l. hollenstein , f. s. n. lobo and s. s. seahra , arxiv:1001.1266 [ gr - qc ] ; q. j. cao , y. x. chen and k. n. shao , arxiv:1001.2597 [ hep - th ] .
e. n. saridakis , arxiv:0905.3532 [ hep - th ] ; m. i. park , arxiv:0906.4275 [ hep - th ] .
m. r. setare , arxiv:0909.0456 [ hep - th ] ; s. mukohyama , arxiv:0905.3563 [ hep - th ] .
h. nastase , arxiv:0904.3604 [ hep - th ] ; r. g. cai , l. m. cao and n. ohta , phys . rev . * d80 * , 024003 ( 2009 ) ; y. s. myung and y. w. kim , arxiv:0905.0179 [ hep - th ] ; r. g. cai , l. m. cao and n. ohta , arxiv:0905.0751 [ hep - th ] ; r. b. mann , jhep * 0906 * , 075 ( 2009 ) ; r. a. konoplya , phys . lett . * b679 * , 499 ( 2009 ) ; s. chen and j. jing , phys . rev . *
d80 * , 024036 ( 2009 ) ; j. chen and y. wang , arxiv:0905.2786 [ gr - qc ] ; s. chen and j. jing , arxiv:0905.1409 [ gr - qc ] ; m. botta - cantcheff , n. grandi and m. sturla , arxiv:0906.0582 [ hep - th ] ; a. castillo and a. larranaga , arxiv:0906.4380 [ gr - qc ] ; j. j. peng and s. q. wu , arxiv:0906.5121 [ hep - th ] .
e. o. colgain and h. yavartanoo , jhep * 0908 * , 021 ( 2009 ) ; h. w. lee , y. w. kim and y. s. myung , arxiv:0907.3568 [ hep - th ] ; s. s. kim , t. kim and y. kim , arxiv:0907.3093 [ hep - th ] ; y. s. myung , arxiv:0908.4132 [ hep - th ] ; c. ding , s. chen and j. jing , arxiv:0909.2490 [ gr - qc ] ; j. z. tang and b. chen , arxiv:0909.4127 [ hep - th ] ; n. varghese and v. c. kuriakose , arxiv:0909.4944 [ gr - qc ] ; d. y. chen , h. yang and x. t. zu , phys . lett .
b * 681 * , 463 ( 2009 ) ; e. kiritsis and g. kofinas , arxiv:0910.5487 [ hep - th ] ; t. harada , u. miyamoto and n. tsukamoto , arxiv:0911.1187 [ gr - qc ] ; m. r. setare and d. momeni , arxiv:0911.1877 [ hep - th ] ; k. izumi and s. mukohyama , arxiv:0911.1814 [ hep - th ] ; d. capasso and a. p. polychronakos , arxiv:0911.1535 [ hep - th ] ; e. kiritsis , arxiv:0911.3164 [ hep - th ] ; b. r. majhi , arxiv:0911.3239 [ hep - th ] ; j. z. tang , arxiv:0911.3849 [ hep - th ] ; r. g. cai and h. q. zhang , arxiv:0911.4867 [ hep - th ] ; j. greenwald , a. papazoglou and a. wang , arxiv:0912.0011 [ hep - th ] ; y. s. myung , arxiv:0912.3305 [ hep - th ] ; m. i. park , arxiv:0905.4480 [ hep - th ] .
h. lu , j. mei and c. n. pope , phys .
* 103 * , 091301 ( 2009 ) .
a. kehagias and k. sfetsos , phys .
* b678 * , 123 ( 2009 ) .
a. ghodsi and e. hatefi , arxiv:0906.1237 [ hep - th ] .
t. harko , z. kovacs and f. s. n. lobo , arxiv:0908.2874 [ gr - qc ] .
t. harko , z. kovacs and f. s. n. lobo , phys . rev . *
d80 * , 044021 ( 2009 ) ; l. iorio and m. l. ruggiero , arxiv:0909.2562 [ gr - qc ] ; l. iorio and m. l. ruggiero , arxiv:0909.5355 [ gr - qc ] ; a. e. mosaffa , arxiv:1001.0490 [ hep - th ] . c. g. boehmer , t. harko and f. s. n. lobo , class .
* 25 * , 045015 ( 2008 ) ; c. g. boehmer , g. de risi , t. harko and f. s. n. lobo , arxiv:0910.3800 [ gr - qc ] .
i. i. shapiro , w. b. smith , m. e. ash and s. herrick , astron .
j. * 76 * , 588 ( 1971 ) ; i. i. shapiro , c. c. counselman and r. w. king , phys . rev .
lett . * 36 * , 555 ( 1976 ) .
d. s. robertson , w. e. carter and w. h. dillinger , nature * 349 * , 768 ( 1991 ) ; d. e. lebach , b. e. corey , i. i. shapiro , m. i. ratner , j. c. webber , a. e. e. rogers , j. l. davis , and t. a. herring , phys .
lett . * 75 * , 1439 ( 1995 ) . | in the present paper we consider the possibility of observationally testing hoava gravity at the scale of the solar system , by considering the classical tests of general relativity ( perihelion precession of the planet mercury , deflection of light by the sun and the radar echo delay ) for the kehagias - sfetsos asymptotically flat black hole solution .
all these gravitational effects can be fully explained in the framework of the vacuum solution of hoava gravity , and it is shown that the analysis of the classical general relativistic tests severely constrain the free parameter of the solution . |
since shor s factorization program was proposed , many studies have been carried out in order to realize the quantum computer @xcite .
nakamura succeeded in the control of the macroscopic quantum state as the solid devices of josephson junctions@xcite .
recently , we have proposed the quantum computer in the asymmetric coupled dot in si nanocrystals@xcite .
the advantage of using the coupled dots attached by the gate electrode is that a quantum state can be controlled by the gate voltage in the solid circuits .
moreover , when the coupled si quantum dots are embedded in the gate insulator of the mosfet@xcite , the quantum state in the two coupled dots ( qubits ) is expected to be detected by the channel current . in the previous paper@xcite , we investigated the case where the sizes of si nanocystals were small such that the energy - levels in the quantum dots were discrete and the two - state system was constituted by the lowest energy - levels of the two quantum dots . here , in this paper , we consider the case where the sizes of the quantum dots are larger ( of the order of a few tens of nm ) and the energy - levels of the individual quantum dots are almost continuous@xcite .
we also assume that the charging energy of each tunneling junction is large enough that the coulomb blockade effects can be seen .
this was realized in the quantum gate of the josephson junction by shnirman@xcite and averin@xcite .
the superconducting state has the advantage from the viewpoint of the decoherence that the qubits are connected by the coherent circuits .
on the other hand , we use the normal state of the capacitively coupled quantum dots , because there exist some predictions that the decoherence time in the normal quantum dot array in semiconductors is not expected to be so short@xcite .
first , in this paper , we show that the capacitively coupled dots can be reduced to a two - state system of interacting two qubits .
next , we show the detecting mechanism of the quantum state by the channel current based on the conventional mosfet model .
asymmetry of the coupled dots is not assumed and we set @xmath0=1 .
the configuration of the quantum dots and capacitances are shown in fig.[fig1 ] .
we assume that the coupling between qubits is weaker than that in a qubit(the full formation of the electrostatic energy is shown in the appendix ) , and we take @xmath1 by this assumption , we can expand the charging energy as a function of @xmath2 and @xmath3 ( @xmath4 ) , by a small @xmath5 , and obtain the electrostatic energy : @xmath6 ^ 2 \nonumber \\
e_c\left[n_b\!+\!\frac{c_7\!-\!c_5}{c_5\!+\!c_7 } n \!-\!\frac{2c_5 c_7}{c_5\!+\!c_7 } v_b \right]^2 \label{eqn : u_1}\end{aligned}\ ] ] where @xmath7 and @xmath8 if we consider the gate voltage region where the electrostatic energy of @xmath9 state crosses that of @xmath10 state described by shnirman@xcite and averin@xcite , the electrostatic energy and the tunneling amplitude , @xmath11 , constitute the hamiltonian of the two - state system as @xmath12 where @xmath13/(c_5+c_7))$ ] and @xmath14/(c_5+c_7))$ ] .
@xmath15 and @xmath16 are pauli matrices . when the capacitances are approximated as @xmath17 where @xmath18 , @xmath19
, @xmath20 is the size of the capacitance and @xmath21 is the radius of each quantum dot , then @xmath22 is of the order of tens of mev .
the mechanism of the controlled not operation is similar to that of the josephson junction@xcite . whether @xmath23 or @xmath24 , the level - crossing gate voltage , @xmath25 , shifts(see eq.([eqn : u_1 ] ) and the controlled not operation
is realized .
the dynamical motion of the excess charge in the two - state system can be easily considered by solving the time - dependent schrdinger equation@xcite , and the excess charge shows the oscillating behavior depending on the energy of the two states .
the time period of the oscillation , @xmath26 , can be simply approximated as @xmath27 .
the polarized charged state of a qubit ( coupled dots ) behaves as the dipole moment under the electric field generated by the gate electrode . by the coupling of the dipole with the electric field
, the two - state system can be described by the bloch equation and the quantum calculation is realized similar to the nmr quantum computer@xcite .
it is well known that the one of the attractive characteristics of the nmr quantum computer is its long decoherence time . here , we comment on the decoherence time of the semiconductor dot array .
the decoherence time in semiconductor dot array has been considered to be short .
this is the reason why shnirman@xcite and averin@xcite used the josephson effects in quantum gates .
the decoherence in this case is considered mainly to originate from the phonon environments , where the interaction between the two - state system and the phonon bath is largely given by a deformation potential@xcite .
the estimated decoherence time is not so short and of the order of @xmath28 secfrom the analysis based on the model of leggett@xcite .
this relatively long decoherence time will be related to the phonon bottleneck derived by zanardi@xcite .
we will have to include the effects of the higher excited energy - levels and temperature for more detailed estimation .
the qubit which changes the charge distribution can be detected by the mosfet structure .
mosfet structure@xcite is considered to be a more efficient detecting devise in semiconductor quantum dots compared to the set structure@xcite .
this is due to the change in the charge distribution in the qubit being detected by the capacitance effects similar to those of the quantum point contact@xcite . in this section ,
we show the detailed detecting mechanism of the mosfet based on the long - channel mosfet model in the case of two qubits .
the qubit system in the mosfet proposed here can be seen as series of single coupled - dot mosfets ( fig.[fig3 ] ) . when bias , @xmath29 , is applied between the source and drain
, the depletion region expands from the source and drain such that the width of the depletion region increases toward the drain .
thus , the channel current differs depending on the positions of the qubits which change the charge distribution .
the channel current between the @xmath30-th qubit and @xmath31-th qubit is given for a small @xmath29 region as@xcite , @xmath32(v_i\!-\!v_{i-1})\!-\ !
} \alpha ( v_i^2\!-\!v_{i-1}^2 ) \right),\ ] ] where @xmath33 ( @xmath34 is the channel width , @xmath35 is the mobility , @xmath36 is the channel length of @xmath30-th qubit where we set @xmath37 , and @xmath38 is the capacitance of the sio@xmath39 ) and @xmath40 where @xmath41 is the charge within the surface depletion region .
@xmath42 is the gate voltage , and the threshold voltage , @xmath43 , is given by @xmath44 where @xmath45 ( @xmath46 is a flat band voltage , @xmath47 is the potential difference between the fermi level and the intrinsic fermi level of substrate ) , and the shift by the change of the charge distribution , @xmath48 in the @xmath30-th qubit .
then , we have the following conditions : @xmath49 in the case of two qubits , with @xmath50 ( @xmath51 is assumed ) , @xmath52 thus , whether ( @xmath53 and @xmath54 ) or ( @xmath55 and @xmath56 ) , the difference of the corresponding currents , @xmath57 is given as @xmath58 this difference can be observed in the nonlinear @xmath59-@xmath29 region and , in the pure linear region where the terms which include @xmath60 disappear , the changed qubits can not be distinguished .
we have investigated the quantum gates of the capacitively coupled quantum dot array in the mosfet structure , and have derived the two - state hamiltonian by the capacitances of the quantum dots .
we also have shown the detecting mechanism of the mosfet structure by analyzing the channel current based on the conventional mosfet model .
the author is grateful to n. gemma , s. fujita , k. ichimura , and j. koga for fruitful discussions .
in this section we show the electrostatic energy of the two qubits in terms of the capacitances of dots and the gate electrodes ( fig .
[ fig1])@xcite .
we assume @xmath61 , @xmath62 , @xmath63 and @xmath64 .
the total electrostatic energy of the two qubits is given by @xmath65 where @xmath66 shows the charge of the @xmath30-th capacitance and we have the relation between @xmath67 and the total charge of the fore dots , @xmath68 , ... , @xmath69 as @xmath70 by minimizing the energy eq.([app1 ] ) at the fixed values of @xmath25 , @xmath71 and @xmath72 , we have @xmath73 ^ 2 \!\!\!+\!\ ! \frac{c_c}{4(\!c_c c_d\!\!-\!\!(\!c_6\!\!-\!\!c_3\!)^2 ) } ( c_7 v^+)^2 \nonumber\end{aligned}\ ] ] where @xmath74 and @xmath75 , @xmath76 , @xmath77/ [ ( c_5 + c_6+\eta c_3)(c_6+c_7+\varrho c_3)-(c_6-c_3)^2]$ ] and @xmath78 . | we investigated the quantum gates of coupled quantum dots , theoretically , when charging effects can be observed .
we have shown that the charged states in the qubits can be observed by the channel current of the mosfet structure . |
we do not have a clear picture of the formation mechanism for the central bulges of spiral galaxies .
their structural and dynamical properties have long suggested the view that bulges are like small elliptical components residing in the center of a large disk .
indeed , bulges obey the d@xmath6
@xmath7 relation and fall on the fundamental plane of elliptical galaxies ( dressler _ et al .
_ 1987 , bender et al .
their surface brightness profiles fall off as @xmath8 ( andredakis _ et al .
_ 1995 ) , with @xmath9 = 4 for early - type galaxies falling to @xmath9 = 1 for sc and later types .
dynamically , bulges are consistent with oblate , isotropic models , like low - luminosity ellipticals ( kormendy & illingworth 1982 , davies _
et al . _
( 1983 ) ) . in a series of papers peletier and balcells have studied optical and near - infrared ( nir ) colour distributions of a complete sample of spiral bulges , using ground - based data ( balcells & peletier 1994 , peletier & balcells 1996 , 1997 ) with the goal of extending the comparison of bulges and ellipticals to their stellar populations .
they find that , once dust is accounted for , the colours of bulges are never redder than those of elliptical galaxies of the same luminosity .
bulge colours are very similar to those of the inner disk , the differences being much smaller than colour differences from galaxy to galaxy .
population models then suggest that the inner disk ( at 2 scale lengths ) must have formed at the same time as , or at most 3 gyr after , the bulge , and that bulge metallicities are lower than those of giant ellipticals .
hst allows us to investigate the centres of bulges with a ten fold increase in angular resolution , yet hst data on the colours of galaxy centres at scales of tens of parsecs is scarce .
colours contain useful information on the stellar populations of the galaxy centres , and allow us to estimate the level of internal extinction .
hst data have shown that dust patches are very common in galaxy nuclei , even in ellipticals ( van dokkum & franx 1995 ) , hence we expect a strong signature of dust in spiral bulges .
nicmos allows an unprecedented look at the inner structure of bulges in the h band where the extinction is approximately a factor 6 lower than in v. nicmos can produce more reliable measurements of surface brightness profiles and isophotal shapes .
here we show the results of the first colour study of the centers of galactic bulges based on hst / nicmos data .
we observed in the nicmos f160w band ( @xmath10-band ) , f814w ( @xmath11 ) and in f450w , a wider version of the @xmath12-band .
this combination of passbands is very suitable to study stellar populations in galaxies with limited amounts of extinction and recent star formation , like centres of early - type spirals . in those galaxies
the observed colours are determined by the underlying old stellar population , perturbed by dust extinction and the light of young , recently formed stars .
extinction by dust affects all optical and near - infrared colours , while recent star formation can only be seen in blue passbands ( see e.g. knapen et al .
1995 ) . in that paper
it is shown that a red colour , like @xmath2 will primarily show the extinction due to dust , as well as old stellar population gradients , while a colour which includes a blue band like @xmath12 or @xmath13 , in this paper @xmath1 , is specially sensitive to recent star formation . in the absence of extinction , b , i and h make it possible , in principle , to measure the age and metallicity of the stellar population ( e.g. aaronson et al .
1978 ; peletier et al . 1990 ; bothun & gregg 1990 ) .
we infer large amounts of centrally concentrated dust within the inner 100 pc . outside this region ,
the colour - colour distribution is quite tight , allowing us to place limits on the age dispersion of galactic bulges . with the hst it is easier to reach a high photometric accuracy , because there is no atmospheric extinction to correct for , the instrumental psf is stable , and the instrument is very well characterized , as compared to ground - based instruments .
this means that we can expect accuracies of 0.020.03 mag or better in optical and near - infrared colours .
such high quality measurements are vital if we are to separate the effects of age and metallicity using colours , even though the large wavelength baseline ( @xmath12-@xmath11-@xmath10 ) minimises the effect of photometric errors .
currently there are three main scenarios for the formation of bulges : bulges form before , contemporaneously with , or after disks ( see the review by wyse _
_ 1997 , and bouwens _ et al .
_ 1999 ) . in the first scenario ( eggen et al .
1962 , larson 1975 , carlberg 1984 ) the formation of bulges is mainly described by collapse of a primordial gas cloud into clumps , which then merge together .
the disk only forms after the last massive merger via gas infall . in a second scenario ,
infall of a gas - rich dwarf galaxy onto a disk produces star formation more or less at the same time in bulge and disk ( pfenniger 1992 ) . in the third scenario of secular evolution of disks ( e.g combes _ et al .
_ 1990 , pfenniger & norman 1990 , norman _ et al .
_ 1996 ) , the bulge is formed by dynamical instabilities of the disk , in which a bar is formed , which then forms a massive central concentration by allowing material to stream towards the central regions . the central mass concentration itself
then tidally disrupts the bar and forms the bulge .
when material is continuing to fall onto the disk sometime later a new bar can be formed , and the whole process will start again and induce more star formation in the bulge . to distinguish between these models one needs to measure the ages of both bulges and disks .
peletier & balcells ( 1996 ) and terndrup _ et al .
_ 1994 , using ground - based data , found that the ages of bulges and inner disks are very similar .
here in this paper we measure the ages of the bulges themselves .
the paper is organized as follows : section 2 gives details of the observations and data reduction .
section 3a discusses the dust content of the galactic bulges .
section 3b describes the analysis of the stellar populations , and in section 4 some implications for our understanding of galactic bulges are discussed .
twenty galaxies were observed with hst in cycle 7 ( summer 1997 ) with wfpc2 ( f450w and f814w ) and nicmos ( f160w ) .
they all are part of the original sample of balcells & peletier ( 1994 ) , which is a complete , @xmath12-magnitude - limited sample of early - type spirals ( type s0-sbc ) with inclinations larger than 50@xmath14 , and for which one side of the minor axis colour profile is approximately featureless as seen from the ground .
the subsample considered here was chosen to include galaxies of types s0-sbc , excluding galaxies that are exactly edge - on , and is biased towards the nearest objects .
the galaxies observed are given in tables 1 and 2 .
their basic parameters : luminosities , effective radii and ground - based colours , are given in andredakis _
et al . _
( 1995 ) and peletier & balcells ( 1997 ) .
the mean absolute @xmath15-band magnitude of the sample is @xmath16 = 22.50 mag , with an rms dispersion of 1 mag ( h@xmath17 = 70 km s@xmath18 mpc@xmath18 ) .
the standard hst - pipeline data reduction was used for the wfpc2 optical data . for ngc 7457
the @xmath11-band observations were taken from the hst archive .
for the @xmath10-band two multiaccum nic2-exposures of in total 256 s were taken , offset from each other by 1@xmath19 , to enable us to remove bad pixels .
the standard stsdas calnica reduction package was used , after which a small fraction of the flatfield was subtracted from the reduced images , to account for a non - zero pedestal level in the raw image .
this fraction was determined by requiring that the final image was smooth , and amounts to about 0.3 - 0.5 adu s@xmath18 , corresponding to about @xmath10=17 mag ( arcsec)@xmath20 . applying this step is only important for the lowest flux levels , but the images improve considerably in quality ( they look like the ground - based images ) .
the final sky background level was determined by comparing the image with the ground - based image in the same band . to compare the @xmath10-band data with the ground - based data in @xmath21 , we used the following transformation , derived from the new gissel96 models of bruzual & charlot ( see leitherer _
1996 ) : @xmath22 the images were calibrated to the stmag system using holtzmann _
et al . _
( 1995 ) , using a constant shift of 0.10 mag to correct to infinite aperture .
the method described in the same paper was used to iteratively account for the colour term in the f450w filter .
the colours were corrected for galactic extinction using the new dust maps of schlegel _ et al . _
( 1998 ) , and the galactic extinction law ( rieke & lebofsky 1985 ) . after this
the bands were k - corrected . this small correction ( from persson _
et al . _
1979 ) was @xmath23 = -5@xmath24 , @xmath25 = @xmath24 and @xmath26 = 0 . to account for the difference in point spread functions ( psf ) when determining the @xmath27
@xmath28 colour maps and profiles , the @xmath27-band image was convolved with the @xmath28-band psf , and the @xmath28-band image with the @xmath27-band psf .
these psfs were determined with the tinytim package ( krist 1992 ) .
we then determined minor axis profiles by averaging azimuthally in wedges of 22.5@xmath14 , centered on the @xmath10-band nucleus .
we then combined the surface brightness profiles with the ground - based profiles to cover the whole spatial range of the bulge , by shifting the ground - based profile on top of the hst - profile between about 3@xmath19 and 6@xmath19 . this way the final profiles have high signal - to - noise everywhere , and the high - quality photometric accuracy of hst data , for which absolute and relative accuracies of a few percent can be expected ( colina _ et al .
we then subtracted these wedge profiles in pairs of bands to obtain colour profiles . in table 1
we list the galaxies and their colours at the center and at one r@xmath3 .
ccccccc & t & r@xmath3 & @xmath29 & @xmath30 & @xmath31 & @xmath32 + 5326 & 1 & 2.16 & 2.510 & 2.089 & 2.117 & 1.796 + 5389 & 0 & 4.67 & 2.992 & 2.082 & 2.208 & 1.823 + 5422 & -2 & 4.51 & 2.581 & 2.092 & 2.170 & 1.809 + 5443 & 3 & 4.70 & 2.894 & 1.941 & 2.408 & 1.554 + 5475 & 0 & 2.58 & 2.342 & 2.065 & 2.014 & 1.774 + 5577 & 4 & 4.14 & 1.910 & 1.862 & 1.792 & 1.665 + 5587 & 0 & 2.13 & 2.600 & 1.993 & 2.086 & 1.840 + 5689 & 0 & 3.72 & 2.355 & 2.058 & 2.231 & 1.816 + 5707 & 2 & 3.40 & 2.720 & 2.069 & 2.144 & 1.722 + 5719 & 2 & 3.88 & 3.494 & 2.252 & 2.758 & 1.972 + 5746 & 3 & 14.6 & 4.117 & 2.040 & 2.910 & 1.749 + 5838 & -3 & 7.50 & 2.600 & 2.091 & 2.184 & 1.767 + 5854 & -1 & 6.03 & 1.872 & 1.840 & 1.821 & 1.659 + 5879 & 4 & 2.29 & 2.653 & 1.932 & 2.353 & 1.704 + 5965 & 3 & 8.23 & 2.946 & 2.051 & 2.406 & 1.755 + 6010 & 0 & 2.43 & 2.543 & 2.086 & 2.105 & 1.843 + 6504 & 2 & 4.98 & 2.452 & 2.064 & 2.062 & 1.764 + 7331 & 3 & 5.0 & 2.356 & 2.047 & 1.978 & 1.766 + 7457 & -3 & 8.55 & 1.951 & 1.920 & 1.823 & 1.620 + 7537 & 4 & 1.48 & 2.928 & 1.948 & 2.560 & 1.969 + + 4472 & & & & 2.357 & & 1.947 + * notes to table 1 : * the morphological types in column ( 2 ) are from the rc3 ( de vaucouleurs et al .
the effective radii of the bulges in column ( 3 ) in arcseconds are from andredakis et al .
( 1995 ) , except for ngc 7331 , where the value is an estimate based on prada et al .
distances in kpc are given in andredakis et al .
( 1995 ) . also given
are the colours of ngc 4472 at 5@xmath19 ( see text ) .
in fig . 1 we show intensity - maps of the inner regions of the galaxies , together with @xmath1 and @xmath2 colour maps with the same scale and orientation , superimposed on @xmath10-band contour maps .
the maps indicate the position of dust lanes .
it is clear from the figure that many , maybe all , galaxies have red nuclei . to show this
more clearly we show all colour profiles obtained in wedges along the dust free side of the minor axis in fig .
this is not the first time that red nuclei have been found in spiral galaxies .
for example , in our ground - based data ( peletier & balcells 1997 ) we show several galaxies with red nuclei .
for that sample however the colour profiles on the least dusty side are generally featureless at radii larger than 1@xmath19 , showing the same logarithmic colour gradients throughout the bulge . inside 1@xmath19 no information is available about the colours from ground - based data . the fact that the psf of the hst data is well known and stable enables us to correct the colour - profiles and maps for most of the instrumental psf effects . in this way one can measure colours down to radii as small as @xmath33 0.10@xmath19 ( the diffraction limit is 0.15@xmath19 ) . at this resolution
the red nuclei are easy to see , and extended in most of our sample .
nuclei can be red because of local dust , large foreground dust lanes , or because of red stellar populations .
one usually can distinguish between dust and stellar population reddening by looking for patchy structures , since they are generally due to dust extinction , or to star formation , in combination with extinction . in table
2 we indicate the nature of the features that we find in the nuclei
. .nuclear features [ cols="<,^,^,^,^,^ , < " , ] * notes to table 2 : * a + sign in column ( 2 ) indicates the presence of a large foreground disk , in column ( 3 ) the presence of a red nuclear dust patch or lane , and in column ( 4 ) the presence of a blue patch near the nucleus . iras fluxes are presented in column ( 5 ) and ( 6 ) .
mdsf in column ( 4 ) means mixed dust and star formation in the whole nuclear area .
the lgc number is the group designation ( by garcia 1993 ) .
we find ( table 2 ) only three galaxies : ngc 6010 , 6504 and 7457 , without nuclear dust features , although in some cases a large foreground dust lane makes nuclear dust difficult to detect .
for these three galaxies we analysed the isophote shapes more carefully .
dust patches make galaxy isophotes irregular in particular the third order fourier terms c3 and s3 ( carter 1978 ) will be non - zero , and generally increasing in amplitude to the blue .
s3 and c3 were found to be significantly non - zero in both @xmath12 and @xmath11 in ngc 6010 and 6504 but not in ngc 7457 .
we conclude that the signature of dust is to be found in almost all galaxies in the sample , including some galaxies indicated with a ( ) sign in column 3 of table 2 .
how much reddening is caused by the dust ?
3 we show the colours at the center ( filled dots ) and at one bulge effective radius ( open dots ) in a @xmath2 vs. @xmath1 colour - colour diagram .
3 ( b ) and ( c ) show that in all cases the galaxy is redder in the centre than at r@xmath34 , by sometimes very large amounts .
since the vector indicating reddening by dust is almost parallel to the the vector indicating changes in metallicity , it is not possible to say exactly how much of the reddening is due to extinction .
lower and upper limits to the internal extinction may be estimated as follows . a plausible _ upper limit _ to the reddening
is derived by assuming that the galaxy at 1 r@xmath3 is dustfree .
this is perhaps reasonable , since at these radii no structure is seen in the colour maps , and there is a small dispersion in the colour - colour plot .
we infer that the central @xmath2 and the @xmath1 colours are reddened by an average of 0.42 @xmath35 0.06 mag , and 0.61 @xmath35 0.11 mag respectively . a
_ lower limit _ to the reddening can be found if one assumes that the stellar populations are never redder than those of the central virgo galaxy ngc 4472 . using the data at 5@xmath19 ( or @xmath36 400 pc ) from peletier _ et al . _
( 1990 ) for ngc 4472 , converted to @xmath1 and @xmath2 as described in the appendix , we find that the @xmath2 colours would be reddened by on the average 0.37 @xmath35 0.06 mag and @xmath1 by 0.41 @xmath35 0.10 mag .
using the galactic extinction law ( rieke & lebofsky 1985 ) these numbers correspond to an average internal extinction a@xmath0 between 0.89 and 1.01 mag , if @xmath2 is used , and between 0.56 and 0.78 mag when one uses @xmath1 .
the estimates using @xmath2 are somewhat higher than those from @xmath1 , because nuclear star formation in some galaxies makes the radial colour difference in @xmath1 smaller , as is seen in ngc 5838 , 5854 and 5475 . in the @xmath1 colour map of these galaxies there
are blue nuclear features which do not appear in @xmath2 .
these patches , presumably caused by young stars , reduce the radial colour variations in blue colours like @xmath1 but not in @xmath2 .
our _ upper limits _ for the amount of dust inferred here are probably somewhat too high , because part of the reddening is caused by stellar population gradients . from the most dustfree galaxies we infer that the colour difference due to stellar populations between center and 1 r@xmath3 is @xmath33 0.1 mag in both @xmath1 and @xmath2 , which implies that colour gradients in the inner parts of bulges are affected much more by dust than by stellar populations . apart from providing high resolution
, hst also has the advantage that the photometric conditions are very stable , so that accurate colours can be determined .
for this reason , and because we have a very large colour - baseline , we can use the colour - colour diagram to infer information about the age - spread in the sample , and about the cause of the stellar population gradients . studying fig . 1
carefully , we see that many physical phenomena are playing a role in these galaxies .
we observe the combined effects of extinction , recent star formation , and old stellar population gradients . to disentangle them
we have displayed the galaxies in fig . 1 according to their morphological type .
although the type ( from de vaucouleurs _ et al . _ 1991 ) has been given based on low - resolution optical observations , we can see that the properties of the galaxies change smoothly as a function of type .
s0 galaxies have small dustlanes , and have rather featureless colour maps .
sb galaxies tend to have strong dustlanes , while the sbc galaxies are considerably bluer , have lower surface brightness , show patchy dust and star formation together , and are rather different from the rest of the galaxies . having established ( in section 3.1 ) that extinction at 1 r@xmath3 is probably negligible , we will now consider what the colours can tell us about the stellar populations .
first we note the small scatter amongst the open symbols in fig .
3a confirming that extinction is not important at 1 r@xmath3 .
if we exclude the three galaxies with the latest hubble type ( the open crossed symbols in fig .
3c ) the stellar populations at r@xmath3 form a rather tight sequence in the @xmath1 vs. @xmath2 plane .
1e shows that the central regions of our 3 sbc are full of dust and regions of recent star formation , we will not consider them further in this analysis of stellar populations . in fig .
3a we also show a set of single stellar population ( ssp ) models of vazdekis _ et al . _
the models displayed here have a salpeter imf , with a reduced number of low - mass stars , to match better the scalo ( 1986 ) imf .
we see that , independent of the amount of extinction , the tightness of the colour - colour relation shows that the luminosity weighted age of the stars is very similar from bulge to bulge .
according to the models the age - spread would be about 1 - 2 gyr , although the bluest ( and generally faintest ) bulges would be somewhat younger .
we derive a mean luminosity weighted age of 9 gyr from these models . while the absolute ages are poorly constrained
, the relative ages are much more robust so we can conclude that the age spread amongst bulges in this sample is small .
( if we use worthey ( 1994 ) s models the inferred luminosity weighted age would be implausibly small : 2 gyr ) .
the galaxy sequence runs parallel to line of constant age suggesting that the colour - variations from galaxy to galaxy are due predominantly to changes in metallicity supporting the view that the colour - magnitude relation for early - type galaxies ( e.g. bower _ et al .
_ 1992 ) is mainly driven by changes in metallicity .
figure 3c shows that the type dependence along the colour - colour relation is very small .
the outlying point is ngc 5854 which has a blue central patch ( probably of recent star formation ) .
ngc 7457 is a faint s0 galaxy that has bluer @xmath37 and @xmath38 colours than an elliptical of comparable luminosity ( balcells & peletier 1994 ) accounting for its rather young inferred age .
the bulk of the galaxies of type sb and earlier occupy a narrow band in @xmath1 vs. @xmath2 . as a comparison we have plotted the colour - colour relation for coma , determined from the data of bower etal ( 1992 ) , converted to our colours as described in the appendix .
we find that the colours of our bulges at 1 r@xmath3 are very similar to those of bright coma galaxies .
our bulges however are somewhat bluer in @xmath2 and redder in @xmath1 , which , according to the stellar population models , could be explained if they are slightly older and somewhat less metal rich .
the colour - colour conversion is rather uncertain , however .
also plotted in fig .
3c are the colour - colour diagrams converted using the theoretical models of vazdekis _ et al . _
( 1996 ) and worthey ( 1994 ) . as can be seen in the figure ,
the difference between the three colour - colour diagrams in @xmath1 and @xmath39 is quite large .
although we argue ( see appendix ) that our empirical calibration can be trusted much more than the theoretical calibrations , one should be careful not to over - interpret the data .
what seems safe to conclude is that coma galaxies are to be found in the shaded area , and that the ages of our bulges are similar to those of the early - type galaxies in coma .
in the previous section we have shown that : 1 .
centers of bulges of early - type spirals are generally dusty . we find that a@xmath0 on the average lies between 0.6 and 1.0 mag , which implies that a@xmath40 should be between 0.1 and 0.2 mag . 2 . at 1 r@xmath3 galactic bulges show a very tight @xmath2 vs. @xmath1 relation , implying that the age spread among bulges of early - type spirals is small ( at most 2 gyr ) .
the colours of bulges of early - type spirals at 1 r@xmath3 are similar to the colours of early - type galaxies in the coma cluster .
this implies that the age difference between nearby bulges and cluster - ellipticals is small , probably smaller than about 2 gyr .
since the extinction observed , is more than a factor 2 ( 0.76 mag ) , if the dust is located in or near the center we can not see through to the other side of the galaxy .
the extinction locally will be very large ( see sadler & gerhard 1985 ) . the fact that this occurs in so many of our galaxies shows that the very central region is almost always optically thick in @xmath12 or @xmath41 .
the situation with dust in the centers of spiral bulges is similar to that of elliptical galaxies .
van dokkum & franx , analyzing wfpc data of nearby elliptical galaxies , found evidence for central extinction in 75% of their sample .
this is more than was previously found from the ground ( ebneter _ et al .
_ 1988 , vron - cetty & vron 1988 ) .
et al . _
( 1994b ) detected dust in 41% of the galaxies in their sample of revised shapley - ames ( sandage & tammann 1981 ) galaxies .
a similar detection rate as found from iras fluxes by knapp _
et al . _
. the higher angular resolution of hst allows us to reach much lower detection limits , we have been able to detect dust in 95 @xmath35 5% of our sample of bulges of early - type spirals . this detection rate is higher than for ellipticals , although our method of finding dust , using colour images , is more sensitive than the method of van dokkum & franx , who determined their dust masses from one @xmath41-band image only .
van dokkum & franx found an average dust mass of 4 @xmath42 10@xmath43 m@xmath44 .
for the bulges analysed here we find an average dust mass of @xmath36 10@xmath45 m@xmath44 , determined using the method of van dokkum & franx : for each dust feature the mass is calculated using m@xmath46 = @xmath47 @xmath48 @xmath49 , with @xmath47 the area of the feature , @xmath48 the mean absorption in the area , and @xmath50 the visual mass absorption coefficient ( sadler & gerhard 1985 ) .
@xmath50 is taken to be 6 10@xmath51 mag kpc@xmath52 m@xmath53 .
rough values of @xmath48 were determined by taking the difference in @xmath1 or @xmath2 of the feature and the dustfree values at 1 r@xmath3 , converted to a@xmath0 using the galactic extinction law ( rieke & lebofsky 1985 ) . if we assume a galactic gas to dust ratio of 130 ( see van dokkum & franx 1995 ) we find the galaxies analysed in this paper have an average of 10@xmath54 m@xmath44 of interstellar material in their nuclear regions .
the origin of the nuclear dust is unclear . at larger
radii kinematic observations of gas , usually associated with dust , in ellipticals show that it is often decoupled from the stellar velocity field ( for a discussion see goudfrooij _
et al . _
this is used to imply that _ large - scale _ dust lanes are of external origin perhaps being accreted during galaxy mergers or interactions .
however , the origin of the small arcsecond - scale dust lanes found in ellipticals , often oriented along the major axis of the stellar body ( goudfrooij _ et al . _
1994b ) , could well be internal .
scaling from the numbers for stellar mass - loss for bright ellipticals given by faber & gallagher , 1976 , we estimate that for typical bulges in our sample the mass will be dposited into the bulge at the rate of @xmath33 0.1 1 m@xmath44 yr@xmath18 , so there is no problem accruing the dust we see .
the amount of dust , the large detected fraction , and the fact that the dust lanes are found parallel to the major axis indicate an internal origin .
the central dust provides a suitable environment for centrally concentrated star formation , probably leading to strong metallicity enhancements in the central 100 pc of bulges . in that regime , it is possible that so much dust is produced that ordinary dust destruction mechanisms are ineffective .
potentially , high resolution spectral observations will show us whether this dust is indeed of internal origin , and what metallicities that are being reached .
the dust however could provide some serious observational difficulties for the determination of the inner slope of the stellar surface density profile of bulges , even in the h band .
several of our bulges have central features which resemble the inner disk in the giant elliptical ngc 4261 ( jaffe _ et al . _
examples are ngc 5326 , ngc 5587 , ngc 5838 and ngc 5854 .
ngc 4261 has a liner spectrum in the center , jaffe _
et al . _ argue that this inner disk might provide the fuel for the agn .
only 4 of the objects studied in our paper however are known to be ( weakly ) active , ngc 5746 , ngc 5838 , ngc 5879 and ngc 7331 , all classified by ho _
et al . _
( 1997 ) as t2 , transition objects between liner and seyfert , with narrow emission lines .
their colours are however entirely consistent with an extincted stellar population .
furthermore , many of the ellipticals with similar features are not active galaxies .
it seems that an inner disk of gas and dust alone is not sufficient to produce an agn .
no hst colour profiles of bulges of spirals have appeared in the literature up to now .
the only paper presenting high - resolution hst colour profiles of ellipticals is carollo _ et al . _
( 1997a ) , who report v and i profiles of 15 galaxies with dynamically decoupled cores using the refurbished wfpc2 .
they find that their galaxies all have very similar @xmath55 gradients between radii of 1.5@xmath19 and 10@xmath19 , while the dispersion is larger between 0.25@xmath19 and 1.5@xmath19 , where some galaxies are seen with gradients that are twice as large , while some others are much smaller , or even negative .
clearly the behavior of @xmath55 is different in their inner 1.5@xmath19 as compared to the area further out .
although carollo _ et al . _
( 1997a ) masked out dusty areas before obtaining the colour profiles by radially averaging the remaining light , this process will probably not have removed all the extinction .
the large spread in colour gradient in the inner region might be a confirmation of significant quantities of dust near the center in almost all elliptical galaxies .
what can we learn about the formation of bulges ?
the fact that the ages of most of the bulges in this paper are so similar and old makes it very difficult to form the bulges of early - type spirals ( s0-sb ) through secular evolution of disks . in this scenario
it is expected that bulges regularly undergo major bursts of star formation , to convert gas that has been funneled to the central area through the presence of a bar into stars .
this would mean that we would expect to find more young bulges and a large spread in bulge ages .
another problem for early - type bulges is that the bulge densities are up to a factor 100 higher than in the center of the disks . to create such high overdensities in the disk ,
the required star formation rates are such that they could easily disrupt the disk ( ostriker 1990 ) .
late type galaxies ( sbc s and later types ) might be different .
we see that the three in our sample have younger ages ( crossed circles in fig .
3 ) , and also the density contrast between bulge and disk is much smaller . for these galaxies in the bulge region
spiral arms , blue star formation regions and dust lanes are seen , as in the disk , telling us that the stellar populations of bulge and disk here are very much the same ( see fig .
although norman _
( 1996 ) use the fact that the fundamental plane has changed little from a redshift of the order of 0.5 to now as an argument against secular - evolution driven bulge building , it is not clear whether bulges of late - type galaxies really lie on the fundamental plane of ellipticals and early - type bulges , since no data is available at present . the fact that the ages of our bulges are all so similar supports the idea of a major episode of star formation in the past , in which most of these bulges ( and also the bright galaxies in large clusters like coma ) were formed .
it is currently thought that cluster ellipticals must have formed at redshifts beyond @xmath24 = 3 ( ellis _ et al .
_ 1997 , stanford _
et al . _
1997 ) , because of the lack of evolution seen in clusters of intermediate redshift .
this corresponds to an age of 10.5 gyr ( @xmath56 = 0.2 , @xmath57 = 0.8 , h@xmath58 = 70 km s@xmath18 mpc@xmath18 ) , or 8.4 gyr ( @xmath56 = 1 , @xmath57 = 0 , h@xmath58 = 70 km s@xmath18 mpc@xmath18 ) ( hogg 1999 ) .
the fact that the colours of the majority of our bulges are similar to the coma galaxies indicates that our bulges ( except maybe the sbc s ) are also old , and formed at redshifts beyond @xmath24 = 3 .
these observations endorse the original model by eggen _
et al . _ ( 1962 ) forming bulges in the beginning during a monolithic , or clumpy , collapse . the fact that the colours of bulges and inner disks are very similar ( terndrup _ et al .
_ 1994 , peletier & balcells 1996 ) then implies that the disk formed very gradually from inside out , with the age of the inner disk similar to the age of the bulge .
can we then say something about the age of the disk ?
would it be possible that the whole disk formed at the same time as the bulge ?
at that point a conflict arises with the star formation history of the universe derived e.g. by madau _
et al . _
( 1996 ) on the basis of data of the hubble deep field , which shows a maximum between @xmath24 = 1 and 3 .
since our bulges are found in a variety of environments , they are in no way special , and if most early - type spirals would form at redshifts beyond 3 this would also imply that the maximum in the hdf would have to go to larger z. however , this problem would be solved if disks of early - type galaxies on the average would be considerably younger than their bulges . with a sufficiently large age difference between bulge and disk the luminosity - weighted age of the galaxy ( which is in general dominated by the disk ) can then be young enough .
there is nothing in the data of terndrup _ et al .
_ 1994 and peletier & balcells 1996 preventing this from happening . alternatively , kauffmann ( 1995 ) and governato _ et al . _
( 1998 ) have pointed out that in biased models of galaxy formation accelerated evolution can be expected in dense regions .
our observations here show that there is no difference in age between the ellipticals in coma and our galactic bulges .
this would mean that our bulges , except for the latest types , must have formed early , beyond @xmath24 = 3 , as well .
on the other hand , we also do not find any dependence of bulge age upon environment .
the bulges are found in all kinds of environments , ranging from isolated to groups of more than 20 members ( see table 1 ) . apparently everywhere in the universe
these intermediate - size galaxies must have started forming early - on .
it appears that the only feasible solution is that bulges of early - type spirals are old , and disks considerably younger .
this is in agreement with abraham _ et .
( 1998 ) , who find for the disk galaxies in the hdf that their bulges are significantly older than their disks ( up to 50% ) .
one has however to take into account that at z @xmath36 1 many large spiral galaxies are found ( lilly _ et al . _
( 1998 ) , whose stellar populations are consistent with a declining activity since at least z=1.5 2 ( abraham _ et al .
semi - analytic galaxy formation models also support this picture . in fig . 4 a histogram is shown of the distribution of @xmath41-band averaged ages of ellipticals and bulges in the simulations of baugh _ et al . _
( 1996 ) ( baugh , private communication ) .
the models find that the large majority of bulges is old .
this is independent of the environment , contrary to elliptical galaxies , which are found to be old in rich clusters , and can be 5 gyrs younger in the field .
bulges are old because the accompanying disk needs time to form without being disrupted .
we have found that bulges of early - type spirals in general are old , that their age - spread is smaller than 2 gyr , and that their colour gradients are mainly due to metallicity gradients in the stellar populations .
these results once more support the idea that bulges and ellipticals are similar objects .
the good agreement between the observational results of this paper and models like the semi - analytical models of baugh _ et al . _
( 1996 ) strongly supports the picture that bulges were formed through monolithic or clumpy collapse , and formed much before theier disks .
however , one can see that bulges of sbc galaxies are different in many respects , and it is likely that these differences will be larger for sc galaxies and later types ( see also the recent hst - study of carollo _ et al . _
( 1997b , 1998 ) .
they are smaller , younger , have lower central surface brightness , and it is not clear whether they fall on the fundamental plane of elliptical galaxies . since the observations until now are not good enough to establish whether these late type bulges can be made from bars , and in this way the galaxy type can be changed by secular evolution , it is clear that it is very important to study the bulges of late type spirals to understand the formation of bulges .
this paper is based on observation with the hubble space telescope the authors acknowledge very useful support from luis colina , massimo stiavelli , jeremy walsh and doris daou at the stsci and st - ecf , and useful discussions with ian smail , harald kuntschner , carlton baugh and carlos frenk .
rld is grateful to durham university for the award of a sir james knott fellowship and to the leverhulme trust for the award of a research fellowship , these awards contributed significantly to this research .
aaronson , m. , cohen , j.g . ,
mould , j. & malkan , m. , 1978 , apj , 223 , 824 aaronson , m. , bothun , g. , mould , j. , huchra , j. , schommer , r.a . & cornell , m.e . , 1986 ,
apj , 302 , 536 abraham , r.g . , ellis , r.s . , fabian , a.c . , tanvir , n.r . & glazebrook , k. , 1998 , mnras , in press ( astro / ph 9807140 ) andredakis , y. , peletier , r.f . &
balcells , m. , 1995 , mnras , 275 , 874 balcells , m. & peletier , r.f . , 1994 , aj , 107 , 135 baugh , c.m . ,
cole , s. & frenk , c.s .
, 1996 , mnras , 282 , l27 baugh , c.m .
, cole , s. , frenk , c.s . & lacey , c.g .
, 1998 , apj , 498 , 504 bender , r. , burstein , d. & faber , s.m . , 1992 ,
apj , 399 , 462 bouwens , r. , cayon , l. & silk , j. , 1999 , apj , in press bothun , g.d . & gregg , m.d . , 1990 , apj , 350 , 73 bower , r.g . ,
lucey , j.r . & ellis , r.s . , 1992 , mnras , 254 , 589 carlberg , r.g . , 1984 , apj , 286 , 403 carollo , c.m . ,
franx , m. , illingworth , g.d .
& forbes , d. , 1997a , apj , 481 , 710 carollo , c.m . ,
stiavelli , m. , de zeeuw , p.t . & mack , j , 1997b , aj , 114 , 2366 carollo , c.m . , stiavelli , m. & mack , j. , 1998 , aj , 116 , 68 carter , d. , 1978 , mnras , 182 , 797 charlot , s. , worthey , g. & bressan , a. , 1996 , apj , 457 , 625 colina , l. , holfeltz , s. & ritchie , c. , 1998 , in _ nicmos and the vlt _
w. freudling and r. hook , 1998 , eso , p. 36 combes , f. , debbasch , f. , friedli , d. , & pfenniger , d. , 1990 , a&a , 233 , 82 davies , r.l . , efstathiou , g. , fall , s.m . ,
illingworth , g. & schechter , p.l .
, 1983 , apj , 266 , 41 de vaucouleurs , g. , de vaucouleurs , a. , corwin , h.g . ,
buta , r.j . , paturel , g. , fouqu , p. 1991
, 3rd reference catalogue of bright galaxies ( rc3 ) , springer , new york djorgovski , s. & davis , m. , 1987 , apj , 313 , 59 dressler , a. , lynden - bell , d. , burstein , d. , davies , r.l . ,
faber , s.m . ,
terlevich , r. & wegner , g. , 1987 , apj , 313 , 42 ebneter , k. , davis , m. & djorgovski , s. , 1988 , aj , 95 , 422 ellis , r. , smail , i. , dressler , a. , et al . , 1997 ,
apj , 483 , 582 _ et al .
_ , esa publications ( astro / ph 9808025 ) faber , s.m . & gallagher , j.s .
iii , 1976 , apj , 204 , 365 garcia , a. , a & as , 100 , 47 goudfrooij , p. , hansen , l. , jrgensen , h.e .
, nrgaard - nielsen , h.u .
, de jong , t & van den hoek , l.b . , 1994a , a&as , 104 , 179 goudfrooij , p. , hansen , l. , jrgensen , h.e . &
nrgaard - nielsen , h.u .
, 1994b , a&as , 105 , 341 governato , f. et al . , 1998 , nature , 392 , 359 ho , l. , filippenko , a.v .
& sargent , w.l.w . , 1997 , apjs , 112 , 315 hogg , d.w . , astro - ph/9905116 holtzmann , j.a . , burrows , c.j . ,
casertano , s. , hester , j.j . ,
trauger , j.t . ,
watson , a.m. & worthey , g. , 1995 , pasp , 107 , 1065 jaffe , w. , ford , h. , ferrarese , l. , van den bosch , f. & oconnell , r.w .
, 1996 , apj , 460 , 214 kauffmann , g. , 1995 , mnras , 274 , 153 kauffmann , g. , 1996 , mnras , 281 , 487 knapen , j.h .
, beckman , j.e . , shlosman , i. , peletier , r.f . ,
heller , c.h . & de jong , r.s . , 1995 , apj , 443 , l73 knapp , g.r . ,
guhathakurta , p. , kim , d .- w . &
jura , m. , 1989 , apjs70 , 329 kormendy , j. & illingworth , g.d . , 1982 , apj , 256 , 460 krist , j. 1992 , tinytim v2.1 user s manual ( stsci ) larson , r.b . , 1975 , mnras , 173 , 671 leitherer , c. , et al . , 1996 , pasp , 108 , 996 lilly , s. , et al . , 1998 ,
apj , 500 , 75 madau , p. , ferguson , h.c . , dickinson , m. et al .
, 1996 , mnras , 283 , 1388 norman , c. , sellwood , j.a . & hasan , h. , 1996 , apj , 462 , 114 ostriker , j.p . , 1990 , in _ evolution of the universe of galaxies _ , asp conf
r. kron ( asp , san francisco ) , p. 25
peletier , r.f . ,
valentijn , e.a . & jameson , r.f . , 1990 ,
a&a , 233 , 62 peletier , r.f . & balcells , m. , 1996 , aj , 111 , 2238 peletier , r.f . & balcells , m. , 1997 , newa , 1 , 349 persson , s.e . ,
frogel , j.a . & aaronson , m. , 1979 , apjs39 , 61 pfenniger , d. , 1992 , in proc .
153 , _ galactic bulges _ , ed . h. dejonghe & h.j .
habing , kluwer , dordrecht , p. 387 pfenniger , d. & norman , c. , 1990 , apj , 363 , 391 prada , f. , gutirrez , c. , peletier , r.f . & mckeith , c.d . , 1996 ,
apj , 463 , l9 rieke , g. & lebofsky , m.j . , 1985 ,
apj , 288 , 618 sadler , e.m .
& gerhard , o.e . , 1985 , mnras , 214 , 177 sandage , a.r .
, tammann , g. , 1981 , _ a revised shapley - ames catalog of bright galaxies _ , carnegie institute of washington scalo , j.m . , 1986 ,
cosmic phys , 11 , 1 schlegel , d.j , finkbeiner , d.p .
& davis , m. , 1998 , apj , in press stanford , s.a . ,
eisenhardt , p.r . & dickinson , m. , 1997 , apj , 492 , 461 terndrup , d.m . , davies , r.l . ,
frogel , j.a . ,
depoy , d.l . & wells , l.a .
, 1994 , apj , 432 , 518 van dokkum , p.g . & franx , m. , 1995 , aj , 110 , 2027 vazdekis , a. , casuso , e. , peletier , r.f .
& beckman , j.e .
, 1996 , apjs , 106 , 307 vron - cetty , m.p . & vron , p. , 1988 ,
a&a , 204 , 28 wyse , r.f.g .
, gilmore , g. & franx , m. , 1997 , araa , 35 , 637 worthey , g. , 1994 , apjs , 95 , 107
in this appendix we describe how a colour - colour relation for the coma cluster was obtained in @xmath1 and @xmath2 .
we started with the excellent data of bower et al .
( 1992 ) in @xmath13 , @xmath41 and @xmath21 . as a first attempt we tried to make the colour conversions using single burst theoretical models for old stellar populations using the salpeter imf .
a fit to the models of vazdekis et al .
( 1996 ) gives there are several reasons to be wary of this approach .
in the first place one knows ( e.g. charlot et al .
1996 ) that the infrared colours of the worthey models for large metallicities are too red by as much as a magnitude .
secondly , the model colours of vazdekis et al . in the infrared , just like those of bruzual & charlot ( leitherer _ et al .
_ 1996 ) do not appear to be very accurate either , since they show abrupt jumps as a function of age and metallicity , due to the rather unknown contribution from stars in their late stages of evolution .
the conversion between @xmath39 and @xmath2 is only of minor importance .
when one compares the results of both methods , one finds that the difference in the colour - colour diagram of coma , when using two independent models , is quite large ( see fig .
this shows that a different , independent way to convert the colours would be very useful , and for that reason we attempted to derive an empirical conversion . in the search of data in @xmath12 or @xmath11 of the coma cluster to combine with bower et al.s dataset
, we have found the data of jrgensen _ et al . _ ( 1994 ) , with photometry in @xmath41 and gunn @xmath65 , and the ph.d .
thesis of steel ( 1998 ) , with @xmath41 and @xmath15 .
since there is no @xmath11-band photometry available , we decided to determine a conversion from @xmath66 to @xmath67 , a band with information very similar to @xmath55 , and then convert @xmath67 to @xmath55 using published stars .
it was also found that if one determines a colour of a galaxy using its integrated magnitudes , the observational error would be much larger than if the same aperture was used both times .
for that reason we preferred to use the @xmath67 data inside 20@xmath19 of steel ( 1998 ) , for which the dispersion in the @xmath67 vs. @xmath68 was only 0.02 mag , rather than the total colours of jrgensen etal .
( 1994 ) . combining the data of steel ( 1998 ) with those of bower et al .
( 1992 ) one obtains with negligible scatter .
the validity of this conversion between @xmath66 and @xmath55 can be for example be established by looking at the @xmath55 - magnitude relation for the elliptical galaxies of goudfrooij _ et al . _
( 1994a ) .
this comparison gives satisfactory results .
finally , we need to obtain a relation between @xmath71 and @xmath72 . since here also there is quite a difference between the theoretical models , we have determined a least squares relation between the observed @xmath71 and jrgensen s @xmath73 , which we first converted to @xmath72 using the conversion given in jrgensen ( 1995 ) : errors are difficult to determine .
we can get an estimate by comparing the data of goudfrooij et al .
( 1994a ) with the @xmath66 data of bower et al . converted to @xmath55 ( figure 5 ) . plotted on the x - axis for the data of bower et al .
are absolute @xmath41-band magnitudes , obtained using a distance modulus of 30.82 for virgo and 34.51 for coma ( aaronson et al .
1986 ) , to which a correction of 0.76 mag has been applied to convert them to integrated magnitudes . for the data of goudfrooij et al .
( 1994a ) we plot their m@xmath0 values .
there are three galaxies in common between the 2 samples , for which the m@xmath0 values agree within 0.10 mag . comparing the two samples we find that for the galaxies between m@xmath0 = 21 and 24 @xmath55 is redder by 0.032 mag on the average in the data of goudfrooij et al .
( 1994a ) .
this shows that our conversion from @xmath66 to @xmath55 is probably reasonable , and based on this , and on the tightness of individual relations , we believe that the error in @xmath55 is at most 0.06 mag , the same for @xmath2 or @xmath39 , and 0.05 mag in @xmath72 , resulting in 0.08 mag in @xmath1 .
we not only applied these transformation to the mean relation for the coma cluster , but also for the giant elliptical ngc 4472 ( see fig . | we present optical and near - infrared colour maps of the central regions of bulges of s0 and spiral galaxies obtained with wfpc2 and nicmos on the hst . by combined use of hst and ground - based data the colour information spans from a few tens of pc to a few kpc . in almost all galaxies the colour profiles in the central 100 - 200 pc
become more rapidly redder .
we attribute the high central colour indices to a central concentration of dust .
we infer an average extinction at the center of a@xmath0 = 0.6 1.0 mag .
several objects show central dust rings or disks at sub - kpc scales similar to those found by others in giant ellipticals . for galactic bulges of types s0 to sb , the tightness of the @xmath1 vs @xmath2 relation suggests that the age spread among bulges of early type galaxies is small , of at most 2 gyr .
colours at 1 r@xmath3 , where we expect extinction to be negligible are similar to than those of elliptical galaxies in the coma cluster , suggesting that these bulges formed at the same time as the bright galaxies in coma .
furthermore the galaxy ages are found to be independent of their environment .
since it is likely that coma was formed at redshift z @xmath4 3 , our bulges , which are in groups and in the field , also must have been formed at this epoch .
bulges of early - type spirals can not be formed by secular evolution of bars at recent epochs , since such bulges would be much younger .
there are three galaxies of type sbc and later , their bulges are younger and could perhaps arise from secular evolution of transient bars .
our results are in good agreement with semi - analytic predictions ( baugh _ et al .
_ 1996 , 1998 , kauffmann 1996 ) , who also predict that bulges , in clusters and in the field , are as old as giant ellipticals in clusters . 2mg@xmath5 |
at metal surfaces there exist specific electronic states not present in the bulk , which can be classified as intrinsic ( crystal - induced ) surface states@xcite and image - potential ( rydberg - like ) states.@xcite intrinsic surface states are originated by the symmetry breaking at the surface , they have their maximum near the surface , and they are classified as tamm@xcite and shockley@xcite states ; in particular , intrinsic shockley surface states typically occur in the gap of free - electron - like s , p bands near the fermi level.@xcite image - potential states appear as a result of the self - interaction that an electron near the surface suffers from the polarization charge it induces at the surface , and they occur in the vacuum region of metal surfaces with a band gap near the vacuum level.@xcite figure [ fig : chulpot ] illustrates shockley and image - potential states in the gap of the @xmath2 projected band structure of the ( 100 ) and ( 111 ) surfaces of cu . if an electron or hole is added to the solid at one of these states , inelastic coupling of the excited quasiparticle with the crystal , which can be experimentally observed through a variety of spectroscopies,@xcite may occur through electron - electron ( e - e ) and electron - phonon ( e - ph ) scattering .
the decay rate due to the e - ph interaction , which is relatively important only in the case of excited shockley holes near the fermi level , has been investigated recently by using the eliashberg function.@xcite accurate many - body calculations of the decay rate due to the e - e interaction were first carried out for image states on the ( 100 ) and ( 111 ) surfaces of cu.@xcite since then , many - body calculations of the e - e decay have been reported for a variety of simple , noble and transition metals .
@xcite nevertheless , existing calculations have been typically performed within the @xmath3 approximation of many - body theory,@xcite with no inclusion of exchange and correlation ( xc ) effects .
exceptions are ( i ) a calculation of the e - e decay rate of image states on the ( 100 ) and ( 111 ) surfaces of cu that incorporates xc effects in an adiabatic local - density approximation ( alda)@xcite and ( ii ) an approximate evaluation of the lifetime of shockley states in the noble metals that incorporates the exchange contribution to the self - energy.@xcite in this paper , we report extensive calculations of the screened interaction , the self - energy , and the e - e inelastic lifetime of shockley and image states on the ( 100 ) and ( 111 ) surfaces of cu that go beyond the @xmath4 approximation .
short - range xc effects are incorporated both in the description of the dynamical screening of the many - electron system [ we go beyond the random - phase approximation ( rpa ) in the evaluation of the screened interaction w ] and in the expansion of the electron self - energy in terms of @xmath5 [ we go beyond the @xmath0 approximation ] .
this is the @xmath1 approximation of many - body theory,@xcite which treats on the same footing xc effects between pairs of electrons within the fermi sea ( screening electrons ) and between the excited electron and the fermi sea .
mahan and sernelius@xcite showed that the inclusion , within the @xmath1 approximation , of the same vertex function in the screened interaction and the numerator of the self - energy yields results for the band - width of a homogeneous electron gas very similar to those obtained in the @xmath3 approximation , due to a large cancellation of vertex corrections .
large cancellations were also observed to occur in the decay rates of image@xcite and bulk@xcite states in the noble metals , by incorporating xc effects in the alda . in the decay of low - energy bulk states below the vacuum level energy transfers @xmath6 are well below the fermi energy and momentum transfers @xmath7 are typically smaller than @xmath8 , @xmath9 being the magnitude of the fermi wave vector , so that one can safely assume that both @xmath10 and @xmath11 are small and xc effects can , therefore , be incorporated in the alda . however , in the case of shockley and image states the alda might lead to spurious results , due to the presence of small local values of the fermi wave vector in a region where the electron density is small .
hence , here we use an adiabatic _ nonlocal _ xc kernel that accurately describes xc effects in the limit of a homogeneous electron gas of arbitrary density and which has been succesful in the description of the xc contribution to the jellium surface energy.@xcite it has been argued in the past that a realistic first - principles description of the electronic band structure is of key importance in the determination of the inelastic lifetime of bulk electronic states in the noble metals.@xcite the main conclusion drawn in ref .
was that in the case of the noble metals deviations from electron dynamics in a free gas of @xmath12 electrons mainly originate in the participation of @xmath13 electrons in the screening of electron - electron interactions .
the role of occupied @xmath13 bands in the dynamics of excited surface - state electrons and holes on silver surfaces was later investigated via a polarizable medium giving rise to additional screening , @xcite and it was concluded that @xmath13 electrons do not participate significantly in the screening of the interaction between surface states ( which are located near the surface ) and the fermi gas of the solid.@xcite in order to investigate the dynamics of shockley and image states on cu surfaces , we assume that the density of valence electrons in the solid varies only along the @xmath14 axis , which is taken to be normal to the surface . hence , our calculations start by solving the single - particle time - independent schrdinger equation of electrons moving in a physically motivated one - dimensional ( 1d ) model potential that is known to correcly reproduce the behaviour of @xmath12 valence states and accurately describes , in particular , the projected bulk band gap and the binding energy of the shockley and the first image state.@xcite the eigenfunctions and eigenvalues of such an effective single - particle hamiltonian are then used to compute the screened interaction , the @xmath1 self - energy , and the e - e decay rates of shockley and image states . for comparison ,
we also compute @xmath3 , @xmath15 , and @xmath16 decay rates , with no inclusion of xc effects , with inclusion of xc effects beyond the rpa in the screened interaction @xmath5 alone , and with inclusion of xc effects beyond the @xmath3 in the expansion of the electron self - energy in terms of the rpa screened interaction @xmath17 , respectively .
our results indicate that ( i ) although the use of the alda leads to spurious results for the screened interaction , a more realistic adiabatic _ nonlocal _ description of xc effects yields inelastic lifetimes of shockley and image states that esentially coincide with those obtained in the alda , and ( ii ) the overall effect of short - range xc is small and @xmath1 linewidths are close to their @xmath3 counterparts , as occurs in the case of low - energy bulk states.@xcite the paper is organized as follows .
explicit expressions for the e - e decay rate of surface - state electrons and holes at solid surfaces are derived in sec .
ii , in the @xmath1 approximation of many - body theory .
the results of numerical calculations of the screened interaction , the self - energy , and the decay rate of shockley and image states on the ( 100 ) and ( 111 ) surfaces of cu are presented in sec .
the summary and conclusions are given in sec .
iv . unless stated otherwise , atomic units ( a.u . ) are used throughout , i.e. , @xmath18 .
let us consider an arbitrary many - electron system of density @xmath19 . in the framework of many - body theory ,
the decay rate ( or reciprocal lifetime ) of a quasiparticle ( electron or hole ) that has been added in the single - particle state @xmath20 of energy @xmath21 is obtained as the projection of the imaginary part of the self - energy @xmath22 over the quasiparticle - state itself@xcite @xmath23 where the @xmath24 sign in front of the integral should be taken to be minus or plus depending on whether the quasiparticle is an electron ( @xmath25 ) or a hole ( @xmath26 ) , respectively , @xmath27 being the fermi energy .
to lowest order in a series - expansion of the self - energy in terms of the frequency - dependent screened interaction @xmath28 , the self - energy is obtained by integrating the product of the interacting green function @xmath29 and the screened interaction @xmath28 , and is therefore called the @xmath0 self - energy . if one further replaces the interacting green function by its noninteracting counterpart @xmath30 , one finds the @xmath15 self - energy and from eq .
( [ eq1 ] ) the following expression for the @xmath15 reciprocal lifetime : @xmath31 where the sum is extended over a complete set of single - particle states @xmath32 of energy @xmath33 ( @xmath34 or @xmath35 ) .
equation ( [ eq : tau ] ) exactly coincides with the result one would obtain from the lowest - order probability per unit time for an excited electron or hole in an initial state @xmath20 of energy @xmath21 to be scattered into the state @xmath32 of energy @xmath33 by exciting a fermi system of interacting electrons from its many - particle ground state to some many - particle excited state.@xcite the interaction @xmath36 entering eq .
( [ eq : tau ] ) can be rigurously expressed as follows @xmath37 @xmath38 representing the bare coulomb interaction and @xmath39 being the time - ordered density - response function of the many - electron system , which for the positive frequencies ( @xmath40 ) entering eq .
( [ eq : tau ] ) coincides with the retarded density - response function of linear - response theory . in the framework of time - dependent density - functional theory ( tddft),@xcite the _ exact _ retarded density - response function
is obtained by solving the following integral equation:@xcite @xmath41({\bf r}_1,{\bf r}_2;\omega)\right\}\chi({\bf r}_2 , { \bf r}';\omega),\end{aligned}\ ] ] where @xmath42 denotes the density - response function of noninteracting kohn - sham electrons , i.e. , independent electrons moving in the effective kohn - sham potential of density - functional theory ( dft ) . the frequency - dependent xc kernel @xmath43({\bf r},{\bf r } ' \omega)$ ]
is the functional derivative of the frequency - dependent xc potential @xmath44({\bf r},\omega)$ ] of tddft , to be evaluated at @xmath19 . in the rpa ,
@xmath43({\bf r},{\bf r}';\omega)$ ] is set equal to zero and eq .
( [ eq : tau ] ) yields the so - called @xmath3 ( or @xmath15-rpa ) reciprocal lifetime .
the xc kernel @xmath43({\bf r},{\bf r}';\omega)$ ] , which is absent in the rpa , accounts for the presence of an xc hole associated to all screening electrons in the fermi sea .
hence , one might be tempted to conclude that the full @xmath15 approximation [ with the formally exact screened interaction @xmath5 of eq .
( [ eq : wrpa ] ) ] should be a better approximation than its @xmath3 counterpart [ with the screened interaction @xmath5 evaluated in the rpa ] .
however , the xc hole associated to the excited hot electron is still absent in the @xmath15 approximation .
therefore , if one goes beyond rpa in the description of @xmath5 , one should also go beyond the @xmath15 approximation in the expansion of the electron self - energy in powers of @xmath5 . by including xc effects both beyond rpa in the description of @xmath5 and beyond @xmath15 in the description of the self - energy,@xcite the so - called @xmath1 approximation
yields a lifetime broadening that is of the @xmath15 form [ see eq .
( [ eq : tau ] ) ] , but with the actual screened interaction @xmath36 of eq .
( [ eq : wrpa ] ) replaced by a new effective screened interaction @xmath45({\bf r},{\bf r}_1;\omega)\right\}\,\chi({\bf r}_1,{\bf r}_2;\omega)\,v({\bf r}_2,{\bf r}'),\end{aligned}\ ] ] which includes all powers in @xmath5 beyond the @xmath15 approximation . in the case of a bounded electron gas that is translationally invariant in two directions , such as the jellium surface or the physically motivated model surface described above , the single - particle states entering eq .
( [ eq : tau ] ) are of the form @xmath46 with energies @xmath47 @xmath48 being a wave vector parallel to the surface and @xmath49 denoting the effective mass in the plane of the surface.@xcite introducing eqs .
( [ bounded1 ] ) and ( [ bounded2 ] ) into eq .
( [ eq1 ] ) , one finds the following expression for the reciprocal lifetime of a quasiparticle ( electron or hole ) that has been added in the single - particle state @xmath50 of energy @xmath51 : @xmath52 where @xmath53 represents the two - dimensional ( 2d ) fourier transform of the self - energy @xmath54 . using the single - particle wave functions and energies of eqs .
( [ bounded1 ] ) and ( [ bounded2 ] ) , the @xmath15 reciprocal lifetime of eq .
( [ eq : tau ] ) yields @xmath55 where @xmath56 , @xmath48 and @xmath57 represent wave vectors parallel to the surface , and @xmath58 denotes the 2d fourier transform of the screened interaction @xmath36 of eq .
( [ eq : wrpa ] ) , i.e. , @xmath59 @xmath60 and @xmath61 being 2d fourier transforms of the bare coulomb interaction and the density - response function of eq .
( [ eq : xalda ] ) , respectively . in the @xmath3 ( or @xmath15-rpa )
approximation , the reciprocal lifetime is also given by eqs .
( [ eq : tau2 ] ) and ( [ eq : wrpa2 ] ) , but with the xc kernel @xmath43({\bf r},{\bf r}';\omega)$ ] entering eq .
( [ eq : xalda ] ) set equal to zero .
using the single - particle wave functions and energies of eqs .
( [ bounded1 ] ) and ( [ bounded2 ] ) , the @xmath1 reciprocal lifetime is also found to be given by eq .
( [ eq : tau2 ] ) , but with @xmath58 replaced by the 2d fourier transform of the effective screened interaction @xmath62 of eq .
( [ eq : walda ] ) , i.e. : @xmath63 where @xmath64(z , z_1;{\bf k},\omega)\right\}\,\chi(z_1,z_2;{\bf k},\omega)\,v(z_2,z';{\bf k}),\end{aligned}\ ] ] @xmath43(z , z_1;{\bf k},\omega)$ ] being the 2d fourier transform of the xc kernel @xmath43({\bf r},{\bf r}';\omega)$ ] . in the @xmath16 approximation ,
the reciprocal lifetime is also given by eqs .
( [ eq : tau3 ] ) and ( [ eq : walda3 ] ) , thereby with full inclusion of the xc kernel entering eq .
( [ eq : walda3 ] ) , but with the xc kernel entering eq .
( [ eq : xalda ] ) set equal to zero .
hence , we note that both @xmath15 and @xmath1 reciprocal lifetimes [ eqs .
( [ eq : tau2 ] ) and ( [ eq : tau3 ] ) ] can be calculated from the knowledge of two basic ingredients : ( i ) single - particle wave functions and energies of the form of eqs .
( [ bounded1 ] ) and ( [ bounded2 ] ) , which are also basic quantities in the evaluation of the noninteracting density - response function @xmath65 , and ( ii ) the xc kernel @xmath43(z , z';{\bf k},\omega)$ ] . for the description of the noninteracting density - response function @xmath65 [ and , therefore , the screened interaction @xmath58 and the effective screened interaction @xmath66 of eqs .
( [ eq : wrpa2 ] ) and ( [ eq : walda3 ] ) , respectively ] single - particle wave functions and energies can safely be taken to be the eigenvalues and eigenfunctions of a _ jellium _ self - consistent kohn - sham hamiltonian .
@xcite nevertheless , the actual band structure of @xmath12 electrons near the surface of noble metals calls for a more realistic description of the single - particle wave functions [ @xmath67 and @xmath68 and energies [ @xmath21 and @xmath33 ] entering eqs .
( [ eq : tau2 ] ) and ( [ eq : tau3 ] ) . hence , in the calculations presented in this paper all the single - particle wave functions and energies ( those entering eqs .
( [ eq : tau2 ] ) and ( [ eq : tau3 ] ) and also those involved in the evaluation of the noninteracting density - response function ) are taken to be the eigenfunctions and eigenvalues of a physically motivated 1d model hamiltonian that accurately reproduces the projected band gap and the binding energy of the shockley and the first image state.@xcite in order to investigate the impact of strong variations of the electron density induced near the surface , and because the excitation energies of interest are typically small ( particularly in the case of shockley holes ) , we consider the following _ adiabatic _ ( @xmath70 ) approximations of the xc kernel @xmath43(z , z';{\bf k},\omega)$]:@xcite if one assumes that dynamic electron - density fluctuations are slowly varying in all directions , the xc kernel @xmath43(z , z';{\bf k},\omega)$ ] is easily found to be given by the following expression:@xcite @xmath71(z , z';{\bf k},\omega)=\bar f^{xc}(n_0(z);k^{3d}=0,\omega=0)\,\delta(z - z').\ ] ] here , @xmath72 is the 3d fourier transform of the xc kernel of a homogeneous electron gas of density @xmath73 , which in the limit as @xmath74 and @xmath75 is known to be the second derivative of the xc energy @xmath76 per particle of a homogeneous electron gas , to be evaluated at the local density @xmath73 .
we use the perdew - wang parametrization @xcite of the difussion monte carlo ( dmc ) xc energy @xmath77 reported by ceperley and alder @xcite .
a more accurate description of short - range xc effects can be carried out by replacing the _ local _ xc kernel @xmath78 entering eq .
( [ alda ] ) by a more accurate still adiabatic but momentum - dependent xc kernel @xmath79 ( thus only assuming that the dynamic density fluctuation is slowly varying in the direction perpendicular to the surface ) , i.e , @xmath80(z , z';{\bf k},\omega)=\bar f^{xc}(n_0(z);k^{3d}=k,\omega=0)\,\delta(z - z').\ ] ] here we exploit the accurate dmc calculations reported by moroni _
et al._@xcite for the static ( @xmath70 ) @xmath81-dependent _ nonlocal _ xc kernel @xmath82 of a homogeneous electron gas .
a parametrization of this data satisfying the well - known small- and large - wavelength asymptotic behaviour was carried out by corradini _
et al . _ ( cdop ) @xcite .
here we still neglect the frequency dependence of the xc kernel ( adiabatic approximation ) , but now we make no assumption on the variation of the dynamic density fluctuation and assume that the _ unperturbed _ density variation @xmath83 $ ] is small within the short range of @xmath43(z , z';{\bf k},\omega)$ ] .
this allows to write @xmath84(z , z';{\bf k},\omega)=\bar f^{xc}(\left[n_0(z)+n_0(z')\right]/2;z , z';k,\omega=0),\ ] ] where @xmath85 represents the 2d fourier transform of the xc kernel @xmath86 of a homogeneous electron gas of density @xmath87 .
an explicit expression for the 2d fourier transform of the cdop parametrization of @xmath88 was reported in ref .
: @xmath89 { \rm e}^{-\beta\left[k_f^2\tilde z^2/4\beta^2 + k^2/k_f^2\right]},\ ] ] where @xmath90 , @xmath91 , @xmath92 , @xmath93 , and @xmath94 are dimensionless functions of the electron density ( see ref . ) , @xmath95 , and @xmath96 .
on the ( 111 ) surface of cu , the @xmath97 shockley state at the center of the surface brillouin zone ( @xmath98 ) lies just below the fermi level , with @xmath99 .
binding energies of the @xmath100 image state on the ( 111 ) and ( 100 ) surfaces of cu ( measured with respect to the vacuum level ) are 0.83 and 0.57 mev , respectively .
effective masses of the @xmath100 image state on cu(111 ) and cu(100 ) are close to the free - electron mass ( @xmath101),@xcite while the effective mass of the @xmath97 shockley state on cu(111 ) is 0.42.@xcite the probability density of the @xmath100 image states on cu(111 ) and cu(100 ) have a maximum at 2.3 and 3.8 @xmath102 , respectively , outside the crystal edge ( @xmath103 ) , which we choose to be located half a lattice spacing beyond the last atomic layer . the @xmath97 shockley state wave function in cu(111 ) , however , is maximum at the crystal edge . we have carried out calculations of the imaginary part of the screened interaction @xmath58 and the effective screened interaction @xmath66 of thin slabs . in order to ensure that our slab calculations are a faithful representation of the actual screened interaction of a semiinfinite system , we have used films up to 50 layers of atoms and 80 interlayer - spacing vacuum intervals , as in the @xmath3 ( @xmath15-rpa ) calculations reported in refs . and .
the impact of xc effects on the imaginary part of the effective screened interaction in the vicinity of the ( 100 ) and ( 111 ) surfaces of cu is illustrated in fig .
[ fig : imw_chul_jel ] , where alda calculations of @xmath104 $ ] ( with full inclusion of xc effects ) are compared to calculations of @xmath105 $ ] with ( alda ) and without ( rpa ) xc effects .
exchange - correlation effects included in the effective screened interaction have two sources , as discussed in section [ gamma ] .
first , there is the reduction of the screening due to the presence of an xc hole associated to all electrons in the fermi sea [ see eq .
( [ eq : xalda ] ) ] , which is included in the calculations represented in fig .
[ fig : imw_chul_jel ] by thick solid lines and _ also _ in the calculations represented by dotted lines .
secondly , there is the reduction of the effective screened interaction itself due to the xc hole associated to each electron [ see eq .
( [ eq : walda3 ] ) ] , which is _ only _ included in the calculations represented in fig .
[ fig : imw_chul_jel ] by thick solid lines .
these contributions have opposite signs and it is the latter which dominates .
existing @xmath1 calculations of the lifetime broadening of image states on cu(100 ) and cu(111 ) were performed with the alda xc kernel that we have used in the calculations represented in fig .
[ fig : imw_chul_jel ] . the error introduced by the use of this _ local _
kernel is small in the interior of the solid , as the wave vectors involved are small ( @xmath106 ) .
however , fig .
[ fig : imw_chul_jel ] shows that the alda leads to spurious ( negative ) results for @xmath104 $ ] near the surface , which is due to the presence of small local values of the fermi wave vector ( @xmath107 ) in a region where the electron density is small .
hence , we have carried out refined alda and adiabatic _ nonlocal _ ( anlda ) calculations of @xmath104 $ ] ( both with full inclusion of xc effects ) , which have been plotted in fig . [ fig : figure4 ] .
this figure clearly shows that the impact of _ nonlocality _ on the effective screened interaction is large near the surface , bringing spurious alda calculations ( thin solid lines ) to a more realistic behaviour near the surface ( thick solid lines ) .
the refined alda scheme partially overcomes the failure of the alda , but a full description of the nonlocality of xc effects near the surface might be needed for a realistic description of the absorption power of solid surfaces .
figure [ multiself ] exhibits @xmath3 ( @xmath15-rpa ) , @xmath15 , and @xmath1 calculations of the imaginary part of the @xmath97 surface - state self - energy @xmath108 , versus @xmath14 , in the vicinity of the ( 111 ) surface of cu , with use ( in the case of the @xmath15 and @xmath1 approximations ) of the adiabatic _ nonlocal _ xc kernel ( anlda ) described in section [ anla ] .
this figure shows that as occurs in the case of the screened interaction xc effects partially compensate each other , leading to an overall effect of no more than @xmath109 . for comparison ,
we have also used ( in the case of the @xmath15 and @xmath1 approximations ) the alda and refined alda kernels described in section [ anla ] , and we have found that although the use of these local or semilocal kernels leads to spurious results for the screened interaction , our more realistic anlda kernel yields self - energies that esentially coincide with those obtained in the alda .
now we focus on the evaluation of the decay rate ( reciprocal lifetime ) of surface - state electrons ( and holes ) at the @xmath100 ( and @xmath97 ) surface - state band edge ( @xmath110 ) of the ( 111 ) and ( 100 ) surfaces of cu .
calculations of the noninteracting density - response function @xmath65 ( and , therefore , the reciprocal lifetime ) require the introduction of complex frequencies of the form @xmath111 , @xmath112 being a positive infinitesimal .
hence , in order to ensure that our numerical calculations yield a converged value of the reciprocal lifetime , we have calculated @xmath113 as a function of the parameter @xmath112 . fig .
[ fig : fig2 ] represents the results we have obtained for the @xmath3 reciprocal lifetimes of shockley and image states on the ( 100 ) and ( 111 ) surfaces of cu , showing that converged results are obtained for a sufficiently small value of @xmath112 . .@xmath4 , @xmath114 , and @xmath1 reciprocal lifetimes , in linewidth units ( mev ) , of an excited surface - state electron ( hole ) at the @xmath100 ( @xmath97 ) surface - state band edge ( @xmath110 ) of the ( 111 ) and ( 100 ) surfaces of cu .
in the case of the @xmath114 and @xmath1 reciprocal lifetimes , both alda and anlda exchange - correlation kernels have been considered . [
cols="<,^,^,^,^,^,^,^,^",options="header " , ] converged calculations of the reciprocal lifetimes of shockley and image states on the ( 100 ) and ( 111 ) surfaces of cu are exhibited in table [ tabled0 ] .
this table shows : ( i ) @xmath3 results , which reproduce previous calculations @xcite , ( ii ) alda @xmath1 results , which in the case of the @xmath100 image state on cu(111 ) and cu(100 ) reproduce the calculations reported in ref .
( alda @xmath1 calculations of the reciprocal lifetime of @xmath97 shockley states had _ not _ been reported before ) , and ( iii ) anlda @xmath1 calculations , never reported before ; for comparison , @xmath15 reciprocal lifetimes are also shown in this table , with use of both the alda and the adiabatic _ nonlocal _ kernel anlda described in section [ anla ] .
differences between our @xmath3 reciprocal lifetime of the @xmath97 shockley state in cu(111 ) ( @xmath115 ) and those reported before@xcite ( @xmath116 ) are simply due to the fact that in our present model we are not accounting for the change of the @xmath14-dependent surface - state wave functions @xmath67 and @xmath117 along the surface - state dispersion curve . as in the case of the self - energy ,
the results shown in table [ tabled0 ] show that ( i ) a realistic adiabatic _ nonlocal _ description of xc effects yields reciprocal lifetimes of shockley and image states that esentially coincide with those obtained in the alda , and ( ii ) the overall effect of short - range xc is small and @xmath1 reciprocal lifetimes are close to their @xmath3 counterparts .
we have carried out extensive calculations of the self - energy and lifetime of shockley and image states on the ( 100 ) and ( 111 ) surfaces of cu , in the framework of the @xmath1 approximation of many - body theory .
this approximation treats on the same footing xc effects between pairs of electrons within the fermi sea ( screening electrons ) and between the excited surface - state electron ( or hole ) and the fermi sea .
we have included xc effects within tddft from the knowledge of an adiabatic _ nonlocal _ xc kernel that goes beyond the local - density approximation , and we have found that these xc contributions ( in the screened interaction @xmath5 and in the expansion of the self - energy in terms of @xmath5 ) have opposite signs and it is the latter which dominates , leading to @xmath1 reciprocal lifetimes that are only slightly lower than their @xmath3 counterparts .
the authors acknowledge partial support by the upv / ehu , the basque unibertsitate eta ikerketa saila , the spanish ministerio de educacin y ciencia ( grant no . csd2006 - 53 ) , and the ec 6th framework network of excellence nanoquanta .
j. e. inglesfield , rep .
. phys . * 45 * , 223 ( 1982 ) .
n. v. smith , rep .
. phys . * 51 * , 1227 ( 1988 ) .
m. echenique and j. b. pendry prog .
sci . * 32 * , 111 ( 1989 ) .
i. e. tamm , z. phys .
* 76 * , 849 ( 1932 ) .
w. shockley phys . rev .
b , * 56 * , 317 ( 1939 ) . p. o. gartland and b. j. slagsvold phys .
b * 12 * , 4047 ( 1975 ) .
w. eberhardt and e.w .
plummer phys . rev .
b * 21 * , 3245 ( 1980 ) .
p.m. echenique and j.b .
pendry j. phys .
c * 11 * , 2065 ( 1978 ) .
v. dose , surf .
rep . * 5 * , 337 ( 1985 ) .
s. d. kevan ( ed . ) , angle - resolved photoemission , vol .
74 of studies in surface science and catalysis , elsevier , amsterdam , 1992 .
m. donath , surf .
rep . * 20 * , 251 ( 1994 ) .
fauster and w. steinmann , in _ photonic probes of surfaces _ , edited by p. halevi , electromagentic waves : recent development in research vol . 2 ( elsevier , amsterdam , 1995 ) .
r. matzdorf , surf .
20 * , 251 ( 1994 ) .
u. hfer , i. l. shumay , ch .
reuss , u. thomann , w. wallauer , and th .
fauster , science * 277 * , 1480 ( 1997 ) .
a. eiguren , b. hellsing , f. reinert , g. nicolay , e. v. chulkov , v. m. silkin , s. hfner , and p. m. echenique , phys .
lett . * 88 * , 066805 ( 2002 ) .
e. v. chulkov , i. sarria , v. m. silkin , j. m. pitarke , and p. m. echenique , phys .
* 80 * , 4947 ( 1998 ) .
i. sarria , j. osma , e. v. chulkov , j. m. pitarke , and p. m. echenique , phys .
b * 60 * , 11795 ( 1999 ) .
j. kliewer , r. berndt , e. v. chulkov , v. m. silkin , p. m. echenique , and s. crampin , science , * 288 * , 1399 ( 2000 ) .
p.m. echenique , j. osma , v.m .
silkin , e.v .
chulkov and j. m. pitarke , appl .
a * 71 * , 503 ( 2000 ) ; p. m. echenique , j. osma , m. machado , v. m. silkin , e. v. chulkov , and j. m. pitarke , prog .
sci . * 67 * , 271 ( 2000 ) .
a. fukui , h. kasai , and a. okiji , surf .
* 493 * , 671 ( 2001 ) .
p. m. echenique , r. berndt , e. v. chulkov , th .
fauster , a. goldmann , and u. hfer , surf .
* 52 * , 219 ( 2004 ) .
l. hedin and s. lundquist , solid state phys .
* 23 * , 1 ( 1996 ) .
f. aryasetiawan and o. gunnarsson , rep .
phys . * 61 * , 237 ( 1998 ) .
p. m. echenique , j. m. pitarke , e. v. chulkov , and a. rubio , chem . phys .
* 251 * , 1 ( 2000 ) .
m. nekovee and j. m. pitarke , comput .
. commun . * 137 * , 123 ( 2001 ) .
g. d. mahan and b. e. sernelius , phys .
lett . * 62 * , 2718 ( 1989 ) .
g. d. mahan , _ many particle physics _
( plenum , new york , 1990 ) . i. g. gurtubay , j. m. pitarke , and p. m. echenique , phys .
b * 69 * , 245106 ( 2004 ) .
j. m. pitarke and j. p. perdew , phys .
b. * 67 * , 045101 ( 2003 ) .
i. campillo , j. m. pitarke , a. rubio , e. zarate , and p. m. echenique , phys .
lett . * 83 * , 2230 ( 1999 ) ; i. campillo , j. m. pitarke , a. rubio , and p. m. echenique , phys .
b * 62 * , 1500(2000 ) . a. garca - lekue , j. m. pitarke , e. v. chulkov , a. liebsch , and p. m. echenique , phys . rev
* 89 * , 096401 ( 2002 ) ; phys .
b * 68 * , 045103 ( 2003 ) .
@xmath13 electrons do play an important role in the case of image states ( but not in the case of shockley states ) on silver surfaces , but this is simply a consequence of the reduction ( in the presence of @xmath13 electrons ) of the surface - plasmon energy that allows to open a new decay channel .
no surface - plasmon decay channel is opened in the case of the cu surfaces , since even in the presence of @xmath13 electrons the cu surface - plasmon energy is too large for this decay channel to be available .
e. v. chulkov , v. m. silkin , and p. m. echenique , surf .
sci . * 391 * , l1217 ( 1997 ) .
j. m. pitarke and i. campillo , nucl .
methods b * 164 * , 147 ( 2000 ) ; j. m. pitarke , v. p. zhukov , r. keyling , e. v. chulkov , and p. m. echenique , chemphyschem * 5 * , 1284 ( 2004 ) .
e. runge and e. k. u. gross , phys .
lett . * 52 * , 997 ( 1984 ) m. petersilka , u. j. gossmann , and e. k. u. gross , phys .
lett . * 76 * , 1212 ( 1996 ) . in the case of the jellium surface ,
the electron motion is free in the plane of the surface and the effective mass coincides , therefore , with that of free electrons . in the case of the model surface
, the effective mass is taken from the actual energy dispersion of the projected band structure .
see , e.g. , a. liebsch , _ electronic excitations at metal surfaces _ ( plenum press , new york , 1997 ) .
j. p. perdew and y. wang , phys . rev .
b * 45 * , 13244 ( 1992 ) .
d. m. ceperley and b. j. alder , phys .
rev . lett . * 45 * , 566 ( 1980 ) .
s. moroni , d. m. ceperley , and g. senatore , phys .
75 * , 689 ( 1995 ) .
m. corradini , r. del sole , g. onida , and m. palummo , phys .
b * 57 * , 14569 ( 1998 ) .
a. goldmann , v. dose , and g. borstel , phys .
b * 32 * , 1971 ( 1985 ) .
s. l. hulbert , p. d. johnson , n. g. stoffel , w. a. royer , and n. v. smith , phys .
b. * 31 * , 6815 ( 1985 ) .
m. g. vergniory , j. m. pitarke , and s. crampin , phys .
b * 72 * , 193401 ( 2005 ) . | we report many - body calculations of the self - energy and lifetime of shockley and image states on the ( 100 ) and ( 111 ) surfaces of cu that go beyond the @xmath0 approximation of many - body theory .
the self - energy is computed in the framework of the @xmath1 approximation by including short - range exchange - correlation ( xc ) effects both in the screened interaction w ( beyond the random - phase approximation ) and in the expansion of the self - energy in terms of w ( beyond the gw approximation ) .
exchange - correlation effects are described within time - dependent density - functional theory from the knowledge of an adiabatic _ nonlocal _ xc kernel that goes beyond the local - density approximation . |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.