article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
in this paper , we study schrdinger operators with any polynomial potential of degree @xmath0 with complex coefficients , under decaying boundary conditions along two rays to infinity in the complex plane , and provide asymptotic expansions of the eigenvalue counting functions and the eigenvalues . then we will use these to reconstruct some coefficients of polynomial potentials from asymptotic expansions of the eigenvalues , and to show that all but finitely many eigenvalues of every @xmath8-symmetric oscillator with a polynomial potential are real and positive . for integers
@xmath0 fixed and @xmath9 , we are considering the non - standard eigenvalue problems @xmath10u(z,\lambda)=\lambda u(z,\lambda),\quad\text{for some $ \lambda\in{\mathbb c}$},\ ] ] with the boundary condition that @xmath11 where @xmath12 is a polynomial of degree at most @xmath13 of the form @xmath14 the anharmonic oscillators @xmath15 with the various boundary conditions are considered in @xcite .
the most studied case is when @xmath16 is even and @xmath17 , for example , see @xcite . in this case
, @xmath15 is an schrdinger operator in @xmath18 .
this is self - adjoint if all the coefficients of the polynomial @xmath19 in @xmath4 are real , and non - self - adjoint if a coefficient of @xmath19 is non - real .
the case when @xmath20 has been studied extensively in recent years in the context of @xmath8-symmetric theory @xcite . throughout this paper
, we use the integer @xmath0 for the degree of the polynomial potential and integer @xmath21 with @xmath1 .
we will use @xmath22 , depending on the potential and the boundary condition , to denote the eigenvalues of @xmath15 , without explicitly indicating their dependence on the potential and the boundary condition .
also , we let @xmath23 be the coefficient vector of @xmath24 . if a nonconstant function @xmath25 satisfies with some @xmath26 and the boundary condition , then we call @xmath27 an _ eigenvalue _ of @xmath15 and @xmath25 an _ eigenfunction of @xmath15 associated with the eigenvalue @xmath27_. also , the _ geometric multiplicity of an eigenvalue @xmath27 _ is the number of linearly independent eigenfunctions associated with the eigenvalue @xmath27 . for each @xmath21 fixed ,
we number the eigenvalues @xmath28 in the order of nondecreasing magnitudes , counting their `` algebraic multiplicities '' , where the integer @xmath29 , depending on the potential and the boundary condition , is due to our method of proof of theorem [ main_thm1 ] . in theorem [ main_thm1 ]
we show that for every large @xmath30 , there exists @xmath22 satisfying below .
however , we do not know the number of eigenvalues `` near '' zero , and this is why we need to have @xmath29 in numbering the eigenvalues .
before we state our main theorems , we first introduce some known facts by sibuya @xcite about the eigenvalues @xmath22 of @xmath15 .
[ main2 ] the eigenvalues @xmath7 of @xmath15 have the following properties . 1 .
the set of all eigenvalues is a discrete set in @xmath31 .
the geometric multiplicity of every eigenvalue is one .
infinitely many eigenvalues , accumulating at infinity , exist .
when @xmath20 the eigenvalues have the following asymptotic expansion @xmath32\quad\text{as $ n$ tends to infinity},\quad n \in { \mathbb n}.\ ] ] this paper contains results on direct and inverse spectral probelms , and their applications to @xmath8-symmetric potential problems .
theorem [ main_thm1 ] below is the main result , regarding asymptotic expansions of `` eigenvalue counting functions '' .
the other results below in the introduction are deduced from theorem [ main_thm1 ] . here , we first introduce the following theorem , regarding asymptotic expansions of a kind of eigenvalue counting functions , where we use multi - index notations with @xmath33 also , we use @xmath34 , @xmath35 and @xmath36 .
[ main_thm1 ] for @xmath37 , the eigenvalues @xmath22 of @xmath15 satisfy @xmath38 where @xmath39 is the largest integer that is less than or equal to @xmath40 , and where the error @xmath41 term is uniform on any compact set of @xmath37 , and @xmath42 where @xmath43 @xmath44 and @xmath45 where @xmath46 is the beta function .
we obtain by investigating the asymptotic expansions of an entire function whose zeros are the eigenvalues .
sibuya @xcite got by using the first order asymptotic expansion of the entire function .
next , we let @xmath47 , @xmath48 , be the eigenvalue counting function , that is , @xmath47 is the number of eigenvalues @xmath27 of @xmath15 such that @xmath49 . then the following theorem on an asymptotic expansion of the eigenalue counting function is a consequence of theorem [ main_thm1 ] .
[ main_thm2 ] let @xmath37 be fixed .
suppose that @xmath50 for @xmath51 .
then @xmath47 has the asymptotic expansion @xmath52 where the error @xmath53 is uniform for any compact set of @xmath37 . in theorem [ ineq_eq ]
below , we show that @xmath54 for all large @xmath30 .
suppose that @xmath55 . then since @xmath56 we see from theorem [ eigen_asy ] below that @xmath57 .
thus , @xmath58 hence , replacing @xmath59 in by @xmath60 , and solving the resulting equation for @xmath61 complete the proof .
next , we improve the asymptotic expansion of the eigenvalues @xmath7 of @xmath62 . in particular , we will prove the following , which essentially invert to get @xmath22 in terms of @xmath61 .
[ eigen_asy ] for each @xmath37 , there exist some constants @xmath63 , @xmath64 , such that @xmath65 where the error term is uniform for any compact set of @xmath37 , and where @xmath66 and @xmath67 , @xmath68 , are defined recurrently by @xmath69 and @xmath70 where @xmath71 .
we note for the first summation in the definition of @xmath72 that @xmath73 implies @xmath74 whenever @xmath75 .
also , for the second summation , we point out that @xmath76 implies @xmath74 whenever @xmath77 .
the asymptotic expansions of the eigenvalues of @xmath15 with @xmath78 and @xmath20 have been studied by a number of people .
for example , maslov @xcite computed the first three terms of asymptotic expansions of @xmath79 , where @xmath22 are the eigenvalues of @xmath80 helffer and robert @xcite considered @xmath81 where @xmath82 are positive integers and where @xmath83 is a _
real _ polynomial of degree at most @xmath84 .
they obtained existence of asymptotic expansions of eigenvalues to all orders , and suggested an explicit way of computing the coefficients of the asymptotic expansion .
in particular , for the case when the potential is @xmath85 , @xmath86 , helffer and robert @xcite computed the first nine terms of the asymptotic expansion of @xmath79 .
fedoryuk @xcite considered with complex polynomial potentials and with for @xmath78 , and showed the existence of asymptotic expansions of the eigenvalues to all orders .
note that there appear to be typographical errors in @xcite .
for example , when @xmath16 is even ( and @xmath17 ) the leading coefficient of the asymptotic expansion of @xmath22 in @xcite is @xmath87 , which is different from @xmath88 found in @xcite and again in theorem [ eigen_asy ] above .
next , we point out some differences between work of fedoryuk @xcite and the present work .
fedoryuk @xcite showed the existence of asymptotic expansion of the eigenvalues to all orders while we do not . on the other hand ,
we treat all decaying boundary conditions with @xmath1 , while fedoryuk @xcite studied the case @xmath78 only . moreover , we computed more coefficients @xmath67 explicitly , and our methods are different from fedoryuk s . here
, we introduce results on inverse spectral problems , but first the following corollary is an easy consequence of theorems [ main_thm1 ] and [ eigen_asy ] , regarding how the coefficients of the asymptotic expansions depend on @xmath37 .
[ corollary_1 ] let @xmath68 be a fixed integer .
then we have the following . 1 .
@xmath89 and @xmath72 are polynomials in @xmath90 .
2 . @xmath89 and @xmath72 do not depend on @xmath91 .
3 . if @xmath92 is a multiple of @xmath16 , then @xmath93 , and @xmath72 does not depend on @xmath94 .
4 . if @xmath92 is not a multiple of @xmath16 , then @xmath89 and @xmath72 depend linearly on @xmath94 .
statements on @xmath89 are direct consequences of the definition of @xmath89 in theorem [ main_thm1 ] .
one can use statements on @xmath89 and induction on @xmath95 to prove statements on @xmath72 .
next , one can reconstruct some coefficients of the polynomial potential from the asymptotic expansion of the eigenvalues .
let @xmath96 be a fixed integer .
suppose that @xmath97 is known whenever @xmath98 and @xmath99 is a multiple of @xmath16 .
if @xmath92 is a multiple of @xmath16 , then the asymptotic expansions of the eigenvalues @xmath7 of @xmath15 of type with an error term @xmath100 uniquely and explicitly determine @xmath94 . from the asymptotic expansion of the eigenvalues , one gets @xmath101 that are explicit polynomials in @xmath102 .
then since we know @xmath97 if @xmath99 is a multiple of @xmath16 , corollary [ corollary_1 ] says that we can find all @xmath102 .
one says that @xmath15 is @xmath8-symmetric if the potential @xmath103 satisfies @xmath104 , @xmath105 , that is equivalent to @xmath106 . here , we prove the partial reality of the eigenvalues @xmath22 of @xmath8-symmetric @xmath15 . but
first , we show the following theorem , regarding monotonicity of modulus of @xmath22 for all large @xmath30 .
[ ineq_eq ] for each @xmath37 there exists @xmath107 such that @xmath108 if @xmath109
. see theorem 3 in @xcite for the proof of the case when @xmath20 .
one can see that proof of theorem 3 in @xcite can be easily adapted for the cases when @xmath110 .
now we are ready to prove the following theorem on the partial reality of the eigenvalues @xmath22 of @xmath15 .
[ main_theorem ] suppose that @xmath106 . then all but finitely many eigenvalues of @xmath15 are real and positive .
hence @xmath50 for all @xmath51 , so that the counting function formula in theorem [ main_thm2 ] is valid .
when @xmath15 is @xmath8-symmetric ( i.e. , @xmath106 ) , @xmath111 is an eigenfunction associated with an eigenvalue @xmath27 if and only if @xmath112 is an eigenfunction associated with the eigenvalue @xmath113 .
thus , the eigenvalues either appear in complex conjugate pairs , or else are real .
so theorem [ ineq_eq ] implies theorem [ main_theorem ] . in recent years
, these @xmath8-symmetric operators have gathered considerable attention , because ample numerical and asymptotic studies suggest that many of such operators have real eigenvalues only even though they are not self - adjoint .
in particular , the differential operators @xmath114 with some polynomial potential @xmath103 and with the boundary condition have been considered in @xcite and references therein .
the rigorous proof of reality and positivity of the eigenvalues of @xmath8-symmetric operators with certain classes of polynomial potentials and with the boundary condition for @xmath20 , was given by dorey , dunning and tateo @xcite in 2001 and by the present author @xcite in 2002 . however , there are some @xmath8-symmetric polynomial potentials that produce a finite number of non - real eigenvalues @xcite for some particular classes of polynomial potentials .
so without any further restrictions on the real coefficients @xmath97 , theorem [ main_theorem ] is the most general result one can expect about reality of eigenvalues of @xmath8-symmetric operators with polynomial potentials .
this paper is organized as follows . in section [ prop_sect ]
, we will introduce work of hille @xcite and sibuya @xcite , regarding properties of solutions of .
also , we introduce entire functions @xmath115 whose zeros are closely related with the eigenvalues of @xmath15 , due to sibuya @xcite ( c. f. , section [ asymp_eigen ] ) . in section [ sec_4 ] , we then provide asymptotics of the entire function @xmath116 as @xmath117 in the complex plane @xcite , improving the asymptotics of @xmath116 in @xcite . in section [ sec_5 ] , we provide asymptotic expansions of @xmath115 as @xmath117 in @xmath31 . in section [ asymp_eigen ] ,
we investigate how the zeros of @xmath118 are related with the eigenvalues of @xmath15 . in sections [ sec_7 ] and [ sec_8 ] , we prove theorem [ main_thm1 ] . in section [ sec_9 ] ,
we prove theorem [ eigen_asy ] .
finally , in the appendix we compute @xmath119 in theorem [ main_thm1 ] , that is originally given in terms of certain integrals .
in this section , we introduce work of hille @xcite and sibuya @xcite about properties of the solutions of .
first , we scale equation because many facts that we need later are stated for the scaled equation .
let @xmath25 be a solution of ( [ ptsym ] ) and let @xmath120 .
then @xmath121 solves @xmath122v(z,\lambda)=0,\ ] ] where @xmath0 and @xmath12 is a polynomial ( possibly , @xmath123 ) of the form @xmath124 when @xmath21 is odd , becomes @xmath125v(z,\lambda)=0.\ ] ] later we will treat the case when @xmath21 is even .
since we scaled the argument of @xmath126 , we must rotate the boundary condition .
we state them in a more general context by using the following definition . _
the stokes sectors _ @xmath127 of the equation ( [ rotated ] ) are @xmath128 see figure [ f : graph1 ] . .
the dashed rays represent @xmath129 it is known from hille @xcite that every nonconstant solution of ( [ rotated ] ) either decays to zero or blows up exponentially , in each stokes sector @xmath127 .
that is , one has the following result .
[ gen_pro ] @xmath130 * for each @xmath131 , every solution @xmath121 of ( [ rotated ] ) is asymptotic to @xmath132^{\frac{1}{2}}\,d\xi\right]\ ] ] as @xmath133 in every closed subsector of @xmath127 . *
if a nonconstant solution @xmath121 of decays in @xmath127 , it must blow up in @xmath134 .
however , when @xmath121 blows up in @xmath127 , @xmath121 need not be decaying in @xmath135 or in @xmath136 .
lemma [ gen_pro ] ( i ) implies that if @xmath121 decays along one ray in @xmath127 , then it decays along all rays in @xmath127 .
also , if @xmath121 blows up along one ray in @xmath127 , then it blows up along all rays in @xmath127 .
this is essentially why we said in the introduction that the boundary conditions with @xmath1 represent all decaying boundary conditions . still with @xmath21 odd , the two rays in map , by @xmath137 , to the rays @xmath138 which are the center rays of the stokes sectors @xmath139 and @xmath140 .
thus , the boundary conditions on @xmath25 become @xmath141 when @xmath21 is even , we let @xmath142 so that becomes @xmath143y(z,\lambda)=0,\ ] ] where @xmath144\ ] ] and hence , @xmath145 .
for these cases , the boundary conditions become @xmath146 next we will introduce sibuya s results , but first we define a sequence of complex numbers @xmath147 in terms of the @xmath97 and @xmath27 , as follows . for @xmath148
fixed , we expand @xmath149 note that @xmath150 do not depend on @xmath27 , so we write @xmath151 for @xmath152 .
so the above expansion without the @xmath153 term still gives @xmath147 for @xmath154 .
we further define @xmath155 if @xmath16 is odd , and @xmath156 if @xmath16 is even .
the following theorem is a special case of theorems 6.1 , 7.2 , 19.1 and 20.1 of sibuya @xcite that is the main ingredient of the proofs of the main results in this paper .
[ prop ] equation ( [ rotated ] ) , with @xmath157 , admits a solution @xmath158 with the following properties . 1 .
@xmath158 is an entire function of @xmath159 and @xmath27 .
2 . @xmath158 and @xmath160 admit the following asymptotic expansions .
let @xmath86 .
then @xmath161,\nonumber\\ f^{\prime}(z , a,\lambda)=&-z^{r_m+\frac{m}{2}}(1+o(z^{-1/2}))\exp\left[-f(z , a,\lambda ) \right],\nonumber\end{aligned}\ ] ] as @xmath4 tends to infinity in the sector @xmath162 , uniformly on each compact set of @xmath163-values . here
@xmath164 3 .
properties and uniquely determine the solution @xmath158 of ( [ rotated ] ) .
4 . for each fixed
@xmath37 and @xmath165 , @xmath166 and @xmath167 also admit the asymptotic expansions , @xmath168\lambda^{-1/4}\exp\left[l(a,\lambda)\right],\label{eq1}\\ f^{\prime}(0,a,\lambda)=&-[1+o(1)]\lambda^{1/4}\exp\left[l(a,\lambda)\right],\label{eq2}\end{aligned}\ ] ] as @xmath117 in the sector @xmath169 , uniformly on each compact set of @xmath37 , where @xmath170 5 .
the entire functions @xmath171 and @xmath172 have orders @xmath173 . in sibuya
s book @xcite , see theorem 6.1 for a proof of ( i ) and ( ii ) ; theorem 7.2 for a proof of ( iii ) ; and theorem 19.1 for a proof of ( iv ) .
moreover , ( v ) is a consequence of ( iv ) along with theorem 20.1 in @xcite .
note that properties ( i ) , ( ii ) and ( iv ) are summarized on pages 112113 of sibuya @xcite .
\(i ) uniformness of the error term in theorem [ main_thm1 ] is essentially due to uniformness of error terms in and .
one can check this by carefully following our proofs . in this paper
, we omit this part of the proof .
\(ii ) throughout this paper , we will deal with numbers like @xmath174 for some @xmath175 , and @xmath176 . as usual , we will use @xmath177\ ] ] and if @xmath178 is specified , then @xmath179=s\left[{{\rm re}\,}(\nu)\frac{2\pi}{m+2}+\arg(\lambda)\right],\quad s\in{\mathbb r}.\ ] ] if @xmath180 then the branch of @xmath181 is chosen to be the negative real axis .
next , we provide an asymptotic expansion of @xmath182 in @xcite . but
first , we recall that for @xmath183 , @xmath184 then @xmath185 also , we define @xmath186 and define @xmath187 see appendix for @xmath119 in terms of some gamma functions as in theorem [ main_thm1 ] .
[ asy_lemma ] let @xmath0 and @xmath37 be fixed .
then there exist constants @xmath188 , @xmath189 , such that @xmath190 as @xmath117 in the sector @xmath191 , uniformly on each compact set of @xmath37 .
see @xcite for a proof .
sibuya @xcite proved the following corollary , directly from theorem [ prop ] , that will be used later in sections [ sec_4 ] and [ sec_5 ] .
[ lemma_decay ] let @xmath37 be fixed .
then @xmath192 as @xmath27 tends to infinity in the sector @xmath193 , and hence @xmath194 as @xmath117 in the sector @xmath195 . in particular , @xmath196 as @xmath117 in any closed subsector of the sector @xmath197 .
in addition , @xmath198 as @xmath117 in any closed subsector of the sectors @xmath199 . based on corollary [ lemma_decay ] , sibuya ( * ? ? ?
* theorem 29.1 ) also proved the asymptotic expansion of the eigenvalues of @xmath15 for @xmath20 . also , sibuya @xcite constructed solutions of that decays in @xmath127 , @xmath131 .
before we introduce this , we let @xmath200 then we have the following lemma , regarding properties of @xmath201 .
[ lemma_25 ] for @xmath37 fixed , and @xmath202 , @xmath203 , and @xmath204 if @xmath205 then @xmath206 next , recall that the function @xmath158 in theorem [ prop ] solves ( [ rotated ] ) and decays to zero exponentially as @xmath207 in @xmath208 , and blows up in @xmath209 .
one can check that the function @xmath210 which is obtained by scaling @xmath211 in the @xmath4-variable , also solves ( [ rotated ] ) .
it is clear that @xmath212 , and that @xmath213 decays in @xmath127 and blows up in @xmath134 since @xmath211 decays in @xmath208 .
since no nonconstant solution decays in two consecutive stokes sectors ( see lemma [ gen_pro ] ( ii ) ) , @xmath214 and @xmath215 are linearly independent and hence any solution of ( [ rotated ] ) can be expressed as a linear combination of these two .
especially , for each @xmath131 there exist some coefficients @xmath216 and @xmath217 such that @xmath218 we then see that @xmath219 where @xmath220 is the wronskian of @xmath221 and @xmath222 . since both @xmath223 are solutions of the same linear equation ( [ rotated ] ) , we know that the wronskians are constant functions of @xmath4 .
also , @xmath224 and @xmath215 are linearly independent , and hence @xmath225 for all @xmath226 . also , the following is an easy consequence of and . for each @xmath227
we have @xmath228 moreover , we have the following lemma that is useful later on .
[ shift_lemma ] suppose @xmath229
. then @xmath230 and @xmath231 , where @xmath232 see sibuya @xcite .
we let @xmath233 , that is , @xmath234
in this section , we introduce asymptotic expansions of @xmath116 as @xmath117 along the rays in the complex plane @xcite . first , we provide an asymptotic expansion of the wronskian @xmath235 of @xmath236 and @xmath221 that will be frequently used later .
[ lemma7 ] suppose that @xmath237 .
then for each @xmath37 , @xmath238\exp\left[l(g^{j}(a),\omega^{2j - m-2}\lambda)+l(a,\lambda)\right],\ ] ] as @xmath117 in the sector @xmath239 next , we provide an asymptotic expansion of @xmath116 as @xmath117 in the sector near the negative real axis .
[ thm_neg ] let @xmath0 , @xmath37 and @xmath240 be fixed
. then @xmath241\exp\left[l(g^{-1}(a),\omega^{-2}\lambda)+l(g(a),\omega^{-m}\lambda)\right],\ ] ] as @xmath242 along the rays in the sector
@xmath243 this is an easy consequence of lemma [ lemma7 ] with @xmath244 and . also , for integers @xmath245 we provide an asymptotic expansion of @xmath116 as @xmath117 in the sector @xmath191 .
[ zero_thm ] let @xmath37 and @xmath246 be fixed .
if @xmath245 then @xmath247\exp\left[l(g^{-1}(a),\omega^{-2}\lambda)-l(a,\lambda)\right]\nonumber\\ & + [ 2\omega^{\frac{1}{2}+\mu(a)+2\nu(a)}+o(1)]\exp\left[l(g(a),\omega^{2}\lambda)-l(a,\lambda)\right],\label{tot_asy}\end{aligned}\ ] ] as @xmath117 in the sector @xmath248 we provide an asymptotic expansion of @xmath116 as @xmath117 along the rays in the upper half plane .
[ lemma_up ] let @xmath245 , @xmath37 and @xmath240 be fixed
. then @xmath249\exp\left[l(g^{-1}(a),\omega^{-2}\lambda)-l(a,\lambda)\right],\ ] ] as @xmath117 in the sector @xmath250 .
also , @xmath251\exp\left[l(g(a),\omega^{2}\lambda)-l(a,\lambda)\right],\ ] ] as @xmath117 in the sector @xmath252 .
we will determine which term in dominates in the upper and lower half planes . since , by , @xmath253 we have @xmath254-\left[{{\rm re}\,}(l(g(a),\omega^{2}\lambda))-{{\rm re}\,}(l(a,\lambda))\right]\nonumber\\ & = k_m\left[\cos\left(-\frac{2\pi}{m}+\frac{m+2}{2m}\arg(\lambda)\right)-\cos\left(\frac{2\pi}{m}+\frac{m+2}{2m}\arg(\lambda)\right)\right]|\lambda|^{\frac{1}{2}+\frac{1}{m}}(1+o(1))\nonumber\\ & = 2k_m\sin\left(\frac{2\pi}{m}\right)\sin\left(\frac{m+2}{2m}\arg(\lambda)\right)|\lambda|^{\frac{1}{2}+\frac{1}{m}}(1+o(1)).\nonumber\end{aligned}\ ] ] thus , the first term in dominates as @xmath117 along the rays in the upper half plane , and the second term dominates in the lower half plane .
this completes the proof . in @xcite
, @xmath255 is used for @xmath256 and asymptotics of @xmath255 are provided .
notice that @xmath257 .
theorem 13 in @xcite implies for the sector @xmath258 theorem 14 in @xcite implies that @xmath247\exp\left[l(g^{-1}(a),\omega^{-2}\lambda)-l(a,\lambda)\right]\nonumber\\ & + [ 2\omega^{1 + 2\mu(a)+4\nu(a)}+o(1)]\exp\left[-l(g^2(a),\omega^{2-m}\lambda)-l(a,\lambda)\right],\nonumber\end{aligned}\ ] ] as @xmath117 in the sector @xmath259 .
one can check that the first term dominates in this sector , by using an argument similar to that in the proof of corollary [ lemma_up ] .
also , theorem 15 in @xcite implies that @xmath260\exp\left[-l(a,\omega^{-m-2}\lambda)-l(^{-2}(a),\omega^{-4}\lambda)\right]\nonumber\\ & + [ 2\omega^{\frac{1}{2}+\mu(a)+2\nu(a)}+o(1)]\exp\left[l(g(a),\omega^{-m}\lambda)-l(a,\omega^{-m-2}\lambda)\right],\nonumber\end{aligned}\ ] ] as @xmath117 in the sector @xmath261 .
one can check that the second term dominates in this sector .
then we replace @xmath27 by @xmath262 to convert the sector here to @xmath263 .
this completes the proof .
[ thm_sector2 ] let @xmath264 and let @xmath37 and @xmath240 be fixed
. then @xmath265\exp\left[l(g^{4}(a),\omega^{-2}\lambda)-l(a,\lambda)\right]\nonumber\\ & -[2i\omega^{\frac{5}{2}}+o(1)]\exp\left[-l(g^2(a),\omega^{-1}\lambda)-l(a,\lambda)\right],\nonumber\end{aligned}\ ] ] as @xmath117 in the sector @xmath266 .
also , @xmath267\exp\left[-l(a,\omega^{-5}\lambda)-l(g^{-2}(a),\omega^{-4}\lambda)\right]\nonumber\\ & + [ 2\omega^{\frac{15}{4}}+o(1)]\exp\left[l(g(a),\omega^{-3}\lambda)-l(a,\omega^{-5}\lambda)\right],\nonumber\end{aligned}\ ] ] as @xmath117 in the sector @xmath268 .
see theorems 14 and 15 in @xcite for a proof .
in this section , we will provide asymptotic expansions of @xmath270 , zeros of which will be closely related with the eigenvalues of @xmath271 . first , we treat the cases when @xmath272 .
let @xmath272 be an integer .
then @xmath270 admits the following asymptotic expansion @xmath273\exp\left[l(g^{-1}(a),\omega^{-2}\lambda)-l(g^{n-1}(a),\omega^{2(n-1)}\lambda)\right],\ ] ] as @xmath117 in the sector @xmath274 also , @xmath275\exp\left[l(g^{-1}(a),\omega^{-2}\lambda)-l(g^{n-1}(a),\omega^{2(n-1)}\lambda)\right]\nonumber\\ & + [ 2\omega^{\frac{2-n}{2}+\mu(g^{-1}(a))}+o(1)]\exp\left[l(g^n(a),\omega^{2n}\lambda)-l(a,\lambda)\right],\label{zero_asy}\end{aligned}\ ] ] as @xmath117 in the sector @xmath276 first we will prove for the sector @xmath277 and the second part of the theorem by induction on @xmath61 .
the case when @xmath278 is trivially satisfied by theorem [ zero_thm ] and corollary [ lemma_up ] since @xmath279 .
suppose that holds in the sector for @xmath280 . from this induction hypothesis
we have @xmath281\exp\left[l(a,\lambda)-l(g^{n-1}(a),\omega^{2(n-1)}\lambda)\right],\label{1steq}\end{aligned}\ ] ] as @xmath117 in the sector @xmath282 that is , @xmath283 also , from lemma [ lemma7 ] if @xmath284 , then we have @xmath285\exp\left[l(g^{j}(a),\omega^{2j - m-2}\lambda)+l(a,\lambda)\right],\ ] ] as @xmath117 in the sector @xmath286 we solve for @xmath287 and set @xmath288 to get @xmath289 set @xmath290 . then since @xmath291 , using , @xmath292\exp\left[l(g^{n}(a),\omega^{2k - m-2}\lambda)+l(g^n(a),\omega^{2n}\lambda)\right]}{\omega^{n-1}[2i\omega^{-\frac{k}{2}}+o(1)]\exp\left[l(g^{k}(a),\omega^{2k - m-2}\lambda)+l(a,\lambda)\right]}\nonumber\\ & + \frac{[2\omega^{-\frac{n-1}{2}+\mu(g^{n-1}(a))}+o(1)]\exp\left[l(a,\lambda)-l(g^{n-1}(a),\omega^{2(n-1)}\lambda)\right]}{\omega^{-1}[2i\omega^{-\frac{k}{2}}+o(1)]\exp\left[l(g^{k}(a),\omega^{2k - m-2}\lambda)+l(a,\lambda)\right]}\nonumber\\ & \times [ 2i\omega^{-\frac{k+1}{2}}+o(1)]\exp\left[l(g^{k}(a),\omega^{2k - m-2}\lambda)+l(g^{-1}(a),\omega^{-2}\lambda)\right]\nonumber\\ & = [ 2\omega^{\frac{2-n}{2}+\mu(g^{-1}(a))}+o(1)]\exp\left[l(g^n(a),\omega^{2n}\lambda)-l(a,\lambda)\right]\nonumber\\ & + [ 2\omega^{\frac{2-n}{2}+\mu(g^{n-1}(a))}+o(1)]\exp\left[l(g^{-1}(a),\omega^{-2}\lambda)-l(g^{n-1}(a),\omega^{2(n-1)}\lambda)\right],\label{asy_sector}\end{aligned}\ ] ] where we used for @xmath293 and for everything else , provided that @xmath27 lies in and that @xmath294 that is , @xmath295 thus , the second part of the theorem is proved by induction .
next in order to prove the first part of the theorem for the sector , we will determine which term in dominates as @xmath117 . to do that
, we look at @xmath296|\lambda|^{\frac{1}{2}+\frac{1}{m}}(1+o(1))\nonumber\\ & = 2k_m\sin\left(\frac{n\pi}{m}\right)\left[\sin\left(\frac{(n-2)\pi}{m}+\frac{m+2}{2m}\arg(\lambda)\right)\right.\nonumber\\ & \qquad\qquad\qquad\qquad\qquad\qquad\left.+\sin\left(\frac{n\pi}{m}+\frac{m+2}{2m}\arg(\lambda)\right)\right]|\lambda|^{\frac{1}{2}+\frac{1}{m}}(1+o(1))\nonumber\\ & = 4k_m\sin\left(\frac{n\pi}{m}\right)\cos\left(\frac{\pi}{m}\right)\sin\left(\frac{(n-1)\pi}{m}+\frac{m+2}{2m}\arg(\lambda)\right)|\lambda|^{\frac{1}{2}+\frac{1}{m}}(1+o(1)),\label{dom_set}\end{aligned}\ ] ] that tends to positive infinity as @xmath117 ( and hence the second term in dominates ) if @xmath297 . we still need to prove for the sector @xmath298 for which we use induction on @xmath61 again . when @xmath278 , holds by lemma [ lemma_up ] .
suppose that in the sector for @xmath280 with @xmath299 . then and with @xmath300 yield @xmath301
if @xmath302 , then @xmath303
. so @xmath304\exp\left[l(g^{n+1}(a),\omega^{2n - m}\lambda)+l(a,\lambda)\right]}\nonumber\\ & + \frac{[2\omega^{\frac{3-n}{2}+\mu(g^{n-1}(a))}+o(1)]\exp\left[l(a,\lambda)-l(g^{n-1}(a),\omega^{2(n-1)}\lambda)\right]}{[2i\omega^{-\frac{n+1}{2}}+o(1)]\exp\left[l(g^{n+1}(a),\omega^{2n - m}\lambda)+l(a,\lambda)\right]}\nonumber\\ & \times [ 2i\omega^{-\frac{n+2}{2}}+o(1)]\exp\left[l(g^{n+1}(a),\omega^{2n - m}\lambda)+l(g^{-1}(a),\omega^{-2}\lambda)\right]\nonumber\\ & = [ -2i\omega^{\frac{3-n}{2}+\mu(g^{-1}(a))+\mu(g^n(a))}+o(1)]\exp\left[-l(g^{n+1}(a),\omega^{2n - m}\lambda)-l(a,\lambda)\right]\nonumber\\ & + [ 2\omega^{\frac{2-n}{2}+\mu(g^{n-1}(a))}+o(1)]\exp\left[l(g^{-1}(a),\omega^{-2}\lambda)-l(g^{n-1}(a),\omega^{2(n-1)}\lambda)\right],\nonumber\end{aligned}\ ] ] where we use the induction hypothesis for @xmath305 and use for @xmath306 and @xmath307 .
next , we use an argument similar to complete the induction step .
thus , the theorem is proved .
next we investigate @xmath308 .
[ bod_thm ] if @xmath245 is an even integer , then @xmath309\exp\left[-l(g^{\frac{m+2}{2}}(a),\lambda)-l(a,\lambda)\right]\nonumber\\ & -[2\omega^{2+\mu(a)+\mu(g^{\frac{m-2}{2}}(a))}+o(1)]\exp\left[-l(g^{\frac{m-2}{2}}(a),\omega^{m-2}\lambda)-l(g^{m}(a),\omega^{m-2}\lambda)\right].\label{thm_eq1}\end{aligned}\ ] ] as @xmath117 in the sector @xmath310
if @xmath245 is an odd integer , then @xmath311\exp\left[-l(g^{\frac{m+1}{2}}(a),\omega^{-1}\lambda)-l(a,\lambda)\right]\nonumber\\ & + [ 2\omega^{\frac{5}{4}}+o(1)]\exp\left[l(g^{m+1}(a),\omega^{-2}\lambda)-l(g^{\frac{m-3}{2}}(a),\omega^{m-3}\lambda)\right].\label{thm_eq2}\end{aligned}\ ] ] as @xmath117 in the sector @xmath312 we will use with @xmath313 , that is , @xmath314 when @xmath16 is even , say @xmath315 , @xmath316 so @xmath317 since @xmath27 lies in , @xmath318 @xmath319\exp\left[l(g^{k+1}(a),\omega^{2k - m}\lambda)+l(a,\lambda)\right]}\nonumber\\ & + \frac{[2\omega^{\frac{3-k}{2}+\mu(g^{k-1}(a))}+o(1)]\exp\left[l(a,\lambda)-l(g^{k-1}(a),\omega^{2(k-1)}\lambda)\right]}{\omega^2 [ 2i\omega^{-\frac{k+1}{2}}+o(1)]\exp\left[l(g^{k+1}(a),\omega^{2k - m}\lambda)+l(a,\lambda)\right]}\nonumber\\ & \times[2\omega^{\frac{3-k}{2}+\mu(g^{2k}(a))}+o(1)]\exp\left[l(g^{k+1}(a),\lambda)-l(g^{2k}(a),\omega^{2(k-1)}\lambda)\right]\nonumber\\ & = [ -2i\omega^{\frac{3-k}{2}+\mu(g^{-1}(a))+\mu(g^{k}(a))}+o(1)]\exp\left[-l(g^{k+1}(a),\lambda)-l(a,\lambda)\right]\nonumber\\ & -[2i\omega^{\frac{3-k}{2}+\mu(g^{k-1}(a))+\mu(g^{2k}(a))}+o(1)]\exp\left[-l(g^{k-1}(a),\omega^{2(k-1)}\lambda)-l(g^{2k}(a),\omega^{2(k-1)}\lambda)\right]\nonumber\\ & = -[2i\omega^{\frac{3-k}{2}+\mu(g^{-1}(a))+\mu(g^{k}(a))}+o(1)]\exp\left[-l(g^{k+1}(a),\lambda)-l(a,\lambda)\right]\nonumber\\ & -[2i\omega^{\frac{3-k}{2}+\mu(a)+\mu(g^{k-1}(a))}+o(1)]\exp\left[-l(g^{k-1}(a),\omega^{2(k-1)}\lambda)-l(g^{2k}(a),\omega^{2(k-1)}\lambda)\right],\nonumber\end{aligned}\ ] ] where we used for @xmath320 and @xmath321 , and for @xmath322 .
finally , we use @xmath323 , to get the desired asymptotic expansion of @xmath324 . next we investigate the case when @xmath16 is odd , say @xmath325 ( so @xmath326 ) . @xmath327 and @xmath328 similarly to the proof of the theorem for @xmath16 even , @xmath329\exp\left[l(g^{k+1}(a),\omega^{2k - m}\lambda)+l(a,\lambda)\right]}\nonumber\\ & + \frac{[2\omega^{\frac{3-k}{2}+\mu(g^{k-1}(a))}+o(1)]\exp\left[l(a,\lambda)-l(g^{k-1}(a),\omega^{2(k-1)}\lambda)\right]}{\omega^{\frac{1}{2 } } [ 2i\omega^{-\frac{k+1}{2}}+o(1)]\exp\left[l(g^{k+1}(a),\omega^{2k - m}\lambda)+l(a,\lambda)\right]}\nonumber\\ & \times[2i\omega^{-\frac{k+1}{2}}+o(1)]\exp\left[l(g^{2k+2}(a),\omega^{2k - m-1}\lambda)+l(g^{k+1}(a),\omega^{-1}\lambda)\right]\nonumber\\ & = [ -2i\omega^{\frac{3-k}{2}+\mu(g^{-1}(a))+\mu(g^{k}(a))}+o(1)]\exp\left[-l(g^{k+1}(a),\omega^{-1}\lambda)-l(a,\lambda)\right]\nonumber\\ & + [ 2\omega^{\frac{2-k}{2}+\mu(g^{k-1}(a))}+o(1)]\exp\left[l(g^{2k+2}(a),\omega^{-2}\lambda)-l(g^{k-1}(a),\omega^{2(k-1)}\lambda)\right]\nonumber\\ & = -[2i\omega^{\frac{3-k}{2}+\frac{m}{2}}+o(1)]\exp\left[-l(g^{k+1}(a),\omega^{-1}\lambda)-l(a,\lambda)\right]\nonumber\\ & + [ 2\omega^{\frac{2-k}{2}+\frac{m}{4}}+o(1)]\exp\left[l(g^{2k+2}(a),\omega^{-2}\lambda)-l(g^{k-1}(a),\omega^{2(k-1)}\lambda)\right],\nonumber\end{aligned}\ ] ] where we use for @xmath320 , and use for @xmath322 and @xmath330 . finally , we use @xmath331 , to get the asymptotic expansion of @xmath324 .
this completes the proof .
the _ order of an entire function _
@xmath332 is defined by @xmath333 where @xmath334 for @xmath335 .
if for some positive real numbers @xmath336 , we have @xmath337\leq m(r ,
g)\leq \exp[c_2 r^{\sigma}]$ ] for all large @xmath338 , then the order of @xmath332 is @xmath339 .
[ zero_free ] let @xmath340 . then the entire functions @xmath270 are of order @xmath173 , and hence they have infinitely many zeros in the complex plane .
moreover , @xmath270 have at most finitely many zeros in the sector @xmath341
in this section , we will relate the eigenvalues of @xmath15 with zeros of some entire function @xmath270 .
suppose @xmath342 is odd with @xmath343
. then becomes by the scaling @xmath120 , and @xmath344 decays in the stokes sectors @xmath345 and @xmath346 .
since @xmath347 and @xmath214 are linearly independent , for some @xmath348 and @xmath349 one can write @xmath350 then one finds @xmath351 also it is easy to see that @xmath27 is an eigenvalue of @xmath15 if and only if @xmath352 if and only if @xmath353 . since @xmath354 ,
by corollary [ zero_free ] @xmath355 has at most finitely many zeros in the sector @xmath356 that is , @xmath357 this is true for all @xmath37 fixed .
next , by symmetry one can show that @xmath355 has at most finitely many zeros in the sector @xmath358 .
we look at @xmath15 with @xmath24 replaced by @xmath359 whose coefficient vector is @xmath360 . then one sees that @xmath353 if and only if @xmath361 .
the latter has at most finitely many zeros in the sector @xmath362 thus , @xmath355 has at most finitely many zeros in the sector @xmath358 , and has infinitely many zeros in the sector @xmath363 since it is an entire function of order @xmath364 .
suppose that @xmath365 is even with @xmath366
. then becomes by the scaling @xmath367 , and @xmath368 decays in the stokes sectors @xmath345 and @xmath136 .
we then see that the coefficient vector @xmath369 of the polynomial @xmath370 becomes @xmath371 now one can express @xmath372 as a linear combination of @xmath214 and @xmath215 as follows .
@xmath373 thus , @xmath27 is an eigenvalue of @xmath15 if and only if @xmath374 .
since @xmath375 by corollary [ zero_free ] , @xmath376 has at most finitely many zeros in the sector @xmath377 that is , @xmath357 this is true for each @xmath37 .
so one can show that @xmath376 has at most finitely many zeros in the sector @xmath358 by symmetry , similar to the cases when @xmath21 is odd .
thus , @xmath378 has infinitely many zeros in the sector @xmath363 .
in this section , we prove theorem [ main_thm1 ] for @xmath379 .
we first treat the case when @xmath21 is odd .
suppose that @xmath380 . recall that when @xmath21 is odd , @xmath27 is an eigenvalue of @xmath15 if and only if @xmath353 .
then since @xmath355 has at most finitely many zeros outside the sector @xmath363 , all the eigenvalues @xmath27 of @xmath15 lie in the sector @xmath363 if @xmath381 is large enough . since @xmath382 we will use to investigate asymptotics of large eigenvalues .
suppose that @xmath353 and @xmath381 is large enough .
then from with @xmath383 , and with @xmath384 and @xmath27 replaced by @xmath385 and @xmath386 , respectively , we have @xmath387&\exp\left[l(g^{k}(a),\omega^{2k}\lambda)-l(g^{-k}(a),\omega^{-2k}\lambda)\right]\nonumber\\ & \times\exp\left[l(g^{k-1}(a),\omega^{2k-2}\lambda)-l(g^{-k+1}(a),\omega^{-2k+2}\lambda)\right]=-\omega^{2\nu(g^k(a))}.\nonumber\end{aligned}\ ] ] next the term @xmath388 $ ] can be absorbed into the exponential function so that we get @xmath389\nonumber\\ & \times\exp\left[l(g^{k-1}(a),\omega^{2k-2}\lambda)-l(g^{-k+1}(a),\omega^{-2k+2}\lambda)+o(1)\right]=-\omega^{2\nu(g^k(a))}.\label{sim_eq1}\end{aligned}\ ] ] for each odd integer @xmath390 , we define @xmath391 where the error term @xmath41 is the same as that in
. then by corollary [ lemma_decay ] , @xmath392 in the sector @xmath363 .
since @xmath393 and @xmath394 , the function @xmath395 maps the region @xmath396 for some large @xmath397 and @xmath363 into a region containing @xmath398 for some large @xmath399 and @xmath400 for some @xmath401 .
so there exists a sequence of @xmath7 in the sector @xmath363 such that @xmath402= -\omega^{2\nu(g^k(a))}$ ] for all large @xmath30 .
so , from , we have @xmath403 thus , since @xmath404 from lemma [ asy_lemma ] , @xmath405 where for @xmath406 , the coefficients @xmath89 are given by @xmath407 since @xmath342 , for @xmath51 , @xmath408 where we used lemma [ lemma_25 ] along with .
notice that shows @xmath409 if @xmath16 is even , then we redefine @xmath410 as the sum of @xmath411 above and @xmath412 ( c. f. , equation ) .
finally , note that @xmath413 this completes the proof .
next we prove theorem [ main_thm1 ] for @xmath379 is even .
proof is very similar to the case when @xmath379 is even .
let @xmath365 for some @xmath414 .
recall that @xmath27 is an eigenvalue of @xmath15 if and only if @xmath415 .
then from , we have @xmath416\nonumber\\ & \times\exp\left[l(g^{k}(\widetilde{a}),\omega^{2k-1}\lambda)-l(g^{-k+1}(\widetilde{a}),\omega^{-2k+1}\lambda)+o(1)\right]=-1,\label{sim_eq2}\end{aligned}\ ] ] where we used @xmath388=\exp[o(1)]$ ] again . like in the case when @xmath379 is odd , from lemma [ asy_lemma ] , @xmath417 where for @xmath406 , @xmath418 where we used @xmath419 and @xmath365 as well .
in this section , we prove theorem [ main_thm1 ] for @xmath78 . we first prove the theorem when
@xmath16 is even , and later , we will treat the cases when @xmath16 is odd . then at the end of the section , we will prove the theorem when @xmath420 , by scaling .
we further divide the case into when @xmath21 is odd and when @xmath21 is even .
suppose that @xmath16 is even and @xmath421 for some @xmath414 .
since @xmath27 is an eigenvalue of @xmath15 if and only if @xmath353 if and only if @xmath422 .
suppose that @xmath422 .
if @xmath381 is large enough , then from , @xmath423\nonumber\\ & \times\exp\left[l(g^{k-1}(a),\omega^{2k-2}\lambda)-l(g^{-k+1}(a),\omega^{-2k+2}\lambda)+o(1)\right]=-\omega^{4\nu(g^{k}(a))}.\nonumber\end{aligned}\ ] ] from lemma [ asy_lemma ] , @xmath424 where for @xmath406 , the coefficients @xmath89 are given by @xmath425 where we used lemma [ lemma_25 ] as well as some other thing as before .
in addition , we used @xmath426 let @xmath427 for some @xmath414 .
then @xmath27 is an eigenvalue of @xmath15 if and only if @xmath415 . if @xmath415 and @xmath381 large enough , then from , @xmath428\nonumber\\ & \times\exp\left[l(g^{k}(\widetilde{a}),\omega^{2k-1}\lambda)-l(g^{-k+1}(\widetilde{a}),\omega^{-2k+1}\lambda)+o(1)\right]=-1.\nonumber\end{aligned}\ ] ] then like before , @xmath429 where for @xmath406 , @xmath430 we divide the case into when @xmath21 is odd and when @xmath21 is even .
let @xmath16 and @xmath431 be odd .
suppose that @xmath27 is an eigenvalue of @xmath15 .
since @xmath27 is an eigenvalue of @xmath15 if and only if @xmath422 , if @xmath381 is large enough , then from , @xmath432\nonumber\\ & \times\exp\left[-l(g^{k+1}(a),\omega^{-2k+1}\lambda)-l(g^{-k+1}(a),\omega^{-2k+2}\lambda)+o(1)\right]=-1.\nonumber\end{aligned}\ ] ] then , from lemma [ asy_lemma ] , @xmath433 where for @xmath406 , @xmath434 where we used @xmath435 let @xmath436 for some @xmath414 . then @xmath27 is an eigenvalue of @xmath15 if and only if @xmath415 . if @xmath415 and @xmath381 large enough , then from , @xmath437\nonumber\\ & \times\exp\left[-l(g^{3k+3}(\widetilde{a}),\omega^{-2k-1}\lambda)-l(g^{k+2}(\widetilde{a}),\omega^{-2k}\lambda)+o(1)\right]=-1.\nonumber\end{aligned}\ ] ] then , from lemma [ asy_lemma ] , @xmath438 where for @xmath406 , @xmath439 theorem [ main_thm1 ] for @xmath440 has been proved .
next we prove theorem [ main_thm1 ] for @xmath420 , by scaling @xmath441 .
suppose that @xmath442 .
if @xmath126 is an eigenfunction of @xmath15 , then @xmath443 solves @xmath444v(z,\lambda ) = \lambda v(z,\lambda),\ ] ] and @xmath445 the coefficient vector of @xmath446 is @xmath447 certainly , @xmath448 also , one can find from that for @xmath449 , @xmath450 moreover , @xmath451 this completes proof of theorem [ main_thm1 ] .
in this section , we prove theorem [ eigen_asy ] . we begin with @xmath452 @xmath453
let @xmath454 introduce the decomposition @xmath455 where @xmath456 then we have @xmath457 thus , @xmath458 and hence @xmath459 thus , one concludes @xmath460 where @xmath461 notice that @xmath462 since @xmath463 .
hence , from along with we have @xmath464 suppose that @xmath465 where @xmath466 and @xmath467 for some @xmath468 .
then from @xmath469 hence , @xmath470 next , @xmath471 we use this in to see that the left hand side of is a power series in @xmath472 . then comparing coefficients of @xmath473 , @xmath474 ,
we have @xmath475 moreover , if @xmath476 ( i. e. , @xmath477 ) then there exists some constant @xmath478 such that @xmath479 now we let @xmath480 where @xmath481 and @xmath482 . if @xmath483 then @xmath484 could be smaller than the error term @xmath485 in , and hence we can not deduce existence of @xmath486 like we do in .
this completes proof of theorem [ eigen_asy ] .
a first few @xmath72 are as follows .
let @xmath0 be an integer .
then @xmath488 suppose that @xmath496 , @xmath497 .
then since @xmath498 we have @xmath499 next , we use the substitution @xmath489 to show @xmath500 finally , we use equations in to complete the proof for @xmath496 , @xmath497 .
the author thanks mark ashbaugh , fritz gesztesy , richard laugesen , boris mityagin , grigori rozenblioum and alexander turbiner for helpful discussions and references .
he also thanks richard laugesen for reading a part of this manuscript and suggestions for improving its presentation . | for integers @xmath0 and @xmath1 , we study the eigenvalue problem @xmath2u(z)=\lambda u(z)$ ] with the boundary conditions that @xmath3 decays to zero as @xmath4 tends to infinity along the rays @xmath5 in the complex plane , where @xmath6 is a polynomial .
we provide asymptotic expansions of the eigenvalue counting function and the eigenvalues @xmath7 .
then we apply these to the inverse spectral problem , reconstructing some coefficients of polynomial potentials from asymptotic expansions of the eigenvalues .
also , we show for arbitrary @xmath8-symmetric polynomial potentials of degree @xmath0 and all symmetric decaying boundary conditions that the eigenvalues are all real and positive , with only finitely many exceptions . _ preprint .
_ = 18pt |
the theory of linear cocycles is a classical and rather developed field of dynamical systems and ergodic theory , whose origins go back to the works of furstenberg , kesten @xcite , furstenberg @xcite and oseledets @xcite .
the simplest examples are the derivative transformations of smooth dynamical systems but the notion of linear cocycle is a lot more general and flexible , and arises naturally in many other situations , e.g. , the spectral theory of schrdinger operators . among the outstanding issues is the problem of simplicity : _
when is it the case that the dimension of all oseledets subspaces is equal to 1 ? _ this was first studied by furstenberg @xcite , guivarch - raugi @xcite and goldsheid - margulis @xcite , who obtained explicit simplicity criteria for random i.i.d .
products of matrices .
bonatti - viana @xcite and avila - viana @xcite have much extended the theory , to include a much broader class of ( hlder continuous ) cocycles over hyperbolic maps .
our purpose in this paper is to initiate the study of simplicity in the context of linear cocycles over partially hyperbolic maps .
the study of partially hyperbolic systems was introduced in the works of brin - pesin @xcite and hirsch - pugh - shub @xcite and has been at the heart of much recent progress in this area . while sharing many of the important features of uniformly hyperbolic systems , partially hyperbolic maps are a lot more flexible and encompass many new interesting phenomena . as we are going to see
, the study of linear cocycles over partially hyperbolic maps introduces a host of new issues .
we will deal with these issues in the context when the partially hyperbolic map is a partially hyperbolic skew - product @xmath1 ( this and other notions mentioned in what follows are defined precisely in section [ s.definitions ] ) .
the case when the map is non - uniformly hyperbolic , that is , when all center lyapunov exponents are non - zero , is better understood .
indeed , the case of @xmath2-dimensional cocycles is covered by viana @xcite and for general @xmath3 it should be possible to combine a symbolic description as in sarig @xcite with the main result of avila , viana @xcite , that deals with cocycles over countable shifts .
for this reason , we focus on the opposite case : namely , we take the skew - product to be mostly neutral along the center direction , meaning that its iterates have bounded derivatives along the verticals @xmath4 . concerning the linear cocycle , we take it to admit strong - stable and strong - unstable holonomies , in the sense of bonatti - gomez - mont - viana @xcite and avila - santamaria - viana @xcite .
the simplicity conditions in our main result , that we are going to state next , may be viewed as extensions of the pinching and twisting conditions in bonatti - viana @xcite and avila - viana @xcite to the present partially hyperbolic setting .
firstly , we call the linear cocycle _ uniformly pinching _ if there exists some fixed ( or periodic ) vertical leaf @xmath5 such that the restriction to @xmath6 of every exterior power of the cocycle admits an invariant dominated decomposition @xmath7 into @xmath8-dimensional subbundles . in particular , this decomposition is continuous and the lyapunov exponents along the factor subbundles @xmath9 are all distinct .
secondly , we say that the linear cocycle is _ uniformly twisting _
if , for any @xmath10-path @xmath11 connecting points @xmath12 and @xmath13 , the push - forward of the decomposition @xmath14 at @xmath15 under the concatenation of strong - stable and strong - unstable holonomies over @xmath11 is in general position with respect to the decomposition @xmath16 at @xmath17 . by the latter , we mean that the image of any sum of @xmath18 subspaces @xmath19 is transverse to the sum of any @xmath20 subspaces @xmath21 , for any @xmath22 .
we say that the linear cocycle is _ uniformly simple _ if it is both uniformly pinching and uniformly twisting . [ teo.a ]
every uniformly simple linear cocycle over a partially hyperbolic skew - product with mostly neutral center direction has simple lyapunov spectrum relative to any invariant probability measure with local partial product structure .
the notion of local partial product structure will be recalled in section [ s.definitions ] , where we will also give a more precise version of the theorem .
indeed , as we will see , the conclusion holds under weaker ( non - uniform ) versions of the pinching and twisting conditions .
moreover , already in the form given in theorem [ teo.a ] , our simplicity criterion holds for a subset of hlder continuous cocycles with non - empty interior .
we will also discuss the issue of density of this subset in section [ s.definitions ] .
* acknowledgements .
* we are grateful to l. backes , f. lenarduzzi , e. pujals and j. yang for numerous discussions .
now let us give the precise definitions of the notions involved , as well as a refined version of theorem [ teo.a ] .
the _ linear cocycle _ defined by a measurable matrix - valued function @xmath23 over an invertible measurable map @xmath24 is the ( invertible ) map @xmath25 given by @xmath26 its iterates are given by @xmath27 where @xmath28 let @xmath29 be an @xmath30-invariant probability measure on @xmath31 such that @xmath32 are integrable . by oseledets
@xcite , at @xmath29-almost every point @xmath33 there exist real numbers @xmath34 and a decomposition @xmath35 into vector subspaces such that @xmath36 for every non - zero @xmath37 and @xmath38 .
the dimension of @xmath39 is called the _ multiplicity _ of @xmath40 . in this work
we assume that the invariant measure @xmath29 is ergodic
. then the lyapunov exponents and the dimensions of the subspaces @xmath39 are constant almost everywhere .
the _ lyapunov spectrum _ of the cocycle is the set of all lyapunov exponents .
the following notion is central to the whole paper : the lyapunov spectrum is _ simple _ if it contains exactly @xmath41 distinct lyapunov exponents or , equivalently , if every lyapunov exponent has multiplicity equal to @xmath8 .
let @xmath42 be any two - sided finite or countable shift .
by this we mean that @xmath43 is the set of two - sided sequences @xmath44 in some set @xmath45 with @xmath46 , and the map @xmath47 is given by @xmath48 let @xmath49 be the distance defined by @xmath50 where @xmath51 , @xmath52 and @xmath53 then @xmath43 is a metric space .
moreover , @xmath47 is a _ hyperbolic homeomorphism _ ( in the sense of @xcite ) , as we are going to explain . given any @xmath54
, we define the _ local stable _ and _ unstable sets _ of @xmath55 with respect to @xmath47 by @xmath56 observe that , taking @xmath57 and @xmath58 , * @xmath59 for any @xmath60 , @xmath61 and @xmath62 ; * @xmath63 for any @xmath60 , @xmath64 and @xmath62 ; * if @xmath65 , then @xmath66 and @xmath67 intersect in a unique point , which is denoted by @xmath68 $ ] and depends continuously on @xmath55 and @xmath69 . by _ partially hyperbolic skew - product _ over the shift map @xmath47 we mean a homeomorphism @xmath70 of the form @xmath71 where @xmath72 is a compact riemann manifold and the maps @xmath73 are diffeomorphisms satisfying @xmath74 where @xmath75 is a constant as in ( i ) - ( ii ) .
we also assume the following hlder condition : there exist @xmath76 and @xmath77 such that the @xmath78-distance between @xmath79 and @xmath80 is bounded by @xmath81 for every @xmath82 .
we say that @xmath30 has _ mostly neutral center direction _ if the family of maps @xmath83 defined for @xmath84 and @xmath54 by @xmath85 have bounded derivatives , that is , if there exists @xmath76 such that @xmath86 [ r.equicontinuity ] clearly , this implies that the @xmath87 is equi - continuous .
when the maps @xmath80 are @xmath88 , equi - continuity alone suffices for all our purposes ( see remark [ equicontinuitysuffices ] ) .
a few comments are in order concerning the scope of the notion of partially hyperbolic skew - product . to begin with ,
the shift @xmath89 may be replaced by a sub - shift @xmath90 associated to a transition matrix @xmath91 . by this
we mean that @xmath92 for every @xmath93 and @xmath94 is the restriction of the shift map @xmath47 to the subset @xmath95 of sequences @xmath44 such that @xmath96 for every @xmath97 .
one way to reduce the sub - shift case to the full shift case is through inducing .
namely , fix any @xmath8-cylinder @xmath98=\{(x_n)_{n\in\integer}\in\hsigma_t : x_0 = i\}$ ] with positive measure and consider the first return map @xmath99 \to [ i]$ ] of @xmath94 to @xmath100 $ ] .
this is conjugate to a full countable shift ( with the return times as symbols ) and it preserves the normalized restriction to the cylinder of the @xmath94-invariant measure .
all the conditions that follow are not affected by this procedure . moreover ,
every linear cocycle @xmath101 over @xmath94 gives rise , also through inducing , to a linear cocycle over @xmath102 whose lyapunov spectrum is just a rescaling of the lyapunov spectrum of @xmath101 . in particular , simplicity may also be read out from the induced cocycle .
moreover , although we choose to formulate our approach in a symbolic set - up , for skew - products over shifts , it is clear that it extends to other situations that are more geometric in nature .
for example , take @xmath103 to be a partially hyperbolic diffeomorphism on a compact @xmath104-dimensional manifold @xmath105 and assume that there exists an embedded closed curve @xmath106 such that @xmath107 and some connected component of @xmath108 is a closed curve .
by @xcite , @xmath102 is conjugate up to finite covering to a skew - product over a linear anosov diffeomorphism of the 2-torus .
thus , using a markov partition for the anosov map , one can semi - conjugate @xmath102 to a partially hyperbolic skew - product over a sub - shift of finite type . in this way
, the conclusions of this paper can be adapted to linear cocycles over such a diffeomorphism .
property is a condition of _ domination _ ( or _ normal hyperbolicity _ , in the spirit of @xcite ) : it means that any expansion and contraction of @xmath79 along the fibers @xmath4 are dominated by the hyperbolicity of the base map @xmath47 . for our purposes ,
its main relevance is that it ensures the existence of strong - stable and strong - unstable `` foliations '' for @xmath30 , as we explain next .
let the product @xmath109 be endowed with the distance defined by @xmath110 where @xmath111 denotes the distance on @xmath43 and @xmath112 is the distance induced by the riemannian metric on @xmath72 .
we consider the _
stable holonomies _
@xmath113 defined for every @xmath55 and @xmath69 with @xmath114 , and
_ unstable holonomies _
@xmath115 defined for every @xmath55 and @xmath69 with @xmath116 .
that these families of maps exist follows from the assumption , using arguments from @xcite ; see for instance @xcite which deals with a similar setting .
we define the _ local strong - stable set _ and the _ local strong - unstable set _ of each @xmath117 to be @xmath118 respectively .
it is easy to check that @xmath119 and analogously on strong - unstable sets for time @xmath120 . throughout ,
we take @xmath29 to be an @xmath30-invariant measure with _ partial product structure _ , that is , a probability measure of the form @xmath121 where : * @xmath122 is a continuous function bounded from zero and infinity ; * @xmath123 is a probability measure supported on @xmath124 ; * @xmath125 is a probability measure supported on @xmath126 ; * @xmath127 is a probability measure on the manifold @xmath72 . for notational convenience ,
we write the boundedness as follows : there exists @xmath128 such that @xmath129 for every @xmath130 , @xmath131 and @xmath132 , @xmath133 , where @xmath134 is defined by @xmath135 observe that when @xmath43 is a finite shift space this is an immediate consequence of compactness and the continuity of @xmath136 .
now define , for @xmath137 , @xmath138 in other words , @xmath139 is the normalization of @xmath140 .
note that the family @xmath141 is a ( continuous ) disintegration of @xmath29 along vertical fibers , that is , with respect to the partition @xmath142 .
the assumption that @xmath29 is invariant under @xmath30 , together with the fact that @xmath143 depends continuously on @xmath55 , implies that @xmath144 we will also see in section [ s.ustates ] that this disintegration is _ holonomy invariant _ : @xmath145 [ r.homoclinic ] in particular , if @xmath146 is a fixed point of the shift map then @xmath147 is invariant under @xmath148 . clearly , it is equivalent to @xmath149 .
moreover , if @xmath150 is a homoclinic point of @xmath146 , that is , a point in @xmath151 , then @xmath152 .
let @xmath153 be a @xmath154-hlder continuous map for some @xmath77 . by this
we mean that there exists @xmath76 such that @xmath155 the linear cocycle defined by @xmath156 over the transformation @xmath157 is the map @xmath158 defined by @xmath159 in what follows we take the cocycle to admit _ strong - stable holonomies _ and _ strong - unstable holonomies _
, in the sense of @xcite .
let us explain this . by _
strong - stable holonomies _ we mean a family of linear transformations @xmath160 , defined for each @xmath161 with @xmath162 and such that , for some constant @xmath163 , 1 .
@xmath164 for every @xmath165 ; 2 .
@xmath166 and @xmath167 for any @xmath168 ; 3 .
@xmath169 .
_ strong unstable holonomies _ @xmath170 are defined analogously , for the pairs @xmath171 with @xmath172 .
it was shown in @xcite that strong - stable holonomies and strong - unstable holonomies do exist , in particular , when the cocycle is
_ fiber bunched_. by the latter we mean that there exist @xmath76 and @xmath173 such that @xmath174 where @xmath75 is a hyperbolicity constant for @xmath30 as in conditions ( i)-(ii ) above . for @xmath175 ,
let @xmath176 denote the space of measurable maps @xmath177 from ( some full @xmath149-measure subset of ) @xmath72 to the grassmannian manifold of all @xmath178-dimensional subspaces of @xmath179 .
for each @xmath180 , consider the following push - forward maps @xmath181 * @xmath182 given by @xmath183 * @xmath184 given , for @xmath185 , by @xmath186 * @xmath187 given , for @xmath188 , by @xmath189 now we state our refined criterion for simplicity of the lyapunov spectrum .
we call the cocycle @xmath190 _ pinching _ if there exists some fixed ( or periodic ) vertical leaf @xmath5 such that the restriction to @xmath6 of every exterior power @xmath191 has simple lyapunov spectrum , relative to the @xmath148-invariant measure @xmath147 ( remark [ r.homoclinic ] ) . in other words ,
the lyapunov exponents @xmath192 are such that , for each @xmath193 and @xmath147-almost every @xmath194 , the sums @xmath195 are all distinct .
it is clear that @xmath190 is pinching if it is uniformly pinching .
take @xmath190 to be pinching and let @xmath196 denote the oseledets decomposition of @xmath190 restricted to @xmath6 .
the maps @xmath197 are defined on a full @xmath147-measure set .
now let @xmath150 be some homoclinic point of @xmath146 and @xmath198 , such that @xmath199 define @xmath200 note that @xmath201 is also defined on a full @xmath147-measure set since the maps @xmath79 and the holonomies @xmath202 and @xmath203 preserve @xmath147 ( remark [ r.homoclinic ] ) .
now let @xmath204 be the @xmath205-matrix defined by @xmath206 we call the cocycle @xmath190 _ twisting _ if , some choice of the homoclinic point @xmath150 , all the algebraic minors @xmath207 of @xmath208 decay sub - exponentially along the orbits of @xmath148 , meaning that @xmath209 and every @xmath210 with @xmath211 .
it is clear that this holds if @xmath190 is uniformly twisting , because in this case the algebraic minors are uniformly bounded .
more generally ( see for instance ( * ? ? ?
* corollary 3.11 ) ) , the property holds whenever the function @xmath212 is @xmath147-integrable .
we say that the cocycle @xmath190 is _ simple _ if it is both pinching and twisting , in the sense of the previous section . that is the case , in particular , if @xmath190 is uniformly pinching and uniformly twisting .
thus theorem [ teo.a ] is contained in the following result : [ teo.b ] if @xmath213 is a simple cocycle then its lyapunov spectrum is simple .
let @xmath214 denote the space of all @xmath154-hlder continuous maps @xmath215 .
the norm @xmath216 defines a topology in @xmath214 that we call _
@xmath154-hlder topology_. we are going to prove in section [ teoremab ] that the uniform pinching and uniform twisting conditions are open with respect to this topology .
thus theorem [ teo.a ] implies [ teo.c ] there is a non - empty open subset of @xmath217 such that the associated linear cocycle @xmath190 over @xmath30 has simple lyapunov spectrum . in many contexts of linear cocycles over _ hyperbolic _
systems , simplicity turns out to be a generic condition : it contains an open and dense subset of cocycles ( precise statements can be found in viana @xcite ) .
this is related to the fact that , in the hyperbolic setting , the pinching and twisting conditions are just transversality conditions , they clearly hold on the complement of suitable submanifolds with positive codimension . at present
, it is unclear how this can be extended to the partially hyperbolic setting .
uniform pinching is surely not a generic condition , in general , but it might be locally generic in some special situations , for instance when @xmath148 is quasi - periodic .
we close this section by outlining the overall strategy of the proof of theorem [ teo.b ] . for every @xmath218
, we want to find complementary @xmath219invariant measurable sections @xmath220 such that the lyapunov exponents of @xmath190 along @xmath221 are strictly larger than those along @xmath222 .
the starting point is to reduce the problem to the case when the maps @xmath79 and the matrices @xmath223 depend on @xmath55 only through its positive part @xmath132 .
this we do in section [ s.martingale2 ] , using the stable holonomies to conjugate the original dynamics to others with these properties . then @xmath157 projects to a transformation @xmath224 on @xmath225 which is a skew - product over the one - sided shift @xmath226 and , similarly , the linear cocycle @xmath227 projects to a linear cocycle @xmath228 over the transformation @xmath30 .
we also denote by @xmath190 and @xmath101 the actions @xmath229 induced by the two linear cocycles on the grassmannian bundles . still in section [ s.martingale2 ] , using very classical arguments , we relate the invariant measures of @xmath30 and @xmath190 with those of @xmath230 and @xmath101 , respectively .
in section [ s.ustates ] we study _
@xmath0-states _ , that is @xmath190-invariant probability measures @xmath231 whose rokhling disintegrations @xmath232 are invariant under unstable holonomies , as well as the corresponding @xmath101-invariant probability measures @xmath233 . here
we meet the first important new difficulty arising from the fact that @xmath30 is only partially hyperbolic .
indeed , in the hyperbolic setting such measures @xmath233 are known to admit _ continuous _ disintegrations @xmath234 along the fibers @xmath235 and this fact plays a key part in the arguments of bonatti - viana @xcite and avila - viana @xcite . in the partially hyperbolic setting ,
the situation is far more subtle : the disintegration @xmath236 along the sets @xmath237 is still continuous , but there is no reason why this should extend to the disintegration @xmath238 along fibers @xmath239 , which is what one really needs . the way we make up for this is by proving a kind of @xmath240-continuity : if @xmath241 in @xmath242 then @xmath243 in @xmath244 .
see proposition [ p.l1 ] for the precise statement .
this also leads to our formulating the arguments in terms of measurable sections @xmath245 of the grassmannian bundle , which is perhaps another significant novelty in this paper .
the properties of such sections are studied in section [ s.zeromes ] .
the key result ( proposition [ p.medidazero ] ) is that , under pinching and twisting , the graph of every invariant grassmannian section has zero @xmath246-measure , for every @xmath247 and any @xmath0-state @xmath231 .
these results build up to section [ s.dirac ] , where we prove that every @xmath0-state @xmath231 has an atomic disintegration .
more precisely ( theorem [ t.dirac ] ) , there exists a measurable section @xmath248 such that , given any @xmath0-state @xmath231 on @xmath249 , we have @xmath250 thus we construct the invariant section @xmath248 in . to find the complementary invariant section @xmath251 , in section [ s.orthogonal ] we apply the same procedure to the adjoint cocycle @xmath252 , that is , the linear cocycle defined over @xmath253 by the function @xmath254 we check ( proposition [ p.twist.adj ] ) that this cocycle @xmath252 is pinching and twisting if and only if @xmath190 is .
so , the previous arguments yield a @xmath252-invariant section @xmath255 related to the @xmath0-states of @xmath252 .
then we just take @xmath256 .
finally , in section [ s.teob ] we check that the eccentricity , or lack of conformality , of the iterates @xmath257 goes to infinity @xmath29-almost everywhere ( see proposition [ p.infinito ] ) and we use this fact to deduce that every lyapunov exponent of @xmath190 along @xmath221 is strictly larger than any of the lyapunov exponents of @xmath190 along @xmath222 . at this stage
the arguments are again very classical .
this concludes the proof of theorem [ teo.b ] .
theorem [ teo.c ] is proven in section [ teoremab ] .
the two appendices contain material that seems to be folklore , but for which we could not find explicit references . in appendix [ a.densel1 ]
we show that continuous maps are dense in the corresponding @xmath240 space , whenever the target space is geodesically convex .
let us start by fixing some terminology , once and for all .
we use @xmath258 to denote the identity transformation in a set @xmath259 .
similarly , @xmath260 will always denote the distance in a metric space @xmath259 .
recall that @xmath261 .
thus we write every @xmath180 as @xmath262 with @xmath263 and @xmath264 . for simplicity , we also write @xmath265 and @xmath266 .
let @xmath267 be the canonical projection given by @xmath268 and let @xmath269 be the one - sided shift . given points @xmath180 and @xmath270 , denote @xmath271 for each @xmath272 . in what follows , @xmath273 is a fixed point of @xmath47 and @xmath274 is a homoclinic point of @xmath146 .
more precisely , we fix @xmath150 and some @xmath275 so that @xmath276 we also consider @xmath277 and the projection @xmath278 . in other words , @xmath279 similarly to , for each @xmath247 let @xmath280 note that @xmath281 is a continuous disintegration of @xmath282 with respect to the partition @xmath283 . in this section we derive some useful properties of these disintegrations and .
for this , we assume that the base dynamics is such that each @xmath73 along the center direction depends only on @xmath284 .
this is no restriction in our setting , as we will see in section [ s.martingale2 ]
. then there exists @xmath224 of the form @xmath285 such that @xmath286 we call the _ extremal center lyapunov exponents _ of @xmath30 the limits @xmath287 the oseledets theorem @xcite ensures that these numbers are well defined at @xmath29-almost every point . in our situation : [ l.lyapunovcentral ] @xmath288 . a trivial consequence of the assumption that the maps @xmath289 have uniformly bounded derivatives . [ equicontinuitysuffices ]
when the maps @xmath289 are @xmath88 , equi - continuity alone suffices to get the conclusion of lemma [ l.lyapunovcentral ] .
this can be shown using pesin theory , as follows .
suppose that @xmath290 .
then we have a pesin unstable manifold defined @xmath29-almost everywhere .
this implies that there exist @xmath54 and @xmath291 such that @xmath292 then , given points @xmath293 and @xmath294 in the unstable manifold and given any @xmath295 , there exists @xmath296 such that @xmath297 .
this implies that the family is not equi - continuous .
the proof for @xmath298 is analogous .
let @xmath299 be the projection @xmath300 , recall that we also assume that @xmath24 admits s - holonomies and u - holonomies and @xmath29 has partial product structure .
that implies that @xmath301 has local product structure in the sense of @xcite .
the map @xmath302 is continuous .
moreover , the disintegration @xmath303 is both @xmath0-invariant and @xmath294-invariant : * @xmath304 for every @xmath305 and * @xmath306 for every @xmath307 . by theorem d in @xcite , there exists a disintegration @xmath308 which is continuous , @xmath0-invariant and @xmath294-invariant . by essential uniqueness ,
@xmath309 for @xmath29-almost every @xmath310 .
since both disintegrations are continuous , it follows that they coincide , and so @xmath311 is continuous , @xmath0-invariant and @xmath294-invariant , as claimed .
[ c.muc ] @xmath312 for every @xmath180 , where @xmath284 .
the assumption that @xmath79 only depends on @xmath284 implies that @xmath313 for every @xmath55 and @xmath69 in the same stable set . by the previous lemma
, this implies that @xmath314 whenever @xmath55 and @xmath69 are in the same stable set .
then , @xmath315 for any @xmath55 with @xmath268 .
we also have [ c.inv.muc ] the disintegration @xmath316 is @xmath230-invariant , in the sense that @xmath317 for every @xmath318 .
we have that @xmath319 for @xmath29-almost every @xmath55 , because @xmath29 is @xmath30-invariant .
since @xmath320 is continuous , the identity extends to _ every _ @xmath180 . by corollary [ c.muc ] this implies that @xmath321 for every @xmath318 .
denote @xmath322 and @xmath323 where @xmath324 denotes both canonical projections @xmath325 and @xmath326 .
recall the functions @xmath327 and @xmath328 defined in and . note that @xmath329 is a disintegration of @xmath29 with respect to the partition @xmath330 of @xmath31 and @xmath331 is a disintegration of @xmath282 with respect to the partition @xmath332 of @xmath31 .
[ r.jacobianos_muc ] let @xmath333 be the jacobian of @xmath334 with respect to @xmath149 . using the observation that @xmath331 is a disintegration of the @xmath230-invariant measure @xmath282
, one easily gets that @xmath335 in particular , these jacobians are uniformly bounded from above and below .
analogously , the fact that @xmath311 is invariant under @xmath294-holonomies and @xmath0-holonomies ensures that the jacobians @xmath336 of those holonomies with respect to @xmath149 are uniformly bounded from above and below .
[ l.jacobian1 ] @xmath337 for every @xmath338 and @xmath339 .
fix some @xmath18-cylinder @xmath340 $ ] and let @xmath341 and @xmath342 . noting that @xmath343 is injective , @xmath344 where @xmath345 is the unique point in @xmath346 and we use the change of variables @xmath347 . using lemma [ c.inv.muc ] , it follows that @xmath348 which concludes the proof .
now we find the jacobian of @xmath349 : [ l.jacobian2 ] @xmath350 , where @xmath351 $ ] .
consequently , the jacobians @xmath352 are continuous and bounded from zero and infinity on every @xmath18-cylinder . given @xmath318 and @xmath353 , let @xmath354 $ ] be the @xmath296-cylinder that contains @xmath310 .
then , @xmath355 since @xmath356 is invariant under @xmath47 , @xmath357 it follows that @xmath358 this proves the first part of the conclusion .
the second part is a consequence , since the local product structure implies that @xmath359 is continuous for every cylinder @xmath360 .
for each @xmath361 , the linear cocycle @xmath213 induces a projective cocycle @xmath362 through @xmath363 denote @xmath364 and , given any borel probability measure @xmath231 on @xmath365 , @xmath366 we will be especially interested in the case when @xmath231 is a @xmath30-invariant probability measure that projects down to @xmath29 under the canonical projection @xmath367 on the first coordinate .
our first step is to show that , up to conjugating the cocycle in a suitable way , we may suppose that : 1 .
the base dynamics @xmath79 along the center direction depends only on @xmath310 ; 2 .
the matrix @xmath223 depend only on @xmath368 .
let us explain how such a conjugacy may be defined using the stable holonomies .
let @xmath369 be fixed .
for any @xmath370 , let @xmath371 and then define @xmath372 then @xmath373 is given by @xmath374 notice that @xmath375 does depend only on @xmath376 ( because @xmath377 does ) .
assume that ( a ) is satisfied .
define @xmath378 and then let @xmath379 define @xmath380 .
then @xmath381 which only depends on @xmath382
. clearly , this procedure does not affect the lyapunov exponents .
from now on , we assume that both ( a ) and ( b ) are satisfied .
then , there exist @xmath383 such that @xmath384 consequently , the map @xmath385 satisfies @xmath386 the following general fact will be used to characterize the @xmath101-invariant probability measures : [ p.invariante ] let @xmath387 be a lebesgue probability space and @xmath103 be a measurable map that preserves @xmath222 .
let @xmath388 be the disintegration of @xmath222 with respect to the partition into pre - images @xmath389 .
let @xmath390 be a measurable skew - product over @xmath102 and , given any probability measure @xmath233 on @xmath391 that projects down to @xmath222 , let @xmath392 be its disintegration with respect to the partition into vertical fibers @xmath393 , @xmath394 . then @xmath233 is invariant under @xmath395 if and only if @xmath396 let us begin with some explanations concerning our notations .
let @xmath397 be the canonical quotient map , assigning to each @xmath394 the element @xmath398 of @xmath399 that contains @xmath310 . by definition ,
the rokhlin disintegration of @xmath222 with respect to @xmath399 is the ( essentially unique ) measurable family @xmath400 of probability measures such that @xmath401 for each @xmath402 and @xmath403 for any bounded measurable function @xmath404 .
let @xmath405 be defined by @xmath406 .
observe that @xmath407 is a bijection ; thus we simply write @xmath408 for each @xmath409 .
observe also that @xmath410 thus , may be rewritten as @xmath411 for any bounded measurable function @xmath404 .
define @xmath412 for each @xmath394 .
start by observing that @xmath413 for each @xmath394 .
next , consider any measurable set @xmath414 . by the definition and the assumption that @xmath222 is invariant under @xmath102 , @xmath415 by , this means that @xmath416 for any measurable set @xmath414 .
this proves that @xmath417 is a disintegration of @xmath418 with respect to the partition into vertical fibers . by the essential uniqueness of the disintegration
, it follows that @xmath419 if and only if @xmath420 for @xmath222-almost every @xmath310 . as an immediate consequence ,
we get : [ c.invariante ] in the conditions of proposition [ p.invariante ] , if @xmath102 is invertible then @xmath233 is invariant under @xmath395 if and only if @xmath421 for @xmath222-almost every @xmath394 .
each @xmath422 must coincide with the dirac mass at @xmath423 .
the next proposition shows that every @xmath190-invariant measure @xmath231 that projects down to @xmath231 may be recovered from the corresponding @xmath101-invariant measure @xmath233 , defined by .
recall that we write @xmath424 for each @xmath425 and @xmath426 .
[ p.limmesa ] take @xmath231 to be @xmath190-invariant .
then , for @xmath29-almost every @xmath427 , the sequence @xmath428 converges to @xmath429 in the weak@xmath430 topology .
let @xmath431 be any continuous function .
define @xmath432 thus @xmath433 .
let @xmath434 be the borel @xmath435-algebra of @xmath436 and , for each @xmath426 , consider the @xmath435-algebra @xmath437 of subsets of @xmath31 .
note that the union @xmath438 generates the borel @xmath435-algebra of @xmath31 . for each @xmath339 ,
let @xmath439 be the disintegration of @xmath282 with respect to the partition @xmath440 of @xmath436 into pre - images under @xmath441 .
since @xmath441 is countable - to - one , we may express the disintegration explicitly in terms of the jacobian : @xmath442 [ l.mukmu ] @xmath443 for @xmath282-almost every @xmath444 and every @xmath426 .
since @xmath233 is @xmath101-invariant , we may use proposition [ p.invariante ] to get that @xmath445 as claimed .
[ l.martingale ] the sequence @xmath446 is a martingale .
it is clear from the definition that every @xmath447 is @xmath448-measurable .
hence , @xmath449 is @xmath450-measurable for every @xmath451 .
we are left to check that @xmath452 for any @xmath450-measurable function @xmath453 and any @xmath451 and @xmath454 . by definition , we may write @xmath455 for some @xmath448-measurable function @xmath456 .
since @xmath29 is @xmath30-invariant and @xmath282 is its projection , @xmath457 ( note that @xmath458 ) . by lemma [ l.mukmu ] and the definition of @xmath459 , @xmath460 combining these identities ,
one gets the property .
so , we may use the martingale convergence theorem to conclude that , given any continuous function @xmath431 , @xmath461 exists for @xmath29-almost every @xmath270 .
so , using the fact that the space @xmath462 of continuous functions is separable , one may find a set @xmath463 with full @xmath29-measure such that @xmath464 is defined for every @xmath465 and @xmath466 .
then , by riesz - markov , for each @xmath465 there exists a probability measure @xmath467 on @xmath468 such that @xmath469 for any @xmath470 .
in other words , @xmath471 converges to @xmath467 in the weak@xmath430 topology .
we are left to prove that @xmath472 for @xmath29-almost every @xmath473 .
let @xmath451 and @xmath474 be any @xmath475-measurable function . taking the limit as @xmath476 in
, we get that @xmath477 for every @xmath466 .
in other words , @xmath478 since @xmath29 is @xmath30-invariant and @xmath479 for every @xmath270 , the right - hand side of the previous identity may be rewritten as @xmath480 finally , this last expression may be rewritten as @xmath481 because @xmath231 is @xmath190-invariant .
the last three identities prove that @xmath482 this relation extends to any function on @xmath483 which is a linear combination of functions @xmath484 such that @xmath485 is @xmath450-measurable for some @xmath353 and @xmath377 is continuous .
since these linear combinations are dense in the set of all bounded measurable functions on @xmath249 , it follows that holds for every bounded measurable functions on @xmath249 .
this means that @xmath486 is a disintegration of @xmath231 .
so , by essential uniqueness , @xmath472 for @xmath29-almost every @xmath473 . [ p.limmesb ] for @xmath29-almost
every @xmath427 , any @xmath487 and any choice of points @xmath488 such that @xmath489 and @xmath490 is contained in some @xmath18-cylinder , @xmath491 for every @xmath492 and @xmath487 , define @xmath493 [ l.cancellation ] for every @xmath426 and @xmath339 , @xmath494 using the definitions of @xmath459 and @xmath495 , together with the fact that @xmath29 is @xmath30-invariant , @xmath496 on the other hand , using lemma [ l.mukmu ] , @xmath497 the claim follows .
it follows from lemma [ l.cancellation ] that , for any @xmath498 , @xmath499 the right - hand side is uniformly bounded ( by @xmath500 ) .
it follows that @xmath501 in view of , this implies that @xmath502 finally , since the jacobian is bounded from above on every @xmath18-cylinder ( lemma [ l.jacobian2 ] ) , we deduce that @xmath503 converges to @xmath504 as @xmath505 , for @xmath29-almost every @xmath425 .
then , using the fact that @xmath462 is separable , one can find a full @xmath29-measure set of values of @xmath473 such that the latter convergence holds for every @xmath466 .
the argument extends immediately to the case when @xmath490 is contained in any finite union of @xmath18-cylinders .
a probability measure @xmath231 on @xmath365 is called a _ @xmath0-state _ of @xmath190 if there exist a disintegration @xmath506 with respect to the partition @xmath507 which is invariant under unstable holonomy : @xmath508 where @xmath509 is some full measure set .
let @xmath367 be the canonical projection .
there is some @xmath190-invariant @xmath0-state @xmath231 that projects down to @xmath29 under @xmath510 .
this is entirely analogous to proposition 4.2 of @xcite , and so we only outline the proof .
the idea is to fix some @xmath54 and define a homeomorphism between the measures in @xmath511 that projects to @xmath512 and the @xmath0-states . using that this space in compact we have that the space of @xmath0-states measures is
also compact and @xmath513-invariant , so we have that any accumulation point of @xmath514 is also a @xmath0-state . in the remainder of this section
, @xmath231 denotes any @xmath190-invariant @xmath0-state that projects down to @xmath29 under @xmath510 , and @xmath506 is taken to be a disintegration as in .
let @xmath299 be the canonical projection @xmath300 and denote @xmath515 .
equivalently , @xmath516 for any measurable set @xmath517 . for each @xmath318 , define @xmath518 to be the normalization of @xmath519 then @xmath520 is a continuous disintegration of @xmath521 with respect to the partition into local stable sets @xmath66 .
the measure @xmath521 satisfies the properties of local product structure , boundedness and continuity in ( * ? ? ? * section 1.2 ) . in
what follows , we recall a few results about this type of measures that we will use later . for each @xmath522 and @xmath487
let the _ backward average _ measure @xmath523 of the map @xmath435 be defined by @xmath524 where @xmath525 is the jacobian of @xmath526 with respect to @xmath349 .
[ l.todos1 ] for any cylinder @xmath527\subset \sigma^+$ ] and @xmath528 , @xmath529 where @xmath530 is the disintegration of @xmath521 with respect to the partition @xmath531 .
analogous to ( * ? ? ?
* lemma 2.6 ) . for every cylinder @xmath532\subset \sigma^+$ ] and @xmath533 , @xmath534 ) & \geq \limsup_n \frac{1}{n}\sum_{k=0}^{n-1}\mu^u_{k , x^u}(\left[j\right])\\ & \geq \liminf_n \frac{1}{n}\sum_{k=0}^{n-1}\mu^u_{k , x^u}(\left[j\right ] ) \geq \frac{1}{\kappa}\mu^u(\left[j\right ] ) .
\end{aligned}\ ] ] analogous to ( * ? ? ?
* lemma 2.7 ) . as a direct consequence , for every cylinder @xmath532\subset \sigma^+$ ] and @xmath533 , @xmath535)\geq \frac{1}{\kappa}\mu^u(\left[j\right]).\ ] ] for every @xmath536 let @xmath537 and for every @xmath538 in the same unstable set let @xmath539 be defined by @xmath540 observe that @xmath541 .
now define @xmath542 as @xmath543 observe that for any @xmath544 , @xmath545 because @xmath231 is a @xmath0-state and @xmath546 is @xmath0-invariant .
it is also easy to see that @xmath542 is a disintegration of @xmath231 with respect to the partition @xmath547 .
the main point with the next corollary is that the conclusion is stated for _ every _ @xmath318 .
[ c.invcont2 ] if @xmath548 is a disintegration of an invariant @xmath0-state @xmath231 then @xmath549 for every @xmath550 , every @xmath318 , and @xmath518-almost every @xmath551 . since @xmath231 is @xmath190-invariant , the equality is true for all @xmath552 and @xmath521-almost all @xmath553 or , equivalently , for @xmath554-almost every @xmath555 and @xmath556-almost every @xmath557 .
consider an arbitrary point @xmath318 .
since @xmath556 is positive on open sets , @xmath310 may be approximated by points @xmath558 such that @xmath559 for every @xmath560 and @xmath561-almost every @xmath555 .
since the conditional probabilities of @xmath231 are invariant under unstable holonomies , it follows that @xmath562 for @xmath561-almost every @xmath555 , where @xmath55 is the unique point in @xmath563 . since the measures @xmath564 and @xmath561 are equivalent , this is the same as saying that the last equality holds for @xmath564-almost every @xmath565 , as claimed .
recall that @xmath566 let @xmath567 and @xmath568 be disintegrations of @xmath233 with respect to the partitions @xmath569 and @xmath570 , respectively .
thus each @xmath571 is a probability measure on @xmath572 and each @xmath573 is a probability measure on @xmath468 .
it is easy to check that @xmath574 may be chosen to be continuous with respect to the weak@xmath430 topology , indeed , we will do that in a while .
the corresponding statement for @xmath575 is false , in general .
however , the main point of this section is to show that the family @xmath576 does have some continuity property : [ p.conv ] let @xmath577 be a sequence in @xmath242 converging to some @xmath318 .
then there exists a sub - sequence @xmath578 such that @xmath579 in the weak@xmath430 topology , for @xmath149-almost every @xmath580
. we will deduce this from a somewhat stronger @xmath240-continuity result , whose precise statement will be given in proposition [ p.l1 ] .
the key ingredient in the proofs is a result about maps on geodesically convex metric spaces that we are going to state in lemma [ l.keylemma ] and which will also be useful at latter stages of our arguments .
[ d.geodesicallyconvex ] a metric space @xmath105 is _ geodesically convex _ if there exists @xmath581 such that for every @xmath582 there exist a continuous path @xmath583\to n$ ] with @xmath584 , @xmath585 and @xmath586$.}\ ] ] geodesically convex metric spaces include convex subsets of a banach space , path connected compact metric spaces and complete connected riemannian manifolds , among other examples . the spaces of maps with values in a geodesically convex metric space are analyzed in appendix [ a.densel1 ] .
[ l.keylemma ] let @xmath587 be a geodesically convex metric space and take @xmath588 to be a probability space such that @xmath72 is a normal topological space , @xmath589 is the borel @xmath435-algebra of @xmath72 and @xmath590 is a regular measure .
let @xmath591 and @xmath592 , with @xmath593 and @xmath194 , be such that @xmath594 uniformly in @xmath194 and @xmath595 and , moreover , the jacobian @xmath596 of each @xmath597 with respect to @xmath590 is uniformly bounded .
then @xmath598 for every bounded measurable map @xmath599 .
take @xmath600 to be sufficiently large that @xmath601 for every @xmath293 and @xmath310 .
then , @xmath602 let @xmath603 be a uniform bound for @xmath604 . by proposition [ p.densel1 ] ,
given @xmath605 there exists a continuous map @xmath606 such that @xmath607 then , by change of variables , @xmath608 then @xmath609 by the continuity of @xmath610 , increasing @xmath611 if necessary , @xmath612 the conclusion follows from these inequalities .
[ l.abs.cont.limit ] let @xmath577 be a sequence in @xmath242 converging to some @xmath613 and let @xmath614 be a sequence of integer numbers going to infinity .
assume that @xmath615 converges uniformly to some @xmath616 and @xmath617 converges to some @xmath557 .
then @xmath102 is absolutely continuous with respect to @xmath149 and it has bounded jacobian . by lemma [ c.inv.muc ]
we have that @xmath618 taking @xmath619 we get that @xmath620 which implies that @xmath621 is uniformly bounded . [ p.l1 ]
let @xmath622 be a continuous function , @xmath623 be a sequence in @xmath242 converging to some @xmath318 and @xmath614 be a sequence of integer numbers such that @xmath615 converges uniformly to some @xmath616 .
then @xmath624 converges to @xmath625 in @xmath626 .
denote @xmath627 . fix @xmath628 and let @xmath629 and @xmath630 note that @xmath631 so , applying lemma [ l.keylemma ] with @xmath632 ( the compact metric space of probability measures on @xmath468 ) , @xmath633 , @xmath634 and @xmath635 we get that the sequence @xmath636 converges to @xmath637 in the space @xmath244 . by the definition of disintegration , @xmath638 and so @xmath639 so
, noting that the integrand goes to zero as @xmath505 , for every @xmath628 , the dominated convergence theorem ensures that @xmath640 as we wanted to prove .
this proposition implies that , given any @xmath641 and any continuous @xmath642 , there exist a sub - sequence @xmath643 such that @xmath644 moreover , since the space of continuous functions is separable , one can use a diagonal argument ( see e.g. the proof of ( * ? ? ?
* proposition 2.1.6 ) ) to construct such a sub - sequence independent of @xmath645 .
in other words , @xmath646 this proves proposition [ p.conv ] .
[ continuidad ] the disintegration @xmath647 is continuous .
let @xmath648 be a continuous function .
given any @xmath649 , we have that @xmath650 by proposition [ p.conv ] , up to restricting to a subsequence , we may suppose that @xmath651 converges to @xmath573 in the weak@xmath430 sense , for @xmath149-almost every @xmath293 .
then @xmath652 for @xmath149-almost every @xmath293 . to get the conclusion it suffices to use this observation in the previous inequality , together with dominated convergence .
[ c.outro ] we have @xmath653 for every @xmath318 and @xmath339 .
the @xmath101-invariance of @xmath233 gives that @xmath653 for @xmath556-almost every @xmath310 , the continuity of the disintegration implies that this extends to every @xmath536 .
let @xmath654 be a basis of a given subspace @xmath655 .
the exterior product @xmath656 depends on the choice of the basis , but its projective class does not .
thus we have a well - defined map @xmath657,\ ] ] which can be checked to be an embedding : it is called the _ plcker embedding _ of @xmath468 .
the image is the projectivization of the _ space of @xmath178-vectors _ @xmath658 which we denote by @xmath659 .
this is a closed subset of @xmath660 , and it is invariant under the action induced on @xmath660 by any linear linear transformation @xmath661
. see ( * ? ? ?
* section 2 ) for more information about @xmath178-vectors .
the _ geometric hyperplane _ @xmath662 associated with a subspace @xmath663 is the set @xmath664 of all subspaces @xmath665 which are _ not _ in general position relative to @xmath177 . in other words , @xmath666 this may also be formulated using the plcker embedding : if @xmath667 is any @xmath668-vector representing @xmath177 , then @xmath664 consists of the subspaces @xmath665 represented by @xmath178-vectors @xmath669 such @xmath670 . in
what follows , let @xmath671 denote the space of measurable sections @xmath672 .
we define the _ dual graph _ of each @xmath673 to be @xmath674 let @xmath231 be any @xmath0-state on @xmath249 , @xmath233 be its projection to @xmath675 and @xmath676 be the rokhlin disintegration of @xmath233 along the fibers @xmath239 ( recall section [ s.l1continuity ] ) .
the purpose of this section is to prove the following fact : [ p.medidazero ] we have @xmath677 for any @xmath0-state @xmath231 , any @xmath678 and every @xmath536 . starting the proof of proposition [ p.medidazero ] ,
recall that each @xmath571 is a probability measure on @xmath572 , and @xmath679 where each @xmath573 is a probability measure on @xmath468 .
recall also that @xmath680 is continuous , by corollary [ continuidad ] .
let @xmath318 be fixed for the time being , and consider the functions @xmath681 [ l.gg ] @xmath682 and @xmath683 are measurable functions .
let @xmath684 be an increasing sequence of finite partitions of @xmath685 such that @xmath686 is the partition into points ( that such a sequence exists is clear , e.g. , because the grassmannian is compact ) .
write @xmath687 and then define @xmath688 where @xmath689 denotes the characteristic function of a measurable set @xmath690 . by the rokhlin disintegration theorem ,
each @xmath691 is a measurable function .
it follows that @xmath692 is measurable for every @xmath296 .
moreover , @xmath693 converges to @xmath395 at every point .
thus , @xmath395 is measurable .
analogously , @xmath694 is measurable for every @xmath296 , and @xmath695 converges pointwise to @xmath102 .
thus @xmath102 is measurable .
for each fixed @xmath194 , the function @xmath696 is upper semicontinuous : if @xmath697 converges to @xmath177 then @xmath698 is contained in a small neighborhood of @xmath699 , for every large @xmath296 , and then @xmath700 can not be much larger than @xmath701 . since @xmath685 is compact , it follows that the set @xmath702 is compact and non - empty ( the supremum in the definition of @xmath102 is attained ) for every @xmath194 .
we borrow the following result from ( * ? ? ?
* theorem iii.30 ) : [ t.measurability ] let @xmath703 be a complete probability space and @xmath259 be a separable complete metric space . denote by @xmath704 the borel @xmath705algebra of @xmath259 .
let @xmath706 be the space of compact subsets of @xmath259 , with the hausdorff topology .
the following are equivalent : 1 .
a map @xmath707 from @xmath708 to @xmath706 is measurable ; 2 .
its graph @xmath709 is in @xmath710 ; 3 .
@xmath711 for any open set @xmath712 .
moreover , any of these conditions implies that there exists a measurable map @xmath713 such that @xmath714 for every @xmath715 .
[ l.supattained ] a given @xmath716 realizes the supremum of @xmath717 if and only if @xmath718 for @xmath149-almost every @xmath194 .
moreover , there exists some @xmath719 that does realize this supremum . by lemma [ l.gg ] ,
the set @xmath720 is a measurable subset of @xmath721 .
compare the second condition in theorem [ t.measurability ] .
thus , from the last claim in the theorem , there exists some measurable map @xmath722 such that @xmath723 for every @xmath194 .
in other words , @xmath724 for every @xmath194 . given any @xmath673 we have @xmath725 thus , @xmath726 does realize the supremum .
moreover , is an equality if and only if @xmath727 for @xmath127-almost every @xmath194 .
so far , we kept @xmath613 fixed .
the next proposition shows that the supremum in lemma [ l.supattained ] is actually independent of @xmath310 .
denote @xmath728 [ p.supmeasure ] @xmath729 for every @xmath318 . given any cylinder @xmath730\subset \sigma$ ] ,
choose a positive constant @xmath731)/\kappa$ ] , where @xmath732 is the constant in .
consider any @xmath733 and @xmath734 .
for each @xmath454 and @xmath735 , define @xmath736 by @xmath737 by corollary [ c.outro ] , @xmath738 and so @xmath739)\sup\lbrace m_x(\graf\cv ) : \cv\in\hsec , x\in [ j ] \rbrace \\ & + ( 1-\mu^u_{k,\tilde{x}}([j]))\gamma .
\end{aligned}\ ] ] by , there exist arbitrary large values of @xmath18 such that @xmath740)\geq c$ ] .
thus @xmath741 \rbrace\rbrace\\ & + ( 1-c)\gamma .
\end{aligned}\ ] ] varying @xmath733 and @xmath742 , we can make the left - hand side arbitrarily close to @xmath11 .
it follows that @xmath743 \rbrace\rbrace \geq \gamma.\ ] ] the converse inequality is obvious . thus , we have shown that the supremum over any cylinder @xmath730 $ ] coincides with @xmath11 .
so , given any @xmath536 we may find a sequence @xmath744 such that the sequence @xmath745 converges to @xmath11 , where ( cf . lemma [ l.supattained ] ) each @xmath746 realizes the supremum at @xmath747 .
moreover , by proposition [ p.conv ] , up to restricting to a subsequence we may assume that @xmath748 for every @xmath293 in some full @xmath127-measure set @xmath749
. then @xmath750 for each fixed @xmath751 , consider a sub - sequence @xmath578 along which the @xmath752 is realized .
it is no restriction to suppose that @xmath753 converges to some @xmath754 as @xmath476 . for any @xmath605 ,
let @xmath755 be the closed @xmath756-neighborhood of @xmath177 .
the fact that @xmath757 for every large @xmath18 implies that @xmath758 ( because @xmath755 is closed ) .
thus , making @xmath759 on the right - hand side , @xmath760 replacing this in , we find that @xmath761 as claimed .
[ p.backgamma ] for any @xmath536 and @xmath678 , we have @xmath762 if and only if @xmath763 for every @xmath764 .
since @xmath233 is invariant under @xmath101 , we have @xmath765 for @xmath282-almost every @xmath318 .
moreover , since the disintegration @xmath236 is continuous , the identity extends to every @xmath318 .
then @xmath766 since the maximum is @xmath11 , it follows from these observations that @xmath767 if and only if @xmath768 for every @xmath769 .
as introduced in section [ ss.jacobians2 ] , let @xmath770 and let @xmath771 be the disintegration of @xmath521 with respect to the partition @xmath772 observe that every @xmath518 is equivalent to @xmath512 .
[ l.almostgamma ] for any @xmath536 and any @xmath678 we have that @xmath773 hence , @xmath774 if and only if @xmath775 for @xmath518-almost every @xmath565 .
suppose there are @xmath776,@xmath678 , @xmath777 and a positive @xmath778-measure subset @xmath708 of @xmath779 such that @xmath780 for each @xmath781 , consider the partition of @xmath782 determined by the cylinders @xmath783^s=[-m;\iota_{-m } , \ldots , \iota_{-1}]^s$ ] , with @xmath784 .
since these partitions generate the @xmath435-algebra of @xmath779 , given any @xmath785 we may find @xmath233 and @xmath360 such that @xmath786^s ) \geq ( 1-\varepsilon ) \hmusigma_y([i]^s).\ ] ] note that @xmath783^s \times \{y\}$ ] coincides with @xmath787 , where @xmath788 .
so , using also lemma [ l.todos1 ] , @xmath789^s ) = j_\mu \sigma^m(x ) \ , \hmusigma_y\big(x \cap [ i]^s).\ ] ] by the previous inequality and lemma [ l.todos1 ] , this is bounded below by @xmath790^s ) & = ( 1-\varepsilon ) \ , ( \hsigma_*^m\hmusigma_x ) \big([i]^s\big)\\ & = ( 1-\varepsilon ) \ , \hmusigma_x \big(w^s_{loc}(x)\big ) = 1 -\varepsilon . \end{aligned}\ ] ] in this way we have shown that @xmath791 fix @xmath785 small enough so that @xmath792 . using corollary [ c.invcont2 ] , we find that @xmath793 for @xmath518-almost every @xmath794 .
it follows that @xmath795 which contradicts the definition of @xmath11 .
this contradiction proves the first part of the lemma .
the second one is a direct consequence , using @xmath796 the proof is complete .
recall ( from section [ ss.pinching_twisting ] ) that @xmath146 is a fixed point of @xmath47 and @xmath150 is a homoclinic associated to @xmath146 . from now on , we denote @xmath797 and @xmath798 .
recall also that @xmath799 is such that @xmath800 . by the pinching hypothesis in section [ ss.pinching_twisting ] , the oseledets decomposition of @xmath101 restricted to @xmath72 has the form @xmath196 , at @xmath149-almost every @xmath194 , with @xmath801 for every @xmath802 .
fix a measurable family @xmath803 of bases of @xmath804 with @xmath805 for every @xmath802 .
the matrices of the iterates @xmath806 relative to these bases are diagonal : @xmath807 we are going to use the associated linear bases of @xmath808 and @xmath809 , defined at @xmath149-almost every @xmath194 by @xmath810 and @xmath811 respectively . by lemma
[ l.supattained ] and proposition [ p.supmeasure ] , we may choose @xmath812 such that @xmath813 .
define @xmath814 for @xmath815 . by proposition [ p.backgamma ]
, we also have @xmath816 for every @xmath815 .
let @xmath817 be a representative of @xmath818 , in the sense that @xmath819 is the projective class of @xmath820 , for each @xmath293 .
then denote @xmath821 for each @xmath815 . expressing @xmath822 in terms of the linear bases of @xmath808 , @xmath823
we find that @xmath824 with @xmath825 .
note that @xmath826 , and so @xmath827 order the multi - indices @xmath828 in such a way that the sums @xmath829 are in increasing order .
let @xmath830 be the first multi - index , in this ordering , for which @xmath831 is not essentially zero .
in what follows we assume that @xmath832 is ergodic for @xmath149 .
then @xmath830 is the same for every @xmath194 in a full @xmath149-measure set .
the non - ergodic case can be reduced to this one by ergodic decomposition .
[ l.semperda ] the section @xmath833 satisfies @xmath834 . by the birkhoff ergodic theorem , @xmath835 for @xmath836-almost every @xmath194 .
so , there exist some @xmath295 such that @xmath837 for every @xmath293 in some full @xmath836-measure set . for any @xmath293 in that set we may consider a sub - sequence @xmath838 such that @xmath839 .
then @xmath840 and so @xmath841 for @xmath149-almost every @xmath293 .
we also have that @xmath842 , and then lemma [ l.supattained ] implies that @xmath843 for @xmath149-almost every @xmath293 , as we claimed .
this means that , from the start , we may take @xmath820 to coincide with one of the invariant sections @xmath844 given by the oseledets decomposition , for @xmath836-almost every @xmath194 . define @xmath845 .
we have that @xmath846 and , by lemma [ l.almostgamma ] , @xmath847 for @xmath123-almost all @xmath848 . for each @xmath849 , define @xmath850 , where @xmath851 is the unique point in @xmath852 . since @xmath231 is a @xmath0-state , and @xmath853 , this implies that @xmath854 denote @xmath855 for each @xmath856 and @xmath857 .
in particular , @xmath858 .
we are going to prove that for a large set of @xmath611s the @xmath859 have no intersection .
[ p.intersection ] there exists @xmath860 such that for every @xmath861 and @xmath295 there exist @xmath862 and @xmath863 , with @xmath864 and @xmath865 for any choice of @xmath866 and @xmath867 .
let @xmath868 be such that @xmath869 is an unitary @xmath870 vector that represents @xmath871 in @xmath808 .
we can write it as @xmath872 then @xmath873 let @xmath874 .
given any @xmath875 and @xmath194 such that @xmath876 , there is some non - zero @xmath877 such that @xmath878 write @xmath879 then can be written as @xmath880 where @xmath881 is the sign of @xmath882 .
this may be written as @xmath883 where @xmath884 @xmath885 so , in order to prove that the intersection is necessarily empty , it suffices to show that has no non - zero solutions , in other words , that @xmath886 .
we are going to use the following fact : [ l.nonzerodet ] let @xmath887 , for @xmath888 and @xmath889 , be measurable functions and suppose there exist @xmath890 such that @xmath891 then for every @xmath861 and @xmath295 there exist @xmath892 and @xmath863 with @xmath864 , such that for any choice of a set @xmath893 with @xmath894 , the matrix @xmath895 has non - zero determinant for every @xmath896 in the proof we are going to use the following simple algebraic fact : [ l.vandermondelike ] let @xmath897 be a square matrix with @xmath898 for every @xmath899 .
then @xmath900 where @xmath901 is defined by @xmath902 the assumption ensures that we may write @xmath903.\ ] ] subtracting the first column from each one of the others , we end up with @xmath904,\ ] ] as claimed .
let us write @xmath905 for @xmath906 and @xmath353 .
the hypothesis implies that there exist @xmath907 and @xmath908 with @xmath909 such that @xmath910 let @xmath911 be fixed and define ( compare ) @xmath912 from , and the observation that @xmath913 , we get that @xmath914 in particular , there exists @xmath915 and @xmath916 with @xmath917 such that @xmath918 ( the second condition follows immediately from and the fact that @xmath919 , but we mention it explicitly , for consistency with what follows )
. next , proceed by induction on @xmath920 : suppose that we have defined an increasing sequence of numbers @xmath921 , a decreasing sequence of sets @xmath922 with @xmath923 , and for every @xmath924 a family of measurable functions @xmath925 with @xmath926 satisfying the following relation : @xmath927 for @xmath928 and @xmath929 .
suppose @xmath930 .
fix , @xmath931 and @xmath926 , define the functions @xmath932 inductively by @xmath933 for @xmath934 , @xmath929 and @xmath935 .
then , arguing as in and using induction on @xmath611 , @xmath936 hence we can find @xmath937 and @xmath938 with @xmath939 such that , for every @xmath940 and @xmath926 , @xmath941 now fix @xmath893 with @xmath894 , and define for @xmath942 the matrix @xmath943 then , in view of the recursive relations , we may apply lemma [ l.vandermondelike ] @xmath41-times to @xmath944 to conclude that @xmath945 this completes our argument .
the twisting condition ( section [ ss.pinching_twisting ] ) , implies that @xmath946 and @xmath947 .
then , by , @xmath948 the pinching condition ensures that these sums are all distinct
. then we may apply lemma [ l.nonzerodet ] to the functions @xmath949 we get that there exist @xmath950 and @xmath863 with @xmath864 such that for every @xmath893 with @xmath894 , the matrix @xmath208 defined in is invertible for every @xmath942 .
let @xmath951 and take @xmath76 large enough that @xmath952 .
consider the sequence of integers @xmath953 given by proposition [ p.intersection ] .
then there exists @xmath863 with @xmath954 such that @xmath955 for every @xmath867 and every @xmath956 .
first , suppose that @xmath957 .
the property means that the sets @xmath958 are @xmath105-wise disjoint for every @xmath867 .
then , @xmath959 this is a contradiction because the measure @xmath960 is a probability .
now we treat the general case . by , @xmath961 for every @xmath611 and @xmath512-almost every @xmath962 . in particular
, we may a sequence @xmath963 with that property .
moreover , let @xmath964 be the matrix defined by a system of equations as in , with coefficients depending on @xmath965 instead of @xmath966 .
keep in mind that , by definition , @xmath967 the sequence @xmath968 converges uniformly to @xmath969 when @xmath476 .
let @xmath970 .
by lemma [ l.keylemma ] ( together with the observation that @xmath240 convergence implies convergence almost everywhere over some subsequence ) , up to restricting to some subsequence of values of @xmath18 we have @xmath971 this proves that @xmath972 converges almost everywhere to @xmath973 .
recall that @xmath886 for every @xmath867 , by lemma [ l.nonzerodet ] .
then , there exist @xmath974 with @xmath975 and @xmath976 such that @xmath977 for every @xmath978 and @xmath979 .
then , applying the previous argument with @xmath980 and @xmath972 instead of @xmath146 and @xmath973 , we get that @xmath981 for every @xmath979 .
thus , again we get a contradiction ( because @xmath982 is a probability ) .
the goal of this section is to prove the following theorem : [ t.dirac ] there exists a measurable map @xmath983 such that , given any @xmath0-state @xmath231 on @xmath249 , we have @xmath984 in particular , there exists a unique @xmath0-state .
we begin by recalling the notion of _ quasi - projective map _ , which was introduced by furstenberg @xcite and extended by goldsheid , margulis @xcite .
see also section 2.3 of @xcite for a related discussion .
let @xmath985 $ ] be the canonical projection from @xmath804 minus the origin to the projective space @xmath986 .
we call @xmath987 a _ projective map _ if there is some @xmath988 that induces @xmath989 through @xmath990 ) = [ p(v)]$ ]
. the space of projective maps has a natural compactification , the space of _ quasi - projective maps _ , defined as follows .
the quasi - projective map @xmath991 induced in @xmath986 by a non - zero , possibly non - invertible , linear map @xmath992 is given by @xmath993 ) = [ q(v)].\ ] ] observe that @xmath991 is well defined and continuous on the complement of the _ kernel _
@xmath994:v\in\ker q\}$ ] .
more generally , one calls @xmath995 a _ projective map _ if there is @xmath988 that induces @xmath989 through @xmath996 . furthermore ,
the quasi - projective map @xmath991 induced in @xmath468 by a non - zero , possibly non - invertible , linear map @xmath992 is given by @xmath997 observe that @xmath991 is well defined and continuous on the complement of the _ kernel _ @xmath998 .
the space of quasi - projective maps inherits a topology from the space of non - zero linear maps , through the natural projection @xmath999 .
clearly , every quasi - projective map @xmath991 is induced by some linear map @xmath1000 such that @xmath1001 .
it follows that the space of quasi - projective maps on any @xmath468 is compact for this topology .
the following two lemmas are borrowed from section 2.3 of @xcite : [ l.kernel ] the kernel @xmath1002 of any quasi - projective map is contained in some hyperplane of @xmath468 .
[ l.topology2 ] if @xmath1003 is a sequence of projective maps converging to some quasi - projective map @xmath1000 of @xmath468 , and @xmath1004 is a sequence of probability measures in @xmath468 converging weakly to some probability @xmath1005 with @xmath1006 , then @xmath1007 converges weakly to @xmath1008 .
recall that , given @xmath1009 and @xmath1010 , we write @xmath1011 for every @xmath194 such that the oseledets subspaces @xmath1012 are defined .
we give ourselves the right to denote by @xmath1013 also the direct sum @xmath1014 thus each @xmath1015 is an element of @xmath176 .
let @xmath1016 be the fixed point of @xmath47 and @xmath553 be a homoclinic point of @xmath146 with @xmath1017 .
fix @xmath275 such that @xmath1018 . for each @xmath1019 , denote @xmath1020 and @xmath1021 . observe that @xmath1022 and , similarly , @xmath1023 .
we take advantage of this fact to simplify the notations a bit in the arguments that follow .
[ nova ] let @xmath1024 . for every sequence @xmath1025
there exists a sub - sequence @xmath1026 such that @xmath1027 for @xmath149-almost every @xmath194 .
we have that @xmath1028 and @xmath1029 so @xmath1030 is equal to @xmath1031 note that @xmath1032 converges uniformly to the identity map @xmath1033 , because @xmath1034 converges to @xmath146 .
let @xmath1035 be a full @xmath149-measure such that the conclusion of the oseledets theorem holds at @xmath15 for every @xmath1036 .
we claim that for any @xmath1036 and every sub - sequence of @xmath1037 that converges , the limit is a quasi - projective transformation @xmath991 that maps every point outside @xmath1002 to @xmath1038 .
this can be seen as follows .
given @xmath1039 and @xmath339 , we may write @xmath1040 with coefficients @xmath1041 .
it follows from the oseledets theorem that @xmath1042 is sub - exponential for every @xmath1043 .
recall that @xmath1044 .
then , the action of @xmath1045 in the projectivization of the exterior power is given by @xmath1046 the quotient of the norms converges to zero for any @xmath1047 .
thus , we have that either @xmath1048 or @xmath1049 .
the latter case means that @xmath669 is in the kernel of the limit .
thus , any limit quasi - projective transformation does map the complement of the kernel to @xmath1050 , as claimed . as an immediate consequence
we get that for any @xmath1036 and every sub - sequence of @xmath1051 that converges , the limit is a quasi - projective transformation that maps every point outside the kernel to @xmath1052 . by remark [ r.equicontinuity ] , the family @xmath1053 is equicontinuous . using arzela - ascoli
, it follows that we can find a sub - sequence of @xmath1054 along which the family @xmath1055 converges to some @xmath1056 .
then , by proposition [ l.open ] , there exists a further subsequence @xmath1026 and a full @xmath149-measure set @xmath908 such that @xmath1057 for every @xmath1058 . by proposition [ p.medidazero ] and lemma [ l.abs.cont.limit ] , there exists a full @xmath149-measure set @xmath1059 such that @xmath1060 gives zero weight to every hyperplane of @xmath468 for every @xmath1061 .
then , by lemma [ l.topology2 ] and the previous observations , @xmath1062 along any sub - sequence such that @xmath1063 converges .
this yields the claim of the proposition .
[ r.replace ] the argument remains valid when one replaces the homoclinic point @xmath150 by any other point in @xmath1064 .
it follows from proposition [ p.limmesa ] that there is a full @xmath1065-measure subset of points @xmath1066 such that @xmath1067 for @xmath127-almost every @xmath194 , @xmath1068 and @xmath1069 .
since the shift is ergodic with respect to the projection of @xmath29 on @xmath43 , one may also require that @xmath1070 for some sub - sequence @xmath1071 .
fix any @xmath180 such that both conditions hold .
let @xmath339 be fixed , for the time being .
then implies that @xmath1072 note also that , by definition ,
@xmath1073 we use once more the fact that @xmath1074 is equi - continuous ( remark [ equicontinuitysuffices ] ) .
using ascoli - arzela , it follows that the exists a sequence @xmath1075 such that @xmath1076 converges to some @xmath1077 . up to further restricting to a sub - sequence if necessary , proposition [ l.open ] ensures that @xmath1078 fix any @xmath194 such that the previous claims are fulfilled .
let @xmath1079 be any sub - sequence of @xmath1080 such that @xmath1081 converges to some quasi - projective map @xmath1082 .
then may be written as @xmath1083 if @xmath1084 then , making @xmath1085 , we may use lemma [ l.topology2 ] and proposition [ nova ] to conclude that @xmath1086 .
this gives the conclusion of theorem [ t.dirac ] under this assumption .
proof of theorem [ t.dirac ] : avoiding the kernel @xmath1087,height=192 ] let us show that we can always reduce the proof to this case . recall that @xmath1088 was chosen such that @xmath1089 .
define @xmath370 by @xmath1090 note that @xmath69 depends on @xmath18 and @xmath611 and @xmath1091 that depends on @xmath18 .
we denote @xmath1092 and @xmath1093 . moreover , @xmath1094 and @xmath1095 for each @xmath426 let @xmath1096 be fixed , for the time being .
we have that @xmath1097 with @xmath1098 .
so , @xmath1099 also @xmath1100 , and so @xmath1101 therefore , by propositions [ p.limmesa ] and [ p.limmesb ] , @xmath1102 where @xmath1103 . the last expression may be rewritten as @xmath1104 making @xmath1105 , @xmath1106 and , restricting to a sub - sequence if necessary , @xmath1107 [ l.reduce ] denote @xmath1108 with @xmath1109 .
then there exists a full @xmath149-measure set @xmath863 and a sub - sequence @xmath1054 such that for every @xmath867 there exists a sub - sequence @xmath1110 of @xmath1111 such that @xmath1112 converges to some quasi - projective transformation @xmath1113 .
moreover , @xmath1114 is not in @xmath1115 if @xmath611 is sufficiently large , depending on @xmath293 . as before denote @xmath1116 and @xmath1117 .
we begin by constructing the sub - sequence @xmath1054 .
note that @xmath1118 and @xmath1119 when @xmath1085 .
first , take a sub - sequence of values of @xmath18 such that @xmath1120 converges uniformly to some @xmath377 . since @xmath1121 converges to the identity map
, @xmath1122 also converges uniformly to @xmath377 .
note that @xmath377 is absolutely continuous with respect to @xmath149 , by lemma [ l.abs.cont.limit ] .
recall that @xmath1123 with @xmath1109 . up to restricting the sub - sequence of values of @xmath18 , we may use lemma [ l.keylemma ] to get that @xmath1124 for every @xmath293 in some full @xmath149-measure set @xmath1125 .
this defines the sub - sequence @xmath1054 in the statement . in what follows ,
all the statements on @xmath18 are meant restricted to this sub - sequence .
the twisting condition implies that @xmath1126 for any @xmath1127 and a full @xmath149-measure set of values of @xmath194 .
in other words , @xmath1128 does not belong to any of the the hyperplanes of @xmath468 determined by the oseledets decomposition at the point @xmath15 .
since @xmath377 and @xmath102 are absolutely continuous , there exists a full @xmath149-measure set @xmath1129 of values of @xmath293 such that holds with @xmath293 replaced by @xmath1130 .
take @xmath1131 .
fix any @xmath896 such that in addition @xmath1132 satisfies the conclusion of the oseledets theorem .
consider any sub - sequence @xmath1079 of @xmath1080 such that @xmath1133 converges to some quasi - projective transformation @xmath1000 when @xmath1134 .
then @xmath1135 converges to @xmath1136 when @xmath1134 .
moreover , @xmath1137 next , observe that @xmath1138 where @xmath1139 and @xmath1140 by lemma [ l.kernel ] , the kernel of @xmath1000 is contained in some hyperplane @xmath1141 of @xmath468 .
hence , @xmath1142 is contained in the hyperplane @xmath1143 , of course .
since we take @xmath194 to be such that the oseledets theorem holds at @xmath15 , the backward iterates @xmath1144 are exponentially asymptotic to some hyperplane section @xmath1145 that is defined by a @xmath668-dimensional sum @xmath1146 of oseledets subspaces .
this remains true for @xmath1147 because @xmath1148 converges exponentially fast to to the identity map , since @xmath1034 converges to @xmath146 exponentially fast .
in other words , using , @xmath1149 exponentially fast as @xmath1085
. then , by , we have that @xmath1150 converges to @xmath1151 .
so , @xmath1152 keep in mind that @xmath1017 and @xmath799 is such that @xmath800 .
recall also ( from section [ ss.reduction ] ) that in the present setting all the local stable holonomies @xmath1153 and @xmath1154 are trivial . in particular
, means that @xmath1155 with @xmath1156 , @xmath1157 and @xmath1158 .
then , by the twisting condition , @xmath1159 can not intersect any sum of the form @xmath1160 with @xmath1161 . in particular , the distance between @xmath1162 is positive .
equivalently , the distance between @xmath1163 is positive .
then @xmath1164 does not intersect @xmath1165 which implies that @xmath1166 .
this gives the conclusion of the lemma .
having established lemma [ l.reduce ] , we can now use the same argument as previously , to conclude that @xmath1167 at @xmath149-almost every point also in this case . to do this , observe that for every @xmath233 and @xmath18 fixed there exist a sub - sequence @xmath1168 of @xmath1169 such that @xmath1170 using a diagonal argument , we may choose @xmath1168 to be independent of @xmath18 and @xmath233 .
fix , once and for all , a full @xmath149-measure subset @xmath1171 such that and the conclusions of lemma [ l.reduce ] and proposition [ nova ] ( more precisely , remark [ r.replace ] ) hold for every @xmath1172 . for
each fixed @xmath1172 , fixing @xmath18 sufficiently large and making @xmath1173 go to infinity ( along the sub - sequence given by lemma [ l.reduce ] ) , we find that @xmath1174 then , making @xmath1175 and using lemma [ l.topology2 ] and proposition [ nova ] , @xmath1176 where @xmath1177 .
thus we proved that @xmath1178 is a dirac measure for @xmath521-almost every @xmath54 and @xmath143-almost every @xmath194 .
note also that the set @xmath509 of points @xmath1179 such that @xmath1180 is a dirac measure is measurable , since the map @xmath1181 is measurable and the set of dirac measures is closed in the weak@xmath430 topology is closed , then @xmath1182 is measurable .
thus we have shown that @xmath1182 has total @xmath29-measure , which completes the proof of theorem [ t.dirac ] .
let @xmath1183 be a linear isomorphism and @xmath1184 .
the @xmath178-dimensional _ eccentricity _ of @xmath587 is defined by @xmath1185 we call _ most expanded _
@xmath178-subspace any @xmath1186 that realizes the supremum .
these always exist , since the grassmannian is compact and the expression depends continuously on @xmath221 .
these notions may be expressed in terms of the polar decomposition of @xmath1187 with respect to any orthonormal basis : denoting by @xmath1188 the eigenvalues of the diagonal operator @xmath1189 , in non - increasing order , then @xmath1190 .
the supremum is realized by any subspace @xmath221 whose image under @xmath72 is a sum of @xmath178-eigenspaces of @xmath1189 such that the product of the eigenvalues is @xmath1191 .
it follows that @xmath1192 , and the most expanded @xmath178-subspace is unique if and only if the eccentricity is strictly larger than @xmath8 .
[ p.exentricidad ] for every @xmath1193 , there exists a set @xmath1194 with @xmath1195 such that @xmath1196 , and the image of the most expanded subspace by @xmath1197 converges to @xmath1198 , restricted to the iterates such that @xmath1199 for the proof , let us recall the following proposition , whose proof can be found in @xcite : [ p.eccentricity ] let @xmath1200 be a weak@xmath430 compact family of probabilities on @xmath468 such that all @xmath1201 give zero weight to every hyperplane .
let @xmath1202 be linear isomorphisms such that @xmath1203 converges to a dirac measure @xmath1204 as @xmath1205 , for some sequence @xmath1206 in n. then the eccentricity @xmath1207 goes to infinity and the image @xmath1208 of the most expanding @xmath178-subspace of @xmath1209 converges to @xmath221 . given @xmath1193
take @xmath1210 to be a compact set , with @xmath1211 and such that the restriction of the map @xmath1212 to @xmath1213 is continuous .
this implies that @xmath1214 is a weak@xmath430 compact subset of the space of probability measures of @xmath468 , and every measure in @xmath1200 gives zero weight to every hyperplane .
moreover , @xmath1215 take @xmath1216
. then the claim follows from proposition [ p.eccentricity ] , with @xmath1217 .
fix a continuous hermitian form @xmath1218 in @xmath1219 .
let @xmath1220 be the adjoint cocycle , defined over @xmath1221 by @xmath1222 where @xmath1223 is the adjoint @xmath1224 of the matrix @xmath1225 with respect to the hermitian form .
in other words , @xmath1223 is characterized by @xmath1226 and @xmath1179 .
we have that @xmath1227 it is also easy to see that @xmath1228 respectively , for any @xmath1229 in the same @xmath1230-unstable set and any @xmath1231 in the same @xmath1230-stable set .
the following result is probably well - known but we include a proof since we could not find one in the literature : [ p.lyap_adjoint ] the cocycles @xmath190 and @xmath252 have the same lyapunov exponents . moreover , if @xmath9 , @xmath1232 are the oseledets spaces of @xmath190 then the oseledets spaces of @xmath252 are , respectively , @xmath1233^{\perp } , j=1 , \dots , k.\ ] ] ( we restrict ourselves to the case when @xmath1234 for every @xmath611 , which is the only case we will actually use ; the extension to the general case is rather straightforward . )
first of all , if @xmath1235 and @xmath1236 then @xmath1237 and so @xmath1238 .
this shows that the decomposition @xmath1239 is invariant under @xmath1240 .
let @xmath1241 and @xmath1242 be unit vectors . then @xmath1243 for @xmath1244 , and @xmath1245 where @xmath1246 so , by the oseledets theorem , @xmath1247 on the other hand , @xmath1248 moreover , by the invariance of @xmath1249 , @xmath1250 thus @xmath1251 putting , and together we get that @xmath1252 which proves the statement .
[ p.twist.adj ] @xmath156 is simple , if and only if , @xmath1240 is simple .
applying proposition [ p.lyap_adjoint ] to the restriction of @xmath190 to the periodic leaf @xmath1253 we get that the cocycle @xmath1240 is pinching , if and only if , @xmath156 is pinching .
moreover , the oseledets decomposition @xmath1254 of @xmath1240 is given by the orthogonal complements of the oseledets subspaces of @xmath1255 : @xmath1256^{\perp}.\ ] ] we are going to use this for proving the twisting property , as follows .
let @xmath1257 and @xmath1258 .
denote @xmath1259 then , for any @xmath1260 , @xmath1261 first , we treat the case @xmath1262 .
write @xmath1263 and let @xmath1264 and @xmath1265 be unit vectors .
the twisting condition means that if @xmath1266 then @xmath1267 we need to deduce the corresponding fact for the adjoint . for this , write @xmath1268 hence @xmath1269 by definition @xmath1270 for every @xmath1271 , where @xmath1272 so , by the oseledets theorem , @xmath1273 then @xmath1274 also as @xmath1275 preserves @xmath1276 this is true for @xmath147-almost everywhere . for @xmath1277
the proof is exactly the same , using the inner product induced on @xmath809 by @xmath1278 , that is , @xmath1279 thus , we have shown that @xmath1255 is twisting if and only if @xmath1280 is twisting .
applying proposition [ p.exentricidad ] to the adjoint cocycle we get : there exists a section @xmath1281 which is invariant under the cocycle @xmath1282 and the unstable holonomies of @xmath1240 .
moreover , given any @xmath1283 there exists @xmath1194 with @xmath1195 such that , restricted to the sub - sequence of iterates @xmath18 such that @xmath1284 in @xmath1285 , the eccentricity @xmath1286 goes to infinity and the image @xmath1287 of the most expanded @xmath178-subspace tends to @xmath1288 as @xmath476 .
the next lemma relates the invariant sections of the two cocycles , @xmath101 and @xmath1282 : [ l.ortogonal ] for @xmath29-almost every @xmath1289 , the subspace @xmath1290 is transverse to the orthogonal complement of @xmath1291 .
recall that the stable holonomies of @xmath156 are trivial .
thus , the unstable holonomies of @xmath1240 are trivial .
so , the fact that @xmath1292 is invariant under unstable holonomies means that it is constant on local stable sets of @xmath30 .
then the same is true about his orthogonal complement @xmath1293 , which means that it only depends on @xmath1294 , where @xmath284 .
so the graph of @xmath1295 over @xmath72 has zero @xmath571-measure : @xmath1296 for @xmath1297-almost every @xmath318 .
hence @xmath1298 , which proves the lemma .
let @xmath1179 , and denote by @xmath1299 the orthogonal complement of @xmath1300 . recall that @xmath1292 is invariant under @xmath1240 , in the sense that @xmath1301 hence , @xmath222 is invariant under @xmath1255 . according to lemma [ l.ortogonal ] , we have that @xmath1302 at @xmath29-almost every point
. we want to prove that the lyapunov exponents of @xmath1255 along @xmath221 are strictly bigger than those along @xmath222 . for that , let @xmath1303 be the oseledets decomposition of @xmath1255 restricted to the two invariant sub - bundles , where @xmath1304 corresponds to the smallest lyapunov exponent among @xmath1305 and @xmath1306 the largest among all @xmath1307 .
denote @xmath1308 and @xmath1309 , and then let @xmath1310 and @xmath1311 be the lyapunov exponents associated to these two sub - bundles , respectively .
define @xmath1312 where @xmath1313 . by the oseledets theorem
@xmath1314 the proof of the following proposition is identical to the proof of proposition 7.3 in @xcite : [ p.infinito ] for every @xmath1193 there exist a set @xmath1194 with @xmath1195 such that for @xmath29-almost every @xmath1179 @xmath1315 restricted to the sub - sequence of values @xmath296 for which @xmath1316 .
so now fix some @xmath1193 and @xmath1285 given by proposition [ p.infinito ] .
let @xmath1317 be the first return map : @xmath1318 then we can define the induced cocycle @xmath1319 @xmath1320 where @xmath1321 .
it is well - known ( see ( * ? ? ?
* proposition 4.18 ) ) that the lyapunov exponents of @xmath395 with respect to @xmath1322 are the products of the exponents of @xmath1323 by the average return time @xmath1324 .
thus , to show that @xmath1325 it suffices to prove the corresponding fact for @xmath395 .
define @xmath1326 then @xmath1327 is a sub - sequence of @xmath1328 such that @xmath1316 .
so , using proposition [ p.infinito ] we conclude that @xmath1329 for @xmath29-almost every @xmath1330 .
we need the following well - known fact : let @xmath1331 be a measurable transformation preserving a probability measure @xmath1005 in @xmath708 , and @xmath1332 be a @xmath1005 integrable function such that @xmath1333 at @xmath1005 almost every point
. then @xmath1334 . applying the lemma to @xmath1335 and @xmath1336
we find that @xmath1337 at @xmath29-almost every point . on the other hand
the relation between lyapunov exponents gives that @xmath1338 this means that @xmath1325 , so there is a gap between the first @xmath178 lyapunov exponents and the remaining @xmath870 ones .
since this applies for every @xmath1184 , we conclude that the lyapunov spectrum is simple .
this completes the proof of theorem [ teo.b ] .
we begin by proving the set of uniformly simple cocycles is non - empty . choose @xmath1339 such that its eigenvalues @xmath1340 satisfy @xmath1341 in particular , the eigenvalues @xmath1340 are all distinct .
then let @xmath1342 be the constant cocycle @xmath1343 .
it is clear that @xmath156 admits ( trivial ) stable and unstable holonomies : @xmath1344 for every @xmath1345 .
let us call @xmath1346 and @xmath1347 , and denote @xmath1348 by the definition of holonomies , @xmath1349 for every @xmath295 let @xmath1350 be the open ball of radius @xmath1351 centered at @xmath55 .
fix @xmath1352 such that @xmath1353 , and consider a hlder continuous function @xmath1354 such that @xmath1355 let @xmath1356 be a @xmath41-by-@xmath41 complex matrix with small norm such that all the minors of @xmath1357 are non - zero .
such a matrix does exist : indeed , the set of matrices with non - vanishing minors is open set in the zariski topology , and so it is dense in the norm topology .
define @xmath1358 . then @xmath1359 moreover , as the perturbation does not affect @xmath156 outside @xmath1360 , @xmath1361 and so @xmath1362 .
as the oseledets subspaces in @xmath6 coincide with the eigenspaces of @xmath1255 , the uniform twisting condition is equivalent to saying that the matrix of @xmath1363 with respect to the eigenbase of @xmath1255 has all its minors different from zero .
thus @xmath1364 is uniformly twisting and uniformly pinching . making @xmath1365 , we can make the cocycle @xmath1364 arbitrary close to @xmath156 .
we need to prove that the set of simple cocycles with strong pinching and twisting is open .
[ l.open ] the set of uniformly simple cocycles is open .
first of all , existence of a dominated decomposition into @xmath8-dimensional subspaces is equivalently to having , for each @xmath1366 , stable and unstable cone fields , @xmath1367 and @xmath1368 , such that @xmath1367 is an @xmath178-dimensional cone and @xmath1368 is a @xmath870 dimensional cone .
let @xmath1369 and @xmath1370 .
if @xmath156 is strongly pinching we can take the cones @xmath1371 and @xmath1372 and these are still , respectively , stable and unstable cone fields for every @xmath973 close to @xmath156 .
moreover , if @xmath1373 is sufficiently close we can choose @xmath756 very small .
thus the spaces @xmath1374 varies continuously with @xmath1373 . by remark 3.4 in @xcite , @xmath1375 in the space
@xmath1376 implies that @xmath1377 then if @xmath1375 the spaces @xmath1378 and @xmath1379 , so for @xmath18 sufficiently large the transversality condition , strong twisting , is also verified . by lemma [ l.open ] ,
the subset of uniformly simple cocycles is open . by theorem [ teo.b ]
every simple cocycle has simple lyapunov spectrum .
this concludes the proof of theorem [ teo.c ] .
let @xmath436 be a normal topological space and @xmath105 be a geodesically convex metric space ( definition [ d.geodesicallyconvex ] ) .
denote by @xmath1381 the set of measurable maps @xmath1382 .
given any regular @xmath435-finite borel measure @xmath282 on @xmath436 , fix any point @xmath1383 and define @xmath1384 when @xmath282 is a finite measure , the choice of @xmath1383 is irrelevant : different choices yield the same space @xmath1385 .
consider any @xmath1394 .
given @xmath605 , fix a set @xmath1395 with finite @xmath282-measure and such that @xmath1396 let @xmath1397 be a countable dense subset of @xmath105 .
the family @xmath1398 covers @xmath105 and , consequently , @xmath1399 is a partition of @xmath105
. then @xmath1400 , @xmath1401 is a partition of @xmath1402 into measurable sets .
fix @xmath1403 large enough that @xmath1404 now define @xmath1389 by @xmath1405 then @xmath1406 and @xmath1407 thus @xmath1408 , which proves the lemma .
let @xmath1411 and @xmath1412 , @xmath1413 be as in the definition of the simple map @xmath294 and @xmath581 be as in definition [ d.geodesicallyconvex ] . denote @xmath1414 for each @xmath1413 , consider a compact set set @xmath1415 such that @xmath1416 .
since the @xmath1417 are are pairwise disjoint , and @xmath436 is assumed to be normal , there exist pairwise disjoint open sets @xmath1418 , @xmath1413 with @xmath1419 .
in particular , we also have @xmath1420 . by the urysohn lemma , there are continuous functions @xmath1421 , @xmath1413 such that @xmath1422 now we use the assumption that @xmath105 is geodesically convex . for each @xmath1413 , fix @xmath1423 \to n$ ] with @xmath1424 and @xmath1425
. then define @xmath1382 by @xmath1426 it is clear that @xmath230 is continuous , because the @xmath1427 are open and pairwise disjoint .
moreover , @xmath1428 if either @xmath1429 all the other values of @xmath310 fall into some of the following cases : h. furstenberg .
boundary theory and stochastic processes on homogeneous spaces . in _ harmonic analysis in homogeneous spaces _ , volume xxvi of _ proc .
pure math .
( williamstown ma , 1972 ) _ , pages 193229 . | criteria for the simplicity of the lyapunov spectra of linear cocycles have been found by furstenberg , guivarch - raugi , goldsheid - margulis and , more recently , bonatti - viana and avila - viana . in all the cases ,
the authors consider cocycles over hyperbolic systems , such as shift maps or axiom a diffeomorphisms . in this paper
we propose to extend such criteria to situations where the base map is just partially hyperbolic .
this raises several new issues concerning , among others , the recurrence of the holonomy maps and the ( lack of ) continuity of the rokhlin disintegrations of @xmath0-states .
our main results are stated for certain partially hyperbolic skew - products whose iterates have bounded derivatives along center leaves .
they allow us , in particular , to exhibit non - trivial examples of stable simplicity in the partially hyperbolic setting . |
millisecond pulsars ( msps ) represent a distinct population of rotation - powered pulsars , characterized by short spin periods , @xmath10 ms , and small intrinsic spin - down rates , @xmath11 , implying relatively low surface magnetic dipole field strengths @xmath12 g and large characteristic spindown ages @xmath13 gyr .
these objects have been studied extensively at radio wavelengths since their discovery @xcite but at x - ray energies , where the bulk of observable radiation is expected , this has become possible only relatively recently , owing to their intrinsic faintness ( @xmath14 ergs s@xmath3 ) .
the _ rosat _
observatory was the first to detect x - ray emission from msps , including 10 in the field of the galaxy and one in the globular cluster m28 ( see * ? ? ? * and references therein ) .
in recent years , the unprecedented sensitivity and spatial resolution of the _ chandra x - ray observatory _ have allowed the detection of x - ray counterparts to radio msps in several globular clusters : 47 tucanae and ngc 6397 @xcite , ngc 6752 @xcite , m4 @xcite , possibly m30 @xcite , and m71 ( elsner et al .
, in preparation ) .
in addition , the _ chandra _ and _ xmm - newton _ telescopes have allowed much more detailed studies of the existing sample of x - ray - detected msps than previously possible @xcite . the x - ray emission detected from msps is observed to be both of non - thermal and thermal character . in the x - ray
bright msps , b182124 and b1937 + 21 , the narrow x - ray pulse profiles indicate the emission is beamed and , therefore , must be non - thermal radiation originating in the pulsar magnetosphere @xcite . alternatively ,
as seen in the `` black widow '' pulsar , b1957 + 20 , non - pulsed non - thermal x - rays can be produced by interaction of the relativistic particle wind from the pulsar with matter from a stellar companion or the interstellar medium @xcite .
thermal emission observed from nearby msps ( see e.g. * ? ? ?
* ; * ? ? ?
* ) , is believed to originate from the magnetic polar caps of the underlying neutron star ( ns ) , heated by a backflow of energetic particles from the pulsar magnetosphere ( see e.g. * ? ? ?
* ; * ? ? ?
* and references therein ) . the detection of 16 msps in the globular cluster 47 tucanae ( ngc 104 , henceforth 47 tuc ) by the _ chandra _
acis - i instrument has granted an excellent opportunity for a detailed study of the physical properties of this class of objects @xcite .
these pulsars were discovered at radio wavelengths by @xcite , @xcite , and @xcite . with further pulse timing observations
, @xcite were able to obtain precise positions for 16 of the 22 msps presently known in this globular cluster @xcite .
in addition , @xcite detected the optical counterpart of 47 tuc w , which does not have an accurate radio timing position , by matching the binary phase and period of an optical candidate with that of the radio pulsar .
finally , timing solutions have recently been obtained for 47 tuc r and y ( freire et al .
, in preparation ) thus increasing the number of msps with known positions in this cluster to 19 .
@xcite have reported in detail on the analysis of the x - ray emission properties of the 47 tuc msps , using the _ chandra _ acis - i observations mentioned above .
however , due to the inherent faintness of these sources ( @xmath2 ergs s@xmath3 ) and the limited exposure time ( 72 kiloseconds of total data obtained resulting in 1 - 30 detected counts for each msp ) , the exact nature of the x - ray emission could not be determined with great confidence , although some interesting constraints could be placed . in the meantime ,
additional x - ray observations of 47 tuc have been carried out , allowing further study of the properties of globular cluster msps as well as msps in general . in this paper , we present spectral and long - timescale variability analyses of the x - ray counterparts of msps in the globular cluster 47 tuc based on deep _ chandra _ acis - s observations .
this paper extends the work presented in @xcite .
the work is organized in the following manner : 2 describes the observations and data reduction procedure ; in 3 we discuss the x - ray spectral properties of the msps , while in 4 we investigate the long - term temporal behavior of the x - ray emission ; in 5 we compare the x - ray and radio pulsar properties .
finally , in 6 we discuss the implications of the results presented and offer conclusions in 7 .
the data set presented here consists of four 65-kilosecond observations of the core of the globular cluster 47 tuc with the s3 chip of the _ chandra _ acis - s detector at the focus .
the first three observations were performed between 2002 september 29 and october 3 , while the fourth was carried out on 2002 october 11 .
these observations have resulted in the detection of @xmath15300 x - ray sources within the half - mass radius of 47 tuc , including x - ray counterparts to all of the 19 msps with known positions . in this present paper
we focus on the counterparts of the radio msps ; for a comprehensive analysis of the other x - ray sources , see @xcite .
the initial data reduction and image processing were performed using the ciao 3.0 software package and are also described in detail by @xcite .
figure 1 shows the combined image of all four observations with @xmath16 circles centered at the radio position of each msp .
the ciao tool wavdetect was able to detect an x - ray source at the position of each of the radio msps , including 47 tuc c , l , m , q , and t , which were not detected or only marginally detected in the 2000 acis - i observations @xcite .
the msps 47 tuc g and i , separated by just @xmath17 , could not be individually resolved , whereas 47 tuc f and s , separated by @xmath18 , were again identified as a single elongated source . in this paper , we also include analyses of msp w , whose optical counterpart was found by @xcite to coincide with the x - ray source w29 @xcite , and the msps r and y , whose recently obtained radio timing solutions ( freire et al . , in preparation )
place them at the positions of the x - ray sources w198 and w82 , respectively . to facilitate the spectral and variability analyses , the counts for each msp in the 0.38 kev energy band
were extracted from circular regions with @xmath16 radii around the radio msp positions .
these circles enclose @xmath19@xmath20% of the total energy for 0.51 kev thermal sources .
a slightly more involved extraction method was required for 47 tuc f and s , whose extraction circles overlap , as evident in figure 1 .
the relative contribution of counts from each msp was determined by maximum likelihood fits of two - dimensional gaussians to the five x - ray sources in the vicinity of the radio positions of f and s using the _ sherpa _ application in ciao .
this procedure was carried out in the 0.34 kev range , as well as in three narrow bands ( 0.30.8 kev , 0.81.5 kev , and 1.54 kev ) to allow an approximate spectral analysis . using this method
we find that @xmath21% of the total counts can be attributed to s. for 47 tuc g and i , this was not possible as their radio positions are only @xmath17 apart .
however , the count rate observed from their position is roughly twice that of a typical msp in the sample and is , thus , consistent with being combined emisson from both msps .
the net counts for most msps were obtained by subtracting the background rate , obtained from three source - free regions on the image , from the total . in general ,
the background noise in the energy range of interest was found to be very low ( @xmath151 background count per pixel over @xmath15260 ks of total exposure time ) .
this was not the case for 47 tuc l , which is located very near the bright x - ray source x9 @xcite .
the background for this msp was obtained by considering a section of an annulus of width @xmath22 around the positions of x9 , at the distance to l , while excluding the @xmath16 extraction circle around the msp . using this background region
we find that @xmath1530% of the gross number of counts are due to x9 . for 47 tuc r , situated @xmath23 from the bright source w24 ,
this procedure was not possible due to a combination of the smaller distance to w24 and the large pixel size of the acis detector .
however , the distribution of counts within the @xmath16 extraction region of r suggests that w24 contributes with no more than 10% to the total counts . in the case of msp o ,
the degree of contamination is difficult to gauge due to severe crowding .
in addition , the x - ray counterpart of 47 tuc o , w39 , is very likely a blend of the msp and a variable source ( see 4 ) .
lcccc c & @xmath24 & @xmath25 & @xmath26 & @xmath27 + d & @xmath28 & @xmath29 & @xmath24 & @xmath30 + e & @xmath31 & @xmath32 & @xmath33 & @xmath34 + f+s & @xmath35 & @xmath36 & @xmath28 & @xmath37 + f & @xmath38 & @xmath39 & @xmath40 & @xmath41 + g+i & @xmath42 & @xmath43 & @xmath44 & @xmath45 + h & @xmath46 & @xmath47 & @xmath48 & @xmath49 + j & @xmath50 & @xmath51 & @xmath52 & @xmath53 + l & @xmath54 & @xmath55 & @xmath56 & @xmath57 + m & @xmath58 & @xmath59 & @xmath48 & @xmath60 + n & @xmath28 & @xmath61 & @xmath62 & @xmath63 + o & @xmath64 & @xmath65 & @xmath66 & @xmath67 + q & @xmath68 & @xmath42 & @xmath69 & @xmath70 + r & @xmath71 & @xmath72 & @xmath73 & @xmath74 + s & @xmath75 & @xmath76 & @xmath77 & @xmath78 + t & @xmath62 & @xmath79 & @xmath80 & @xmath81 + u & @xmath62 & @xmath35 & @xmath33 & @xmath82 + w & @xmath29 & @xmath83 & @xmath84 & @xmath85 + y & @xmath73 & @xmath86 & @xmath87 & @xmath88 +
to get a general sense of the x - ray properties of the msps , we have computed the hardness ratios hr1=(medium counts)/(soft counts ) and hr2=(hard counts)/(medium counts ) for each object and constructed an x - ray color - color diagram .
the soft ( 0.30.6 kev ) , medium ( 0.61.5 kev ) , and hard ( 1.54 kev ) energy bands were selected so as to reduce the errors in hr2 , allowing us to discriminate between the two favored spectral models : thermal or non - thermal ( power - law ) .
the net counts in the three bands for each msp are listed in table 1 , while the resulting x - ray color - color diagram is shown in figure 2 . for comparison ,
we have also plotted values of hr1 and hr2 for simple absorbed models of blackbody , neutron star hydrogen atmosphere @xcite , and power - law spectra for a range of temperatures and photon indices , respectively .
the _ chandra _ acis - i observations of 47 tuc revealed that the x - ray colors of the majority of the msps are consistent with those of soft , presumably thermal , sources with blackbody temperatures of @xmath89 kev and emission radii of @xmath90 km @xcite .
the x - ray color - color diagram for the acis - s data ( fig .
2 ) shows that , with the exception of 47 tuc j and w , the msps are indeed soft sources although the positions of the msps on the diagram are suggestive of a more complex spectrum , such as a composite thermal and non - thermal ( power - law ) spectrum or a multi - temperature thermal spectrum . as @xcite and @xcite argue , although consistent with the x - ray colors of the msps , a pure thermal bremsstrahlung spectrum is ruled out as it would require unrealistically large plasma densities , which would be inconsistent with the observed dispersion measures .
the fivefold improvement in counting statistics , resulting from the longer exposure time and the better soft response of the acis - s compared to the acis - i observations , permits a more precise determination of the nature of the x - ray emission via spectral fitting for the 47 tuc msps . for this purpose
, we used the total counts extracted for each msp from the four observations to generate source and background spectra , along with the corresponding response matrix and auxiliary files . the msp pairs 47 tuc f and s , and g and i , were treated as single sources as the individual spectra could not be separated . for the fainter msps
, the extracted counts were grouped into energy bins containing at least 10 counts , while for the brighter msps each bin contained at least 15 counts .
the subsequent least - squares fits to the spectra were performed using the xspec software package and were restricted to photon energies between 0.3 and 8 kev . in the fitting process
, we considered the following spectral models as plausible physical descriptions of the msp emission : a pure blackbody ( bb ) , an unmagnetized neutron star hydrogen atmosphere ( nsa ; * ? ? ?
* ) , a pure power - law ( pl ) , a composite thermal and power - law ( bb+pl or nsa+pl ) , and a two - temperature thermal model ( bb+bb or nsa+nsa ) . throughout this analysis
the hydrogen column density was fixed at the nominal value for 47 tuc ,
@xmath91 @xmath92 @xcite . for the nsa model we assumed a value of @xmath93^{-1/2}-1=0.31 $ ] for the gravitational redshift at the ns surface , appropriate for a @xmath94 m@xmath95 , @xmath96 km star .
thus , the only free parameters for the bb and nsa fits were the effective temperature and bb / nsa normalization , while for the pl model the free parameters were the photon index and the normalization .
lcccccl [ ] c & bb & @xmath97 & @xmath98 & - & @xmath99 & @xmath100 + d & bb & @xmath101 & @xmath102 & - & @xmath103 & @xmath104 + e & bb & @xmath105 & @xmath106 & - & @xmath107 & @xmath108 + f+s & bb & @xmath109 & @xmath110 & - & @xmath111 & @xmath112 + g+i & bb & @xmath113 & @xmath114 & - & @xmath115 & @xmath116 + h & bb & @xmath117 & @xmath118 & - & @xmath119 & @xmath120 + j & bb+pl & @xmath121 & @xmath122 & @xmath123 & @xmath124 & @xmath125 ( @xmath126 ) + l & bb & @xmath127 & @xmath128 & - & @xmath129 & @xmath130 + m & bb & @xmath131 & @xmath132 & - & @xmath133 & @xmath134 + n & bb & @xmath135 & @xmath136 & - & @xmath137 & @xmath138 + o & bb+pl & @xmath139 & @xmath140 & @xmath141 & @xmath142 & @xmath143 ( @xmath144 ) + q & bb & @xmath145 & @xmath132 & - & @xmath146 & @xmath147 + r & bb & @xmath148 & @xmath149 & - & @xmath150 & @xmath151 + t & bb & @xmath152 & @xmath153 & - & @xmath154 & @xmath155 + u & bb & @xmath156 & @xmath157 & - & @xmath119 & @xmath158 + w & bb+pl & @xmath159 & @xmath160 & @xmath161 & @xmath162 & @xmath163 ( @xmath164 ) + y & bb & @xmath165 & @xmath132 & - & @xmath166 & @xmath167 + lcccccl c & nsa & @xmath168 & @xmath169 & - & @xmath170 & @xmath171 + d & nsa & @xmath172 & @xmath173 & - & @xmath174 & @xmath175 + e & nsa & @xmath176 & @xmath177 & - & @xmath178 & @xmath179 + f+s & nsa & @xmath180 & @xmath181 & - & @xmath182 & @xmath183 + g+i & nsa & @xmath184 & @xmath185 & - & @xmath186 & @xmath187 + h & nsa & @xmath188 & @xmath189 & - & @xmath190 & @xmath191 + j & nsa+pl & @xmath192 & @xmath193 & @xmath194 & @xmath195 & @xmath196 ( @xmath197 ) + l & nsa & @xmath198 & @xmath199 & - & @xmath200 & @xmath201 + m & nsa & @xmath202 & @xmath203 & - & @xmath204 & @xmath205 + n & nsa & @xmath206 & @xmath207 & - & @xmath208 & @xmath209 + o & nsa+pl & @xmath210 & @xmath211 & @xmath212 & @xmath213 & @xmath214 ( @xmath215 ) + q & nsa & @xmath216 & @xmath217 & - & @xmath218 & @xmath219 + r & nsa & @xmath220 & @xmath221 & - & @xmath222 & @xmath223 + t & nsa & @xmath224 & @xmath225 & - & @xmath226 & @xmath227 + u & nsa & @xmath228 & @xmath229 & - & @xmath230 & @xmath231 + w & nsa+pl & @xmath232 & @xmath233 & @xmath234 & @xmath235 & @xmath236 ( @xmath237 ) + y & nsa & @xmath238 & @xmath239 & - & @xmath240 & @xmath241 + from the spectral fits we find that many of the 47 tuc msp spectra are well described by a purely thermal spectrum ( bb or nsa ) . for the bb model
the fits yielded effective temperatures and radii of @xmath242 k and @xmath243 km , respectively , while in the case of the nsa model we obtained @xmath244 k and @xmath245 km .
tables 2 and 3 list the best fit model , the corresponding values for the fitted parameters , and the unabsorbed flux for each msp for the bb and nsa model fits , respectively .
figure 3 shows representative thermal spectra for the 47 tuc msps .
it is apparent from the values of @xmath246 listed in tables 2 and 3 , as well as figure 3 , that the spectra of several msps are poorly fit by a single temperature model , especially at energies above @xmath151.5 kev .
in addition , we find the fits to be unacceptable for a pl or a composite pl+bb / pl+nsa model as well . motivated by the results for the nearby msps j04374715 @xcite and j0030 + 0451 @xcite we attempted fitting a two - temperature polar cap model to the brighter thermal - emitting 47 tuc msps , namely , d , e , l , and r. we found that a two - temperature spectrum can account for the excess at @xmath2471.5 kev apparent in figure 3 , resulting in an improved fit , although the limited count statistics do not allow us to reliably constrain the emission areas .
nonetheless , if we fix the two temperatures to values comparable to those of j04374715 and j0030 + 0451 we obtain very similar results for the effective radii of the 47 tuc msps to those obtained for the two field msps .
specifically , in the case of the nsa model , for @xmath248 k and @xmath249 k @xcite the fits yielded @xmath250 km and @xmath251 km , respectively , while for a bb model with @xmath252 k and @xmath253 k @xcite , we obtain @xmath254 m and @xmath255 m. if we extend the analogy further , it seems plausible that the thermal emission from all 47 tuc msps consists of two components as well .
this imples that the derived values of @xmath256 and @xmath257 listed in tables 2 and 3 for the single component fit represent averages , weighted based on the relative temperatures and areas of the two emitting regions . for the fainter msps , such as 47 tuc c , m ,
n , q , and t , a pure power - law model , with a fairly steep spectrum ( @xmath258 ) , also yields an acceptable fit .
however , we believe that this likely stems from the fact that a two - temperature spectrum appears power - law - like in the energy range of interest ( see e.g. fig . 2 in * ? ? ? * ) .
we also note that a broken power - law model , with a break at @xmath259 kev , is also a good description of the observed emission from the 47 tuc msps . in principle
, such a spectrum could arise due to a deficit of radiating high - energy particles in the pulsar magnetosphere .
a break in the spectrum is observed in young normal pulsars though at much higher energies ( a few gev , see e.g. , * ? ? ?
. however , evidence against a purely non - thermal model comes from optical observations of psr j0030 + 0451 , an msp with a qualitatively similar spectrum to those of the 47 tuc pulsars , which show that such a model grossly overestimates ( by a factor of @xmath15500 ) the optical flux upper limits when extrapolated to lower energies @xcite . for the most luminous msps in the sample , namely 47 tuc j , o , and w , a composite nsa+pl or bb+pl model yields an acceptable fit , with the pl component contributing @xmath1570% , @xmath1550% , and @xmath1575% of the total flux , respectively
the spectral fits , shown in figure 4 , were found to be acceptable for photon indices @xmath260 .
we note that although an f - test does not indicate that a composite spectrum is statistically prefered over a pure pl ( but with a steeper photon index of @xmath261 ) , the fainter thermal component is very likely genuine , given that the inferred values for @xmath256 and @xmath257 are very similar to those of the other 47 tuc msps .
the non - thermal x - rays from msp w , and probably msp j and o as well , most likely originate in an intrabinary shock ( see 4 ) although for j and o at least a portion of this emission may originate in the pulsar magnetosphere . for 47 tuc o , which is near the crowded center of the globular cluster
, the pl component may also be due to a background x - ray source .
the spectra of all other 47 tuc msps may , in fact , contain a faint magnetospheric emission component , which is undetectable in our observations due to the limited number of counts .
it is important to note that , in order to model the nsa and bb spectra of msps properly , one needs to take into account the rotation of the ns , which induces a modulation in the projected area of the polar cap(s ) at the spin period and , in the case of a nsa , in the observed spectrum as well due to the energy dependence of the limb darkening effects of a ns atmosphere ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
therefore , the derived temperatures and radii represent time - averaged values integrated over the rotation period of the pulsar . as a consequence , the spectral fits may underestimate the true flux by up to a factor of @xmath152 , even if we consider the effects of gravitational redshift and bending of light near the ns surface ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
unfortunately , in order to account for these effects , one requires prior knowledge of the relative orientation of the pulsar spin and magnetic axes and the line of sight , which is unavailable .
although the limited time resolution of the acis - s observations ( 3.2 seconds ) precludes a timing analysis at the millisecond level , these observations in conjunction with the 2000 acis - i observations , permit us to investigate the temporal behavior of the msp emission over a large range of timescales from hours to years . as expected , for the bulk of the sample we find no statistically significant x - ray variability .
for msp o , however , there seems to be a gradual decrease in the photon count rate , by a factor of @xmath15@xmath262 in total , especially at medium energies ( 0.61.5 kev ) , from the second to the fourth acis - s observation .
there is no known physical mechanism which could account for such a substantial variation of the flux on timescales of days from an old msp such as 47 tuc o. on the other hand , such behavior is typical of cataclysmic variables and chromospherically active binaries with @xmath2 ergs s@xmath3 .
thus , we conclude that this variation is most likely due to blending of msp o with such an x - ray source .
this seems likely as o is at the heart of the cluster ( see fig .
1 ) , where the number density of x - ray sources is large .
the msp r is found to be marginally variable between observations , while msp u exhibits possible variability within the third observation @xcite . in both cases ,
the apparent variability is likely spurious considering the probability of spurious detection and the number ( 19 ) of trials .
for the 12 binary msps , we can also examine whether there exist any variations in the x - ray flux as a function of orbital position . in this sense , 47 tuc j , o , r , and w are of particular interest as they are in very compact binaries with periods of 2.9 , 3.3 , 1.6 , and 3.2 hours , respectively .
the former three msps are bound to very low mass degenerate companions ( @xmath263 m@xmath95 ) and are eclipsed regularly at radio wavelengths for 10 - 25% of their orbits @xcite .
msp w , on the other hand , has a significantly more massive ( @xmath264 m@xmath95 ) main - sequence secondary @xcite and undergoes eclipses for about 3540% of the entire orbit @xcite . for this analysis ,
the observations for the binary msps were first folded at the binary period using the latest values for the orbital periods @xmath265 and the epochs of ascending node @xmath266 ( * ? ? ? * freire et al .
unpublished data ) .
the net counts were then grouped in phase bins sized so as to allow detection of any large - amplitude variations in the x - ray photon count rate at the eclipse phases .
for the same purpose , we also performed kolmogorov - smirnov and cramer - von mises tests on the folded but unbinned data , using an integer number of binary orbits . for 47 tuc j ,
o , and r we find no detectable variation in the x - ray count rate at any orbital phase and are only able to set very crude upper limits on the amplitude of variation of @xmath26760 - 80% .
in contrast , 47 tuc w exhibits dramatic variations in the x - ray flux ( significant at the 99.9% level ) as a function of orbital phase . this behavior can be naturally explained by the existence of a swept - back shocked stream of gas formed by interaction of the pulsar wind with matter from the irradiated companion issuing through the inner lagrange point .
the properties of this intriguing system and its possible connection to low mass x - ray binaries are described in depth by @xcite .
lcc c & @xmath268 & @xmath269 + d & @xmath270 & @xmath271 + e & @xmath272 & @xmath273 + f & @xmath274 & @xmath275 + g(+i ) & @xmath2761.72 & @xmath277 + h & - & @xmath278 + i & @xmath2767.14 & - + j & @xmath279 & @xmath280 + l & @xmath281 & @xmath282 + m & - & @xmath283 + n & @xmath284 & @xmath285 + o & @xmath286 & @xmath287 + q & @xmath288 & @xmath285 + r & @xmath289 & @xmath290 + s & @xmath291 & @xmath292 + t & @xmath293 & @xmath294 + u & @xmath295 & @xmath296 + w & - & @xmath297 + y & @xmath298 & @xmath285 + ) for the msps in 47 tuc ( filled symbols ) and all other msps detected in x - rays ( open symbols ) .
the x - ray luminosities due to polar cap ( thermal ) , magnetospheric , and shock emission are shown as circles , squares , and triangles , respectively . for 47 tuc j and o , and j0437 - 4715
the individual emission components are plotted separately .
the dashed line corresponds to the best fit for the 47 tuc msps with thermal spectra , while the dotted line shows the linear relation @xmath299 .
the numerical labels for the open symbols correspond to the values listed in the first column of table 5.,scaledwidth=42.0% ]
rotation - powered pulsars , including msps , appear to exhibit a linear relation between their x - ray luminosity @xmath300 and rotational spindown luminosity @xmath301 , where @xmath302 is the ns moment of inertia typically assumed to be @xmath303 g @xmath304 , with @xmath305 @xcite .
however , @xcite have shown that the dependence of @xmath4 on @xmath306 for the 47 tuc msps may be significantly flatter , with @xmath307 . using the present data
, we can re - examine the @xmath6 relation for the 47 tuc msps .
however , the measured values of @xmath308 and , hence , @xmath306 for these msps are significantly affected by acceleration in the gravitational potential of the globular cluster . in order to obtain the intrinsic period derivative @xmath309 , we first determined the three - dimensional position of each msp relative to the cluster center by considering the projected radius and assuming that the observed differences in dispersion measure within the msp sample are solely due to a spread in distance along the line of sight for an assumed uniform intra - cluster plasma @xcite . the cluster acceleration term for each msp
was then computed assuming a king model for the cluster using a central velocity dispersion of @xmath310 km s@xmath3 , core radius @xmath311 , and distance @xmath312 kpc @xcite .
subtracting the acceleration term from the observed @xmath308 yields the intrinsic spindown rate @xmath309 , which was then used to compute @xmath306 .
the resulting values and their uncertainties are listed in table 4 . in this calculation we have not included msp h , which exhibits anomalous variations in @xmath308 , and msp w , which currently has no measured @xmath308 .
figure 5 shows the plot of @xmath4 versus @xmath306 for the sample of msps in 47 tuc based on the values listed in table 4 .
the distance used to compute @xmath4 for the 47 tuc msps was taken to be 4.5 kpc . for the msp pair
f and s , the combined @xmath300 was apportioned based on the procedure described in 2 . in the case of g
and i the individual contribution of each msp to the observed @xmath300 is unknown . also plotted are all msps detected in x - rays in the field of the galaxy as well as in the globular clusters m4 , m28 , m30 , ngc 6397 , and ngc 6752 .
the latest values for the parameters of these msps are summarized in table 5 . here
, we have chosen to compute the x - ray luminosity @xmath313 in a broad energy band ( 0.110 kev ) instead of the _ rosat _ band ( 0.12.4 kev ) in order to take into account the substantial portion of the total flux ( up to @xmath1580% ) that is present beyond 2.4 kev for the msps with hard spectra . in this analysis , we have classified the msps based on the three distinct types of x - ray emission : ( i ) non - thermal magnetospheric , ( ii ) non - thermal shock , and ( iii ) thermal polar cap emission . in the latter category we have also included the msps with undetermined spectral properties as they have very similar luminosities to the pulsars with thermal spectra . in addition , all thermal luminosities have been corrected for the gravitational redshift , assuming @xmath314 . for the msps with known multi - component spectra , such as psr j04374715 and 47 tuc j and o
, the emission has been decomposed into the individual constituents .
the thermal luminosities of the 47 tuc msps alone follow the relation @xmath315 . in this fit
we have taken into account the errors in both @xmath300 and @xmath306 listed in table 4 . as evident in figure 5 , due to the large uncertainties
, the 47 tuc sample appears consistent with both the linear relation and a much flatter trend , such as @xmath307 , predicted by the @xcite polar cap heating model .
it is interesting to note that both field and cluster msps with thermal spectra occupy the same region of the @xmath6 diagram .
this suggests that in terms of x - ray properties the two populations are indistinguishable .
curiously , despite the fundamentally different emission mechanisms , all msps appear to follow the linear relation @xmath316 , and have an efficiency of converting @xmath306 to @xmath4 covering a surprisingly narrow range , with @xmath317 .
this is not expected given that in systems such as 47 tuc j , o , and w , b1957 + 20 , and j17405340 , whose emission most likely originates in an intrabinary shock , @xmath4 is determined , in part , by parameters unrelated to the physics of the pulsar , such as the binary separation , the properties of the companion star , and the volume of the shocked region .
the small range in @xmath318 may stem from the fact that in all cases the high - energy emission is , ultimately , driven by the same underlying particle acceleration and pair production processes occurring in the pulsar magnetosphere .
however , the fact that msps with @xmath319 ergs s@xmath3 appear to have thermal spectra , whereas those with @xmath320 ergs s@xmath3 show strong non - thermal emission , suggests a profound difference in the conditions in the pulsar magnetosphere ( see 6.3 ) .
next , we consider how @xmath4 depends on the pulsar magnetic field as it governs the physical processes responsible both for magnetospheric emission , polar cap heating , and , to some degree , shock emission . in figure @xmath321 , we plot @xmath4 versus @xmath322 g , the magnetic field at the light cylinder ( @xmath323 ) for a simple dipole .
the msps with thermal spectra are consistent with @xmath324 . as with the @xmath325 relation ,
the fit is not particularly constraining due to the large uncertainties .
nonetheless , as noted by @xcite , it is striking that b182124 and b1937 + 21 , which exhibit very luminous non - thermal ( magnetospheric ) x - ray emission , have the highest magnetic fields at the light cylinder , with @xmath326 g. it may be the case that these msps do , in fact , have heated polar caps but the resulting emission is several orders of magnitude fainter than the magnetospheric emission . alternatively , the energetics of the system may not favor substantial polar cap heating ( see , e.g. * ? ? ?
pulsars with @xmath327 g , on the other hand , appear to have x - ray emission that is typically either thermal or of indeterminate ( but probably thermal ) character . in the case of the dipole magnetic field at the surface @xmath328 g , the @xmath329 relation ( fig .
@xmath330 ) seems even less defined with a large scatter for both the complete msp sample and the 47 tuc msps .
the thermal @xmath300 for the 47 tuc msps is best fit by @xmath331 .
the apparent lack of correlation between @xmath4 and @xmath332 may either indicate that the small scale magnetic field structure near the surface of the polar caps deviates from that of a pure dipole ( see e.g. , * ? ? ? * ) or that both the thermal and non - thermal x - ray emission properties of msps are determined by the conditions much higher in the magnetosphere ( e.g. , at the light cylinder ) . versus magnetic field strength at the light cylinder ( @xmath333 ) for the msps in 47 tuc ( filled symbols ) and all other msps detected in x - rays ( open symbols ) .
( _ b _ ) @xmath334 versus surface magnetic field strength ( @xmath332 ) .
the dashed line corresponds to the best fit for the 47 tuc msps with thermal spectra .
the various symbols are defined as in figure 5.,scaledwidth=46.0% ] llccccc 1 & j0030 + 0451 & 4.86 & @xmath2761.0 & 0.3 & 3.6 & 1 , 2 + 2 & j0218 + 4232 & 2.32 & 7.78 & 2.7 & 4.2 & 3 , 4 + 3 & j04374715 & 5.76 & 1.86 & 0.139 & 12.9 ( 3.0 ) & 5 , 6 + 4 & j0751 + 1807 & 3.48 & 0.73 & 1.1 & 0.47 & 7 , 8 + 5 & j1012 + 5307 & 5.26 & 0.97 & @xmath2470.77 & 1.3 & 9 , 8 + 6 & j10240719 & 5.16 & @xmath2760.30 & @xmath2760.20 & 0.2 & 10 , 11 , 12 + 7 & j17441134 & 4.07 & 0.71 & 0.357 & 0.2 & 10 , 13 , 12 + 8 & b1937 + 21 & 1.56 & 10.6 & 3.6 & 3.3 & 11 , 14 + 9 & b1957 + 20 & 1.61 & 1.15 & 2.5 & 0.9 & 11 , 15 + 10 & j21243358 & 4.93 & 1.30 & 0.25 & 3.6 & 10 , 11 , 12 + 11 & b162026 ( m4 ) & 11.07 & @xmath150.67 & 1.73 & 0.03 & 16 , 17 , 18 + 12 & j17405340 ( ngc 6397 ) & 3.60 & 3.90 & 2.55 & 0.24 & 19 , 20 , 21 + 13 & b182124 ( m28 ) & 3.05 & 162 & 5.5 & 4.6 & 22 , 23 + 14 & j19116000c ( ngc 6752 ) & 5.28 & 0.22 & 4.1 & 0.01 & 24 + 15 & j21402310a ( m30 ) & 11.02 & @xmath276209 & 9.0 & @xmath2670.007 & 25 , 26 + in comparing theoretical predictions and observed relations it is necessary to examine and quantify all the effects that act to modify the observed parameters under consideration .
first , it is important to emphasize that the x - ray luminosities plotted in figures 5 and 6 represent observed , not true quantities .
this is especially critical for the non - thermal magnetospheric emission , which is subject to beaming and , hence , viewing angle effects .
therefore , the measured value of @xmath4 may greatly underestimate the true x - ray luminosity . as for the thermal emission , a major source of uncertainty
is introduced by the lack of knowledge of the masses and radii of the sample of msps .
these parameters affect @xmath4 through the gravitational redshift and bending of light . in particular , the true thermal luminosity , as measured at the ns surface , is larger by a factor of @xmath335 than the luminosity measured by a distant observer . in our analysis ,
we have corrected for this effect by assuming the canonical values of @xmath336 m@xmath95 and @xmath337 km . however , for a reasonable range of ns masses , 1.2 - 2.3 m@xmath95 , based on the results of @xcite and @xcite , we find that the true @xmath4 may differ from the value used here by up to a factor of @xmath151.8 .
in addition , for ns radii in the range @xmath338 km , the derived @xmath300 may deviate from the true value by @xmath26720% .
the parameters @xmath306 , @xmath332 , and @xmath333 also depend on the msp mass and radius , as @xmath339 and @xmath340 .
for the same range of ns masses ( @xmath341 m@xmath95 ) , the assumption of 1.4 m@xmath95 may result in a deviation of up to @xmath1560% for @xmath306 and @xmath1525% for @xmath342 and @xmath333 , from the true values .
this implies that a spread in masses within the sample of msps would tend to increase the scatter about the derived relations in both coordinates ( e.g. @xmath300 and @xmath306 ) and perhaps result in a skewed trend .
conversely , for a sample of msps that have very similar masses but are systematically more or less massive than the canonical 1.4 m@xmath95 , the result will be an offset in the derived intercept , with little effect on the slope of the log - log relation under consideration ( e.g. @xmath343 ) . as for @xmath344 , for the range @xmath338 km we find that @xmath306 , @xmath332 , and @xmath333 may differ by up to a factor of @xmath152.6 .
the effect of the ns radius on the scatter in the trends discussed above is uncertain due to lack of knowledge of the true ns equation of state ( eos ) .
for instance , in certain theoretical models of the ns eos a small change in @xmath345 does not result in an appreciable change in @xmath344 , while for others the opposite is the case ( see e.g. , * ? ? ? * for details ) .
for the 47 tuc msp , the parameters derived from @xmath308 are subject to an additional uncertainty arising from lack of knowledge of the magnitude of acceleration of the msps by the gravitational potential of the globular cluster .
it is important to note that the model - dependent values derived for @xmath309 are sensitive to the assumptions made regarding the gravitational potential and gas density profile of the cluster . for simplicity , in this paper
we have derived @xmath309 for each msp assuming a simple constant gas density profile .
however , it is likely that in reality the gas distribution of 47 tuc is more complex .
for instance , the nearly identical dispersion measures of the msps near the cluster center , namely 47 tuc f , l , o , and s @xcite , may be interpreted as being due to the presence of a cavity in the plasma around the cluster core carved out by the winds of these pulsars .
hence , obtaining accurate values of @xmath309 may require implementation of more elaborate models of the gas distribution , which will be explored in detail elsewhere .
another potential source of systematic uncertainty is the presence of an observational selection effect .
it is probable that so far pulsar observations have sampled only the most luminous and nearby msps .
the existence of x - ray faint msps , both in globular clusters and the field of the galaxy , would mean that the observed narrow range of @xmath300 for the msps with thermal spectra is simply an artifact of the limited sample of very low luminosity msps .
given the large uncertainties discussed above , any conclusions drawn for the apparent relations between @xmath4 and quantities derived from the pulsar spin parameters should be taken with much caution , as it is not possible , at present , to properly account for all effects that act to modify the observed trends .
the spectral fits described in 3 show that the majority of 47 tuc msps exhibit emission consistent with a thermal model , with no indication of any magnetospheric emission .
it is probable that the msps with seemingly purely thermal emission do , in fact , have a fainter magnetospheric x - ray component ( @xmath276@xmath346 ergs s@xmath3 ) .
however , it remains unclear what determines the relative strengths of these two x - ray production mechanisms .
although viewing angle dependences and beaming of the magnetospheric radiation can explain why thermal emission dominates in many msps , it is more likely that the relative strength of the two components is determined by the poorly understood details of the magnetospheric emission and polar cap heating mechanisms .
@xcite and @xcite have raised the possibility that the very narrow range of thermal luminosities and small emission areas imply the existence of a strong multipole magnetic field near the surface . however , the small range in @xmath334 _ is _ consistent with a simple dipole field if one considers general relativistic effects near the surface . in particular ,
gravitational bending of light causes @xmath34775% of the ns surface to be visible at any given time ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
as a consequence , the degree of modulation in the flux which arises due to rotation of the ns is greatly reduced .
this , in turn , diminishes the scatter in @xmath4 for the sample of msps that would otherwise result from different orientations of the magnetic and spin axes relative to the line of sight .
thus , although entirely plausible , the existence of a multipole magnetic field need not be invoked to account for the small range in @xmath4 .
we note that the inferred radii of the thermally emitting areas obtained in 3 are indeed somewhat smaller than the expected size of a magnetic polar cap for a simple dipole field , @xmath348 ( e.g. , @xmath349 km for a 3 ms pulsar ) .
this likely arises due to the fact that we have fitted most msp spectra with a single temperature model , where in reality a two - temperature model is more appropriate , as suggested by the results of 3 . in adittion ,
due to variation of the projected area of the polar caps , induced by the rotation of the star , fits to the time - integrated spectrum result in an underestimate of the effective emission area .
however , even with a two temperature spectrum and a correction factor for the effective area , in the case of a bb model the total emitting region is still substantially smaller than the expected polar cap area .
if the ns surface does , in fact , radiate like a blackbody , this discrepancy may be indicative of either non - uniform polar cap heating or deviations from a pure dipole field .
for instance , the polar cap model of harding & muslimov predicts that the heating rate by positrons should be highest near the rim of the polar cap ( cf .
fig . 7 of * ? ? ?
* ) , resulting in an annulus - shaped emitting region with an area significantly smaller than that of the entire polar cap .
@xcite have argued that a small emission area can arise if the msp is an aligned rotator , resulting in a so - called `` squeezed '' polar cap .
the outer gap model of @xcite also predicts that the heated area should be significantly smaller than the whole polar cap due to the presence of a small scale multipole field near the surface .
finally , this result may imply that a blackbody is simply not a valid approximation of the ns surface .
the discrepancy may , thus , be a consequence of the fact that simplified blackbody models result in higher derived temperatures and smaller emission areas due to the fact that the x - ray spectra of light - element atmospheres are harder than blackbody spectra @xcite .
the results of this study as well as recent _ chandra _ and _
xmm - newton _ observations of nearby field msps @xcite have revealed that there are no clear systematic differences between the x - ray properties of globular cluster and field msps , which could arise , for instance , due to different evolutionary paths .
for example , the nearby field msps such as j04374715 and j0030 + 0451 , are quite similar to the bulk of 47 tuc msps in terms of x - ray luminosity and spectral properties .
also , b182124 in the globular cluster m28 , is akin to b1937 + 21 , characterized by large x - ray and spindown luminosities ( @xmath350 ergs s@xmath3 , @xmath351 ergs s@xmath3 ) , hard non - thermal spectra , and very narrow radio and x - ray pulse profiles @xcite .
these similarities imply that if , in fact , multiple binary exchanges and accretion episodes occur for globular cluster msps , they may not significantly alter the elementary emission properties of the pulsar .
@xcite have shown that the binary msp 47 tuc w exhibits large - amplitude x - ray variability at the binary period .
this modulation can be interpreted as being due to geometric occultations of an x - ray emitting intrabinary shock by the @xmath150.15 m@xmath95 main - sequence companion . on the other hand , in the three other systems in which radio eclipses are observed , namely 47 tuc j , o , and r
, we detect no such variability .
the persistence of x - ray emission during the radio eclipses confirms that at radio frequencies j , o , and r are not eclipsed by their respective binary companions but rather by a tenuous ionized envelope formed due to mass loss by the companion driven by the pulsar wind , as in the case of psrs b1957 + 20 and j20510827 @xcite . the unique x - ray behavior of 47 tuc w is most certainly the result of the presence of a main - sequence secondary star instead of the typical very - low mass ( @xmath150.03 m@xmath95 ) degenerate companion found in `` black widow''-like systems such as j , o , and r. while the latter three systems are expected end products of low - mass x - ray binary evolution , 47 tuc w is likely the result of an exchange encounter in which the original very low - mass companion was ejected from the binary ( * ? ? ? * and references therein ) .
however , even if the current companion of msp w is roche - lobe filling , the existence of an intrabinary shock implies that another accretion phase can not commence as the energetic pulsar wind is constraining the outflow of mass from the companion ( see * ? ? ? * for details ) .
in 5 we have found that the @xmath352 relation for the thermally emitting msps has a slope of @xmath8 .
although such a trend is consistent with the prediction of the polar cap heating model of @xcite , @xmath9 , we can not exclude the possibility that the 47 tuc msps follow the linear @xmath353 relation . despite the large uncertainties , for the present sample of msps detected in x - rays we still observe a general trend of decrease of @xmath4 as a function of spin down luminosity and magnetic field strength .
if these parameters are , in turn , determined by the age of the msp , it is tempting to speculate that figures 5 and 6 represent evolutionary diagrams .
here we consider a possible evolutionary path of a msp described within the framework of the @xcite electron - positron ( @xmath354 ) pair production model .
for relatively young and energetic msps , such as psr b182124 ( @xmath355 yr ) , @xmath354 pairs are produced via curvature radiation ( cr ) .
these pairs , in turn , produce the observable x - ray/@xmath356-ray and radio non - thermal magnetospheric emission . although some polar cap heating by a backflow of energetic particles may occur in these objects , the non - thermal component ( @xmath15@xmath357 ergs s@xmath3 ) is several orders of magnitude brighter , making it difficult to observe any emission from the polar caps .
as the msp ages , it undergoes rotational spindown and perhaps magnetic field decay , which result in a decline in the cr pair production rate . at a certain stage ,
the magnetic field becomes too low to permit cr pair production so inverse compton scattering ( ics ) of soft x - rays by accelerated electrons becomes an important channel for @xmath354 pair production .
ics pair production results in substantial polar cap heating by a backflow of positrons from the magnetosphere .
since the polar cap heating rate has a weaker dependence on the spindown luminosity ( @xmath358 ) than magnetospheric emission ( @xmath359 ) , the thermal emission will eventually become dominant over the magnetospheric emission , as is observed in the majority of 47 tuc msps and many field msps .
at much later times , the msp is expected to cross the ics pair production death line ( see fig . 1 in * ? ? ? * ) and will no longer be observable as a pulsar , both in the radio and x - ray . in the end , as the polar caps cool , the msp will fade away from view at x - ray energies as well .
we have presented a spectral analysis of the x - ray emission for the complete sample of known msps in 47 tuc .
the x - ray spectra of the majority of msps are well described by a thermal model , a bb or a ns hydrogen atmosphere , with @xmath360 k , @xmath361 km , and @xmath2 ergs s@xmath3 ( 0.38 kev ) .
the small observed scatter in the thermal @xmath4 is consistent with a simple dipole magnetic field configuration .
a two temperature thermal model , reminiscent of that observed from some nearby field msps , is also consistent with the emission from the bulk of 47 tuc msps and , in some cases , results in an improved fit .
such a spectrum could arise due to non - uniform polar cap heating or lateral diffusion of heat accross the ns surface away from the regions directly heated by particles from the magnetosphere .
the msps j , o , and w show a prominent non - thermal component in their spectra .
at least part of the non - thermal x - rays are very likely due to interaction of the relativistic particle wind of the pulsar with matter from the companion .
all other 47 tuc msps exhibit no evidence of a significant non - thermal x - ray component .
our re - examination of the @xmath325 relation has revealed that regardless of the emission mechanism responsible for production of the observed x - rays , the conversion efficiency of @xmath306 to @xmath4 appears to be , in all cases , @xmath15@xmath362 .
all of the 47 tuc msps , with the exception of 47 tuc w , display no large amplitude variability in their x - ray emission on timescales of days , weeks , and years .
the msp w exhibits dramatic variations at the binary period , which can be attributed to the presence of an intrabinary shock that is eclipsed by the secondary star .
the lack of long term variations in the flux of most msps could potentially serve as a criterion for classification of unidentified soft x - ray sources @xcite .
given that most of the @xmath15@xmath363 msps presently known have spin parameters similar to those of the 47 tuc msps , we expect many of them to have predominantly thermal x - ray spectra .
moreover , for these msps , gravitational bending of light near the ns surface ensures that at least one of the antipodal heated polar caps is always observable , while for a wide range of angles between the magnetic and spin axes and the line of sight both polar caps are visible simultaneously for a large portion of the msp spin period ( see e.g. * ? ? ? * ) .
this effect invariably results in a 100% duty cycle , even if the magnetospheric emission beams never sweep across the earth or if the radio emission is very faint or heavily absorbed / scattered .
therefore , many msps with heated polar caps may , in principle , be detectable at x - ray energies even if they can not be observed as radio pulsars .
we would like to thank peter edmonds for insightful discussions , don lloyd for use of his ns hydrogen atmosphere model , and the referee natalie webb for many helpful comments .
this work was supported by nasa _
grants go2 - 3059a and g02 - 3059b .
the research presented here has made use of the nasa astrophysics data system .
richer , h. b .. fahlman , g. g .. ibata , r. a. , pryor , c. , bell , r. a. , bolte , m. , bond , h. e. , harris , w. e. , hesser , j. e. , holland , s. , ivanans , n. , mandushev , g. , stetson , p. b. , & wood , m. a. 1997 , , 484 , 741 | we present spectral and long - timescale variability analyses of _ chandra x - ray observatory _ acis - s observations of the 19 millisecond pulsars ( msps ) with precisely known positions in the globular cluster 47 tucanae .
the x - ray emission of the majority of these msps is well described by a thermal ( blackbody or neutron star hydrogen atmosphere ) spectrum with a temperature @xmath0 k , emission radius @xmath1 km , and luminosity @xmath2 ergs s@xmath3 . for several msps
, there is indication that a second thermal component is required , similar to what is seen in some nearby field msps .
the observed radiation most likely originates from the heated magnetic polar caps of the msps .
the small apparent scatter in @xmath4 is consistent with thermal emission from the polar caps of a global dipole field although the small emission areas may imply either a more complex small - scale magnetic field configuration near the neutron star surface or non - uniform polar cap heating .
the radio - eclipsing binary msps 47 tuc j , o , and w show a significant non - thermal ( power - law ) component , with spectral photon index @xmath5 , which most likely originates in an intrabinary shock formed due to interaction between the relativistic pulsar wind and matter from the stellar companion .
we re - examine the x - ray spindown luminosity relation ( @xmath6 ) and find that for the msps with thermal spectra @xmath7 , where @xmath8 . due to the large uncertainties in both parameters
the result is consistent with both the linear @xmath6 relation and the flatter @xmath9 predicted by polar cap heating models . in terms of x - ray properties
, we find no clear systematic differences between msps in globular clusters and in the field of the galaxy .
we discuss the implications of these results of the present understanding of the x - ray emission properties of msps . |
in the early 1960 s , i.m .
lifshitz published his famous analysis of the low - energy tails of the density @xmath12 of states of an electron in a medium with randomly scattered immobile impurities @xcite .
he demonstrated that , in @xmath2-dimensions , the spectrum @xmath12 of the random schrdinger operator decays as @xmath13 this exponential decay is in striking contrast to the polynomial decay associated with a periodic schrdinger operator , and is caused by the presence of arbitrarily large , albeit rare , regions without impurities .
a decade later balagurov and vaks @xcite and donsker and varadhan @xcite published their celebrated work on trapping kinetics in a medium with randomly placed immobile traps ( called @xmath4 ) .
they found that the long - time decay of the survival probability @xmath14 of a particle ( called @xmath1 ) diffusing in such a @xmath2 dimensional medium follows the stretched - exponential law @xmath15 where @xmath3 is the mean density of traps , @xmath16 is the particle diffusion coefficient , and @xmath17 is the @xmath2-dependent constant @xcite @xmath18 here @xmath19 is the first zero of the bessel function @xmath20 and @xmath21 $ ] denotes the volume of a @xmath2-dimensional sphere of unit radius .
the superscript @xmath4 on the survival probability emphasizes that the traps @xmath4 are immobile .
balagurov and vaks obtained the decay form in eq .
( [ tr ] ) exactly for @xmath22 . while they furthermore deduced this behavior for general @xmath2 by noticing the close mathematical connection between the trapping problem and the lifshitz problem , donsker and varadhan
were the first to obtain the rigorous exact solution of this essentially many - body problem in @xmath2 dimensions .
they determined not only the time dependence but also the decay coefficient @xmath17 of eq .
( [ ll ] ) . note that this law stands in stark contrast to the purely exponential decay predicted by standard chemical kinetics for the reaction @xmath23 of mobile particles and traps , and even to the predictions of the smoluchowski approach based on a reaction - diffusion equation , @xmath24 , & d=1\\ \exp [ - 4 \pi \rho d t/\ln(d t / a^2 ) ] , & d=2 \\ \exp ( - 4 \pi \rho d t ) , & d=3 \end{cases}\ ] ] where @xmath25 is the sum of the diffusion coefficients of the particle and the traps @xcite . in mathematical analogy with the source of the lifshitz tail ,
the decay law ( [ tr ] ) arises from the presence of arbitrarily large regions without traps in which the particle can diffuse for a relatively long time before being trapped .
although the survival probability of a particle in any one such region is purely exponential , the ensemble average over a random distribution of such trap - free voids produces the anomalously slow decay . the decay law in eq .
( [ tr ] ) has been also generalized to trapping reactions on fractals and in inhomogeneous structures @xcite , to @xmath1 particles performing subdiffusive motion @xcite or attached to the extremities of polymer chains @xcite , and also to agglomerated random distributions of traps distributed on immobile polymer chains in solution @xcite or in clusters @xcite .
the history of this problem and many other results have been summarized in several reviews ( see , e.g. , @xcite ) .
the survival probability eq .
( [ tr ] ) [ as well as the lifshitz tail result of eq .
( [ dens ] ) ] is valid only when the traps ( or impurities ) are strictly immobile .
indeed , if one allows them to diffuse , no matter how small the diffusion coefficient , the particle survival probability is described by a faster decay law . as proved by bramson and lebowitz @xcite
( see also @xcite ) , when both species diffuse , the survival probability of the @xmath1 particle at long times is given by @xmath26 the time dependences are the same as in the smoluchowski problem , but the constants @xmath27 and @xmath28 are in general different , may depend on the diffusion coefficients of the particle and of the traps ( but see below ) , and were not determined in this original work . when traps diffuse , the many - body trapping effects captured by eq .
are thus no longer applicable and the @xmath1 particle survival probability decays with time according to the faster time dependencies in eq .
( [ bramsleb ] ) .
the underlying fluctuation mechanism governing the trapping dynamics has changed when the traps are allowed to move . while the time dependence in the decay laws ( [ bramsleb ] ) is consistent with smoluchowski - like results , which in fact represent a two - body approximation , the decay amplitudes @xmath29 are not simply functions of the sum of the diffusion coefficients .
in particular , the exact form ( including the coefficients ) of the leading large-@xmath30 behavior for @xmath22 and @xmath31 has only recently been found by bray and blythe @xcite and , surprisingly , @xmath32 and @xmath33 depend only on the diffusion coefficient @xmath34 of the traps and are independent of @xmath16 , the diffusion coefficient of the @xmath1 particle !
this implies that the survival probability of the diffusing @xmath1 particle in this scenario is asymptotically identical to that of an @xmath1 particle that remains still , that is , to @xmath35 .
this remarkable result has subsequently been extended to systems in which both the particles and the traps move subdiffusively @xcite .
in particular , it is again found that the survival probability of a subdiffusive @xmath1 particle in a sea of subdiffusive traps is identical to that of and @xmath1 particle that remains still ( as in the bramson - lebowitz scenario ) , @xmath36 . however , it is still not clear what happens when traps move subdiffusively and the particle is diffusive . in one dimension
it was shown in @xcite that the asymptotic survival probability of the diffusing particle is the same when surrounded by subdiffusive traps characterized by a mean square displacement that grows as @xmath5 with @xmath37 as it is for a stationary particle , _ provided the subdiffusive traps move sufficiently rapildy _
( @xmath38 ) , again akin to the bramson - lebowitz scenario .
however , few results seem to be currently available for @xmath39 and/or if the subdiffusive traps are too slow .
" an interesting question is then the following : will the trapping tails associated with immobile traps withstand sufficiently slow sub - diffusion , thus leading to a survival probability such as that of the donsker - varadhan result , or will they switch to appropriately rescaled bramson - lebowitz forms ?
this is the main question that we pose in this paper . in this pursuit
, we follow the general idea @xcite of constructing lower and upper bounds on the survival probability of a diffusive particle in a sea of subdiffusive traps whose motion is described by a continuous time random walk that starts at time @xmath40 @xcite .
if these bounds coincide asymptotically , then we can extract an exact asymptotic result for the survival probability of the particle .
we write the asymptotic survival probability in the form @xmath41 where the constants @xmath42 and @xmath43 depend on @xmath11 and on dimension , and explore whether a convergence of upper and lower bounds provides the exponents @xmath43 and perhaps even the exponential prefactor @xmath42 .
we broadly anticipate our results by noting that if the subdiffusive particles are sufficiently slow , specifically if @xmath7 , these bounds lead to a donsker - varadhan behavior , and the lifshitz or trapping tails thus remain unchanged . if @xmath8 and @xmath44 , then the bounds lead to a bramson - lebowitz behavior that is , a survival probability behavior associated with a stationary particle ( note that , albeit for a different system , this result is implicit in @xcite ; that work addresses diffusive particles in a fractal medium , while here we are considering subdiffusive motion in euclidean geometries ) .
however , if @xmath8 and @xmath45 we are not able to establish a unique asymptotic behavior .
the detailed results are exhibited later in the paper . in sec .
[ model ] we formulate the model of moving particle and traps . in sec .
[ lowerbound ] we calculate a lower bound on the survival probability of the particle , and in sec . [ upperbound ] we obtain two upper bounds .
the consequences of these bounds on the survival probability are collected in sec .
[ collection ] .
a brief recapitulation of the results is given in the concluding section .
in this section we formulate the general model that allows us to highlight the approximations made to obtain upper and lower bounds on the survival probability .
consider a @xmath2-dimensional system of volume @xmath46 containing a single _ diffusive _ @xmath1 particle of radius @xmath47 and @xmath48 randomly moving point traps @xmath4 ( a finite trap radius would add nothing interesting to the problem for our purposes ) .
the initial position of @xmath1 is the origin , and one realization of its trajectory is denoted by @xmath49 .
the starting points @xmath50 of the traps are randomly ( poisson ) distributed throughout the volume , and a trajectory of the @xmath51th trap is denoted by @xmath52 .
next we define @xmath53 the indicator function @xmath54 $ ] of the event that the @xmath1 particle has not met any of the @xmath4s up to time @xmath30 for a given realization of their trajectories can then be written as @xmath55=\prod_{k=1}^k \psi[{\bm a}_t,{\bm b}_t^{(k ) } ] = \prod_{k=1}^k \exp\left[- \int^t_0 v({\bm a}_{\tau } - { \bm b}_{\tau}^{(k ) } - { \bm b}_0^{(k ) } ) d\tau\right].\ ] ] consequently the @xmath1 particle survival probability is @xmath56 \right\}\right)\right\},\ ] ] where the symbol @xmath57 denotes an average over all @xmath1 particle trajectories @xmath49 that start at the origin .
the symbol @xmath58 denotes an average over all @xmath4 trajectories @xmath59 whose starting point is @xmath60 .
we have labeled each of the latter trajectories by a trap label @xmath51 , but since the traps move independently we can omit the label .
furthermore , we go to the thermodynamic limit @xmath61 while keeping the ratio @xmath62 fixed .
this leads to the @xmath1 survival probability @xmath63\right\}}\big\},\ ] ] where @xmath64 $ ] is the functional of the trajectories @xmath65 and @xmath66 @xmath67 = \int_{r^d }
d{\bm b}_{0 } \left(1 - e^{- \int^t_0 v({\bm a}_{\tau } - { \bm b}_{\tau } - { \bm b}_0 ) d\tau}\right).\ ] ] the exact problem has thus been reduced to an effective two - body problem involving a single @xmath1 particle and a single @xmath4 trap .
nevertheless , it unfortunately does not seem possible to evaluate the survival probability exactly from this expression , the main mathematical difficulty being that the average over all possible trajectories @xmath66 has to be performed for each fixed @xmath65 . only after this average has been performed can one then go on to carry out the further average over the @xmath1 particle trajectories .
this appears to be a non - tractable mathematical problem , and recourse has to be made to controllable approximations .
we do it here by constructing lower and upper bounds on @xmath0 and identifying conditions and regimes where these converge asymptotically .
a lower bound was originally devised in @xcite for diffusive particles and traps and was extended in @xcite to the case of ( sub)diffusive particles and traps .
we adapt this method to the current situation in which the traps @xmath4 perform continuous - time random walks , so that we may restate this bound in the language introduced above . following bray and blythe ,
we introduce a notional spherical volume @xmath68 of radius @xmath69 centered at the origin , and pick only those realizations of initial trap distributions for which this volume is completely devoid of traps . the probability that the region will initially be empty of traps is @xmath70 we furthermore introduce the probability @xmath71 that the @xmath1 particle does not leave the notional volume @xmath68 during time @xmath30 @xcite , @xmath72 thirdly
, we introduce the probability that no @xmath4 particle enters the notional volume up to time @xmath30 , that is , the probability that an immobile @xmath2-dimensional target of radius @xmath69 survives up to time @xmath30 in the presence of a concentration @xmath3 of traps all of which perform subdiffusive motion @xcite : @xmath73 here @xmath74 is the gamma function , and @xmath34 is the anomalous diffusion coefficient of the traps , that is , the coefficient in the mean square displacement relation @xmath75 .
since the functional in eq.([l ] ) is positive definite , the two latter constraints on the trajectories of @xmath1 and the @xmath4s naturally lead to a lower bound on the survival probability @xmath0 .
furthermore , for these constrained trajectories the functional in eq.([w ] ) is strictly equal to zero , and hence eq .
( [ l ] ) restricted in this way is equal to unity . as a consequence , the probabilities associated with the random processes @xmath65 and @xmath66 subject to these constraints can simply be factored , immediately leading to the lower bound on @xmath0 @xmath76 finally , we note that this lower bound @xmath77 is in fact a family of lower bounds dependent on the radius @xmath69 of the notional volume separating particle and traps . this radius can be chosen to give the best lower bound , that is , the maximal lower bound , which we simply denote as @xmath78 .
the optimal radii are shown explicitly in appendix [ appa ] , and are shown to depend on dimensionality and on @xmath11 .
the associated optimal lower bounds for the survival probability of @xmath1 , which also depend on dimensionality and on @xmath11 , then immediately follow .
we thus have @xmath79 where , for @xmath80 , @xmath81 for @xmath31 we have @xmath82 for @xmath10 the results are @xmath83 there is thus a change in behavior of the lower bound in all dimensions when @xmath11 crosses the value @xmath84 .
an upper bound on the survival probability @xmath0 of the diffusive particle was recently derived on the basis of the so - called pascal principle , " which states that the @xmath1 particle survives longer if it remains still than if it moves .
the problem with a static @xmath1 particle and moving traps is the so - called target problem , " and consequently we label this upper bound on the survival probability as @xmath85 .
the pascal principle was conjectured in @xcite and was proved in @xcite for @xmath80 for conventional diffusive motion . in
@xcite the statement was proved for a rather general class of random walks on @xmath2-dimensional lattices .
a similar statement was introduced more than a decade earlier in @xcite in the context of excitation energy migration .
this upper bound is given by eq .
( [ ctrw - target ] ) obtained in @xcite if we set @xmath86 , the radius of the @xmath1 particle . for visual ease
we explicitly rewrite eq .
( [ ctrw - target ] ) with this replacement , @xmath87 to clearly tie together the various notations introduced so far , we note that @xmath88 and @xmath89 however , this upper bound when associated with the lower bound is not always sufficiently tight to provide the desired information about the asymptotic survival probability of the diffusive particle .
we thus introduce and alternative new upper bound , which in some cases is lower than the above . in the next section
we then explicitly pick the bounds to be used and exhibit the information that can be obtained from them .
this new upper bound is based on what we will call the anti - pascal principle . "
we will show that the worst possible strategy for traps in their search for a target is to remain immobile .
random motion , even uncorrelated with the motion of the target , enhances the probability to encounter the target . in other words , the diffusing particle @xmath1 survives longer if the traps remain still .
the problem with a moving @xmath1 particle in a sea of static traps is the so - called trapping problem , " and so we label this upper bound on the survival probability as @xmath90 . again , to clearly tie together various notations we note that @xmath91 , and the upper bound just introduced then says that @xmath92}\big\}.\ ] ] to prove eq .
we make use of jensen s inequality for convex functions , which for our model can be stated as @xmath93\big\ } } \leq e^{(b)}\big\ { e^{- \rho w[{\bm a}_{t } - { \bm b}_{t } ] } \big\}\ ] ] ( note that this inequality is generally used to derive a _
lower _ bound . indeed , when applied to the @xmath94 average in eq .
( [ 3 ] ) it yields a lower bound which is exactly the smoluchowski - type result @xcite . ) consequently , we have the following upper bound : @xmath95 } \right\ } \right\ } \equiv e^{(c)}\left\{e^{- \rho w[{\bm c}_t ] } \right\},\ ] ] where @xmath96 denotes an average with respect to the trajectories @xmath97 of a fictitious " particle @xmath98 of radius @xmath47 which starts its motion at @xmath99 . note that @xmath100 $ ] has a clear geometric interpretaion - it defines the volume swept by the fictitious particle @xmath98 within the time interval @xmath101 and is thus an analog of the so - called wiener sausage for conventional diffusive motion .
clearly , @xmath100 $ ] is a _
non decreasing _ function of time @xmath30 or , more precisely , of the number @xmath102 of jumps made by the fictitious particle within the time interval @xmath101 .
this number has two contributions , @xmath103 , where @xmath104 is the number of jumps made by particle @xmath1 and @xmath105 the number of jumps made by a trap .
setting @xmath106 , i.e. , supposing that the trap is immobile , clearly diminishes the total number of jumps and hence diminishes @xmath100 $ ] . in this way we tighten
the bound in eq.([4 ] ) to arrive at the desired inequality ( [ imm ] ) we have thus established the second upper bound ( [ imm ] ) for the survival probability of @xmath1 . the maximal information about the survival probability is thus obtained from the lower bounds presented in eqs .
( [ lbtight])-([optlower3d ] ) and the smaller of the upper bounds ( [ uppertarget ] ) with ( [ upper1 ] ) and ( [ imm ] ) with ( [ tr ] ) .
as expressed in eq .
( [ assume ] ) , or bounds on this exponent , as a function of the trap subdiffusion exponent @xmath11 .
first panel : @xmath22 ; second panel : @xmath31 .
third panel : @xmath107 .
the solid lines are upper bounds ; specifically , the thick solid lines are pascal upper bounds ( immobile particle ) while the thin solid lines are anti - pascal upper bounds ( immobile traps ) .
the solid lines with superimposed circles are lower bounds .
thus the first and second panels show that for @xmath108 the pascal ( donsker - varadhan ) and lower bound lines coincide for @xmath7 .
the behavior changes when @xmath11 crosses the value @xmath84 , leading to a coincidence of the anti - pascal ( bramson - lebowitz ) and lower bound lines . for @xmath109 the asymptotic exponent
is thus obtained for all values of @xmath11 .
the third panel for @xmath107 shows convergence of the upper ( pascal ) and lower bounds for @xmath110 .
however , when @xmath111 the situation is uncertain . the exponent @xmath43 now lies in the triangular region delimited by the lower bound and the pascal ( @xmath112 ) or anti - pascal ( @xmath113 ) upper bound . ] in this section we collect our detailed results , exhibited by displaying values for the constants @xmath43 and @xmath42 in eq .
( [ assume ] ) when the upper and lower bounds converge asymptotically , and providing bounds for the exponent @xmath43 when they do not .
the results for the exponent @xmath43 as a function of the subdiffusive trap exponent @xmath11 are shown for integer dimensions @xmath22 , @xmath114 , and @xmath115 in fig . [ fig1 ] .
even before displaying the explicit results we note the following : * when @xmath116 the asymptotic survival probability of the diffusive @xmath1 particle in the sea of subdiffusive traps is the same as it is for _ immobile traps_. the question asked in the introduction , whether trapping tails associated with immobile traps withstand sufficiently slow sub - diffusion , is thus answered in the affirmative , with a precise dimension - dependent characterization of what is sufficiently slow . "
this behavior , obtained by the asymptotic convergence of the bounds @xmath78 and @xmath117 , is shown for the exponent @xmath43 by the thick solid lines with superimposed circles in fig .
[ fig1 ] . in this regime
the survival probability is thus of the donsker - varadhan form .
* when @xmath11 crosses the value @xmath84 there is a kind of dynamical phase transition .
for @xmath108 the survival probability of the diffusive particle in the sea of now more rapidly moving subdiffusive traps decays as it would if the _ particle remains immobile_. this is indicated by the thin solid lines with superimposed circles in the first two panels of fig .
[ fig1 ] , and results from the asymptotic convergence of the bounds @xmath78 and @xmath85 . in this regime
the survival probability is thus of the bramson - lebowitz form generalized to subdiffusive traps . * for @xmath10 and @xmath118
the situation is left somewhat uncertain : we are only able to bound the decay exponent but not determine it uniquely , because neither upper bound converges asymptotically to the lower bound .
all we can say is that the survival probability decay exponent @xmath43 lies in the triangular region bounded by the thin solid line with superimposed circles ( associated with the lower bound ) , the thick solid line ( associated with @xmath90 ) , and the thin solid line ( associated with @xmath119 ) , as indicated in the figure caption . .collected results for the asymptotic survival probability exponent @xmath43 and prefactor @xmath42 . [ cols="^,^,^,^,^ " , ] [ table ] finally , we collect the explicit results described above in table [ table ] .
the reported results are for the exponents @xmath43 and the prefactors @xmath42 in eq .
( [ assume ] ) .
again , we note that the results for @xmath43 are sketched explicitly for @xmath120 in fig .
in conclusion , we have rigorously determined the survival probability of a particle diffusing in a @xmath2-dimensional medium in the presence of a concentration of traps performing subdiffusive random motion .
we have arrived at our results by calculating a lower bound and two alternative upper bounds for the survival probability .
one of the upper bounds , based on the pascal principle , " is obtained by assuming the particle to remain immobile .
the other , based on the anti - pascal principle , " is found by assuming that the traps remain immobile .
we can then choose the tighter ( lower ) upper bound .
results for the asymptotic survival probability of the particle can thus be extracted if the lower bound and one of the upper bounds converge asymptotically . following this procedure ,
we have shown that when the dynamical exponent @xmath11 characterizing the growth of the second moment of the displacement of the traps is less than @xmath84 , that is , if the traps move sufficiently slowly , the decay of the survival probability of the diffusing particle in any dimension is given exactly by the donsker - varadhan result obtained for immobile traps . when the traps move more rapidly than this , i.e. , when @xmath8 , then in dimensions @xmath9 the survival probability of the particle is identical to that of a stationary @xmath1 particle in a sea of mobile traps . for higher dimensions , @xmath10 , our results do not uniquely determine the survival probability of the particle in this @xmath11 regime , but they do provide tighter bounds than previously known .
we close by noting the well - known difficulties that may stand in the way of reaching and verifying some asymptotic results via numerical simulations ( see @xcite and references therein ) .
such simulations continue to pose an interesting challenge .
the research of s.b.y . has been supported by the ministerio de educacin y ciencia ( spain ) through grant no .
fis2007 - 60977 ( partially financed by feder funds ) .
k.l . gratefully acknowledges the support of the national science foundation under grant no .
the research of o.b . and g.o .
is partially supported by agence nationale de la recherche ( anr ) under grant `` dyoptri - dynamique et optimisation des processus de transport intermittents '' .
in this appendix we explicitly show the optimal radius @xmath69 that leads to the maximal lower bound on the survival probability of the diffusive particle in the presence of a sea of subdiffusive point traps . combining the expressions in eqs .
( [ void ] ) , ( [ mea ] ) and ( [ ctrw - target ] ) and differentiating their product with respect to @xmath69 , we find that the optimal @xmath69 depends on dimensionality and on @xmath11 . for @xmath80
we find @xmath121 for @xmath31 we obtain @xmath122 for @xmath10 we find @xmath123 i. m. lifshitz , sov .
jetp * 17 * , 1159 ( 1963 ) ; sov .
* 7 * , 549 ( 1965 ) .
i. m. lifshitz , s. a. gredescul , and l. a. pastur , _ an introduction to the theory of disordered systems _ ( wiley , new york , 1988 ) .
balagurov and v. g. vaks , sov .
jetp * 38 * , 968 ( 1974 ) .
m. d. donsker and s. r. s. varadhan , commun .
pure appl . math .
* 28 * , 525 ( 1975 ) ; * 32 * , 721 ( 1979 ) .
a. m. berezhkovskii , yu . a. makhnovskii , and r. a. suris , chem
. phys . * 137 * , 41 ( 1989 ) .
a. blumen , j. klafter , and g. zumofen , phys .
b * 28 * , 6112 ( 1983 ) ; j. klafter , a. blumen , and g. zumofen , j. stat . phys . * 36 * , 561 ( 1984 ) s. b. yuste and k. lindenberg , phys
e * 72 * , 061103 ( 2005 ) ; s. b. yuste , j. j. ruiz - lorenzo , and k. lindenberg , phys .
e * 74 * , 046119 ( 2006 ) .
j. j. ruiz - lorenzo , s. b. yuste , and k. lindenberg , j. phys . : condens .
matter * 19 * , 065120 ( 2007 ) .
s. b. yuste and k. lindenberg , phys .
e * 76 * , 051114 ( 2007 ) .
g. oshanin , o. bnichou , m. coppey , and m. moreau , phys .
e * 66 * , 060101 ( 2002 ) .
g. oshanin , s. nechaev , a. m. cazabat , and m. moreau , phys .
e * 58 * , 6134 ( 1998 ) ; s. nechaev , g. oshanin and a. blumen , j. stat . phys . * 98 * , 281 ( 2000 ) . s. f. burlatsky and g. s. oshanin , phys .
lett . a * 145 * , 61 ( 1990 ) ; g. s. oshanin , a. v. mogutov and s. f. burlatsky , phys .
a * 149 * , 55 ( 1990 ) ; g. oshanin , m. moreau , and s. f. burlatsky , adv .
colloid interface sci . * 49 * , 1 ( 1994 ) .
a. m. berezhkovskii , yu . a. makhnovskii , and r. a. suris , j. stat
. phys . * 57 * , 333 ( 1989 ) ; * 65 * , 1025 ( 1991 ) .
j. w. haus and k. w. kehr , phys .
* 150 * , 263 ( 1987 ) .
f. den hollander and g. h. weiss , in _ contemporary problems in statistical physics _
, edited by g. h. weiss ( siam , philadelphia , 1994 ) .
m. bramson and j. lebowitz , phys .
lett . * 61 * , 2397 ( 1988 ) ; j. stat .
phys . * 62 * , 297 ( 1991 ) . s. f. burlatsky and a. a. ovchinnikov , sov .
jetp * 65 * , 908 ( 1987 ) .
a. j. bray and r. a. blythe , phys .
lett . * 89 * , 150601 ( 2002 ) ; r. a. blythe and a. j. bray , phys .
e * 67 * , 041101 ( 2003 ) .
a. blumen , j. klafter and g. zumofen , in _ optical spectroscopy of glasses _ , edited by i. zschokke ( reidel , dordrecht , 1986 ) .
a. j. bray , s. n. majumdar , and r. a. blythe , phys .
e * 67 * , 060102(r ) ( 2003 ) .
m. moreau , g. oshanin , o. benichou , and m. coppey , phys .
e * 67 * , 045104(r ) ( 2003 ) ; phys .
e * 69 * , 046101 ( 2004 ) .
s. f. burlatsky , g. s. oshanin , and a. a. ovchinnikov , phys .
a * 139 * , 241 ( 1989 ) . | we study the long - time tails of the survival probability @xmath0 of an @xmath1 particle diffusing in @xmath2-dimensional media in the presence of a concentration @xmath3 of traps @xmath4 that move sub - diffusively , such that the mean square displacement of each trap grows as @xmath5 with @xmath6 . starting from a continuous time random
walk ( ctrw ) description of the motion of the particle and of the traps , we derive lower and upper bounds for @xmath0 and show that for @xmath7 these bounds coincide asymptotically , thus determining asymptotically exact results .
the asymptotic decay law in this regime is exactly that obtained for immobile traps .
this means that for sufficiently subdiffusive traps , the moving @xmath1 particle sees the traps as essentially immobile , and lifshitz or trapping tails remain unchanged . for @xmath8 and @xmath9
the upper and lower bounds again coincide , leading to a decay law equal to that of a stationary particle .
thus , in this regime the moving traps see the particle as essentially immobile . for @xmath10 , however , the upper and lower bounds in this @xmath11 regime no longer coincide and the decay law for the survival probability of the @xmath1 particle remains ambiguous . |
in 1881 , newcomb @xcite noticed that the first digit distribution of numerical data is not uniform but rather logarithmic .
he had observed that the pages of logarithm tables were more worn out for smaller digits such as @xmath1 and @xmath2 than for larger ones , and concluded that `` [ the ] law of probability of the occurrence of the numbers is such that all mantissae of their logarithms are equally probable . ''
he explicitly tabulated the probability of occurrence of the first and second digits .
apparently unaware of newcomb s results , in 1938 , benford @xcite found the same first digit phenomenon , and explicitly gave the formula for the probability of a number having the first digit @xmath3 , @xmath4 where @xmath5 is used to represent the base @xmath6 logarithm .
he gathered empirical evidence for formula ( [ eq : ben ] ) by collecting thousands of numbers from diverse datasets , such as the area of the riverbeds , atomic weights of elements , etc .
it has been shown ( see for example @xcite ) that the only non - trivial digit distribution that is left invariant under scale change of the underlying distribution is the benford distribution .
scale - invariance means that a collection of data has the same digit distribution when multiplied by a constant . in his seminal paper @xcite
, hill showed that the appropriate domain for the significant digit probability is the smallest collection of positive real subsets that contains all the infinite sets of the form @xmath7 .
this set denoted by @xmath8 , is the smallest sigma algebra generated by @xmath9 , @xmath10 , , where @xmath11 is the @xmath12-th significant - digit function .
@xmath13 and @xmath14 for @xmath15 .
for example , @xmath16 and @xmath17 .
observe that @xmath18 for all @xmath12 and @xmath3 . within this framework , for a random variable @xmath19 ,
the benford s first digit distribution law can be stated as @xmath20 in general @xcite , a random variable is benford if for all @xmath21 , all @xmath22 and all @xmath23 for @xmath24 , @xmath25 for example , the probability of having digits @xmath26 , @xmath27 and @xmath28 as the first , second and third significant digits , respectively , is @xmath29 hill defined base - invariance , and proved that base - invariance as well as scale - invariance imply benford distribution of digits .
leemis et .
al @xcite investigated several examples of symmetric and non - symmetric distributions that lead to benford distributed random variables .
this paper is organized as follows . in section [ mod1 ] ,
we use the mod 1 map to show that the well - known base- and scale - invariance of benford distributed random variables are consequences of @xmath30 , where @xmath31 is the associated density function . in section [ nben ]
, we introduce the concept of @xmath0-digit benford distributed random variables which are guaranteed to obey the log distribution in their first @xmath0 digits , and give a general construction method for such variables .
unless otherwise specified , throughout this paper , we assume that the base is @xmath6 .
any positive real number @xmath32 can be written as @xmath33 for some @xmath34 , where @xmath35
. then the first digits of @xmath32 and @xmath36 are the same : @xmath37 .
let us assume that @xmath32 is given by a random variable @xmath19 with the density function @xmath38 .
let @xmath39 be the random variable with the density function @xmath40 . for any real function @xmath41
, we define @xmath42 as @xmath43 [ l : one ] [ dagger ] the probability of @xmath44 having its first digit @xmath3 is @xmath45 let us consider the real numbers starting with the digit @xmath3 , i.e. , @xmath46 .
these numbers belong to the set @xmath47 now consider the random variable @xmath39 with the density function @xmath40 .
the logarithmic function maps @xmath48 into @xmath49 .
this set modulo @xmath1 is just @xmath50 as illustrated in figure ( [ f : mod1 ] ) for the case @xmath51 .
.,scaledwidth=90.0% ] consequently @xmath52 clearly , if @xmath30 , then one obtains benford s law @xmath53 .
the uniformity of @xmath31 resonates with newcomb s pioneering observation in 1881 @xcite that the `` probability of the occurrence of the numbers is such that all mantissae of their logarithms are equally probable . ''
lemma [ l : one ] can be readily generalized for the probability of a sequence of prescribed digits @xmath54 , where @xmath55 and @xmath56 for @xmath57 : @xmath58 scale - invariance means that a collection of data does not change its digit distribution when multiplied by a constant .
for example , suppose that the prices of goods are benford distributed , then scale - invariance implies that these prices remain benford distributed regardless of the currency in which they are converted .
[ l : scale - tr ] scaling of a random variable @xmath19 is equivalent to a translation of @xmath59 .
let @xmath60 be a random variable with the density function @xmath40 and let @xmath61 be the random variable generated by the translation of @xmath60 by @xmath62 units , i.e. , @xmath63 and @xmath64 , where @xmath65 is the density function of @xmath61 .
define @xmath66 and @xmath67 and let @xmath38 and @xmath68 be their corresponding density functions . in terms of cumulative distribution functions , we have @xmath69 the converse is immediate .
due to modular arithmetics , a translation of @xmath40 will result in a wrap - around effect in @xmath31 .
hence , scaling of @xmath19 induces a wrap - around of @xmath31 .
for example , let @xmath70 be the random variable with the density function @xmath71 , and let @xmath61 be its translation by @xmath62 .
the effect of scaling of @xmath19 on @xmath31 is shown in figure ( [ f : wrap ] ) .
of a translation by t=0.65 of the @xmath71.,scaledwidth=60.0% ] by lemma ( [ dagger ] ) , we observe that @xmath72 , i.e. , the first digit distributions of @xmath19 and @xmath73 are not the same , indicating that @xmath19 is not scale - invariant .
one can see that only a uniform @xmath31 remains unchanged under the wrap - around effect induced by a scaling of @xmath19 .
[ t : ben - scale ] only the random variables characterized by @xmath74 are scale - invariant .
by lemma ( [ l : scale - tr ] ) , scaling of @xmath19 yields a wrap - around effect of @xmath31 . for a benford distribution ,
we want the same areas under the @xmath31 and @xmath75 over the intervals @xmath76 @xmath77 . here
@xmath75 denotes the mod 1 projection of @xmath65 which is the shifted @xmath40 .
clearly , a uniform @xmath31 will satisfy this constraint . any other function when translated by an arbitrary amount , will not simultaneously keep the areas under @xmath31 preserved over all of these intervals .
thus , the only @xmath31 left unchanged under all translations is the uniform one .
we are now arriving at two important results first proved by hill @xcite concerning scale- and base - invariance .
scale - invariance implies benford s law . in order to have scale - invariance , we must have @xmath74 . from lemma ( [ dagger ] ) ,
one obtains benford s law .
base - invariance means that for any base @xmath78 , the probability of having @xmath3 as the first digit is @xmath79 scale - invariance implies base - invariance .
scale - invariance requires @xmath74 .
then , for an arbitrary base @xmath78 , integrating over the logarithmic intervals @xmath80 , we get benford s distribution @xmath81 using the insight gained by the mod 1 map analysis , we can construct an infinite class of benford distributed random variables with non - compact support .
to this end , let us consider @xmath60 with the following density function @xmath82,\ ] ] as illustrated in figure [ f : series ] . with non - compact support yielding uniform @xmath31.,scaledwidth=45.0% ] here
@xmath83 is the heaviside step function . since the series @xmath84 , one obtains @xmath30 and thus , @xmath66 is benford distributed .
a similar construction can be used for any convergent series .
we define a random variable that has a logarithmic distribution in its first @xmath0 digits to be a @xmath0-digit benford variable .
more precisely , such a variable is benford in the first @xmath0 digits , but it is not guaranteed to have a logarithmic distribution beyond the @xmath0-th digit .
let @xmath85 .
a random variable is @xmath0-digit benford if for all @xmath22 and all @xmath86 , @xmath87 , @xmath88 to show that such variables exist , we use ( [ e : five ] ) to construct a @xmath31 that satisfies ( [ e : benf ] ) for any @xmath85 .
that is , we search for @xmath31 such that @xmath89 we proceed by the partition @xmath90 of the @xmath91 interval . then an example of @xmath31 which yields a @xmath1-digit benford distribution is given by @xmath92 as illustrated in figure [ f : sine ] . yielding a @xmath1-digit benford random variable.,scaledwidth=45.0% ]
one can readily check that @xmath93 . by theorem ( [ t : ben - scale ] ) any density function @xmath40 whose @xmath31 is non - uniform is not scale - invariant .
thus the example of equation ( [ e : non - u ] ) while yielding a @xmath1-digit benford distributed variable , is not scale- nor base - invariant . following this idea , one can generalize the example given in ( [ e : non - u ] ) in order to produce a @xmath31 which satisfy ( [ e : star ] ) with an arbitrary large @xmath0 .
for such a construction , let @xmath94 be a partition of the interval @xmath91 .
let @xmath95 \to [ 0,\infty)$ ] be a probability density function .
for @xmath96 such that @xmath97 , let us define @xmath98 it is easy to see that for all @xmath99 , @xmath100 therefore , we have that @xmath31 behaves as the uniform density ( @xmath101 for all @xmath102 ) when integrated over intervals of the form @xmath103 with @xmath104 .
using mathematica @xcite , we modeled a 1-digit benford random variable by generating its discrete approximation with @xmath105 data points . as a concrete example we took @xmath66 where @xmath60 has the density function @xmath106 given by ( [ e : non - u ] ) . from @xmath107
, we obtain the density function for @xmath19 : @xmath108 the 100-bin histogram of these data points is shown in figure [ f : subfig1 ] along with the graph of the density function @xmath38 represented by dashed lines .
the mathematica generated data closely follow the actual curve of the density function , as we can see in [ f : subfig1 ] . in figure
[ f : subfig2 ] , we show the histogram of the first digit distribution of the experimental points versus the theoretical benford probabilities . the numerical results are recorded in table [ t : data ] . note the accuracy of the mathematica generated numbers .
.theoretical vs. experimental ( mathematica generated ) probabilities for the 1-digit benford random variable ( [ e : f - non - u ] ) [ cols="^,^,^,^,^,^,^,^,^,^",options="header " , ]
in this paper , we demonstrated the importance of mod 1 map in analyzing the digit distribution of random variables .
in particular , we have shown that a uniform @xmath31 implies scale- and base - invariance .
we have introduced the concept of @xmath0-digit benford , and gave a concrete example of a 1-digit benford random variable .
furthermore , we have shown how to use the mod 1 map to construct a @xmath0-digit benford variable starting with any density function .
we generated , with the help of the computer algebra system mathematica , a @xmath105 data points discrete approximation of a 1-digit benford variable , and found close agreement between the first digit probabilities of the model against the expected theoretical probabilities .
we are grateful to the anonymous referee for the valuable and constructive suggestions that helped improve this work .
we also would like to thank prof .
berger for very useful discussions . | the scope of this paper is twofold .
first , to emphasize the use of the mod 1 map in exploring the digit distribution of random variables . we show that the well - known base- and scale - invariance of benford variables are consequences of their associated mod 1 density functions being uniformly distributed .
second , to introduce a new concept of the @xmath0-digit benford variable .
such a variable is benford in the first @xmath0 digits , but it is not guaranteed to have a logarithmic distribution beyond the @xmath0-th digit .
we conclude the paper by giving a general construction method for @xmath0-digit benford variables , and provide a concrete example . |
large - scale diffuse soft x - ray emission ( galactic coronae ) has been detected around various types of nearby galaxies ( e.g. , @xcite and references therein ) .
they include elliptical galaxies containing active galactic nuclei ( agn ) ( e.g. , @xcite ) or not ( e.g. , @xcite ) , normal disk galaxies with little star formation ( sf ) ( e.g. , @xcite ) , galaxies with active sf spread widely over their disk ( e.g. , @xcite ) , nuclear starburst galaxies ( e.g. , @xcite ) , dwarf galaxies ( e.g. , @xcite ) , and interacting or merging galaxies ( e.g. , @xcite ) , etc .
the presence of such galactic coronae is a manifestation of the energetic interplay between galaxies and their surroundings , although the exact origin of the x - ray emission remains very uncertain . in x - ray observations of _ active sf _ galaxies ,
the coronal luminosity is observed to be tightly ( nearly linearly ) correlated with the sf rate ( sfr ; e.g. , @xcite ) .
furthermore , the coronal gas often shows metallicity patterns consistent with being enriched by massive stellar objects [ e.g. , massive stellar winds and core collapsed ( cc ) sne ] ( e.g. , @xcite ) .
these strongly suggest that the coronal gas in star forming galaxies is mainly generated by sf feedback .
in addition , deep observations of some nuclear starburst galaxies have revealed a high - temperature ( @xmath1 ) thermal component , which is usually not detected because of its low density and hence low emissivity , compared to the dominant soft x - ray - emitting gas of typical temperature of a few @xmath2 @xcite .
this high - temperature component is thought to be produced by the sn - driven superwinds themselves , which carry the bulk of the sn feedback energy and interact with the cool interstellar medium ( ism ) , producing much of the soft x - ray emission at the interfaces @xcite , although this view is still far from being widely accepted ( e.g. , see @xcite ) . in contrast , the diffuse soft x - ray luminosity ( @xmath3 ) of sf - quiescent _ elliptical _ galaxies is shown to be closely correlated with the stellar mass ( @xmath4 ) .
furthermore , the slope of the @xmath5 relation is shown to be far larger than unity , especially for massive galaxies ( e.g. , @xcite ) , strongly suggesting that gravity plays an important role in heating and/or confining the coronal gas .
metallicity measurements further indicate that the hot gas in elliptical galaxies ( maybe except for those located in cluster environments ) is primarily enriched by old stars ( via type ia sne and stellar mass loss , e.g. , @xcite ) .
the modeling of this stellar feedback shows that galactic scale outflow are most likely responsible for removing a large fraction of the sn energy , and chemically - enriched mass from central regions of low- and intermediate - mass elliptical galaxies ( e.g. , @xcite ) . but both the dynamics and ultimate fate of the outflows are sensitive to the feedback history , as well as to the gravitational confinement of the galaxies ( e.g. , @xcite ) .
the origin of the coronal gas remains very uncertain for relatively _ quiescent disk _ galaxies ( such as the milky way and m31 ) .
indeed , the hierarchical galaxy formation scenario definitively predicts the presence of coronae around galaxies with individual halo masses greater than a few @xmath6 ( e.g. , @xcite ) .
such coronae are formed primarily from the ( virial ) shock - heating and subsequent gravitational compression of gas accreted from the intergalactic medium ( igm ) .
but non - gravitational heating is also expected to be important .
observationally , non - nuclear - starburst galaxies , for example , seem to follow the same @xmath7 relation as nuclear starburst ones @xcite . even in passively evolving bulge - dominated disk galaxies ( e.g. , s0 galaxies ) which are often quiescent
, non - gravitational feedback can still be significant , primarily in forms of ia sne and mass - loss of evolved low - mass stars ( e.g. , @xcite ) .
it is thus of great interest to check if the coronae around quiescent disk galaxies represent the interplay between gravitational and non - gravitational effects , and what roles different heating mechanisms play in producing galactic coronae in and around various types of galaxies .
we herein present a comprehensive _ chandra _ data analysis for a sample of 53 nearby highly - inclined _ disk _ galaxies , together with multi - wavelength measurements of other galaxy properties , such as the sfr , stellar mass , and dynamical mass .
our sample includes all kinds of disk galaxies and covers a large range of galaxy properties . in comparison , similar existing studies of nearby highly - inclined disk galaxies , based on _ chandra _ and _ xmm - newton _ observations , all have sample sizes less than about 10 galaxies @xcite ( fig .
[ fig : sample ] ) .
elliptical galaxies are relatively well studied ( e.g. , @xcite ) and the results will be compared with those from the present study in the companion paper ( @xcite , hereafter paper ii ) .
the large sample size of the disk galaxies in the present study will enable us to examine the similarities and differences between various types of galaxies . by conducting a multi - variable statistical analysis
, we will also examine how the coronal properties are affected by other processes , such as the environmental effects ( e.g. , @xcite ) and the cool - hot gas interaction ( e.g. , @xcite ) .
this analysis of archival _ chandra _ data forms part of a long - term multi - wavelength investigation of galactic halos and their interplay with galactic disks .
in particular , an extensive radio survey of 35 nearby highly - inclined disk galaxies , which represents a volume - limited sample , is being carried out with the _ evla _ ( the _ chang - es _ project ; @xcite ) .
16 of these galaxies are included in the present _ chandra _ analysis , which represents the first step to provide the x - ray data coverage of the _ chang - es _ galaxies .
this archival study represents a necessary preparation for proposing new x - ray observations for deep exposures or for additional galaxies . with the superb spatial resolution ,
_ chandra _ observations are ideal for detecting and subtracting point - like sources , mapping out galactic coronae , particularly in regions close to the galactic disks .
an analysis program , using existing and new _ xmm - newton _ observations , is also planned , which can provide complementary capabilities ( e.g. , greater counting statistics and field coverage ) and are better suited for spectral analysis of large - scale x - ray emission .
the present paper utilizes mainly the _ chandra _ data for the detection and excision of x - ray point sources and the characterization of the galactic coronae in and around the sample galaxies .
we report the x - ray measurements of the coronae in this paper . in [ paperisec : sampleselection ]
we introduce the sample selection criteria and the definitions of subsamples .
the data reduction and analysis are detailed in
[ paperisec : datereduction ] , providing uniform measurements of the diffuse soft x - ray properties for all the sample galaxies .
notes on individual galaxies of distinct properties are given in
[ paperisec : individual ] .
we discuss uncertainties related to the data analysis in
[ paperisec : discussion ] .
the main results are summarized in [ paperisec : summary ] .
paper ii presents ( 1 ) a detailed correlation analysis between the x - ray properties ( e.g. , luminosity , temperature , and abundance ratio ) and multi - wavelength galaxy properties ( e.g. , sf properties , galaxy mass , and morphology ) for different types of galaxies ; ( 2 ) a comparison of our measurements ( e.g. , the coronal luminosity and/or temperature ) with those from cosmological simulations or _ chandra
_ observations of elliptical galaxies ; ( 3 ) an estimate of the x - ray radiation efficiency of the galactic feedback ; and ( 4 ) an investigation of the origin of the coronal emission .
we are aiming at understanding the relative importance of various physical processes in generating galactic coronae in and around different types of disk galaxies .
we thus need a large sample of galaxies with all types of sf properties , masses , morphologies , and formation environments . limited by the availability of existing _
observations , our sample selection for the present study is thus related to , but not limited by , the _ chang - es _
sample @xcite , with the following criteria : \(1 ) _ morphological classification . _
the present study includes both spiral and lenticular galaxies with optical morphological type codes of @xmath8 ( table [ table : sampleselectionpara ] ) .
this means that some interacting galaxies may be included .
these galaxies are sometimes observed to have quite different x - ray properties from the isolated ones ( e.g. , @xcite ) , and may provide us with a laboratory to study how galaxy interaction could regulate the coronal properties .
\(2 ) _ inclination angle . _ with the inclination angle restricted to @xmath0 , the sample includes not only nearly perfect edge - on galaxies , but also some moderately inclined ones . for the latter ones ,
additional inclination corrections may be needed to uniformly compare them with other galaxies in the sample .
but in general , the selection of only highly inclined disk galaxies minimizes the confusion of the soft x - ray emission from discrete sources in galactic disks ( young stars , supernova remnants , etc . ) .
the present study thus preferentially samples the diffuse soft x - ray emission from regions outside galactic disks with dense cool gas .
\(3 ) _ distance .
_ all the selected galaxies have distances @xmath9 so that the bulk of luminous point sources can be individually detected and that the disk / halo emission can be reasonably separated .
the distances of the galaxies are obtained based on various redshift independent methods , as listed in table [ table : distance ] .
\(4 ) _ extinction .
_ all the galaxies are selected to have the foreground column density @xmath10 , minimizing the galactic extinction in soft x - ray .
\(5 ) _ size .
_ all the galaxies have optical diameters @xmath11 so that a single _ chandra _ observation could typically cover the bulk of the coronal emission .
@xmath12 is also limited to @xmath13 so that the disk and halo can be well separated .
\(6 ) _ chandra data . _
each of the selected galaxies has a total non - grating _
chandra_/acis exposure @xmath14 .
we only include observations with individual exposures @xmath15 ( table [ table : chandradata ] ) .
galaxies with only multiple snapshot observations , even with a total exposure @xmath15 , are not included , except for ngc 660 ( @xmath16 ) and ngc 4666 ( @xmath17 ) , which are known to have high sfrs and are expected to have strong diffuse x - ray emission .
\(7 ) _ agn contamination .
_ we do _ not _ select our sample against galaxies containing agn .
nevertheless , we have excluded one galaxy ( ngc 2992 ) from our original sample , because the scattered x - ray light from its agn ( wings of point - spread function , as well as a severe ccd readout steak ) is too bright to allow for an effective study of the galactic corona ( the thermal component is @xmath18 of the total diffuse x - ray emission in the near - nuclear region of @xmath19 ; @xcite ) .
other galaxies may contain agn ; but their presence does not seriously affect the data reduction .
lcccccccccccccc ic2560 & sbb & @xmath20 & 65.63 & 29.2 & 6.51 & 3.55 & 2.13 & - & @xmath21 & - & - & - + m82@xmath22 & sd & @xmath23 & 79.4 & 3.53 & 3.98 & 11.0 & 2.15 & 0.674 & @xmath24 & 0.16 & - & - + ngc24 & sc & @xmath25 & 70.11 & 9.08 & 2.27 & 6.18 & 2.57 & 0.480 & @xmath26 & 0.12 & - & - + ngc520@xmath27 & sa & @xmath28 & 77.45 & 27.8 & 3.27 & 4.09 & 2.55 & 0.699 & @xmath29 & 0.25 & - & - + ngc660 & sba & @xmath30 & 78.85 & 14.7 & 4.86 & 4.57 & 2.71 & 0.710 & @xmath31 & 0.12 & - & - + ngc891@xmath22 & sb & @xmath32 & 88 & 9.95 & 7.64 & 13.0 & 4.31 & 0.697 & @xmath33 & 0.55 & 71.12 & 81.66 + ngc1023@xmath34 & e - s0 & @xmath35 & 76.7 & 11.6 & 7.16 & 7.40 & 2.42 & 0.911 & @xmath36 & 0.57 & - & - + ngc1380 & s0 & @xmath37 & 90 & 21.2 & 1.31 & 4.58 & 2.07 & 0.882 & - & 1.54 & - & - + ngc1386@xmath38 & s0-a & @xmath39 & 90 & 15.3 & 1.37 & 3.59 & 2.67 & 0.776 & - & 1.36 & 2.93 & @xmath400.40 + ngc1482@xmath22 & s0-a & @xmath41 & 63.59 & 19.6 & 3.69 & 2.46 & 1.72 & 0.860 & @xmath42 & 0.31 & 31.82 & 5.16 + ngc1808@xmath43 & saba & @xmath44 & 83.91 & 12.3 & 2.70 & 5.42 & 2.96 & 0.690 & @xmath45 & 0.30 & - & - + ngc2787@xmath46 & s0-a & @xmath47 & 66.21 & 13.0 & 4.32 & 3.24 & 1.79 & 0.891 & @xmath48 & 0.06 & 0.24 & 7.70 + ngc2841 & sb & @xmath49 & 68 & 14.1 & 1.45 & 6.90 & 2.08 & 0.792 & @xmath50 & 0.13 & - & - + ngc3079@xmath22 & sbcd & @xmath51 & 82.5 & 16.5 & 0.789 & 8.18 & 6.34 & 0.512 & @xmath52 & 0.29 & 93.56 & 89.35 + ngc3115@xmath46 & e - s0 & @xmath53 & 81.6 & 9.77 & 4.32 & 7.10 & 2.36 & 0.899 & @xmath54 & 0.08 & 0.05 & 5.74 + ngc3198 & sc & @xmath55 & 70 & 14.5 & 1.02 & 6.46 & 3.51 & 0.421 & @xmath56 & 0.15 & - & - + ngc3384@xmath34 & e - s0 & @xmath57 & 90 & 11.8 & 2.73 & 5.24 & 2.21 & 0.875 & 17 & 0.54 & - & - + ngc3412 & s0 & @xmath58 & 71.92 & 11.5 & 2.59 & 3.96 & 1.83 & 0.850 & - & 0.52 & 0.06 & @xmath400.22 + ngc3521 & sabb & @xmath59 & 65.5 & 11.2 & 4.06 & 8.32 & 1.86 & 0.698 & @xmath60 & 0.19 & 77.23 & 90.74 + ngc3556@xmath61 & sbc & @xmath62 & 67.51 & 10.7 & 0.794 & 3.98 & 2.40 & 0.568 & @xmath63 & 0.15 & 14.58 & 49.39 + ngc3628@xmath22 & sb & @xmath64 & 79.29 & 13.1 & 2.22 & 11.0 & 3.21 & 0.675 & @xmath65 & 0.39 & - & - + ngc3877@xmath66 & sc & @xmath67 & 83.24 & 14.1 & 2.22 & 5.36 & 4.36 & 0.654 & @xmath68 & 1.53 & 10.28 & 12.64 + ngc3955 & s0-a & @xmath69 & 90 & 20.6 & 4.74 & 4.15 & 3.44 & 0.522 & 86 & 0.08 & - & - + ngc3957 & s0-a & @xmath70 & 90 & 27.5 & 3.63 & 3.24 & 5.32 & - & - & 0.48 & - & @xmath4043.32 + ngc4013 & sb & @xmath32 & 90 & 18.9 & 1.39 & 4.89 & 3.96 & 0.829 & 182 & 1.34 & - & - + ngc4111 & s0-a & @xmath71 & 84.2 & 15.0 & 1.40 & 1.78 & 2.79 & 0.815 & @xmath72 & 1.09 & 0.17 & 7.21 + ngc4217 & sb & @xmath32 & 81.05 & 19.5 & 1.23 & 5.43 & 3.37 & 0.750 & @xmath73 & 0.95 & - & - + ngc4244@xmath22 & sc & @xmath74 & 88 & 4.37 & 1.67 & 16.2 & 2.24 & 0.412 & @xmath75 & 0.39 & 0.60 & 26.77 + ngc4251@xmath38 & s0 & @xmath76 & 74.56 & 19.6 & 1.84 & 2.34 & 1.91 & 0.800 & - & 1.20 & @xmath400.35&@xmath400.57 + ngc4342@xmath77 & e - s0 & @xmath78 & 90 & 16.8 & 1.60 & 1.25 & 1.96 & 0.964 & - & 2.64 & - & @xmath400.76 + ngc4388@xmath79 & sb & @xmath80 & 82 & 17.1 & 2.60 & 5.38 & 4.18 & 0.575 & @xmath81 & 1.56 & 6.07 & 5.67 + ngc4438@xmath82 & sa & @xmath83 & 73.29 & 14.4 & 2.66 & 9.16 & 2.28 & 0.756 & @xmath84 & 2.67 & - & - + ngc4501 & sb & @xmath85 & 61 & 15.7 & 2.48 & 8.65 & 1.98 & 0.623 & @xmath86 & 2.04 & 83.00 & 23.39 + ngc4526@xmath34 & s0 & @xmath87 & 82.8 & 17.2 & 1.65 & 6.95 & 2.81 & 0.890 & @xmath88 & 2.45 & 6.03 & 24.56 + ngc4565@xmath66 & sb & @xmath89 & 90 & 11.1 & 1.30 & 16.7 & 5.79 & 0.677 & @xmath90 & 1.00 & 38.94 & 109.75 + ngc4569 & saba & @xmath91 & 66 & 9.86 & 2.49 & 9.12 & 2.39 & 0.605 & @xmath92 & 1.15 & - & - + ngc4594@xmath93 & sa & @xmath94 & 78.5 & 9.77 & 3.77 & 8.45 & 1.72 & 0.878 & @xmath95 & 0.32 & @xmath4015.65 & 3.50 + ngc4631@xmath96 & sbcd & @xmath97 & 85 & 7.62 & 1.27 & 14.5 & 6.58 & 0.386 & @xmath98 & 0.41 & - & - + ngc4666 & sabc & @xmath99 & 69.58 & 15.7 & 1.74 & 4.97 & 2.52 & 0.641 & @xmath100 & 0.54 & 79.70 & 43.80 + ngc4710 & s0-a & @xmath101 & 90 & 16.8 & 2.14 & 4.39 & 3.88 & 0.766 & 160 & 2.00 & 8.47 & 0.77 + ngc5102@xmath102 & e - s0 & @xmath103 & 90 & 3.18 & 4.33 & 9.71 & 2.62 & 0.638 & 90 & 0.17 & @xmath400.08 & 2.54 + ngc5170@xmath104 & sc & @xmath105 & 90 & 22.5 & 6.95 & 7.96 & 5.65 & 0.705 & 245 & 0.20 & - & - + ngc5253@xmath106 & s ? &
@xmath107 & 66.7 & 4.07 & 3.87 & 5.01 & 2.36 & 0.300 & @xmath108 & 0.18 & - & 2.34 + ngc5422 & s0-a & @xmath109 & 90 & 30.9 & 1.16 & 2.81 & 4.79 & 0.900 & - & 0.36 & - & 34.87 + ngc5746@xmath104 & sabb & @xmath32 & 83.9 & 24.7 & 3.27 & 7.24 & 6.73 & 0.769 & @xmath110 & 0.83 & - & - + ngc5775@xmath111 & sbc & @xmath112 & 83.22 & 26.7 & 3.48 & 3.70 & 4.38 & 0.660 & @xmath113 & 0.67 & 67.59 & 128.79 + ngc5866@xmath114 & s0-a & @xmath115 & 90 & 15.3 & 1.46 & 6.31 & 2.32 & 0.785 & - & 0.24 & 7.71 & @xmath403.00 + ngc6503@xmath22 & sc & @xmath116 & 73.5 & 5.27 & 4.09 & 5.94 & 3.00 & 0.559 & @xmath117 & 0.08 & 2.84 & 13.91 + ngc6764@xmath118 & sbbc & @xmath119 & 63.86 & 26.2 & 6.07 & 2.49 & 2.04 & 0.571 & @xmath120 & 0.08 & - & - + ngc7090@xmath121 & sbc & @xmath122 & 90 & 6.28 & 2.76 & 8.15 & 5.00 & 0.451 & 102 & 0.13 & - & - + ngc7457 & e - s0 & @xmath123 & 73.05 & 13.2 & 5.56 & 3.95 & 1.75 & 0.815 & - & 0.13 & 0.05 & @xmath400.73 + ngc7582@xmath124 & sbab & @xmath125 & 68.17 & 23.0 & 1.93 & 6.95 & 2.19 & 0.661 & @xmath126 & 0.39 & - & - + ngc7814 & sab & @xmath127 & 70.59 & 18.1 & 3.71 & 4.37 & 2.32 & 0.872 & @xmath128 & 0.12 & @xmath4014.56 & 17.10 [ table : sampleselectionpara ] lcllcllcl ic2560@xmath22 & @xmath129 & tully - fisher & ngc3521@xmath22 & @xmath130 & tully - fisher & ngc4594@xmath93 & @xmath131 & sbf + m82@xmath27 & @xmath132 & trgb & ngc3556@xmath22 & @xmath133 & tully - fisher & ngc4631@xmath96 & @xmath134 & trgb + ngc24@xmath22 & @xmath135 & tully - fisher & ngc3628@xmath82 & @xmath136 & tully - fisher & ngc4666@xmath22 & @xmath137 & tully - fisher + ngc520@xmath34 & @xmath138 & tully - fisher & ngc3877@xmath22 & @xmath139 & tully - fisher & ngc4710@xmath34 & @xmath140 & tully est + ngc660@xmath38 & @xmath141 & tully - fisher & ngc3955@xmath34 & @xmath142 & tully est & ngc5102@xmath114 & @xmath143 & trgb + ngc891@xmath43 & @xmath144 & pnlf & ngc3957@xmath34 & @xmath145 & tully est & ngc5170@xmath106 & @xmath146 & gclf + ngc1023@xmath46 & @xmath147 & sbf & ngc4013@xmath22 & @xmath148 & tully - fisher & ngc5253@xmath79 & @xmath149 & cepheids + ngc1380@xmath61 & @xmath150 & snia & ngc4111@xmath93 & @xmath151 & sbf & ngc5422@xmath34 & @xmath152 & tully est + ngc1386@xmath66 & @xmath153 & sbf & ngc4217@xmath22 & @xmath154 & tully - fisher & ngc5746@xmath22 & @xmath155 & tully - fisher + ngc1482@xmath34 & @xmath156 & tully est & ngc4244@xmath96 & @xmath157 & trgb & ngc5775@xmath34 & @xmath158 & tully est + ngc1808@xmath22 & @xmath159 & tully - fisher & ngc4251@xmath93 & @xmath160 & sbf & ngc5866@xmath93 & @xmath161 & sbf + ngc2787@xmath34 & @xmath162 & tully est & ngc4342@xmath34 & @xmath140 & tully est & ngc6503@xmath77 & @xmath163 & trgb + ngc2841@xmath79 & @xmath164 & cepheids & ngc4388@xmath102 & @xmath165 & tully - fisher & ngc6764@xmath22 & @xmath166 & tully - fisher + ngc3079@xmath22 & @xmath167 & tully - fisher & ngc4438@xmath82 & @xmath168 & tully - fisher & ngc7090@xmath22 & @xmath169 & tully - fisher + ngc3115@xmath46 & @xmath170 & sbf & ngc4501@xmath104 & @xmath171 & snia & ngc7457@xmath93 & @xmath172 & sbf + ngc3198@xmath79 & @xmath173 & cepheids & ngc4526@xmath61 & @xmath174 & snia & ngc7582@xmath38 & @xmath175 & tully - fisher + ngc3384@xmath46 & @xmath176 & sbf & ngc4565@xmath106 & @xmath177 & gclf & ngc7814@xmath22 & @xmath178 & tully - fisher + ngc3412@xmath46 & @xmath179 & sbf & ngc4569@xmath111 & @xmath180 & tully - fisher & & & [ table : distance ] lcccclcccclcccc ic2560 & 1592 & acis - s & 9.96 & 9.63 & ngc3384 & 4692 & acis - s & 10.03 & 9.47 & ngc4569 & 5911 & acis - s & 39.65 & 29.63 + & 4908 & acis - s & 55.67 & 54.24 & & 11782 & acis - s & 29.04 & 28.67 & ngc4594 & 1586 & acis - s & 18.75 & 18.12 + m82 & 361 & acis - i & 33.68 & 32.14 & ngc3412 & 4693 & acis - s & 10.09 & 9.96 & & 9532 &
acis - i & 86.04 & 83.34 + & 1302 & acis - i & 15.71 & 15.36 & ngc3521 & 4694 & acis - s & 10.07 & 6.59 & & 9533 &
acis - i & 90.15 & 87.00 + & 2933 & acis - s & 18.25 & 17.83 & & 9552 & acis - i & 72.41 & 71.39 & ngc4631 & 797 & acis - s & 59.97 & 56.45 + ngc24 & 9547 & acis - s & 43.78 & 41.99 & ngc3556 & 2025 & acis - s & 60.12 & 57.98 & ngc4666 & 4018 & acis - s & 5.00 & 4.74 + ngc520 & 2924 & acis - s & 49.96 & 32.19 & ngc3628 & 2039 & acis - s & 58.70 & 53.90 & ngc4710 & 9512 & acis - s & 30.16 & 29.57 + ngc660 & 1633 & acis - s & 1.94 & 1.91 & & 2918 & acis - s & 22.26 & 21.08 & ngc5102 & 2949 & acis - s & 34.66 & 32.77 + & 4010 & acis - s & 5.13 & 4.87 & & 2919 & acis - s & 22.76 & 22.28 & ngc5170 & 3928 &
acis - i & 33.45 & 32.58 + ngc891 & 794 & acis - s & 51.56 & 29.03 & ngc3877 & 767 & acis - s & 19.16 & 18.71 & ngc5253 & 2032 & acis - s & 57.36 & 49.72 + & 4613 & acis - s & 120.40 & 100.32 & & 768 & acis - s & 23.75 & 22.45 & & 7153 & acis - s & 69.12 & 67.66 + ngc1023 & 4696 & acis - s & 10.37 & 9.52 & & 952 & acis - s & 20.02 & 19.57 & & 7154 & acis - s & 67.49 & 62.09 + & 8197 & acis - s & 48.88 & 47.04 & & 1971 & acis - s & 29.55 & 26.82 & ngc5422 & 9511 & acis - s & 18.15 & 17.92 + & 8198 & acis - s & 50.38 & 49.13 & & 1972 & acis - s & 29.09 & 13.82 & & 9772 & acis - s & 18.64 & 18.20 + & 8464 & acis - s & 48.18 & 47.16 & ngc3955 & 2955 & acis - s & 19.96 & 18.73 & ngc5746 & 3929 &
acis - i & 37.29 & 35.82 + & 8465 & acis - s & 45.76 & 44.17 & ngc3957 & 9513 & acis - s & 38.03 & 36.55 & ngc5775 & 2940 & acis - s & 58.96 & 58.02 + ngc1380 & 9526 & acis - s & 42.18 & 40.05 & ngc4013 & 4739 & acis - s & 80.11 & 75.86 & ngc5866 & 2879 & acis - s & 34.18 & 30.58 + ngc1386 & 4076 & acis - s & 19.89 & 19.64 & ngc4111 & 1578 & acis - s & 15.00 & 14.76 & ngc6503 & 872 & acis - s & 13.36 & 11.27 + ngc1482 & 2932 & acis - s & 28.56 & 23.78 & ngc4217 & 4738 & acis - s & 73.66 & 69.37 & ngc6764 & 9269 & acis - s & 20.18 & 19.72 + ngc1808 & 3012 & acis - s & 43.43 & 39.52 & ngc4244 & 942 & acis - s & 49.78 & 46.21 & ngc7090 & 7060 & acis - s & 26.41 & 24.53 + ngc2787 & 4689 & acis - s & 31.24 & 29.22 & ngc4251 & 4695 & acis - s & 10.18 & 9.85 & & 7252 & acis - s & 31.02 & 25.47 + ngc2841 & 6096 & acis - s & 28.58 & 26.05 & ngc4342 & 4687 & acis - s & 38.75 & 31.49 & ngc7457 & 4697 & acis - s & 9.11 & 5.63 + ngc3079 & 2038 & acis - s & 26.92 & 25.79 & ngc4388 & 1619 & acis - s & 20.23 & 19.75 & & 11786 & acis - s & 29.04 & 28.08 + ngc3115 & 2040 & acis - s & 37.45 & 34.75 & ngc4438 & 2883 & acis - s & 25.40 & 24.68 & ngc7582 & 436 & acis - s & 13.62 & 6.93 + & 11268 & acis - s & 41.10 & 40.19 & ngc4501 & 2922 & acis - s & 18.10 & 17.28 & ngc7814 & 11309 & acis - s & 59.11 & 57.74 + & 12095 & acis - s & 76.65 & 74.50 & ngc4526 & 3925 & acis - s & 44.11 & 38.00 & & & & & + ngc3198 & 9551 & acis - s & 62.42 & 60.64 & ngc4565 & 3950 & acis - s & 59.95 & 50.31 & & & & & [ table : chandradata ] lccccccccc ic2560 & @xmath181 & @xmath182 & @xmath183 & @xmath184 & @xmath185 & @xmath186 & @xmath187 & @xmath188 & @xmath189 + m82 & @xmath190 & @xmath191 & @xmath192 & @xmath193 & @xmath194 & @xmath195 & @xmath196 & @xmath197 & @xmath198 + ngc24 & @xmath199 & @xmath200 & @xmath201 & @xmath202 & @xmath203 & @xmath204 & @xmath205 & @xmath206 &
+ ngc520 & @xmath207 & @xmath208 & @xmath209 & @xmath210 & @xmath211 & @xmath212 & @xmath213 & @xmath214 & @xmath215 + ngc660 & @xmath216 & @xmath217 & @xmath218 & @xmath219 & @xmath220 & @xmath221 & @xmath222 & @xmath223 & @xmath224 + ngc891 & @xmath225 & @xmath226 & @xmath227 & @xmath228 & @xmath229 & @xmath230 & @xmath231 & @xmath232 & @xmath233 + ngc1023 & @xmath234 & @xmath235 & & & & & & & + ngc1380 & @xmath236 & @xmath237 & @xmath238 & @xmath238 & @xmath239 & @xmath240 & @xmath241 & @xmath242 & + ngc1386 & @xmath243 & @xmath244 & @xmath245 & @xmath246 & @xmath247 & @xmath248 & @xmath249 & @xmath250 & @xmath251 + ngc1482 & @xmath252 & @xmath253 & @xmath254 & @xmath255 & @xmath211 & @xmath256 & @xmath257 & @xmath258 & @xmath259 + ngc1808 & @xmath260 & @xmath261 & @xmath262 & @xmath263 & @xmath264 & @xmath265 & @xmath196 & @xmath266 & @xmath267 + ngc2787 & @xmath268 & @xmath269 & @xmath270 & @xmath271 & @xmath272 & @xmath273 & @xmath274 & @xmath275 & @xmath276 + ngc2841 & @xmath277 & @xmath278 & @xmath238 & @xmath238 & @xmath279 & @xmath280 & @xmath281 & @xmath282 & @xmath283 + ngc3079 & @xmath284 & @xmath285 & @xmath286 & @xmath287 & @xmath288 & @xmath289 & @xmath290 & @xmath291 & @xmath292 + ngc3115 & @xmath293 & @xmath294 & & & & & & & + ngc3198 & @xmath295 & @xmath296 & @xmath238 & @xmath297 & @xmath298 & @xmath299 & @xmath300 & @xmath301 & @xmath302 + ngc3384 & @xmath303 & @xmath304 & & & & & & & + ngc3412 & @xmath305 & @xmath306 & & & & & & & + ngc3521 & @xmath307 & @xmath308 & @xmath309 & @xmath310 & @xmath311 & @xmath312 & @xmath313 & @xmath314 & @xmath315 + ngc3556 & @xmath316 & @xmath317 & @xmath318 & @xmath319 & @xmath320 & @xmath321 & @xmath322 & @xmath222 & @xmath323 + ngc3628 & @xmath324 & @xmath325 & @xmath326 & @xmath327 & @xmath212 & @xmath328 & @xmath329 & @xmath330 & @xmath331 + ngc3877 & @xmath332 & @xmath333 & @xmath334 & @xmath335 & @xmath336 & @xmath337 & @xmath338 & @xmath339 & @xmath340 + ngc3955 & @xmath341 & @xmath342 & @xmath343 & @xmath344 & @xmath345 & @xmath346 & @xmath347 & @xmath188 & @xmath348 + ngc3957 & @xmath349 & @xmath350 & @xmath238 & @xmath351 & @xmath352 & @xmath353 & @xmath354 & @xmath355 & + ngc4013 & @xmath356 & @xmath357 & @xmath238 & @xmath358 & @xmath359 & @xmath360 & @xmath361 & @xmath362 & @xmath108 + ngc4111 & @xmath363 & @xmath364 & & & & & & & @xmath365 + ngc4217 & @xmath356 & @xmath366 & & & & & & & @xmath367 + ngc4244 & @xmath368 & @xmath369 & @xmath370 & @xmath238 & @xmath371 & @xmath372 & @xmath373 & @xmath374 & + ngc4251 & @xmath375 & @xmath376 & & & & & & & + ngc4342 & @xmath377 & @xmath378 & & & & & & & + ngc4388 & @xmath379 & @xmath380 & @xmath381 & @xmath382 & @xmath383 & @xmath384 & @xmath385 & @xmath386 & @xmath387 + ngc4438 & @xmath388 & @xmath389 & @xmath238 & @xmath201 & @xmath390 & @xmath391 & @xmath392 & @xmath393 & @xmath394 + ngc4501 & @xmath234 & @xmath395 & @xmath396 & @xmath397 & @xmath398 & @xmath399 & @xmath400 & @xmath213 & @xmath401 + ngc4526 & @xmath402 & @xmath403 & @xmath404 & @xmath405 & @xmath406 & @xmath407 & @xmath408 & @xmath409 & @xmath410 + ngc4565 & @xmath411 & @xmath412 & @xmath413 & @xmath414 & @xmath415 & @xmath416 & @xmath417 & @xmath418 & @xmath419 + ngc4569 & @xmath420 & @xmath421 & @xmath422 & @xmath423 & @xmath424 & @xmath425 & @xmath426 & @xmath427 & @xmath428 + ngc4594 & @xmath429 & @xmath430 & @xmath431 & @xmath432 & @xmath433 & @xmath434 & @xmath435 & @xmath436 & @xmath437 + ngc4631 & @xmath438 & @xmath439 & @xmath440 & @xmath441 & @xmath442 & @xmath443 & @xmath444 & @xmath445 & @xmath446 + ngc4666 & @xmath447 & @xmath448 & @xmath449 & @xmath450 & @xmath451 & @xmath452 & @xmath453 & @xmath454 & @xmath455 + ngc4710 & @xmath456 & @xmath457 & @xmath458 & @xmath459 & @xmath460 & @xmath461 & @xmath462 & @xmath463 & @xmath464 + ngc5102 & @xmath465 & @xmath466 & @xmath238 & @xmath238 & @xmath467 & @xmath468 & @xmath469 & @xmath470 & @xmath471 + ngc5170 & @xmath472 & @xmath473 & @xmath238 & @xmath458 & @xmath474 & @xmath475 & @xmath476 & @xmath477 & @xmath478 + ngc5253 & @xmath479 & @xmath480 & @xmath481 & @xmath482 & @xmath483 & @xmath484 & @xmath485 & @xmath486 & @xmath487 + ngc5422 & @xmath488 & @xmath489 & & & & & & & + ngc5746 & @xmath490 & @xmath491 & @xmath238 & @xmath201 & @xmath492 & @xmath493 & @xmath494 & @xmath495 & @xmath496 + ngc5775 & @xmath497 & @xmath498 & @xmath499 & @xmath500 & @xmath501 & @xmath502 & @xmath257 & @xmath503 & @xmath504 + ngc5866 & @xmath505 & @xmath506 & @xmath507 & @xmath508 & @xmath509 & @xmath510 & @xmath511 & @xmath512 & @xmath513 + ngc6503 & @xmath388 & @xmath514 & @xmath515 & @xmath516 & @xmath517 & @xmath518 & @xmath519 & @xmath520 & @xmath521 + ngc6764 & @xmath522 & @xmath523 & @xmath524 & @xmath525 & @xmath526 & @xmath527 & @xmath318 & @xmath528 & @xmath529 + ngc7090 & @xmath530 & @xmath531 & @xmath238 & @xmath532 & @xmath533 & @xmath534 & @xmath535 & @xmath300 & + ngc7457 & @xmath536 & @xmath537 & & & & & & & + ngc7582 & @xmath538 & @xmath539 & @xmath540 & @xmath541 & @xmath542 & @xmath543 & @xmath544 & @xmath545 & + ngc7814 & @xmath465 & @xmath546 & & & & & & & [ table : basicparaii ] in total ,
53 galaxies are selected for the present study , which form the largest _
sample of nearby disk galaxies .
key parameters of the sample galaxies and the _ chandra _ observations are summarized in tables [ table : sampleselectionpara]-[table : basicparaii ] .
references for those galaxies which have been studied previously for diffuse x - ray emission with the _ chandra _ data are given in table [ table : sampleselectionpara ] ; 20 galaxies are presented for the first time .
the sample coverages of the ir luminosity and warmth ( @xmath547 ) , morphological type , local galaxy number density , and stellar mass are shown in fig .
[ fig : sample ] . in this paper
, we utilize the spearman s rank order coefficient ( @xmath548 ; by definition , @xmath549 ) to describe the goodness of a correlation .
this simple goodness description works for any pair of variables that show a monotonic correlation , insensitive to its exact form .
we consider @xmath550 or @xmath551 as a tight or weak correlation , and @xmath552 as no correlation .
we further characterize important tight correlations with simple expressions , together with the corresponding dispersion measurements [ the root mean square ( @xmath553 ) around the fitted relations ] .
all errors are quoted at the @xmath554 confidence level .
there are several commonly - used sf tracers ( for reviews , e.g. , see @xcite ) . for example , the ultraviolet ( uv , @xmath555 ) emission and the hydrogen recombination lines ( h@xmath556 , h@xmath557 , p@xmath556 , etc . )
are the products of massive stars and are often adopted as direct sf indicators .
however , such relatively short wavelength radiation is often highly attenuated by dust in a highly inclined galactic disk . here
, we estimate the sfr of an individual galaxy using its total ir luminosity ( table [ table : basicparaii ] ) , which is mainly from the dust - reprocessed young stellar emission and is little affected by the extinction .
we further check if such a simple estimation of the sfr is consistent with other extinction - free sf tracers , such as the radio continuum emission ( e.g. , @xcite ) and the molecular gas mass ( e.g. , @xcite ) . here
we use the mass of the molecular gas ( instead of that of the total cold gas including atomic gas , as sometimes considered in similar analysis , e.g. , @xcite ) , because it shows a better correlation with the sf activity ( e.g. , @xcite ) .
the relations between different sf tracers are shown in fig .
[ fig : sftracer ] , and can be expressed with the following equations : @xmath558 @xmath559 despite the tight correlations , all these sf tracers have their own uncertainties .
for example , ir emission can be produced by circum - stellar dust in early - type galaxies @xcite ; radio continuum emission can be produced by agn ( e.g. , @xcite ) ; sf inactive galaxies may still host significant amount of molecular gas @xcite .
furthermore , even if the current sfr is accurately estimated , the different sf history of different types of galaxies also mean that their stellar mass built - up and so feedback history may be different ( e.g. , for s0 galaxies , @xcite ) .
it is still not clear if this difference in sf history can significantly affect the observed coronal properties , or their relations to other galaxy properties ( e.g. , the @xmath7 and @xmath560 relations ; @xcite ) .
the above uncertainties , either in the measurement of sfr or the different sf histories , may result in additional scatters in the following correlation analysis ( paper ii ) .
in addition to the global sfr , another key property of sf is its specific intensity ( sfr per unit area or unit mass ) , which is often traced by the ir warmth ( @xmath547 , a measurement of dust temperature ) @xcite .
it is not surprising that the sfr correlates well with the ir warmth ( fig .
[ fig : sample]a ) , since a galaxy with a higher sfr also tends to have a higher sf intensity
. however , this correlation has a large dispersion , which suggests that other factors such as the sf mode ( e.g. , nuclear starburst vs. disk - wide star forming ) may affect the observed ir properties .
we will thus use both the ir luminosity and warmth to define the starburst subsample ( [ paperisubsubsec : subsample ] ) .
galaxy mass is often described in two ways : the stellar mass ( responsible for old stellar feedback ) and the gravitational mass ( related to the gravitational confinement ) , which are closely correlated with each other ( e.g. , @xcite ) .
we use the _
2mass _ k - band luminosity @xcite and the color - dependent mass - to - light ratio @xcite to estimate the stellar mass ( @xmath4 , table [ table : basicparaii ] ) .
the application of this method is somewhat inaccurate for edge - on galaxies , especially starburst ones , because the extinction can be strong , even in near - ir .
the uncertainty in the mass estimate can lead to scatters in the subsequent statistical analysis .
the baryon mass ( @xmath561 ) is estimated from the rotation velocity ( @xmath562 in table [ table : sampleselectionpara ] ) together with the baryon tully - fisher relation , which represents a tight correlation between the total baryon mass ( though only contributions from stars and cool ism are included ; fig .
[ fig : galaxymass ] ) and the rotation velocity of spiral galaxies ( @xcite ) .
the relation implies an approximate constancy of the baryon to gravitational mass ratio in such galaxies .
our inferred @xmath561 of a galaxy is thus expected to be proportional to its gravitational mass .
the rotation velocity is obtained from the _ hyperleda _ database and is estimated from the 21 cm line widths and/or the rotation curves ( generally from h@xmath556 emission ) , except for m82 ( see [ paperisec : individual ] for the reason ) . for data homogenization , the epidemic method as detailed in @xcite is used .
we emphasize that the accuracy of the mass obtained this way may be affected by other factors : ( 1 ) the galaxy inclination correction of @xmath562 in _ hyperleda _ is conducted using the inclination angle defined in optical ( @xmath563 in table [ table : sampleselectionpara ] ) , which may be different from those of or h@xmath556 gases .
( 2 ) some galaxies in our sample are apparently undergoing tidal interactions with their companions .
the gas in such a galaxy may be strongly disturbed .
consequently , their @xmath562 and mass may not follow the tully - fisher relation . ( 3 ) s0 galaxies may not be well described by the tully - fisher relation , which is defined only for spiral galaxies .
we have double - checked the published rotation curves of some galaxies in our sample ( e.g. , @xcite ) and found that their flattening velocities are consistent with @xmath562 as given in _
we thus consider that the estimated @xmath561 values are generally reliable , except for a few extreme individuals ( see later discussion in [ paperisec : individual ] ) , and sufficiently good for a statistical comparison as is intended in paper ii .
we further define some subsamples of galaxies for later in - depth comparisons .
@xcite separate starburst and non - starburst galaxies by @xmath564 , based on the assumption that galaxies with more intense and/or compact sf tend to have higher dust temperature . from fig . [ fig : sample]a , we see that some non - starburst galaxies with low ir luminosities can also have high @xmath547 ratios . in the present work ,
we define starburst galaxies as those with both @xmath565 and @xmath566 .
other galaxies are defined as non - starburst galaxies . according to this definition ,
a couple of dwarf galaxies with high @xmath547 ratio and low @xmath567 are not considered as starbursts ; the @xmath547 ratios of these low @xmath567 galaxies tend to be contaminated by other non - sf - related heating processes ( e.g. , due to heating by old stars and/or weak agns ; @xcite ) .
we use the local galaxy number density ( @xmath568 ) to characterize the galaxy environment ( fig .
[ fig : sample]b ; table [ table : sampleselectionpara ] ) .
galaxies with @xmath569 are defined as being in the field , while those with @xmath570 as being clustered . we further separate galaxies with different morphological types .
early- and late - type disk galaxies are separated by @xmath571 or roughly the `` sa '' type ( fig .
[ fig : sample]b ) .
the _ chandra _ data used in the present work are summarized in table [ table : chandradata ] .
the data are comprised of 72 acis - s and 7 acis - i observations of the 53 galaxies .
we reprocess the data using ciao ( _ chandra _ interactive analysis of observations ) v.4.2 and the corresponding _ chandra _ calibration database ( caldb ) .
we reduce all the data in a _ uniform _ manner , which allows for comparison among them with minimal calibration biases .
we start our calibration with the level=1 raw data .
we remove afterglow , representing residual charges from the interaction of cosmic rays in a ccd using the ciao tool _
`` dmtcalc''_. we flagged out hot pixels with _
we then generate a new level=2 event file , using _
`` acis_process_events''_. the _ `` eventdef '' _ and _ `` check_vf_pha '' _ parameters are set according to the observational modes , while the tgain ( acis time - dependent gain ) and cti ( charge transfer inefficiency ) calibrations are conducted following the official recommendations of the _ chandra _ data analysis guide to account for the time - dependent degradation of the ccd quantum efficiency .
the remaining events are filtered for good grade , status , and time , and the steak events are also removed . such basic calibrations result in a new level=2 event file for further analysis .
we clean the light curve of each observation in the 2.5 - 7 kev band , which is sensitive to background flares ( short periods of significantly enhanced background count rate ; @xcite ) .
the light curve is extracted from regions located in the same ccd chips as those used for the later scientific analysis , with bright point sources excluded .
we initially identify significant background flares using a @xmath572 clip , check the light curves individually , and change the threshold slightly if necessary .
we then filter the data , removing the time intervals with these background flares , resulting in the effective exposure time of the good time intervals ( @xmath573 ) as listed in table [ table : chandradata ] . for some galaxies ,
there are multiple observations .
observations have taken on m82 , but we only use the three observations analyzed by @xcite , which are sufficient for our measurements of the global coronal properties .
we merge multiple observations together to maximize counting statistics for imaging and spectral analysis . for the present study
, we find that the default astrometric accuracy is good enough ( typically @xmath574 ) . for spatial analysis
, we use the ciao tool _
`` reproject_events '' _ and _ `` dmmerge '' _ to combine multiple event files into a single one . while for spectral analysis
`` specextract '' _ to extract spectra from individual observations and stack them .
we utilize acis stowed data to estimate the quiescent instrument background @xcite .
we find the stowed background data set which best matches an individual observation , according to its observing date .
the exposure of the stowed data is adjusted so that the resultant counts rate in the 10 - 12/10 - 14 kev band [ for the front / back - illuminated ( fi / bi ) chips ] to be the same as that of the target observation .
we generate exposure maps for spatial analysis .
an exposure map accounts for the variation of the effective area across the detector field of view ( fov ) and the telescope dithering . unlike a spectral response file
, an exposure map must be averaged in a relatively broad band .
we produce an exposure map using a power law spectral weighting ( power law index=1.7 ) , which is a reasonable characterization for the cosmic background and the typical point - like sources . for an accurate count flux calculation ,
for example , we make the exposure corrections in up to four relatively narrow bands ( see [ paperisubsubsec : pointsource ] ) . for image construction and rough spatial analysis , we may use corrections in broader bands . for quantitative calculations ,
more accurate spectral response files are adopted ( [ paperisubsec : spec ] ) .
most of the discrete x - ray sources are detected with very limited counting statistics in a typical _ chandra _ observation .
furthermore , the detection threshold can strongly depend on the point - spread function ( psf ) , the local background , and the effective exposure , which all vary significantly across the fov .
our procedure ( refer to @xcite for details ) includes the map detection and maximum likelihood analyses of individual sources , as well as the conventional wavelet detection , and gives a better iterative estimation of the local background ( than using the usual wavelet source detection tools alone ) , which is critical in detecting faint point sources , especially in a crowded field . using this procedure
, we perform source detection in the broad ( b , @xmath575 ) , soft ( s , @xmath576 ) and hard ( h , @xmath577 ) bands . this multi - band detection may help us to reduce false detection and to roughly identify some sources with peculiar hardness ratio .
the detected sources will be used for the subsequent image construction ( [ paperisubsubsec : image ] ) , spatial ( [ paperisubsec : spatial ] ) , and spectral analyses ( [ paperisubsec : spec ] ) . to map out the diffuse x - ray emission ,
we remove the detected discrete sources from the images ( count , background , and exposure ) .
a circular region of twice the 90% energy enclosed radius ( eer ) is excluded around each source of a count rate @xmath578 . for a brighter source ,
the removal radius is further multiplied by a factor of @xmath579 . generally about 96% of the source counts
are excluded in such a removal @xcite .
the images , with or without point sources subtracted , are further smoothed with the ciao tool _
`` csmooth '' _ ( with a gaussian kernel ) , adopting adaptive smoothing scales to match the preset range of the signal - to - noise ratio .
the stowed - background and exposure maps are smoothed with the smoothing scale ( kernel size ) map adaptively calculated for the count image in the broad band .
such an image construction is performed with the same scale map for different bands and is shown as a tri - color image of each galaxy in fig .
[ fig : imagea ] to highlight the hardness variation .
we emphasize that such smoothed images are only used to show the outstanding diffuse x - ray emission features , while all the quantitative measurements are based on the data without any smoothing .
we define the vertical range that contains most of the significant coronal emission .
this range can be used uniformly for spectral extraction and for obtaining the overall luminosities of the coronae .
we extract the surface brightness profile in the 0.5 - 1.5 kev band for each galaxy along its minor ( vertical ) axis .
the horizontal range used to extract such a profile is mostly @xmath580 , which typically contains bulk of the coronal emission ( fig .
[ fig : imagea ] ) , and is adjusted for only a few galaxies with exceptionally large or small diffuse x - ray extensions ( e.g. , ngc 4342 , fig .
[ fig : imagea]ck , cl ) .
we fit each vertical profile with an exponential function plus a linear function with a slope to characterize the background .
to minimize effects due to the absorption and/or enhanced emission in the galactic disk , our fit excludes an interval that corresponds to a significant valley and/or peak ( if there is any ) in the profile at the galactic major axis ( shown as the inner box in fig .
[ fig : imagea ] for some galaxies , if presented ) .
the best fits to the profiles of all the galaxies are shown in fig .
[ fig : profilesa ] , while the fitted scale heights ( @xmath581 ) are listed in table [ table : scaleheight ] .
such a simple vertical profile extraction and fitting are _ not _ aimed to specify the spatial distribution of the coronae ( in fact , the unresolved stellar contribution can be important in some bulge - dominated galaxies , [ paperisubsec : spec ] ) , but instead , to roughly characterize their extensions , which will chiefly be used for a uniform definition of the spectral analysis regions .
lcc ic2560 & @xmath582 & @xmath583 + m82 & @xmath584 & @xmath585 + ngc24 & @xmath586 & @xmath587 + ngc520 & @xmath588 & @xmath589 + ngc660 & @xmath590 & @xmath591 + ngc891 & @xmath592 & @xmath593 + ngc1023 & @xmath594 & @xmath595 + ngc1380 & @xmath596 & @xmath597 + ngc1386 & @xmath598 & @xmath599 + ngc1482 & @xmath600 & @xmath601 + ngc1808 & @xmath602 & @xmath589 + ngc2787 & @xmath603 & @xmath604 + ngc2841 & @xmath605 & @xmath606 + ngc3079 & @xmath607 & @xmath608 + ngc3115 & @xmath609 & @xmath610 + ngc3198 & @xmath611 & @xmath612 + ngc3384 & @xmath613 & @xmath614 + ngc3412 & @xmath615 & @xmath616 + ngc3521 & @xmath617 & @xmath618 + ngc3556 & @xmath619 & @xmath507 + ngc3628 & @xmath620 & @xmath621 + ngc3877 & @xmath606 & @xmath338 + ngc3955 & @xmath622 & @xmath623 + ngc3957 & @xmath624 & @xmath625 + ngc4013 & @xmath626 & @xmath627 + ngc4111 & @xmath588 & @xmath628 + ngc4217 & @xmath511 & @xmath629 + ngc4244 & @xmath630 & @xmath631 + ngc4251 & @xmath632 & @xmath633 + ngc4342 & @xmath634 & @xmath635 + ngc4388 & @xmath532 & @xmath636 + ngc4438 & @xmath637 & @xmath638 + ngc4501 & @xmath639 & @xmath640 + ngc4526 & @xmath641 & @xmath642 + ngc4565 & @xmath643 & @xmath644 + ngc4569 & @xmath645 & @xmath646 + ngc4594 & @xmath647 & @xmath648 + ngc4631 & @xmath649 & @xmath650 + ngc4666 & @xmath651 & @xmath652 + ngc4710 & @xmath653 & @xmath654 + ngc5102 & @xmath655 & @xmath656 + ngc5170 & @xmath657 & @xmath658 + ngc5253 & @xmath659 & @xmath660 + ngc5422 & @xmath661 & @xmath662 + ngc5746 & @xmath663 & @xmath664 + ngc5775 & @xmath665 & @xmath666 + ngc5866 & @xmath667 & @xmath589 + ngc6503 & @xmath668 & @xmath669 + ngc6764 & @xmath670 & @xmath671 + ngc7090 & @xmath638 & @xmath672 + ngc7457 & @xmath673 & @xmath674 + ngc7582 & @xmath675 & @xmath676 + ngc7814 & @xmath677 & @xmath678 [ table : scaleheight ] we extract the spectrum of a galaxy from the region with the same horizontal extension and disk exclusion as those used for extracting its vertical surface brightness profile ( [ paperisubsec : spatial ] ) , except for several individuals with distorted diffuse x - ray morphology ( for these galaxies , the regions are selected to cover the bulk of the diffuse x - ray emission , e.g. , ngc 520 and ngc 1386 ; fig .
[ fig : imagea ] ) .
the outermost vertical extension is taken to be @xmath679 , except for ngc 4342 ( @xmath680 to limit the region on the s3 chip ) and ngc 4244 ( @xmath681 is not well constrained due to the low counting statistics ; the region is taken to cover the bulk of the diffuse x - ray emission ) .
ccd edges and gaps are filtered out to minimize the uncertainty in the instrument spectral response , which may be affected by the dithering of the observations .
we limit the on - source and background regions to the s3 chip for acis - s observations and to the i0 - 3 chips for acis - i observations .
[ fig : imagea ] includes the outlines of the adopted spectral extraction regions , although they may be cut short by filtered ccd edges .
if different spectral regions are adopted , we correct for the derived luminosities by assuming the soft x - ray intensity profile can be characterized by the exponential model described in [ paperisubsec : spatial ] .
the regions typically include the most luminous part of the coronae , except for the residual of some disk absorption or emission features ( fig .
[ fig : profilesa ] ) and some very extended extraplanar features ( e.g. , ngc 2841 ; fig .
[ fig : imagea]ba , bb ) .
these residual features typically only account for a small fraction of the coronal emission and will be individually noted in [ paperisec : individual ] .
generally , they can not seriously affect our statistical comparisons of the global coronal properties within such a large sample .
we extract the on - corona ( source ) and background spectra , as well as the corresponding response files using the ciao tool _
`` specextract '' _ , which is officially recommended for extended sources .
the _ `` specextract '' _ script automatically runs a set of ciao tools .
in particular , it has the ability to determine when _ `` mkacisrmf '' _ should be used in place of _ `` mkrmf '' _ in building the rmf ( the response matrix file ) , based on the gain and cti correction adopted for the event file , as well as the type and focal plane temperature of the ccd .
the rmf and arf ( ancillary response files ) files are weighted using a detector weight map ( wmap ) in 0.3 - 2 kev , created for the source and background files , respectively .
the sky background can be complicated in some fields ( e.g. , @xcite ) .
therefore , we prefer a local background subtraction in most cases .
most of our sample galaxies are small enough to be covered by a single ccd chip , allowing for background regions selected within the same ccd , but still beyond @xmath679 .
for a few galaxies with very extended coronae ( filling the entire ccd , e.g. , m82 ; fig .
[ fig : imagea]ac , ad ) , the fov of the _ chandra _ observations is too small to allow for a local background subtraction . in such a case
, we use the blank - sky background files , which are reprojected and exposure adjusted to match the 10 - 12/10 - 14 kev ( for fi / bi ccds ) count rate in the source spectrum .
the coronal emission in such cases are strong .
the uncertainty in the background estimate should not cause significant problem . for each galaxy with good counting statistics ,
we rebin the background subtracted spectrum to achieve a minimum signal - to - noise ratio ( @xmath682 ) for spectral fitting with @xmath683 statistic . to model the spectra of the galactic coronae
, we need to subtract the residual contributions from discrete sources . in
_ x - ray faint _ galaxies , after removing the brightest x - ray point sources , the unresolved stellar contribution may still be non - negligible , or may even dominate the hot gas emission .
@xcite have successfully calibrated this individually faint stellar source component in m32 , which is not massive enough to hold a galactic corona , or by directly resolving them in the galactic ridge ( also see @xcite ) .
this component is mainly comprised of cataclysmic variables ( cvs ) and coronal active binaries ( abs ) , which individually have a luminosity of @xmath684 ( hereafter the cv+ab component ) . the cv+ab contribution can be characterized with a @xmath685 thermal plasma plus a @xmath686 power law @xcite .
the characterization of this cv+ab component ( or its x - ray spectral contribution ) scaled with the stellar mass enables us to quantitatively isolate the truly diffuse hot gas emission in x - ray faint , especially bulge - dominated galaxies ( e.g. , @xcite ) . as a result , we can
just simply include a fixed accumulated cv+ab component in the spectral modeling of an _ early - type _ disk galaxy .
exceptions are made for two individuals , ngc 520 and ngc 660 , which show highly distorted morphologies with warped disks @xcite ; the old stellar population and hence the cv+ab contribution are too uncertain to quantify .
in fact , the diffuse x - ray emission is strongly elongated along the galaxies minor axes , indicating that the stellar contribution is not important and has a significantly different distribution from their stellar light ( fig .
[ fig : imagea]ag - aj ) .
similarly , we also neglect the stellar contribution for late - type galaxies , which is expected to be small and to be largely absorbed by the gas - rich galactic disks .
in addition , we also include a residual contribution from the removed sources due to counts spilled out of the removal circles and from x - ray binaries below our detection limit for each galaxy .
this contribution is expected to be small because of the flat luminosity function of x - ray binaries at low luminosities ( e.g. , @xcite ) . as is usually done , the spectral shape of this component is assumed to be a simple power law , which may also approximately account for the possible residual instrumental and/or sky background .
this component is typically less important in soft x - ray , e.g. , usually more than an order of magnitude lower than the hot gas emission in the 0.5 - 1.5 kev band ( figs .
[ fig : speca ] and [ fig : spec2 t ] ) .
we fit a coronal spectrum with a 1-t thermal plasma model ( xspec model mekal or vmekal ; @xcite ) ( tables [ table : specmodel]-[table : specpara1tmetal ] ) .
elements are divided into two groups : o - like ( c , n , o , ne , mg , si , s , ar , ca ) and fe - like ( the other elements ) ; the respective abundances of o - like elements are fixed to be solar . for 14 spectra with relatively high counting statistics ,
we fit the fe / o abundance ratio ( see [ paperisubsubsec : thermalchemicalstates ] for details ) . for other spectra
, we simply set the ratio to be 0.3 for late - type galaxies @xcite and 1 for early - type ones @xcite .
all the spectral components are subject to the galactic foreground absorption ( wabs ; the @xmath687 values are listed in table [ table : sampleselectionpara ] ) .
although strong variation in the absorption column on scales of several tens of pc may be expected in superwinds , such an uncertainty in @xmath687 has very limited effects on the modeling of the accumulated spectra of the galactic coronae @xcite . for galaxies with particularly thick cool gas disks ( e.g. , ngc 5775 ; @xcite ) ,
there is also uncertainty in the residual x - ray absorption from the emission right behind the disks . for most of our sample galaxies ,
the spectral quality is not sufficiently high to allow for directly constraining @xmath687 . for galaxies with good counting statistics
, we fit the @xmath687 value and found it is typically not significantly different from the galactic foreground value .
we thus fix @xmath687 at the galactic value , assuming that the intrinsic absorption of the extraplanar emission is not significant , although it may be important for galaxies with other orientations @xcite .
we also test a 2-t model for several _ late - type _ galaxies to characterize the temperature range of the coronae ( [ paperisec : introduction ] ) , while the abundances are fixed for simplicity . for 16 galaxies ,
such a modeling shows evidence for the presence of distinct different temperature components .
we list the parameters of the two thermal components in table [ table : specpara2 t ] and show the best - fit spectra in fig .
[ fig : spec2 t ] .
we emphasize that with the limited quality , the spectral data themselves are not sufficient to justify a specific decomposition of various components , which is selected to represent the physical plausibility ( e.g. , the presence of optically - thin plasma ) and the best calibration ( e.g. , the cv+ab contribution ) .
we also fix as many parameters as possible ( e.g. , the cv+ab component , and in some cases the abundances ) .
but when a component ( e.g. , the power law component used to characterize the residual background and stellar contributions ) is hard to quantify with reasonable certainty , we choose to fit it so that its uncertainty can be propagated into the estimation of other spectral parameters . however , in almost all cases , we find that the power law component is less important ( than the thermal and/or the cv+ab component ) in soft x - ray and thus has very limited effect on the modeling of the galactic coronae . typically , each spectral model fit contains a total of only 4 - 6 free model parameters ( i.e. , the power law index and normalization , one or two temperatures plus the corresponding normalizations of the thermal plasma ) , which can be usefully constrained with reasonably good counting statistics of a spectrum ( table [ table : specmodel ] ) . for a corona with inadequate counting statistics for the above spectral fit
, we estimate the luminosity simply from the background - subtracted 0.5 - 1.5 kev count rate detected in the same region as would be used for a spectral extraction .
this conversion assumes a thermal plasma of a typical 0.3 kev temperature .
the other parameters ( abundances and the cv+ab component ) are adopted in the same way as those used in the spectral fit cases .
but the power law component is omitted , because its contribution is typically negligible , compared to the uncertainty in the luminosity estimate .
we compare the results ( the directly measured hot gas properties , such as luminosity , temperature , abundance ratio , and emission measure ; table [ table:1tspec ] ) to those from previous analyses of the same _ chandra _ data ( 33 galaxies ; table [ table : sampleselectionpara ] ) .
most of the results are consistent with each other within the quoted errors or with suitable corrections ( e.g. , for the differences in the spectral extraction regions and/or the subtraction of stellar sources ) .
several galaxies with significantly inconsistent results are individually noted in [ paperisec : individual ] .
the errors of the hot gas parameters as listed in tables [ table : specmodel]-[table : specpara2 t ] are statistical only ( i.e. , from the spectral fits ) .
systematic uncertainties ( e.g. , in the assumed spectral components and models ) , which are difficult to quantify , will be discussed in
[ paperisec : discussion ] , and will be reflected in the dispersions of subsequent correlation relations of the parameters ( paper ii ) . [
cols="^,^,^,^,^,^,^,^,^ " , ] lccccccc ic2560 & @xmath688 & 0.3 & @xmath689 & @xmath690 & @xmath691 & @xmath692 & @xmath693 + m82 & @xmath694 & @xmath695 & @xmath696 & @xmath697 & @xmath698 & @xmath699 & @xmath700 + ngc0024 & @xmath701 & 0.3 & @xmath702 & @xmath703 & @xmath704 & @xmath705 & @xmath706 + ngc0520 & @xmath707 & @xmath708 & @xmath709 & @xmath710 & @xmath711 & @xmath712 & @xmath713 + ngc0660 & @xmath714 & @xmath715 & @xmath716 & @xmath717 & @xmath718 & @xmath719 & @xmath720 + ngc0891 & @xmath721 & @xmath722 & @xmath723 & @xmath724 & @xmath725 & @xmath726 & @xmath727 + ngc1023 & @xmath728 & @xmath729 & @xmath730 & @xmath731 & @xmath732 & @xmath733 & @xmath734 + ngc1380 & @xmath735 & @xmath736 & @xmath737 & @xmath738 & @xmath739 & @xmath740 & @xmath741 + ngc1386 & @xmath742 & @xmath743 & @xmath744 & @xmath745 & @xmath746 & @xmath747 & @xmath748 + ngc1482 & @xmath749 & @xmath750 & @xmath751 & @xmath752 & @xmath753 & @xmath754 & @xmath755 + ngc1808 & @xmath756 & @xmath757 & @xmath758 & @xmath759 & @xmath760 & @xmath761 & @xmath762 + ngc2787 & @xmath763 & @xmath764 & @xmath765 & @xmath766 & @xmath767 & @xmath768 & @xmath769 + ngc2841 & @xmath770 & @xmath771 & @xmath772 & @xmath773 & @xmath774 & @xmath775 & @xmath776 + ngc3079 & @xmath777 & @xmath778 & @xmath779 & @xmath780 & @xmath781 & @xmath782 & @xmath783 + ngc3115 & @xmath784 & @xmath785 & @xmath786 & @xmath787 & @xmath788 & @xmath789 & @xmath790 + ngc3198 & @xmath791 & 0.3 & @xmath792 & @xmath793 & @xmath794 & @xmath795 & @xmath796 + ngc3384 & @xmath797 & 0.3 & @xmath798 & @xmath799 & @xmath800 & @xmath801 & @xmath802 + ngc3412 & @xmath803 & 0.3 & @xmath804 & @xmath805 & @xmath806 & @xmath807 & @xmath808 + ngc3521 & @xmath809 & @xmath810 & @xmath811 & @xmath812 & @xmath813 & @xmath814 & @xmath815 + ngc3556 & @xmath816 & @xmath736 & @xmath817 & @xmath818 & @xmath819 & @xmath820 & @xmath821 + ngc3628 & @xmath822 & @xmath823 & @xmath824 & @xmath825 & @xmath826 & @xmath827 & @xmath828 + ngc3877 & @xmath829 & @xmath830 & @xmath831 & @xmath832 & @xmath833 & @xmath834 & @xmath835 + ngc3955 & @xmath836 & @xmath837 & @xmath838 & @xmath839 & @xmath840 & @xmath841 & @xmath842 + ngc3957 & @xmath843 & 0.3 & @xmath844 & @xmath845 & @xmath846 & @xmath847 & @xmath848 + ngc4013 & @xmath849 & 0.3 & @xmath850 & @xmath851 & @xmath852 & @xmath853 & @xmath854 + ngc4111 & @xmath855 & @xmath856 & @xmath857 & @xmath858 & @xmath859 & @xmath860 & @xmath861 + ngc4217 & @xmath862 & 0.3 & @xmath863 & @xmath864 & @xmath865 & @xmath866 & @xmath867 + ngc4244 & @xmath868 & 0.3 & @xmath869 & @xmath870 & @xmath871 & @xmath872 & @xmath873 + ngc4251 & @xmath874 & 0.3 & @xmath875 & @xmath876 & @xmath877 & @xmath878 & @xmath879 + ngc4342 & @xmath880 & @xmath881 & @xmath882 & @xmath883 & @xmath884 & @xmath885 & @xmath886 + ngc4388 & @xmath887 & @xmath888 & @xmath889 & @xmath890 & @xmath891 & @xmath892 & @xmath893 + ngc4438 & @xmath894 & @xmath895 & @xmath896 & @xmath897 & @xmath898 & @xmath899 & @xmath900 + ngc4501 & @xmath901 & @xmath902 & @xmath903 & @xmath904 & @xmath905 & @xmath906 & @xmath907 + ngc4526 & @xmath908 & @xmath909 & @xmath910 & @xmath911 & @xmath912 & @xmath913 & @xmath914 + ngc4565 & @xmath915 & @xmath916 & @xmath917 & @xmath918 & @xmath919 & @xmath920 & @xmath921 + ngc4569 & @xmath922 & @xmath923 & @xmath924 & @xmath925 & @xmath926 & @xmath927 & @xmath928 + ngc4594 & @xmath929 & @xmath930 & @xmath931 & @xmath932 & @xmath933 & @xmath934 & @xmath935 + ngc4631 & @xmath936 & @xmath937 & @xmath938 & @xmath939 & @xmath940 & @xmath941 & @xmath942 + ngc4666 & @xmath943 & @xmath944 & @xmath945 & @xmath946 & @xmath947 & @xmath948 & @xmath949 + ngc4710 & @xmath950 & @xmath951 & @xmath952 & @xmath953 & @xmath954 & @xmath955 & @xmath956 + ngc5102 & @xmath957 & 0.3 & @xmath958 & @xmath959 & @xmath960 & @xmath961 & @xmath962 + ngc5170 & @xmath963 & 0.3 & @xmath964 & @xmath965 & @xmath966 & @xmath967 & @xmath968 + ngc5253 & @xmath969 & @xmath970 & @xmath971 & @xmath972 & @xmath973 & @xmath974 & @xmath975 + ngc5422 & @xmath976 & 0.3 & @xmath977 & @xmath978 & @xmath979 & @xmath980 & @xmath981 + ngc5746 & @xmath982 & @xmath983 & @xmath984 & @xmath985 & @xmath986 & @xmath987 & @xmath988 + ngc5775 & @xmath989 & @xmath990 & @xmath991 & @xmath992 & @xmath993 & @xmath994 & @xmath995 + ngc5866 & @xmath996 & @xmath997 & @xmath998 & @xmath999 & @xmath1000 & @xmath1001 & @xmath1002 + ngc6503 & @xmath1003 & @xmath1004 & @xmath1005 & @xmath1006 & @xmath1007 & @xmath1008 & @xmath1009 + ngc6764 & @xmath1010 & @xmath1011 & @xmath1012 & @xmath1013 & @xmath1014 & @xmath1015 & @xmath1016 + ngc7090 & @xmath1017 & @xmath1018 & @xmath1019 & @xmath1020 & @xmath1021 & @xmath1022 & @xmath1023 + ngc7457 & @xmath1024 & 0.3 & @xmath1025 & @xmath1026 & @xmath1027 & @xmath1028 & @xmath1029 + ngc7582 & @xmath1030 & @xmath1031 & @xmath1032 & @xmath1033 & @xmath1034 & @xmath1035 & @xmath1036 + ngc7814 & @xmath1037 & 0.3 & @xmath1038 & @xmath1039 & @xmath1040 & @xmath1041 & @xmath1042 [ table:1tspec ] lcccccccc m82 & @xmath1043 & @xmath1044 & @xmath1045 & @xmath1046 & @xmath1047 & @xmath1048 & @xmath1049 & @xmath1050 + ngc891 & @xmath1051 & @xmath823 & @xmath1052 & @xmath1053 & @xmath1054 & @xmath1055 & @xmath1056 & @xmath1057 + ngc1482 & @xmath1058 & @xmath1059 & @xmath1060 & @xmath1061 & @xmath1062 & @xmath1063 & @xmath1064 & @xmath1065 + ngc1808 & @xmath1066 & @xmath1067 & @xmath1068 & @xmath1069 & @xmath1070 & @xmath1071 & @xmath1072 & @xmath1073 + ngc2841 & @xmath1074 & @xmath1075 & @xmath1076 & @xmath1077 & @xmath1078 & @xmath1079 & @xmath1080 & @xmath1081 + ngc3079 & @xmath1082 & @xmath1067 & @xmath1083 & @xmath1084 & @xmath1085 & @xmath1086 & @xmath1087 & @xmath1088 + ngc3556 & @xmath1089 & @xmath1090 & @xmath1091 & @xmath1092 & @xmath1093 & @xmath1094 & @xmath1095 & @xmath1096 + ngc3628 & @xmath1097 & @xmath1098 & @xmath1099 & @xmath1100 & @xmath1101 & @xmath1102 & @xmath1103 & @xmath1104 + ngc4342 & @xmath1105 & @xmath1106 & @xmath1107 & @xmath1108 & @xmath1109 & @xmath1110 & @xmath1111 & @xmath1112 + ngc4526 & @xmath1113 & @xmath1114 & @xmath1115 & @xmath1116 & @xmath1117 & @xmath1118 & @xmath1119 & @xmath1120 + ngc4594 & @xmath1121 & @xmath930 & @xmath1122 & @xmath1123 & @xmath1124 & @xmath1125 & @xmath1126 & @xmath1127 + ngc4631 & @xmath1128 & @xmath1129 & @xmath1130 & @xmath1131 & @xmath1132 & @xmath1133 & @xmath1134 & @xmath1135 + ngc5253 & @xmath1136 & @xmath1137 & @xmath1138 & @xmath1139 & @xmath1140 & @xmath1141 & @xmath1142 & @xmath1143 + ngc5866 & @xmath1144 & @xmath1145 & @xmath1146 & @xmath1147 & @xmath1148 & @xmath1149 & @xmath1150 & @xmath1151 [ table : specpara1tmetal ] lcccccccc ngc520 & @xmath1152 & @xmath1153 & @xmath1154 & @xmath1155 & @xmath1156 & @xmath1157 & @xmath1158 & @xmath1159 + ngc2841 & @xmath1160 & @xmath1161 & @xmath1162 & @xmath1163 & @xmath1164 & @xmath1165 & @xmath1166 & @xmath1167 + ngc3079 & @xmath1168 & @xmath1169 & @xmath1170 & @xmath1171 & @xmath1172 & @xmath1173 & @xmath1174 & @xmath778 + ngc3521 & @xmath1175 & @xmath1176 & @xmath1177 & @xmath1178 & @xmath1179 & @xmath1180 & @xmath1181 & @xmath1182 + ngc3556 & @xmath1183 & @xmath1184 & @xmath1185 & @xmath1186 & @xmath1187 & @xmath1188 & @xmath1189 & @xmath1190 + ngc3628 & @xmath1191 & @xmath1192 & @xmath1193 & @xmath1194 & @xmath1195 & @xmath1196 & @xmath1197 & @xmath1198 + ngc3877 & @xmath1199 & @xmath1200 & @xmath1201 & @xmath1202 & @xmath1203 & @xmath1204 & @xmath1205 & @xmath1206 + ngc4388 & @xmath1207 & @xmath1208 & @xmath1209 & @xmath1210 & @xmath1211 & @xmath1212 & @xmath1213 & @xmath1214 + ngc4501 & @xmath1215 & @xmath1216 & @xmath1217 & @xmath1218 & @xmath1219 & @xmath1220 & @xmath1221 & @xmath1222 + ngc4565 & @xmath1223 & @xmath1224 & @xmath1225 & @xmath1226 & @xmath1227 & @xmath1228 & @xmath1229 & @xmath1230 + ngc4569 & @xmath1231 & @xmath1232 & @xmath1233 & @xmath1234 & @xmath1235 & @xmath1236 & @xmath1237 & @xmath1238 + ngc4631 & @xmath1239 & @xmath1240 & @xmath1241 & @xmath1242 & @xmath1243 & @xmath1244 & @xmath1245 & @xmath1246 + ngc5253 & @xmath1247 & @xmath1248 & @xmath1249 & @xmath1250 & @xmath1251 & @xmath1252 & @xmath1253 & @xmath916 + ngc5775 & @xmath1254 & @xmath1255 & @xmath1256 & @xmath1257 & @xmath1258 & @xmath1259 & @xmath1260 & @xmath1261 + ngc6764 & @xmath1262 & @xmath1263 & @xmath1264 & @xmath1265 & @xmath1266 & @xmath1267 & @xmath1268 & @xmath1269 + ngc7582 & @xmath1270 & @xmath1271 & @xmath1272 & @xmath1273 & @xmath1274 & @xmath1275 & @xmath1276 & @xmath1214 [ table : specpara2 t ]
we herein describe multi - wavelength properties of individual galaxies , concentrating on those that show distinct properties and/or are not well studied in literatures . _
m82 _ : as a member of the m81 group , m82 is known to be surrounded by large - scale structures of atomic gas , produced by tidal interaction @xcite .
the rotation curve of m82 declines outward without a significant flattening part @xcite . the @xmath562 value of @xmath1277 obtained from _
hyperleda _ is likely measured at a projected radius of @xmath1278 and may be strongly affected by the tidal interaction .
indeed , the rotation curve measured with the stellar absorption lines peaks at @xmath1279 with a peak velocity of @xmath1280 ( as adopted in the present work ; table [ table : sampleselectionpara ] ) , where the tidal effect should be negligible @xcite .
the surrounding gas may also help to confine the outflow and hence enhance the x - ray emission , as manifested most vividly by the presence of an h@xmath556 and x - ray ridge , @xmath1281 or 11.6 kpc north of the galactic disk @xcite .
_ ngc 520 _ : this system represents an ongoing merger between two galaxies : one gas - rich and the other gas - poor @xcite . only the nucleus and its surrounding region of the gas - rich one are bright in x - ray . the diffuse emission , in particular ,
is elongated in the direction perpendicular to the gas - rich disk @xcite ( also see fig . [
fig : imagea]ag , ah ) .
we thus adjust the spectrum extraction region to cover the bulk of this dominant diffuse x - ray feature ( fig .
[ fig : imagea]ag )
. both the number of x - ray bright point sources and the diffuse x - ray emission , relative to its sfr , are smaller than other merging galaxies , such as the mice and antennae galaxy systems .
these distinct x - ray properties are suggested to be a result of a merger between gas - rich and gas - poor galaxies , instead of between gas - rich and gas - rich ones @xcite .
_ ngc 660 _ : this galaxy has a severely tidally - disturbed disk surrounded by a gas - rich polar ring .
the stellar population of the polar ring is young , with a considerable fraction consisting of blue and red supergiants @xcite .
although continuous sf is evidenced in the polar ring , it shows no significant soft x - ray enhancement in the short _ chandra _ observations ( table [ table : chandradata ] ; fig .
[ fig : imagea]ai ) .
bright co emission exhibits a clear central concentration , similar to those seen in nuclear starburst galaxies @xcite .
polarized extraplanar optical emission is attributed to the light scattered by dust grains residing up to @xmath1282 from the stellar disk @xcite , which are likely entrained in an energetic outflow , as revealed by the corona elongated along the minor axis of the galactic disk ( fig .
[ fig : imagea]ai , aj ) . _ ngc 1380 _ : this is a lenticular galaxy @xmath1283 from the center of the fornax cluster .
the measured temperature of @xmath1284 is significantly lower than that of the surrounding intra - cluster medium ( icm ) @xcite .
therefore , the coronal gas likely arises from the galaxy itself .
the temperature obtained from the _ chandra _ observation is lower than the _ rosat _ value of @xmath1285 @xcite , which is likely due to contamination from point sources or the icm .
the coronal emission is bright and extended toward southeast , with a blob of diffuse x - ray emission located nearby .
this very extended x - ray - emitting blob is centered on a small spiral galaxy , which is , however , not detected as a compact x - ray source ( a bright point - like x - ray source cxo j033639.4 - 345847 is indeed detected @xmath1279 off the center and is apparently not related ; fig . [
fig : imagea]ao ) .
if the blob is associated with the spiral galaxy and if its distance is the same as that of ngc 1380 , the x - ray to k - band luminosity ratio , @xmath1286 , would then be much higher than typical isolated galaxies .
in addition , if the diffuse x - ray is dominated by thermal emission from hot gas , the hardness ratio of the blob suggests a temperature of @xmath1287 , likely higher than that of ngc 1380 ( @xmath1288 , table [ table:1tspec ] ) .
we thus speculate that the blob is more likely to be a background galaxy group , if not a local enhancement of the fornax cluster happening to coincide with a background galaxy .
_ ngc 1386 _ : this galaxy is also a member of the fornax cluster . it hosts a low luminosity seyfert-2 agn with fe k - shell lines detected in x - ray @xcite .
the diffuse soft x - ray emission is faint , extended , and apparently highly distorted , especially in the outer part ( fig .
[ fig : imagea]aq , ar ; but the distorted outermost contour in fig .
[ fig : imagea]ar is close to the background level , may or may not relate to the galaxy itself , thus is not included in the spectral analysis ) .
analysis of the diffuse x - ray emission extracted from the box in fig .
[ fig : imagea]aq shows that the thermal component has a temperature of @xmath1289 , significantly lower than the icm value @xcite .
therefore , the bulk of the hot gas in the galaxy vicinity is likely associated with the galaxy itself and reshaped in the cluster environment .
_ ngc 2787 _ : this is a low mass s0 galaxy rich in atomic gas , although the molecular gas mass is relatively low @xcite .
most of the atomic gas is located in a ring - like structure with a diameter of @xmath1290 @xcite , significantly larger than the extent of the stellar light ( fig .
[ fig : imagea]aw ) .
the unresolved soft x - ray emission shows a round morphology and has a spectrum consistent with primarily stellar in origin ( fig .
[ fig : speca]aj ; @xcite ) .
_ ngc 2841 _ : this is a massive spiral galaxy .
some plume - like diffuse x - ray emission features are apparently connected to the disk , some of which may form a limb - brightened `` x - shaped '' structure around the galaxy s nucleus , similar as those often found in nuclear starburst galaxies @xcite .
the coronal luminosity obtained here is significantly lower than the estimation using _ rosat _ observation @xcite , which is clearly due to the differences in removing point sources and in separating the disk and halo components .
_ ngc 3079 _ : this galaxy hosts a kpc - scale nuclear superbubble , whose diffuse x - ray emission is closely associated with various h@xmath556 filaments @xcite .
it is still not clear whether the nuclear starburst or the seyfert 2 agn @xcite is the dominant driving source of the bubble .
an observation of the galaxy and its two companions ( ngc 3073 and mcg 9 - 17 - 9 ) reveals that ngc 3073 exhibits an elongated tail remarkably aligned with the nucleus of ngc 3079 , and is probably shaped by the ram pressure stripping of the outflowing gas from ngc 3079 @xcite .
ngc 3079 is thought to have both nuclear - concentrated and disk - wide sfs and have the diffuse x - ray characteristics of both types of galaxies ( `` x - shaped '' and disk - wide extended ) @xcite
. _ ngc 3115 _ : this galaxy contains primarily an old stellar population , shows little on - going sf , and is extremely poor in cold gas @xcite .
the galaxy hosts a massive nuclear bh @xcite ; its bondi radius is readily resolved by _
chandra _ @xcite .
spectral analysis of the unresolved x - ray emission shows that it mainly consists of unresolved stellar sources , with little contribution from hot gas ( fig .
[ fig : speca]am ) .
previous detection of the @xmath1285 hot gas from short _ chandra _ observations @xcite is more likely due to the contamination of unresolved point sources , i.e. , the cv+ab component ( [ paperisubsubsec : spectraanalysis ] ) .
_ ngc 3384 _ : this galaxy is a member of the m96 group . its collision with m96 is thought to be the major triggering mechanism for the very large - scale , ring - like , star forming gaseous structure ( the `` leo ring '' , @xcite ) .
the galaxy has a very low rotation velocity ( @xmath1291 , table [ table : sampleselectionpara ] ) which may be highly affected by the interaction .
its @xmath4 is significantly higher than @xmath561 ( fig .
[ fig : galaxymass]b ) ; the latter may _ not _ be physical .
_ ngc 3628 _ : this galaxy is a member of the leo triplet , and has a huge optical and tidal tail ( @xmath1292 extension ) @xcite , which is the result of the interaction with its companions ngc 3627 and ngc 3623 .
clumpy sf regions are detected in this tidal tail , with a sf age of @xmath1293 @xcite .
apparently extended x - ray emission around this nuclear starburst galaxy was detected in _
einstein _ @xcite and _ rosat _ @xcite observations . in higher resolution _
images ( fig .
[ fig : imagea]bq , br ) , the diffuse x - ray emission is dominated by a vertically elongated component , which is associated with the h@xmath556 emission on various scales , probably representing a well collimated nuclear outflow @xcite .
_ ngc 4244 _ : this is a quiescent low - mass late - type galaxy , with very weak sf , no nuclear activity , and a small bulge .
even with a nearly perfect edge - on orientation ( @xmath1294 , table [ table : sampleselectionpara ] ) , the galaxy shows little extraplanar diffuse x - ray emission in a deep _ chandra _ observation ( @xmath1295 , table [ table : chandradata ] ; fig .
[ fig : imagea]cg , ch ; @xcite ) .
the detected emission comes from primarily some patches , with a total luminosity of @xmath1296 , putting the galaxy into the x - ray faintest group of the present sample .
_ ngc 4342 _ : this is a low - mass galaxy located in the virgo cluster . as a dwarf galaxy with little sf
, it appears to be unusually bright in x - ray and hosts an exceptionally massive nuclear bh @xcite .
it is suggested to reside in a massive dark matter halo .
the very extended and lopsided morphology of the diffuse x - ray emission , as well as the sharp northeastern boundary ( or a cold front ) ( fig .
[ fig : imagea]ck , cl ) , suggests that the galaxy is moving supersonically in the icm @xcite . _
ngc 4388 _ : this galaxy is also located in the virgo cluster .
it hosts a seyfert 2 nucleus @xcite .
the extended diffuse soft x - ray emission is correlated well with the ionization cone found in optical - line emission @xcite .
the relatively high temperature ( @xmath1297 ; table [ table:1tspec ] ) indicates that the gas represents a mixture of the stripped ism and the surrounding icm @xcite .
_ ngc 4438 _ : it belongs to an interacting galaxy pair in the virgo cluster @xcite and is much more x - ray luminous than galaxies with similar sfr or stellar mass .
in addition , the diffuse x - ray emission of ngc 4438 seems to connect both members of the pair ( fig .
[ fig : imagea]co , cp ) and is apparently correlated with many cool gas features ( @xcite ) .
the sf and agn are not particularly active in the galaxy , so the very luminous and asymmetric x - ray emission is likely produced and shaped by other mechanisms , such as cool - hot gas interaction , ram - pressure stripping , and turbulent mixing ( e.g. , @xcite ) .
_ ngc 4565 _ : this galaxy is inactive in sf @xcite and hosts a weak seyfert 2 agn @xcite .
the diffuse soft x - ray emission consists of three components ( fig .
[ fig : imagea]cu , cv ) with distinct morphologies : 1 ) a round one apparently associated with the small stellar bulge , 2 ) an elongated one along the disk , and 3 ) an lopsided one extending from the disk toward northeast .
the very asymmetric , large - scale , diffuse x - ray morphology of this third component is probably due to the ram pressure that the galaxy is subject to against the ambient igm @xcite .
_ ngc 4569 _ : this is one of the largest and most deficient galaxies in the virgo cluster @xcite .
it shows strong evidence for ongoing icm - ism interaction .
line synthesis observations have showed that the galaxy has lost more than 90% of its gas @xcite .
the diffuse soft x - ray emission extends to the west ( fig .
[ fig : imagea]cw , cx ) , coinciding with a giant and h@xmath556 arm , which likely represents the gas stripped from the disk about 300 myr ago @xcite .
_ ngc 4594 _ : this , often known as the `` sombrero '' galaxy , is a massive , bulge - dominated sa galaxy .
it contains a quiescent nucleus and little sf , but is very luminous in x - ray ( @xmath1298 , table [ table:1tspec ] ) .
quantitative calculations indicate that the observed energy , mass , and metal contents of the corona are much less than expected from stellar feedback @xcite . _
ngc 4631 _ : strong starburst is evidenced in this galaxy , apparently triggered by interactions with neighboring galaxies .
multi - wavelength observations @xcite reveal that the diffuse x - ray morphology resembles the radio halo , indicating a close connection between the outflows of hot gas , cosmic ray , and magnetic field from the galactic disk . enhanced diffuse x - ray emission is also apparently enclosed by numerous h@xmath556-emitting loops blistered out from the central disk of the galaxy .
_ ngc 4666 _ : this galaxy hosts a compton - thick agn and intense sf @xcite , which are thought to be caused by gravitational interaction with the neighboring galaxy ngc 4668 @xcite . the galaxy is in many respects very similar to ngc 4631 , showing a strong galactic superwind as revealed by multi - wavelength observations @xcite .
observation has a too short exposure ( @xmath1299 , table [ table : chandradata ] ) to resolve the structure of the superwind , but still shows prominent vertical - elongated diffuse x - ray emission ( fig .
[ fig : imagea]de , df ) . _ ngc 5170 _ : this is a massive quiescent spiral galaxy , with a small stellar bulge and a perfect edge - on orientation ( table [ table : sampleselectionpara ] ) , thus good for testing the accretion scenario of the corona formation ( e.g. , @xcite ) .
@xcite analyzed the _
chandra _ data , but failed to detect any significant extraplanar diffuse x - ray emission ( most of the x - ray emission in fig .
[ fig : imagea]dk , dl can be attributed to the residual contribution from point sources ) .
the upper limit is consistent with our measurement , accounting for the different spectral regions adopted in the two analyses .
_ ngc 5253 _ : in our sample , this galaxy has the highest specific sfr ( @xmath1300 ; table [ table : sampleselectionpara ] ) @xcite , compared to m82 , which has @xmath1301 .
however , ngc 5253 appears to be relatively x - ray faint and has a less vertically extended corona than those of typical starburst galaxies ( fig .
[ fig : imagea]dm , dn ) , probably because of its shallow gravitational potential . _ ngc 5746 _ : this galaxy , similar as ngc 5170 , is also a massive quiescent spiral galaxy , but has a slightly smaller inclination angle and a larger stellar bulge ( table [ table : sampleselectionpara ] ) .
previous analysis of the _ chandra _ data also failed to detect the extraplanar hot gas @xcite .
there is some apparently diffuse x - ray emission in the nuclear region , most of which may be nonthermal in origin ( fig .
[ fig : speca]bh ) .
_ ngc 5775 _ : this galaxy is interacting with its companion ngc 5774 ( and probably also with another small companion ic 1070 ) , which probably triggers the active sf over the disk .
observation shows that gas transfer from ngc 5774 to ngc 5775 is ongoing through two `` bridges '' @xcite .
the diffuse x - ray emission shows characteristics of both nuclear starburst and disk - wide sf , i.e. , with both vertically and horizontally elongated components ( fig .
[ fig : imagea]ds , dt ) .
a less smoothed x - ray image shows a 10 kpc - diameter shell - like feature in the southern halo , but the current x - ray data ( lack of spectral information on the `` shell '' ) can not be used to say much about its nature @xcite . _ ngc 5866 _ : this galaxy resembles ngc 3115 in stellar mass , morphological type , stellar population , and environment ( tables [ table : sampleselectionpara ] , [ table : basicparaii ] ) , but is much richer in cool gas and significantly more x - ray luminous @xcite
. detailed spectral analysis shows a prominent fe l - shell line feature , but no significant oxygen line , suggesting a high fe / o abundance ratio @xcite .
optical observation reveals various extraplanar dusty filaments , which are suggested to be cool gas outflows produced by type ia sne .
_ ngc 6764 _ : this galaxy hosts both agn and nuclear starburst .
the two kpc - scale radio bubbles spatially coincide with the diffuse x - ray emission with a high fitted temperature ( @xmath1302 ) @xcite .
in addition , the vertical profile peaks sharply toward the galactic disk ( fig . [ fig : profilesa]ca ) . it is therefore most likely that the hot gas is heated by the agn jet / ism interactions . _ ngc 7090 _ : this is a bulge - less , low - mass galaxy with a patchy dust lane . the sf is weak and scattered over the disk , but the radio continuum emission is very extended and asymmetric @xcite .
an h@xmath556 image reveals a faint halo in the form of knots , and morphologically coincident with the extended radio continuum emission @xcite . both the gas distribution and the kinematics are distorted .
but there is no companion galaxy that might have caused these distortions @xcite .
the diffuse x - ray emission is weak and concentrated only in the central region .
a large fraction of the emission can be attributed to the residual emission from point sources ( figs .
[ fig : imagea]ec , ed , [ fig : speca]bm ; @xcite ) .
_ ngc 7582 _ : this galaxy hosts a prominent dust lane being spatially coincident with the excess of x - ray absorption .
two `` hotspots '' with highly ionized oxygen and neon line emissions ( stronger and emission than and ) are detected in soft x - ray , and are thought to be ionized by the leaking photons from the agn instead of from the disk sf regions @xcite .
in addition to the directly fitted parameters of a galactic corona from the spectral analysis ( luminosity , temperature , and emission measure ) ( [ paperisubsec : spec ] ) , we further derive several potentially useful parameters : the electron number density ( @xmath1303 ) , total mass ( @xmath1304 ) , thermal energy ( @xmath1305 ) , and radiative cooling timescale ( @xmath1306 ) ( table [ table:1tspec ] ) .
the estimation of these parameters depends on the uncertain spatial distribution of the x - ray - emitting gas .
we parameterize this uncertainty into an unknown effective volume filling factor @xmath1307 of the gas ( e.g. , @xmath1308 if the distribution is uniform ) . in high - resolution optical and x - ray observations of starburst galaxies ( e.g. , ngc 253 , @xcite ) ,
the soft x - ray emission is often observed to be spatially correlated with h@xmath556 emitting filaments .
the hot gas is thus apparently produced at the interface between sn - driven winds and the cooler ambient ism instead of from the wind material itself @xcite .
therefore , at least in such starburst galaxies , the volume filling factor of the x - ray emitting gas is likely far below unity and can only be poorly constrained in few galaxies with deep and high - resolution x - ray observations ( e.g. , @xcite ) .
the filling factor represents a large uncertainty in the estimation of the derived hot gas parameters .
we obtain the average electron number density and the total hot gas mass from the emission measure , assuming a cylinder with its radius and height as defined for the spectral extraction . using the mass and the average temperature
, we further estimate the total thermal energy contained in the volume .
we finally use the extinction - corrected luminosity from the best - fit plasma model and the total thermal energy to infer the radiative cooling timescale of the corona .
a few systematic corrections are needed to obtain uniform estimates of the parameters for the coronae .
_ first _ , the fields excluded from the removal of the sources and ccd edges are sometimes non - negligible .
for each of the galaxies , we correct for the missed luminosity by the ratio of the areas before and after the removal .
_ second _ , we correct for the loss of the coronal luminosity due to the filtered out `` disk '' range , by extrapolating the vertical intensity profile into the disk , using the best - fit exponential model as obtained in [ paperisubsec : spatial ] .
the correction is then the ratio of the accumulated model fluxes , with and without the extrapolation .
these two corrections are merged to obtain a total correction factor for the coronal luminosity ( @xmath1309 ) .
similarly , we obtain the corresponding correction factors for the derived parameters , based on the fitted intensity profiles .
we further extrapolate @xmath1303 and @xmath1306 into the disk , assuming constant temperature and metal abundances . in this case , the soft x - ray intensity is assumed to be proportional to @xmath1310 .
@xmath1304 and @xmath1305 are calculated in the same volume as that for the luminosity calculation .
all these correction factors are listed in table [ table : correctionfactor ] . except for the systematical corrections discussed above
, all the directly measured or derived parameters may be affected by other factors .
therefore , their physical interpretation may not be straightforward .
here we discuss the measurement uncertainties and the physical meanings of the derived parameters ( @xmath1303 , @xmath1304 , @xmath1305 , @xmath1306 ) .
our results will be based primarily on the directly measured parameters @xmath3 and @xmath1311 , which will be more thoroughly discussed in the next subsection ( [ paperisubsec : contaminations ] ) .
_ density _ :
for each galaxy , the coronal gas density , @xmath1303 , is inferred from the emission measure and further interpolated into the galactic disk following the soft x - ray intensity profile .
physically , the emission measure anti - correlates with the absolute abundance which is poorly constrained in low - resolution soft x - ray spectroscopy .
therefore , the absolute value of @xmath1303 is meaningful only if both the absolute abundance and the volume filling factor @xmath1307 can be well constrained .
fortunately , except for a few extreme cases ( e.g. , @xcite ) , most of the galaxies are not observed to have very super- or sub - solar abundances ( e.g. , @xcite ) .
the assumption of solar metallicity of oxygen ( [ paperisubsec : spec ] ) then only has limited effects on the measured @xmath1303 .
furthermore , due to the possible existence of higher temperature hot gas @xcite , even if a pressure balance is reached , as often assumed in the estimation of @xmath1303 ( e.g. , @xcite ) , the soft x - ray - emitting gas of a galactic corona is still expected to have a volume filling factor far below unity @xcite .
therefore , @xmath1303 in table [ table:1tspec ] could be adopted as an lower limit to the real hot gas density ( so we have included @xmath1307 in the table ) . _ hot gas mass and thermal energy _ : also due to the unknown volume filling factor , @xmath1304 and @xmath1305 as listed in table [ table:1tspec ] are only upper limits to the real coronal mass and thermal energy . in general , the coronal gas only takes a small fraction of the total galactic baryonic mass , typically @xmath1312 ( tables [ table : basicparaii ] and [ table:1tspec ] ) , even for the most x - ray luminous galaxies .
furthermore , much of the feedback material may be stored in a higher temperature phase . in hydrodynamical simulations of galactic scale sn feedbacks
, the majority of the feedback energy from either starburst cores or sf inactive galactic bulges is supposed to be stored in a volume - filling hot ( @xmath1313 ) phase of ism , which triggers the fast galactic wind and transport large amount of energy ( but relatively small amount of mass ) out of the host galaxy @xcite . deep _
chandra _ and _ xmm - newton _
observations do have revealed such a high - temperature component through the detection of @xmath13146.7 kev fek@xmath556 emission line in the starburst core of the nearby superwind galaxy m82 @xcite
. however , in most of the cases , the properties of this high - temperature component is still poorly constraint observationally because of its low density and hence low emissivity ( e.g. , @xcite and references therein ) .
this means we have detected only a small fraction of the feedback material ( the `` missing galactic feedback problem '' , see @xcite and references therein ) , which prevents us from quantitatively understanding the galactic feedback .
_ radiative cooling timescale _ : there are two possible fates of the coronal gas , either blown out into the intergalactic space or fall back to the galactic disk .
the radiative cooling timescale ( @xmath1306 ) and the dynamical timescale of the global gas flows describe the relative importance of these two possibilities .
the speed of the galactic outflow , if there is any , can not be directly measured with the existing x - ray data . however , in simulations of hot gas outflows around starburst galaxies @xcite or low / intermediate - mass galactic spheroids @xcite , the outflow velocity can often reach several @xmath1315 or even @xmath1316 .
this velocity corresponds to a dynamical timescale of @xmath1317 , assuming a corona size of @xmath1318 .
although @xmath1306 is proportional to @xmath1319 , it is _ unlikely _ to be globally smaller than this dynamical timescale ( the typical value of @xmath1306 is @xmath1320 , table [ table:1tspec ] ) .
therefore , in starburst and/or low / intermediate - mass galaxies ( most of our sample galaxies ) , outflow likely plays a dominant role .
nevertheless , the hot gas entrained in the outflow often has a broad velocity range . even in the most energetic cases ,
a significant fraction of the hot gas may not have sufficiently high velocities ( typically a few @xmath1321 ) to overcome the galaxy s gravitational potential @xcite .
this gas tends to interact with the surrounding cool ism and lost its energy radiatively , largely responsible for the observed soft x - ray emission .
it is thus still possible that a significant fraction of the coronal gas may cool down within @xmath1306 and fall back to the galactic disk .
lccccc ic2560 & 1.736 & 2.985 & 1.568 & 0.335 + m82 & 1.707 & 2.698 & 1.507 & 0.371 + ngc24 & 1.004 & 1.777 & 1.004 & 0.563 + ngc520 & 1.018 & 2.244 & 1.018 & 0.446 + ngc660 & 1.006 & 2.244 & 1.006 & 0.446 + ngc891 & 1.707 & 2.748 & 1.368 & 0.364 + ngc1023 & 1.017 & 2.244 & 1.017 & 0.446 + ngc1380 & 1.030 & 2.244 & 1.030 & 0.446 + ngc1386 & 1.010 & 2.244 & 1.010 & 0.446 + ngc1482 & 2.019 & 2.965 & 1.483 & 0.337 + ngc1808 & 2.834 & 3.280 & 1.734 & 0.305 + ngc2787 & 1.015 & 2.244 & 1.015 & 0.446 + ngc2841 & 1.021 & 2.244 & 1.021 & 0.446 + ngc3079 & 1.304 & 2.415 & 1.128 & 0.414 + ngc3115 & 1.062 & 2.244 & 1.062 & 0.446 + ngc3198 & 1.071 & 2.244 & 1.071 & 0.446 + ngc3384 & 1.757 & 2.591 & 1.242 & 0.386 + ngc3412 & 1.003 & 2.244 & 1.003 & 0.446 + ngc3521 & 1.300 & 2.244 & 1.300 & 0.446 + ngc3556 & 2.281 & 2.672 & 1.456 & 0.374 + ngc3628 & 1.583 & 2.510 & 1.488 & 0.398 + ngc3877 & 3.566 & 3.447 & 2.004 & 0.290 + ngc3955 & 1.007 & 2.244 & 1.007 & 0.446 + ngc3957 & 1.005 & 2.244 & 1.005 & 0.446 + ngc4013 & 1.608 & 2.692 & 1.305 & 0.371 + ngc4111 & 1.785 & 2.816 & 1.378 & 0.355 + ngc4217 & 3.733 & 3.204 & 1.879 & 0.312 + ngc4244 & 1.003 & 2.244 & 1.003 & 0.446 + ngc4251 & 1.008 & 2.244 & 1.008 & 0.446 + ngc4342 & 1.021 & 1.777 & 1.021 & 0.563 + ngc4388 & 1.651 & 2.717 & 1.312 & 0.368 + ngc4438 & 1.571 & 2.432 & 1.343 & 0.411 + ngc4501 & 4.301 & 3.263 & 2.090 & 0.306 + ngc4526 & 2.109 & 2.958 & 1.495 & 0.338 + ngc4565 & 1.233 & 2.244 & 1.233 & 0.446 + ngc4569 & 2.044 & 2.631 & 1.347 & 0.380 + ngc4594 & 1.897 & 2.540 & 1.636 & 0.394
+ ngc4631 & 1.965 & 2.689 & 1.595 & 0.372 + ngc4666 & 3.214 & 3.082 & 1.830 & 0.324 + ngc4710 & 1.010 & 2.244 & 1.010 & 0.446 + ngc5102 & 1.036 & 2.244 & 1.036 & 0.446 + ngc5170 & 1.318 & 2.244 & 1.318 & 0.446 + ngc5253 & 1.683 & 2.671 & 1.293 & 0.374 + ngc5422 & 1.006 & 2.244 & 1.006 & 0.446 + ngc5746 & 1.337 & 2.244 & 1.337 & 0.446 + ngc5775 & 2.792 & 3.210 & 1.908 & 0.312 + ngc5866 & 1.525 & 2.637 & 1.267 & 0.379 + ngc6503 & 1.007 & 2.244 & 1.007 & 0.446 + ngc6764 & 7.976 & 4.084 & 2.762 & 0.245 + ngc7090 & 1.010 & 2.244 & 1.010 & 0.446 + ngc7457 & 1.002 & 2.244 & 1.002 & 0.446 + ngc7582 & 1.625 & 2.549 & 1.284 & 0.392 + ngc7814 & 4.906 & 3.748 & 2.225 & 0.267 [ table : correctionfactor ] except for the systematical corrections already adopted in
[ paperisubsec : derivedparameters ] , there are still some other uncertainties which can hardly be quantified . we discuss in this subsection how they may affect our results .
the multi - phase nature of the hot gas in nearby galaxies is suggested by both observations ( e.g. , @xcite ) and simulations ( e.g. , @xcite ) .
however , the temperature distribution of this multi - phase gas is usually robustly characterized with the low quality x - ray spectra ( e.g. , @xcite ) .
the fitted temperatures ( often with a simple 1-t thermal plasma model ) are thought to be significantly biased from the commonly used definitions of average temperature ( e.g. , luminosity - weighted or emission - measure - weighted temperature ; @xcite ) .
therefore , we need to test how such possible bias may affect our results .
the multi - temperature thermal plasma is most likely produced by the interaction between the sn - driven superwind and the ambient cool ism @xcite .
we thus test a 2-t plasma model for late - type galaxies ; early - type ones are typically poor in cool gas and are expected to have a relatively simple thermal state .
[ fig : spec2 t ] presents the spectra of 16 galaxies for which the 2-t plasma model gives significantly different results from the 1-t model , while table [ table : specpara2 t ] lists the fitted parameters . in fig .
[ fig : test2t]a , we compare the _ luminosities _ of the thermal emission obtained from the 1-t ( @xmath3 ) and 2-t ( @xmath1322 ) models . while the two luminosities show a tight correlation , @xmath1322 is systematically larger than @xmath3 , because the power law component tends to be suppressed in a 2-t fit , especially at high energies ( fig .
[ fig : spec2 t ] ) . nevertheless , this systematical difference is only about 25% , typically comparable to or even less than the measurement uncertainty in the luminosity of a corona .
therefore , even if the 2-t plasma model represents a better description of the coronal thermal state , the luminosity obtained from the 1-t fit is still a good approximation of the coronal emission .
we also check the adequacy of the 1-t model in characterizing the _ thermal state _ of a corona .
we define the luminosity weighted temperature of the 2-t model as : @xmath1323 , where subscripts @xmath1324 and @xmath1325 refer to the low - t and high - t components . as shown in fig .
[ fig : test2t]b , @xmath1326 is generally consistent with @xmath1311 , showing no significant bias within the errors .
the large errors of @xmath1326 for some galaxies are due to the uncertainty in the decomposition of the two thermal components .
therefore , the temperature obtained from the 1-t model typically gives an adequate characterization of the mean thermal state of a corona with the limited quality of the data .
the galactic coronae , and even the igm , are thought to be metal enriched by various types of stellar ejecta ( e.g. , massive stellar wind , planetary nebula , and sn ejecta ) .
therefore , the metal abundance pattern of the coronal gas can be affected by many factors , such as the competition of different metal sources and how different phases of gas mix with each other ( e.g. , @xcite ) .
different abundance ratios ( e.g. , fe / o ) have been observed in different types of galaxies .
some of the ratios are significantly different from the solar values , indicating recent enrichment ( e.g. , @xcite ) .
however , the absolute metal abundance of the coronal gas is typically poorly constrained because the continuum , dominated by the bremsstrahlung of free elections from hydrogens , can hardly be detected in soft x - ray at the low resolution of a typical ccd spectrum ; grating spectra , as can be obtained for some relatively compact bright galaxies with deep _ xmm - newton _
rgs observations , can not help much , because the continuum emission from point - like sources can not be easily subtracted ( e.g. , @xcite ) .
the fe / o ratio ( fe and o are often employed as the most important products of ia and cc sne , representing the feedbacks from old and young stellar populations ) can be readily measured ; but its interpretation can still be complicated .
for example , @xcite show that the uncertainty in the thermal state ( e.g. , the need for a second temperature component to represent the temperature range of the hot gas ) can significantly affect the measured iron abundance ( the fe - bias ) , but little on the oxygen abundance . in a typical 1-t model ,
this effect can increase the uncertainty of the measured fe / o value by as large as @xmath1327 .
all the above uncertainties in the metal abundance measurement may strongly affect the estimation of other coronal properties ( e.g. , @xcite ) .
we then check how the assumption of fe / o ratio in the 1-t model ( [ paperisubsubsec : spectraanalysis ] ) affects the measurements of the luminosity and temperature .
[ fig : testmetal ] shows that the luminosity and temperature with or without the fe / o ratio fixed are generally consistent with each other .
the only exception is ngc 5866 . for this galaxy ,
the temperature fitted with the abundance ratio fixed is substantially greater than that if it is allowed to vary .
indeed , the galaxy has been reported in a detailed study to have a strong fe l line bump while no detectable oxygen lines @xcite .
more sensitive spectroscopic observations are needed to understand such an unusual abundance pattern in this cool - gas - rich s0 galaxy .
in addition to the uncertainty in the thermal and chemical states , a probably more important factor in the measurement of hot gas properties arises from the residual stellar contribution .
the model of the stellar component ( cv+ab , as well as the power law accounting for both the unresolved lmxb and the residual background ) has both thermal and power law components with different relative contributions in different galaxies ( [ paperisubsec : spec ] ) .
in addition , the specific contribution of faint stellar sources in soft x - ray ( per stellar mass ) may also vary in different types of galaxies ( e.g. , due to different sf history ) .
this uncertainty can be as large as @xmath1328 @xcite .
therefore , for galaxies in which a large fraction of the `` diffuse '' x - ray emission comes from unresolved stellar sources , the hot gas properties obtained from the spectral analysis can have large systematic errors ( e.g. , for ngc 3115 , [ paperisec : individual ] , @xcite ) .
our focus here is at characterizing the average properties of the coronae ( luminosity , average temperature , and fe / o ratio ) , instead of studying the specific thermal and chemical states ( temperature distribution and absolute abundance of different elements ) .
therefore , we believe that a simplified 1-t model with fixed abundance ratio plus stellar contributions is an optimal choice for a unform , statistical analysis of the galactic coronae
. our measurement may also depend on the inclination angle of a galaxy .
the diverse of inclination of the present sample may affect the volume estimate , the filtering of the galactic disk , and so the correction of the measurement .
we have checked the dependence of the measured hot gas properties on the inclination angle ( @xmath563 ) to investigate any possible inclination bias . as an example ( the most significant dependence on @xmath563 ) , fig .
[ fig : test_incli]a shows that @xmath1303 seems to be slightly correlated with @xmath563 ; but the correlation is only marginal ( @xmath1329 ) and probably dominated by only a few galaxies with high @xmath1303 values .
[ fig : test_incli]b further indicates that the difference between highly and moderately inclined galaxies ( divided by @xmath1330 ) may be real , but a simple geometric correction can not account for this difference ( fig .
[ fig : test_incli]c ) .
this possible systematical bias , if not dominated by several extreme cases , could be attributed to the contamination from outer disk emission in moderately - inclined galaxies ( [ paperisubsec : spatial ] , [ paperisubsec : spec ] ) .
nevertheless , such potential inclination biases are generally small , compared to the measurement uncertainty itself , as well as uncertainties caused by the intrinsic diversities in the coronal morphologies ( e.g. , the lopsidedness ) or physical / chemical states .
therefore , we have made _ no _ attempt to systematically correct for any inclination effects on our measurements .
we will instead compare high and moderate inclined galaxies in the subsequent statistical analysis ( paper ii ) .
we further consider the potential effects of agn on the _ coronal _ properties .
agn may affect our analysis in two different ways , either via the scattered light of the central x - ray bright point source , or via various forms of feedback ( ejection or heating of the surrounding gas caused by the radiation , winds and/or jets , @xcite ) . as presented in
[ paperisec : sampleselection ] , none of our sample galaxy has nuclear source bright enough that its scattered light ( wings of psf and ccd readout steak ) may dominate the truly diffuse coronal x - ray emission .
in fact , we benefit from the large inclination angle in the sample selection that most of the x - ray emission from agn may have already been blocked by the cool gas in the disk . on the other hand , various forms of feedback from the agn may affect the galaxy environment significantly ( e.g. , @xcite and references therein ) , so potentially affect our x - ray analysis of the galactic coronae .
generally , there are two major modes of agn feedback ( e.g. , @xcite ) .
the _ radiative mode _
( also known as the quasar or wind mode ) mainly operates when the accreting black hole is close to the eddington limit .
although it is sometimes suggested that the radiation from agn could be responsible for some extended x - ray features even as far as several kpc from the nucleus , by the photo - ionization and fluorescence of the gas ( e.g. , @xcite ) , this mode typically has the most significant effect on the dense cool gas in the inner part of a galaxy .
little evidence is currently revealed for the direct interaction between these relatively homogenous wind and the large scale hot gas .
we thus neglect it from the direct effects on the corona .
the other mode is known as the _ kinetic mode _ ( or jet / radio mode ) .
there are many observational evidence ( e.g. , x - ray jet and filaments ) for the direct interaction between the jet and the hot halo ( e.g. , @xcite ) .
in addition , many agn blowout superbubbles are also detected in the hot surrounding medium of _ massive _ galaxies located in the center of cool core clusters , which plays an important role in shaping the circumgalactic environment ( e.g. , @xcite ) .
there are also some lines of evidence that the diffuse x - ray morphology is correlated with the radio / x - ray luminosities of agns in normal elliptical galaxies , which may indicate that the coronal gas is disturbed by agn and thus departs from hydrostatic equilibrium with the gravitational potential @xcite .
such a radio mode agn feedback may even be energetic enough to unbind a significant fraction of the igm in galaxy groups @xcite , so explain the breaking of self - similarity observed in the scaling relations between the gravitational mass and some hot gas properties of massive clusters of galaxies ( e.g. , @xcite ) .
in spite of these lines of observational evidence , which are mostly for massive elliptical galaxies , we do not find much evidence for prominent agn feedback features in our sample of disk galaxies .
there are several exceptions ( [ paperisec : individual ] ) , such as the @xmath1314kpc scale limb - brightened nuclear bubble of ngc 3079 @xcite , the small nuclear bubble of ngc 4438 @xcite , the coincidence between the radio bubble and diffuse x - ray emission of ngc 6764 @xcite , etc . in most cases , such structures do not significantly affect our measurements of the global properties of the relevant individual coronae .
furthermore , agn feedback is often _ energetically _ less important ( than sn feedback ) in galaxies with small or intermediate mass bulges ( e.g. , @xcite ) , as is the case for most of the present sample galaxies ( refer to [ paperisec : sampleselection ] for details ) .
we thus believe that agn feedback only have limited _ direct _ effect on the coronal x - ray emission .
in addition to the _ direct _ effect on the coronal x - ray emission , agn feedback could also have strong effects on various galaxy properties , which may then affect the coronae _
indirectly_. for example , the intense flux of photons and particles produced by the agn may sweep the cool gas in the galaxy , or at least keep the gas hot and reduce radiative cooling , therefore quench the sf ( e.g. , @xcite ) .
strong agn feedback can also reduce the accretion and stifle the fuel supply onto the supermassive bh , then terminates the nuclear activity ( e.g. , @xcite ) .
all these changing of galaxy properties ( e.g. , sf and nuclear activity ) may have potential effects on the coronal properties , e.g. , by changing the stellar feedback rate or the global accretion rate from the igm ( e.g. , @xcite and references therein ) .
nevertheless , here we only consider the direct interaction between agn and the coronal gas ( such as heating , mass injection , and compression , etc . ; e.g. , @xcite ) , while leaving the discussions of the scatters caused by various indirect effects to the correlation analysis between the coronal properties and various galaxy properties ( paper ii )
. it may be helpful to statistically compare the coronal x - ray luminosity and temperature measurements of the galaxies with or without agn ( fig .
[ fig : testagn ] ; the classification of the galaxies plotted in the figure is based on @xcite ) .
we find that the median values of @xmath3 and @xmath1311 for these two subsamples do not show significant differences ( @xmath1331 higher in @xmath3 for galaxies with agns and no noticeable difference in @xmath1311 ) .
in contrast , starburst and non - starburst galaxies are indeed distinguishable statistically ( different by 0.7 dex in @xmath3 and 0.16 kev in @xmath1311 ) .
we thus conclude that _ agns are generally less important than sf in regulating the coronal properties of our sample galaxies_. there are growing lines of evidence that the cxe ( also called charge transfer ) makes a non - negligible contribution to the x - ray line emission in various sorts of astrophysical objects ( see @xcite for a recent review ) .
the cxe describes a process in which a highly ionized ion ( e.g. , ) takes one or more electrons from another atom ( like h , @xmath1332 , or he ) , which can be represented by : @xmath1333 .
such a process will produce an excited ion @xmath1334 , which emits x - ray photons as it decays to the ground state .
extensive studies have been carried out on the solar wind cxe and its contribution to the x - ray emission of some solar system objects , as well as to the galactic soft x - ray background ( e.g. , @xcite ) .
indeed , the cxe has further been speculated to be important in many other types of astrophysical objects ( e.g. , @xcite ) . according to the diffuse x - ray emission from nearby galaxies
, it has been shown that a substantial fraction of it can not arise from optically - thin thermal plasma , but may originate in the cxe ( e.g. , @xcite ) .
this scenario is also consistent with the spatial correlation between some cool / hot gas features , especially in active sf galaxies ( e.g. , @xcite ) .
existence of the cxe will not only affect the spectral modeling of specific line emissions , but more important to this work , also potentially significantly enhance the global soft x - ray emission .
the cxe produces only line emission , which can not be well separated from the thermal emission in a low resolution ccd spectrum . for example , the relative strengthes of the forbidden ( @xmath1335 ) and inter - combination ( @xmath1336 ) lines to the resonance line ( @xmath1337 ) of the k@xmath556 triplet is often adopted as a spectral diagnostic for the presence of the cxe @xcite . while the cxe is closely related to the presence of cool gas , the amount of cool gas can not be used as a direct diagnostic parameter of the cxe strength .
this is because the amount of cool gas is often tightly correlated with the sfr ( e.g. , fig . [
fig : sftracer]b ) , or other cool - hot gas interaction processes , such as mass loading and turbulent mixing , which all enhance radiative cooling ( e.g. , @xcite ) .
nevertheless , since all these processes are more or less related to sf , we may expect that some enhanced cool / hot gas features , especially in low - density environments ( where the thermal emission is relatively weak ) , may originate in the cxe .
a few apparent candidates for significant cxe contributions present in our sample .
the analysis of the _ xmm - newton _ rgs spectra of m82 has shown that the contributions of the cxe are about 90% , 50% , and 30% to the , , and triplets , respectively @xcite .
although the grating spectral analysis is focused chiefly on the central region ( typically @xmath1338 from the nucleus ) , it is still expected that a considerable fraction of the coronal soft x - ray line emission may arise from the cxe . in particular , m82 is located within a complex of gas , formed apparently from the tidal interaction among members of the m81 group @xcite ; a large number of interfaces , generated by expected strong interaction between the gas and the galactic superwind , may lead to strong cxe emission .
similarly , for the disk - wide starburst galaxy , ngc 4631 , the measured line ratio is also consistent with a cxe origin @xcite .
another possible example of a significant cxe contribution is ngc 4438 .
this galaxy is unusually x - ray bright ( @xmath1339 , compared to @xmath1340 of m82 and @xmath1298 of ngc 4594 ; table [ table:1tspec ] ) for its moderate stellar mass ( @xmath1341 , compared to @xmath1342 of ngc 4594 ) and sfr ( @xmath1343 , compared to @xmath1344 of m82 ) .
it is located in the virgo cluster and shows tidal interaction with a companion galaxy ngc 4435 ( [ paperisec : individual ] ) .
in addition , the cool / hot gas features are also spatially correlated @xcite . in contrast , galaxies undergoing tidal interaction and cool gas transfer from companions , but located in non - cluster environments ( e.g. , ngc 5775 ) , do not have such well - matched cool / hot gas features ( e.g. , @xcite ) .
we thus speculate that the enhanced x - ray emission may at least partly arise from the cxe between the stripped cool gas and the hot icm . due to the scarcity of high - quality , high - resolution spectroscopic observations and the lack of suitable spectral models @xcite , the contribution of cxe to the total diffuse x - ray luminosity
can hardly be constrained in individual galaxies or in a sample . however , at least in some cases , this component appears to be non - negligible .
we have conducted a systematical data reduction and analysis of the _ chandra _ data of 53 nearby disk galaxies , of which 20 are presented for the first time to study the _ diffuse _ soft x - ray emission .
this sample covers broad ranges of sfr ( @xmath1345 ) , stellar mass ( @xmath1346 ) , and other galaxy properties .
the sample size is much larger than those of previous studies of similar types of galaxies ( typically @xmath1347 galaxies ) .
we have conducted uniform data calibration of all the galaxies .
in particular , discrete point - like sources are detected and removed , while significant contributions from unresolved stellar sources are carefully estimated for different types of galaxies , allowing for a spectral analysis of the truly diffuse hot gas emission . based on these calibrations ,
we have provided measurements of various diffuse soft x - ray properties of the galactic coronae , including both the directly measured parameters from the spatial and spectral analysis ( scale height , luminosity , and temperature ) , and some derived parameters ( density , mass , thermal energy , and radiative cooling timescale ) .
notable properties are highlighted for individual galaxies .
our products ( images , spectra , brightness profiles , and measured hot gas parameters ) form a comprehensive database for the study of the galactic coronae .
we have also discussed various corrections and complications in the measurements . in particular , we have investigated the effects caused by uncertainties in the thermal and chemical states of the hot gas . we find that a simple 1-t model with fixed abundance ratio plus stellar contributions is typically sufficient to characterize the overall properties of the coronal emission .
the different inclination angles of our sample galaxies may cause a small systematical bias due to the potential contamination from the disk emission / absorption in moderately inclined galaxies .
but this bias can not play a significant role in determining the coronal properties .
several galaxies in our sample host agns ; some of them may even be responsible for distinct diffuse x - ray features .
but in general , agn feedback is less important than sf feedback for our sample of disk galaxies ; the presence of the diffuse x - ray features related to agn does not significantly affect our measurements of the global coronal properties .
the cxe may significantly contribute to the soft x - ray emission in some of the galaxies , either active in sf or not .
in particular , the unusually high x - ray luminosity of the tidally disturbed gas - rich virgo cluster member ngc 4438 is most likely due to the cxe .
further insights into the origin of the coronae can be obtained from a statistical analysis of the measurements , together with other galaxy parameters , which is the topic of paper ii .
we thank the referee for his / her constructive comments and suggestions , which are very helpful in the revision of the paper .
this work is supported by nasa through the cxc / sao grant ar0 - 11011b , and the adap grant nnx12ae78 g . | we present a systematical analysis of the _ chandra _ observations of 53 nearby highly - inclined ( @xmath0 ) disk galaxies to study the coronae around them .
this sample covers a broad range of galaxy properties : e.g. , about three orders of magnitude in the sfr and more than two orders of magnitude in the stellar mass .
the _ chandra _
observations of the _ diffuse _ soft x - ray emission from 20 of these galaxies are presented for the first time .
the data are reduced in a uniform manner , including the excision / subtraction of both resolved and unresolved stellar contributions .
various coronal properties , such as the scale height and luminosity , are characterized for all the sample galaxies . for galaxies with high enough counting statistics
, we also examine the thermal and chemical states of the coronal gas .
we note on galaxies with distinct multi - wavelength characteristics which may affect the coronal properties .
the uniformly processed images , spectra , and brightness profiles , as well as the inferred hot gas parameters , form a large x - ray database for studying the coronae around nearby disk galaxies .
we also discuss various complications which may cause biases to this database and their possible corrections or effects , such as the uncertainty in the thermal and chemical states of hot gas , the different galactic disk inclination angles , the presence of agn , and the contribution of the emission from charge exchange at interfaces between hot and cool gases .
results from a detailed correlation analysis are presented in a companion paper , to gain a more comprehensive statistical understanding of the origin of galactic coronae . |
in this appendix we rationalize a close qualitative connection between classical dynamics of multidimensional motion in a potential with forbidden islands and a geodesic flow in a compact manifold with non - trivial topology . for the sake of simplicity
we consider the case of two coordinates .
the motion is restricted to a disk - like region punctured by forbidden islands .
boundaries of the accessible region correspond to the lines where the total energy coincides with the potential energy . to employ the original argument of sinai
@xcite we glue the accessible region of the configuration space to its antipode replica along the boundary . a reflection from the boundary
can be thought of as continuing motion in the other component .
the resulting surface is compact and its topology is characterized by genus @xmath39 ( the number of handles attached to a sphere ) corresponding to the number of original islands .
configuration space reflecting on its boundaries ( left ) .
reflections can be viewed as transitions to the antipode surface component glued to the original one along the boundaries ( right ) .
trajectories on the deformed smooth version of the resulting compact surface are shown as solid and dashed lines , which correspond to the original and antipode components , respectively .
[ islands ] , width=268 ] in fig . [ islands ]
we show both the original billiard - like motion in the configuration space and the equivalent motion in two components assembled together that form a compact surface of genus @xmath134 .
as mentioned in the main text trajectories of a particle moving in a potential resemble the geodesic lines in curved space with the metric @xmath38 .
the regions of the configuration space that correspond to negative curvature work as a defocusing lenses , causing instability ( divergence of close trajectories ) . for example , close trajectories diverge every time they approach the boundary of the classically inaccessible island or pass a region of a potential local maximum that belongs to the accessible region . passing through the stable regions with positive curvature can not compensate for the instability , since stability reflects oscillatory , rather than converging features in the dynamics of the trajectory deviation .
in particular , existence of unstable regions combined with ergodicity ensures exponential divergence of trajectories over long enough times . according to the gauss - bonnet theorem , in the @xmath40 case
the average curvature is negative , which implies regions of instability .
response functions can be obtained starting with the driven liouville equation @xmath135 .
the phase space distribution @xmath136 is represented as a functional series in @xmath1 with the initial distribution @xmath137 being the zero - order contribution . in our case of interest
the observable coincides with the dipole moment @xmath13 that represents coupling to the probe field .
the linear and second - order classical response functions are expressed via the poisson brackets as @xmath138 we start with utilizing the fact that the equilibrium distribution @xmath137 depends on energy @xmath139 only , and make a transformation based on the fdt .
this brings us to eq .
( [ second - order ] ) .
the problem of free motion in a compact surface of constant negative curvature possesses dynamical @xmath58 symmetry .
ds means that the flow is determined by the @xmath57 generator @xmath140 , it also conserves the stable and unstable directions ; this is expressed by the relation @xmath141 .
any phase - space distribution can be decomposed into a direct sum of irreducible representations ; only principal series representations of @xmath58 contribute to the linear and second - order response @xcite .
ds implies that the distributions in different representations evolve independently .
distributions in irreducible representation can be implemented as functions in a circle with the basis of angular harmonics @xmath65 .
the evolution in the circle is determined by eq .
( [ equation - on - circle ] ) that , given the initial condition @xmath142 , can be solved using the method of characteristics : @xmath143 implementing the correspondence of natural scalar products in @xmath144 and in the circle , we find the expansion coefficients in eq .
( [ evolution - expansion ] ) @xmath145 these coefficients represent the main _ dynamical ingredient _ in the response function calculation .
we represent them in terms of the hypergeometric functions .
alternative representations have been derived in the context of two - point correlations@xcite .
the linear response function is determined by the dynamical part alone : @xmath146 calculation of the second - order response function is more involved . substituting the poisson brackets ( [ poisson - bracket ] ) into eq .
( [ first - order0 ] ) the @xmath0 signal adopts a form @xmath147 all terms in parentheses are further expanded in irreducible representations and angular harmonics .
the resulting contributions include two coefficients @xmath148 and a triple product given by eq .
( [ abqrs ] ) .
the latter is the _
geometrical ingredient _ that appears in nonlinear response functions .
the simplest expression for the second - order response function that incorporates different collective resonances , originates from two terms in the decomposition @xmath149 . then the response function reads @xmath150 \times \\ & \nonumber \biggl[t_2\frac{\partial}{\partial t_2}-n\biggr ] ( a_{n}^*(t_2;p)a_{n+1}(t_1;r ) ) \\ &
\nonumber -\biggl[\biggl(n+\frac{1}{2}-p\biggr)a^{pqr}_n-\biggl(n+\frac{1}{2}-r\biggr)a^{pqr}_{n+1}\biggr ] \times \\ & \nonumber \biggl[t_2\frac{\partial}{\partial t_2}+n+1\biggr ] ( a_{n+1}^*(t_2;p)a_{n}(t_1;r ) ) \biggr\ } \,.\end{aligned}\ ] ] it turns out that the resulting second - order response function may be represented as a double expansion over rp resonances .
the expansion is nontrivial : one encounters diverging series if the expansion of quantities @xmath151 is performed first .
fortunately , we have found an efficient scheme for numerical evaluation of the series@xcite .
the procedure also demonstrates that the second - order response is indeed expanded in the rp resonances .
in this appendix we present some details on @xmath0 spectra that characterize the second - order response of a chaotic dynamical system considered in this letter . experimental data on @xmath0 time - domain spectroscopy that probes the response function @xmath152
is usually presented using the so - called @xmath0 spectra that constitute a numerical @xmath0 fourier transform of the response function with respect to @xmath153 and @xmath154 .
the absolute value of the @xmath0 spectra for our dynamical system are given in fig .
[ 2dfourier ] , whereas in fig .
[ realimaginary ] we present the real and imaginary parts .
spectra : ( a ) real part , ( b ) imaginary part with the single resonance @xmath127 ; ( c ) real part , ( d ) imaginary part of linear combination of terms with two different resonances @xmath128 and @xmath129 .
[ realimaginary ] , width=244 ] as follows from the general picture presented in the main text , for a given unperturbed dynamics the response function depends on the expansion of the dipole function @xmath13 in the eigenmodes of the laplace operator : each eigenmode provides with an oscillation in the signal .
since the dipole is generally a smooth function of the system coordinates , the number of eigenmodes that participate in the expansion is typically small . to study the important features of the signals we consider the cases of one and two oscillation in the signal . in the first case
we see a diagonal peak with a stretched feature along @xmath131 direction ; such stretching never occurs due to a periodic orbit in an integrable system even if it is coupled to a harmonic bath . in the second case
we see both diagonal and cross peaks accompanied by the stretched features . | we consider classical response in a strongly chaotic ( mixing ) system .
as opposed to the case of stable dynamics , the nonlinear classical response in a chaotic system vanishes at large times .
the physical behavior of the nonlinear response is attributed to the exponential time dependence of the stability matrix .
the response also reveals certain features of collective resonances which do not correspond to any periodic classical trajectories .
we calculate analytically linear and second - order response in a simple chaotic system and argue on the relevance of the model for interpretation of spectroscopic data .
time - domain femtosecond spectroscopy constitutes a powerful tool that probes electronic and vibrational coherent dynamics of complex molecular systems in condensed phase @xcite .
spectroscopic signals are directly related to optical response functions that carry detailed information on the underlying dynamical phenomena . at room temperatures the complexity often originates from slow strongly anharmonic vibrational modes that can be treated within the framework of classical mechanics .
a number of studies have been devoted to the classical response in stable ( integrable ) dynamical systems @xcite .
the nonlinear response functions have been shown diverge linearly with time , while the divergence can be eliminated by invoking a fully quantum description @xcite .
although integrable dynamics represents only an approximation for realistic physical situations , weak deviations from the integrability do not eliminate the unphysical behavior of the response functions .
this follows from the fact that the divergence is related to quasiperiodic motions on invariant tori @xcite . according to the kolmogorov - arnold - moser ( kam ) theory ,
most the invariant tori are not destroyed by small perturbations that break down integrability @xcite . yet
, stable dynamics would be typical for a close to equilibrium situation . at larger energies a generic situation would correspond to dynamical behavior with chaotic features @xcite .
moreover , unstable ( hyperbolic ) dynamics may be more common due to the stability of chaos with respect to perturbations . it has been argued based on results of numerical analysis @xcite that chaotic dynamics appears to observe the convergence of the classical response functions . in spite of apparent importance , and to the best of our knowledge the problem of the nonlinear response in strongly chaotic systems has never been addressed using analytical methods .
we note that analytical calculations of the response are rarely feasible in nonintegrable systems , whereas numerical simulations of chaotic dynamics are complicated by the exponential divergence of stability matrices @xcite . in this letter
we show that ( i ) the classical response of the chaotic system exhibits decay and oscillations as a function of times between probing pulses , and ( ii ) the fourier transform of @xmath0 second order response function reveals broad and asymmetric peaks as signatures of chaos .
this is the main result or our study .
a strongly chaotic ( mixing ) system is characterized by a special spectrum of the liouville operator @xcite .
the spectrum consists of complex ruelle - pollicott ( rp ) resonances that determine the asymptotic oscillations and decay of the correlations .
this yields the linear response function directly related to the two - point correlation functions , in agreement with the fluctuation dissipation theorem ( fdt ) .
nonlinear response , however , turns out to be more involved still with the noticeable effect of the resonances . a system driven by a time - dependent external field @xmath1
is described by the hamiltonian @xmath2 with the the function @xmath3 in phase space representing the dipole ( see , e.g. ref . ) .
the response functions @xmath4 that depend on @xmath5 time intervals describe the expansion of the measured signal @xmath6 in a functional series in @xmath7 .
the second - order response function reads @xmath8 where @xmath9 is the equilibrium distribution and @xmath10 is the liouville operator of the unperturbed system . a schematic picture of second - order response formation in a chaotic system that employs the liouville phase space evolution is shown in fig .
[ fettuccine ] .
is presented by two regions @xmath11 and @xmath12 ( dark red and light green ) . as time
elapses the distribution elongates along unstable direction and contracts along stable one .
[ fettuccine ] , width=268,height=192 ] propagating the observable @xmath13 in eq .
( [ second - order ] ) backward in time we interpret the second - order response as an overlap of the distributions @xmath14 and @xmath15 with @xmath16 and the vector field @xmath17
. since @xmath18 , we can for simplicity represent the function @xmath13 by two separate regions in the phase space where it adopts positive and negative values . in the case of hyperbolic dynamics evolution during time @xmath19 changes the regions shapes . for @xmath20 ,
the shape becomes similar to ribbon - like fettuccine : elongated along the unstable direction by a factor @xmath21 , narrowed along the stable one @xmath22 ( with @xmath23 being the lyapunov exponent ) and unchanged along the flow . since @xmath24 results from reverse dynamics , the distributions @xmath25 and @xmath24 are elongated along the unstable and stable directions , respectively
therefore , the overlap is represented by a large set of @xmath26 small disconnected regions with the volume @xmath27 . since @xmath13 is represented by a positive and negative regions ,
the distribution @xmath24 consists of two positively and negatively `` charged '' fettuccine and the cancellations result in a signal determined by a typical fluctuation proportional to @xmath28 , and the overlap integral attains a factor @xmath29 that turns out to be exponentially small .
yet , this is not the end of the story since the derivative @xmath30 of a sharp feature along the stable direction can create exponentially large @xmath31 factors .
this is the liouville space signature of the exponentially growing components of the stability matrix , which affects the response starting with second order due to fdt@xcite .
however , the divergent terms cancel out : decomposing @xmath32 into the direction along the flow , and unstable and stable components , we note that only the last term is potentially dangerous . calculating the dangerous component of the overlap integral by parts we arrive at two contributions : the overlap of @xmath33 with @xmath25 , and the overlap of @xmath24 with @xmath25 weighted with @xmath34 .
the first contribution contains an additional @xmath35 factor , the second one provides an additional factor independent of time .
this results in a physical @xmath36 large time asymptotics of the nonlinear response function .
trajectories of a particle in arbitrary potential @xmath37 are known to be the same as for a free motion in a curved space with the metric @xmath38 @xcite .
although the potential generates nonuniform motion along the trajectories , one can expect chaotic behavior due to the exponentially growing separation between the trajectories .
in addition , when the motion is finite , the accessible part of the configurational space at a given energy can be multiply connected . in the simplest case of two coordinates the motion occurs inside a disk - like region punctured by @xmath39 forbidden islands .
some fraction of trajectories approaches the boundaries so close that this can be qualified as reflection . utilizing the original argument of sinai @xcite reflection
can be interpreted as continuing motion on the antipode replica of the accessible region glued to its original counterpart via the boundaries ( see appendix [ app : reflection ] and ref . ) .
the resulting compact surface has topology of a sphere with @xmath39 handles ( riemann surface of genus @xmath39 ) . in the @xmath40 case
the average curvature is negative , which results in unstable ( hyperbolic ) dynamics . to rationalize the described above qualitative picture of response in a chaotic system
we calculate the linear and second - order classical response functions for free motion in a riemann surface @xmath41 of constant negative ( gaussian ) curvature .
this model that allows an exact solution has been serving as a prototype of classical chaos , as well as an example of semiclassical quantization @xcite .
the classical free - particle hamiltonian @xmath42 depends on the absolute momentum value @xmath43 only , and @xmath44 includes the two coordinates and momentum direction angle @xmath45 .
therefore , the smooth compact @xmath46 manifold @xmath47 , which represents the subspace of phase space with fixed energy , is preserved in classical dynamics .
hereafter , we will use dimensionless units so that @xmath48 the curvature @xmath49 , and @xmath50 ( when energy is fixed ) .
our model allows for an exact solution due to strong dynamical symmetry ( ds ) . to describe
ds we adopt an agreement used in differential geometry by identifying a first - order differential operator of differentiating along the vector field with the vector field itself .
we denote by @xmath51 the vector field that describes the phase space dynamics , and set @xmath52 , @xmath53 $ ] .
a simple local calculation yields @xmath54=-\sigma_1 $ ] and @xmath55=-k\sigma_{z}$ ] which implies that in the constant negative curvature case the vector fields @xmath56 , @xmath51 , and @xmath53 $ ] form the lie algebra @xmath57 .
the group @xmath58 action in the reduced phase space @xmath47 is obtained by integrating of the @xmath57 algebra action .
ds with respect to the action of the group @xmath59 does not mean symmetry in a usual sense , i.e. that the system dynamics commutes with the group action , but rather reflects the fact that the poisson bracket is given by a bi - vector field @xmath60 in particular , the vector field that determines classical dynamics is given by an element @xmath51 of the corresponding lie algebra @xmath57 , whereas the stable and unstable directions are determined by @xmath61 .
ds implies that the space of phase - space distributions constitutes a representation of @xmath58 and , being decomposed into a sum of irreducible representations , provides a set of uncoupled evolutions .
unitary irreducible representation of @xmath58 are well - known and can be conveniently implemented in terms of functions @xmath62 in a circle @xcite .
principal series representations relevant for our calculations are labeled by imaginary parameter @xmath63 with the liouville operator @xmath64 .
the aforementioned decomposition identifies the angular harmonic @xmath65 in a circle with a phase - space distribution @xmath66 with given angular momentum @xmath67 .
commutation law in @xmath57 entails @xmath68 with two anti - hermitian conjugated ladder operators @xmath69 , @xmath70 .
any relevant distribution can be decomposed in the modes @xmath71 , where @xmath63 adopts discrete values ( the spectrum ) .
the spectrum @xmath72 can be related to the eigenvalues of the laplacian operator in @xmath41 by identifying @xmath73 and @xmath74 .
the dipole @xmath13 is a function in @xmath41 , i.e. a function in phase space that does not depend on both @xmath43 and momentum direction @xmath45 , hence @xmath75 , and @xmath13 can be expanded as a sum over the principal series representations @xmath76 .
the integrands in the response function ( [ second - order ] ) include the result of action of the evolution operator @xmath77 on the components @xmath78 of the dipole momentum and their derivatives resulting from the poisson brackets .
to simplify the calculations , we transform the expressions of response functions in a way that the evolution of derivatives of @xmath78 does not appear .
we use the representation in the circle to find the expansion of @xmath79 over basis vectors @xmath80 , @xmath81 since @xmath82 represents a function obtained as the result of the action of the evolution operator on @xmath83 , it can be found by solving the equation @xmath84 supplemented with initial condition @xmath85 . in the principal series representation with @xmath86 , @xmath87 , and
the equation takes the following form : @xmath88 the solution can be easily found , and thus after the integration @xmath89 we obtain @xmath90 : @xmath91,\end{aligned}\ ] ] where @xmath92 is the gauss hypergeometric function .
for details see appendix [ app : calculation ] and ref . .
the linear response function can be obviously expressed via @xmath93 .
the calculation involves only the first part of the poisson bracket , and the result is conveniently represented similarly to the fluctuation - dissipation relation @xmath94 @xcite .
for large @xmath19 the linear response function shows damped oscillations @xmath95 .
the expansion in powers of @xmath96 corresponds to rp resonances @xcite .
only even rp resonances contribute to the response function .
the expansion is a converging series in @xmath96 if @xmath97 , i.e. @xmath98 . since the dipole @xmath99 can decomposed in irreducible representations characterized by different values of @xmath63 , the linear response function constitutes a linear combination of contributions with coefficients @xmath100 .
the second - order response function can be calculated using eqs .
( [ second - order ] ) and ( [ poisson - bracket ] ) @xcite .
the calculation can be substantially simplified by propagating the observable @xmath13 in eq .
( [ second - order ] ) backwards in time and making use of @xmath101
. then we apply eq .
( [ evolution - expansion ] ) to decompose @xmath102 in @xmath66 and @xmath103 in @xmath104 .
the integration over the reduced phase space includes an integral over the momentum direction @xmath45 that results in vanishing of all terms with @xmath105 .
the second order response function consists of several contributions , @xmath106 of similar form , e.g. : @xmath107 with the matrix elements @xmath108 we further establish recurrent relations for coefficients ( [ abqrs ] ) that allow expressing all of them via few `` initial conditions '' , e.g. @xmath109 .
for instance , in the simplest case @xmath110 , the recurrent relations read @xmath111 , and all coefficients @xmath112 are expressed via @xmath113 due to the symmetry @xmath114 @xcite . in this case
we find the asymptotic behavior @xmath115 for @xmath116 .
the summation over @xmath117 is performed numerically .
the set of coefficients @xmath118 , as well as the spectrum @xmath72 are attributes of a particular riemann surface .
constant curvature riemann surfaces of genus @xmath39 are classified by the so - called moduli spaces whose dimensions grow linearly in @xmath39 .
therefore , any finite set @xmath119 of spectral elements , as well as the relevant set of coefficients @xmath118 can be implemented for some particular riemann surface , and hereafter we will treat them as independent parameters .
the convergence of the series over @xmath117 is ensured by the dependence @xmath120 for large @xmath117 .
the series is almost sign - alternating , and the dependence of the summand magnitude on @xmath117 becomes smoother with increasing @xmath117 , independently of @xmath19 .
this allows for an effective procedure that evaluates the series , since it does not involve numerical summation of the exponentially increasing with @xmath19 number of terms .
the procedure also works in any order of @xmath121 and @xmath122 obtained from the hypergeometric expansions of @xmath90 @xcite . in ref .
we extend our study of optical response by adding langevin noise to classical deterministic dynamics .
we identify the rp resonances as the eigenvalues of the related fokker - planck operator @xmath123 in the weak - noise limit @xmath124 .
we demonstrate that a standard spectral decomposition of the response functions in the eigenmodes of the fokker - planck operator @xmath125 has a well - defined limit at @xmath126 that reproduces our expression for the classical response functions .
fourier transform of the second order response function : ( a ) single resonance @xmath127 , ( b ) linear combination of terms with two resonances @xmath128 and @xmath129 .
linear plots show cross - sections of the spectra at @xmath130 .
[ 2dfourier ] , width=259 ] the absolute value of the @xmath0 fourier transform of the second - order response function presented in fig .
[ 2dfourier ] shows diagonal and cross peaks , as well as a stretched along the @xmath131 direction feature ( real and imaginary parts are presented in appendix [ app : pictures ] ) .
the latter originates from time - domain damped oscillations with variable period and can be interpreted as a signature of chaos ( instability ) in the underlying dynamics .
also note that in our chaotic case the peak frequencies may not be attributed to any particular periodic motions , although they can be expressed in terms of all periodic orbits via the dynamical @xmath43-function @xcite .
therefore , they can be referred to as collective chaotic resonances .
our results can be applied to the interpretation of spectroscopic data .
classical chaos is quite generic for dynamics in large molecules and molecular networks coupled via hydrogen bonds . even for a small number of vibrational modes , e.g. in small systems of hydrogen bonds @xcite , the shape of the effective potential energy can lead to chaotic dynamics .
moreover , the chaos may be caused by just some regions with nonuniform curvature of the potential surface @xcite . in the present letter we studied classical nonlinear response in a strongly chaotic ( mixing ) system .
detailed analysis performed for free motion in a riemann compact surface of constant negative curvature demonstrated that calculation of the response based on liouville space dynamics does not have fundamental difficulties .
we demonstrated that peaks in @xmath0 spectra are not necessarily attributed to periodic or quasiperiodic motions , but rather can have collective nature .
we argue that collective resonances should be accompanied by stretched features in @xmath0 spectra that constitute a spectroscopic signature of underlying dynamical chaos . 99 a. tokmakoff , m. j. lang , d. s. larsen , g. r. fleming , v. chernyak , and s. mukamel , phys . rev . lett . * 79 * , 2702 ( 1997 ) .
m. c. asplund , m. t. zanni , and r. m. hochstrasser , proc .
natl .
acad .
sci .
usa * 97 * , 8219 ( 2000 ) .
m. d. fayer , ann .
rev .
phys . chem . *
52 * , 315 ( 2001 ) .
d. m. jonas , ann .
rev .
phys . chem . *
54 * , 425 ( 2003 ) .
a. stolow and d. m. jonas , sciense * 305 * , 1575 ( 2004 ) . j. a. leegwater and s. mukamel , j. chem .
phys . * 102 * , 2365 ( 1995 ) .
w. g. noid , g. s. ezra , and r. f. loring , j. phys .
chem .
b * 108 * , 6536 ( 2004 ) .
m. kryvohuz and j. cao , phys .
rev . lett . * 95 * , 180405 ( 2005 ) ; * 96 * , 030403 ( 2006 ) ; j. chem .
phys.*122 * , 024109 ( 2005 ) .
m. c. gutzwiller , _ chaos in classical and quantum mechanics _ , springer verlag , 1990 .
v. i. arnold , _ mathematical methods of classical mechanics _ , springer verlag , 1989 .
n. l. balazs and a. voros , phys .
rep . * 143 * , 109 ( 1986 ) .
y. g. sinai , russ .
math .
surv . * 25 * , 137 ( 1970 ) . c. dellago and s. mukamel , phys .
rev .
e * 67 * , 035205(r ) ( 2003 ) ; j. chem .
phys . * 119 * , 9344 ( 2003 ) .
d. ruelle , phys .
rev .
lett . * 56 * , 405 ( 1986 ) .
s. roberts and b. muzykantskii , j. phys .
a : math . gen . * 33 * , 8953 ( 2000 ) .
s. mukamel , v. khidekel , and v. chernyak , phys .
rev .
e * 53 * , r1 ( 1996 ) .
p. c. martin , _ mesurements and correlation functions _
( gordon and breach , new york , 1968 ) .
a. a. kirillov , _ elements of the theory of representation of groups _ , springer verlag , 1986 .
s. lang , @xmath132 , addison - wesley , 1975 .
f. l. williams , _ lectures on the spectrum of _ @xmath133 .
s. v. malinin and v. y. chernyak , to be published .
s. v. malinin and v. y. chernyak , to be published .
j. d. eaves , j. j. loparo , c. j. fecko , s. t. roberts , a. tokmakoff , and p. l. geissler , proc .
natl .
acad .
sci .
usa * 102 * , 13019 ( 2005 ) .
l. casetti , c. clementi , and m. pettini , phys .
rev .
e * 54 * , 5969 ( 1996 ) . |
the use of latent variables has several interesting applications in statistics , econometrics and related fields like quantitative marketing .
models like tobit , and probit , ordered probit and multinomial probit are good examples . references and examples
could be found in geweke , keane and runkle ( 1997 ) and allenby and rossi ( 1999 ) , especially for the multinomial probit .
when the estimation process uses some simulation technique , in particular in bayesian analysis , the need of drawing a sample from the distribution of the latent variable naturally arises .
this procedure _ augments _ the observed data @xmath2 with a new variable @xmath3 which will be referred as a _ latent _ variable .
it is usually referred to as _ data augmentation _ ( tanner and wong ( 1987 ) , tanner ( 1991 ) , gelfand and smith ( 1990 ) ) and in other contexts as _ imputation _ ( rubin 1987 ) .
the data augmentation algorithm is used when the likelihood function or posterior distribution of the parameter given the latent data @xmath4 is simpler than the posterior given the original observed data @xmath5 . if the distribution of the variables @xmath4 is a multivariate normal in @xmath6 dimensions and @xmath7 is the indicator function of the set @xmath1 , the drawing is made from the normal distribution truncated to @xmath1 ( and/or @xmath8 ) .
the goal of this paper is to produce exact ( or perfect ) samples from random variables with distributions supported on a @xmath6-dimensional box @xmath1 ; we call _ box _ the cartesian product of @xmath6 bounded intervals .
we construct a discrete - time stationary markov process @xmath9 in the state space @xmath10 whose time - marginal distribution at any time @xmath11 ( that is , the law of @xmath12 ) has a given density distribution @xmath13 with support in @xmath1 .
the construction is then implemented to perfectly simulate normal vectors of reasonable large dimension truncated to bounded boxes .
the approach is also useful to show uniqueness of the invariant measure for a family of processes in a infinite - dimensional space @xmath14^{z^d}$ ] with truncated gaussian distributions and nearest - neighbor interactions ( in preparation , see also section [ s6 ] ) .
the process is the gibbs sampler popularized by geman and geman ( 1985 ) and used for the truncated normal case by geweke ( 1991 ) and robert ( 1995 ) in a _ markov chain monte carlo ( mcmc ) _ algorithm to obtain samples with distribution approximating @xmath13 . to describe the process in our setting ,
let @xmath15 and take a box @xmath16 where @xmath17 are bounded intervals : @xmath18\subset\r$ ] , @xmath19 .
if @xmath20 is the configuration at time @xmath21 , then at time @xmath11 a site @xmath22 is chosen uniformly in @xmath23 , say @xmath24 and @xmath25 is substituted by a value @xmath26 chosen with the density @xmath13 conditioned to the values of the other coordinates ; that is with the density @xmath27 given by @xmath28 where @xmath29 , @xmath30 , @xmath31 .
note that @xmath32 does not depend on @xmath25 .
the distribution in @xmath10 with density @xmath13 is reversible measure for this dynamics .
we construct a stationary gibbs sampler @xmath9 as a function of a sequence @xmath33 of independent identically distributed and uniform in @xmath34 $ ] random variables and the updating schedule @xmath35 .
we define ( backwards ) stopping times @xmath36 $ ] such that @xmath37 is measurable with respect to the field generated by the uniform random variables and the schedule in @xmath38 $ ] .
the configuration at time @xmath11 depends on the uniform random variables and the schedule in @xmath39 $ ] .
in fact we construct simultaneously processes @xmath40 } , -\infty$ ] @xmath41 . for each fixed
@xmath42 and @xmath43 , @xmath40},\,s\le t<\infty)$ ] is the gibbs sampler starting at time @xmath42 with configuration @xmath43 . for each fixed @xmath11 , @xmath40},\,s\le t,\,\zeta\in\cb)$ ]
is a _ maximal coupling _ ( see thorisson ( 2000 ) and section [ s2 ] ) .
for each fixed time interval @xmath44 $ ] the process @xmath45 } , s\le t'\le t)$ ] is a function of the uniform random variables and the schedule in @xmath38 $ ] and the initial @xmath43 .
the crucial property of the coupling is that @xmath46 does not depend on @xmath43 .
the construction is a particular implementation of the _ coupling from the past ( cftp ) _ algorithm of propp and wilson ( 1996 ) to obtain samples of a law @xmath47 as a function of a finite ( but random ) number of uniform random variables . for an annotated bibliography on the subject see the web page of wilson ( 1998 ) .
a markov process having @xmath47 as unique invariant measure is constructed as a function of the uniform random variables .
the processes starting at all possible initial states run from negative time @xmath42 to time @xmath48 using the uniform random variables @xmath49 and a fixed updating schedule .
if at time 0 all the realizations coincide , then this common state has distribution @xmath47 .
if they do not coincide then start the algorithm at time @xmath50 using the random variables @xmath51 and continue this way up to the moment all realizations coincide at time 0 ; @xmath52 is the maximal @xmath53 such that this holds .
the difficulty is that unless the dynamics is monotone , one has to effectively couple all ( non countable ) initial states .
murdoch and green ( 1998 ) proposed various procedures to transform the infinite set of initial states in a more tractable finite subset . in the normal case , for some intervals and covariance matrices , the maximal coupling works without further treatment .
moeller ( 1999 ) studies the continuous state spaces when the process undergoes some kind of monotonicity ; this is not the case here unless all correlations are non negative .
philippe and robert ( 2003 ) propose an algorithm to perfectly simulate normal vectors conditioned to be positive using coupling from the past and slice sampling ; the method is efficient in low dimensions .
consider a @xmath6-dimensional random vector @xmath54 with normal distribution of density @xmath55 where the covariance matrix @xmath56 is a positive definite matrix , @xmath57 is the mean vector and @xmath58 in this case we say that @xmath59 .
we denote @xmath60 the truncation of @xmath61 to the box @xmath62 ; its density is @xmath63 where @xmath64=\int_{\cb}f({\bx})d{\bx}$ ] is the normalizing constant .
we have tested the algorithm in some examples .
in particular for the matrix @xmath65 where @xmath66 is the identity matrix and @xmath67 . for the box @xmath68^d$ ] and for @xmath69 our method is faster than the rejection method .
the simulations indicate that the computer time for the rejection method grows exponentially with @xmath6 while it grows like @xmath70 for our method . using a simple program in matlab in desktop microcomputer the method permits to simulate up to dimension @xmath71 for the interval @xmath68^d$ ] . for bigger boxes or boxes
apart from the mean , for instance @xmath72^d$ ] , our method is less efficient but in any case it is much faster than the rejection one . since our method uses several times the same uniform variables , to compare the computation time we have counted the number of times each of the one - dimensional uniform random variable is used .
it is expected that if the box @xmath1 is sufficiently small and the covariance matrix has a `` neighbor structure '' , the number of steps may grow as @xmath73 . in this case
one should be able to simulate the state of an infinite dimensional gaussian field in a finite box .
correlated normal vectors can be mapped into independent standard normal vectors using the cholesky transformation ( see chapter xi in devroye ( 1986 ) ) . in the case of truncated normals to a box , the transformation maps the box into a set .
one might simulate standard normals and reject if they do not belong to the transformed box .
the difficulties are similar to those of the rejection method : as the dimension grows , the transformed set has small probability and the expected number of iterations grows exponentially with the dimension .
when the truncating set is not a box , our method generally fails as the coupling event for each coordinate has probability zero in general .
we illustrate this with an example in section [ s6 ] .
another possibility , also discussed by devroye ( 1986 ) is to compute the marginal of the first coordinate , then the second marginal conditioned to the first one and so on .
the problem here is that the computation of the marginals conditioned to all the other coordinates be in @xmath1 may be as complicated as to compute the whole truncated vector . in section [ s2 ]
we define coupling and maximal coupling . in section [ s3 ]
we describe the stationary theoretical construction of the gibbs sampler and the properties of the coupling .
section [ s4 ] is devoted to the pseudocode of the perfect simulation algorithm . in section [ s5 ]
we compute the functions entering the algorithm for the truncated normal case .
in section [ s6 ] we give some examples and compare our perfect simulation algorithm with the rejection one based on the uniform distribution .
a _ coupling _ of a family of random variables @xmath74 with @xmath75 a label set , is a family of random variables @xmath76 with the same marginals ; that is , such that @xmath77 an event @xmath78 is called _ coupling event _ , if @xmath79 are identical in @xmath78 .
that is , @xmath80 we consider continuous random variables @xmath81 in @xmath82 with densities
@xmath83 satisfying @xmath84 for any coupling event @xmath78 .
this is always true if @xmath75 is countable and in the normal case treated here . when there is a coupling event @xmath78 such that identity holds in the coupling is called _
maximal_. see thorisson ( 2000 ) for a complete treatment of coupling , including historical quotations .
a natural maximal coupling of these variables is the following .
let @xmath85 be the corresponding cumulative distribution functions and define @xmath86 let @xmath87 .
let @xmath88 be a random variable uniformly distributed in @xmath34 $ ] and define @xmath89 where @xmath90 , the _ generalized inverse _ of @xmath91 , is defined by @xmath92 the marginal distribution of @xmath93 is @xmath94 : @xmath95 the process @xmath96 is a maximal coupling for the family @xmath97 .
we use this coupling to construct the dynamics starting from different initial conditions .
then we show how to compute @xmath98 and @xmath99 in the normal case . in the sequel we use a family of couplings .
let @xmath100 be a decreasing sequence of parameter sets : @xmath101 for all @xmath102 .
let @xmath103}(x)\equiv 0 $ ] and for @xmath102 , @xmath104}(x):=\int_{-\infty}^{x}\inf_{\lambda\in \lambda_\ell}g_\lambda(y)dy;\qquad r_{[\ell]}:=r_{[\ell]}(\infty)\ , .
\end{aligned}\ ] ] assume @xmath105}= 1 $ ] .
define @xmath106}(x ) = r_{[\ell]}(x)-r_{[\ell-1]}(x)+ r_{[\ell-1]}$ ] and for @xmath107 define @xmath108}(x)= g_\lambda(x ) - r_{[\ell]}(x ) + r_{[\ell]}$ ] and @xmath109}^{-1}(u ) \,\one\{u\in[r_{i-1},r_i)\ } \,+\ , \hat g_{\lambda,[\ell]}^{-1}(u ) \,\one\{u > r_{[\ell]}\ } \,.\ ] ] then @xmath110 is a maximal coupling of @xmath111 for each @xmath112 .
the goal is to define a stationary process @xmath9 in @xmath1 with the gibbs sampler dynamics associated to a density @xmath13 in @xmath1 .
the time marginal of the process at any time will be the distribution with the density @xmath13 .
the gibbs sampler is defined as follows .
assume @xmath113 is known , then at time @xmath11 choose a random site @xmath22 uniformly in @xmath23 independently of `` everything else '' .
then set @xmath114 if @xmath115 and for @xmath29 choose @xmath116 with the law @xmath13 conditioned to the values of @xmath117 at the other coordinates : @xmath118 where @xmath119 , @xmath120 $ ] , is the conditional density of the @xmath121th component of @xmath13 given the other coordinates given in .
reversibility follows from the identity @xmath122 for any @xmath123 and @xmath124 continuous functions from @xmath1 to @xmath82 , where @xmath20 and @xmath30 .
in particular the law with density @xmath13 is invariant for the gibbs sampler .
it is indeed the unique invariant measure for the dynamics , as we show in theorem [ p31 ] . to define a stationary version of the process
we construct a family of couplings for each time interval @xmath44 $ ] .
the coupling starts with all possible configurations at time @xmath42 and uses random variables associated to each time to decide the updating .
all couplings use the same updating variables . to reflect the evolution of all possible configurations
we introduce a process on the set of boxes contained in @xmath1 .
the initial configuration of this process is @xmath1 at time @xmath42 but at times @xmath125 can be ( and will be as @xmath42 decreases ) reduced to a point .
let @xmath126 be a @xmath6-tuple of closed bounded intervals of @xmath82 , @xmath127 , @xmath128 .
we abuse notation calling @xmath126 also the box @xmath129 notice that @xmath126 has dimension less than @xmath6 when @xmath130 is just a point for some @xmath131 .
define @xmath132 these functions depend on @xmath126 only through @xmath133 .
@xmath134 is the probability that if the @xmath121-th coordinate is updated when the other coordinates belong to the set @xmath126 , then a coupled event is attained for the @xmath121 coordinate .
this means that for any configuration @xmath135 , the updated value of the @xmath121-th coordinate will be the same , say @xmath136 ; its law is given by @xmath137 . in this case
the updated interval @xmath138 is also reduced to the point @xmath136 . if the coupled event is not attained @xmath138 is kept as the interval @xmath17 .
the updating random variables consist on two families : @xmath139 , a family of independent variables with uniform distribution in @xmath34 $ ] and @xmath140 , a family of independent variables with uniform distribution in @xmath23 and independent of @xmath141 .
now for each @xmath142 we define a process @xmath143},\,t\ge s)$ ] taking values on boxes contained in @xmath1 as function of @xmath144 .
the initial state is @xmath145}=\cb$ ] and later each coordinate @xmath121 is either a ( random ) point or the full interval @xmath17 .
more precisely , for @xmath142 and @xmath146 set @xmath147}= 0\ , , \qquad r_{[s , s]}(x)= 0\ , , \qquad d_{[s , s]}(x)=0\ , , \qquad \eta_{[s , s]}=\cb.\end{aligned}\ ] ] fix @xmath148 and assume @xmath149}(x)$ ] , @xmath149}$ ] and @xmath150}$ ] are defined if @xmath151 .
then for @xmath152 and @xmath153 , define @xmath154}&=&r_{\kappa(t)}(\eta_{[s , t-1 ] } ) \\
r_{[s , t]}(x)&=&r_{\kappa(t)}(x|\eta_{[s , t-1 ] } ) , \\
d_{[s , t]}(x)&=&r_{[s+1,t]}+r_{[s , t]}(x)-r_{[s+1,t]}(x)\,\nonumber \\\label{oi } \eta_{[s , t]}(k)&=&\left\ { \begin{array}{lll } \eta_{[s , t-1]}(k)\,,\qquad \mbox{if}\;k\neq \kappa(t),\\ \\ \eta_{[s+1,t]}(k ) \,{\bf 1}\{u_t\leq r_{[s+1,t]}\}\,+ \,d_{[s , t]}^{-1}(u_t)\,{\bf 1}\{r_{[s+1,t]}<u_t\leq r_{[s , t]}\}\ , \\ \qquad\qquad\qquad + \,b(k)\,{\bf 1}\{u_t > r_{[s , t]}\}\ , , \qquad\mbox { if}\;k=\kappa(t ) ; \end{array } \right.\,.\end{aligned}\ ] ] in words : @xmath29 defines the coordinate to be updated at time @xmath11 .
@xmath149}$ ] is the probability that the coupling event is attained at coordinate @xmath22 for all the processes starting at times smaller than or equal to @xmath42 .
the coupling event is attained for the process starting at @xmath42 when @xmath155}$ ] . in case
the coupling event is attained , the value of coordinate @xmath121 is given by @xmath156}(u_t)$ ] , for @xmath157 given by the biggest @xmath158 such that @xmath155}$ ] ( second term in ) .
this value is the same for each @xmath159 ( first term in ) .
when the coupling event is not attained ( that is , for @xmath160 ) , the coordinate @xmath121 is kept equal to the full interval @xmath17 ( third term in ) .
it follows from this construction that @xmath161}(\cdot),\,r_{[t-1,t]},\,d_{[t-1,t]})$ ] does not depend on @xmath141 and , for @xmath162 , @xmath163}(\cdot),\,r_{[s , t]},\,d_{[s , t ] } ) \hbox { is a function of } ( ( u_n,\kappa(n)),\,n = s+1,\dots , t-1).\\ & & \eta_{[s , t ] } \hbox { is a function of } ( r_{[s , t]}(\cdot),\,r_{[s , t]},\,d_{[s , t ] } ) \hbox { and } ( u_t,\kappa(t))\,.\end{aligned}\ ] ] the @xmath121th coordinate of @xmath150}$ ] is either the interval @xmath17 or a point .
the process is monotonous in the following sense : @xmath164}\subset\eta_{[s , t]}\hbox { for all } s < t\,.\ ] ] in particular , if some coordinate is a point at time @xmath11 for the process starting at @xmath42 , then it will be the _ same _ point for the processes starting at previous times : @xmath165}(k ) \hbox { is a point , then } \eta_{[s-1,t]}(k ) = \eta_{[s , t]}(k),\qquad
\hbox { for all } s < t\,.\ ] ] for each time @xmath11 define @xmath166}(k ) \hbox { is a point for all } k\}\,.\ ] ] using we conclude that @xmath167 is a stopping time for the filtration generated by @xmath168 : the event @xmath169 is a function of @xmath170 .
assume @xmath171 almost surely for all @xmath11 and define the process @xmath172 in @xmath1 by @xmath173 } \in \cb\,.\ ] ] noticing that @xmath171 is equivalent to @xmath174}\to 1 $ ] as @xmath175 , we get the following explicit expression for @xmath12 : @xmath176}^{-1}(u_t)\,{\bf 1}\{r_{[n+1,t]}<u_t\leq r_{[n , t]}\}\ , , \qquad\mbox { if}\;k=\kappa(t ) ; \end{array } \right.\,.\end{aligned}\ ] ] for each @xmath177 and @xmath178 we now construct the process @xmath40},\ , t\ge s)$ ] , the gibbs sampler starting with @xmath43 at time @xmath42 .
let @xmath179 and @xmath180}(x|\zeta ) = r_{[s , t]}+ g_{\kappa(t)}(x|\zeta ) - r_{[s , t]}(x)\,.\ ] ] for each @xmath178 define @xmath181}=\zeta$ ] and for @xmath182 , @xmath183 } ( k)&=&\left\ { \begin{array}{lll } \zeta^\zeta_{[s , t-1]}(k)\,,\qquad \mbox{if}\;k\neq \kappa(t),\\ \\ \sum_{n = s}^{t-1 } \one\{u_t\in[r_{[n+1,t]},r_{[n , t]}]\}\ ,
d^{-1}_{[n+1,t]}(u_t)\\ \qquad\qquad \qquad\qquad + \ ; \one\{u_t > r_{[s , t]}\ } \,\hg_{[s , t]}^{-1}(u_t|\zeta^\zeta_{[s , t-1]})\ , , \qquad\mbox { if } \;k=\kappa(t ) \end{array } \right.\,.\end{aligned}\ ] ] for each fixed @xmath11 and @xmath121 , @xmath184}(k),\ , \zeta\in\cb),\,s\le t)$ ] is a maximal coupling among the processes starting with all possible configurations @xmath43 at all times @xmath158 .
[ p31 ] assume @xmath171 almost surely for all @xmath185 .
then the process @xmath9 defined in is a stationary gibbs sampler related to @xmath13 .
the distribution of @xmath12 at each time @xmath11 is the distribution with density @xmath13 in @xmath1 .
this distribution is the unique invariant measure for the process and @xmath186 } \neq \zeta_t ) \le \p(\tau(t)<s)\,,\ ] ] where @xmath187}$ ] is the process starting with configuration @xmath43 at time @xmath42 .
stationarity follows from the construction .
indeed , for all @xmath158 and @xmath188 , @xmath150}$ ] as a function of @xmath189 is identical to @xmath190}$ ] as a function of @xmath191 .
the function @xmath192\times [ 0,1]^{\{-\infty,\dots , t-1\}}\to b(\kappa(t))$ ] defined in by @xmath193}^{-1}(u)\,{\bf 1}\{r_{[n+1,t]}<u\leq r_{[n , t]}\}\ ] ] satisfies for @xmath29 @xmath194 where we used that @xmath117 is function of @xmath195 and @xmath196 .
this is sufficient to show that @xmath12 evolves according to the gibbs sampler .
to finish we need to show that the distribution with density @xmath13 is the marginal law of @xmath12 . since the updating is performed with the conditioned distribution , the distribution with density @xmath13 is invariant for the gibbs sampler .
since @xmath187}=\zeta_{[s , t]}$ ] if @xmath197 , then follows .
this also implies that @xmath198 } = \zeta_t$ ] almost surely . in particular @xmath187}$ ]
converges weakly to the law of @xmath12 as @xmath199 .
the same is true for @xmath200}$ ] .
taking @xmath43 random with law @xmath13 , this proves @xmath12 must have law @xmath13 .
then taking @xmath43 with law @xmath201 invariant for the dynamics , we conclude @xmath202 and the uniqueness of the invariant measure follows .
[ s22 ] a sufficient condition for @xmath171 _ a.s . _
is that @xmath203 for all @xmath121 . in this case @xmath204
decays exponentially fast with @xmath205 .
@xmath206 @xmath207 and @xmath208 , the last time in the past the coordinates have been successively updated to a point independently of the other coordinates . at @xmath209 and successive times @xmath210 , each coordinate of @xmath211}$ ] has been reduced to a point . @xmath212 is finite for all @xmath11 because the event @xmath213 and @xmath208 occurs for infinitely many @xmath42 with probability one .
the same argument shows the exponential decay of the tail of @xmath214 .
[ [ remark ] ] remark + + + + + + the velocity of convergence of the gibbs sampler to equilibrium depends on the values @xmath215 . in turn , these values depend on the size of the box @xmath1 and the correlations of the distribution @xmath13 . strongly correlated vectors produce small values @xmath216 and hence slow convergence to equilibrium .
the efficiency of the algorithms discussed in the next sections will also depend on these values .
the construction of section [ s3 ] is implemented in a perfect simulation algorithm .
let @xmath126 be a box contained in @xmath1 . for each @xmath19 , let @xmath217 and @xmath218 as in and .
when @xmath219 we denote by @xmath220 the value of @xmath221 .
let @xmath22 be the updating schedule .
it can be either a family of iid chosen uniformly in @xmath23 or the periodic sequence @xmath222_{{\rm mod } \
d}+1 $ ] . for
each pair @xmath223 of boxes contained in @xmath1 and @xmath224 , let @xmath225 $ ] be the function defined by @xmath226 let @xmath227 be boxes contained in @xmath1 , @xmath228 $ ] , and @xmath224 . the auxiliary _ coupler function _
@xmath229 is defined by ` function ` @xmath230 ` : ` + @xmath231 + @xmath232 + ` if ` @xmath233 ` is not a point in ` @xmath82 ` and ` @xmath234 ` , ` + ` update the ` @xmath121`th coordinate of ` @xmath126 ` : ` + ` `
@xmath235 + ` end ` + @xmath236 + ` return(`@xmath237 ` ) ` [ [ perfect - simulation - algorithm ] ] * perfect simulation algorithm : * + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + perform the following steps @xmath238 + ` start(`@xmath239 ` ) ` + ` while ` @xmath239 ` is not a point in ` @xmath1 + ` ` @xmath240 + ` start(`@xmath241 ` ) ` + ` ` @xmath242 + ` while ` @xmath243 ` update ` @xmath244 ` using the coupler function ` @xmath229 ` : ` + ` ` @xmath245 + ` ` @xmath246 + ` end ` + ` end ` + ` return(`@xmath239 ` ) ` where , for each @xmath247 , `` start(@xmath248 ) '' is the following subroutine ` start(`@xmath241 ` ) : ` + @xmath249 + ` generate ` @xmath250 ` a uniform random variable in ` @xmath34 $ ] + ` if ` @xmath251 ` , compute the ` @xmath252`th coordinate of ` @xmath241 ` : ` + ` ` @xmath253 + ` end `
in this section we implement the construction in the normal case .
we start with elementary facts of the one dimensional normal distribution . [
[ one - dimension ] ] one dimension + + + + + + + + + + + + + let @xmath254 be the standard normal density in @xmath82 and @xmath255 the corresponding distribution function .
let @xmath256 be real numbers .
a random variable @xmath3 with mean @xmath0 and variance @xmath257 has _ truncated normal distribution _ to the interval @xmath14 $ ] if its density is given by @xmath258 where @xmath259 and @xmath260 .
this distribution is called @xmath261 .
the truncated normal is the law @xmath262 conditioned to @xmath14 $ ] .
let @xmath263 let @xmath264 $ ] be a family of normal distributions truncated to the interval @xmath14 $ ] : @xmath265 .
let @xmath266 and @xmath267 be defined by @xmath268 and @xmath269 let @xmath270 $ ] and @xmath271\\ r(i ) & = & r(b|i)\,.\end{aligned}\ ] ] the next proposition , proven later , gives an explicit way of computing the functions participating in the multicoupling .
it says that the infimum of normal densities with same variances @xmath272 and means in the interval @xmath270 $ ] coincides with the normal with mean @xmath273 up to @xmath274 and with the normal with mean @xmath275 from this point on .
[ p22 ] the integrand in has the following explicit expression @xmath276 as a consequence,@xmath277{\bf 1}\{x < x(i,\sigma)\}\nonumber\\ & & \qquad+\,\frac{1}{a^{+}}\left[\phi\left(\frac{x(i,\sigma)-\mu^+}{\sigma}\right ) -\phi\left(\frac{a-\mu^+}{\sigma}\right)\right]{\bf
1}\{x\geq x(i,\sigma)\}\nonumber\\ & & \qquad+\,\frac{1}{a^{-}}\left[\phi\left(\frac{x-\mu^-}{\sigma}\right ) -\phi\left(\frac{x(i,\sigma)-\mu^-}{\sigma}\right)\right]{\bf 1}\{x\geq x(i,\sigma)\}. \nonumber\end{aligned}\ ] ] in particular , @xmath278 + \frac{1}{a^{-}}\left[\phi\left(\frac{b-\mu^-}{\sigma}\right ) -\phi\left(\frac{x(i,\sigma)-\mu^-}{\sigma}\right)\right]\ ] ] and for each @xmath279 $ ] there exists a unique @xmath280 $ ] such that @xmath281 .
[ [ remark-1 ] ] remark + + + + + + @xmath282 .
indeed , @xmath283 because @xmath284 . if @xmath285 , then all densities coincide unless @xmath286 ; by hypothesis this trivial case is excluded . [ [ multivariate - normal ] ] multivariate normal + + + + + + + + + + + + + + + + + + + let @xmath287 , @xmath288 be the vector conditioned to @xmath10 and @xmath289 .
the law of @xmath290 conditioned to @xmath291 is the truncated normal @xmath292 , where @xmath293 we see that @xmath294 does not depend on @xmath25 and @xmath295\,.\end{aligned}\ ] ] where @xmath296 [ [ remark-2 ] ] remark + + + + + + since @xmath297 for all @xmath121 , @xmath171 almost surely for all @xmath185 .
this implies that the truncated multivariate normal case falls under theorem s [ p31 ] hypothesis and our algorithm works .
[ [ back - to - one - dimension .- minimum - of - truncated - normals ] ] back to one dimension .
minimum of truncated normals + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we prove proposition [ p22 ] . to simplify notation write @xmath298 instead of @xmath299 .
observing that @xmath300 it is sufficient to prove for @xmath301 . the proof is based in the following elementary lemmas . [ s14 ]
let @xmath302 .
then @xmath303 where @xmath304 [ s16 ] for all @xmath305 it holds @xmath306 [ [ proof - of - proposition - p22 ] ] proof of proposition [ p22 ] + + + + + + + + + + + + + + + + + + + + + + + + + + let @xmath305 .
by lemma [ s14 ] we have @xmath307 hence @xmath308 if and only if @xmath309\,.\end{aligned}\ ] ] but by lemma [ s16 ] the interval in is empty for all @xmath310 .
this implies that@xmath311 , from where @xmath312 holds for @xmath301 . the corresponding identity ( [ marzo1 ] )
follows by applying lemma [ s14 ] to @xmath313 and @xmath314 .
in this section we compare our method with the rejection method in some examples .
the conclusion is that the rejection method may be better than ours in dimension 2 for some regions but ours becomes better and better as dimension increases .
then we show why the method does not work when the region is not a box .
this discards the following tempting approach : multiply the target vector by a matrix to obtain a vector with iid coordinates .
the transformed vector is much easier to simulate ; the problem is that it is now conditioned to a transformed region .
when the region is not a box , the conditioned vector is not an iid vector and a coupling must be performed .
we show here that in general the corresponding coupling event has probability zero .
[ [ the - rejection - method ] ] the rejection method + + + + + + + + + + + + + + + + + + + + we compare our algorithm with the following rejection algorithm : simulate a uniformly distributed vector @xmath315 in @xmath316 $ ] . if @xmath317 , then accept @xmath318 .
we use @xmath319 one - dimensional uniform random variables for each attempt of the rejection algorithm .
let @xmath320 be the expected number of one - dimensional uniform random variables generated by our perfect simulation algorithm , each counted the number of times that it is used .
let @xmath321 be the expected number of one - dimensional uniform random variables needed by the rejection algorithm to produce an accepted value .
we divide the experiments in two parts : @xmath322 and @xmath15 . @xmath323 $ ] .
let @xmath324 and @xmath325 .
we consider two cases .
( i ) boxes @xmath326\times[0,1]+(x_1,x_2)$ ] , @xmath327 and @xmath328 .
( ii ) boxes @xmath329\times[0,r]$ ] ( type @xmath330 ) and @xmath331\times[0,r]$ ] ( type @xmath332 ) , @xmath333 .
the results are shown in table [ uno ] for ( i ) and in figure [ dos ] for ( ii ) .
@xmath334 @xmath335 $ ] example 1 .
fix @xmath336 and @xmath337 , where @xmath66 is the identity matrix and @xmath338 .
we consider two cases : ( i ) boxes @xmath339^d$ ] , @xmath340 and ( ii ) boxes @xmath341^d$ ] , @xmath342 .
the results are shown in figure [ tres ] and table [ cuatro ] for ( i ) and ( ii ) , respectively .
@xmath343 @xmath335 $ ] example @xmath332 .
let @xmath336 and @xmath344 be the matrix , with `` neighbor structure '' , @xmath345 , with @xmath346 .
we consider boxes of the form @xmath347^d$ ] .
the results are shown in figure [ cinco ] .
[ [ strongly - correlated - variables ] ] strongly correlated variables + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the algorithm slows down very fast with the dimension when the variables are strongly correlated .
an associate editor proposes to consider a gaussian vector truncated to the box @xmath348^d$ ] , with mean @xmath349 and covariance matrix @xmath350 where @xmath66 is the identity matrix and @xmath351 , for small values of @xmath352 . as a measure of the speed of the algorithm
we have computed the coefficient @xmath99 , shown in table [ x2 ] , as function of @xmath352 and the dimension @xmath6 .
@xmath353 in @xmath322 the coupling time @xmath52 for this example is a geometric with mean @xmath354 , because a it suffices that one of the uniforms is smaller than @xmath99 to get the coupling in the next step . in general , as explained in lemma [ s22 ] , @xmath355 , whose expectation is of the order @xmath356 .
[ [ other - regions ] ] other regions + + + + + + + + + + + + + let @xmath357 be a @xmath6-dimensional multinormal random vector with zero mean and nonsingular covariance matrix @xmath358 .
let @xmath359\times\cdots\times[a_d , b_d]$ ] .
if we rotate and scale the coordinates to obtain iid @xmath360 s , i.e. , if we put @xmath361 , then the box constraints are of the form @xmath362 , where @xmath363 and @xmath364 . when the constraints are of this form , the conditional distribution of @xmath365 has a probability density function whose support depends on the condition @xmath366 . in such case
the maximal coupling in our approach may have a coupling event with zero probability .
indeed , it could happen that @xmath367 this implies that @xmath368 for all @xmath25 . to see this in @xmath322 ,
observe that the transformation takes the rectangular box into a parallelogram @xmath369 with sides not in general parallel to the coordinate axes .
we need to simulate the standard bi - dimensional normal @xmath4 conditioned to @xmath369 .
it is still true that the law of @xmath370 conditioned on @xmath371 is normal truncated to the slice of the parallelogram @xmath372 .
it is a matter of geometry that there are different @xmath26 and @xmath373 such that @xmath374 ; this implies the infimum above is zero .
take for instance @xmath324 , @xmath375 and @xmath5 conditioned to the box @xmath347\times[0,1]$ ] .
then @xmath376 , and the transformed box is a parallelogram @xmath369 with vertices @xmath377 .
we have proposed a construction ( simulation ) of a multidimensional random vector conditioned to a bounded box using a version of the coupling from the past algorithm of the gibbs sampler dynamics . in the case of normal distribution ,
we have given a explicit construction taking advantage of the bell shape of the density and the fact that the distribution of one coordinate given the others is a truncated normal with variance independent of the other coordinates .
this allows to define the coupling set @xmath78 in function of the extreme possible values of the mean and to obtain positive probability for each coupling set @xmath78 as soon as the box is bounded .
one of the consequences of the approach is to show that the gibbs sampler is ergodic in the box if the coupling event has probability uniformly bounded below in each one of the coordinates .
this corresponds to the condition @xmath203 for all coordinate @xmath121 . when the uniform random variable used to update site @xmath121 at some time
falls below @xmath203 , this coordinate coincides for all processes .
the method can be used to show ergodicity of the process in infinite volume when the inverse of the covariance matrix has a neighbor structure ; for instance for tridiagonal matrices . as an example , consider the formal density in @xmath14^{\z^d}$ ] : @xmath378 which corresponds to an infinite volume gaussian field , each coordinate truncated to @xmath14\subset \r$ ] ; @xmath379 is a parameter ( inverse temperature in statistical mechanics ) .
the gibbs sampler can be performed in continuous time and the coupling can be done as in the finite case .
however the updating of site @xmath121 depends only on the configurations in sites @xmath380 such that @xmath381 .
each site gets a definite value ( and not an interval ) when the corresponding uniform random variable falls below a certain value @xmath382)$ ] ( corresponding to @xmath216 , but now it is constant in the coordinates ) . to determine the value of site @xmath121 at time @xmath11 one explores backwards the process ; calling @xmath88 the random variable used for the last time ( say @xmath383 ) before @xmath11 to update site @xmath121 there are two cases : @xmath384 or @xmath385 . if @xmath384 , we do not need to go further back , the value is determined . if @xmath385 , we need to know the values of @xmath386 neighbors at time @xmath383 . repeating the argument
, we construct a `` oriented percolation '' structure which will eventually finish if @xmath387 .
this value is obtained by dominating the percolation structure with a branching process which dies out with probability @xmath99 and produces @xmath386 offsprings with probability @xmath388 .
the value of @xmath99 depends on the length of the interval @xmath14 $ ] and the strength of the interaction governed by @xmath389 .
one can imagine that for small intervals and small @xmath389 things will work .
we thank christian robert and havard rue for discussions .
propp , j. g. and wilson , d. b. ( 1996 ) exact sampling with coupled markov chains and applications to statistical mechanics .
proceedings of the seventh international conference on random structures and algorithms ( atlanta , ga , 1995 ) .
_ random structures algorithms * 9 * _ , no . 1 - 2 , 223252 .
d. b. wilson .
annotated bibliography of perfectly random sampling with 77arkov chains . in d.
aldous and j. propp , editors , _ microsurveys in discrete probability _ , volume 41 of _ dimacs series in discrete mathematics and theoretical computer science _ , pages 209220 .
american mathematical society , 1998 .
updated versions can be found at http://dimacs.rutgers.edu/126dbwilson/exact . | the target measure @xmath0 is the distribution of a random vector in a box @xmath1 , a cartesian product of bounded intervals .
the gibbs sampler is a markov chain with invariant measure @xmath0 .
a `` coupling from the past '' construction of the gibbs sampler is used to show ergodicity of the dynamics and to perfectly simulate @xmath0 . an algorithm to sample vectors with multinormal distribution truncated to @xmath1
is then implemented . |
in this paper we show how to solve the inverse problem for ifsm in the case of trajectories of stochastic processes in @xmath0 .
the method is based on the solution of the inverse problem for ifsm due to forte and vrscay @xcite .
this is an extension of classical ifs methods which can be used for approximating a given element of @xmath1 thus in particular trajectories of stochastic processes on this space .
the final goal of this approach is simulation . indeed ,
several methods are currently available to simulate paths of stochastic processes and in particular of the brownian motion .
paths of the bm can be simulated using the properties of the increments of the process like in the euler scheme @xcite , or as the limit of a random walk or via @xmath0 decomposition like the kac - siegert / karnounen - loeve series @xcite . in this paper
we first propose a ifsm ( iterated function systems with maps ) operator whose fixed point is the trajectory of the bm .
we then use this representation of the process to simulate its trajectories .
the resulting simulated trajectories are self - affine , continuous and fractal by construction .
this fact produces more realistic trajectories than other schemes in the sense that their geometry is closer to the one of the true bm s trajectories .
the ifsm trajectory of the bm can then be used to generate more realistic solutions of stochastic differential equations .
the paper is organized as follows : section [ sec : ifsm ] recalls the theory of ifsm on @xmath0 , section [ sec : l2 ] recalls some details for stochastic processes with trajectories in @xmath0 and the link with the ifsm theory .
section [ sec : bm ] presents the application of the ifsm theory to the problem of simulation with particular attention to the case of the brownian motion .
the basic idea of iterated function systems ( ifs ) can be traced back to some historical papers but the use of such systems to construct fractals and other similar sets was first described by hutchinson ( 1981 ) .
the fundamental result on which the ifs method is based is banach theorem .
the mathematical context is the following : given @xmath2 in a complete metric space @xmath3 , find a contractive operator @xmath4 that admits a unique fixed point @xmath5 such that @xmath6 is small enough .
in fact if one is able to solve _ the inverse problem _ with arbitrary precision , it is possible to identify @xmath2 with the operator @xmath7 which has it as fixed point .
the fundamental theorems on which the ifs method is based on are the following : ( banach theorem ) let @xmath8 be a complete metric space ; suppose there exists a mapping @xmath4 such that @xmath9 for all @xmath10 and some @xmath11 .
@xmath12 is said to be the contractivity factor of @xmath7 .
then there exits a unique @xmath5 such that @xmath13 and for any @xmath14 we have @xmath15 when @xmath16 .
( collage theorem ) let @xmath8 be a complete metric space . given @xmath14 suppose that there exists a contractive map @xmath7 with contractivity factor @xmath11 such that @xmath17
. if @xmath18 is the fixed point of @xmath7 then @xmath19 .
let @xmath8 be a complete metric space and @xmath20 be two contractive mappings with fixed points @xmath21 and @xmath22 .
then @xmath23 where @xmath24 and @xmath25 is the contractivity factor of @xmath26 .
we are going to use a particular class of ifs operators , known as ifsm ( ifs with maps ) , introduced by forte and vrscay in 1994 .
let @xmath27 $ ] , @xmath28 be the lebesgue measure on @xmath29 ( the borel @xmath30-algebra ) and for any integer @xmath31 let @xmath32 denote the linear space of all real valued functions @xmath33 such that @xmath34 is integrable on @xmath35 . to build a contraction map @xmath7 on @xmath1 we need @xmath36-map contractive ifs i.e. a set of maps @xmath37 and a set of functions ( grey level maps ) @xmath38 with @xmath39 .
the operator @xmath7 corresponding to the @xmath36 map ifsm(@xmath40,@xmath41 ) is @xmath42 where the prime means that the sum operates on all those terms for which @xmath43 ( the inverse function of @xmath44 ) is defined .
let us define the following two sets @xmath45 @xmath46 @xcite let @xmath47 be an ifsm such that @xmath48 and @xmath49 for @xmath50 . then @xmath51 and for any @xmath52 we have @xmath53 where @xmath54 given @xmath55 the inverse problem consists of finding the operator @xmath7 such that @xmath56 in @xcite it is proved that this problem can be reduced to the determination of grey level maps @xmath57 which minimize the collage distance @xmath58 @xmath59 @xmath60 in the special case when * @xmath61 i.e. the sets @xmath62 `` tile '' @xmath63 * @xmath64 for @xmath65 we say that the maps @xmath44 are nonoverlapping .
later , in the applications , we will assume that , for @xmath66 , * @xmath67 * @xmath68 * @xmath69 , @xmath70 the collage distance becomes @xmath71 @xmath72 @xmath73 where @xmath74 @xmath58 is a quadratic form in @xmath75 and @xmath76 , that is @xmath77 where @xmath78 .
the matrix @xmath79 is symmetric and @xmath80 @xmath81 and @xmath82 . as in @xcite
we add an additional constraint in order to guarantee that the minimum of this quadratic form exists on a compact subset of feasible parameters @xmath75 and @xmath76 .
the additional constraint is @xmath83 the maps @xmath44 are choosen in an infinite set @xmath84 of fixed affine contraction maps on @xmath63 which has the @xmath28-dense and nonoverlapping property ( in the sense of the following definition ) ; when @xmath85 the corresponding @xmath44 is superfluous and the @xmath86-th term can be dropped from .
we say that @xmath84 generates a @xmath28-dense and nonoverlapping family @xmath87 of subsets of @xmath63 if for every @xmath88 and every @xmath89 there exists a finite set of integers @xmath90 , @xmath91 , @xmath66 , such that * @xmath92 * @xmath93 * @xmath94 if @xmath95 let @xmath96 be the @xmath36 truncations of @xmath40 .
let @xmath97 the @xmath36 vector of affine grey level maps .
let @xmath98 be the solution of the previous quadratic optimization problem and @xmath99 .
it can be shown that @xmath100 may be arbitrarly small when @xmath101 ( see @xcite ) .
let @xmath102 be a probability space and @xmath103 be a sequence of @xmath30-algebras such that @xmath104 .
let @xmath105 be a stochastic process in @xmath1 , that is a sequence of random variables @xmath106-adapted ( that is each variable @xmath107 is @xmath106-measurable ) .
given @xmath108 a trajectory of the process is the function @xmath109 belonging to @xmath1 . for a given @xmath107 , the trajectory of the stochastic process , the aim of the inverse problem consists in finding the parameters of the ifsm such that @xmath107 is the solution of the equation @xmath110 in this case the coefficients of the matrix @xmath79 and the vector @xmath111 of the previous section become @xmath112 and if @xmath113 it becomes @xmath114 the other elements in the matrix @xmath79 can be calculated as @xmath115 and @xmath116 for the vector @xmath111 @xmath117 and @xmath118 in the nonoverlapping case , we have * @xmath119 , @xmath65 , and @xmath120 , @xmath121 * @xmath122 , @xmath121 , @xmath65 and @xmath123 * @xmath124 , @xmath121 , @xmath65 and @xmath125 it also holds this self - similarity property .
let @xmath126 be the solution of the inverse problem with a set of nonverlapping maps @xmath44 and suppose that @xmath127 .
then @xmath128 for all @xmath129 .
in fact we have @xmath130 @xmath131 we suppose that a.e .
@xmath107 is an element of a subspace @xmath132 of @xmath1 and that @xmath107 is a zero - mean process .
let @xmath133 be the covariance function of this process that is @xmath134.\ ] ] and assume @xmath135 if @xmath136 comprises the entire spectrum of eigenvalues of @xmath133 , where @xmath137 and the associated orthonormal eigenfunctions @xmath138 form a complete set of the subspace @xmath132 then the kac and siegert decomposition holds : @xmath139 we also have that @xmath140 are uncorrelated random variables with mean @xmath141 and variance @xmath142 .
the following theorem states some properties of this decomposition .
suppose that @xmath143 and @xmath133 satisfies the properties above .
then * \i ) @xmath144 * \ii ) @xmath145 as @xmath146 a.s . *
\iii ) @xmath147 are with mean @xmath141 and variance @xmath142 * \iv ) @xmath148 where @xmath149 * \v ) @xmath150 ^ 2\to 0 $ ] for each t as @xmath146 * \vi ) @xmath151 with @xmath149 uncorrelated with mean @xmath141 and variance @xmath142 .
in the literature there several methods of simulation of the trajectory of the browian motion , i.e. the stochastic process @xmath153\}$ ] , such that @xmath154 a.s . , @xmath155 is distributed with gaussian law with zero mean and variance @xmath156 , and with independent increments . in this case
, the trajectory is obtained simulating the increments of @xmath157 in the following way : @xmath158 , @xmath159 , where the @xmath160 s are independent @xmath161 random variables . in
the other points the trajectory is built by linear interpolation of these simulated data .
karhunen - love / kac - siegert decomposition of @xmath157 is better for pathwise simulation @xmath162 with @xmath163 @xmath164 a basis of orthogonal functions and @xmath160 s are @xmath161 the trajectory generated by euler method is too simple and regular to mimic the roughness of the bm ; moreover the simulated path is stochastically equivalent to the true trajectory only on the points of the grid used in the simulation .
the kac - siegert decomposition of the bm is a pathwise approximation which can lead to a too smooth path ( see figure [ fig1 ] ) ; our idea is to use ifsm for generating fractal trajectories of the bm .
there are applications in finance ( for instance pricing of american options ) in which the whole path matters ; our ifsm approch produces a global approximation of the trajectory preserving the geometric fractal nature of the target .
this method can be also used to simulate paths of solutions of stochastic differential equations driven by brownian motion ( e.g. diffusion processes ) replacing the linear behaviour of the euler trajectory with a fractal object . )
pseudo - random gaussian numbers were used . ] for the solution of the inverse problem for the bm , we choose the so - called wavelet type maps @xcite , that is : @xmath165 with @xmath166 and @xmath167 for each fixed @xmath168 , the family @xmath169 is a set of nonoverlapping maps . for these maps @xmath170 . we organize them as follows @xmath171 to simulate a trajectory of @xmath157 with non overlapping maps we then need to simulate the joint distribution of all this objects 1 .
@xmath172 2 .
@xmath173 3 .
@xmath174 4 .
@xmath175 ) } b(t)dt$ ] 5 . @xmath176 but it appears to be still a too difficult problem . in practice , it is preferable to use all the above maps and not only the subset of non - overlapping maps . in this case , we need simulate the value of the trajectory of the brownian motion on a fixed grid ( using one of the known methods ) and we use these points to approximate the integrals in the quadratic form .
we then solve the constrained quadratic programming problem using standard algorithms ( see e.g. @xcite ) .
figure [ fig2 ] bottom shows an example of trajectory generated using the ifsm approach using wavelet type maps for @xmath177 , @xmath178 .
figure [ fig2 ] top represents the euler trajectory built on 50 gaussian terms which has been used to build the ifsm .
as one can notice the ifsm path shows more `` fractal '' complexity then the corresponding euler path .
we have proposed a new method to generate paths of the brownian motion .
these ifsm paths seems to mimic more closely the fractal nature of the trajectory of the brownian motion then existing schemes . at current stage
we are not able to show formal property of the ifsm path in terms of strong and weak approximation ( see @xcite ) .
open source software for generating ifsm trajectories written in c and r language @xcite is available via ` ifs ` package at http://cran.r-project.org for free download .
forte , b. , vrscay , e.r .
( 1998 ) , `` inverse problem methods for generalized fractal transforms '' , in _ fractal image encoding and analysis _ , nato asi series f , vol .
y. fisher , springer verlag , heidelberg . | several methods are currently available to simulate paths of the brownian motion . in particular , paths of the bm can be simulated using the properties of the increments of the process like in the euler scheme , or as the limit of a random walk or via @xmath0 decomposition like the kac - siegert / karnounen - loeve series . in this paper
we first propose a ifsm ( iterated function systems with maps ) operator whose fixed point is the trajectory of the bm .
we then use this representation of the process to simulate its trajectories .
the resulting simulated trajectories are self - affine , continuous and fractal by construction .
this fact produces more realistic trajectories than other schemes in the sense that their geometry is closer to the one of the true bm s trajectories .
the ifsm trajectory of the bm can then be used to generate more realistic solutions of stochastic differential equations . *
ams subject classification : * * keywords : * iterated function systems , brownian motion , simulation of stochastic differential equations |
. , - -. . . , , . - , . . , . , . ,
. [ f : metro ] . ( 52,53 ) ( 25,28)(0,0 ) ( 23,52)(0,0)@xmath4 ( 49,26)(0,0)@xmath1 ( 11,0)(0,0)[b] 1 ( 39,0)(0,0)[b] 2 , . , . ( , , ) .
, , , . [ s : clock ] ( ) .
. ,
[ s : l ] .
( 23,49 ) ( 13,26)(0,0 ) ( 0,8)(0,0)[l]@xmath5 ( 0,44)(0,0)[l]@xmath6 ( 4,0)(0,0)[b]1 ( 22,0)(0,0)[b]2 , 1 2 @xmath5 , ( . [ f : sig1 ] ) .
, : * , @xmath7 , . , ; . * @xmath8 , . , . . * @xmath8 . . , . ( ) .
, ( ) ( [ s : p ] ) .
, . - , , . , -. . ( 50,36 ) ( 25,18)(0,0 ) ( 23,9)(0,0)@xmath9 ( 39,21)(0,0)@xmath10 @xmath9 . ( @xmath10 ) , ( . [ f : bomb ] ) .
, @xmath9 , . @xmath9 , ( ) , ( , ) .
_ _ @xmath9 .
, , @xmath9 , . ( ) @xmath9 .
_
@xmath9 .
_
@xmath9 ( . [ f : cone1 ] ) .
@xmath9 , @xmath9 . ( 42,42 ) ( 21,21)(0,0).,title="fig : " ] ( 21,37)(0,0) ( 21,5)(0,0) ( 41,21)(0,0)[r] ( 18,21)(0,0)@xmath9 ( ) . [ f : cone ] .
; ; . , . ( 40,48 ) ( 20,20)(0,0 )
, , . , _ _. - , , . ( 25,51 ) ( 12.5,24)(0,0 ) ( 5,51)(0,0)[t]1 ( 24,51)(0,0)[t]2 ( 3,24)(0,0)@xmath5 ( 22,37)(0,0)@xmath6 ( 6.5,14)(0,0)@xmath11 ( 2.5,8)(0,0)@xmath9 , ( @xmath9 , . [ f : sig2 ] ) .
0 @xmath9 .
1 @xmath5 ( ) ; 2 @xmath6 ( ) .
@xmath12 @xmath11 1 2 .
( 29.6,51 ) ( 14.8,24)(0,0 ) ( 5.8,51)(0,0)[t]1 ( 17.2,51)(0,0)[t]2 ( 28.6,51)(0,0)[t]3 ( 3.8,21)(0,0)@xmath5 ( 9,28)(0,0)@xmath6 ( 26,37.5)(0,0)@xmath13 ( 3,12)(0,0)@xmath14 ( 10.5,10.5)(0,0)@xmath15 ( 8,16)(0,0)@xmath16 ?
, . ( @xmath9 , . [ f : sig3 ] ) .
@xmath17 @xmath18 . , ( [ l : phi ] ) ( ) .
1 @xmath5 ( ) ; 2 @xmath4 ( @xmath10 ) , ; 1 @xmath6 ( . [ f : l ] ) .
, @xmath19 ; , @xmath20 , 1 2 . , @xmath21 ( 32.5,66 ) ( 16.25,31.5)(0,0 ) ( 5,66)(0,0)[t]1 ( 31.5,66)(0,0)[t]2 ( 3,24)(0,0)@xmath5 ( 3,54)(0,0)@xmath6 ( 21.5,39)(0,0)@xmath4 ( 17,39)(0,0)@xmath10 ( 2.5,8)(0,0)@xmath9 ( 6.5,14)(0,0)@xmath11 @xmath10 1 , . @xmath10 1 @xmath22 $ ] , . , @xmath10 1 . 1 , ; 1 @xmath10 . , , @xmath10 1 @xmath23 ( [ l : t12 ] ) , @xmath24 : @xmath25 , 1 ( .. @xmath11 ) .
@xmath26 , @xmath27 @xmath1 ( @xmath9 @xmath10 ) ; @xmath28 @xmath1 , .. @xmath9 @xmath10 .
@xmath4 , @xmath9 @xmath10 2 ( ) .
( [ l : x2 ] ) , @xmath29 @xmath30 .
( ) .
( [ l : csh ] ) , 2 1 @xmath31 ( @xmath32 , 1 @xmath33 ) .
, , ; @xmath11 , ( [ l : phi ] ) .
( 22.4,69 ) ( 11.2,33)(0,0 ) ( 3.4,69)(0,0)[t]1 ( 20.2,69)(0,0)[t]2 ( 0.4,21)(0,0)@xmath5 ( 0.4,57)(0,0)@xmath6 ( 10.5,24)(0,0)@xmath34 ( 17,46)(0,0)@xmath35 ( 4.4,15)(0,0)@xmath11 ( 23,39)(0,0)@xmath10 ( 0.4,8)(0,0)@xmath9 @xmath26 , @xmath27 @xmath10 1 @xmath36 , @xmath37 2 , @xmath11 1 ? . [ f : l2 ] @xmath38 @xmath39 @xmath40 @xmath41 . , @xmath1 : @xmath42 .
( [ s : lf ] ) .
@xmath1 @xmath9 @xmath10 .
( . [ f : cone1 ] ) : * @xmath43 , ( @xmath10 @xmath9 ) . , , @xmath26 ( @xmath9 @xmath10 ) . * @xmath44 , ( @xmath10 @xmath9 ) . ( , ) , , . * @xmath45 ( @xmath10 @xmath9 )
. , , @xmath27 ( .. @xmath9 @xmath10 ) .
, @xmath28 ( [ l : x2 ] ) .
@xmath46 ?
@xmath47 @xmath48 .
@xmath49 @xmath1 ( @xmath50 , 1 , 2 , 3 ) @xmath51 : @xmath52 @xmath53 ( ) 0 3 .
@xmath1 @xmath2 , @xmath54 .
: @xmath55 , ( ) , @xmath56 , , . , . , . @xmath1 @xmath2 , ( . [ f : ort ] ) : @xmath57 , @xmath58 , @xmath54 ; , @xmath2 . , : , , . ( 43,43 ) ( 21,21)(0,0 ) ( 17,40 ) ( 40,17 ) ( 27,38)(0,0)@xmath1 ( 38,27)(0,0)@xmath2 ( 22.25,27)(0,0)@xmath11 ( 27,21.75)(0,0)@xmath11 @xmath1 @xmath4 ( @xmath59 ) , @xmath11 , ( [ l : csh ] ) @xmath60 ( . [ f : proj ] ) .
@xmath61 @xmath62 .
@xmath63 ; @xmath64 .
( 42,28 ) ( 21,13.5)(0,0 ) .,title="fig : " ] ( 17,25 ) ( 40,2 ) ( 30,23)(0,0)@xmath1 ( 22.5,12)(0,0)@xmath11 ( 20,22)(0,0)[r]@xmath65 ( 29,5)(0,0)[t]@xmath66 @xmath1 @xmath2 ( , ) @xmath1 ( @xmath67 ) , @xmath2 @xmath1 ( @xmath68 ) : @xmath69 @xmath11 @xmath1 @xmath2 .
@xmath70 ; , ( @xmath64 ) .
`` '' @xmath10 , @xmath9 @xmath4 .
@xmath59 , ( . [ f : hyp]a ) .
, , ; . @xmath71 ( 104,45 ) ( 21,24)(0,0 ) ; ( b ) @xmath72.,title="fig : " ] ( 21,0)(0,0)[b]a ( 19,26)(0,0)@xmath9 ( 29,39.5)(0,0)@xmath10 ( 33,38)(0,0)@xmath1 ( 19.5,36)(0,0)@xmath4 ( 22.5,29)(0,0)@xmath11 ( 83,24)(0,0 ) ; ( b ) @xmath72.,title="fig : " ] ( 83,0)(0,0)[b]b ( 81,26)(0,0)@xmath9 ( 98,32)(0,0)@xmath10 ( 96,35)(0,0)@xmath1 ( 95,22)(0,0)@xmath73 ( 88,25.5)(0,0)@xmath11 `` '' ( @xmath72 , . [ f : hyp]b ) , , . @xmath74 ( [ m : hypt ] ) ( [ m : hypr ] ) @xmath11 ( . [ f : ort ] ) .
( 42,42 ) ( 21,21)(0,0).,title="fig : " ] ( 23,3)(0,0)[l]@xmath75 ( 34,24.5)(0,0)[l]@xmath11 ( 37.5,29.5)(0,0)[l]@xmath76 `` '' ( [ m : hypr ] ) , 0 @xmath75 , @xmath77 @xmath78 ( . [ f : hyp2 ] ) .
@xmath11 @xmath76 @xmath78 @xmath11 , @xmath11 0 , @xmath76
@xmath75 . - : , @xmath11 @xmath76 , @xmath79 @xmath80 @xmath75 @xmath78 ( ) .
( ) `` '' @xmath59 ( . [ f : hyp3]a ) .
, , . `` '' @xmath72 ( . [ f : hyp3]b ) , , . . ( 100,51 ) ( 20,27)(0,0 ) ; ( b ) @xmath72.,title="fig:",width=151 ] ( 80,27)(0,0 ) ; ( b ) @xmath72.,title="fig:",width=151 ] ( 20,0)(0,0)[b]a ( 80,0)(0,0)[b]b @xmath81 @xmath82 , ( . [ f : tri ] ) .
`` '' : @xmath9 @xmath83 @xmath84 , @xmath10 @xmath83 @xmath85 .
, @xmath9 @xmath10 , @xmath9
@xmath83 @xmath83 @xmath10 : @xmath86 , @xmath1 @xmath2 ; @xmath1 @xmath2 , @xmath87 .
, @xmath88 @xmath89 @xmath90 @xmath89 ( . [ f : proj ] ) ; , @xmath91 @xmath92 @xmath90 @xmath92 .
- : ( [ m : xy ] ) .
( 27.75,49 ) ( 15.5,24.5)(0,0 ) ( 0,2)(0,0)[l]@xmath9 ( 0,47)(0,0)[l]@xmath10 ( 0,22.6)(0,0)[l]@xmath84 ( 0,27.7)(0,0)[l]@xmath93 ( 0,34.5)(0,0)[l]@xmath85 ( 19.5,24.5)(0,0)@xmath83 , @xmath9 @xmath10 , . ( [ m : tri ] ) , @xmath90 . @xmath9 @xmath10 , , . @xmath94 , @xmath95 , .. , @xmath96 @xmath1 : @xmath97 : @xmath98 ( , @xmath99 , @xmath100 ) .
1 , @xmath94 , .. : @xmath101 .
, 1 ( [ s : v ] ) .
@xmath95 : @xmath102 .
, @xmath11 ( . [ f : l3 ] ) .
@xmath103 @xmath11 .
@xmath1 : @xmath104 , @xmath105 .
@xmath11 : @xmath106 @xmath1 ( [ m : x ] ) : @xmath107 ( [ l : l ] ) .
( 82,82 ) ( 41,41)(0,0 ) ( 38.5,43)(0,0)@xmath9 ( 64.5,78)(0,0)@xmath10 ( 67,76)(0,0)@xmath1 ( 38.5,76)(0,0)@xmath26 ( 65,38.5)(0,0)@xmath27 ( 38,59.5)(0,0)@xmath94 ( 60,37.5)(0,0)@xmath95 ( 52,72)(0,0)@xmath36 ( 53,44)(0,0)@xmath37 ( 50,65.5)(0,0)@xmath108 ( 62,53.5)(0,0)@xmath109 ( 42.5,47)(0,0)@xmath11 ( 47,42)(0,0)@xmath11
( 42,44 ) ( 21,21)(0,0 ) ( 20,21)(0,0)[r]@xmath110 ( 22,26)(0,0)[r]@xmath111 ( 25.5,26)(0,0)[l]@xmath112 ( 37,42)(0,0)[l]@xmath113 @xmath110 .
@xmath114 @xmath115 , @xmath116 ( . [ f : v ] ) .
@xmath117 @xmath118 @xmath110 , . , , : @xmath119 @xmath120 @xmath121 , @xmath122 , . ( 42,42 ) ( 21,21)(0,0 ) ( 33,26)(0,0)[l]@xmath123 ( 27,33)(0,0)[b]@xmath124 ( 22.5,26)(0,0)@xmath11 ( 26,22)(0,0)@xmath11 @xmath125 ( . [ f : a ] )
. @xmath126 .
( .. ) @xmath11 @xmath127 ( [ m : arc ] ) ( @xmath64 ) .
@xmath128 , @xmath129 , , @xmath1 .
@xmath4 @xmath130 @xmath131 @xmath1 @xmath132 . @xmath133 @xmath134 , @xmath4 @xmath135 , @xmath11 ( . [ f : a ] ) .
, , ( [ v : xt ] ) : @xmath136 @xmath137 @xmath118 , @xmath110 . ( [ v : vt ] ) : @xmath138 @xmath139 ; . . @xmath73 @xmath140 . ( . [ f : spiral]a ) @xmath141 ( 83,65 ) ( 8.5,37)(0,0 ) ( 21.5,13)(0,0)@xmath27 ( 8,7)(0,0)@xmath28 ( 9,63.5)(0,0)@xmath26 ( 11.5,0)(0,0)[b]a ( 68,38)(0,0 ) ( 68,0)(0,0)[b]b , , , @xmath142 @xmath26 @xmath143 . @xmath144 @xmath145 @xmath11 @xmath26 .
( [ v : sp1 ] ) : @xmath146 @xmath4 : @xmath147 @xmath148 ( @xmath149 , . [ s : p ] ) .
, ( . [ f : spiral]b ) : @xmath150 ( ( [ v : sp3 ] ) ) .
@xmath151 , @xmath152 , @xmath153 .
( ) @xmath154 , @xmath155 ( ( [ v : spv ] ) ) .
@xmath156 .
@xmath157 . , @xmath156 , : @xmath158 `` '' ( [ p : p2 ] ) , ( . [ f : hyp3]a ) , . : @xmath159 .
, @xmath160 .
. 0 ; . , , , . @xmath160 ( . [ f : cone ] ) .
@xmath161 , @xmath93 , @xmath162 .
@xmath163 .
, @xmath164 , @xmath113 @xmath165 .
@xmath156 @xmath166 @xmath167 .
@xmath168 ( [ m : tri ] ) @xmath169 .
1 ?
( [ p : cons ] ) @xmath170 : @xmath171 .
@xmath172 , @xmath173 @xmath174 @xmath175 .
@xmath176 : @xmath177 ; @xmath178 , @xmath179 @xmath180 , @xmath181 . . @xmath93 @xmath182 . @xmath183 . @xmath176 @xmath184 , @xmath185 @xmath186 : @xmath187 @xmath188 : @xmath189 .
@xmath190 , @xmath191 , @xmath192 , @xmath193 , @xmath194 . @xmath195 , @xmath196 , @xmath197 ,
@xmath198 . @xmath199 . @xmath199 @xmath200 ( @xmath201 ) ?
@xmath202 @xmath203 .
: @xmath204 , @xmath205 @xmath206 @xmath207 , .. @xmath208 @xmath209 . . @xmath210 @xmath184 .
@xmath211 , , @xmath212 .
( [ p : k ] ) @xmath213 , ( . [ f : wave]a ) .
, ( , , ( [ p : k ] ) ) , @xmath184 ( . [ f : wave]b ) .
@xmath214 ( ) , . , @xmath184 ( . [ f : wave]c ) .
( 110,33 ) ( 15,18)(0,0 ) .,title="fig : " ] ( 55,18)(0,0 ) .,title="fig : " ] ( 95,18)(0,0 ) .,title="fig : " ] ( 15,0)(0,0)[b]a ( 17,26)(0,0)@xmath184 ( 55,0)(0,0)[b]b ( 61,27)(0,0)@xmath184 ( 56,23)(0,0)@xmath11 ( 60,19)(0,0)@xmath11 ( 95,0)(0,0)[b]c ( 105,26)(0,0)@xmath184 @xmath124 @xmath215 .
( [ p : k ] ) @xmath216 @xmath217 .
( ) .
. ( ) , ( [ p : k ] ) @xmath184 , @xmath218 @xmath219 .
, , .
@xmath156 , @xmath220 .
, . , @xmath221 . ( @xmath222 ) , @xmath223 @xmath224 : @xmath225 @xmath223 @xmath224 @xmath226 @xmath227 .
@xmath4 , @xmath221 @xmath228 , @xmath229\ , , \end{split } \label{clock : t}\ ] ] @xmath230 , @xmath231 .
@xmath221 @xmath232 @xmath222 @xmath233 .
, @xmath221 @xmath222 , . , . .
@xmath26 , @xmath27 @xmath1 @xmath234 @xmath10 . [ f : l ] , @xmath235 , @xmath236 .
@xmath1 @xmath237 @xmath238 .
@xmath1 , , @xmath239 @xmath240 ; , @xmath241 , @xmath242 , ( . [ f : cone1 ] ) .
@xmath243 , @xmath244 @xmath10 2 ( . [ f : l2 ] ) @xmath245 , @xmath246 1 ( [ l : l0 ] ) : @xmath247 @xmath28 ( [ lf : x2 ] ) . ( [ l : l ] ) , . `` '' @xmath248 ( . [ f : hyp]a ) ; @xmath249 , `` '' @xmath250 ( . [ f : hyp]b ) @xmath251 ( [ m : sp ] ) @xmath256 @xmath94 , @xmath95 @xmath257 @xmath258 @xmath1 ( . [ f : lf ] ) : @xmath259 : @xmath260 ( 52,32 ) ( 26,16)(0,0 ) ( -1,20)(0,0)@xmath252 ( 5.7,13.3)(0,0)@xmath261 ( 44,20)(0,0)@xmath253 ( 54,30)(0,0)@xmath262 ( 21,23)(0,0)@xmath94 ( 29.3,24.7)(0,0)@xmath263 ( 43,1)(0,0)@xmath95 ( 44.7,9.3)(0,0)@xmath264 ( 33,22)(0,0)@xmath1 @xmath11 , @xmath265 ( . [ f : lf2 ] ) , @xmath1 ( [ lf : l ] ) . | ( ) . , , , , . . ( ) . , . @xmath0 , @xmath1 , @xmath2 . `` '' @xmath3/ ; . ( 19472002 ) , , . + + |
the galaxies of the local group ( lg ) and its immediate surroundings are of particular interest since their proximity permits us to study many aspects of galaxy evolution in great detail .
the galaxy census of the lg and its surroundings is still incomplete .
a number of new nearby dwarf galaxy candidates were discovered in recent years and have largely been confirmed ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
. for recent discussions and reviews of the properties of galaxies in the local group and its surroundings , see @xcite .
a recent review of the evolution and chemistry of dwarf irregular galaxies is given in @xcite .
as the closest _ bona fide _ dwarf group , the antlia - sextans group @xcite may hold clues to the star formation histories and chemical evolution of galaxies that presumably experienced little interaction and that evolved in relative isolation . in the present study , which is part of a larger effort to gain a comprehensive picture of dwarf galaxy evolution in different environments @xcite
, we measure nebular abundances in two of the dirr galaxies in antlia - sextans , namely sextans a and b ( see table [ tbl : general ] for basic properties of both galaxies ) .
we determine nebular abundances both for hii regions and for planetary nebulae ( pne ) in these galaxies .
the most massive , most luminous galaxy in antlia - sextans , the dirr ngc3109 , will not be discussed here ; for its nebular abundances , we refer to @xcite and @xcite and references therein . with the telescopes presently available , spectroscopic abundance determinations for individual stars can only be carried out for luminous , massive young supergiants beyond the lg , so stellar abundance estimates for older stellar populations have to rely on photometry ( see grebel , harbeck , & gallagher 2003 for results ) .
emission - line spectroscopy , on the other hand , can already be obtained with medium - sized telescopes .
emission - line spectra of hii regions provide a convenient means of obtaining present - day gas phase abundances in different areas of a galaxy .
emission - line spectroscopy of planetary nebulae makes abundances of older populations ( from a few 100 million to many billion years ) accessible .
sextans a is at a distance of 1.32 mpc @xcite .
deep hst imaging permitted photometric derivations of the recent and intermediate - age star formation and enrichment histories , but for ages of several gyr the uncertainties become large @xcite .
the best fit is obtained for the assumption of a roughly constant metallicity of @xmath41.4 dex ( with an rms scatter of 0.15 dex at all ages ) throughout the lifetime of sextans a @xcite . as is typical for dirr galaxies @xcite , star formation amplitude variations of factors of two to three occurred in the history of sextans a. the star formation rate inferred for sextans
a reveals recent increases starting a few gyr ago @xcite , whereas the galaxy appears to have been more quiescent at earlier times .
spectroscopic abundance determinations for three a supergiants yield @xmath5[fe / h]@xmath6 @xcite . the heavy element gas abundance for sextans a was first determined by ( * ? ?
* skh89 hereafter ) .
they presented only an empirical estimate of o / h ( @xmath7 log(o / h)@xmath8 ) , which has a characteristic rms uncertainty of @xmath350% .
one of the goals of our current study is to re - determine nebular abundances with higher accuracy .
sextans b is a dirr at a distance of 1.36 mpc @xcite . like in sextans a , the star formation history has likely been complex , with the long periods of relatively low star formation rates , punctuated by periods of short star - forming bursts @xcite . as far as its global properties
are concerned , sextans b is considered to be a `` twin '' of sextans a , but sextans b is of particular interest because the previous measurements of the nebular oxygen abundance for this galaxy by @xcite , skh89 and ( * ? ? ?
* mam90 hereafter ) differ considerably .
here we re - measure the hii region oxygen abundances with additional high quality spectroscopy and check whether the apparent discrepancy remains , and compare these abundances with the abundance of the candidate planetary nebulae ( pne ) found by @xcite .
the contents of this paper are organized as follows .
[ txt : obs_and_red ] gives the description of all observations and data reduction . in
[ txt : results ] we present our results , and discuss them in [ txt : disc ] .
the conclusions drawn from this study are summarized in
[ txt : summ ] . for the remainder of this paper ,
we adopt the revised solar value of the oxygen abundance , 12+@xmath1(o / h ) = 8.66 @xcite .
spectrophotometric observations of hii regions and pne in the dwarf irregular galaxies sextans a and sextans b were conducted with the new technology telescope ( ntt ) , a 3.5-meter telescope at the european southern observatory ( eso ) , la silla , on the nights of 2004 february 12 14 .
all spectral observations were made with the low - resolution long - slit spectroscopy ( rild ) mode of the emmi ( eso multi - mode instrument ) used in conjunction with two mosaic 2048@xmath94096 ccd detectors . with a pixel scale of 0166
the effective field of view was 99@xmath991 . using a binning factor of 2 ,
our final spatial sampling was 033 pixel@xmath10 .
we used a @xmath11 slit for all spectral observations .
grisms # 5 ( spectral range 3800 7000 , a final scale perpendicular to the slit of 1.63 pixel@xmath10 and a spectral resolution ( fwhm ) of @xmath37 ) and # 6 ( spectral range 5750 8670 , a final scale perpendicular to the slit of 1.44 pixel@xmath10 and a fwhm of @xmath36.5 ) were used to cover the spectral range with all the lines of interest .
the seeing during the observations was very stable and varied from night to night in the range of 04 to 06 .
total exposure times were 60 or 90 minutes for each grism .
each exposure was broken up into 23 subexposures , 30 minutes each , to allow for removal of cosmic rays .
spectra of he
ar comparison arcs were obtained to calibrate the wavelength scale .
the spectrophotometric standard stars feige 56 , gd 108 , gd 50 and g 60 - 54 @xcite were observed with a 10 slit width at the beginning , middle and end of each night for flux calibration .
one or two - minute h@xmath12 acquisition images were obtained before the spectroscopic observations in order to select an optimal position of the slit .
we tried to place the slit over the knots with maximal h@xmath12 flux to increase the probability of detecting the [ oiii ] @xmath134363 emission line , and thus being able to calculate the electron temperature directly . for sextans
a we use the hii region nomenclature published by @xcite .
one of our slits crossed the extended regions hii@xmath417 and 19 and the compact region hii@xmath420 ( see figure [ fig : sexa_ha ] for more details ) .
these hii regions correspond to the regions 1 , 3 , and 2 of skh89 . additionally , this slit position crossed the extended regions hii@xmath418 , 22 , and 24 @xcite .
thanks to the good seeing and resulting high angular resolution , @xmath305 , hii@xmath417 was well resolved into two compact bright knots with an angular separation of @xmath315 ( called hereafter hii@xmath417a and 17b ) , which seem to be located at the intersection of two shell - like structures ( figure [ fig : sexa_ha_b ] ) .
the second slit position went through the pn .
the coordinates of the pn are given in @xcite , and a finding chart was published in @xcite .
this pn is called hii@xmath413 in @xcite .
the second slit position also crosses hii@xmath42 and 7 in the west , hii@xmath415 in the central part of the galaxy , and the northern edge of hii@xmath424 in the east ( see also figure [ fig : sexa_ha ] ) . in sextans
b we adopt the hii region nomenclature suggested by @xcite . for the observed pn candidate in this galaxy
we use the nomenclature of @xcite , who named it pn3 .
one slit position went through hii@xmath41 , 4 , 5 , and the edge of 2 , and the other one crossed hii@xmath42 , 6 , and 11 ( see figure [ fig : sexb_ha ] for more details ) . the two - dimensional spectra were bias subtracted and flat - field corrected using iraf .
cosmic ray removal was done with filter / cosmic task in midas .
we use the iraf software routines identify , reidentify , fitcoord , transform to perform the wavelength calibration and to correct each frame for distortion and tilt . to derive the sensitivity curve
, we fitted the observed spectral energy distribution of the standard stars with a high - order polynomial .
all sensitivity curves observed during each night were compared and we found the final curves to have a precision better than 2% over the whole optical range , except for the region blueward of @xmath134000 where the sensitivity drops off rapidly .
all two - dimensional spectra obtained with the same grism for the same object were averaged .
one - dimensional ( 1d ) spectra were then extracted from averaged frames using the iraf apall routine . for the hii - regions 1d spectra
were extracted from areas along the slit , where @xmath14 1.5@xmath15 ( @xmath15 is the dispersion of the noise statistics around this line ) . for the pne ,
1d spectra were extracted to get the total light .
the resulting reduced and extracted spectra of the pne in sextans a and sextans b obtained with grism # 5 are shown in fig .
[ fig : sexa_pn_spec ] , and part of the reduced and extracted spectrum of the pn in sextans a obtained with grism # 6 is shown in fig .
[ fig : sexa_pn_spec_red ] .
the reduced and extracted 1d spectra of hii regions in sextans a , obtained with grism # 5 and in which the [ o iii ] @xmath134363 line was detected , are shown in fig .
[ fig : sexa_hii_spec ] . fig .
[ fig : sexb_hii_spec ] shows the reduced and extracted 1d spectra obtained with grism # 5 of hii regions in sextans b in which the [ o iii ] @xmath134363 line was detected .
after 1d - spectra were extracted , we used a method for measuring emission - line intensities described in detail in @xcite : ( 1 ) the software is based on the midas command language ; ( 2 ) the continuum noise estimation was done using the absolute median deviation ( amd ) estimator ; ( 3 ) the continuum was determined with the algorithm from @xcite ; ( 4 ) the programs dealing with the fitting of emission / absorption line parameters are based on the midas fit package ; ( 5 ) every line was fitted with the corrected gauss - newton method as a single gaussian superimposed on the continuum - subtracted spectrum .
some close lines were fitted simultaneously as a blend of two or more gaussian features : the h@xmath12 @xmath136563 and [ n ii ] @xmath166548,6584 lines , [ s ii ] @xmath166716,6731 lines , and [ o ii ] @xmath167320,7330 lines ; ( 6 ) the final errors in the line intensities , @xmath17 , include two components : @xmath18 , due to the poisson statistics of line photon flux , and @xmath19 , the error resulting from the creation of the underlying continuum and calculated using the amd estimator ; ( 7 ) all uncertainties were propagated in the calculation of abundances and are accounted for in the accuracies of the presented element abundances .
the emission lines he i @xmath135876 , [ o i ] @xmath136300 , [ s iii ] @xmath136312 , [ o i ] @xmath136364 , h@xmath12 @xmath136563 , [ n ii ] @xmath166548,6584 , he i @xmath136678 , and [ s ii ] @xmath166716,6731 were usually detected independently in both grism spectra .
their intensities and equivalent widths were averaged after applying weights inversely proportional to the errors of the measurements . the electron temperature , number densities , ionic and total element abundances for oxygen were calculated in the same manner as in @xcite , where the [ oii ] @xmath133727,3729 doublet was also not observed , and o@xmath20/h@xmath20 was calculated using intensities of the [ oii ] @xmath137320,7330 lines .
the contribution to the intensities of the [ oii ] @xmath137320,7330 lines due to recombination was not taken into account since this is negligible compare to other uncertainties ( namely , @xmath21 as follows from @xcite ) .
total element abundances for ne ,
n , s and ar were derived after correction for unseen stages of ionization as described in @xcite and @xcite : ( 1 ) the electron temperatures for different ions were either calculated from the relations in @xcite or using the equations following from the hii photoionization models of @xcite : @xmath22 , \label{eq : toii}\ ] ] @xmath23 @xmath24 @xmath25 @xmath26 @xmath27 @xmath28 where t@xmath29=t@xmath29/10@xmath30k .
( 2 ) the ionization correction factors icf(a ) for different elements were calculated using the equations from @xcite : @xmath31 @xmath32 @xmath33 @xmath34\}^{-1 } , \label{eq : icfs}\ ] ] @xmath35 @xmath36\}^{-1 } , \label{eq : icfar1}\ ] ] @xmath37^{-1 } , \label{eq : icfar2}\ ] ] where @xmath38 = o@xmath20/o . in those cases where the [ ar iv ] @xmath134740 line was not detected ,
equation [ eq : icfar2 ] was used instead of equation [ eq : icfar1 ] .
the he abundance was calculated following the method described in detail by @xcite and @xcite .
the line [ nii ] @xmath135755 was detected in the spectrum of the sextans a pn , which allowed us to determine @xmath39 directly .
this was done in the same way as for @xmath40 using the @xcite equation : @xmath41 10^{1.086/t_{\rm e}}\ ] ] here the constants @xmath42 , @xmath43 , and @xmath44 for different temperatures were taken from @xcite and interpolated using an iterative procedure .
@xmath45 is a term describing the dependence on the electron density @xmath46 and @xmath47 .
additionally , a strong nebular he ii @xmath134686 emission line is detected in the spectrum of the sextans a pn , implying the presence of a non - negligible amount of o@xmath48 . in this case
its abundance was derived from the relation taken from @xcite : @xmath49 after that , the total oxygen abundance is equal to @xmath50
the observed emission line intensities @xmath51 , and those corrected for interstellar extinction and underlying stellar absorption @xmath52 are presented in table [ tbl : sexta_hii_int ] for the three hii regions in sextans a , in which the [ oiii ] @xmath134363 line was detected : hii@xmath417a , hii@xmath417b and hii@xmath420 .
the h@xmath53 equivalent width @xmath54(h@xmath53 ) , the absorption equivalent widths @xmath54(abs ) of the balmer lines , the h@xmath53 flux , and the extinction coefficient @xmath55(h@xmath53 ) ( this is a sum of the internal extinction in sextans a and foreground extinction in the milky way ) are also listed there . in table
[ tbl : sextb_hii_int ] we present similar data for the three hii regions hii@xmath41 , 2 and 5 in sextans b. the range of @xmath55(h@xmath53 ) measured in the hii regions of sextans a corresponds to values of b - band extinction a@xmath56 ranging from 020 to 054 .
accounting for the milky way foreground extinction of 019 in the direction to this galaxy @xcite the internal extinction values are in the range of 001 to 035 , implying a low dust content in these hii regions . for sextans
b , the values of a@xmath56 calculated from the spectra for hii@xmath45 and 1 are consistent to within the uncertainty in the foreground extinction in the direction of this galaxy , 014 .
the a@xmath56 for hii@xmath42 corresponds to an additional internal extinction of 035 . in table
[ tbl : sexta_hii_ab ] we give the derived physical conditions and ion and element abundances for the above mentioned three hii regions in sextans a. the corresponding data for three hii regions in sextans b are presented in table [ tbl : sextb_hii_ab ] .
as evident from these tables , we detected the faint [ oiii ] @xmath134363 line at the level of 4 to 8 @xmath57 and the [ oii ] @xmath137325 line at the level of 4 to 10 @xmath57 .
this results in o / h uncertainties in the individual hii regions on the order of 0.05 to 0.11 dex .
since the flux uncertainties of faint lines of other ions are in general higher , the resulting errors of n / h , s / h and ar / h in the individual hii regions lie in the range of 0.1 to 0.2 dex .
the line [ siii ] @xmath136312 is not detected in hii@xmath417b of sextans a , and the derived value of s / h should therefore be treated as a lower limit .
hence we do not list s / h for this hii region .
the line [ neiii ] @xmath133868 is detected in only one hii region , due to the significant drop in sensitivity below @xmath3@xmath134000 .
all three hii regions in sextans a from table [ tbl : sexta_hii_ab ] show oxygen abundances consistent with each other to within the uncertainties , and hence a weighted average value of 12+log(o / h)[email protected] can be derived for this galaxy based on our data .
for the two hii regions in sextans b ( hii@xmath41 and hii@xmath42 from table [ tbl : sextb_hii_ab ] ) , the oxygen abundances also are very similar and within their rather small rms uncertainties can be treated as equal .
their weighted mean o / h value corresponds to 12+@xmath1(o / h)[email protected] .
the average relative abundances n / o , s / o and ar / o for all three hii regions in sextans a ( @[email protected] , @[email protected] , and @[email protected] respectively ) and for two low - metallicity hii regions in sextans b ( @[email protected] , @[email protected] , and @[email protected] ) are quite consistent with the average values for the most metal - poor hii galaxies obtained by @xcite and confirmed in many subsequent observations ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , implying production in massive stars .
the relative abundances of n / o , ne / o , s / o and ar / o for the third hii region in sextans b analyzed here are also consistent with the mean values determined by @xcite for their high - metallicity subsample of hii galaxies .
we present the observed and balmer line extinction corrected emission line intensities for the two observed pne in table [ tbl : sextab_pn_int ] .
our fluxes for the emission lines [ o iii ] @xmath135007 and h@xmath12 in both pne are consistent to within the observational uncertainties with the fluxes measured by @xcite and @xcite .
the measured @xmath55(h@xmath53 ) in the sextans a and sextans b pne corresponds to a@xmath56 of 066 and 071 , respectively ( a@xmath58 of 050 and 054 ) . after accounting for the galactic contribution
, this suggests an intrinsic extinction a@xmath59 of 047 and 057 , which is higher than the values for hii regions in these galaxies .
this result is consistent with the fact that the progenitors of pne produce additional dust in the late stages of their evolution . in table
[ tbl : sextab_pn_ab ] we give the derived physical conditions , ion and element abundances for both pne . for the pn in sextans
a the quality of the spectral data is comparable to the quality of the hii region spectra .
the resulting uncertainties in he / h , o / h and n / h are at the level of 0.030.05 dex . the values of o / h and n / h for the pn appear to be significantly higher than observed in the hii regions of this galaxy ( by a factor of @xmath33 and @xmath3250 , respectively ) . with @xmath1(n
/ o ) = 0.39 and he / h = [email protected] , according to the definition of ( * ? ? ?
* @xmath1(n / o)@xmath600.1 ) or to the definition in ( * ? ? ?
* @xmath1(n / o)@xmath600.6 and he / h @xmath14 0.10 ) , the pn in sextans a should be classified as a type i. since the line fluxes of pn3 in sextans b are about a factor of 23 ( apart [ nii ] ) lower than for the sextans a pn , the resulting signal - to - noise for the spectrum of pn3 in sextans b is significantly less .
therefore , from this spectrum it was only possible to derive values of o / h and n / h with final uncertainties of 0.16 and 0.18 dex , respectively .
the observed properties of pn3 allow us to exclude a classification as a pn of type i. however , the quality of the pn spectrum is not sufficient to perform the detailed comparison of its abundances with those of the hii regions in sextans b. the self - consistent age - dating of pne progenitors requires the determination of the bolometric luminosity and the effective temperature of the pn central star in combination with photoionization modelling and pn evolutionary tracks @xcite .
this is well beyond the scope of our paper .
however , based on our data , it is possible to make some rough estimates , which we summarize in table [ tbl : pn_par ] .
the effective temperatures ( t@xmath61 ) of the pn progenitors can be estimated by the method developed by @xcite . since the method is applicable for the case of optically thick nebulae , we have to first check whether our pne are optically thick .
the criteria of this were suggested by @xcite on the results of analysis of the dereddened i(3727)/i(h@xmath53 ) , i(4686)/i(h@xmath53 ) and i(6584)/i(h@xmath12 ) line ratios for the large sample of pne covered a wide range of optical thickness . according to these criteria , the dereddened line ratios i(4686)/i(h@xmath53 ) and i(6584)/i(h@xmath12 ) that reflect the zanstra temperature ratio tr = t@xmath62(heii)/t@xmath62(h ) @xcite , show that our pne are certainly optically thick , since : ( a ) pn in sextans a has [email protected] and is the case 1 following @xcite classification ( class a i from * ? ? ? * ) ; ( b ) pn3 in sextans b has [email protected] and i(4686)@xmath640.9 and is the case 2 following @xcite classification ( class a ii from * ? ? ?
* ) . from the spectrum of pn3 in sextans
b we estimated a 2@xmath15 upper limit for the flux of heii @xmath134686 line as 8% of i(h@xmath53 ) . after that , for the calculation of the zanstra temperature we used the equation suggested by @xcite : @xmath65 @xmath66 @xmath67 where @xmath68 is the flux of heii @xmath134686 in the units of i(h@xmath53 ) with i(h@xmath53 ) = 100 .
the total luminosities of the pn central stars were derived from the relation given in @xcite , using the h@xmath53 absolute fluxes , extinction @xmath55(h@xmath53 ) and distances : @xmath69 masses were derived from the theoretical evolutionary tracks of @xcite for z=0.001 ( 1/20 of z@xmath70 ) .
the ages were derived using the evolutionary lifetimes of the various phases of the progenitor stars from @xcite , also for z=0.001 .
in addition , since the pne 1d spectra were extracted taking the total light of a nebula , we synthesized the @xmath71 colors of the pne using zeropoints from the spectrum of vega @xcite .
these colors were transformed into the standard johnson - cousins @xmath72 system as specified by @xcite and @xcite .
the @xmath73-band was not included in table [ tbl : pn_par ] since , due to the drop in sensitivity , the spectra become very noisy below @xmath3@xmath134000 .
hii region abundances mainly provide information about @xmath12-process elements , produced predominantly in short - lived massive stars . in contrast to hii regions ,
some elemental abundances in pne are affected by nucleosynthesis in the pn progenitor stars .
it is well known that newly synthesized material can be dredged up by convection in the envelope , significantly altering the abundances of he , c , and n in the surface layers during the evolution of 18 m@xmath74 stars on the giant branch and asymptotic giant branch ( agb ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in addition , if during the thermally pulsing phase of agb evolution convection `` overshoots '' into the core , significant amounts of @xmath75o can be mixed into the inter - shell region and may be convected to the surface @xcite . in combination ,
all of the above factors mean that only the ne , s , and ar abundances , observed in both hii regions and pne , can be considered as reliable probes of the enrichment history of galaxies , unaffected by the immediately preceding nucleosynthesis in the progenitor stars .
therefore , in further discussions of pn progenitor metallicity we do not use o / h , he / h , and n / h as reliable metallicity indicators and base our conclusions on chemical evolution only on pne abundances of ne , s , and ar .
the value of 12+log(o / h)[email protected] , derived here as the average of three hii regions in sextans a , is in good agreement with the earlier derived value 12+log(o / h)[email protected] from skh89 .
however , the previous value was based on spectra of four separate hii regions with only a marginal detection of [ oiii ] @xmath134363 in one of them , such that skh89 used a combined estimate for all four hii regions based on the strong - line empirical relation of @xcite .
it is worth noting
that the use of modern empirical calibrations of the strong - line relation @xcite for the published data from skh89 gives 12+@xmath1(o / h ) in the range of 7.7 to 8.0 .
recently @xcite measured the present - day stellar metallicity of sextans a using three a - type supergiants with ages of @xmath310 myr .
their average @xmath12-element abundance , derived from mg i lines , relative to the solar value is @xmath64[@xmath12(mg i)/h]@xmath14 = 1.09 @xmath2 0.02 @xmath20.19 , which corresponds to an oxygen abundance of 12+@xmath1(o / h ) = 7.57 dex .
this is in excellent agreement with our average o / h for the three hii regions , which are roughly located in the same part of sextans a. thus , the element abundances in these hii regions and in the three a supergiants located in other parts of sextans a ( see figure 1 from * ? ? ?
* ) do not provide any evidence of metallicity inhomogeneity in this galaxy . since the abundances of the @xmath12-elements s and ar in the sextans a pn should not be affected by nucleosynthesis in its progenitor star , these elements can be used as good tracers of the metallicity of the interstellar gas in sextans a at the epoch when the pn progenitor was formed . according to our estimate
, this occurred some 1.6 gyr ago ( table [ tbl : pn_par ] ) . due to the limited accuracy of s / h and ar / h in the hii regions , and especially in the pn ( in the latter the errors are on the order of about 40%
50% ) , the quantitative result on the rate of the interstellar medium ( ism ) enrichment needs further improvement .
however , the values for relative enrichment , derived independently for sulfur , ( s / h)@xmath76/(s / h)@xmath77 = 3.17 ( @xmath78)@xmath21.05 , and argon , ( ar / h)@xmath76/(ar / h)@xmath77 = 2.99 ( @xmath79)@xmath20.54 , are consistent with each other .
we give the estimates of the uncertainties of these ratios in the above form since , due to the large relative errors of the values in the denominators , 1@xmath15 intervals are highly non - symmetric ; the values in parenthesis give the uncertainties in the denominators , while the last term describes the uncertainties in the numerators .
the weighted average value is @xmath33.1@xmath80 or equivalently , [email protected] dex .
this rather large enrichment in @xmath12-elements may be caused by the increase in the star formation rate during the last @xmath312 gyr .
when we transform the above values to o / h and take the @xmath12-element enrichment into account , we find 12+@xmath1(o / h)@xmath81 = [email protected] as the value prior to enrichment .
the comparison of element abundances in three hii regions and in the pn suggests that the region of sextans a where the pn progenitor formed was as metal - deficient as i zw 18 is now @xcite .
moreover , assuming the element abundance pattern in the gas of which the pn progenitor formed to be typical of very low - metallicity hii galaxies , we conclude that the progenitor star , with an estimated mass of 1.5m@xmath74 , enriched the material of this pn by a factor of @xmath310 in oxygen ( from the value of 12+@xmath1(o / h)=7.04 to the current value of 8.02 ) , and by a factor of @xmath3750 in nitrogen . previously , @xcite observed , according to their description , hii@xmath45 ( although the object indicated on their finding chart seems to be closer to hii@xmath410 ) .
they did not detect [ oiii ] @xmath134363 and presented a lower limit on o / h , equivalent to 12+@xmath1(o / h)@xmath827.4 .
later , four hii regions ( hii@xmath41 , 2 , 5 , and 10 ) were observed by skh89 .
these authors failed to detect [ oiii ] @xmath134363 . to determine o / h
, they used the empirical strong lines relation ( which has an internal accuracy of 50% ) by @xcite and derived a lower limit for o / h in hii@xmath45 .
their resulting value for the oxygen abundance of sextans b was 12+@xmath1(o / h)=7.56 .
subsequently , mam90 detected [ oiii ] @xmath134363 in the spectrum of hii@xmath45 at a level of [email protected] of i(h@xmath53 ) and derived its o / h as 12+@xmath1(o / h)=8.12 , with no cited uncertainty .
we estimate that their rms is @xmath83 dex .
it is worth noting here that the empirical value of skh89 for hii@xmath45 is not in good agreement with the value derived via the t@xmath0-method in mam90 . in his compilation , @xcite adopted an average value of 12+@xmath1(o / h)[email protected] for sextans b. in our observations of sextans b hii regions , the temperature - sensitive line [ oiii ] @xmath134363 was detected , and hence o / h was reliably measured , in only three of them . for hii@xmath41 and hii@xmath42 , which are next to each other and which are situated at the western part of the brighter central one - kpc region ( figure [ fig : sexb_ha ] ) ,
the oxygen and other element abundances are very similar . for the third hii region ( hii@xmath45 in figure [ fig : sexb_ha ] , at @xmath30.6 kpc ne ) the oxygen abundance ( 12+@xmath1(o / h)[email protected] ) is larger than for the former two by a factor of [email protected]
. the ratios of the other elements are similar : ( n / h)@xmath84/(n / h)@xmath85 , ( ar / h)@xmath84/(ar / h)@xmath86 , and ( s / h)@xmath84/(s / h)@xmath87 .
all these ratios are consistent to within their rms uncertainties .
their mean abundances imply an enrichment by a factor of @xmath32.5 relative to the element abundances in hii@xmath41 and 2 .
these differences are significant for each of the four elements at a confidence level corresponding to 4 @xmath88 , where @xmath88 is the combined rms error of the difference ( on a linear scale ) between the abundances in hii@xmath45 and the average on hii@xmath41 and 2 .
we have checked whether our results for hii@xmath45 are consistent with the o / h value from mam90 . for this
we recalculated their o / h in our system , using their relative line intensities , and obtained 12+@xmath1(o / h)=8.06 .
this is consistent with our value to within @xmath32@xmath88 .
the main difference comes from our higher ( by a factor of 1.5 ) relative intensity of [ oiii ] @xmath134363 , which is detected in our spectrum at an 8@xmath15 level compared to a 4@xmath15 detection in mam90 .
therefore , we consider our element abundances to be more reliable .
we also note that the apparent problem with the small value of @xmath1(n / o ) = @xmath41.8 in this hii region that was pointed out by @xcite from the data of mam90 , does not exist according to our data .
while the disagreement between the o / h values of skh89 and mam90 could ( in principle ) be considered as an indication of o / h inhomogeneity , our data provide the first _
firm _ evidence of such chemical inhomogeneity in the interstellar medium of sextans b on scales of @xmath30.51 kpc .
in contrast to the pn in sextans a , the accuracy of the o / h and n / h ratios measured in pn3 in sextans b allows us to formulate only very preliminary conclusions .
o / h and n / h in this pn are very close to the average values derived for sextans b hii regions 1 and 2 ( the difference is @xmath890.5 @xmath88 ) .
the measured relative abundance log(n / o ) = [email protected] for this pn is well consistent with the average values for the most metal - poor hii galaxies @xcite .
this suggests that the enrichment in o and n is sufficiently small and in general does not exceed the uncertainty of log(n / o ) . adopting the 95% confidence level for the possible o and n abundances in pn3 ,
we conclude that the maximum o and n enrichment since the epoch of the pn progenitor s formation does not exceed @xmath30.4 dex . summarizing the results on the element abundances in sextans
b , we conclude that our new spectroscopy confirms earlier conclusions ( based on lower accuracy measurements ) that some of the nearby dirr galaxies reveal considerable abundance inhomogeneities among their hii regions .
this is an important finding since dirr galaxies are usually believed to be chemically well - mixed and homogeneous at a given age . to confidently determine the average ism metallicity in a well - resolved dirr galaxy one thus needs to measure its abundances in several regions .
various processes ( such as metal - enriched outflows , self - enrichment on a time - scale of a few myr , or infalling metal - poor intergalactic hi clouds ) can in principle be responsible for the local deviations of the ism metallicity from some average galactic value
. they may result in either an excess or a deficiency as compared to the galactic mean .
note that in the lower - mass dwarf spheroidal galaxies there are indications for chemical inhomogeneities in terms of possible radial gradients among their old stellar populations @xcite .
the available data on hii regions of sextans b , taken at face value , are inconclusive in terms of constraining the probable cause of the inhomogeneities .
it is unfortunate that older tracers of the chemical evolutionary history of galaxies like sextans b are so difficult to measure . for the time
being , only photometric estimates are available for the old population(s ) , based either on isochrones or on globular cluster red giant branch fiducials . for sextans b it would be useful to observe more hii regions at a sufficient depth to detect the [ oiii ] @xmath134363 line , in order to explore what level of present - day metallicity is characteristic of the entire galaxy .
the other three hii regions visible in our long - slit spectra will require observations with an 8-m class telescope to reach an adequate signal to noise level .
it is worth noting that the census of hii region metallicities in dirr galaxies in the lg and its surroundings is rather unsatisfactory . only in the magellanic clouds
can the number and spatial extent of hii regions with measured element abundances be considered sufficient for investigations of the chemical inhomogeneity problem ( see , e.g. , * ? ? ?
* ; * ? ? ?
but even in the smc , the situation could be improved ; while in the former study 19 hii regions were studied across much of the radial extent of the clouds , @xmath35@xmath90 ( @xmath35 kpc ) , the uncertainties of the individual o / h values seem too large to exclude possible variations at the level of @xmath30.3 dex ( this level of variations we see in sextans b ) . in the latter study the errors of o / h are reasonably small , and they are consistent for all three measured hii regions to within 20% ( or 0.08 dex ) , but all three regions are located close to each other within a circle @xmath320 pc in diameter . for most of the other nearby dirr galaxies , o / h values are based on old data , generally of poor to moderate signal to noise for [ oiii ] @xmath134363 .
alternatively ( or in addition ) , in many cases only one to two hii regions per galaxy were measured ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
thus the existing data for most nearby dirr galaxies do not allow one to draw a definite conclusion on chemical inhomogeneities in the ism at a level of @xmath30.3 dex across the entire galaxy .
the general claim by @xcite that there is `` no evidence for significant dispersion of oxygen abundances in any local group dirrs in which multiple hii regions have been studied '' should therefore be treated with some caution ; this statement probably applies only to variations with an amplitude of @xmath30.4 dex or higher .
we are aware of only a few similar occurrences in nearby dirr galaxies : in the lg dirr galaxies smc and wlm and in the m 81 group dwarf holmberg ii .
star clusters in the smc appear to differ in their metallicity at a given age @xcite , which seems to indicate differences in the chemical composition in different regions of this galaxies at a given time . in wlm
, @xcite discovered that one of two studied blue supergiants has a clearly higher [ o / h ] than the value measured for the hii region wlm hm@xmath47 , situated 300 pc away from this star .
the supergiant exceeds the [ o / h ] of the ism by @xmath910.7 dex at a 3@xmath15 level , a curious difference for which an explanation still needs to be found @xcite . for holmberg
ii we estimated from the data of @xcite that the weighted average of 12+@xmath1(o / h ) for four hii regions where [ oiii ] @xmath134363 was measured is [email protected] , while for the hii@xmath46 region the 2@xmath15 lower limit is indicated as 7.97 ; in this galaxy several hii regions show low abundances , while one appears to be enriched by a factor of @xmath32 .
@xcite , in their analysis of dispersal and mixing of oxygen in the ism of gas - rich galaxies , gave a fairly low characteristic time of 1.5 myr for local mixing of freshly ejected nucleosynthesis products due to raleigh - taylor and kelvin - helmholtz instabilities in star - forming regions .
while these processes are likely to be very efficient in mixing freshly released processed heavy elements into the ambient ism , detailed numerical simulations would be desirable that take the specific conditions in our target galaxies into account . in their analysis of a large sample of hii galaxies , @xcite
also conclude that some self - enrichment should play a role in these objects in order to match all observed correlations .
galaxies of different types generally follow a trend of increased global metallicity with increased luminosity , although offsets exist between different galaxy types ( e.g. , * ? ? ?
* ggh03 hereafter ) .
the lower o / h in two hii regions of sextans b gives a better agreement with the empirical relation between o / h and the blue luminosity for dwarf irregular galaxies ( e.g. , skh89 ) , while the higher value of o / h in the hii@xmath45 region shifts sextans b well above the 12+@xmath1(o / h ) vs @xmath92 relation .
thus , the hypothesis of a significant overabundance in the region hii@xmath45 , related to localized metal pollution , sounds more realistic .
however , as discussed earlier , so far such cases appear to be quite rare .
this is in some aspects similar to the detection of a significant nitrogen overabundance ( on the spatial scale of @xmath350 pc ) in the central starburst of a nearby dwarf starburst galaxy , ngc 5253 @xcite , again supporting very short characteristic timescales of local enrichment and subsequent dispersal as suggested by @xcite .
this may then imply that we caught the evolution of hii@xmath45 during the short period when fresh @xmath12-elements were just mixed into the @xmath310@xmath93 k medium via stellar winds and supernova ejecta , but prior to their dilution via the mechanisms discussed , e.g. , by @xcite and @xcite .
if this is the case , the detection of chemical inhomogeneities in the star - forming regions across the body of a dwarf galaxy is facilitated by the high spatial resolution of our study ( @xmath36 pc ) , which avoids the effective smearing of the localized short time - scale enrichment in more distant actively star - forming galaxies due to limited angular / spatial resolution .
obviously , this also implies that galaxies with many hii regions offer a better chance to detect inhomogeneities than dirrs with only one or two hii regions . in order to better understand possible reasons for the observed metal excess in hii@xmath45
, more detailed and higher angular resolution studies of this region would be helpful .
if the excess is indeed related to the recent effective mixing of the hii region s gas with matter from a hot bubble , some tracers of such a process could be imprinted in the ionized gas kinematics , or in additional ionization and excitation compared to more typical cases .
how do the spectroscopic results compare with the photometric enrichment history derived by @xcite ? using synthetic color - magnitude diagram ( cmd ) techniques , these authors found a fairly constant average stellar metallicity throughout the galaxy s history of [ fe / h ] @xmath91 @[email protected] .
they suggest that most of the enrichment occurred more than 10 gyr ago . for the old red giant branch ggh03 quote an approximate photometric mean metallicity of @xmath94 dex , estimated from comparison with globular cluster red giant branch fiducials .
@xcite s star formation history implies that more than half of the stellar mass formed during that early period .
however , @xcite also caution that the early star formation history can not be accurately inferred from the available hst data .
approximately 10 gyr ago the average star formation ( sf hereafter ) rate dropped for a period of @xmath37.5 gyr by a factor of @xmath320 with only about 5% of all stellar mass forming during that intermediate - age period . presumably , the low sf intensity would have been accompanied by little chemical enrichment . about 12 gyr ago
the sf again increased drastically , resulting in the formation of about 30% of the stellar mass of sextans a. this latter episode of active star formation should have been accompanied by a significant production of metals .
the present - day spectroscopic metallicity of [ fe / h ] @xmath91 @xmath41.1 dex ( from hii regions and young a supergiants ) is consistent with an enrichment of about 0.4 dex compared to the average stellar metallicity from @xcite , and with an enrichment of @xmath95 dex compared to the mean metallicity of the old red giants in sextans a @xcite .
one can not exclude , however , that due to the powerful energy release over the period of increased star formation during the past 12 gyr a significant fraction of freshly synthesized heavy elements was blown away .
thus , the metallicity in the young population and ism may indicate only a lower limit of the true metal production during this period .
in the above scenario , the results for the star formation history and metal enrichment of sextans a , derived from photometric data , are compatible with our conclusion of [email protected] dex enrichment according to the comparison of the hii region abundances and the data on s and ar abundances in the sole pn of sextans a. as we show in table [ tbl : pn_par ] the pn progenitor was a low - mass star with @xmath31.5 m@xmath70 and an age of @xmath31.6 gyr , i.e. , it formed at that time when the star formation rate and the accompanying enrichment showed a marked increase , according to @xcite . for sextans
b , the deepest published cmd is based on hst data @xcite but only covers the upper red giant branch .
hence no attempt was made to derive the star formation history of sextans b from these data .
however , the cmd of @xcite allows us to draw the following qualitative conclusions : sextans b possesses a substantial intermediate - age to old population with ages exceeding @xmath96 gyr as evidenced by its very well populated , prominent red giant branch .
our estimate of the age of the observed pn progenitor ( see table [ tbl : pn_par ] ) also supports the idea that sextans b had increased sf about 6 gyr ago .
ggh03 suggest a mean photometric metallicity of @xmath97 dex for the old population through comparison of the red giant branch with globular cluster fiducials .
the cmd of @xcite also shows the presence of a large number of luminous asymptotic giant branch stars above the tip of the red giant branch , and an extended group of young blue main - sequence stars and blue - loop stars .
red supergiants appear to be present as well .
to summarize , the cmd shows the typical features also known from many other dirr galaxies , which support the scenario of star formation continuing over a hubble time ( e.g. , grebel 1999 ) . in earlier work based on much shallower ground - based data
, @xcite studied the recent star formation history of sextans b up to 1 gyr and found moderate star formation activity , probably intermittent .
they provide a rough photometric estimate of the young population s stellar metallicity of @xmath98 dex . while they claim that the oxygen abundance in the hii@xmath410 region of 12+@xmath1(o / h)=8.1 ( mam90 ) is in the perfect agreement with their stellar metallicity estimate , this is no longer correct . in the modern calibration
the above oxygen abundance corresponds to [ o / h][email protected] .
the current ism metallicity , as measured here by o / h in the two low - metallicity hii regions , corresponds to [ o / h ] = @[email protected] .
the latter , however , is in agreement with the estimate of the young population s stellar metallicity from @xcite .
we estimated in section [ txt : disc_abun_sextb ] that the maximum o and n enrichment during the last few gyr does not exceed @xmath30.4 dex for sextans b. if this is true , one can suggest that most of the enrichment occured in sextans b more than 6 gyr ago , as would also be suggested by the very low metallicity of the old red giant branch in sextans b derived by ggh03 .
the qualitative star formation history of sextans b over the total cosmological time , based on all previously available cmd data is summarized by @xcite , @xcite , and @xcite .
it may have shown a pronounced early star formation rate during the first few gyr with a subsequent period of possibly lower star formation activity and a potential increase at more recent times ( say , 12 gyr ago ) .
it should be emphasized that there is still very little quantitative information on the star formation history of sextans b.
the general similarity of the sf histories of both sextans a and b and other nearby dwarf irregulars is important in the aspect of dwarf galaxy evolution in small galaxy groups .
this likely implies that the sf drivers are similar in the lg and this particular dwarf group .
another question concerns the nature of significant differences in sf histories during the last several gyr ; since there appears to be little correlation between the sf histories of sextans a and b , it is tempting to suggest that the sf during this period was mainly due to intrinsic processes , unrelated to the tidal action from a nearby neighbor .
indeed , the estimate of tidal forces between the two dwarfs , based on their projected distance of @xmath3250 pc , total masses of @xmath3(48)@xmath910@xmath99 m@xmath70 and holmberg radii of @xmath30.80.9 kpc @xcite results in a relative strength of tides of the order of 10@xmath100 .
however , this issue needs more careful study including the morphology and kinematics of gas and stars in these galaxies , and more importantly , constraints on their orbits . while the sf histories of dwarf galaxies of the same morphological type exhibit similar overall properties and trends , they differ in the details
. indeed , each dwarf has its own unique evolutionary history ( e.g. , grebel 1997 ) .
these differences become most pronounced on scales of a fraction of a gyr , but may encompass time scales of several gyr . in the absence of detailed , deep cmds and detailed information on the chemical evolution of such galaxies , the pn census may serve as a possible indicator . since the absolute magnitudes @xmath101 of sextans a and b are very similar ( see table [ tbl : general ] ) , and since their global sf histories appear to be similar , we should then expect that both galaxies also have a comparable number of pne .
instead , surveys for pne in the two dirrs have revealed a difference by a factor of 5 in the pn census .
this may imply that sextans b experienced a significantly higher sf rate during the aforementioned time interval .
similar conclusions the likely differing sf histories in sextans a and b during last few gyr were reached by @xcite .
in this paper we presented new determinations of the chemical abundances of six hii regions in sextans a and b , and of one pn in each of these galaxies . based on the data and discussion presented in the paper ,
the following conclusions can be drawn : \1 .
we confirmed with good accuracy the low ism metallicity of both galaxies , with 12+@xmath1(o / h ) = [email protected] in three hii regions in sextans a , and 12+@xmath1(o / h ) = [email protected] in two hii regions in sextans b. the element abundance ratios of o , n , s , and ar are well consistent with the expected patterns of very metal - poor hii galaxies .
new high accuracy chemical abundances in the hii regions of sextans a and b allowed us to probe the present - day metallicity homogeneity across the bodies of the respective galaxies .
while our statistics are still poor ( three hii regions per galaxy ) , we find that in sextans a the measured abundances in all three hii regions show no differences exceeding 0.1 dex ( at the 63% confidence level ) .
moreover , the metallicities of three a - supergiants studied by @xcite are in good agreement with those of the hii regions . \3 . in sextans
b one hii region ( hii@xmath45 ) is significantly enriched , with an excess in o , n , s , and ar of a factor of [email protected] relative to the mean value of the two other hii regions studied here .
this is strong evidence for chemical inhomogeneity in a dirr galaxy .
whether this implies a general chemical inhomogeneity among populations of comparable age in sextans b , and thus a metallicity spread at a given age , or whether we happen to see the short - lived effects of freshly ejected nucleosynthesis products prior to their dispersal and mixing with the ambient interstellar medium will require further study .
chemical abundances derived for the pn in sextans a show that this is a type i object with a highly elevated nitrogen abundance : n / o @xmath32.5 .
its o / h is about a factor of 3 higher than in the hii regions , which implies significant self - pollution by the pn progenitor .
the abundances of s and ar indicate that the ism metallicity was @xmath30.5 dex lower at the time of the formation of the pn progenitor , compared to that currently measured in hii regions . in this case
the pn progenitor enriched the material by a factor of @xmath310 in oxygen , and by a factor of @xmath3750 in nitrogen .
the element abundances of pn3 in sextans b are consistent with those of the two hii regions with low metallicity .
comparison of the element abundances for this pn and the two nearby hii regions implies that the maximum o and n enrichment during the last few gyr ( the estimated age of the progenitor star of pn3 ) did not exceed @xmath30.4 dex . \6 . despite the overall similarity of sextans
a and b , their star formation history and enrichment history over the last few gyr looks quite different , as evidenced by both the number of hii regions and pne , and their abundances . the overall chemical enrichment experienced by sextans a and b
as judged from the estimated mean metallicity of their old red giant branch populations and their hii regions has spanned at least 0.8 dex .
this number is a lower limit since it refers to the mean red giant metallicity and necessarily neglects the still unknown metallicity of the most metal - poor giants in the two dirrs .
the authors are pleased to thank l.a .
pustilnik for consultations on plasma instabilities .
we thank also the anonymous referee for useful comments and suggestions .
s.a.p . and a.g.p .
acknowledge financial support and hospitality of mpia during part of this work .
ekg was supported by the swiss national science foundation through the grants 200021 - 101924/1 and 200020 - 105260/1 .
this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
lcc alternate names & ugca 205 , ddo 75 & ugc 5373 , ddo 70 + @xmath12 ( j2000.0 ) & 10@[email protected]@xmath104 & 10@[email protected]@xmath104 + @xmath105 ( j2000.0 ) & @xmath404@[email protected]@xmath108 & + 05@[email protected]@xmath108 + distance ( kpc ) & 1320@xmath240@xmath109 & 1360@xmath270@xmath110 + v@xmath74,opt ( km s@xmath111 ) & 328@xmath25@xmath112 & 301@xmath21@xmath113 + v@xmath74,radio ( km s@xmath111 ) & 325@xmath23@xmath30 & 303@xmath22@xmath114 + b@xmath115 ( mag ) & 11.86@xmath116 & 11.85@xmath110 + a@xmath59 ( mag)@xmath117 & 0.19 & 0.14 + ( m - m@xmath118 ) ( mag ) & 25.62@xmath116 & 25.67@xmath116 + m@xmath59 & @xmath119 & @xmath120 + lcccccc 4101 h@xmath105 & [email protected] & [email protected] & & & & + 4340 h@xmath121 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 4363 [ o iii ] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 4861 h@xmath53 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 4959 [ o iii ] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 5007 [ o iii ] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 5876 he i & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 6300 [ o i ] & [email protected] & [email protected] & & & & + 6312 [ s iii ] & [email protected] & [email protected] & [email protected] & [email protected] & & + 6548 [ n ii ] & [email protected] & [email protected] & & & & + 6563 h@xmath12 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 6584 [ n ii ] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 6678 he i & [email protected] & [email protected] & [email protected] & [email protected] & & + 6717 [ s ii ] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 6731 [ s ii ] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 7065 he i & [email protected] & [email protected] & & & & + 7136 [ ar iii ] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 7325 [ o ii ] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + & & & & & & + c(h@xmath53 ) dex & [email protected] & [email protected] & [email protected] + ew(abs ) & [email protected] & [email protected] & [email protected] + ew(h@xmath53 ) & 2c 43@xmath2 1 & 2c16@xmath21 & 2c13@xmath21 + f(h@xmath53 ) & [email protected] & [email protected] & [email protected] + lcccccc 3868 [ ne iii ] & [email protected] & [email protected] & & & & + 3889 he i + h8 & [email protected] & [email protected] & [email protected] & [email protected] & & + 3967 [ ne
iii ] + h7 & [email protected] & [email protected] & & & & + 4101 h@xmath105 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 4340 h@xmath121 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 4363 [ o iii ] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 4471 he i & [email protected] & [email protected] & & & [email protected] & [email protected] + 4740 [ ar iv ] & [email protected] & [email protected] & & & & + 4861 h@xmath53 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 4959 [ o iii ] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 5007 [ o iii ] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 5876 he i & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 6300 [ o i ] & & & & & [email protected] & [email protected] + 6312 [ s iii ] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 6548 [ n ii ] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 6563 h@xmath12 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 6584 [ n ii ] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 6678 he i & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 6717 [ s ii ] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 6731 [ s ii ] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 7065 he i & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 7136 [ ar iii ] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 7325 [ o ii ] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + & & & & & & + c(h@xmath53 ) dex & [email protected] & [email protected] & [email protected] + ew(abs ) & [email protected] & [email protected] & [email protected] + ew(h@xmath53 ) & 2c60@xmath21 & 2c70@xmath21 & 2c66@xmath21 + f(h@xmath53 ) & [email protected] & [email protected] & [email protected] + lccc @xmath122(oiii)(k ) & 15,000@xmath21400 & 19,460@xmath22180 &
19,040@xmath22750 + @xmath122(oii)(k ) & 13,760@xmath21230 & 15,520@xmath21620 &
15,390@xmath22070 + @xmath122(siii)(k ) & 14,160@xmath21170 & 17,860@xmath21800 &
17,500@xmath22280 + @xmath123(sii)(@xmath124 ) & 10@xmath210 & 35@xmath225 & 390@xmath290 + & + o@xmath125/h@xmath125(@xmath910@xmath113 ) & [email protected] & [email protected] & [email protected] + o@xmath126/h@xmath125(@xmath910@xmath113 ) & [email protected] & [email protected] &
[email protected] + o / h(@xmath910@xmath113 ) & [email protected] & [email protected] & [email protected] + 12+log(o / h ) & [email protected] & [email protected] & [email protected] + & + n@xmath125/h@xmath125(@xmath910@xmath116 ) & [email protected] & [email protected] & [email protected] +
icf(n ) & 2.19 & 1.87 & 2.16 + n / h(@xmath910@xmath116 ) & [email protected] & [email protected] & [email protected] + 12+log(n / h ) & [email protected] & [email protected] & [email protected] + log(n / o ) & [email protected] & [email protected] & [email protected] + & + s@xmath125/h@xmath125(@xmath910@xmath116 ) & [email protected] & [email protected] &
[email protected] + s@xmath126/h@xmath125(@xmath910@xmath116 ) & [email protected] & [email protected] & + icf(s ) & 1.22 & 1.19 & 1.22 + s / h(@xmath910@xmath116 ) & [email protected] & [email protected] & +
12+log(s / h ) & [email protected] & [email protected] & + log(s / o ) & [email protected] & [email protected] & + & + ar@xmath126/h@xmath125(@xmath910@xmath116 ) & [email protected] & [email protected] & [email protected]
+ icf(ar ) & 1.44 & 1.48 & 1.45 + ar / h(@xmath910@xmath116 ) & [email protected] & [email protected] & [email protected] + 12+log(ar / h ) & [email protected] & [email protected] & [email protected] + log(ar / o ) & [email protected] & [email protected] & [email protected] + lccc @xmath122(oiii)(k ) & 12,760@xmath2640 & 17,230@xmath21620 & 17,450@xmath21160 + @xmath122(oii)(k ) & 12,590@xmath2590 & 14,730@xmath21320 &
14,820@xmath2940 + @xmath122(siii)(k ) & 12,680@xmath2530 & 16,000@xmath21340 &
16,180@xmath2960 + @xmath123(sii)(@xmath124 ) & 10@xmath210 & 10@xmath210 & 10@xmath210 + & + o@xmath125/h@xmath125(@xmath910@xmath113 ) & [email protected] & [email protected] &
[email protected] + o@xmath126/h@xmath125(@xmath910@xmath113 ) & [email protected] & [email protected] & [email protected] + o / h(@xmath910@xmath113 ) & [email protected] & [email protected] & [email protected] + 12+log(o / h ) & [email protected] & [email protected] & [email protected] + & + n@xmath125/h@xmath125(@xmath910@xmath116 ) & [email protected] & [email protected] &
[email protected] + icf(n ) & 2.46 & 1.71 & 1.46
+ n / h(@xmath910@xmath116 ) & [email protected] & [email protected] & [email protected] + 12+log(n / h ) & [email protected] & [email protected] &
[email protected] + log(n / o ) & [email protected] & [email protected] & [email protected] + & + ne@xmath126/h@xmath125(@xmath910@xmath113 ) & [email protected] & & + icf(ne ) & 1.69 & & + ne / h(@xmath910@xmath113 ) & [email protected] & & + 12+log(ne / h ) & [email protected] & & + log(ne / o ) & [email protected] & & + & + s@xmath125/h@xmath125(@xmath910@xmath116 ) & [email protected] & [email protected] &
[email protected] + s@xmath126/h@xmath125(@xmath910@xmath116 ) & [email protected] & [email protected] &
[email protected] + icf(s ) & 1.252 & 1.158 & 1.100 + s / h(@xmath910@xmath116 ) & [email protected] & [email protected] & [email protected] + 12+log(s / h ) & [email protected] & [email protected] & [email protected] + log(s / o ) & [email protected] & [email protected] & [email protected] + & + ar@xmath126/h@xmath125(@xmath910@xmath116 ) & [email protected] & [email protected] &
[email protected] + ar@xmath127/h@xmath125(@xmath910@xmath116 ) & [email protected] & [email protected] & [email protected]
+ icf(ar ) & 1.18 & 1.55 & 1.82 + ar / h(@xmath910@xmath116 ) & [email protected] & [email protected] & [email protected] + 12+log(ar / h ) & [email protected] & [email protected] & [email protected] + log(ar / o ) & [email protected] & [email protected] & [email protected] + lcccc 4101 h@xmath105 & [email protected] & [email protected] & [email protected] & [email protected] + 4340 h@xmath121 & [email protected] & [email protected] & [email protected] & [email protected] + 4363 [ o iii ] & [email protected] & [email protected] & [email protected] & [email protected] + 4686 he ii & [email protected] & [email protected] & & + 4740 [ ar iv ] & [email protected] & [email protected] & & + 4861 h@xmath53 & [email protected] & [email protected] & [email protected] & [email protected] + 4959 [ o iii ] & [email protected] & [email protected] & [email protected] & [email protected] + 5007 [ o iii ] & [email protected] & [email protected] & [email protected] & [email protected] + 5199 [ n i ] & [email protected] & [email protected] & & + 5755 [ n ii ] & [email protected] & [email protected] & & + 5876 he i & [email protected] & [email protected] & [email protected] & [email protected] + 6300 [ o i ] & [email protected] & [email protected] & & + 6312 [ s iii ] & [email protected] & [email protected] & & + 6364 [ o i ] & [email protected] & [email protected] & & + 6548 [ n ii ] & [email protected] & [email protected] & & + 6563 h@xmath12 & [email protected] & [email protected] & [email protected] & [email protected] + 6584 [ n ii ] & [email protected] & [email protected] & [email protected] & [email protected] + 6678 he i & [email protected] & [email protected] & & + 6717 [ s ii ] & [email protected] & [email protected] & & + 6731 [ s ii ] & [email protected] & [email protected] & & + 7065 he i & [email protected] & [email protected] & [email protected] & [email protected] + 7136 [ ar iii ] & [email protected] & [email protected] & & + 7320 [ o ii ] & [email protected] & [email protected] & [email protected] & [email protected] + 7330 [ o ii ] & [email protected] & [email protected] & & + & & & & + c(h@xmath53 ) dex & [email protected] & [email protected] + f(h@xmath53 ) & [email protected] & [email protected] + lcc @xmath122(oiii)(k)@xmath117 & 19,020@xmath2860 & 23,300@xmath24860 + @xmath122(oii)(k ) & 15,380@xmath2650 & 16,450@xmath22770 + @xmath122(nii)(k)@xmath117 & 13,750@xmath2530 & + @xmath122(siii)(k ) & 17,490@xmath2710 & + @xmath123(sii)(@xmath124 ) & 1918@xmath21434 & 2000 + & & + o@xmath125/h@xmath125(@xmath910@xmath113 ) & [email protected] & [email protected]
+ o@xmath126/h@xmath125(@xmath910@xmath113 ) & [email protected] & [email protected] + o@xmath127/h@xmath125(@xmath910@xmath113 ) & [email protected] & + o / h(@xmath910@xmath113 ) & [email protected] & [email protected] + 12+log(o / h ) & [email protected] & [email protected] + & & + n@xmath125/h@xmath125(@xmath910@xmath114 ) & [email protected] & [email protected]
+ icf(n ) & 3.54 & 2.51 + n / h(@xmath910@xmath113 ) & [email protected] & [email protected] + 12+log(n / h ) & [email protected] & [email protected] + log(n / o ) & [email protected] & [email protected] + & & + s@xmath125/h@xmath125(@xmath910@xmath116 ) & [email protected] & + s@xmath126/h@xmath125(@xmath910@xmath116 ) & [email protected] & + icf(s ) & 1.38 & + s / h(@xmath910@xmath116 ) & [email protected] & +
12+log(s / h ) & [email protected] & + log(s / o ) & [email protected] & + & & + ar@xmath126/h@xmath125(@xmath910@xmath116 ) & [email protected] & + ar@xmath127/h@xmath125(@xmath910@xmath116 ) & [email protected] & + icf(ar ) & 1.08 & + ar / h(@xmath910@xmath116 ) & [email protected] & + 12+log(ar / h ) & [email protected] & + log(ar / o ) & [email protected] & + & & + he / h & [email protected] & + 12+log(he / h ) & [email protected] & + & + lcc t@xmath61 ( k ) & 180,000 & @xmath6395,000 + log10(t@xmath61 ) & 5.26 & @xmath634.98 +
l / l@xmath74 & 6500 & 3400 + log(l / l@xmath74 ) & 3.80 & 3.53 + m / m@xmath128 & @xmath31.5@xmath117 & @xmath631.0@xmath117 + t@xmath129 ( gyr)@xmath110 & @xmath31.6@xmath117 & @xmath825.7@xmath117 + & & + v ( mag)@xmath130 & 2241@xmath2016 & 2271@xmath2027 + r ( mag)@xmath130 & 2157@xmath2012 & 2251@xmath2038 + i ( mag)@xmath130 & 2343@xmath2029 & 2240@xmath2064 + a@xmath131 ( mag)@xmath132 & 050 & 026 + | we present the results of high quality long - slit spectroscopy of planetary nebulae ( pne ) and hii regions in the two dwarf irregular galaxies sextans a and b , which belong to a small group of galaxies just outside the local group .
the observations were obtained with the eso new technology telescope ( ntt ) multi - mode instrument ( emmi ) . in sextans
a we obtained the element abundances in its only known pn and in three hii regions with the classical t@xmath0 method .
the oxygen abundances in these three hii regions of sextans a are all consistent within the individual rms uncertainties , with the average 12+@xmath1(o / h ) = [email protected] .
the oxygen abundance of the pn in sextans a is , however , significantly higher : 12+@xmath1(o / h ) = [email protected] .
this pn is even more enriched in nitrogen and helium , suggesting a classification as a pn of type i. the pn abundances of s and ar , which are presumably unaffected by nucleosynthesis in the progenitor star , are well below those in the hii regions , indicating lower metallicity at the epoch of the pn progenitor formation ( @xmath31.5 gyr ago , according to our estimates based on the pn parameters ) . in sextans
b we obtained spectra of one pn and six hii regions .
element abundances with the t@xmath0 method could be derived for the pn and three of the hii regions . for
two of these hii regions , which have a separation of only @xmath370 pc in projection , the oxygen abundances do not differ within the rms uncertainties , with a mean of 12+@xmath1(o / h ) = [email protected] .
the third hii region , which is about 0.6 kpc northeast from the former two , is twice as metal - rich , with 12+@xmath1(o / h ) = [email protected] .
this suggests considerable inhomogeneity in the present - day metallicity distribution in sextans b. whether this implies a general chemical inhomogeneity among populations of comparable age in sextans b , and thus a metallicity spread at a given age , or whether we happen to see the short - lived effects of freshly ejected nucleosynthesis products prior to their dispersal and mixing with the ambient interstellar medium will require further study . for the pn we measured an o / h ratio of 12+@xmath1(o / h)[email protected] , consistent with that of the low - metallicity hii regions .
we discuss the new metallicity data for the hii regions and pne in the context of the published star formation histories and published abundances of the two dwarf irregular ( dirr ) galaxies .
both dirrs show generally similar star formation histories in the sense of continuous star formation with amplitude variations , but differ in their detailed enrichment time scales and star formation rates as a function of time . if we combine the photometrically derived estimates for the mean metallicity of the old red giant branch population in both dirrs with the present - day metallicity of the hii regions , both dirrs have experienced chemical enrichment by at least 0.8 dex ( lower limit ) throughout their history . |
interfaces and the interaction of particles with them play a crucial role on small scales in biology and technology . for instance , biopolymers such as proteins or enzymes diffusing in biological cells intermittently bind to cellular membranes , or individual bacteria forming a biofilm on a surface use bulk excursions to efficiently relocate . similarly the exchange between a liquid phase with a solid surface is an important phenomenon in the self - assembly of surface layer films and is a ubiquitous process in emulsions .
this bulk - mediated surface diffusion , schematically shown in fig . [ scheme ] , was previously analyzed in terms of scaling arguments and simulations @xcite , and was unveiled in field cycling nmr experiments in porous glasses @xcite .
moreover , effects of bulk - surface interchange were reported on proton transport across biological membranes @xcite .
recent studies are concerned with effects of bulk - surface exchange on reaction rates in interfacial systems @xcite and with surface diffusion of coppper atoms in nanowire fabrication @xcite .
the remarkable finding of the bulk mediated surface diffusion model is that the effective surface motion is characterized by a cauchy propagator @xcite @xmath0 where @xmath1 denotes a scaling property ignoring multiplicative constants , and @xmath2 is a dimensional factor .
the associated stochastic transport is of superdiffusive nature @xcite , @xmath3 here we present a strictly analytical approach to this process .
our findings corroborate the previous scaling results for superdiffusion , however , we also derive the cutoffs to this behavior : at sufficiently long distances , the cauchy propagator turns over to a gaussian wing .
moreover at longer times the effective surface diffusion becomes subdiffusive , due to the fact that the particle spends less and less time on the surface .
normalized to the time - dependent surface coverage the effective surface diffusion turns over from superdiffusion to normal diffusion . .
following dissociation from the surface , thicker ( blue ) lines depict excursions into the bulk volume , in which the diffusion constant is @xmath4
. eventually , the particle rebinds to the surface.,width=302 ]
we start with a derivation of the coupling between surface and bulk in a discrete random walk process along the @xmath5 coordinate perpendicular to the surface ( fig . [ scheme ] ) .
let @xmath6 with @xmath7 denote the number of particles at site @xmath8 of this one - dimensional lattice with spacing @xmath9 .
the number of particles on the surface at lattice site @xmath10 are termed @xmath11 .
the exchange of particles is possible only via nearest neighbor jumps .
each jump event between bulk sites @xmath7 is associated with the typical waiting time @xmath12 . for the exchange between the surface and site
@xmath13 we then have the following law @xmath14 where @xmath15 is the characteristic time for desorption from the surface .
the probability of adsorption to the surface here is one , i.e. , a particle adds to the surface population automatically when moving from site 1 to site 0 .
the exchange site at @xmath13 between the surface at @xmath10 and the next bulk site @xmath16 is governed by the balance relation @xmath17 finally , the bulk sites @xmath18 are governed by equations of the form @xmath19 etc .
let us define the number of `` bulk '' particles at the surface site @xmath10 through @xmath20 this trick allows us to formulate the exchange equation also for site @xmath13 in a homogeneous form .
namely , from eq .
( [ exch ] ) we have @xmath21 moreover , from eqs .
( [ exch1 ] ) we find @xmath22 for @xmath23 .
let us now take the continuum limit .
for that purpose we make a transition from @xmath24 as the number of surface particles , and @xmath25 for the bulk concentration of particles .
the factor @xmath9 may be viewed as the lattice constant measuring the distance between successive sites along the @xmath5 axis . as
the bulk density is a function of @xmath5 while the surface particles are all assembled at @xmath26 , we need this length scale @xmath9 to match dimensionalities between @xmath27 and @xmath28 which are @xmath29=\mathrm{cm}^{-2}$ ] and @xmath30=\mathrm{cm}^{-3}$ ] .
expansion of the right hand side of eq .
( [ exch2 ] ) yields the surface - bulk coupling @xmath31 similarly from eq .
( [ exch3 ] ) we obtain the bulk diffusion equation @xmath32 finally , the boundary condition @xmath33 stems from our definition ( [ exch4 ] ) .
this completes the description of the particle exchange between surface and bulk , as well as the diffusion of bulk particles along the @xmath5 axis .
we now turn back to the full three - dimensional problem and add to above equations the @xmath34 directions ( see fig . [ scheme ] ) . as
the motion in the perpendicular @xmath5 direction is fully independent , we may simply adjust the laplacian to three dimensions , ending up with the diffusion equation @xmath35 with diffusivity @xmath4 .
for the surface diffusion along the @xmath36 and @xmath37 coordinates with @xmath26 , we end up with the two - dimensional diffusion equation with diffusivity @xmath38 , plus the bulk - surface exchange term : @xmath39 these two equations are valid in the range @xmath40 and @xmath41 , and are supplemented by the initial condition @xmath42 indicating that initially the particles are all concentrated on the surface at @xmath43 .
moreover we observe the boundary condition for @xmath28 , @xmath44 and the bulk initial condition @xmath45 in what follows , for simplicity of notation we use a unit initial concentration @xmath46 .
the surface and bulk diffusivities can be expressed in terms of the lattice constant @xmath9 and the typical waiting times between jumps in the bulk , @xmath12 , and on the surface , @xmath47 , through @xmath48 and @xmath49 respectively . in the continuum limit , both @xmath9 and @xmath12 ( or @xmath47 ) tend to zero such that the diffusion constants remain finite . in many physical systems
bulk diffusion is considerably faster , such that @xmath50 . above we also introduced the coupling parameter @xmath51 of physical dimension @xmath52=1/\mathrm{cm}$ ] .
small values of @xmath53 at fixed @xmath9 and @xmath4 correspond to slow bulk - surface exchange . for consistency
we derive the overall number of particles . to this end
we define the number particles on the surface , @xmath54 , and in the bulk , @xmath55 through the relations @xmath56 and @xmath57 from integration of eq .
( [ surf_diff ] ) we find @xmath58 similarly , eq . ( [ bulk_diff ] ) yields @xmath59 combination of eqs .
( [ surf_numb ] ) and ( [ bulk_numb ] ) produces @xmath60 thus the overall number of particles is conserved , as it should be .
to solve the set of coupled equations ( [ surf_diff ] ) and ( [ bulk_diff ] ) for the specified initial and boundary value problem , we start by defining the two - dimensional green s function @xmath61 then , the solution of eq .
( [ surf_diff ] ) becomes @xmath62 with initial condition ( [ surf_init ] ) we thus find @xmath63 where @xmath64}\ ] ] is the fourier - laplace transform of the surface green s function @xmath65 . here and in the following
we denote the laplace and fourier transform of a function by explicit dependence on the image variables , that is , @xmath66 the bulk particle density according to eq .
( [ bulk_diff ] ) is given by the formal expression @xcite @xmath67 from this expression one can indeed show that , despite the factor @xmath5 , the coupling equation ( [ bound ] ) is fulfilled , see appendix [ app1 ] .
now , eq . ( [ surf_fl ] ) requires the derivative of expression ( [ bulk_fl ] ) with respect to @xmath5 , evaluated at @xmath26 . to find that expression , we first differentiate @xmath68,\ ] ] and then calculate its fourier - laplace transform @xmath69 t\right)}{t^{3/2}}\right\}_{z=0},\ ] ] due to the convolution nature of expression ( [ expr ] ) .
the laplace transform is evaluated by help of the shift theorem , yielding @xmath70t\right)}{t^{3/2}}\right\ } = \mathscr{l}\left\{\frac{1}{t^{3/2}}\right\}_{s\to s+d_b[k_x^2+k_y^2]}= -\sqrt{4\pi\left(s+d_b[k_x^2+k_y^2]\right)}.\ ] ] we thus finally obtain @xmath71 insertion of this equation into expression ( [ surf_fl ] ) delivers the solution for the surface density in fourier - laplace space , @xmath72}},\ ] ] where the fourier - laplace transform of the green s function @xmath65 was defined in eq .
( [ greens ] ) .
after some transformations we arrive at the exact closed form expression @xmath73+\chi\sqrt{s+ d_b\left[k_x^2+k_y^2\right]}},\ ] ] with the rescaled coupling parameter @xmath74 of dimension @xmath75=1/\mathrm{sec}^{1/2}$ ] .
relation ( [ result_fl ] ) is the main result of this work , and we now consider the consequences to the surface motion effected by the bulk mediation .
we first determine the number of particles on the surface , @xmath76 whose laplace transform is @xmath77 inverse laplace transformation then yields the exact expression @xmath78 where we use the complementary error function @xmath79 at short times @xmath80 , this leads to the initial decay @xmath81 of the number of surface particles , eventually turning into the long time behavior @xmath82 the asymptotic @xmath83 decay stems from the returning dynamics to the origin @xmath26 of a brownian motion along the @xmath5 coordinate , i.e. , it is proportional to the normalization factor of a one - dimensional brownian motion . the additional prefactor @xmath84 rescales time with respect to the efficiency of the surface - bulk exchange .
we now turn to the surface dynamics , as quantified by the effective mean squared displacement along the surface . in the laplace domain , @xmath85n(k_x , k_y , s)\big|_{k_x = k_y=0}\\ & = & \frac{4d_s}{s(\sqrt{s}+\chi)^2}+\frac{2\chi d_b}{s^{3/2}(s^{1/2}+\chi)^2},\end{aligned}\ ] ] where @xmath86 . from this expression
we obtain the limiting behaviors at short and long times .
thus , we observe that the short time limit @xmath87 in laplace domain corresponds to @xmath88 , such that @xmath89 this translates into the asymptotic time evolution @xmath90^{1/2}\right).\ ] ] as long as the ratio @xmath91 is sufficiently large , there is a superdiffusive component @xmath92 winning over the normal surface diffusion proportional to @xmath38 .
this is exactly the famed bulk mediated superdiffusion originally obtained from scaling arguments by bychuk and oshaugnessy @xcite .
note that this diffusional enhancement is accompanied by an almost constant number of surface particles , compare eq .
( [ numb_short ] ) .
conversely , at long times @xmath93 ( or @xmath94 in laplace domain ) we find @xmath95 corresponding to the temporal behavior @xmath96^{1/2}\right).\ ] ] somewhat surprisingly , at sufficiently long times the bulk contribution to the effective mean squared displacement dominates over the surface contribution for arbitrary ratio @xmath91 , giving rise to _ subdiffusive _ behavior .
this subdiffusion occurs due to the ongoing loss of surface particles into the bulk , see eq .
( [ numb_long ] ) . instead of considering the surface
mean squared displacement @xmath97 we introduce the normalized effective surface mean squared displacement @xmath98 this quantity can be interpreted as the surface mean squared displacement covered by an individual particle that effectively stays on the surface and does not fully escape to the bulk . at long times
this quantity has the limiting form @xmath99 the long time diffusion corrected for the number of escaping particles displays normal diffusion , albeit with the _ bulk diffusivity_. in the following we neglect contributions from the surface diffusion proportional to @xmath38 , in order not to overburden the presentation .
the interesting behavior is due to the bulk mediation with weight @xmath4 .
we quantify the motion in terms of fractional order moments , before embarking for the surface propagator .
we now derive an exact expression for the @xmath100-th order moments ( @xmath101 ) @xmath102 to this end we utilize the following integral on the plane : @xmath103 where we used polar coordinates @xmath104 and @xmath105 corresponding to the two - dimensional vectors @xmath106 and @xmath107 .
thus we identify @xmath108 with the definition @xmath109 ^ 2.\ ] ] with this trick we can rephrase the @xmath100th order moment ( [ qth ] ) as follows , @xmath110\frac{d\mathbf{k}}{k^{2+q}}.\ ] ] the integral over @xmath111 of the surface density is but the number of surface particles , such that @xmath112,\ ] ] where we replaced the fourier cosine transform of @xmath113 by the real part of the exponential fourier transform . with
the fourier - laplace transform ( [ result_fl ] ) of the surface propagator with @xmath38 set to zero , for the laplace transform of the @xmath100th order moment we obtain @xmath114\frac{d\mathbf{k}}{k^{2+q}}=2\pi k(q)\left[\frac{1}{s+\chi\sqrt{s}}- \frac{1}{s+\chi\sqrt{s+d_bk^2}}\right]\frac{dk}{k^{1+q}}\\ \nonumber & = & 2\pi k(q)\frac{\chi}{s+\chi\sqrt{s}}\int_0^{\infty}\frac{\sqrt{s+d_bk^2}- \sqrt{s}}{s+\chi\sqrt{s+d_bk^2}}\frac{dk}{k^{1+q}}\\ \nonumber & = & 2\pi k(q)\frac{\chi d_b}{s+\chi\sqrt{s}}\int_0^{\infty}\frac{1}{\left(\sqrt{s } + \sqrt{s+d_bk^2}\right)\left(s+\chi\sqrt{s+d_bk^2}\right)}\frac{dk}{k^{q-1}}\\ \nonumber & = & 2\pi k(q)\frac{\chi^2}{s+\chi\sqrt{s}}\left(\frac{d_b}{s}\right)^{q/2}\int_0 ^{\infty}\frac{1}{\left(\chi+\chi\sqrt{1+y^2}\right)\left(\sqrt{s}+\chi\sqrt { 1+y^2}\right)}\frac{dy}{y^{q-1}}\\ & = & 2\pi k(q)\frac{\chi d_b^{q/2}}{s^{(1+q)/2}\left(s-\chi^2\right)}\big\ { i_1(q)-i_2(q)\big\ } , \label{qth_1}\end{aligned}\ ] ] where on the way we introduced the substitution @xmath115 .
the two integrals @xmath116 are defined by @xmath117 and @xmath118 now we analyze the temporal behavior of the @xmath100th order moment at short and long times . at long times @xmath93 ( or @xmath119 ) we see that @xmath120 and @xmath121 corresponding to the time evolution @xmath122 taking normalization by the number of surface particles into account , we find @xmath123 that is , at long times the surface diffusion exhibits normal scaling behavior .
the more interesting case is the short time behavior corresponding to the limit @xmath87 ( or @xmath124 ) .
here we consider three separate cases : \(i ) the case @xmath125 : since both integrals @xmath126 and @xmath127 converge , we may simply neglect @xmath128 .
then @xmath129 such that after laplace inversion we find @xmath130 \(ii ) the case @xmath131 : now we should take into account both integrals , @xmath132 to estimate the main contribution from this difference , we split it into three parts , namely @xmath133 and evaluate each contribution separately .
we find @xmath134 and then @xmath135 finally , @xmath136 thus the main contribution come from eqs .
( [ app_int1 ] ) and ( [ app_int2 ] ) , and @xmath137 with this estimate the laplace transform of the @xmath100th order moment , eq .
( [ qth_1 ] ) , has the leading order behavior @xmath138 this corresponds to @xmath139 in the time domain .
\(iii ) the case
@xmath140 : this special case requires some care .
we start with the substitution @xmath141 in eqs .
( [ int1 ] ) and ( [ int2 ] ) .
then the integrals @xmath116 become @xmath142 and @xmath143 we can now rewrite eq .
( [ qth_1 ] ) in the form @xmath144 where we used @xcite @xmath145 for @xmath146 the integral in the parenthesis of eq .
( [ help ] ) becomes @xcite @xmath147 thus , for eq .
( [ help ] ) we find @xmath148 so that we find the @xmath149 behavior @xmath150 after laplace inversion , @xmath151 at long times , @xmath93 , consistent with eq .
( [ long_qth ] ) .
conversely , for @xmath152 we employ @xcite @xmath153 thus we find @xmath154.\ ] ] we thus obtain the limiting form at large @xmath155 @xmath156 back - transformed this results in @xmath157 at short times @xmath87 . summarizing our results for @xmath100th order moments , at short times @xmath87
the effective surface diffusion exhibits anomalous scaling : the @xmath100th order moment of the radius scales like @xmath158 for @xmath131 and @xmath159 for @xmath125 , while the first moment includes a logarithmic contribution , @xmath160 , consistent with the earlier results in ref
. @xcite . at long times
@xmath93 the @xmath100th order moments scale normally with time , proportional to @xmath161 .
we now turn to the behavior of the surface propagator . if we neglect surface diffusion ( i.e. , @xmath162 ) the surface propagator from eq .
( [ result_fl ] ) becomes @xmath163 we perform an inverse laplace transformation along the bromwich path : @xmath164 where on the way we introduced the substitutions @xmath165 and @xmath166 .
with @xmath167 we finally arrive at the fourier transform of the surface propagator , @xmath168\exp\left(\left[\frac{1}{2}+\sqrt{\kappa^2+\frac{1 } { 4}}\right]t\chi^2\right)\mathrm{erfc}\left(\left[\frac{1}{2}+\sqrt{\kappa^2 + \frac{1}{4}}\right]\chi\sqrt{t}\right)\right.\\ & & + \left . \left(\sqrt{\kappa^2+\frac{1}{4}}-\frac{1}{2}\right)\exp\left(\left[\frac{1 } { 2}-\sqrt{\kappa^2+\frac{1}{4}}\right]t\chi^2\right)\mathrm{erfc}\left(\left [ \frac{1}{2}-\sqrt{\kappa^2+\frac{1}{4}}\right]\chi\sqrt{t}\right)\right\}. \label{kt_prop}\end{aligned}\ ] ] in compact form , @xmath169 with @xmath170 we now calculate the surface propagator @xmath113 in the limits of short and long times . at short times
@xmath80 we distinguish between the central part of the propagator and its wings . starting with the central part , we thus focus on distances @xmath171 . due to @xmath172 ,
this means that we consider @xmath173 , and therefore @xmath174 .
consequently we find @xmath175 , @xmath176 , and thus @xmath177 then , the first term in the curly brackets of eq .
( [ compact ] ) is ( note that @xmath178 ) @xmath179 the error function in the second term of eq .
( [ compact ] ) is approximately 2 .
thus , the second term in the curly brackets of eq .
( [ compact ] ) prevails , and we obtain @xmath180 the fourier inversion is accomplished with the help of the integral @xcite @xmath181 and we obtain @xmath182 such that @xmath183 this is exactly the two - dimensional cauchy distribution obtained by bychuk and oshaugnessy from scaling arguments @xcite .
we see that indeed radius and time are coupled linearly in superdiffusive fashion , @xmath184 .
however , as we will see this cauchy form holds only for the central part of the surface propagator .
it assumes steeper wings such that at all times any moment of the propagator exists .
we are quantifying these wings below .
the number of particles populating this cental cauchy domain of the propagator , i.e. , within a circle of radius @xmath185 is @xmath186 i.e. , at sufficiently short times a major portion of the particles are contained in the cauchy domain . indeed ,
these particles show the superdiffusive mean squared displacement @xmath187 consistent with result ( [ msd_asymp ] ) for @xmath162 , the prefactor differing by approximately a factor @xmath188 .
the outer part of the propagator at short times is found for @xmath189 .
this region can be divided into two subregions , namely , @xmath174 corresponding to @xmath190 , and @xmath191 corresponding to @xmath192 .
however , it is easy to check that in both of these sub - regions we have @xmath193 and @xmath194 , due to the combination of the conditions of short times ( @xmath195 ) and large radii ( @xmath196 ) .
thus both error functions in eq .
( [ compact ] ) are expanded at small values of their arguments , @xmath197 with @xmath198 and @xmath199 , respectively .
our result is @xmath200 plus terms of order @xmath201 and higher .
we conclude that at the wings the central cauchy distribution is truncated by gaussian tails whose dispersion grows like @xmath202 .
this is the limit @xmath93 .
as this implies @xmath203 we may expand the error function as @xmath204 we consider the shape of the propagator in the main part excluding the region near the origin , i.e , we have @xmath205 in this domain we may use @xmath206 , @xmath207 , and @xmath208 . since @xmath209 , we have @xmath194 , and thus the second error function becomes @xmath210 the surface propagator thus assumes the shape ( recall that @xmath211 ) @xmath212 equivalently , this means that @xmath213 inverse fourier transform produces the propagator at long times , @xmath214 from this we obtain the gaussian form @xmath215)}{4 \pi d_bt}.\ ] ] the number of particles contained in this part of the long - time propagator follows from @xmath216 comparison to eq .
( [ numb_long ] ) shows that these are virtually all particles still on the surface .
thus the gaussian ( [ gauss ] ) indeed dominates the behavior at long times .
this statement is also corroborated from calculation of the surface mean squared displacement from eq .
( [ gauss ] ) in the domain @xmath217 : @xmath218.\end{aligned}\ ] ] again , up to a small correction , this term coincides with our exact result for the long time behavior , eq .
( [ momlong ] ) .
complementing our earlier studies of bulk mediated surface diffusion on a cylinder @xcite , we presented a strictly analytical treatment of the bulk mediated surface diffusion problem on a flat surface .
while our analysis corroborates previous scaling results of superdiffusion and the cauchy form of the surface propagator , we find corrections to this behavior sufficiently far away from the origin , and at long times .
thus , the cauchy domain is rectified by a gaussian decay of the extremities of the propagator which we derive explicitly . as a consequence finite spatial moments of any order remain finite . at long times
a gaussian regime takes over , assembling virtually all particles .
our analysis is based on the exact representation of the surface propagator in fourier - laplace space from which systematic expansions are obtained . for the number of surface particles we find an initial stagnation , followed at longer times by a decrease proportional to @xmath219 .
the mean squared displacement @xmath220 turns from the initial superdiffusive behavior proportional to @xmath202 to a subdiffusive form @xmath221 .
the latter corresponds to normal diffusion , albeit with the bulk diffusivity , if normalized to the decay of the surface particles .
our analysis is complemented with the derivation of fractional order moments for the surface propagation .
rm thanks the academy of finland for financial support within the fidipro scheme .
in this appendix we derive the lvy flight behavior recovered in the main part of the text based on a complementary approach developed in refs .
we start with the coupled diffusion equation along the surface ( equivalent to eq .
( [ surf_diff ] ) ) as @xmath222 where the last term represents rebinding events of particles that previously desorbed from the surface .
the kernel @xmath223 in the convolution is the probability density to attach at @xmath34 at time @xmath224 after desorption from the origin at time zero .
together , the last two terms of eq .
( [ eq : astart ] ) represent jumps along the surface mediated by the bulk .
note that we have implicitly assumed that all particles start at the surface at @xmath225 .
otherwise an additional term would be needed to represent particles binding time to the surface for the first time . to obtain @xmath226 we need to solve the bulk diffusion equation , eq .
( [ bulk_diff ] ) , with a single particle released from the boundary at the origin at time zero .
this can be achieved by choosing a boundary condition @xmath227 this equation corresponds to the current of particles into the surface , the second term on the right hand side representing the initial release . supplementing this by boundary conditions of vanishing density at infinity ,
@xmath228 as @xmath229 , and the initial condition of finding no particles in the bulk , @xmath230 , we obtain @xmath231 from the solution in terms of the flux into the boundary after the initial release : @xmath232 . to solve eq .
( [ bulk_diff ] ) we again fourier transform along @xmath36 and @xmath37 ( @xmath233 and @xmath234 ) and laplace transform in time ( @xmath235 ) . together with the initial condition and the boundary conditions at infinity it is easily found that @xmath236 where @xmath237 .
insertion of eq .
( [ eq : nbsol ] ) into eq .
( [ eq : aboundary ] ) leads to a bulk density at the boundary which is @xmath238 multiplication by the rate constant @xmath239 then leads to @xmath240 we are now ready to obtain the surface density @xmath27 in fourier - laplace space . from eq .
( [ eq : astart ] ) with the initial condition @xmath241 we obtain @xmath242 } , \label{eq : a6}\ ] ] in which eq .
( [ eq : wbulk ] ) can be directly inserted . to obtain the expression in eq .
( [ result_fl ] ) of the main text one takes the limit of the desorption rate constant @xmath243 and adsorption rate constant @xmath239 going to infinity while keeping their ratio @xmath53 fixed , which leads to @xmath244\to \mu d_b\sqrt{k^2+s / d_b}.\ ] ] together with @xmath245 and @xmath246 this makes eq .
( [ eq : a6 ] ) equivalent to eq .
( [ result_fl ] ) . as a byproduct of the approach of this appendix we remark that we can now obtain the distribution of bulk mediated jump lengths .
these are given by @xmath247 . from eq .
( [ eq : wbulk ] ) we find [ performing the angular part of the inverse fourier transform similarly to eq .
( [ cauchy ] ) ] @xmath248,\ ] ] where @xmath249 are the struve functions and @xmath250 are the bessel functions of the second kind .
if we expand @xmath251 asymptotically at large @xmath252 we obtain a power law @xmath253 thus we see that the cauchy propagator ( [ cauchy ] ) , arising for @xmath162 in the short time regime where almost all particles are still bound on the surface , is connected with a lvy flight behavior based on a power law jump length distribution .
we start by rewriting eq .
( [ bulk_fl ] ) in the form @xmath254 we know that the gaussian is a limiting representation of the @xmath255 function : @xmath256 let us now introduce the new variable @xmath257 , such that relation ( [ app1_1 ] ) turns into @xmath258 rearranging terms , @xmath259 we are now in the position to take the limit @xmath260 , producing ( we omit the prime in @xmath261 ) @xmath262 which directly reduces to the expression @xmath263 finally , with the substitution @xmath264 with the jacobian @xmath265 , the remaining integral represents the normalization of a gaussian , @xmath266 and the equivalence with the coupling equation ( [ bound ] ) is established .
h. s. carslaw and j. c. jaeger , conduction of heat in solids , 2nd edition ( clarendon press , oxford , uk , 1959 ) , chapter 2 ; a. n. tikhonov and a. a. samarskii , equations of mathematical physics ( pergamon press , oxford , uk , 1963 ) , chapter 3 . | we consider the effective surface motion of a particle that intermittently unbinds from a planar surface and performs bulk excursions .
based on a random walk approach we derive the diffusion equations for surface and bulk diffusion including the surface - bulk coupling . from these exact dynamic equations
we analytically obtain the propagator of the effective surface motion .
this approach allows us to deduce a superdiffusive , cauchy - type behavior on the surface , together with exact cutoffs limiting the cauchy form .
moreover we study the long - time dynamics for the surface motion . |
the result reported by @xcite was of a difference in the faraday rotation measure @xmath3 between two lines of sight to two components of an extragalactic radio source ( 3c228 ) .
these lines of sight were separated by an angular distance @xmath4 , which corresponds to a linear separation in the corona between the two lines of sight , @xmath5 , where @xmath6 is the distance to the sun . in the observations reported in @xcite @xmath7 arcseconds and @xmath8 km .
the observations were made when the line of sight to 3c228 passed through the corona at heliocentric distances from @xmath9 .
the technique is illustrated in figure 1 of @xcite .
the fundamental physical relation used in the technique is expressed by equation ( 3 ) of @xcite @xmath10 where the integral is around an amperian loop through the corona , consisting of the two lines of sight , closed by imaginary line segments which join the two lines of sight , at locations infinitely separated from the corona . in this formula , @xmath11 is a collection of atomic constants which arise in the description of faraday rotation , defined as @xmath12 in si units , @xmath13 is the electron density , @xmath14 is the vector magnetic field in the corona , and @xmath15 is an incremental step around the amperian loop .
use of ampere s law @xmath16 in equation ( 1 ) shows that the differential rotation measure @xmath3 is directly related to the current within the amperian loop defined by the two lines of sight .
the transition from the middle to the right term in equation ( 1 ) involves an approximation , in which the position - dependent plasma density in the integrand is replaced by an effective mean density @xmath17 .
this approximation is discussed at length in @xcite , where arguments for its plausibility are presented . as is the case with ampere s law , the differential faraday rotation measurement provides information on the _ net current _ within the amperian loop . in general , we expect both positive and negative currents to be flowing within the loop .
the electrical current flowing in a given location in the corona could be much larger than that deduced by the arguments above , and contained in equation ( 7 ) of @xcite .
@xcite presented the following results from two observing sessions with the very large array radiotelescope .
each session lasted approximately 8 hours . 1 .
in one of the two sessions , there was a confident detection of a @xmath3 event , with a corresponding inferred electrical current of @xmath0 amperes .
2 . in the second observing session
, a marginal @xmath3 event was detected with a value of @xmath18 ( which may well be considered an upper limit ) of @xmath19 amperes .
3 . during a several hour period of good data quality ,
no significant @xmath3 events were detected , with a corresponding upper limit to the current of @xmath20 amperes .
although the data from the earlier investigation of @xcite have not been reanalysed in this manner , examination of figure 11 of that paper shows no clear evidence of a @xmath3 event in several more hours of vla observation .
all of this indicates that detection of a clear differential rotation measure event between lines of sight separated by @xmath21 km is relatively rare .
the aforementioned observations will represent the observational constraints imposed on the theory developed in the next section . and width @xmath22 ( noted in the figure as `` t '' and `` l '' ) , within a `` domain '' of width @xmath23 .
adapted from @xcite.,width=313 ]
in this section , i discuss the implications of the results from @xcite for coronal heating .
the presence of electrical currents indicates that joule heating will occur as well .
i will calculate an estimate of the average volumetric heating rate of a system of currents which could produce the observations discussed in section 2
. this calculation will be highly model dependent , as well as dependent on assumptions regarding the nature of the current sheets .
since the subsequent discussion will introduce many assumed parameters of the coronal current sheets , some of which are poorly constrained , i will follow the once - common practice in physics and astronomy literature of including a glossary of physical variables .
this is contained in table 1 .
ll variable & definition + @xmath24 & area of the amperian loop defined by the experiment + @xmath25 & ratio of current sheet width to domain size , @xmath26 + @xmath27 & joule heating per current sheet ( watts ) + @xmath28 & joule heating per domain , due to all current sheets + @xmath29 & domain - averaged volumetric heating rate + @xmath30 & domain - averaged volumetric heating rate for turbulent current sheets + @xmath31 & domain - averaged volumetric heating rate for deterministic current sheets + @xmath32 & electrical resistivity of a plasma , @xmath33 + @xmath34 & spitzer resistivity ( equation ( 16 ) ) + @xmath35 & fraction of sheets with positive current density within amperian loop + @xmath36 & fraction of sheets with negative current density within amperian loop + @xmath18 & electrical current per sheet + @xmath37 & total current within the amperian loop + @xmath38 & current density within one current sheet + @xmath39 & transverse separation of two lines of sight in the corona + @xmath22 & width of a current sheet + @xmath23 & transverse scale of a domain which contains a few current sheets + @xmath40 & extent of domain along the large scale magnetic field + @xmath41 & anisotropy of the current sheets ( scale length along magnetic field / scale length across field ) + @xmath42 & number of current sheets within one domain + @xmath43 & number of sheets with positive current density within amperian loop + @xmath44 & number of sheets with negative current density within amperian loop + @xmath45 & total number of current sheets within the amperian loop defined by the experiment + @xmath46 & effective line - of - sight depth of the coronal plasma + @xmath47 & electrical conductivity of a plasma + @xmath48 & thickness of a current sheet + @xmath49 & electron drift speed + @xmath50 & volume of a current sheet + @xmath51 & volume of a domain + @xmath52 & extent of a current sheet along the large scale magnetic field + the following analysis assumes that the current is contained in a number of thin current sheets within the amperian loop .
this model is illustrated in figure 1 .
a coordinate system is defined by having one axis ( the @xmath53 axis ) coincide with that of the large scale coronal magnetic field .
i assume that the current sheets are extended along the large scale field , as is the case in quasi-2d magnetohydrodynamics @xcite .
the current sheet properties have a weaker dependence on the coordinate along the large scale field than on the coordinates in a plane perpendicular to that field .
i begin by assuming that one can define a `` domain '' which has a scale @xmath23 perpendicular to the large scale coronal magnetic field , and which contains a small integer number @xmath42 of current sheets . in the analysis which follows in sections 3.1 and 3.2
, i will assume that all current sheets are identical .
the current sheet properties which are introduced are obviously to be understood as mean values from a distribution .
figure 1 illustrates such a domain .
the extent of the domain in the direction perpendicular to the plane defined by figure 1 is @xmath54 , with @xmath55 .
the current sheets have a width @xmath22 , a thickness @xmath48 , and an extension along the large scale field @xmath56 .
the picture which has been drawn so far is consistent with the original view of @xcite .
it is also consistent with results from studies of 2d mhd turbulence , which show that turbulent evolution results in the formation of isolated , intense sheets of current and vorticity .
the development which follows is based on results from @xcite , which contains an extensive bibliography to the literature where these ideas were developed earlier , most importantly @xcite . in the case of current sheets which arise from 2d mhd turbulence ,
the number of positive and negative current sheets should be equal , and the expectation value of the current in an amperian loop is zero .
the detection of net currents ( via differential faraday rotation ) would then be interpreted as a statistical fluctuation of the total current about the zero expectation value . in what follows
, i will refer to these as `` turbulent current sheets '' .
it is also possible that the physics of the corona selects current sheets with a preferred sign of the current density , at least for that portion of the corona which is probed in a faraday rotation experiment .
this situation is referred to as that of `` deterministic current sheets '' , and is discussed in section 3.2 . within this model ,
i assume that the properties of the individual current sheets are essentially the same as in the turbulent model , but that there is a preference for one sign of current density .
it should be noted that the turbulent current sheet model is based on analytic and numerical solutions of the equations of 2d magnetohydrodynamics , whereas the deterministic model is plausible but ad - hoc .
i begin with the view that the current sheets arise as the evolution of 2d , or quasi-2d mhd turbulence @xcite . in this case
, the domain size @xmath23 may be plausibly identified with the outer scale of the turbulence .
the joule heating in each current sheet @xmath27 is @xmath57 where @xmath50 is the volume of a single sheet , given by @xmath58 . the joule heating from all the current sheets in the domain @xmath28 is then given by @xmath59 where @xmath42 is the number of current sheets per domain .
the mean volumetric heating rate in the domain @xmath29 , which is taken to be the overall volumetric heating rate , is @xmath60 where @xmath61 is the volume of a domain .
using the fact that the current per sheet is @xmath62 , we have @xmath63\ ] ] the question now arises as to how to relate the current in an individual current sheet , @xmath18 , with the total current @xmath37 within the amperian loop .
this relation will depend on the model for the current sheets . for the remainder of this subsection , i will adopt the turbulence model in which there are , on average , equal numbers of positive and negative current sheets , and statistical fluctuations are responsible for @xmath64 . let @xmath45 be the total number of current sheets within the amperian loop .
we then identify the measured current @xmath37 with the rms fluctuation in the total current contained within the loop , @xmath65 the total number @xmath45 is given by @xmath66 where @xmath24 is the area of the amperian loop , @xmath39 is the spacing between the lines of sight , introduced in section 1 , and @xmath46 is the effective line - of - sight extent of the coronal plasma . equation ( 8) is for the simplest case , in which the amperian loop is perpendicular to the large scale field . in the general case , a cosine of an orientation angle would be introduced in the numerator .
this detail is ignored in the present discussion .
substitution of equations ( 7 ) and ( 8) into equation ( 6 ) yields the volumetric heating rate in terms of the measured total current @xmath37 , @xmath67 \left ( \frac { i_{obs}^2}{ls_{los } } \right)\ ] ] as a final approximation , i assume that the extension of the domain and that of the current sheet along the large scale field direction are described by the same anisotropy index @xmath68 , @xmath69 .
use of these relations gives us the basic expression for the average volumetric heating rate due to turbulent current sheets in terms of the observed parameter @xmath37
@xmath70 \left ( \frac{i_{obs}^2}{ls_{los}}\right)\ ] ] this expression factors itself neatly into three terms , each contained within brackets .
the first is determined by the resistivity in the plasma and the domain properties .
the second is determined by properties of the current sheets , specifically their thickness .
the final term collects properties of the observations , such as the inferred total current and the parameters of the lines of sight . in this subsection ,
i consider the possibility that the current sheets are not entirely random , and that there may be some preference for one sign of the current density , probably determined by the polarity of the large - scale coronal field .
the total net current from the sun must obviously be zero . however , it is possible for a net current to exist in a limited region probed by a radio remote sensing measurement .
we assume that the properties of the individual current sheets can be described as previously , so that the equations of section 3.1 up to , and including equation ( 6 ) , are valid .
however , in the present case , there will be a different relationship between the total current @xmath37 and the current of an individual sheet , @xmath18 .
if there is a preference for current sheets of one sign of the current density , we can write @xmath71 where @xmath43 is the number of sheets with positive current within the amperian loop , and @xmath44 is the number of sheets with negative current density .
the individual sheet current @xmath18 is then taken as an absolute magnitude , with the sign of the current assumed in @xmath43 and @xmath44 .
if we introduce probabilities that the current densities will be positive or negative by @xmath72 , we have an expression for the current in a single current sheet , @xmath73 the total current sheet number @xmath45 is the same as that defined in equation ( 8) .
substitution of equation ( 12 ) into ( 6 ) , and algebraic manipulation gives the volumetric heating rate in the case of `` deterministic '' current sheets @xmath74 \left ( \frac{1}{t_c(1 - 2f_-)^2 } \right ) \left ( \frac{i_{obs}^2}{l^2 s_{los}^2 } \right)\ ] ] where the expression has again been factored into terms which contain , respectively , characteristics of the plasma , the current sheets , and the observations .
the expressions for the volumetric heating rate in the two models of the current sheets , equations ( 10 ) and ( 13 ) respectively , appear quite different in form , and it is natural to ask which is the larger for realistic input parameters . in other words ,
given a measurement of @xmath37 , would greater joule heating result if the current were distributed in a random set of turbulent current sheets as described in section 3.1 , or in a set of sheets with predominantly one sign of the current density , as discussed in section 3.2 .
let the heating rate expression for a turbulent set of current sheets as given in equation ( 10 ) be noted by @xmath30 , and that due to a systematic set of sheets with preferentially one sign of the current density ( equation ( 13))as @xmath31 .
equations ( 10 ) and ( 13 ) can be easily manipulated into the following form @xmath75 \left ( \frac{\lambda^2}{l s_{los}}\right)\ ] ] if one assumes that the first term in square brackets on the right hand side of this equation is of order unity , then the relative heating rate depends on the ratio of the domain area to that of the amperian loop .
the precise value of this ratio depends on the circumstances of the observations , as well as the value of @xmath23 .
an estimate of its value in the case of the observations of @xcite is given at the end of the next section .
in what follows , i discuss the case of turbulent current sheets , then briefly note that the conclusions would not be significantly different for the deterministic case .
equation ( 10 ) is now used to estimate the coronal heating rate from turbulent current sheets .
the variables in the last term ( @xmath76 ) are observational parameters and are known .
the calculation will be carried out for the conditions characteristic of the large @xmath3 event of august 16 , 2003 , in which the inferred current was @xmath0 amps .
the line of sight to the radio source had a minimum heliocentric distance of @xmath77 .
coronal plasma properties characteristic of this distance will be used in the calculation below .
a similar analysis at other times in the observations of @xcite , when there were only upper limits to the current , would obviously yield lower values for the joule heating rate .
the calculation also requires estimates of @xmath32 , @xmath23 , and @xmath48 .
[ [ resistivity ] ] resistivity + + + + + + + + + + + for the resistivity @xmath32 , the spitzer resistivity is used , which is based on coulomb collisions of current - carrying electrons with ions and other electrons .
it is certain to be a drastic underestimate , in that the true resistivity is almost certainly determined by collisionless processes .
however , the spitzer resistivity can be derived from fundamental principles , which is not true of other estimates , and it can serve as a lower limit to the true resistivity .
an informal discussion of the possible role of collisionless processes in determining the resistivity is given in section 4 below .
the spitzer resistivity is the reciprocal of the conductivity given by @xcite @xmath78 in this equation , and equation ( 16 ) below , @xmath23 stands for the coulomb logarithm rather than the domain size as used otherwise .
the electron temperature is @xmath79 .
all other terms in equation ( 15 ) have been defined , or are obvious fundamental physical constants .
equation ( 15 ) can be used to write the spitzer resistivity in a `` suitable for observers '' form as @xcite @xmath80 where the thermal energy @xmath81 is now given in electron volts . for approximate coronal conditions
i choose a value for the coulomb logarithm of @xmath82 @xcite . with an assumed coronal temperature of @xmath83
k , appropriate for closed - field regions ( electron thermal energy @xmath84 ev in equation ( 16 ) ) , the resistivity is @xmath85 ohm - m , or about 35 times the resistivity of silver . [ [ domain - size ] ] domain size
+ + + + + + + + + + + + i will take the domain size @xmath23 to be the outer scale of the turbulence in the relevant part of the corona .
there are two estimates in the literature for this outer scale .
the first is mean spacing between flux tubes which expand into the corona .
this estimate was introduced by @xcite , and subsequently used by @xcite and @xcite .
the formula used by @xcite is @xmath86 where @xmath87 is the magnetic field strength in gauss . for the magnetic field in the corona
, we use the recent estimate of @xcite which was obtained from faraday rotation measurements very similar to those of @xcite .
they found that the magnetic field could be represented by an inverse square dependence on the heliocentric distance , with a normalizing value of @xmath88 g at @xmath89 . at a heliocentric distance of 6.7 @xmath2 ,
the estimated magnetic field is @xmath90 g , and the corresponding value of @xmath23 is @xmath91 m. in a more recent theoretical study of coronal heating and solar wind acceleration , @xcite argue for a smaller value of the domain size ( their parameter @xmath92 which serves as the outer scale of the turbulence ) , which physically corresponds to the diameter of the photospheric flux tubes rather than their separation . adopting the estimate for
@xmath93 from @xcite would reduce our value of @xmath23 by about a factor of 4 from the estimate of equation ( 17 ) .
a second estimate of the outer scale of coronal turbulence comes from power spectra of fluctuating doppler shifts of a spacecraft transmitter @xcite .
these estimates , which result from measurements rather than plausible theoretical arguments , give outer scales from a few tenths of a solar radius to a solar radius or more at heliocentric distances of @xmath94 .
the values reported by @xcite and @xcite are several times larger than that given by equation ( 17 ) .
it is obvious that the factor of 4 smaller value for @xmath23 advocated by @xcite is in more serious disagreement with the observational value of @xcite and @xcite .
a resolution of this matter would warrant a paper in its own right , but for the present work we use equation ( 17 ) . as may be seen from the heating rate expression in equation ( 10 )
, lower values of the domain size @xmath23 generate higher values of the heating rate @xmath29 .
[ [ current - sheet - thickness ] ] current sheet thickness + + + + + + + + + + + + + + + + + + + + + + + + for the current sheet thickness @xmath48 , i choose the ion inertial length @xmath95 where @xmath96 is the alfvn speed , and @xmath97 is the proton ion cyclotron frequency .
this would seem to be both plausible and a good lower limit to what the current sheet thickness can be .
once again , equation ( 10 ) shows that use of a minimum plausible value for @xmath48 leads to an upper limit to the heating rate @xmath29 . to calculate the ion inertial length , the plasma density profile given by equation ( 6 ) of @xcite ( based on radio propagation measurements of the corona ) and the magnetic field model of @xcite are used .
these yield the following formula for the estimated current sheet thickness @xmath98 where @xmath99 is the heliocentric distance in units of a solar radius . for @xmath100 , @xmath101 m. [ [ observational - parameters ] ] observational parameters + + + + + + + + + + + + + + + + + + + + + + + + the observed parameters in equation ( 10 ) are contained in the term in the third set of brackets .
the observed current @xmath102 amps and the separation of the lines of sight @xmath39 is 33,000 km @xcite .
there remains the value for the effective thickness of the plasma along the line of sight .
i use the expression from @xcite @xmath103 use of the above parameters with an impact parameter @xmath104 gives a heating rate @xmath105 watts / m@xmath106 . to determine the significance of this number , i compare it to theoretical estimates of @xcite and @xcite .
these papers utilize cgs units , and report heating rates in power per unit mass . the volumetric heating rate given above is then @xmath107 ergs / cm@xmath106/sec , and is converted to a heating rate / unit mass @xmath109 @xmath110 where i have again used the power law density model of equation ( 6 ) of @xcite in obtaining the mass density at @xmath111 .
@xcite calculate heating rates as a function of heliocentric distance .
since their calculations are self - consistent , their heating rates may be considered to be those which are required by the observed heating of the corona and acceleration of the solar wind .
examination of figure 8 of @xcite shows that the heating rate per unit mass at a heliocentric distance of @xmath112 is in the range of @xmath113 ergs - sec@xmath114-gm@xmath114 , depending on the assumed amplitude of the photospheric velocity fluctuations .
we therefore conclude that the heating rate given by equation ( 10 ) is lower than values which are required to account for coronal heating by at least a factor of @xmath1 , if the input parameters used here are valid .
if the outer scale to the turbulence is a factor of @xmath115 less than the value given by equation ( 17 ) , as recommended by @xcite , the ratio of mass heating rates would be about @xmath116 . in either case
, this huge mismatch means that exercises with fine tuning the parameters in the model would be a fool s errand .
it should be noted that the ratio @xmath117 which appears in equation ( 14 ) is of the order of unity , within a factor of several either larger or smaller depending on the assumed outer scale of the turbulence .
the conclusion on the magnitude of joule heating would not be changed by adopting the non - turbulent current sheet model of section 3.2 .
there are two possible conclusions to be drawn from the calculations of this section . 1 . in view of the large disparity between the calculated joule heating rate and that which is required for a significant contribution to the thermodynamics of the corona , the currents which may have been observed are irrelevant for coronal heating .
this argument would seem to be strengthened by the fact that i used the largest detected value of @xmath37 from the two days of observation .
other intervals would have provided smaller values for @xmath37 or upper limits thereto , yielding smaller values of @xmath29 .
2 . a more likely explanation , in my opinion ,
is that these current systems do play an important role in coronal heating , but that role is underestimated in the calculations presented here , because they are based on the spitzer resistivity . according to this viewpoint , the analysis of sections 3.1 and 3.2 is valid , but a correct calculation would require an appropriate , and much larger value of the resistivity .
this is clearly speculation until it can be demonstrated that a much larger resistivity ( by orders of magnitude ) characterizes the coronal plasma at @xmath118 .
in this section , i consider the second of the possibilities listed immediately above , i.e. that the resistivity could be sufficiently enhanced in these coronal current sheets to make joule heating a thermodynamically important process .
an obvious way for this to happen is a plasma instability that produces high levels of fluctuating electric or magnetic fields , which scatter the current - carrying electrons and enhance the resistivity . to assess this possibility
, we need to examine the magnitude of the electron drift speed within the current sheets .
equations ( 7 ) and ( 8) give the relationship between the observed current in the amperian loop , @xmath37 , and the current in a single sheet , @xmath18 . using these equations and the identity immediately before equation ( 6 ) , we have for the current density in a single sheet @xmath119 and for the electron drift speed @xmath120 where @xmath121 is the fundamental electric charge . to simplify equation ( 22 ) i adopt a set of plausible assumptions .
i assume that the width of a current sheet will be some fraction of the domain size , @xmath122 with @xmath25 probably having a value between 0.1 and 0.5 . from section 3 , we already have estimates of other parameters ( e.g. @xmath48 , @xmath22 , @xmath123 ) in equation ( 21 ) at the fiducial heliocentric distance of @xmath77 .
this yields the following estimate for the electron drift speed , @xmath124 for this expression to be meaningful in the context of plasma instabilities , we need to compare it with a characteristic plasma speed . an obvious choice is the electron thermal speed @xmath125 @xcite .
i use a value of @xmath126 k , which is characteristic of closed - magnetic - field regions in the corona .
open field regions would have a lower temperature and lower thermal speed .
we then have for the drift speed to thermal speed ratio @xmath127 as mentioned in the definition of @xmath25 immediately above , and the discussion of @xmath42 in section 3 , @xmath25 is a number which is probably less than unity , but not by a large factor , and @xmath42 is an integer which is probably larger than unity , but not much greater .
their product should therefore be of order unity .
this calculation then suggests the plausibility of electron drift speeds of order the thermal speed in these current sheets . if the drift speed is of the order of the electron thermal speed , it is also of order or larger than the ion acoustic speed .
it is necessary to stress that this calculation has contained products of several parameters ( such as @xmath128 , etc ) which are imperfectly known , so the net result presented here is similarly uncertain .
however , the conclusion of this section is that a current - driven instability , which would lead to high levels of fluctuating electric and magnetic fields , is a possibility .
this calculation has also been carried out for the case of `` deterministic '' current sheets discussed in section 3.2 .
the details of the calculation are not presented here , but the final result is that the drift - to - thermal speed ratio is somewhat smaller ( a factor of about 0.17 ) , but not enough to alter the qualitative conclusion stated above .
the possible existence of substantial electron drift speeds , comparable to the electron thermal speed , raises the possibility of an interesting observational diagnostic of such current sheets .
@xcite pointed out that an electron distribution carrying a current will have its distribution function distorted in the direction of current flow , and accordingly have a more populated tail than a distribution which carries no current .
this additional tail component makes the plasma more effective at collision excitation of ions to excited states whose energy above the ground state is a few times the electron thermal energy .
@xcite suggested that turbulent current sheets might reveal themselves via enhanced emission line glow as the excited ions radiatively de - excite .
an important parameter determining the intensity of the line radiation is @xmath129 this parameter is approximately equal to the square of the drift speed to thermal speed ratio .
when @xmath130 becomes of the order of a few tenths , the line emission can be substantially enhanced relative to the current - free value ( see figure 9 of * ? ? ?
@xcite found that @xmath131 for turbulence in the interstellar medium , so turbulent enhancement of emission line radiation probably does not occur there .
however , the results presented in this section suggest that this mechanism is much more likely to occur in the solar corona . a complete discussion of the consequences of a current - driven plasma instability for the resistivity within coronal current sheets is beyond the scope of the present paper .
i will only briefly refer to some results in the literature which indicate the effect may be significant .
the issue of instabilities due to high electron drift speeds was discussed in @xcite , who cited results from @xcite .
the remarks made there are still relevant to the present discussion .
the summary of the work of @xcite presented in @xcite is that an electron drift speed greater than @xmath132 could be sufficient for excitation of obliquely - propagating electrostatic ion cyclotron waves .
the role of a current - driven instability in enhanced resistivity was discussed by @xcite .
@xcite developed a fluid theory to explain observations of laboratory z - pinches which showed these structures to be larger than expected on the basis of theory with a spitzer conductivity .
the unexpectedly large size of z pinches suggested that enhanced transport coefficients were present .
@xcite found that the electron drift speed exceeded the ion acoustic speed in the outer edges of the z pinch .
his results ( see figure 3 of @xcite ) showed that the effective resistivity due to lower hybrid waves could exceed the spitzer resistivity by 3 - 4 orders of magnitude .
this enhancement is approaching that needed for thermodynamic relevance of coronal currents , as discussed in section 3.4 .
enhanced resistivity has also been hypothesized to play an important role in magnetic reconnection , allowing reconnection to proceed at a faster rate and produce heating of the plasma . @xcite discussed experiments showing the presence of magnetic fluctuations within the reconnection current sheet on the mrx experiment .
@xcite identify these fluctuations as obliquely - propagating waves arising due to a cross - field current which has a drift speed exceeding the alfvn speed .
@xcite found an enhancement of the resistivity , estimated from the wave force on the electrons , which exceeds the spitzer resistivity by a factor of several .
the relatively modest enhancement of the resistivity relative to that required for coronal relevance , or the results discussed by @xcite , can be attributed to the nearly collisional dynamics of the mrx experiment .
the discussion in @xcite has recently been superceded by the results of @xcite , who now favor perpendicularly - propagating , unstable waves which nonlinearly couple to magnetosonic waves .
it is the magnetosonic waves which determine the resistivity .
this anomalous resistivity is higher than estimated in @xcite .
the above - cited studies are not intended to correspond in detail to the case of joule heating of the solar corona and the highly enhanced resistivity that would be required there .
the investigations of @xcite , @xcite and @xcite are of importance in showing that current filaments with drift speeds comparable to or exceeding the ion acoustic speed can produce wave and turbulence fields that substantially enhance the resistivity .
current sheets or filaments with drift speeds approaching the electron thermal speed would be even more subject to instability , and to a wider range of unstable modes . a final point to be considered in this section
is whether the resistivity in the coronal current sheets could plausibly reach levels necessary for important joule heating .
the arguments in the previous paragraphs have shown that current - driven instabilities could quite plausibly be present , but could they enhance the resistivity by several orders of magnitude ?
a very general expression for the resistivity is @xmath133 where @xmath134 is a collision frequency of some sort .
if collisionless scattering of electrons determines the resistivity , @xmath134 could be as large as the frequency of a high frequency plasma mode . in what follows
, i choose the lower hybrid frequency @xmath135 in this identity , @xmath97 and @xmath136 are , respectively , the ion and electron cyclotron frequencies .
i use the lower hybrid frequency as a proxy for a plasma mode frequency which is above the ion cyclotron frequency , and do not claim that the unstable waves in coronal current sheets are necessarily lower hybrid waves . in support of this approach , @xcite presents estimates of the anomalous collision frequency which are of order @xmath137 , with a multiplicative constant dependent on the drift speed to ion acoustic speed ratio ( see equation ( 2 ) of * ? ? ?
using the coronal magnetic field model of @xcite , we estimate a lower hybrid frequency @xmath138 khz at a heliocentric distance @xmath139 .
substitution of this collision frequency into equation ( 25 ) ( again using the same estimate of the electron density @xmath123 used in equation ( 22 ) ) gives an anomalous resistivity of 6.27 ohm - m .
this exceeds the spitzer resistivity calculated following equation ( 16 ) by approximately 7 orders of magnitude .
this enhancement in the resistivity is comparable to , and in fact exceeds , the factor by which our calculated heating rate must be increased in order to be relevant for the thermodynamics of the corona @xcite .
this brief calculation then suggests that self - enhanced resistivity in coronal current sheets could lead to thermodynamically - relevant levels of joule heating in the corona . 1
. radioastronomical observations reported by @xcite are consistent with coronal currents flowing through amperian loops defined by adjacent lines of sight to different components of a radio source .
there are estimates of currents of @xmath0 and @xmath19 amperes , respectively , on two days .
another interval of high quality data on one of the days yielded an upper limit to the differential faraday rotation , and a corresponding upper limit to the current of @xmath140 amperes .
these data are used as input for a calculation of joule heating of the solar corona .
two models are developed to calculate the joule heating associated with the observed currents . in both models ,
the current is envisioned as being in thin , intense current sheets stretched out along the large - scale coronal magnetic field .
the first views the sheets as arising in the evolution of quasi-2d magnetohydrodynamic turbulence .
the other assumes that current sheets will arise in the coronal plasma , and could show a preference for one sign of the current density .
these derivations are given in sections 3.1 and 3.2 , and provide the formulas for the volumetric heating rates given in equations ( 10 ) and ( 13 ) .
3 . use of these formulas , with observational data from @xcite and plausible independent coronal data , yield an estimated heating rate of @xmath141 watts / m@xmath106 ( @xmath142 ergs / cm@xmath106/sec in cgs units ) . the corresponding heating rate per unit
mass is @xmath143 ergs / gm / sec .
this appears to be smaller than the level necessary to be significant for coronal heating by at least a factor of @xmath1 .
the conclusion to be drawn from point ( 3 ) is that either these currents are irrelevant for coronal heating , or that the true resistivity in the corona exceeds the spitzer value by several orders of magnitude .
resolution of this matter obviously lies in a better understanding of the resistivity in a collisionless plasma .
the same model used to estimate the volumetric and mass heating rates is also used to estimate the electron drift speed in the current sheets .
this drift speed could be comparable to the electron drift speed , and in excess of the ion acoustic speed .
accordingly , current - driven instabilities might be present in these sheets , and the waves driven unstable by these currents might enhance the resistivity to significant levels .
this contention is supported by works in the literature which have shown enhancement of resistivity by current - driven instabilities .
this work was supported at the university of iowa by grant atm03 - 54782 from the national science foundation .
i thank dr .
steven cranmer of the harvard - smithsonian center for astrophysics for extensive commentary and suggestions on an earlier draft of this paper .
i also thank dr .
gary of the los alamos national laboratory and dr .
jack scudder of the university of iowa for discussions .
chittenden , j.p : the effect of lower hybrid instabilities on plasma confinement in fiber z pinches , phys .
plasmas 2 , 1242 , 1995 cranmer , s.r . and
ballegooijen , a.a . : on the generation , propagation , and reflection of alfven waves from the solar photosphere to the distant heliosphere , astrophys
. j. suppl .
s. 156 , 265 , 2005 cranmer , s.r .
, van ballegooijen , a.a . , and edgar , r.j . :
self - consistent coronal heating and solar wind acceleration from anisotropic magnetohydrodynamic turbulence , astrophys .
s. 171 , 520 , 2007 drummond , w.e . and
rosenbluth , m.n . : anomalous diffusion arising from microinstabilities in a plasma , phys .
fluids 5 , 1507 , 1962 efimov , a.i . , bird , m.k . ,
chashei , i.v . , and samoznaev , l.n .
: outer scale of solar wind turbulence deduced from two - way coronal radio sounding experiments , adv . space res .
33 , 701 , 2004 gudiksen , b.v . and nordlund , : an ab initio approach to the solar coronal heating problem , astrophys .
j. 618 , 1020 , 2005 ; _ erratum : _ astrophys . j. 623 , 600 gurnett , d.a . and bhattacharjee , a. : introduction to plasma physics , cambridge university press , 2005 , p430 hollweg , j.v . , bird , m.k . , volland , h. , edenhofer , p. , stelzried , c.t . , and seidel , b.l . : possible evidence for coronal alfven waves , j. geophys .
87 , 1 , 1982 ingleby , l.d . ,
spangler , s.r . , and whiting , c.a .
: probing the large - scale plasma structure of the solar corona with faraday rotation measurements , astrophys . j. 668 , 520 , 2007 krall , n.a . and
trivelpiece , a.w . :
principles of plasma physics , mcgraw - hill , 1973 , p294 kulsrud , r. , ji , h. , fox , w. , and yamada , m. : an electromagnetic drift instability in the magnetic reconnection experiment and its importance for magnetic reconnection , phys .
plasmas 12 , 082301 , 2005 mancuso , s. and spangler , s.r . :
coronal faraday rotation observations : measurements and limits on plasma inhomogeneities , astrophys . j. 525 , 195 , 1999 nicholson , d.r . : introduction to plasma theory , john wiley and sons , 1983 , p4 parker , e.n . : topological dissipation and the small - scale fields in turbulent gases , astrophys . j. 174 , 499 , 1972 peter , h. , gudiksen , b.v . , and nordlund , .
: forward modeling of the corona of the sun and solar - like stars : from a three - dimensional magnetohydrodynamic model to synthetic extreme - ultraviolet spectra , astrophys .
j. 638 , 1086 , 2006 sakurai , t. and spangler , s.r . : the study of coronal plasma structures and fluctuations with faraday rotation measurements , astrophys .
j. 434 , 773 , 1994 spangler , s.r .
: magnetohydrodynamic turbulence and enhanced atomic processes in astrophysical plasmas , phys .
plasmas 5 , 3006 , 1998 spangler , s.r . : two dimensional magnetohydrodynamics and interstellar plasma turbulence , astrophys . j. 522,879 , 1999 spangler , s.r . : the amplitude of magnetohydrodynamic turbulence in the inner solar wind , astrophys . j. 576,997 , 2002 spangler , s.r . : a technique for measuring electrical currents in the solar corona , astrophys .
j. 670 , 841 , 2007 wang , y , kulsrud , r. , and ji , h. : new results on the perpendicularly - propagating electromagnetic lower hybrid drift instability ( lhdi ) , 49th meeting of the aps division of plasma physics , paper tp8.012 , 2007 wang , y , kulsrud , r. , and ji , h. : nonlinear mode coupling calculation on the perpendicularly - propagating electromagnetic instabilities in the mrx , 50th meeting of the aps division of plasma physics , paper gp6.022 , 2008 wohlmuth , r. , plettemeier , d. , edenhofer , p. , bird , m.k .
, efimov , a.i . , samoznaev , l.n . , and chashei , i. : radio frequency fluctuation spectra during the solar conjunctions of the ulysses and galileo spacecraft , space sci . rev .
97 , 9 , 2001 zank , g.p . and
matthaeus , w.h . : the equations of reduced magnetohydrodynamics , j. plasm .
48 , 85 , 1992 | recent radioastronomical observations of faraday rotation in the solar corona can be interpreted as evidence for coronal currents , with values as large as @xmath0 amperes @xcite .
these estimates of currents are used to develop a model for joule heating in the corona .
it is assumed that the currents are concentrated in thin current sheets , as suggested by theories of two dimensional magnetohydrodynamic turbulence .
the spitzer result for the resistivity is adopted as a lower limit to the true resistivity .
the calculated volumetric heating rate is compared with an independent theoretical estimate by @xcite .
this latter estimate accounts for the dynamic and thermodynamic properties of the corona at a heliocentric distance of several solar radii .
our calculated joule heating rate is less than the cranmer et al estimate by at least a factor of @xmath1 .
the currents inferred from the observations of @xcite are not relevant to coronal heating unless the true resistivity is enormously increased relative to the spitzer value . however , the same model for turbulent current sheets used to calculate the heating rate also gives an electron drift speed which can be comparable to the electron thermal speed , and larger than the ion acoustic speed .
it is therefore possible that the coronal current sheets are unstable to current - driven instabilities which produce high levels of waves , enhance the resistivity and thus the heating rate . in a recent paper ,
@xcite reported radioastronomical observations which were consistent with the presence of coronal currents in the range of hundreds of megaamperes to a few gigaamperes .
this measurement was made using faraday rotation observations of a radio source occulted by the corona , and the coronal plasma probed was at heliocentric distances of 5.2 to 6.7 @xmath2 .
in the present paper , i discuss the implications of these observations for the process of coronal heating by joule heating .
as discussed in @xcite the currents reported ( and summarized in section 2 below ) correspond to the _ net _ current within an amperian loop defined by the two , closely - spaced lines of sight through the corona to the different parts of the radio source .
the measured net current could be , and probably is , a residual due to numerous current filaments with alternate positive and negative current density within the amperian loop .
this topic is of interest because joule heating has been identified as the primary mechanism for heating the closed - field part of the corona @xcite .
the purpose of this paper is to make model - dependent estimates of the heating rate due to joule dissipation of these currents .
as expected , the calculation involves introduction of several `` imponderables '' , i.e. physical characteristics of the turbulence in the corona which are poorly constrained by observations , but which play an important role in coronal heating .
i feel this exercise is worthwhile in identifying coronal parameters which are important in coronal heating , so that they can be targeted for future observational investigations .
the outline of this paper is as follows . in section 2 ,
i briefly summarize the observational results of @xcite which were the basis of the estimates of the coronal current .
section 3 is the most important part of the paper ; it introduces a model for current - carrying coronal turbulence , and identifies the most important characteristics of this turbulence .
a glossary of the variables and parameters introduced in this discussion is given in table 1 .
this model is used to obtain an estimate of the volumetric heating rate due to joule heating .
section 4 briefly considers the possibility that current densities in these sheets could be large enough to generate turbulence via current - driven instabilities , and thus produce high wave levels which enhance the resistivity and thus the joule heating rate .
section 5 summarizes what has been learned from this exercise and presents conclusions . |
given a bounding class @xmath2 ( the definition of which we recall below ) and group with word - length function @xmath3 , one may construct the rapid decay algebra @xmath8 .
this algebra was introduced in @xcite ; it is a frchet algebra whenever the bounding class @xmath2 is equivalent to a countable class ( there are no known cases when this does nt occur ) .
when @xmath9 , the class of polynomial bounding functions , this algebra is precisely the rapid decay algebra @xmath10 introduced by jolissaint in @xcite , @xcite . because @xmath8 is a rapid decay algebra formed using weighted @xmath5-norms one has , associated to each conjugacy class @xmath4 , summands @xmath11 of @xmath12 , where @xmath13 , and thus projection maps @xmath14 and similarly for cohomology .
recall that a conjugacy class @xmath15 is _ elliptic _ if @xmath15 has finite order , and _ non - elliptic _ if @xmath16 has infinite order . as in @xcite
, one may then posit - for a given bounding class @xmath2 and group with word - length @xmath3 - the following generalization of the strong bass conjecture : for each non - elliptic conjugacy class @xmath15 , the image of the composition @xmath17 is zero .
when @xmath9 , the subalgebra @xmath18 is smooth in @xmath7 by @xcite , @xcite ; in this case the above conjecture can be restated as : for each non - elliptic conjugacy class @xmath15 , the image of the composition @xmath19 is zero .
the conjecture @xmath5-srbc is certainly the most important special case of the first more general conjecture ; as shown in the appendix of @xcite these conjectures follow from the rational surjectivity of an appropriately defined baum - connes assembly map .
thus any counter - example to one of these conjectures would , in turn , provide a counterexample to a baum - connes type conjecture holding for the corresponding rapid decay algebra .
it is a result due to burghelea that one has decompositons @xmath20}}){{\,\cong\,}}\underset{<x>\in
< g>}{\oplus } hh_*({{\mathbb{c}[g]}})_{<x>},\quad hc_*({{\mathbb{c}[g]}}){{\,\cong\,}}\underset{<x>\in < g>}{\oplus } hc_*({{\mathbb{c}[g]}})_{<x>}\ ] ] and for each conjugacy class @xmath15 an isomorphism @xmath20}})_{<x>}{{\,\cong\,}}h_*(bg_x;\c)\ ] ] where @xmath21 denotes the centralizer of @xmath22 .
this last fact allows one to derive corresponding descriptions of @xmath23}})_{<x>}$ ] ; specifically , for non - elliptic classes one has @xmath24}})_{<x>}{{\,\cong\,}}h_*(b(g_x/(x));\c)\ ] ] where @xmath25 denotes the infinite cyclic subgroup generated by @xmath16 , and @xmath26 the quotient group .
the standard way of arriving at this result is via the isomorphism of sets @xmath27 where @xmath28 is the set of elements conjugate to @xmath16 , and the correspondence is given by @xmath29 .
when @xmath30 is equipped with word - length @xmath31 , this map becomes a @xmath2-g - morphism of weighted right @xmath2-g - sets in the sense of @xcite , but one whose inverse is in general unbounded .
this leads to the notion of @xmath2-solvable conjugacy bound , introduced in @xcite - the condition that the map in ( [ eqn : fundmap ] ) is a @xmath2-g - isomorphism - which in turn allows for the identification of the summands @xmath11 in terms of the rapid decay homology of @xmath32 or @xmath33 .
unfortunately , this condition is very restrictive , as it requires the group to have a conjugacy problem which is solvable in @xmath2-bounded time .
in particular , when @xmath9 , it is difficult to go much beyond hyperbolic groups with this restriction in place ( the main result of @xcite was verification of a relative version of the @xmath5-bc conjecture in the presence of relative hyperbolicity ) . with this in mind , we may now state our main result , by which the solvable conjugacy bound constraint of @xcite is removed . the consequence is a complete characterization of the conjugacy class summands of the hochschild and cyclic ( co)-homology groups of @xmath8 for any group with word - length @xmath3 and bounding class @xmath2 . recall that the of a subgroup with length function @xmath34 of a group with length function @xmath35 is the function @xmath36 defined by @xmath37 .
let @xmath3 be a discrete group equipped with proper word - length function and @xmath2 a countable bounding class .
then for each conjugacy class @xmath4 there are isomorphisms in topological hochschild ( co)-homology @xmath38 these isomorphisms imply the existence of isomorphisms in topological cyclic ( co)-homology @xmath39 here the weighting on @xmath32 and @xmath33 comes from the induced word - length function on @xmath21 and the corresponding quotient word - length function on @xmath26 , while @xmath40 , @xmath41 denote the @xmath2-bounded ( co)-homology groups as defined in @xcite ( and reviewed below ) . for the cyclic groups , the @xmath42$]-module and comodule structures of the terms on the right are exactly as in the case of the group algebra for both elliptic and non - elliptic classes ( compare @xcite ) .
finally , @xmath43 refers to the distortion of the infinite cyclic subgroup @xmath44 of @xmath30 ; the condition @xmath45 means that the distortion of @xmath44 is bounded above by a function in @xmath2 .
this theorem allows for the construction of a class of groups satisfying the @xmath2-srbc conjecture , modeled on the class of groups originally considered by the first author in @xcite and independently by chadha and passi in @xcite , and extended to a slightly larger class by emmanouil in @xcite .
it also allows for direct verification of the @xmath5-srbc conjecture for a large and important class of groups , as given by the following theorem .
recall that a group @xmath30 satisfies the if @xmath46}})_{<x>}\to hc^{*+2}({{\mathbb{c}[g]}})_{<x>}$ ] is a nilpotent operator for all non - elliptic classes @xmath4 .
similarly , @xmath30 satisfies the if @xmath47 is a nilpotent operator for all non - elliptic classes @xmath4 . by @xcite , if @xmath30 satisfies the @xmath2-nilpotency condition , then @xmath2-srbc is true for @xmath30 . finally , @xmath3 is @xmath2-isocohomological if the comparison map in cohomology @xmath48 is an isomorphism ( @xcite ) .
let @xmath3 be a semihyperbolic group in the sense of @xcite . if @xmath30 satisfies the nilpotency condition
, then it satisfies the @xmath2-nilpotency condition for any bounding class @xmath2 containing @xmath49 .
in particular , if @xmath26 has finite cohomogical dimension over @xmath50 for each non - elliptic class @xmath15 , then @xmath30 satisfies @xmath2-srbc for all bounding classes @xmath2 containing @xmath49 .
by @xcite the centralizer subgroups @xmath51 are quasi - convex subgroups of the ( linearly ) combable group @xmath30 , hence each @xmath51 is combable and quasi - isometrically embedded into @xmath30 . by @xcite , @xcite , or @xcite
each @xmath51 is @xmath2-isocohomological for any bounding class containing @xmath49 .
moreover , @xmath44 is quasi - isometrically embedded in @xmath21 - hence also in @xmath30 - for each conjugacy class @xmath4 , implying its distortion in @xmath30 is at most polynomial .
thus for each non - elliptic conjugacy class @xmath52 we have isomorphisms @xmath53 by burghelea s result together with theorem a above , the nilpotency condition for @xmath30 implies the @xmath2-nilpotency condition for @xmath3 whenever @xmath2 is a bounding class containing @xmath49 .
let @xmath54 denote the set of non - decreasing functions @xmath55 .
suppose @xmath56 is a function of sets , and @xmath57 .
@xmath2 is weakly closed under @xmath58 if for each @xmath59 there is @xmath60 with @xmath61 .
a bounding class is a subset @xmath57 satisfying 1 .
it contains the constant function @xmath62 .
it is weakly closed under the operations of taking positive rational linear combinations .
it is weakly closed under the operation @xmath63 for @xmath64 , where @xmath65 denotes the linear bounding class @xmath66 .
a bounding class is if it is weakly closed under the operation of composition of fuctions . naturally occurring classes besides @xmath65 are @xmath67 , @xmath68 the set of polynomials with non - negative rational coefficients , and @xmath69 .
bounding classes were discussed in detail in @xcite .
the bounding classes we will be considering will all contain @xmath65 .
recall that a function @xmath70 , with @xmath71 discrete , is if the preimage of any bounded subset is finite .
a @xmath72 will refer to a countable discrete set @xmath71 together with a proper function @xmath73 .
a morphism between weighted sets @xmath74 is if @xmath75 given two weighted sets @xmath72 and @xmath76 , the cartesian product @xmath77 carries a canonical weighted set structure , @xmath78 with weight function @xmath79 .
suppose further that @xmath71 is a right @xmath30-set and @xmath80 is a left @xmath30-set .
let @xmath81 denote the quotient of @xmath77 by the relation @xmath82 .
there is a natural weight , @xmath83 , on @xmath81 induced by the weight @xmath84 on @xmath77 as follows .
@xmath85 ) = \min_{g \in g } r ( xg , g^{-1 } x ' ) .\ ] ] given a weighted set @xmath72 and @xmath86 , the seminorm @xmath87 on @xmath88 is given by @xmath89 .
we will mainly be concerned with the case @xmath90 is a discrete group endowed with a length function @xmath31 .
the length function @xmath31 is called a word - length function ( with respect to a generating set @xmath91 ) if @xmath92 and there is a function @xmath93 with @xmath94 when @xmath91 is finite , taking @xmath95 produces the standard word - length function on @xmath30 .
recall that a simplicial object in a category @xmath96 is a covariant functor @xmath97 , from the simplicial category into @xmath96 .
that is , for each @xmath98 @xmath99 , and there exist degeneracy morphisms in @xmath100 and face morphisms in @xmath101 which satisfy the usual simplicial identities . the augmented simplicial category , @xmath102 , is obtained by adding another object @xmath103 $ ] to @xmath104 , and a single face morphism in @xmath105 , [ 0])$ ] . an augmented simplicial object in a category @xmath96 is a covariant functor @xmath106 .
this consists of a simplicial set @xmath107 , @xmath98 , as well as @xmath108 with an augmentation @xmath109 .
an augmented simplicial group with word - length @xmath110 consists of an augmented simplicial group @xmath111 , with @xmath112 a word - length function on @xmath113 for all @xmath114 .
if @xmath2 is a bounding class , the augmented simplicial group with word - length @xmath110 is @xmath2-bounded if all face and degeneracy maps , including the augmentation map , are @xmath2-bounded with respect to the word - lengths , @xmath115 .
then @xmath110 is a if i ) @xmath116 is a countably generated free group with @xmath117-valued word - length metric @xmath112 generated by a proper function on the set of generators for @xmath113 for all @xmath98 , and ii ) @xmath111 admits a simplicial set contraction @xmath118 which is a @xmath2-bounded set map in each degree .
every countable discrete group @xmath30 admits a type @xmath2 resolution @xmath111 with @xmath119 .
in fact , starting with @xmath30 , the resolution can always be constructed so that the face and degeneracy maps , the augmentation , and the simplicial contraction are all linearly bounded ( * ? ? ?
* appendix ) . the cyclic bar construction , @xmath120 .
let @xmath30 be a countable group and let @xmath31 be a word - length function on @xmath30 .
we define @xmath121\mapsto g^{n+1}\}_{n\geq 0}$ ] with @xmath122 the simplicial weight is given by @xmath123 .
@xmath124 is a @xmath2-bounded simplicial set .
the bar resolution , @xmath125 . recall that the non - homogeneous bar resolution of @xmath30 is @xmath126\mapsto g^{n+1}\}_{n\geq 0}$ ] with @xmath127 & = [ g_0,\dots , g_ig_{i+1},\dots , g_n],\ , 0 \leq i \leq n-1,\\ \partial_n[g_0,\dots , g_n ] & = [ g_0,\dots , g_{n-1}],\\ s_j[g_0,\dots , g_n ] & = [ g_0,\dots , g_j,1,g_{j+1},\dots , g_n]\end{aligned}\ ] ] the simplicial weight function on @xmath125 is given by @xmath128 ) = \sum_{i=0}^n l(g_i)$ ] .
the left @xmath30-action is given , as usual , by @xmath129 = [ gg_0,g_1,\dots , g_n]$ ] .
note that with respect to the given weight function and action of @xmath30 , @xmath125 is a @xmath2-bounded simplicial @xmath30-set for any @xmath2 .
for @xmath2 a bounding class and @xmath72 a weighted set , @xmath130 will denote the collection of @xmath2-bounded functions on @xmath71 .
that is , @xmath130 consists of all functions @xmath131 such that there is @xmath132 with @xmath133 for all @xmath134 .
dually , @xmath135 will consist of all @xmath131 such that for all @xmath132 , the sum @xmath136 is finite .
this is a completion of the collection of finitely supported chains on @xmath71 with respect to a family of seminorms on @xmath135 given by @xmath137 for @xmath132 .
when @xmath72 is @xmath3 , a discrete group with a word - length function , @xmath8 is the @xmath2-rapid decay algebra of @xmath30 .
when @xmath2 is equivalent to a countable bounding class the seminorms give @xmath8 the structure of a frchet algebra . for @xmath138 a weighted simplicial set and @xmath2 a bounding class , set @xmath139 , the completion of @xmath140
if each of the face maps of @xmath141 are @xmath2-bounded , then the boundary maps @xmath142 extend to @xmath2-bounded maps @xmath143 .
the homology of the resulting bornological complex @xmath144 is the @xmath2-bounded homology of @xmath141 , @xmath145 .
similarly the cohomology of the bornological cochain complex @xmath146 is the @xmath2-bounded cohomology of @xmath141 , @xmath147 . in the case of the weighted cyclic
bar construction , @xmath120 we have identifications from @xcite , @xmath148 a group with length function @xmath3 has @xmath2-cohomological dimension ( @xmath2-cd ) @xmath149 if there is a projective resolution of @xmath150 over @xmath8 of length at most @xmath151 .
here we do not require finite generation over @xmath8 in any degree , but as in @xcite we require a @xmath2-bounded @xmath150-linear contracting homotopy .
we denote by @xmath152 the @xmath153 functor on the category of bornological @xmath8-modules .
that is , for a bornological @xmath8-module , @xmath154 is obtained by taking a projective resolution of @xmath150 over @xmath8 , apply @xmath155 , and take the cohomology of the resulting cochain complex .
we have the following analogue of lemma viii.2.1 of @xcite . for a group with length function @xmath3 and @xmath2 a composable bounding class , the following are equivalent .
1 . @xmath2-cd @xmath156 .
@xmath157 for all @xmath158 .
3 . @xmath159 .
if @xmath160 is any sequence of bornological @xmath8-modules admitting a @xmath2-bounded @xmath150-linear contracting homotopy with each @xmath161 projective then @xmath6 is projective .
if @xmath162 is any sequence of bornological @xmath8-modules admitting a @xmath2-bounded @xmath150-linear contracting homotopy with each @xmath163 free then @xmath6 is free .
\(1 ) iff ( 2 ) iff ( 3 ) iff ( 4 ) as in @xcite needing virtually no modification .
( 4 ) iff ( 5 ) follows from the bornological eilenberg swindle .
there is always , for each conjugacy class @xmath4 , an isomorphism of simplicial sets @xmath164 ( here @xmath21 is the centralizer of the element @xmath165 . )
this map induces an isomorphism @xmath166 when @xmath16 is not central in @xmath30 , this isomorphism depends on the particular choice of representative for each conjugacy class .
recall from @xcite that a conjugacy class @xmath15 has a if there is an @xmath60 such that , for all @xmath167 the set of elements in @xmath30 conjugate to @xmath16 , there is a @xmath168 with @xmath169 and @xmath170 .
this is equivalent to the statement that the natural map @xmath171 defined by @xmath172 is a @xmath2-bounded isomorphism .
when @xmath173 has a @xmath2-bounded conjugator length , the map @xmath174 is a @xmath2-bounded isomorphism of weighted simplicial sets , which in turn induces an isomorphism of cohomology groups @xmath175 we are interested in the case where there is no such conjugacy bound ( which is most of the time ) . in this case
, the map @xmath176 still exists as a @xmath2-bounded map and an isomorphism of simplicial sets , however a choice of @xmath2-bounded inverse may not exist .
assume given a type @xmath2 simplicial resolution @xmath177 .
associated to this simplicial type resolution is a simplicial chain complex equipped with augmentation map @xmath178\mapsto { { \mathcal{b}}}c_*(n^{cy}_{\bullet}(\gamma_n))\}_{n\ge 0 } { \twoheadrightarrow}{{\mathcal{b}}}c_*(n^{cy}_{\bullet}(g)).\ ] ] by theorem 1 of @xcite , this simplicial complex is of resolution type .
consequently , the double complex associated to @xmath179 maps via the augmentation map to @xmath180 by a map inducing an isomorphism in @xmath2-bounded homology and cohomology . now consider the following diagram of simplicial sets : @xmath181 \ar@<-1ex>[d ] \ar@<-2ex>[d ] & \\
n^{cy}_{\bullet}(\gamma_1 ) \ar[d ] \ar@<-1ex>[d ] \ar@<-1ex>[u ] \ar@<-2ex>[u ] & \\
n^{cy}_{\bullet}(\gamma_0 ) \ar@<-1ex>[u ] \ar@{>>}[d ] \ar@{<-}[r]_-{{{\,\cong\,}}}^-{\phi } & \underset{{y \in < \gamma_0>}}{\coprod } ( \gamma_{0})_y \backslash \gamma_0 \underset{\gamma_0}{\times } e(\gamma_0)_{\bullet } \ar@{>>}[d ] \\
n^{cy}_{\bullet}(g ) & \underset{{<x >
\in < g>}}{\coprod } g_{x } \backslash g \underset{g}{\times } eg_{\bullet } \ar[l ] } \ ] ] for the map on the right , explicit centralizer subgroups @xmath21 and @xmath182 are chosen as follows : for each @xmath4 , fix a basepoint @xmath16 of the set @xmath183 . then for each @xmath184 , we choose a basepoint @xmath185 of minimal word - length so that the augmentation map @xmath186 induces a map of basepointed sets @xmath187 the map @xmath58 is a linearly bounded isomorphism with linearly bounded inverse ( hence a @xmath2-bounded isomorphism for all @xmath2 ) , as @xmath188 is a free group equipped with word - length metric . this isomorphism will allow us to construct a simplicial resolution of @xmath189 .
define the bisimplicial set @xmath190 to be : @xmath191 the face maps and degeneracy maps between @xmath192 and @xmath193 , for @xmath194 are given by the corresponding maps @xmath195 in @xmath196 induced by the face and degeneracy maps of @xmath197 .
the face and degeneracy maps involving @xmath198 are determined by @xmath199 , @xmath200 for @xmath201 and @xmath202 , @xmath203 .
these maps satisfy the simplicial identities and are all @xmath2-bounded .
in particular we have a @xmath2-bounded bisimplicial @xmath30 set @xmath204 \mapsto x_{\bullet n } \}$ ] .
there is a @xmath2-bounded isomorphism of bisimplicial @xmath30-sets @xmath205
\mapsto x_{\bullet n } \}_{n \geq 0 } \to \ { [ n ] \mapsto n^{cy}_{\bullet } ( \gamma_n ) \}_{n \geq 0}\ ] ] as the augmentation @xmath206 is surjective , the induced map @xmath207 is a surjective morphism of weighted simplicial sets , which moreover is bounded .
our aim is to show that the induced and given weight functions on the set to the right are equivalent . to this end , we define a section @xmath208 of @xmath209 by using a set - theoretic splitting @xmath210 of the augmentation @xmath186 which is length - minimizing among all basepoint - preserving maps ; precisely , if @xmath211 is the basepoint of @xmath183 , then i ) @xmath212 should be the basepoint of @xmath213 , and ii ) @xmath214 should have minimal word - length among all basepoints @xmath215 for which @xmath216 .
because of the manner in which the basepoints in @xmath188 were chosen , such a section always exists , and moreover it can be chosen so as to be length - preserving on basepoints .
of course , it can always be chosen so as to be length - preserving away from basepoints . [ lem : bddsplitting ] the section @xmath208 induces a splitting @xmath217 of @xmath209 on the level of graded sets which is @xmath2-bounded .
denote by @xmath218 the proper length function on @xmath188 and by @xmath31 the proper length function on @xmath30 .
fix a conjugacy class @xmath219 and consider an element of @xmath220 , @xmath221 $ ] .
we may assume @xmath222 is a minimal weight representative of the class , and that @xmath223 is a minimal length element of the coset @xmath224 , so the weight of @xmath221 $ ] is @xmath225 .
given that we have arranged the basepoints for the sets @xmath226 so as to make the augmentation basepoint - preserving , the induced map @xmath227 on the right has the form @xmath228 = [ g_x \epsilon(\gamma ) \times ( \epsilon(\gamma_0 ) , \epsilon(\gamma_1 ) , \ldots , \epsilon(\gamma_n))]\ ] ] where @xmath229 . the weight of this class is no more than @xmath230 . if @xmath231 is a bounding function for @xmath232 , then we see that the weight of this class is no more than @xmath233 .
thus @xmath227 is bounded by @xmath234
. for each @xmath151-simplex @xmath235 \in g_x \backslash g \underset{g}{\times } eg_n$ ] , fix a minimal weight representative @xmath236 with @xmath237 of minimal length in the coset @xmath238 .
the desired section on @xmath151-simplices is defined by @xmath239 = [ ( \gamma_0)_{\tilde{s}(x ) } \tilde{s}(g ) \times ( \tilde{s}(g_0 ) , \tilde{s}(g_1 ) , \ldots , \tilde{s}(g_n))].\ ] ] note that the collection of maps @xmath217 will not , in general , define a map of simplicial sets , because the original section @xmath208 can not be chosen to be a homomorphism unless @xmath30 itself is free .
however , it is clear that @xmath240 is a set - theoretic splitting of @xmath227 on @xmath151-simplicies which moreover is bounded . to see this , note that the weight of @xmath241 $ ] can be no greater than @xmath242
thus if @xmath208 is bounded by @xmath243 , then @xmath240 is bounded by @xmath244 .
as we have indicated , there are two basic approaches to endowing @xmath189 with a weight function .
one could use the weight induced by the length structure on @xmath30 itself , or one could use the weights induced by the surjection @xmath227 .
however , the existence of a bounded section guaranteed by the previous lemma shows that these two methods yield @xmath2-equivalent weight structures .
[ thm : res ] if @xmath177 is a type @xmath2 resolution of @xmath30 , then @xmath245 is a @xmath2-bounded augmented simplicial set for which the bisimplicial set @xmath204 \to x_{\bullet , n } \}_{n \geq 0}$ ] yields isomorphisms in homology and cohomology @xmath246 the sections coming from the previous lemma , together with those arising from the type @xmath2 resolution , imply that the augmented simplicial topological chain complex @xmath247\mapsto \begin{cases } { { \mathcal{b}}}c_*(x_{\bullet n})\quad n\ge 0\\ { { \mathcal{b}}}c_*\left(\underset{{<x > \in g}}{\coprod } g_x \backslash g \underset{g}{\times } eg_\bullet\right)\quad
n = -1 \end{cases}\ ] ] is of resolution type .
the result then follows as in ( * ? ? ?
* theorem 2 ) .
there are isomorphisms in @xmath2-bounded homology and cohomology @xmath248 the weighted bisimplicial set @xmath190 is @xmath2-boundedly isomorphic to the weighted bisimplicial set @xmath204 \to n^{cy}_{\bullet}(\gamma_n)\}_{n \geq 0}$ ] , and the augmentation map @xmath249 induces an isomorphism in @xmath2-bounded homology and cohomology .
the result follows then from theorem [ thm : res ] .
[ cor : globalisom ] there is an isomorphism in @xmath2-bounded cohomology @xmath250 by ( * ? ?
1.4.5 ) there is an isomorphism @xmath251 where each centralizer subgroup @xmath21 is equipped with the induced word - length function coming from the embedding into @xmath30 .
the previous corollary gives the result .
for each non - elliptic conjugacy class @xmath173 with @xmath252 , there is an isomorphism @xmath253 where @xmath26 is equipped with the word - length function induced by the projection @xmath254 .
the isomorphism of corollary [ cor : globalisom ] splits over conjugacy classes @xmath255 when @xmath256 , the connes - gysin sequences in @xcite for the summand @xmath257 along with this isomorphism for @xmath258 gives the result . in the case
@xmath259 , it is almost certainly true that @xmath260 . to conclude this , however , one needs to know that if @xmath261 is embedded into a group with distortion greater than @xmath2 , then @xmath262 is trivial for @xmath263 ( where @xmath264 is equipped with the word - length function induced by the embedding ) .
this is conjectured to be true , but is currently unverified .
as noted , the results of this section were previously known only for groups with conjugacy classes satisfying @xmath2-bounded conjugacy length bounds of @xcite .
they may be interpreted as saying that , even when the conjugacy problem for @xmath30 can not be solved in an appropriately bounded timeframe , or even solved at all , it can always be solved `` up to bounded homotopy '' .
we start with a technical lemma .
we follow the setup of ( * ? ? ?
* theorem 1.1.12 ) .
denote by @xmath269 the length function on @xmath30 , which by restriction is the length function on @xmath270 , and by @xmath271 the quotient length function on @xmath272 .
as @xmath267 , there is a finite length free resolution of @xmath150 over @xmath273 , with @xmath2-bounded @xmath150-linear splittings .
denote this resolution by @xmath274 , let @xmath275 be a free resolution of @xmath150 over @xmath276 , and set @xmath277 .
consider the spectral sequence associated to the column filtration .
for @xmath285 , @xmath286 , so @xmath287 for @xmath285 .
similarly , by the finiteness condition on @xmath288 , @xmath289 for @xmath290 .
consequently , @xmath283 whenever @xmath291 . as a consequence , @xmath292 for @xmath293 for all coefficients @xmath279 .
let @xmath96 be the collection of all countable groups @xmath30 which satisfy the nilpotency condition , and let @xmath294 be the collection of all groups with length function @xmath3 which satisfy the @xmath2-nilpotency condition .
any group @xmath295 satisfies @xmath2-srbc , just as any @xmath296 satisfies sbc .
moreover if @xmath3 satisfies @xmath2-srbc , then @xmath30 satisfies srbc .
recall that a group with length function @xmath297 is said to be ( @xmath2-ic ) if the inclusion @xmath298}}\to { { \mathcal{h}_{{\mathcal{b}},l}\left(g\right)}}$ ] induces isomorphisms @xmath299 for all @xmath300 .
the group is ( @xmath2-sic ) if the maps @xmath301 are isomorphisms for every @xmath300 and every bornological @xmath8-module @xmath279 .
suppose the discrete group @xmath30 lies in @xmath96 and @xmath31 is a proper length function on @xmath30 .
if for each non - elliptic conjugacy class @xmath173 , the centralizer @xmath302 is @xmath2-ic and the embedding @xmath303 is at most @xmath2-distorted , then @xmath304 .
as shown in the introduction this class contains all semihyperbolic groups @xmath30 with word - length function @xmath31 , which satisfy the nilpotency condition .
we remark that this , in turn , contains all word - hyperbolic and finitely generated abelian groups . 1 .
@xmath30 has finite @xmath2-cd .
2 . for every non - elliptic conjugacy class @xmath173 , @xmath307 has finite @xmath2-cd in the induced length function .
3 . for every non - elliptic conjugacy class @xmath173
, the distortion of @xmath44 as a subgroup of @xmath30 is @xmath2-bounded .
we follow the proof of theorem 4.3 of @xcite .
( 1 ) follows from @xcite after noting that if @xmath3 is a group with length function and @xmath312 , any projective resolution of @xmath150 over @xmath8 is also a projective resolution of @xmath150 over @xmath313 .
abcd j. alonso , m. bridson .
`` semihyperbolic groups '' .
london math .
soc._(3 ) * 70 * , pp .
56 114 , 1995 .
_ cohomology of groups_. graduate texts in math .
87 , springer verlag , 1972 .
d. burghelea .
`` the cyclic homology of the group rings '' .
_ comm . math _ * 60 * , pp .
354 365 , 1985 g.k .
chadha , i.b.s . passi .
`` centralizers and homological dimension '' .
alg . _ * 22*(14 ) , pp . 57035708 , 1994 . i. emmanouil .
`` on a class of groups satisfying bass conjecture '' .
_ * 132 * , pp . 307 330 , 1998 .
`` nilpotency of connes periodicity operator and the idempotent conjectures '' . _
k - theory _ * 9 * pp .
59 75 , 1995 .
r. ji , c. ogle , b. ramsey .
`` relatively hyperbolic groups , rapid decay algebras , and a generalization of the bass conjecture '' .
_ jncg _ , * 4 * ( 1 ) , pp . 83 124 , 2010 .
r. ji , c. ogle , b. ramsey .
`` @xmath2-bounded cohomology and applications '' .
preprint , arxiv:1004.4677v4 .
p. jolissaint .
_ les fonctions dcroissance rapide dans les @xmath314-algbres rsduites de groupes_. thesis , univ . of geneva .
p. jolissaint .
`` k - theory of reduced @xmath314-algebras and rapidly decreasing functions on groups '' .
_ k - theory _ * 2 * , pp .
723 735 , 1989 .
`` polynomially bounded cohomology and discrete groups '' .
_ jpaa _ * 195 * , pp .
173 209 , 2005 . c. ogle .
`` polynomially bounded cohomology and the novikov conjecture '' .
preprint , arxiv:1004.4680v4 .
`` combable groups have group cohomology of polynomial growth '' . _ quart .
_ * 57 * , pp . 241 261 , 2006 . | we show that the technical condition of solvable conjugacy bound , introduced in @xcite , can be removed without affecting the main results of that paper .
the result is a burghelea - type description of the summands @xmath0 and @xmath1 for any bounding class @xmath2 , discrete group with word - length @xmath3 and conjugacy class @xmath4 .
we use this description to prove the conjecture @xmath2-srbc of @xcite for a class of groups that goes well beyond the cases considered in that paper .
in particular , we show that the conjecture @xmath5-srbc ( the strong bass conjecture for the topological @xmath6-theory of @xmath7 ) is true for all semihyperbolic groups which satisfy srbc , a statement consistent with the rationalized bost conjecture for such groups . |
experiments with the clas detector at thomas jefferson national accelerator facility ( jlab ) have accumulated extensive and accurate data of meson electroproduction reactions on protons .
detailed experimental data have now become available for the @xmath5 , @xmath6 and @xmath7 exclusive channels @xcite .
in addition first electroproduction data for the channels with smaller cross sections , such as the @xmath8 , and @xmath9 , have also been obtained with clas @xcite .
review of these clas experimental results may be found in the papers @xcite .
these data combined will allow us to determine the electromagnetic transition amplitudes for the majority of nucleon resonances ( @xmath10 ) in a wide range of photon virtualities @xmath2 from 0.2 to 5.0 gev@xmath3 .
the @xmath2 evolution of these amplitudes contains information on the relevant degrees of freedom in the nucleon resonance structure at varying distance scales .
it also allows to explore the strong interaction mechanisms responsible for baryon formation and their relationship to qcd @xcite .
the @xmath11 and @xmath12 exclusive channels are two major contributors to the @xmath13 and @xmath14 reactions in the resonance excitation region .
they are strongly coupled through hadronic interactions of the final @xmath11 and @xmath12 states @xcite thus an understanding of the reaction mechanisms for these two channels is vital for extracting the @xmath15-@xmath10 transition amplitudes at various photon virtualities .
it is also a necessary first step towards the exploration of @xmath16 s in the mass region above 1.5 gev using other exclusive channels with smaller cross sections , such as @xmath8 , @xmath17 , and @xmath9 .
the cross sections of these weaker channels can be affected considerably by the @xmath11 and @xmath12 channels because of unitarity condition .
the mechanisms of the @xmath11 electroproduction have been extensively studied .
the most comprehensive recent studies are by the efforts at mami @xcite , gwu @xcite and jlab @xcite .
in contrast , the study of @xmath12 electroproduction mechanisms is still limited . in this work
we try to improve the situation by analyzing the recently published clas data of @xmath0 channel @xcite .
the world data on @xmath7 electroproduction in the resonance excitation region were rather scarce before the clas experiments .
they were limited to integrated cross sections , various invariant mass distributions , and @xmath18 angular distributions @xcite .
furthermore , the large binning in @xmath4 and @xmath2 does not allow the determination of @xmath16 parameters from these data .
first detailed measurements of @xmath0 electroproduction cross sections with clas were reported in @xcite . the new clas @xmath0 data @xcite use binnings of @xmath19 mev and @xmath20 0.05 gev@xmath21 , and are much more precise than the previous data .
for the first time nine sets of 1-fold differential cross sections were determined in the region of 1.3 @xmath1 w @xmath1 1.6 gev and 0.25 @xmath1 q@xmath3 @xmath1 0.60 gev@xmath3 .
they consist of invariant mass distributions of the final @xmath22 , @xmath23 , and @xmath24 and the angular distributions for all final state particles , as will be described in the section [ xsect ] .
all cross sections measured in this reaction and in others clas experiments may be found in @xcite .
several models have been developed for analyzing the double pion photoproduction @xcite and electroproduction @xcite reactions in the nucleon resonance excitation region , beginning with the pioneering effort of @xcite .
they are based on the tree - diagrams of effective meson - baryon lagrangians and have been reasonably successful in describing the very limited data of the fully integrated cross sections and the invariant mass distributions .
it remains to be seen , to what extent these models can describe the recent clas data .
a comparison between these model predictions and @xmath12 cross sections ( e.g. @xcite ) clearly showed the necessity for further improvements of reaction models in order to isolate the resonant contribution and to determine nucleon resonance parameters from the data .
comprehensive clas data on various differential @xmath0 cross open up new opportunities for phenomenological data analysis . by studying the kinematic dependencies of the differential cross section and their correlations we are able to establish the presence and strength of the relevant reaction mechanisms and to achieve reliable separation of resonant and non - resonant contributions .
in this paper we apply a phenomenological approach developed in the past several years by the jefferson laboratory - moscow state university ( jm ) collaboration @xcite , to analyze the clas data @xcite on @xmath0 electroproduction at @xmath4 @xmath1 1.6 gev . within the jm model developed up to 2005 ,
called jm05 , the major part of @xmath0 production at @xmath4 @xmath1 1.6 gev is due to contributions from the @xmath25 isobar channels . the @xmath26 ( 1232 )
resonance is clearly seen in all @xmath23 mass distributions at w @xmath27 1.4 gev , while other mass distributions do not show any structures .
the contributions from all other isobar channels @xmath28 , @xmath29 , @xmath30 and @xmath31 included in jm05 @xcite are negligible in the near - threshold and sub - threshold regions . thus these channels are not included to this work . the @xmath32 production amplitude within jm05
can then be written as @xmath33 with @xmath34g_{\delta}\gamma_{\delta,\pi n}\ , \label{eq : pid - t}\end{aligned}\ ] ] where @xmath35 is the propagator of the @xmath26 intermediate state , and the vertex function @xmath36 describes the @xmath37 decay .
explicit expressions for these amplitudes may be found in appen .
contributing mechanisms in equations ( [ eq : full - t],[eq : pid - t])are illustrated in fig .
[ mech_low_06 ] .
the diagram @xmath38 in the second row is the resonant term @xmath39 in eq.([eq : pid - t ] ) .
it is parameterized as a breit - wigner form @xcite and calculated from all well established @xmath10 , @xmath40 states with masses less than 2.0 gev which have hadronic decays to the @xmath11@xmath41 final states .
the non - resonant term @xmath42 is calculated from the well established born terms of @xmath43 @xcite .
their amplitudes are presented in appendix i. the additional contact term @xmath44 was introduced in @xcite to account phenomenologically for the other possible production mechanisms through the @xmath25 intermediate states , as well as for fsi effects .
parametrization of these amplitudes may be found in appendix ii .
the diagrams in the bottom of fig .
[ mech_low_06 ] represents the direct term @xmath45 in eq.([eq : full - t ] ) which was introduced in @xcite to describe the direct @xmath46 mechanisms .
it was parameterized in jm05 @xcite in terms of a contact vertex and a particle - exchange amplitude .
the parameterization of this term in jm05 will be given explicitly in section [ jlabmsugen ] .
reaction in the cm frame of the final hadrons ( first assignment presented in the section [ xsect ] ) .
the top plot shows @xmath18 spherical angles @xmath47 and @xmath48 while the bottom plot shows angle @xmath49[p'\pi^{+}]}$ ] between two planes : one of them ( plane a ) is defined by the 3-momenta of the initial proton and the final @xmath18 , a second ( plane b ) is defined by the 3-momenta of the two others final hadrons @xmath50 and p.,title="fig:",width=302 ] reaction in the cm frame of the final hadrons ( first assignment presented in the section [ xsect ] ) .
the top plot shows @xmath18 spherical angles @xmath47 and @xmath48 while the bottom plot shows angle @xmath49[p'\pi^{+}]}$ ] between two planes : one of them ( plane a ) is defined by the 3-momenta of the initial proton and the final @xmath18 , a second ( plane b ) is defined by the 3-momenta of the two others final hadrons @xmath50 and p.,title="fig:",width=302 ] production mechanisms , seen in the previous clas @xmath0 data @xcite .
the data on @xmath18 angular distributions at w=1.49 gev , @xmath2=0.65 @xmath51 ( left ) and at w=1.49 gev , @xmath2=0.95 @xmath51 ( right ) are shown by points with error bars . the full calculation
are shown by solid lines.the dashed lines correspond to the contributions from jm05 direct 2@xmath41 production @xcite.,width=340 ] in this paper , we have further improved the jm05 model in order to fit the new clas data on @xmath0 electroproduction cross sections .
the clas @xmath0 data allowed us to extend considerably our knowledge of the direct production term @xmath52 of eq.([eq : full - t ] ) in the low w and low q@xmath3 region .
the resulting model , called jm06 , describes successfully all available clas and world @xmath0 electroproduction data at w @xmath1 1.6 gev and @xmath2 from 0.25 to 0.6 gev@xmath3 .
a good description of the data of nine differential cross sections with rather different manifestation of contributing mechanisms in these observables enables us to establish all essential @xmath0 electroproduction mechanisms in this kinematic domain .
in particular , it allows us to isolate the resonant contributions for determining the electrocouplings of the p@xmath53 and d@xmath54 states . within jm06 ,
we find that these two n@xmath55 states are the main contributors to the resonant part of the @xmath0 cross sections at w @xmath1 1.6 gev . in section [ xsect ]
, we define the kinematics of @xmath0 production .
the jm06 model will be specified in section [ jlabmsugen ] .
the fitting procedures are explained in section [ fit ] .
section [ results ] is devoted to discuss the results .
a summary and outlook are given in section [ concl ] .
at a given invariant mass @xmath4 and photon virtuality @xmath56 , the cross section of the @xmath32 reaction can be written as @xmath57 where @xmath58 is virtual photon energy in the lab .
frame , @xmath59 is the virtual photon flux defined by the momenta of incoming and outgoing electrons @xmath60}$ ] and @xmath61 is the phase space volume of the five independent variables in the c.m .
system of the final @xmath62 state .
there are many possible choices @xcite of the five independent variables . defining @xmath63 , @xmath64 , and @xmath65 as invariant mass variables of the three possible two - particle pairs in the @xmath7 system , we adopt the following three assignments : 1 .
@xmath66[p\pi^{-}]}$ ] , where @xmath67 ( @xmath47 , @xmath48 ) are the final @xmath68 spherical angles with respect to the direction of virtual photon , and @xmath69p\pi^{-}]}$ ] is the angle between the plane b defined by the momenta of the final @xmath70 pair and the plane a defined by the initial proton and the final @xmath18 ( see fig . [ kinematic ] ) ; 2 .
@xmath71[\pi^{+}\pi^{-}]}$ ] , where @xmath72 ( @xmath73 , @xmath74 ) are the final proton spherical angles with respect to the direction of virtual photon , and @xmath75[\pi^{+}\pi^{-}]}$ ] is the angle between the plane b defined by the momenta of @xmath76 pair and the plane a defined by the momenta of the initial and final protons ; 3 .
@xmath77[p\pi^{+}]}$ ] , where @xmath78 ( @xmath79 , @xmath80 ) are the final @xmath81 spherical angles with respect to the direction of virtual photon , and @xmath82[p\pi^{+}]}$ ] is the angle between the plane b
defined by the momenta of the final @xmath83 pair and the plane a defined by the initial proton and the final @xmath50 .
the emission angles for the final state particles in the case of first assignment are shown in fig .
[ kinematic ] .
this choice of the kinematical variables is most suitable for describing @xmath0 electroproduction through the @xmath84 intermediate state , which is the dominant contributor of all isobar channels in the kinematical region covered by the considered data . for the others
assignments the emission angles of the final hadrons are similar to the ones given in fig .
[ kinematic ] .
the relations between the momenta of the final state hadrons and the five variables of the first assignment can be found in @xcite .
the limited statistics of the available data does not allow the use of correlated five - fold differential cross sections in the analysis .
instead , we use the differential cross sections depending on only one final kinematic variable , obtained from integrating the 5-fold differential cross sections over four other independent kinematic variables .
these differential cross sections were obtained with reasonable statistical accuracy in @xcite .
they are defined as : .
@xmath85[p'\pi^{+}]}\ ] ] @xmath86[p'\pi^{+}]}\ ] ] @xmath87[p'\pi^{-}]}\ ] ] for invariant mass distributions , and @xmath88[p'\pi^{+}]}\ ] ] @xmath89[p'\pi^+ ] } ) } = \int\frac{d^{5}\sigma}{d^{5}\tau_{2 } } dm_{p\pi^+}dm_{\pi^+\pi^-}d\omega_{\pi^-}\end{aligned}\ ] ] for angular distributions .
other distributions for the angles @xmath90 , @xmath91 , and @xmath92[p'\pi^-]}$ ] , @xmath93[\pi^+\pi^-]}$ ] are similarly defined .
overall nine independent 1-fold differential cross section in each bin of @xmath4 and @xmath2 were included in the analysis .
as a starting point , we described the recent clas data @xcite at 1.3 @xmath1 w @xmath1 1.6 gev and 0.25 @xmath1 q@xmath3 @xmath1 0.60 gev@xmath3 within the jm05 model @xcite , which is briefly outlined in section i. the relations between jm model amplitudes and charged double pion electroproduction cross sections are presented in appendix iv .
here we only specify the input to our calculations and indicate improvements incorporated in jm06 for the analysis presented in this paper .
the nucleon resonances included in the resonant term @xmath94 of eq.([eq : pid - t ] ) are listed in table [ nstlist ] .
the @xmath95 state observed in the analysis of clas @xmath0 electroproduction data @xcite is also included there .
the @xmath16 hadronic decay widths , branching fraction to @xmath25 and @xmath96 final states were taken in part from analysis of hadroproduction data @xcite ( numbers in table [ nstlist ] ) .
hadronic parameters for other states needed in the calculations of resonant amplitude @xmath94 ( symbols `` var '' in table [ nstlist ] ) were taken from analysis @xcite of the clas data @xcite , which covered 1.4 @xmath1 @xmath4 @xmath1 2.0 gev and 0.5 @xmath1 @xmath2 @xmath1 1.5 gev@xmath3 .
the initial values of the electromagnetic form factors , @xmath97 , @xmath98 , @xmath99 were estimated from interpolations @xcite of clas and world data to the considered @xmath56 region covered by the data @xcite and further adjusted in fitting the data . with these specifications , it was found that the fits at low @xmath4 and @xmath56 @xcite are only sensitive to the first two @xmath16 s listed in the table [ nstlist ] .
the contributions from the higher masses @xmath16 , @xmath40 states are not varied in the fits of recent clas data @xcite collected at 1.3 @xmath1 w @xmath1 1.6 gev and 0.25 @xmath1 q@xmath3 @xmath1 0.60 gev@xmath3 . .[nstlist ] list of resonances included and theirs hadronic properties : total decay widths @xmath100 , branching fractions ( bf ) to @xmath101 and @xmath96 final states .
the quoted values are taken from review of particle properties .
the quantities labeled as `` var . ''
correspond to the variable parameters fit to the clas @xmath0 data @xcite . [ cols="^,^,^,^,^ " , ] here we present the initial values of the parameters of @xmath102 and @xmath103 resonances used in computing the various isobar channel contributions .
they were obtained from interpolation of previously available clas / world @xmath10 electrocoupling data @xcite to the kinematical area covered by the recent clas data @xcite .
the starting values of hadronic parameters were taken from previous analysis of the clas @xmath12 electroproduction data within the framework of jm05 model @xcite .
interpolated electrocoupling values were further adjusted to measured differential cross sections . only listed in the tables [ elpar],[hadpar ] resonances have a measurable impact on observables in the clas data .
the contributions from other resonances are well inside the data uncertainties .
listed in the tables below resonance parameters were varied as it was described in the section [ fit ] these parameters should be used in combination with non - resonant jm06 parameters in order to describe superposition of resonant and non - resonant mechanisms in various isobar channels .
they can not be interpreted as resonance parameter values derived from the data fit , since no attempt was made to improve the determined @xmath16 parameters as well as to evaluate the uncertainties of @xmath16 parameters , related to the accuracies of measured cross sections .
non - resonant helicity amplitudes in @xmath25 isobar channels were expanded over the set of partial waves , corresponding to the quantum states with the total angular momenta @xmath104 and it s projections @xmath105 and @xmath58 onto the directions along the initial and final particle momenta in the c.m , frame .
the defined partial waves @xmath106 are related to the full helicity amplitude @xmath107 as : @xmath108 where @xmath110 is the pion polar emission angle in the cm frame .
clas data analysis showed that at @xmath4 @xmath1 1.55 gev the partial wave basis may be restricted to @xmath111=5/2 . contribution from partial waves with @xmath104 @xmath27 5/2 are inside the data uncertainties .
the helicity amplitudes depend on the @xmath4 , @xmath2 variables , describing the initial state and on the @xmath110 , @xmath112 variables of the final state , where @xmath112 stands for running mass of the @xmath26 isobar . after integration ( [ jopa1 ] ) the partial waves @xmath106 depend on @xmath4 , @xmath2 and @xmath112 variables only .
all helicity amplitudes and partial waves , discussed here , contain the factor @xmath113 , which describes their dependence on the pion azimuthal emission angle @xmath114 .
this exponential factor is retained for any reaction dynamic , being a consequence of the rotational invariance of the production amplitudes . instead of helicity representation ,
@xmath115 representation is frequently used for the description of the @xmath25 final state .
the partial waves @xmath116 in helicity representation may be expressed through the partial waves @xmath117 , in @xmath115 representation as : the transition coefficients @xmath119 are given by the products of clebsch - gordan coefficients and the factor accounting for different wave function normalizations in the @xmath115 and helicity representations the clebsch - gordan coefficient @xmath121 describes the spinless @xmath41 and the @xmath26 of 3/3 spin , which are coupled to the total spin @xmath122=3/2 . since the quantization axis is along the @xmath41 momenta , the total spin projection is -@xmath123 .
the clebsch - gordan coefficient @xmath124 corresponds to the orbital angular momentum @xmath125 and the total spin @xmath122=3/2 of the @xmath25 system coupled to the total angular momentum @xmath104 .
the projection of the orbital angular momentum is equal to zero because of choice of the quantization axis mentioned above . | recent clas data on the @xmath0 electroproduction off protons at 1.3@[email protected] gev and 0.25@xmath1@[email protected] gev@xmath3 have been analyzed using a meson - baryon phenomenological model . by fitting nine 1-fold differential cross section data for each @xmath4 and @xmath2 bin ,
the charged double pion electroproduction mechanisms are identified from their manifestations in the observables .
we have extracted the cross sections from amplitudes of each of the considered isobar channels as well as from their coherent sum .
we also obtained non - resonant partial wave amplitudes of all contributing isobar channels which could be useful for advancing a complete coupled - channel analysis of all meson electroproduction data . |
coherent states ( css ) defined as the right eigenstates of the harmonic oscillator annihilation operator , i.e. , @xmath0 play an important role in quantum optics and modern physics @xcite . along the generalization of these states , nonlinear css @xcite or @xmath1-css @xcite have been introduced . according to this formalism @xmath1-deformed annihilation and creation operators , respectively defined as @xmath2 and @xmath3 where @xmath4 and @xmath5 are bosonic annihilation , creation and number operators , respectively .
the intensity - dependent function @xmath6 is responsible for the nonlinearity of the states .
nonlinear css @xmath7 are then defined as the right eigenstates of @xmath1-deformed annihilation operator , @xmath8 the fock space representation of these states is explicitly given by @xmath9!},\qquad\nonumber\\ & \mathbf{\mathcal{n}}(|z|^2)&=\left ( \sum_{n=0}^{\infty } \frac{|z|^{2n}}{n ! \,([f(n)]!)^2}\right)^{-1/2},\qquad
z\in \mathbb{c } \end{aligned}\ ] ] where @xmath10!=f(1)f(2) ... f(n)$ ] and by convention @xmath11!\doteq 1 $ ] .
note that we have confined ourselves to the special case of real valued function @xmath6 .
roknizadeh et al derived a hamiltonian associated to a nonlinear system based on action identity requirement of nonlinear coherent states as @xmath12 @xcite .
so , the eigenvalue equation for any one dimensional physical system with known discrete eigenvalues may be given by @xmath13 , where @xmath14 .
accordingly , one simply has @xmath15 associated to solvable quantum systems . in this way
the nonlinear css may be deduced corresponding to any solvable quantum system as @xcite : @xmath16!}}|n\rangle.\ ] ] but it is remarkable that these states are not temporally stable like the original nonlinear css in ( [ bast nonlin ] ) . moreover , an analytical representation of gazeau - klauder css ( gkcss ) corresponding to any hamiltonian with discrete ( nondegenerate ) eigenvalues is defined as @xcite @xmath17 where @xmath18 is a normalization constant .
these states satisfy the following requirements : ( i ) continuity of label , ( ii ) resolution of the identity , ( iii ) temporal stability and ( iv ) action identity .
the last condition requires @xmath19!$ ] .
so the state in ( [ bast gk z , alpha ] ) can be written as @xmath20!}}|n\rangle.\ ] ] it is also established in @xcite that gkcss are a class of nonlinear css with nonlinearity function @xmath21 , which are temporally stable , i.e. , @xmath22 along this work , gazeau - klauder squeezed states associated to solvable quantum systems have been introduced by one of us @xcite . on the other hand , recently there has been much interest in the superpositions of css @xcite . due to the quantum interference between the coherent components , such superposition states may exhibit various nonclassical properties such as squeezing and sub - poissonian statics .
nonclassical states of light are central to quantum optics .
their importance comes from potential applications on advanced optics , as teleportation @xcite , quantum computation @xcite , quantum communication @xcite , quantum cryptography @xcite , quantum lithography @xcite , etc .
a number of schemes have been proposed for generating such states @xcite . among them , a method has been presented for generating superpositions of css of a cavity field via degenerate raman interaction @xcite . in this paper , we have followed the same approach and suggested a theoretical scheme for generating the gkcss which are temporally stable , via the intensity - dependent degenerate raman interaction .
it is worth to mention that , we have used the intensity - dependent modified effective hamiltonian which describes appropriately the dynamics of the interaction between atom and field .
this is while , to the best of our knowledge , no scheme for generation of gkcss may be found in the earlier literature .
the degenerate raman interaction describes the interaction between a degenerate @xmath23-type three - level atom ( as shown in fig . 1 ) and a single - mode radiation field characterized by bosonic operators @xmath24 .
various properties of this system have been previously discussed @xcite .
the hamiltonian for such a system in the interaction picture is given by : @xmath25 where we have set @xmath26 , @xmath27 and @xmath28 are the two degenerate lower states and @xmath29 is the upper state of the atom , the coupling constant of the transition between @xmath29 and @xmath30 with the cavity field is denoted by @xmath31 , @xmath32 is detuning , where @xmath33 is the cavity field frequency , @xmath34 is the energy of the two lower states and @xmath35 is the energy for the upper level of the atom . on the other side , to describe the degenerate raman interaction , an effective hamiltonian was derived by agarwal in @xcite , when the atomic transition between the upper and lower levels is far from the frequency of the field mode . indeed ,
after adiabatically eliminating the upper level , the author introduced the effective hamiltonian as : @xmath36 where @xmath37 indicates the effective atom - field coupling constant .
in fact , in the effective hamiltonian approach , the two lower levels are coupled with a single - mode field through a virtual upper level . as a result
, the degenerate raman interaction is reduced to an effective two - level system .
also , the effective hamiltonian given in ( [ effective hamiltonian1 ] ) leads to the elimination of time - dependent phase factors in ( [ interaction hamiltonian ] ) .
these factors are related to the ac stark shifts of energy levels @xcite .
the hamiltonian that represents the stark shifts of energy levels is given by : @xmath38 where @xmath39 describes the stark shift of the state @xmath40 .
hence , one may modify the effective hamiltonian given in ( [ effective hamiltonian1 ] ) , so that it includes the ac stark effects represented in ( [ ac stark ] ) . in this way , for the modified effective hamiltonian
, one has then @xcite @xmath41 inserting @xmath42 and @xmath43 respectively from ( [ effective hamiltonian1 ] ) and ( [ ac stark ] ) in ( [ sum hamiltonian ] ) , the modified effective hamiltonian can be written as follows : @xmath44 it is found that , the modified effective hamiltonian @xmath45 is more suitable than the well - known effective hamiltonian @xmath42 in studying the time evolution of the systems .
xu and zhang in @xcite showed that , in the degenerate raman interaction , far off - resonant ( enough large value of @xmath46 ) , the modified effective hamiltonian has been given in ( [ modified effective hamiltonian ] ) , can be applied instead of the hamiltonian in the interaction picture in ( [ interaction hamiltonian ] ) .
in fact , starting from a unique state vector at @xmath47 , both of the hamiltonians lead to the same state vector of atom - field system at any time @xmath48 .
altogether , several other mathematical physics methods were suggested to obtain the modified effective hamiltonian in @xcite .
generalization of the single - mode to two - mode degenerate raman process has been recently done in @xcite . but , to achieve our aim of the paper , we should generalize the single - mode degenerate raman interaction in a rather different way .
indeed , following the path of knight in @xcite and specifically zheng in @xcite , we replace the bosonic operators @xmath24 in ( [ modified effective hamiltonian ] ) with @xmath49 , i.e. , the generalized @xmath1-deformed ladder operator .
this type of generalization may be arisen from the same procedure of degenerate raman interaction , briefly illustrated in the previous section , with the difference that the interaction between the same @xmath23-type atom and a single - mode radiation field is now considered to be intensity - dependent .
henceforth , analogously to ( [ interaction hamiltonian ] ) one may suggest the following form for the hamiltonian of such a system in the interaction picture @xmath50 in the same way , analogously to ( [ modified effective hamiltonian ] ) , for the intensity - dependent modified effective hamiltonian we suggest @xmath51 clearly , inserting @xmath52 in ( [ tamim interaction hamiltonian ] ) and ( [ tamim effective hamiltonian ] ) recovers the usual degenerate raman interaction ( intensity independent coupling ) introduced in ( [ interaction hamiltonian ] ) and ( [ modified effective hamiltonian ] ) , respectively . this type of generalization ( of degenerate raman interaction ) to intensity - dependent interaction can be frequently found in the literature , for instance , whenever one imposes nonlinearity on the standard jcm @xcite .
before we proceed , to establish the above formal approach more explicitly , the following discussion may be offered .
as mentioned , when @xmath46 is enough large , the equivalence of the introduced hamiltonian in ( [ interaction hamiltonian ] ) and ( [ modified effective hamiltonian ] ) are demonstrated in @xcite .
our aim at this stage is to prove that , in the nonlinear degenerate raman interaction , under particular conditions , the nonlinear modified effective hamiltonian in ( [ tamim effective hamiltonian ] ) can be applied appropriately instead of the nonlinear hamiltonian in the interaction picture introduced in ( [ tamim interaction hamiltonian ] ) . to investigate the validity of our proposal ,
we will obtain the time evolution of the state vector of the atom - field system by using ( [ tamim interaction hamiltonian ] ) and ( [ tamim effective hamiltonian ] ) , and continue with showing that for enough large value of @xmath46 , these two state vectors are equal .
for this purpose , we suppose that the atom is initially in a superposition of @xmath28 and @xmath27 : @xmath53 and the field is in the arbitrary state : @xmath54 where @xmath55 determines the initial state of the field .
the state of the atom - field system at @xmath47 is thus given by : @xmath56 the time evolution of the state of the atom - field system , governed by @xmath57 introduced in ( [ tamim interaction hamiltonian ] ) , is the solution of : @xmath58 where @xmath59 , denotes the state of the atom - field system after time @xmath48 .
the following general solution may be considered as : @xmath60 .
\end{aligned}\ ] ] the problem is now determining the time dependent coefficients which may be obtained by lengthy but straightforward calculations with the following final results : @xmath61 @xmath62 @xmath63 where we have set @xmath64 @xmath65 @xmath66 @xmath67 @xmath68 with the following definitions : @xmath69 @xmath70 on the other side , if we use @xmath71 introduced in ( [ tamim effective hamiltonian ] ) for obtaining the time evolution of the state of the atom - field , we have to solve the equation @xmath72 where @xmath73 denotes the state of the atom - field system after time @xmath48 .
the general solution may be given by @xmath74 \end{aligned}\ ] ] with @xmath75 @xmath76 here we have set @xmath77 @xmath78 @xmath79 when @xmath80 namely @xmath46 be enough large and @xmath81 which means that evolving time is not too long , one has : @xmath82 @xmath83 therefore , by using the two above approximations we have : @xmath84 @xmath85 and consequently , @xmath86 as it is observed , we established that in the nonlinear degenerate raman interaction , whenever @xmath46 be enough large , the state vectors given in ( [ halate system t barhamkonesh ] ) and ( [ halate system t moaser ] ) will be equal .
therefore , in the nonlinear degenerate raman interaction , when @xmath46 be enough large , the nonlinear modified effective hamiltonian is equivalent to the nonlinear hamiltonian in the interaction picture .
now , in the remainder of the paper , we want to present our theoretical scheme for generating temporal stable gkcss via the introduced intensity - dependent degenerate raman interaction , while the atom - field interaction is described by the modified effective hamiltonian given in ( [ tamim effective hamiltonian ] ) .
it is worth to notice that , without loss of generality , @xmath87 is assumed in ( [ tamim effective hamiltonian ] ) for continuing the calculational works , thus one has @xmath88 .
so , one arrives at the modified effective hamiltonian as follows : @xmath89 using the expression of @xmath90 , it may be readily seen that the effective hamiltonian in ( [ tamim modified effective hamiltonian ] ) can be explicitly re - written as : @xmath91 inserting @xmath92 in ( [ modified effective hamiltonian ] ) and then comparing the recasted relation with ( [ tamime effective hamiltonian , rewrite ] ) , the above development can be , in a sense , considered simply as generalizing @xmath93 to @xmath94 .
this result may be viewed as intensity - dependent atom - field coupling . now
, suppose that the cavity field is initially prepared in the nonlinear cs introduced in ( [ bast nonlin ] ) . also , the flux of the atoms injected into the cavity is such a low that there exists at most one atom at a time inside the cavity .
if the first atom be initially in a superposition of @xmath27 and @xmath28 , such that @xmath95 the state of the atom - field system at @xmath47 is thus given by : @xmath96 using the time evolution of the state of atom - field described by : @xmath97 with @xmath98 in ( [ tamime effective hamiltonian , rewrite ] ) and the initial state in ( [ first atom+field at t=0 ] ) , the state of the atom - field system after time @xmath48 is given by : @xmath99 |g , n\rangle\nonumber \\ & + & [ ( 1+\varepsilon_{1 } ) e^{2i\lambda n f^2(n)t}+(1-\varepsilon_{1})]|e , n\rangle\ \}.
\end{aligned}\ ] ] the atomic velocities can be controlled such that every atom interacts for a fixed time @xmath100 , with the radiation field inside the cavity .
suppose that the atom which is exiting the cavity is detected in the state @xmath27 , so the cavity field certainly collapses to @xmath101 |n\rangle , \end{aligned}\ ] ] where @xmath102 is an appropriate normalization factor that may be determined .
inserting @xmath103 from ( [ bast nonlin ] ) in ( [ ( first atom)field at tau ] ) one can rewrite the field state as @xmath104!}\;e^{2i\lambda n f^2(n)\tau}|n\rangle\nonumber\\ + & ( 1-\varepsilon_{1 } ) & \mathbf{\mathcal{n}}(|z|^2 ) \sum_{n=0 } ^{\infty } \frac{z^n}{\sqrt{n!}\ ; [ f(n)]!}\;|n\rangle\}. \end{aligned}\ ] ] it is worth mentioning the fact that the normalization coefficient @xmath18 for both nonlinear cs and gkcs is exactly the same .
noticing that @xmath105 , at last the state in ( [ vazeh ] ) can be written as @xmath106 , \end{aligned}\ ] ] where @xmath107 is a gkcs introduced in ( [ final bast gk z , alpha ] ) with @xmath108 .
it is clear from ( [ ( first atom)field at tau , temporall+nonlinear ] ) that we have arrived at a superposition of gkcs and initial nonlinear cs .
now , if one prepares the atom in the initial state ( [ first atom ] ) with @xmath109 , namely the probabilities of the presence of the atom in @xmath27 and @xmath28 are equal , the cavity field will collapse to @xmath110 .
so , we have successfully generated a temporal stable state from initial nonlinear cs @xmath7 that is not temporally stable .
now , assume that the second atom is injected into the cavity , while it is initially in the state described by @xmath111 if the second atom also interacts for a time @xmath100 , with the cavity field which is described by ( [ ( first atom)field at tau ] ) and then , it is detected in the state @xmath27 , the cavity field collapses to @xmath112 where again @xmath113 is a normalization constant . similar to the procedures we followed in ( [ ( first atom)field at tau])-([(first atom)field at tau , temporall+nonlinear ] ) , the state in ( [ ( second atom)field at tau ] ) can be re - written as @xmath114 , \end{aligned}\ ] ] where we have set @xmath115 .
now , once again when the first and second atoms are initially prepared in @xmath116 , i.e. , @xmath117 , the cavity field would be in @xmath118 , i.e. , in the gkcs . by further iteration of the above procedure
it is easy to show that if the @xmath119th atom is initially in the state @xmath120 and we enter @xmath121 atoms into the cavity , one by one , we can generate a superposition of @xmath121 temporally stable gkcss and the initial state of the cavity field @xmath7 that is not temporally stable .
when all of @xmath121 atoms that are exiting one by one from the cavity are detected in the state @xmath27 , the state of the cavity field is given by @xmath122 , \end{aligned}\ ] ] where @xmath123 is an appropriate normalization constant and @xmath124 .
again , if @xmath125 @xmath126 , then @xmath127 and @xmath128 . as a result , the cavity field would be in @xmath129 which is exactly in the gkcss family .
we end this section with noticing that , the initial state of the cavity , i.e. , nonlinear cs may be generated experimentally . as an evidence recall that the generation of photon - added coherent states as a well - known class of nonlinear css reported by zavatta et al in @xcite .
also , a theoretical scheme for generation of any class of nonlinear css in a micromaser using the intensity - dependent jaynes - cummings model was introduced in @xcite .
in summary , using the intensity - dependent degenerate raman interaction procedure has been introduced in the present paper and assuming the cavity field to be initially in the nonlinear cs ( which is not temporally stable ) , the atoms that are in a superposition of the states @xmath28 and @xmath27 are sent inside the cavity , one by one . the atomic velocity can be controlled such that every atom interacts with the cavity field for a specific time interval @xmath100 .
if each of the atoms , which is exiting from the cavity is detected in the state @xmath27 , the cavity field collapses to a superposition of gkcss and initial nonlinear cs . if @xmath121 atoms are sent inside the cavity , superposition of @xmath121 gkcss and the initial state of the cavity will be generated .
preparing the initial state of the atom , such that the probabilities of being every atom in states @xmath28 and @xmath27 are equal , then the generated states reduce to a gkcs .
in fact , we have generated temporally stable gkcs from a nonlinear cs which does not have this property .
in addition to these , we have found a more deep insight to the physical foundation of gkcss , where we observed that according to our proposal , the parameter @xmath130 in the phase factor of gkcss in ( [ final bast gk z , alpha ] ) explicitly depends on @xmath93 ( the effective coupling constant of the atom - field ) and @xmath100 ( the time of the interaction between atom and field ) .
altogether , our proposal may be considered as the first step in producing the gkcss .
we hope that the proposed theoretical scheme will stimulate further better approaches to study the gkcss , experimentally .
v. buzek , phys .
a. * 39 * ( 1989 ) 3196 ; r.a .
zait , phys .
a. * 319 * ( 2003 ) 461 ; s. sivakumar , int . j. theor . phys . * 43 * ( 2005 ) 2405 ; v.i .
koroli , v.v .
zalami , j. phys .
b : at . mol .
. phys . * 42 * ( 2009 ) 035505 ; a - s .
f. obada , h.a .
hessian , a - b .
a. mohamed , j. phys .
* 43 * ( 2010 ) 025305 . | a theoretical scheme is presented for generating gazeau - klauder coherent states ( gkcss ) via the generalization of degenerate raman interaction with coupling constant to intensity - dependent coupling .
firstly , we prove that in the intensity - dependent degenerate raman interaction , under particular conditions , the modified effective hamiltonian can be used instead of hamiltonian in the interaction picture , for describing the atom - field interaction .
we suppose that the cavity field is initially prepared in a nonlinear cs , which is not temporally stable .
as we will observe , after the occurrence of the interaction between atom and field , the generated state involves a superposition of gkcss which are temporally stable and initial nonlinear cs .
under specific conditions which may be prepared , the generated state just includes gkcs .
so , in this way we produced the gkcs , successfully .
* keywords : * generation of gazeau - klauder coherent state , nonlinear coherent states , degenerate raman interaction , modified effective hamiltonian . *
pacs : * 03.65.bz , 03.65.-w , 42.50.-p |
heavy quarks play important roles in probing properties of the quark - gluon plasma ( qgp ) .
a single heavy quark in the qgp interacts with the medium particles and eventually it will be kinetically thermalized as is described by the langevin equation @xcite : @xmath0 if one can observe the degree of thermalization in a given time scale , one can derive the drag force of the heavy quark . when put in the qgp
, a quarkonium , which is a bound state of a heavy quark - antiquark pair , gets unstable at sufficiently high temperature because the plasma constituents screen the color charges and turn the confining potential into the debye screened potential @xcite : @xmath1 the debye screening in the qgp and the related issues are still under active theoretical and conceptual developments @xcite .
if one can observe the melting " of quarkonium , one can infer the temperature provided that there is a reliable theory for the melting .
these heavy quark probes are fundamentally different from each other : the drag force is a classical description for the non - potential irreversible force while the melting is a quantum phenomenon of bound - states in the screened potential .
moreover it is natural to expect that each heavy quark in a quarkonium experiences the irreversible drag process . in a theoretical point of view , it is desirable to have a unified quantum description for both of them , whose classical limit yields the langevin dynamics in the single heavy quark case . in other words , we pursue the unified quantum description for in - medium forces in the qgp . the theoretical framework needed for this purpose
is that of open quantum systems .
the basics of the open quantum systems @xcite can be summarized as follows .
when hilbert space consists of direct product of two hilbert spaces @xmath2 , the hamiltonian for the total system is in general @xmath3 and the density matrix of the total system evolves according to the von - neumann equation : @xmath4.\end{aligned}\ ] ] the dynamics of the open quantum system is described by the master equation for the reduced density matrix : @xmath5.\end{aligned}\ ] ] in the case of the heavy quarks in the qgp , the system degrees of freedom are the heavy quarks and the environmental degrees of freedom are the light quarks and the gluons . in this article , we review the perturbative derivation of the master equation for the heavy quark systems at high temperature @xcite .
the description of heavy quark systems as open quantum systems can be derived on the basis of the closed - time path formalism of non - equilibrium quantum field theory @xcite .
applied to qcd , the partition function is @xmath6 & = & \int \mathcal { d}[\bar \psi , \psi , \bar q , q , a ] \langle \bar\psi,\bar q , a \ ( s=-\infty)|{\bm\rho}_{\rm tot}|\psi , q , a \ ( s=\infty)\rangle\nonumber\\ & & \times\exp \left[i\int_{\mathcal c}ds \int d^3x\left\ { { \mathcal l}_{\rm qcd}(\bar \psi , \psi,\bar q , q , a ) + \bar \eta \psi+\bar \psi \eta \right\}\right],\end{aligned}\ ] ] where @xmath7 and @xmath8 are the heavy and light quark dirac fields , @xmath9 is the gluon field , and gauge fixing is implied in the qcd lagrangian @xmath10 . here
@xmath11 denotes the closed - time path contour which starts from @xmath12 , extends to @xmath13 , and then returns to the original @xmath12 . the initial density matrix ( in the schrdinger picture )
is defined as @xmath14 : at @xmath12 , a system in arbitrary state @xmath15 starts to interact with the environmental degrees of freedom in thermal equilibrium .
states @xmath16 and @xmath17 are the coherent states , satisfying for example @xmath18 .
the dynamics of the open quantum system is described by the influence functional @xcite obtained after integrating out the environmental degrees of freedom .
the influence functional @xmath19 $ ] is expressed as a functional of heavy quark color currents @xmath20 on the forward time axis @xmath21 and @xmath22 on the backward time axis @xmath23 where @xmath24 , and is expanded in terms of the currents with coefficients being the correlation functions of the gluons in thermal equilibrium . in the leading - order perturbation and in the non - relativistic limit , @xmath25 ( or @xmath26 ) consists only of four - fermi terms whose coefficients are the leading - order gluon two - point function @xmath27 .
the long - time dynamics of heavy quarks is summarized in the following effective action @xmath28 : @xmath29,\\ s_{\rm fv}^{\rm lonr}&=&-\frac{1}{2}\int dtd^3xd^3y ( j^{a0}_{1}(t,\vec x),j^{a0}_{2}(t,\vec x ) ) \left ( \begin{array}{cc } v(\vec x-\vec y ) & -id(\vec x-\vec y)\\
-id(\vec x-\vec y ) & -v^*(\vec x-\vec y ) \end{array } \right ) \left ( \begin{array}{c } j^{a0}_{1}(t,\vec y ) \\
j^{a0}_{2}(t,\vec y ) \end{array } \right)\nonumber\\ & & -\int dtd^3xd^3y \frac{\vec\nabla_x d(\vec x-\vec y)}{4 t } \cdot \left ( \vec j^{a}_{1,\rm nr}(t,\vec x)j^{a0}_2(t,\vec y)+j^{a0}_1(t,\vec x)\vec j^{a}_{2,\rm nr}(t,\vec y ) \right),\end{aligned}\ ] ] where @xmath30 , @xmath31 , and @xmath32 denotes the non - relativistic limit of the heavy quark color current .
with this effective action @xmath28 , the partition function is given by @xmath33&=&\int \mathcal { d}[\bar \psi_1 , \psi_1,\bar \psi_2 , \psi_2 ] \langle\bar\psi_1(-\infty)|{\bm\rho}_{\rm s}|\psi_2(-\infty)\rangle\\ & & \times \exp\bigl[is_{1 + 2 } + i\int d^4x\left ( \bar \eta_1 \psi_1+\bar \psi_1 \eta_1-\bar
\eta_2 \psi_2-\bar \psi_2 \eta_2 \right)\bigr].\nonumber\end{aligned}\ ] ] from eq . , the time evolution of @xmath34 is obtained as a functional schrdinger equation " with an effective hamiltonian @xmath35 derived from @xmath28 .
however , the physically relevant matrix element of the density matrix is not @xmath34 but @xmath36\equiv \langle q^*_{1(c)}|{\bm\rho}_{\rm s}(t)|{q}_{2(c)}\rangle$ ] , where @xmath37s are the non - relativistic pauli spinors for the heavy ( anti)quarks .
it needs to take following technical procedures to obtain the functional schrdinger equation for @xmath36 $ ] . in these procedures ,
introduction of conjugate fields @xmath38 on @xmath23 allows a symmetric notation between @xmath39 and @xmath40 . 1 .
the operator ordering in @xmath35 is determined according to the time ordering of the fields in the path integral formula .
the functional schrdinger equation for @xmath41 $ ] is obtained from @xmath35 with substitution @xmath42 3 .
the above substitution gives ultraviolet divergent contributions in the four - fermi terms arising from the coulomb potential at the origin , e.g. @xmath43\left[q^*_{1(c)}(\vec y)t^a\frac{\delta}{\delta q^*_{1(c)}(\vec y)}\right]\nonumber\\ & & = \int d^3x c_{\rm f}v(\vec 0 ) \left[q^*_{1(c)}(\vec
x)\frac{\delta}{\delta q^*_{1(c)}(\vec x)}\right]+\cdots,\end{aligned}\ ] ] which needs to be renormalized .
4 . instead of ( ii ) and ( iii ) , it is equivalent to derive first the renormalized effective hamiltonian @xmath44 , in which the operators are reordered in the normal order and the ultraviolet divergences are renormalized , and then do the substitution ( ii ) to obtain a functional differential operator @xmath45 $ ] .
we adopt ( iv ) and obtain the following renormalized effective hamiltonian @xmath44 : @xmath46\\ % & & + \int d^3x \left [ a^*m\left ( { \tilde { \bm q}}_{2(c)}^{\dagger } { \tilde { \bm q}}_{2(c ) } \right ) + { \tilde { \bm q}}_{2(c)}^{\dagger}\left(-\frac{\nabla^2}{2m}\right){\tilde { \bm q}}_{2(c ) } \right]\nonumber\\ % & & + \frac{1}{2}\int d^3xd^3y \ { \rm n}\left\ { ( \bm j^{a0}_{1}(\vec x),\bm j^{a0}_{2}(\vec x ) ) \left ( \begin{array}{cc } v(\vec x-\vec y ) & -id(\vec x-\vec y)\\
-id(\vec x-\vec y ) & -v^*(\vec x-\vec y ) \end{array } \right ) \left ( \begin{array}{c } \bm j^{a0}_{1}(\vec y ) \\
\bm j^{a0}_{2}(\vec y ) \end{array }
\right)\right\}\nonumber\\ % & & + \int d^3xd^3y \frac{\vec\nabla_x d(\vec x-\vec y)}{4t}\cdot { \rm n}\left\ { \vec { \bm j}^{a}_{1,\rm nr}(\vec x)\bm j^{a0}_{2}(\vec y ) + \bm j^{a0}_{1}(\vec x)\vec { \bm j}^{a}_{2,\rm nr}(\vec y ) \right\ } + \cdots , \nonumber\end{aligned}\ ] ] where @xmath47 , @xmath48 , and @xmath49 denotes taking the normal - ordered product . from @xmath44 ,
the functional schrdinger equation is derived as @xmath50 = h^{\rm ren}_{1 + 2}[q^*_{1(c)},\tilde q^*_{2(c ) } ] \rho_{\rm s}\left[t , q^*_{1(c)},\tilde q^*_{2(c)}\right].\end{aligned}\ ] ] the calculation of the gluon two - point functions @xmath51 at leading order in the hard thermal loop resummed perturbation theory yields
@xmath52 = -\frac{g^2\omega_{\rm d}^2t}{2\pi } \int^{\infty}_{0}dk\frac{\sin kr}{r(k^2+\omega_{\rm d}^2)^2},\\ \label{eq : part - a } a&= & 1+\frac{c_{\rm f}}{2m}\frac{g^2}{4\pi}(-\omega_{\rm d}-it ) , \ \omega^2_{\rm d}\equiv g^2t^2 \left(\frac{n_{\rm c}}{3}+\frac{n_{\rm f}}{6}\right).\end{aligned}\ ] ] at this order of perturbation , the effect of @xmath53-channel scattering between a heavy quark and hard medium particles is taken into account .
this interaction contributes to the diffusion process of the heavy quarks as well as the debye screening of the color charges .
the former is given by the imaginary part of @xmath54 and @xmath55 while the latter is by the real part of @xmath54 , and thus these in - medium forces are both contained in the renormalized effective hamiltonian @xmath44 .
the master equation for the heavy quark systems is obtained from the functional schrdinger equation eq . .
the matrix element of @xmath56 between the heavy quark coherent states @xmath57 and @xmath58 is given as a functional of the heavy quark fields by @xmath59 & = & \langle q^*_{1(c)}|\bm\rho_{\rm s}(t)| \tilde q_{2(c)}^*\rangle,\\ \langle q_{1(c)}^*|&\equiv & \langle \omega |\exp\left[-\int d^3x \left\ { \bm q_{(c)}(\vec x)q^*_{1(c)}(\vec x ) \right\}\right],\\ \exp\left[-\int d^3x \left\ { \tilde q_{2(c)}^*(\vec x)\bm q_{(c)}^{\dagger}(\vec x ) \right\}\right]|\omega\rangle.\end{aligned}\ ] ] here @xmath60 is the heavy quark vacuum state . by functionally differentiating @xmath61 $ ] , we obtain , for example , the reduced density matrix for a single heavy quark system : @xmath62\big |_{q_{1,2(c)}^*=0},\end{aligned}\ ] ] and that for a heavy quark - antiquark system : @xmath63\big |_{q_{1(c)}^*=\tilde q_{2(c)}^*=0}.\nonumber\end{aligned}\ ] ] the master equation for each of these reduced density matrices is therefore obtained from the functional schrdinger equation for @xmath61 $ ] in eq . .
for example , let us examine the real - time dynamics of a single heavy quark system at high temperature .
the master equation for a single heavy quark system reads @xmath64 and that for the color - traced reduced density matrix @xmath65 is given by @xmath66 interestingly , taking short distance limit @xmath67 in the master equation eq .
reduces it to the well - known master equation of the caldeira - leggett model for the quantum brownian motion @xcite . using the master equation eq .
, the following ehrenfest relations for averaged expectation values @xmath68 are derived : @xmath69 to be explicit , @xmath70 which is consistent with the leading - order perturbative calculation of the drag force @xcite . in this way
, we can derive the consequences of the classical langevin dynamics through the quantum ehrenfest relations .
in this article , we reviewed the derivation of the master equation for the reduced density matrix of heavy quark systems at high temperature .
the derivation is based on the influence functional formalism applied to the heavy quark systems as the open quantum systems and is explicitly shown in the leading - order perturbation . the essential information is contained in the renormalized effective hamiltonian @xmath44 , from which the master equation is derived . note that this method is applicable to systems with any finite number of heavy quarks and is not limited to the leading - order perturbation ;
once the renormalized effective hamiltonian is derived , it is straight forward to obtain the master equation . as explained in ref .
@xcite , the forward propagator of heavy quarks can also be derived from the renormalized effective hamiltonian .
the complex potential between heavy quark - antiquark pair is defined in terms of the forward propagator and thus can also be derived from our approach to yield consistent result with refs .
@xcite in the leading - order perturbation .
+ here we list several next steps of this approach : * derivation of the renormalized effective hamiltonian in the next - to - leading order of perturbation .
we expect the next - to - leading order contribution will introduce new processes , namely the so - called gluo - dissociation processes @xcite .
* consideration of how to extend this formalism to the general non - perturbative situation .
one of the challenges is to understand how to describe heavy - light meson bound states solely in terms of heavy quark fields , or to examine what is the correct choice of the system variables as the open quantum system in such cases . *
phenomenological application to the dimuon spectra from bottomonia decays at the lhc @xcite .
if one has a reliable theory of melting , one can study the physics of deconfinement by comparing with the experimental data .
this work was supported by the sasakawa scientific research grant from the japan science society . 9 b. svetitsky , phys .
d * 37 * , 2484 ( 1988 ) .
t. matsui and h. satz , phys . lett .
b * 178 * , 416 ( 1986 ) .
m. laine , o. philipsen , p. romatschke and m. tassler , jhep * 0703 * , 054 ( 2007 ) [ hep - ph/0611300 ] ; a. beraudo , j. -p .
blaizot and c. ratti , nucl .
a * 806 * , 312 ( 2008 ) [ arxiv:0712.4394 [ nucl - th ] ] . n. brambilla , j. ghiglieri , a. vairo and p. petreczky , phys .
d * 78 * , 014017 ( 2008 ) ; n. brambilla , m. a. escobedo , j. ghiglieri , j. soto and a. vairo , jhep * 1009 * , 038 ( 2010 ) [ arxiv:1007.4156 [ hep - ph ] ] .
a. rothkopf , t. hatsuda and s. sasaki , phys .
lett . * 108 * , 162001 ( 2012 ) [ arxiv:1108.1579 [ hep - lat ] ] .
j. l. albacete , y. v. kovchegov and a. taliotis , phys .
d * 78 * , 115007 ( 2008 ) [ arxiv:0807.4747 [ hep - th ] ] ; j. noronha and a. dumitru , phys . rev .
lett . * 103 * , 152304 ( 2009 ) [ arxiv:0907.3062 [ hep - ph ] ] ; t. hayata , k. nawa and t. hatsuda , arxiv:1211.4942 [ hep - ph ] .
y. akamatsu and a. rothkopf , phys .
d * 85 * , 105011 ( 2012 ) [ arxiv:1110.1203 [ hep - ph ] ] .
breuer and f. petruccione , _ the theory of open quantum systems _
( oxford university , new york , 2002 ) .
y. akamatsu , phys .
d * 87 * , 045016 ( 2013 ) [ arxiv:1209.5068 [ hep - ph ] ] .
j. s. schwinger , j. math .
* 2 * , 407 ( 1961 ) ; l. v. keldysh , zh . eksp .
teor . fiz . * 47 * , 1515 ( 1964 ) [ sov .
jetp * 20 * , 1018 ( 1965 ) ] ; p. m. bakshi and k. t. mahanthappa , j. math .
phys . * 4 * , 1 ( 1963 ) ; j. math .
phys . * 4 * , 12 ( 1963 ) .
r. p. feynman and f. l. vernon , jr .
, annals phys .
* 24 * , 118 ( 1963 ) .
a. o. caldeira and a. j. leggett , physica * 121a * , 587 ( 1983 ) . g. d. moore and d. teaney , phys .
c * 71 * , 064904 ( 2005 ) [ hep - ph/0412346 ] ; s. caron - huot and g. d. moore , phys .
lett . * 100 * , 052301 ( 2008 ) [ arxiv:0708.4232 [ hep - ph ] ] ; jhep * 0802 * , 081 ( 2008 ) [ arxiv:0801.2173 [ hep - ph ] ] .
d. kharzeev and h. satz , phys .
b * 334 * , 155 ( 1994 ) [ hep - ph/9405414 ] ; x. -m .
xu , d. kharzeev , h. satz and x. -n .
wang , phys .
c * 53 * , 3051 ( 1996 ) [ hep - ph/9511331 ] .
n. borghini and c. gombeaud , arxiv:1103.2945 [ hep - ph ] ; eur .
j. c * 72 * , 2000 ( 2012 ) [ arxiv:1109.4271 [ nucl - th ] ]
. s. chatrchyan _ et al . _
[ cms collaboration ] , phys .
lett . *
107 * , 052302 ( 2011 ) [ arxiv:1105.4894 [ nucl - ex ] ] ; jhep * 1205 * , 063 ( 2012 ) [ arxiv:1201.5069 [ nucl - ex ] ] ; arxiv:1208.2826 [ nucl - ex ] . | we derive the quantum dynamics of heavy quark systems in quark - gluon plasma as open quantum systems .
the scatterings of a heavy quark with the hard medium particles give rise to debye screened qcd force and drag force and its fluctuation .
we present a unified quantum description of these in - medium qcd forces at high temperature in the leading - order perturbation on the basis of the influence functional formalism . |
many of the drift sections in spear 3 will contain wigglers or undulators . in the present machine ,
there are 7 insertion devices ( i d ) . in this paper
, we review the magnetic parameters for each of these i d s , estimate linear and higher order wiggler effects , evaluate radiation effects , and discuss correction of wiggler focusing using quadrupole trim windings .
tracking results show that dynamic aperture reduction due to wiggler effects is modest .
all optics calculations were done with the mad code @xcite , and the tracking simulations with the lego code @xcite .
in spear 3 , the horizontally deflecting i d s reside in dispersion free straights where the unperturbed @xmath0 functions are @xmath1/@xmath210.1/4.8 m. the magnetic field parameters for each i d are given in table 1 , where @xmath3 is the number of periods , @xmath4 the period length , @xmath5 the peak field.=10 , @xmath4=0.18 , ( 2 ) @xmath3=15 , @xmath4=0.12 , ( 3 ) @xmath3=24 , @xmath4=0.07 , ( 4 ) @xmath3=30 , @xmath4=0.06 , and ( 5 ) elliptically polarized undulator ( epu ) . here , we chose the @xmath3=10 , @xmath4=0.18 wiggler . ]
future devices are likely to be @xmath610-period wigglers similar to beamlines 9 and 11
. these devices will replace the wigglers on beamlines 4 and 7 or illuminate new beam lines .
the main optical effects of the wigglers on the spear 3 lattice are also summarized in table 1 .
the parameter definitions and relevant formulas are listed below : ttttttttttttttttt = ttttttttttttt = bend radius @xmath7 + wiggler parameter @xmath8 + integrated focusing @xcite @xmath9 + linear tune shift @xmath10 + maximum @xmath0-beat @xmath11 + amplitude dependent @xmath12 + tune shift @xcite @xmath13 $ ] where @xmath14=5.23 is the vertical betatron tune , @xmath15 the vertical emittance , and @xmath16=3 gev the beam energy .
.wiggler parameters and main optics effects .
[ cols="<,^,^,^,^,^,^,^",options="header " , ] [ tab : mult2 ] without correction of the wiggler focusing , the vertical dynamic aperture reduces by about 20% with all wigglers active . with corrected wiggler focusing
, linear wiggler effects alone do not reduce the aperture .
the study of wiggler multipole errors was based on measured systematic field errors in the wiggler on beamline 11 .
a numerical multipole fit to the measured data was performed and the results for the integral of normal and skew systematic multipole field are shown in table 3 , where ` @xmath17 ' is the multipole order starting with dipole , @xmath18 , and @xmath19 , @xmath20 are the skew and normal field coefficients defined as @xmath21 the effect of 7 wigglers on dynamic aperture is shown in fig .
[ fig : aper ] , with the aperture calculated at the i d locations ( @xmath1/@xmath210.1/4.8 m ) for machine simulations with all errors in the ring magnets . without multipole errors in the wigglers but with corrected wiggler focusing , the aperture is not affected . with the systematic multipole errors included in the 7 wigglers ,
the dynamic aperture reduction is about 10% ( 2 mm ) .
the wiggler - intrinsic high order terms ( eqn . [ eq : bx ] ) were not included in the previous simulation , but can further reduce the dynamic aperture . this field does not have the usual form of ( @xmath22-@xmath23 ) multipole expansion normally used in lattice codes . to accurately study the effect of the octupole - like and higher order terms in eqn .
[ eq : bx ] , one needs a symplectic model of this field .
we used an approximation to evaluate the main effect of these terms . since the non - linear @xmath24 field generates kicks only in @xmath23-plane , the effect on dynamic aperture
is expected to be mostly in the vertical plane , as it was observed in simulations for other machines @xcite .
one can notice that the @xmath24 field can be reproduced at @xmath22=0 by the normal field of eqn .
[ eq : bxby ] ( @xmath25=0 ) . using this approach , we used normal systematic multipoles in the form of @xmath26 to produce the non - linear terms of the wiggler field @xmath24 on the vertical axis ( eqn .
[ eq : bx ] ) , and performed tracking to evaluate vertical aperture near @xmath22=0 .
simulation of the octupole - like and dodecapole - like ( @xmath27 ) fields with 7 wigglers showed only modest reduction of vertical dynamic aperture , from 11 mm to 8 - 9 mm .
this aperture still remains well outside the 6 mm wiggler physical aperture .
tracking simulations with 7 wigglers in spear 3 predict small impact on dynamic aperture when linear optics effects are locally compensated with individually controlled quadrupole magnets .
the 18 mm horizontal dynamic aperture is sufficient for 10 mm injection oscillations and long touschek lifetime , and the vertical aperture remains well outside the 6 mm wiggler physical aperture . with all wigglers active , the increase in energy spread is negligible ( @xmath282% ) , and the beam horizontal emittance decreases to about 15.3 nm@xmath29rad at 3 gev .
we would like to thank m. cornacchia , h. wiedemann and b. hettel for useful discussions , suggestions and guidance .
99 h. grote , f. c. iselin , cern / sl/90 - 13 ( ap ) rev.4 ( 1994 ) .
y. cai _ et al .
_ , slac - pub-7642 ( 1997 )
. h. wiedemann , `` particle accelerator physics ii , '' springer - verlag , berlin , heidelberg ( 1995 ) .
l. smith , lbl - esg tech .
note-24 ( 1986 ) .
k. halbach , _ nim _ , a*187 * , p. 109
a. ropert , cern 90 - 03 , p. 142 | the spear 3 upgrade lattice will provide much reduced beam emittance to increase the brightness of synchrotron radiation beams from wigglers and undulators .
seven existing insertion devices will be used in the lattice . in this paper
we review the wiggler parameters , outline the wiggler compensation scheme , and evaluate wiggler effect on the optics and dynamic aperture .
slac
pub8233 + august 1999 + * effect of insertion devices in spear 3[multiblock footnote omitted ] * * abstract * |
the observation of absorbing and emitting spectral lines is the most important method of investigation in astrophysics .
for this reason the radiative transfer problems in spectral lines are considered in huge numbers of papers . at first only the intensity of spectral line was considered . the angular distribution of scattered radiation was assumed isotropic .
we remind only some of references in the time order ( busbrige 1953 ; thomas 1957 ; jefferies & white 1958 ; ivanov 1962a , b ; avrett 1965 ; ivanov & nagirner 1966 ; nagirner & ivanov 1966 ; abramov , dykhne & napartovich 1967 ; hummer 1968 ; rybicki & hummer 1991 ) . in particular , the problem of scalar spectral line transfer was presented in book by ivanov ( 1973 ) in full .
the detailed discussion of scalar spectral line transfer in stellar atmospheres is given in the book by mihalas ( 1982 ) .
the allowance of non - isotropic scattering of polarized spectral line radiation is more difficult problem .
the number of papers devoted to this problem is less than ones devoted to problems without polarization ( see , for example , rees & saliba 1982 ; faurobert - sholl & frisch 1989 ; faurobert - sholl et al .
1997 ; hovenir et al . 2004 ; milic & faurobert 2012 ; faurobert et al .
it should be also noted the series of papers by ivanov 1995 , 1996 ; ivanov et al .
1997a , b , and the papers by dementyev 2008 , 2009 , 2010 .
the short review of the problem with the allowance of the radiation polarization was presented by nagirner ( 1985 ) .
the polarization of spectral line depends on the influence of magnetic field ( see faurobert 1987 ; faurobert - sholl 1991 , 1993 ; stenflo 1994 ; trujillo bueno & landi deglinnocenti 1996 ; nagendra et al .
2002 ; the books by fluri 2003 , and by landi deglinnocenti and landolfi 2004 ) .
we present here the transfer equation for polarized spectral line radiation in media considering three types particles - atoms with resonant level , atoms without resonant level and dust grains .
these media exist near the centre of seyfert galaxies in the forms of gas - dusty torus and the equatorial flow ( see , for example , the reviews by marin 2014 , 2016 ) .
usually one considers the vectorial radiative transfer equation for the intensity @xmath9 and the stokes parameter @xmath10 which describe axially symmetric problems ( parameter @xmath11 .
this equation ( see eq.(1 ) in the next section ) contains two parameters : @xmath6 and @xmath12 describing the absorption in continuum nearby the wavelength of spectral line and the absorption @xmath13 of the spectral line radiation due to collisional destruction of spectral line per scatterings of atoms with one another .
one assumes that parameters @xmath6 and @xmath13 are independent with one another and are independent of frequency @xmath7 .
first we introduce the ( column ) vector with the components ( @xmath16 ) , where @xmath17 and @xmath18 [ erg/@xmath19 hz s sr ] are the intensity and the stokes parameter , respectively .
we consider the axially symmetric problem , when the stokes parameter @xmath20 . in problems depending on azimuthal angle @xmath21
our transfer equation describes the azimuthally independent part of radiation , which usually is the most large part of the total radiation ( see chandrasekhar 1960 ) . for investigation of multiple scattering of resonant
radiation one uses the vectorial transfer equation of the general form ( see , for example , ivanov et al .
1997a , 1997b ; dementyev 2008 ) : @xmath22 @xmath23 @xmath24 where @xmath7 is the frequency of light , @xmath25 .
the angle @xmath26 is the angle between the line of sight @xmath27 and the normal to the plane - parallel atmosphere @xmath28 .
the dimensionless extinction factor of a spectral line is @xmath29 .
the function @xmath1 describes the shape of spectral line emitted from excited atom .
the factor @xmath30 , where the terms @xmath31 and @xmath32 are the cross - sections of scattering by resonant atom and the absorption , respectively .
the cross - section of absorption @xmath32 describes collisional destruction of excited atoms .
this cross - section is assumed to be independent of frequency @xmath7 .
this assumption is true if the energy of atom s collisions is much higher than the energy corresponding to the width of resonant line .
parameter @xmath13 is the degree of absorption .
usually one assumes that @xmath13 is independent of frequency .
the optical depth @xmath33 takes into account the mean scattering factor in a line .
parameter @xmath34 is the number density of resonant atoms .
the matrix @xmath35 is the probability that radiation with frequency @xmath36 and direction @xmath37 after scattering acquires the frequency @xmath7 and direction @xmath27 .
the source term @xmath38 describes the non - polarized isotropic radiation due to the plank thermal radiation .
frequently one takes @xmath39 ( see , for example , ivanov 1973 ) . the transfer equation ( 1 ) ,
discussed above , describes the radiative transfer in medium consisting of resonant atoms .
the allowance of non - resonant atoms gives rise to additional term in parameter @xmath6 independent of frequency .
usually one assumes that @xmath6 and @xmath13 are independent with one another , and are independent of frequency .
of course , this assumption has an approximate character . below we consider more successively three - component medium consisting of atoms ( or molecules ) with resonant level , atoms without the resonant level and dust grains .
the number densities of these components are @xmath34 , @xmath40 and @xmath41 , respectively .
the corresponding cross - sections are @xmath42 , @xmath43 , @xmath44 and @xmath45 .
the cross - section of extinction by resonant atoms is @xmath46 .
we know that the direct propagation of radiation is described by the relation : @xmath47 , where @xmath48 is the free path of radiation .
the value of @xmath48 is related with particular free paths by the known formula : @xmath49 @xmath50 to derive the scattering term in transfer equation for resonant radiation , we are take into account that the incident radiation can collide with various particles according to corresponding probabilities @xmath51 and @xmath52 . to describe the light scattering on resonant atoms
we are take into account only the @xmath53 - probability . clearly , the sum of all probabilities are to be equal to unity .
it is also clear from physical point of view , that the probabilities are proportional to the values @xmath54 and @xmath55 with the proportionality factor @xmath56 . the normalization condition - the sum of all probabilities is equal to unity , gives @xmath57 .
thus , the probability @xmath53 acquires the form : @xmath58 @xmath59 @xmath60 where the parameters @xmath3 and @xmath4 are : @xmath61 @xmath62 we consider the symmetric form of spectral line , i.e. @xmath63 .
this gives @xmath64 .
if the resonant photons collide with resonant atom , only part of them - @xmath65 , undergoes the resonant scattering .
thus , the total probability to undergo the resonant scattering is equal to production : @xmath66 @xmath67 as a result , we obtain the following radiative transfer equation for spectral line : @xmath22 @xmath68 @xmath69 here we take @xmath70 ; @xmath1 describes the shape of the line ( see below ) , @xmath0 is dimensionless extinction factor .
the source term describes the scattered isotropic plank s radiation @xmath71 .
the factor @xmath6 in our case is described by the expression : @xmath72 @xmath73 parameter @xmath2 is independent of frequency , whereas @xmath74 depends on @xmath7 .
the numerical calculations , presented in ch .
6 , demonstrate that the dependence of @xmath75 on frequency is very important ( for example , the maximum difference between line shapes is @xmath76 for linearly increasing sources , and is @xmath77 for homogeneous sources at @xmath78 ) .
note , that for the polarization the difference is higher . thus , we can not assume that the @xmath79 - factor is independent of frequency .
real parameters independent of frequency are @xmath3 and @xmath4 .
it is important that the parameters @xmath3 and @xmath4 appear in radiative transfer equation ( 5 ) as a sum @xmath80 .
this is quite naturally because both parameters have the same physical meaning - they describe absorption of the line intensity by various mechanisms .
note , that frequently one consideres the case of the absence of absorption ( @xmath81 ) . in this case
the equation ( 5 ) coincides with equation ( 1 ) , where @xmath82 and @xmath81 are taken .
the cross - section of resonant scattering in optical range is of the order @xmath83 @xmath84 ( see , for example , gasiorovich 1996 ) .
the cross - section of scattering of non - resonant radiation is much lesser @xmath85 @xmath84 .
remind , that the thomson cross - section of scattering on free electrons is @xmath86 @xmath84 , larger than atom s cross - section . the cross - section @xmath87 , where @xmath88 is the radius of a grain .
this means that the factor @xmath89 is very small compared with the term @xmath90 , if the number density of non - resonant atoms is not much higher than the number density of grains .
usually such inequality does not hold and the term @xmath89 can not be taken into account .
the normalized function @xmath1 describes the shape of line .
frequently one uses the limit forms - gaussian or doppler @xmath91},\ ] ] @xmath92 and the lorentz @xmath93 here @xmath94 is central frequency of a resonant line , @xmath95 is doppler s width of a line : @xmath96 where the thermal velocity is determined by the temperature @xmath97 , and the turbulent velocity is determined as a mean value of chaotic macroscopic motions @xmath98 .
the value @xmath99 depends on widths of the atomic energy levels .
the gaussian form of a line arises as a result of doppler s frequency shifts due to thermal and turbulent motions of atoms and molecules .
this form usually corresponds to the centre of a line .
the lorentz form characterize the wings of a line which are frequently blanketed by line environment .
introducing dimensionless frequencies @xmath100 or @xmath101 , we derive from eqs.(7 ) and ( 8) the usually used functions : @xmath102 @xmath103 the expressions ( 10 ) and ( 11 ) present the limit cases of function @xmath104
. there exists the voigt formula which presents expressions ( 10 ) and ( 11 ) as a limit cases : @xmath105 here parameter @xmath106 is the ratio of line width due to emission and atomic impacts to the doppler width .
if @xmath107 , then @xmath108^{-1 } \to\delta(x - y)$ ] , and expression ( 12 ) transforms to formula ( 10 ) ( see ivanov 1973 ) .
the matrix @xmath35 in general has very complex form ( see mckenna 1985 ; landi deglinnocenti & landolfi 2004 ) .
this is the reason why one uses model of full redistributed frequencies : @xmath109 apparently , that formula ( 13 ) corresponds to the case when during the lifetime of atomic level many impacts occur .
the matrix @xmath110 has the form : @xmath111 and @xmath112 the matrix @xmath113 is ( see ivanov et al .
1996 ) : @xmath114 here parameter @xmath115 depends on quantum numbers of transition atomic levels . for simplest case of dipole transition @xmath116 .
our calculations in ch 6 correspond to this case .
note that the superscript t will be used for matrix transpose . using the dimensionless frequencies @xmath117 and identity ( 15 ) , the transfer equation ( 5 ) can be written in the form : @xmath118 @xmath119 @xmath120 this equation can be presented as : @xmath121 @xmath122 . \label{18}\ ] ] the vector @xmath123 is determined by the expression : @xmath124
radiative transfer equation ( 18 ) can be written in the form : @xmath126 @xmath127 where the vector @xmath128 is introduced by the expression : @xmath129 the vector @xmath130 , according to eq.(20 ) , can be written as : @xmath131 @xmath132 @xmath133 @xmath134 substituting these expressions into eq.(19 ) , and taking into account that matrix @xmath135 is even function of @xmath136 , we derive the expression : @xmath137 @xmath138 @xmath139 using the formulas ( 21 ) and ( 23 ) , we obtain the integral equation for @xmath140 : @xmath141 the free term @xmath142 has the form : @xmath143 the matrix kernel of integral equation ( 24 ) is the following : @xmath144 @xmath145 where @xmath146 note that the matrix @xmath147 is symmetric : @xmath148 .
this property gives rise to symmetry of kernel @xmath149 .
the explicit form of matrix @xmath150 is the following : @xmath151
according to the standard theory of integral equations ( see , for example , smirnov 1964 , tricomi 1957 ) , the solution of eq . ( 24 ) can be presented in the form : @xmath152 where the resolvent matrix @xmath153 obeys the integral equation : @xmath154 @xmath155 @xmath156 this equation can be obtained from eq.(24 ) by simple process of successive iterations .
taking the matrix transpose in this series of iterations and using the property @xmath157 , we obtain the relation : @xmath158 it follows from eq.(30 ) that @xmath159 this equation coincides formally with eq.(24 ) , and , consequently , @xmath160 can be written in the form : @xmath161 @xmath162 @xmath163\,\hat{\psi}(\mu^2 ) , \label{33}\ ] ] where @xmath164 is diagonal unit matrix and @xmath165 denotes the laplace transform : @xmath166 according to relation ( 31 ) we have @xmath167 . transposing eq.(33 ) , we obtain the following equation for @xmath168 : @xmath169 @xmath170 @xmath171 . \label{35}\ ] ] differentiation of first eq.(30 ) over @xmath172 and the second - over @xmath173 gives rise to the equation : @xmath174 @xmath175 the expression in parentheses is proportional to matrix @xmath176 .
this matrix can be separated out expression ( 36 ) .
the remainder part obeys the equation ( 32 ) , i.e. is equal to @xmath160 . as a result ,
one can obtain the following relation ( see , sobolev 1969 ) : @xmath177 this relation means that the resolvent matrix @xmath153 can be obtained , if we know two auxiliary matrices , depending on one variable - @xmath178 or @xmath179 .
the laplace transform over @xmath178 and over @xmath179 of the matrix @xmath180 we denote as : @xmath181 taking this double laplace transform in eq.(37 ) , we obtain the relation : @xmath182 .
\label{39}\ ] ] taking the laplace transform over variable @xmath173 with the parameter @xmath183 in eq.(33 ) and using the relation ( 39 ) , we obtain the matrix equation : @xmath184 @xmath185 where we introduce new matrices : @xmath186 @xmath187 it is seen from eq.(40 ) , that to know @xmath188 we are to know @xmath189 .
taking @xmath190 in eq.(40 ) , we derive the separate equation for @xmath191 : @xmath192 @xmath193 thus , the system ( 42 ) plays the main role in resolvent matrix approach . taking the transpose of eq.(42 )
, we derive the equation for matrix @xmath189 : @xmath194 @xmath195 the effective methods of numerical calculation of @xmath196 -matrix are presented in papers krease & siewert ( 1971 ) , rooij et al.(1989 ) and dementyev ( 2009 ) . using the algebraic technique presented in silantev et al.(2015 ) , we derive the following linear matrix equation for @xmath197 : @xmath198 @xmath199 here z is arbitrary number .
taking @xmath200 one may obtain : @xmath201 @xmath202 it is known that in the milne problem the equation for characteristic number @xmath203 arises .
according to this equation the determinant of right part in eq.(45 ) is to be equal to zero .
the right part of eq.(45 ) depends on the function @xmath197 .
the linear equation ( 45 ) transforms the characteristic equation to the expression : @xmath204 this equation does not depend on complex function @xmath197 .
according to eq . ( 22 ) , the vector @xmath205 , describing the outgoing radiation , has the form : @xmath206 @xmath207 i.e. this expression is proportional to the laplace transform of @xmath140 over variable @xmath208 .
the vector @xmath128 is presented in eq.(29 ) .
the laplace transform of this vector can be written as : @xmath209 @xmath210{\bf g}(\tau ) .
\label{48}\ ] ] thus , expression ( 47 ) acquires the form : @xmath211 @xmath212 .
\label{49}\ ] ] eq.(49 ) is general expression for outgoing radiation , where the term @xmath71 characterizes the distribution of sources of non - polarized radiation in an atmosphere .
explicit form of @xmath142 can be obtained from the model of an atmosphere ( see eq.(25 ) ) .
it appears , the distribution @xmath71 ( i.e. the model of atmosphere ) can be approximated in the form : @xmath213 for source function @xmath214 the expression for @xmath205 acquires comparatively simple form . in this case
the expression for @xmath205 depends on @xmath215 and @xmath216 , does not on total matrix @xmath217 .
the sources of types @xmath218 are related with exponential source by the simple formula : @xmath219 it means that the sources of type ( 50 ) can be considered on the base of the exponential source .
let us consider this case in detail .
taking @xmath214 in eq.(25 ) , we obtain @xmath220 @xmath221 \left ( \begin { array}{c } 1 \\ 0 \end{array } \right ) .
\label{52}\ ] ] taking into account formula ( 39 ) for double laplace transform of matrix @xmath217 and eqs.(42 ) and ( 43 ) , we derive from eq.(52 ) the following expression : @xmath222 @xmath223 according to relation ( 51 ) the emerging radiation @xmath224 for sources @xmath225 can be derived from the expression : @xmath226 @xmath227_{h=0}\left ( \begin { array}{c } 1 \\ 0 \end{array } \right ) \equiv\frac{\varphi(x)p_{res}(x)}{\alpha(x)}\times\ ] ] @xmath228 @xmath229 \right ) _ { h=0}\left ( \begin { array}{c } 1 \\ 0 \end{array } \right ) .
\label{54}\ ] ] to calculate @xmath230 , @xmath231 , @xmath232 etc .
one can use transposed eq.(40 ) with @xmath233 in the form : @xmath234 @xmath235\hat{h}^t\left ( \frac{1}{h}\right ) .
\label{55}\ ] ] homogeneous source @xmath236 corresponds to @xmath237 . in this case
we obtain from eq .
( 53 ) : @xmath238 @xmath239 the emerging radiation for the source @xmath240 can be derived according to eq.(54 ) : @xmath241 @xmath242 \left ( \begin { array}{c } 1 \\ 0 \end{array } \right ) .
\label{57}\ ] ] here @xmath243 is the matrix : @xmath244 analogously we can derive formulas for outgoing radiation for sources of type @xmath245 etc .
thus , we see that the emerging radiation for sources of type ( 50 ) can be calculated , if we calculated the matrix @xmath246 and the moments of of this matrix of type ( 58 ) .
it is easy to check ( see , for example , ivanov 1973 ) that eq.(42 ) for matrix @xmath246 can be presented as the equation for variable @xmath247 .
the emerging radiation ( see eqs . ( 53 ) , ( 54 ) , ( 56 ) and ( 57 ) ) depends on the product @xmath248 .
we prefer to derive the equation for @xmath249 . in this case
we are not related with the complex problem of approximation of @xmath250 from the value @xmath251 . from eqs.(27 ) and ( 42 ) we derive the following equation for @xmath252 : @xmath253 @xmath254 recall , that this equation depends on one parameter @xmath2 : ( @xmath255 and @xmath256 ) .
the definitions of these parameters are presented in eq.(4 ) .
the solution of eq.(59 ) can be carry out by iteration method , analogous to known chandrasekhar s ( 1960 ) method ( see also dementyev ( 2009 ) ) . for values of parameter @xmath6
we chose @xmath257 @xmath258 .
the shapes of resonance line @xmath259 are presented in fig.1 ( for homogeneous @xmath236 and linearly increasing @xmath240 sources ) and in fig.2 ( for exponentially decreasing sources @xmath260 and @xmath261 . in calculations
we assume that all coefficients @xmath262 and @xmath263 are equal to unity . here
the function @xmath264 is the flux of intensity @xmath265 : @xmath266 the value @xmath264 presents the intensity flux as the function of the dimensionless frequency @xmath117 .
recall , that the flux @xmath264 one observes from the distant spherical stars ( really one observes the value @xmath267 , where @xmath268 is the distance to the star and @xmath269 is the radius of a star ) . the linear polarization of radiation from spherical star is equal to zero due to axial symmetry of the problem .
the figures 1 and 2 - demonstrate the dependence of spectral line shape on the parameter @xmath6 . in these figures we present the ratio @xmath259 .
the homogeneous sources with small parameters @xmath270 and @xmath271 give rise to the minimum of outgoing radiation at @xmath272 . the resonance line near @xmath272 looks like an absorption line .
for @xmath270 the maximum value of @xmath273 is @xmath274 and corresponds to @xmath275 . for @xmath276
the maximum of the line is considerably less @xmath277 and corresponds to @xmath278 . for @xmath279 and @xmath280 the shape of the line looks like usual line profile @xmath281 with the maximum at @xmath272 . for @xmath282 and @xmath280
the value @xmath273 is slightly less than @xmath283 .
the linearly growing sources are result in qualitatively the same behavior , but the maxima of the lines are more profound ( @xmath284 at @xmath285 for @xmath270 and @xmath286 at @xmath287 for @xmath276 ) .
small maxima @xmath288 at @xmath289 and @xmath290 at @xmath291 arise for @xmath292 and @xmath293 , respectively . for @xmath294 and @xmath280
the lines look like @xmath281 . with the increasing of parameter @xmath6
the shapes of outgoing lines became narrower than @xmath283 .
the existence @xmath104 in eq.(47 ) means that the intensity of resonance line tends to zero for @xmath295 .
why the flux of the resonance line has the maximum for small values of absorption parameter @xmath6 ?
let us bear in mind that the radiation density @xmath296 due to diffusion of radiation increases with the growing @xmath173 .
the emerging radiation with the @xmath117- frequency mainly corresponds to the optical depth @xmath297 . for the centre of a line @xmath298 . in this region
the density of radiation is less than that for @xmath299 : @xmath300 . the radiation with @xmath301 escapes proportionally to small value @xmath302 whereas the radiation in promediate value , say @xmath303 , is proportional to @xmath304 . as a result ,
the peak of radiation arises in promediate @xmath117-value at @xmath305 . for larger values of @xmath117 the radiation density @xmath296
is not such large to compensate the absorption factor @xmath306 and the intensity of outgoing radiation tends to zero . the value of the peak is higher for linearly increasing source @xmath307
however , in this case the peak is more acute than that in the case of homogeneous source .
the exponential sources @xmath308 are close to the boundary of a medium and the diffusion of radiation does not give rise to large radiation density far from the boundary . as a result
, the peaks do not arise .
the fig.2 shows this .
the absolute values of intensity flux can be obtained from the function @xmath273 if we know the flux @xmath309 and the constants @xmath310 : @xmath311 the upper indices denote the type of problem : @xmath312 denotes the problem with homogeneous sources of radiation , @xmath313 denotes the case of linearly distributed sources , @xmath314 and @xmath315 denote exponential sources @xmath316 and @xmath317 , respectively .
the values @xmath318 we present in tables 1 and 2 for a number of parameters @xmath6 .
the value @xmath318 presents the flux in the centre of a spectral line . in these tables
we also present the integrals : @xmath319 it is seen , that the values @xmath309 and @xmath320 decrease with the increasing of parameter @xmath6 . .
the values of @xmath321 , @xmath322 , and @xmath323 , @xmath324 for homogeneous sources and sources @xmath325 . [ cols="<,<,<,<,<",options="header " , ]
in this paper we derived the radiative transfer equation for polarized resonance line in three component medium , consisting of resonant atoms , non - resonant atoms and grain particles .
this equation generalize usual transfer equation .
most important difference is that the absorption factor in scattering term of our equation depends on frequency .
the important feature of our equation is that it depends on one parameter @xmath2 .
the parameter @xmath3 describes the absorption line s energy due to destruction of excited level by collisions of atoms .
the parameter @xmath4 describes the extinction of resonant radiation due to scattering on non - resonant atoms and due to absorption by dust grains .
we derived the solution of transfer equation using the matrix resolvent technique .
this method allows us to obtain the intensity and polarization of radiation outgoing from semi - infinite atmosphere for sources of non - polarized radiation of homogeneous , linearly increasing and exponentially decreasing types . to obtain these values
one must to solve the system of non - linear equations like those for chandrasekhar h - functions . as a result , we present the intensity and polarization degrees for models of gas - dusty atmospheres , mentioned above . our theory and numerous figures and graphics can be applicated to seyfert galaxies , where gas - dusty media exist , and for estimations of resonance line intensity and polarization emerging from accretion discs .
this research was supported by the basic research program p-7 of prezidium of russian academy of sciences , the program of the department of physical sciences of russian academy of sciences no 2 .
60 abramov yu.yu . ,
dykhne , a.m. , napartovich , a.p . , ap * 3 * , 215 ( 1967 ) avrett , e. in proc .
second harvard - smithsonian conference on stellar atmospheres , smithsonian astrophys .
. special report * 174 * ( 1965 ) busbrige , i.w .
, mnras * 113 * , 52 ( 1953 ) chandrasekhar , s. : radiative transfer .
dover , new york ( 1960 ) dementyev , a. v. astron .
l. * 34 * , 574 ( 2008 ) dementyev , a. v. ap * 52 * , 545 ( 2009 ) dementyev , a. v. ap * 53 * , 419 ( 2010 ) dolginov , a.,z . , gnedin , yu . n. , silantev , n. a. : propagation and polarization of radiation in cosmic media , gordon & breach , new york , ( 1995 ) faurobert , m. a & a * 178 * , 269 ( 1987 ) faurobert - sholl , m. a & a * 246 * , 469 ( 1991 ) faurobert - sholl , m. a & a * 268 * , 765 ( 1993 ) faurobert - sholl , m. , frisch h. a & a * 219 * , 338 ( 1989 ) faurobert - sholl , m. , frisch , h. , nagendra , k.n .
a & a * 322 * 896 ( 1997 ) faurobert , m. , milic i. , atanockovic. o. a & a * 559 * 68 ( 2013 ) fluri , d.m . : radiative transfer with polarized scattering in the magnetized solar atmosphere . cuvillier verlag , gottingen ( 2003 ) frisch , h. in : solar polarization ( eds .
stenflo j.o .
, nagendra k. n. ) , kluwer acad .
, dordrecht , p.135 ( 1996 ) gasiorovich , s. : quantum physics .
john wiley & sons , ny ( 1996 ) hovenier , j.w . , van der mee , m. , domke , h. : transfer of polarized light in planetary atmospheres .
basic concepts and practical methods .
dordrecht : kluwer acad .
( 2004 ) hummer , d.g .
mnras * 138*,73 ( 1968 ) ivanov , v.v .
uchenie zapiski leningrad university * 307 * , 52 ( 1962a)(in russian ) ivanov , v.v . astron.zh.*39 * , 1020 ( 1962b ) ivanov , v.v .
, nagirner d.i .
ap * 2 * , 75 ( 1966 ) ivanov , v.v . :
radiative transfer in spectral lines
. national bureau of standards , washington , ( 1973)(translation from russian edition of 1969 ) ivanov , v.v .
, a & a * 303 * , 609 ( 1995 ) ivanov , v.v .
, a & a * 307 * , 319 ( 1996 ) ivanov , v. v. , kasaurov , a. m. , loskutov , v. m. a & a * 307 * , 332 ( 1996 ) ivanov , v.v .
, grachev s.i . ,
loskutov , v. m. a & a * 318 * , 609 ( 1997a ) ivanov , v.v .
, grachev s.i . ,
loskutov , v. m. a & a * 321 * , 968 ( 1997b ) jefferies , j.t . ,
white , o.r .
j. * 132 * , 767 ( 1958 ) kriese , j. t. , siewert , c. e. , astrophys j. * 164 * , 389 ( 1971 ) landi deglinnocenti , e. , landolfi , m. : polarization in spectral lines .
dordrecht : kluwer acad . publ . ( 2004 ) marin , f. mnras * 441 * , 551 ( 2014 ) marin , f. arxiv : 1605.02904v ( 2016 ) , ( sent to mnras ) mckenna , s.j .
apss , * 108 * , 31 ( 1985 ) mihalas , d. : stellar atmospheres .
w.h.freeman and company , san francisco ( 1982 ) milic , i. , faurobert , m. a & a * 539 * , 10 ( 2012 ) nagendra , k.n .
, frisch , h. , faurobert , m. a & a * 395 * 305 ( 2002 ) nagirner , d.i . ,
ivanov , v.v .
ap * 2 * , 1 ( 1966 ) nagirner , d.i .
ppic . conf .
* 118 * ( 1985 ) rees , d. e. , saliba , g. , a&a , * 115 * , 1 , ( 1982 ) rooij , de w. a. , bosma , p. b. , hooff , van , j.p.c . ,
a & a * 226 * , 347 ( 1989 ) rybicki , g.b . , hammer , d.g .
a & a , * 245 * 171 ( 1991 ) silantev , n.a .
astron . rep . *
57 * , 587 ( 1980 ) silantev , n.a . , alekseeva g.a . ,
novikov , v.v . ap&ss
* 357 * , 53 ( 2015 ) smirnov , v.i .
: the course of higher mathematics , vol .
4 . integral equations and partial differential equations , pergamon press , new york ( 1964 ) sobolev , v.v .
: course in theoretical astrophysics .
nasa technical translation f-531 , washington ( 1969 ) stenflo , j. o. : solar magnetic fields .
polarized radiation diagnostics .
dordrecht , kluwer acad .
( 1994 ) thomas , r.n .
astrophys . j.*125 * , 260 ( 1957 ) tricomi , f. g. : integral equations , intersci .
publishers , inc . , new york ( 1957 ) trujillo bueno , j. , landi deglinnocenti , e .. in : solar polarization ( eds .
stenflo j.o .
, nagendra k. n. ) , kluwer acad .
publ.,dordrecht , p.49 ( 1996 ) | it is assumed that in seyfert galaxies the gas - dusty medium exits near the centre in the form of a molecular and dusty torus and equatorial flow .
these objects have spectral lines emission of hydrogen , helium and other atoms .
we derived the spectral line radiative transfer equation for such media .
this equation has dimensionless extinction factor of the form : @xmath0 , where @xmath1 describes the shape of spectral line emerging from excited atom and @xmath2 .
small factor @xmath3 is proportional to absorption cross - section , describing the collisional destruction of the resonance photons .
the factor @xmath4 describes the extinction of the line radiation due to scattering on non - resonant atoms and due to absorption in dust grains .
the term of transfer equation describing the scattering on resonant atoms is proportional to @xmath5 .
thus , the radiative transfer equation depends on only one parameter @xmath6 , independent of frequency @xmath7 . using the known method of resolvent matrices
, we obtain the exact solution of vectorial radiative transfer equation for various sources of non - polarized radiation in semi - infinite atmosphere .
homogeneous , linear increasing and exponentially decreasing sources are considered .
we present the intensity and linear polarization values for radiation emerging from semi - infinite medium for different values of parameter @xmath6 .
we consider the doppler shape of a spectral line which is a good approximation for central part of the spectral line .
the @xmath6 - dependence of shape of the line , angular distribution of radiation and the polarization degree is presented .
* keywords * : radiative transfer , spectral lines , scattering , gas - dusty atmosphere @xmath8 e - mail : [email protected] |
a well known difficulty of the bethe - salpeter equation @xcite is the existence ( for some interactions at least ) of abnormal solutions , corresponding to excitations in the relative time - energy degree of freedom @xcite .
wick suggested that these solutions could be an artefact due to the ladder approximation , and would disappear if the higher - order irreducible graphs could be included @xcite .
mugibayashi showed however that the abnormal solutions remain present in the exact static model @xcite , which is simple enough to be completely solved .
recently , the conclusions of mugibayashi were questioned by jallouli and sazdjian @xcite , so that the road to wick s interpretation is open again .
this possible conspiration of the higher - order contributions for suppressing the abnormal solutions reminds us a similar cancellation of the higher - order terms at the one - body ( or one large mass ) limit of some 3d reductions of the bethe - salpeter equation , leaving only a simple dirac or klein - gordon equation @xcite .
a way of relating these two facts can be found in our previous study of a zero - d reduction of the 1d bethe - salpeter equation for the static model at the ladder approximation @xcite .
we showed that the reduction of the bethe - salpeter equation does not in general kill the possible abnormal solutions . despite the elimination of the relative time ,
they are recovered via the multiplicity of solutions due to the complicated dependence of the final potential on the total energy .
this was the starting point of the present work .
it can be easily verified that this dependence of the reduced equation potential on the total energy is a necessary condition to the presence of abnormal solutions of any reduction of a bethe - salpeter equation . as this dependence disappears from some
well chosen 3d reductions of the exact bethe - salpeter equation at the one - body limit ( and also from the zero - d reduction of the exact static model obtained by letting the mass of the lighter particle go also to infinity ) , the 3d equations have no abnormal solution at this limit .
conversely , these abnormal solutions are present in the reduction of the static model bethe - salpeter equation at the ladder approximation , and we do not see why they would not be a fortiori present in the case of finite masses . does the absence of abnormal solutions for a 3d reduction imply their absence for the original bethe - salpeter equation itself ?
we shall show that the existence of a solution of the bethe - salpeter equation for a given total energy implies in general the existence of a solution of the reduced equation for this same total energy ( there could be convergence problems , in principle manageable by analytic extensions ) .
however , the reduction being based on a constraint fixing the value of the relative energy , a solution of the bethe - salpeter equation will not lead to a corresponding solution of the reduced equation if it happens to vanish for this value .
this accident does clearly not happen with the normal solutions of the exact bethe - salpeter equation , and , from the study of the static model , we can deduce that it does not happen with at least some of the abnormal solutions of the ladder approximated bethe - salpeter equation .
for the possible abnormal solutions of the exact bethe - salpeter equation , we have of course no example to examine ( as we are just trying to demonstrate that such solutions do not exist ! ) .
we can only conclude that these solutions , if they exist and survive at the one - body limit , are not transmitted to the 3d reductions we use . in section 2 , we present our 3d reduction of the exact or ladder approximated bethe - salpeter equation for two fermions , the existence condition for this reduction and how to invert it . in section 3
, we compute the one - body limit of this reduction in the qed case , the one - body limit of the bethe - salpeter equation itself and verify that the reduction of the limit is the limit of the reduction . in section 4 , we get the static model by letting the mass of the former light fermion become also infinite , and compare again the zero - d reduction of the zero - body limit with the limit of the reduction .
we also verify that the existence condition for the reduction is satisfied at least by the solutions of the ladder approximated bethe - salpeter equation for the static model . in section 5
, we consider the bethe - salpeter equation and its 3d reduction for equal masses , and get the static model directly by letting this common mass go to infinity .
this leads to the same bethe - salpeter equation as in the preceding section , but to another zero - d reduction , based on a different choice of the fixed relative energy .
section 6 is devoted to conclusions .
we shall write the bethe - salpeter equation for the bound states of two fermions as : @xmath0 where @xmath1 is the bethe - salpeter amplitude , function of the positions @xmath2 or of the momenta @xmath3 of the fermions , according to the representation chosen .
the operator @xmath4 is the bethe - salpeter kernel , given in a non - local momentum representation by the sum of the irreducible two - fermion feynman graphs .
the operator @xmath5 is the free propagator , given by the product @xmath6 of the two individual fermion propagators : @xmath7 where the @xmath8 are the dirac free hamiltonians @xmath9 the bethe - salpeter kernel @xmath10 should contain charge renormalization and vacuum polarization graphs , while the propagators @xmath11 should contain self - energy terms ( which can be transferred to @xmath12 @xcite ) . in this work ,
we consider only the free fermion propagators in @xmath13 and the `` skeleton '' graphs in @xmath14 , and we hope that the inclusion of the various corrections would not change our conclusions
. we shall define the total ( or external , cm , global ) and relative ( or internal ) variables : @xmath15 @xmath16 and give a name to the corresponding combinations of the free hamiltonians : @xmath17 in order to keep our formula s compact , we do not write the arguments of the functions and operators . when needed , we work in the momentum representation ( most often , we indicate only the relative energy explicitly ) . the free propagator @xmath5 will be approached by a carefully chosen expression @xmath18 , combining a constraint fixing the relative energy , and a global 3d propagator .
the argument of the constraint and the inverse of the propagator should be combinations of the operators used in the free equations ( at last approximately and for the positive - energy solutions ) .
we shall choose breit s propagator @xcite @xmath19 with @xmath20 @xmath21 we shall write the free propagator as the 3d approached propagator , plus a rest : @xmath22 the bethe - salpeter equation becomes then the inhomogeneous equation @xmath23 with @xmath24 solving ( formally ) the inhomogeneous equation ( [ 2.26 ] ) and putting the result into ( [ 2.27 ] ) , we get @xmath25 where @xmath26 obeys @xmath27 the reduction series ( [ 2.29 ] ) re - introduces in fact the reducible graphs into the bethe - salpeter kernel , but with @xmath5 replaced by @xmath28 . equation ( [ 2.28 ] ) is a 3d equivalent of the bethe - salpeter equation .
it depends on the choice of @xmath18 and is not manifestly covariant ( of course , we can always write it in the center of mass frame ( assuming @xmath29 ) and then replace everything by its covariant equivalent )
. we can write it as a pair of coupled equations by multiplying it by @xmath30 and @xmath31 : @xmath32 @xmath33 the relative time can be completely eliminated by making @xmath34 which gives @xmath35 @xmath36 the @xmath37 spectrum of ( [ 2.34 ] ) should be that of the original bethe - salpeter equation ( [ 2.1 ] ) , without the possible states for which the expression used in ( [ 2.27 ] ) to define the wave function @xmath38 vanishes . the normal bound state spectrum , connected to the nonrelativistic spectrum , will be obtained at first order by keeping only the first term @xmath39 ( or an approximation of it ) in @xmath40 the remaining of @xmath41 being treated as a perturbation .
if the bethe - salpeter equation has also an abnormal spectrum , it should , at least implicitly , remain present also in the 3d reduction ( [ 2.34 ] ) , via the @xmath42dependence of @xmath43 writing @xmath44\psi'\label{2.36}\ ] ] and diagonalizing the hamiltonian @xmath45\psi'_i(p_0)\,=\,e_i(p_0)\psi'_i(p_0)\label{2.37}\ ] ] we get the energy spectrum as the set of the roots of the equations @xmath46 for each value of @xmath47 we should find the energy of the normal solution , plus possibly a set of abnormal energies .
this scheme is in general too complicated to be used in actual numerical calculations .
we know three exceptions : when we want only to compute the normal solutions by perturbations .
in the static model for two fermions in qed at the ladder approximation ( one photon exchange ) @xcite . in this case
, @xmath48 is not an operator anymore , it is simply a function of the energy and the coupling constant .
the reduced equation ( a simple numerical equation ) gives the only normal solution and the even abnormal solutions ( for the odd abnormal solutions , @xmath49 vanishes , as the relative energy is fixed to zero in the reduction performed in @xcite )
. the unique function @xmath50 is divided into sectors separated by poles and containing each one solution .
the reduction series converges only in the first sector containing the normal solution , but can be analytically extended by using the pad approximants method .
at the one - body or zero - body limits of the exact equation , since the final 3d potential reduces then to a simple compact term , as we shall see in section 3 .
the 3d reduction performed above becomes impossible when the 3d wave function @xmath51 defined by ( [ 2.27 ] ) , is the null function .
this gives the condition @xmath52 or , using the bethe - salpeter equation itself : @xmath53 the expression @xmath54 when projected on the subspaces @xmath55 becomes a product of factors which can not vanish in the usual range of the bound states @xmath56 the reduction condition becomes thus simply @xmath57 the solutions @xmath58 of the bethe - salpeter equation for which @xmath59 will thus not lead to solutions of the 3d equation .
this would be the case for the solutions antisymmetrical in @xmath60 which may appear when the masses are equal .
we have of course the possibility of performing a reduction based on a different constraint , like putting a fermion on the mass shell , as gross @xcite .
the bethe - salpeter amplitude is given in terms of the wave function by @xmath61 at @xmath62 we can factorize an operator @xmath63 out of @xmath64 the 3d potential @xmath65 out of @xmath66 and use the 3d equation ( [ 2.34 ] ) to recover the relation ( [ 2.38 ] ) . to be definite
, we shall work in the qed framework , but our results will be more or less directly adaptable to other interactions , and/or to the two - boson and boson - fermion systems @xcite .
the modified kernel @xmath41 is then , in feynman s gauge : @xmath67 @xmath68_b\,\beta_1\beta_2\bigg)(\gamma_1\ ! \cdot\ ! \gamma_2)+\gamma_1^\mu g_{01}(q_1)(\gamma_1\ ! \cdot\ ! \gamma_2)g_{02}(q'_{1c})\gamma_{2\mu}\bigg]+ .... \label{2.41}\ ] ] the various contributions are symbolized by the graphs of figure 1 . in the reducible graphs , we use @xmath69 instead of the @xmath70 we would use in the calculation of the scattering amplitude , so that the graph b , for example , symbolizes the sum of the second and third terms of ( [ 2.41 ] ) .
the external lines are there only to indicate the initial momenta @xmath71 and the final momenta @xmath72 of the fermions 1 and 2 .
the internal momenta of fermion 1 are @xmath73 and the momenta of the photons leaving the fermion 1 line are @xmath74 all momenta will be written in terms of the fixed external momenta and of @xmath75 which we choose as integration variables . with these conventions , the different graphs of a given order differ only by the insertion points of the photon lines on the fermion 2 line .
the internal momenta of fermion 2 will vary accordingly , and we shall denote them by @xmath76 with a supplementary index identifying the corresponding graph .
the momenta of the intermediate fermion lines are thus : @xmath77 @xmath78 @xmath79
we shall now take the one - body limit @xmath80 of the 3d equation in qed .
we shall first define e as a finite part of @xmath81 : @xmath82 for @xmath83 , we have , in any part of ( [ 2.41 ] ) : @xmath84 this is also the asymptotic form to give to @xmath85 and @xmath86 in ( [ 2.41 ] ) , because of the constraint . combining this result with ( [ 3.1 ] ) , we get @xmath87 for the intermediate energies , we shall define more practical variables . considering , to be definite , the @xmath88 diagrams , we have @xmath89 @xmath90 @xmath91 @xmath92 in the ladder diagram ( figure 1 ) : @xmath93 @xmath94 for the 4!-1 other diagrams of order @xmath88 , we have simply to permute @xmath95 in ( [ 3.10 ] ) .
this leads to an equation with only the @xmath96 components , where @xmath97 can be replaced by @xmath98 @xcite : @xmath99 @xmath100\,g_{01}(q_1)\beta_1\right.\ ] ] @xmath101\bigg[{1\over \omega_1+i\epsilon}+2i\pi\delta(\omega_1 ) \bigg]\right.\ ] ] @xmath102 with @xmath103 @xmath104 since @xmath105 we see that the second - order contributions cancel mutually .
it can be shown that similar cancellations occur at each order @xcite , so that we remain with the first term of @xmath65 ( a coulomb potential ) only .
this result can also be obtained in other gauges .
equation ( [ 2.37a ] ) has then only one root , corresponding to the normal solution . at the ladder aproximation ,
however , we have only the terms in @xmath106 with nothing to cancel them . equation ( [ 2.37a ] ) can then have several roots .
we have actually computed them at the zero - body limit ( @xmath107 static model ) @xcite , and we do not see how they could not be present for a finite @xmath108 it must be noted that , if we make the instantaneous approximation without making the ladder approximation , the cancellation does not occur either : we remain with a dependence of the potential on the total energy , and possibly again with several roots of ( [ 2.37a ] ) . in order to get the @xmath109 limit of the bethe - salpeter equation ( [ 2.1 ] ) , we shall write @xmath110 writing @xmath111 we get , at the @xmath109 limit : @xmath112 the 3d reduction of ( [ 3.16 ] ) is obtained by writing @xmath113 @xmath114 here , @xmath115 is obtained from @xmath116 by a simple change of the sign of one @xmath117 ( as it is also the case in gross reduction @xcite ) .
the consequences of this change are not infinitesimal , as the integral of @xmath118 with respect to @xmath119 is zero for the positive values of @xmath120 putting the iterations of ( [ 3.17 ] ) in ( [ 3.18 ] ) gives @xmath121 @xmath122 @xmath123\psi'\label{3.19}\ ] ] with the exact kernel , the higher - order terms of @xmath124 ( crossed diagrams ) cancel the higher - order terms of the expansion ( [ 3.19 ] ) ( we get ( [ 3.11b ] ) with the first term of ( [ 3.11c ] ) ) . at the ladder approximation ,
the kernel is given by @xmath125 equations ( [ 3.19]-[3.20 ] ) give the same results as equations ( [ 3.11b]-[3.11c ] ) , both for the exact model and for the ladder approximation . in order to perform this 3d reduction , we must have @xmath126 ( using ( [ 3.16 ] ) and ( [ 3.18 ] ) ) .
this implies that @xmath127 must have a pole at origin ( accordingly , ( [ 2.38 ] ) shows that @xmath128 must be infinite at the @xmath129 limit if we want to get a nonvanishing @xmath130 ) .
this is certainly true for the normal solutions , for which we can in principle compute @xmath131 by perturbations around a nonrelativistic approximation
. we shall verify below that it is also true for the abnormal solutions of the ladder approximated bethe - salpeter equation at the zero - body @xmath132 limit .
the @xmath133 limit of the reduction ( [ 3.11b]-[3.11c ] ) will be performed after writing @xmath134 . at the ladder approximation , we get @xmath135@xmath136 @xmath137 @xmath138\,\psi'(\vec p_1).\label{4.2}\ ] ] with the exact kernel , we must keep only the first term . since we expect to get the static model , we shall search for solutions in @xmath139\label{4.3}\ ] ] where @xmath140 is a set of three continuus quantum numbers which can be interpretated as the conserved relative position of two infinitely heavy fermions . as the only remaining 3-momenta are @xmath141 etc
.... in the propagators of the photons we shall perform the integrations with respect to the @xmath142 . using the relations
@xmath143 @xmath144\over\omega^2-\vec k^2+i\epsilon}\,=\ , { -2\pi^2\over\vert\vec x\vert}\,\exp[\,i\vert\omega \vec x\vert ] , \label{4.4}\ ] ] removing a common factor @xmath145 \,$ ] and choosing @xmath146 which means taking @xmath147 as the energy unit and rendering @xmath148 and @xmath149 dimensionless ( the energy will thus be @xmath150 from now on ) , we get @xmath151\ , { 1\over(-\omega_1+w+i\epsilon)(\omega_1-i\epsilon ) } \,\exp[i\vert \omega_1\!-\!0\vert]\]]@xmath152\ , { 1\over(-\omega_2+w+i\epsilon)(\omega_2-i\epsilon ) } \,\exp[i\vert \omega_2\!-\!\omega_1\vert]\,\ ] ] @xmath153+ ... \bigg ] \label{4.5}\ ] ] with the exact kernel , we have simply @xmath154 the @xmath133 limit of the bethe - salpeter equation ( [ 3.16 ] ) will be , after writing @xmath134 : @xmath155 we shall search for solutions in @xmath156.\label{4.7}\ ] ] the only 3-momenta in @xmath157 being again the @xmath158 of the photon propagators ( in both the exact and the ladder approximated models ) , we shall again perform the integrations with respect to the @xmath159 defining @xmath160 and choosing @xmath146 we get @xmath161 at the ladder approximation , we have @xmath162 the zero - d reduction is obtained , as in section 3.2 , by changing the sign of the second @xmath163 of ( [ 4.8 ] ) , leaving a term in @xmath164 as the initial term of an iteration .
this gives ( [ 4.5 ] ) at the ladder approximation , and @xmath165 for the exact model .
the condition for this reduction is @xmath166 which implies , via ( [ 4.8 ] ) , that @xmath167 must have a pole at origin .
we think that the absence of abnormal solutions of the 3d reduction of the exact bethe - salpeter equation at the one - body limit is probably the consequence of their absence for the original finite - masses bethe - salpeter equation itself .
there remains however the possibility that such solutions exist and disappear as a consequence of the one - body limit , of the 3d reduction , or of the combination of both .
it is of course impossible to study directly the behaviour of these solutions , the very existence of which is questioned .
we know anyway that the abnormal solutions of the wick - cutkosky model survive the one - body limit @xcite .
we know also that the abnormal solutions of the ladder - approximated static model survive the zero - d reduction we performed in @xcite .
this last zero - d reduction is however based on the choice @xmath168 ( just between the two poles at @xmath169 and @xmath170 and treats the two fermions in a symmetrical way , while the present zero - d reduction is based on the choice @xmath169 , in order to get the very compact result @xmath171 in the exact model .
we must thus check if the condition for this new zero - d reduction is also satisfied by the abnormal solutions of the ladder approximated static model .
if we write @xmath172 we get a more symmetrical bethe - salpeter equation : @xmath173\,u(\omega ) .
\label{4.11}\ ] ] where the sign before the @xmath163 comes from the fact that we are interested in negative values of @xmath148 ( we had @xmath174 equations ( [ 4.8 ] ) or ( [ 4.11 ] ) are the bethe - salpeter equation for the static model , at the ladder approximation , in the relative energy representation . in a previous work @xcite , we studied equation ( [ 4.11 ] ) in the relative time representation , performing a wick rotation and a zero - d reduction
. we could meet our previous results by performing a fourier transform of ( [ 4.11 ] ) , but it turns out to be also possible without leaving the relative energy representation . writing @xmath175 @xmath176 we get @xmath177\over t^2 - 1+i\epsilon}\ , \,\,v(t ) .
\label{4.14}\ ] ] the condition for the zero - d reduction ( [ 4.5 ] ) is that @xmath178 must not vanish at @xmath179 if we derive twice equation ( [ 4.14 ] ) , we get the differential equation @xmath180v(t)\,=\,0 . \label{4.15}\ ] ] although we are still in the relative energy representation , equation ( [ 4.15 ] ) and the complex conjugated equation ( [ 4.14 ] ) ( written thus for @xmath181 ) are exactly the equations we studied previoulsly in the relative time representation @xcite ( the fourier transform of the propagator takes the form of the potential , and vice - versa ) .
we found that we had to solve the differential equation ( [ 4.15 ] ) , and , using the wick rotation in ( [ 4.14 ] ) , that the solutions had to be square integrable along the imaginary axis .
equation ( [ 4.15 ] ) being regular and symmetric around the origin , the solutions are regular at origin and can be made even or odd .
in fact , for each eigenvalue @xmath182 we found either an even or an odd solution , but not both simultaneously . a zero - d reduction of ( [ 4.11 ] ) , with @xmath183 demands @xmath184 so that the odd solutions of the bethe - salpeter equation were not solutions of the reduced equation ( we did not miss them , as they do not satisfy the normalisation condition deduced , like the homogeneous bethe - salpeter equation for the amplitudes , from the inhomogeneous bethe - salpeter equation for the propagator ) . to see if a solution of ( [ 4.14 ] ) leads to a solution of the reduced equation ( [ 4.5 ] ) , we must verify that it does not vanish at @xmath185 the solutions @xmath186 of ( [ 4.15 ] ) , a fuchs equation around this point , can be written as linear combinations of two solutions of the form @xmath187 in the @xmath188 domain of the @xmath189 complex plane . at @xmath190 @xmath191 is zero , and @xmath192 is @xmath193 by definition a nonzero arbitrary constant .
as written above , we know also that we must build even and odd solutions around @xmath194 in order to build these solutions , we must in general use a linear combination of @xmath191 and @xmath195 which will differ from zero at @xmath190 unless built with @xmath196 alone , which should then be even or odd when re - expanded into a series of powers of @xmath197 although this would imply an infinity of cancellations between the coefficients @xmath198 we did not exclude this possibility a priori .
we checked numerically that it does not happen for the 20 first eigenvalues computed in ref .
we can take directly the @xmath199 limit of the 3d reduction .
we have and keep @xmath200 we shall write @xmath201 @xmath202 @xmath203 @xmath204 the fermion propagators , in the second - order terms of the exact model , will give a contribution proportional to @xmath205\,{1\over { 1\over2}w-\omega_1+i\epsilon } \label{5.5}\ ] ] with @xmath206 ( at the ladder approximation , we would not get the crossed term ) .
in contrast with what happens with the asymmetrical zero - body limit , these terms do not combine to zero . in coulomb
s gauge , however , the only relative energy dependence lies in the fermion propagators , and the integral of ( [ 5.5 ] ) with respect to @xmath207 is zero .
the integral of the crossed term is indeed zero , as the poles lie both on the same ( upper ) half part of the @xmath208-complex plane .
the integrals of the two other contributions combine to zero ( they correspond to the modified ladder term , @xmath209 which was designed for that ) . at higher orders , we find the products of the lower fermion propagators @xmath210 where @xmath211 corresponds to the ith lower fermion propagator of the graph @xmath212 it can be shown that the @xmath211 are combinations of the @xmath213 with coefficients @xmath214 for the uncorrected ladder graphs , @xmath215 for the crossed graphs , at least one @xmath216 appears only with minus signs in the lower fermion propagators , so that the integral with respect to this @xmath216 will be zero .
we remain then with the modified ladder graphs ( built with the ladder term plus the @xmath217 ) which give also zero by integration .
we conclude that , in coulomb s gauge , both the exact kernel and the ladder approximation give the normal solution only .
this confirms our result above ( and that of @xcite ) about the absence of abnormal solutions in the exact static model , without contradiction with our result about the ladder approximation , which is not supposed to be gauge - invariant .
writing @xmath218 in the bethe - salpeter equation at the ladder approximation , and taking the @xmath219 limit , we get directly the symmetrical equation ( [ 4.11 ] ) .
the zero - d reduction of ( [ 4.11 ] ) , using ( [ 4.16 ] ) , gives @xmath151\ , \left\{{1\over { 1\over4 } w^2-\omega_1 ^ 2-i\epsilon}+2i\pi \,{\delta(\omega_1)\over w}\right\}\,\exp[i\vert \omega_1\!-\!0\vert]\]]@xmath152\ , \left\{{1\over { 1\over4 } w^2-\omega_2 ^ 2-i\epsilon}+2i\pi \,{\delta(\omega_2)\over w}\right\}\,\exp[i\vert \omega_2\!-\!\omega_1\vert]\,\]]@xmath220+ ... \bigg ] \label{5.7}\ ] ] the terms between brackets correspond to the two first terms ( modified ladder part ) of ( [ 5.5 ] ) .
we pointed out that the cancellation mechanism which transforms some well chosen 3d reductions of the exact bethe - salpeter equation into a simple dirac equation ( for two fermions in qed ) at the one - body limit , implies that these reductions have no abnormal solution at this limit , nor at the zero - body limit ( static model ) .
conversely , this cancellation does not occur at the ladder approximation , so that the abnormal solutions are present in this case .
our argumentation is based on the fact that the abnormal solutions , when present , are implicitly preserved by the 3d reduction , via the dependence of the potential on the total energy , and should thus disappear when this potential reduces to a simple compact potential , like a coulomb potential .
this cancellation mechanism in the 3d reductions at the one - body limit occurs in many situations ( 2 fermions , 2 bosons , 1 boson + 1 fermion with various interactions ) @xcite , providing many new examples for the discussion of the abnormal solutions problem .
the 3d reduction must be based on a constraint which puts , at the one - body limit , the heavy particle on its mass shell .
putting directly one particle on its mass shell , as in the 3d reduction of gross @xcite without taking the one - body limit , is however not enough .
these facts and the result of ref .
@xcite about the absence of abnormal solutions in the exact static model , support wick s conjecture that the abnormal solutions are spurious consequences of the ladder approximation , and would disappear in the exact model .
however , it remains possible that the exact model has abnormal solutions , which are killed by the one - body limit , the 3d reduction ( if the corresponding bethe - salpeter amplitudes vanish for the relative energy fixed in the reduction ) , or the combination of both .
this does clearly not happen with the normal solutions , and our study of the static model has shown that it does not happen either with the abnormal solutions at the ladder approximation .
the main problem of the existence of abnormal solutions of the exact , finite masses bethe - salpeter equation remains thus open .
what we have shown in this work is that the higher - order contributions to the 3d potential of well chosen reduced equations for the exact model cancel mutually at leading order in each of the masses .
the dependence on the total energy , which is related to the presence of abnormal solutions , is thus much less important than at the ladder approximation , and possibly too small to allow for more than one root of @xmath221 it could perhaps be possible to eliminate this dependence completely order by order , using the large freedom which we still have in the choice of the 3d reduction .
we made clear anyway that , although the non - ladder contributions can be treated as correction terms in the study of the normal solutions , they modify completely ( if not annihilate ) the abnormal solutions obtained at the ladder approximation .
the detection and computation of abnormal solutions goes thus beyond the validity of the ladder approximation .
salpeter and h.a .
bethe : _ phys .
_ * 84 * ( 1951 ) 1232 .
m. gell - mann and f. low : _ phys .
* 84 * ( 1951 ) 350 .
g. c. wick : _ phys .
_ * 96 * ( 1954 ) 1124 .
cutkosky : _ phys .
* 96 * ( 1954 ) 1135 .
y. ohnuki , y. takao and h. umezawa : _ prog .
* 23 * ( 1960 ) 273 .
n. nakanishi , prog .
* 43 * ( 1969 ) 1 ; , edited by o. hara , s. ishida and s. naka ( world scientific singapore 1992 ) , p.109 ; ( 1965 ) b1182 , * 139 * ( 1965 ) b1401 , _ ibid . _ ab1 ( e )
. n. mugibayashi : _ prog .
* 25 * ( 1961 ) 803 .
h. jallouli and h. sazdjian : _ journal of physics g _ * 22 * ( 1996 ) 1119 .
f. gross : _ phys .
* 186 * ( 1969 ) 1448 , _ phys .
rev . _ * c26 * ( 1982 ) 2203 .
w. dietrich : _ phys .
_ * d1 * ( 1970 ) 3345 .
neghabian and w. glckle : _ can .
_ * 61 * ( 1983 ) 85 .
j. mourad and h. sazdjian : _ j. math .
* 35 * ( 1994 ) 6379 ; _ j. phys
* 21 * ( 1995 ) 267 .
j. bijtebier and j. broekaert : _ journal of physics g _ * 22 * ( 1996 ) 559 ; * 22 * ( 1996 ) 1727 ; k.m .
maung , j.w .
norbury and d.h .
kahana : proceedings of the international conference on quark confinement and the hadron spectrum , p. 297
( world scientific 1994 ) . j. bijtebier and j. broekaert : _ nuovo cimento _ * a107 * ( 1994 ) 1275 .
h. sazdjian : _ j. math
_ * 28 * ( 1987 ) 2618 , * 29 * ( 1988 ) 1620 .
g. s. adkins : application of the bound state formalism to positronium , in w. johnson , p. mohr and j. sucher eds .
_ relativistic , quantum electrodynamic , and weak interaction effects in atoms , aip conference proceedings 189 _ ( new york 1989 ) .
g. breit : _ phys .
rev . _ * 34 * ( 1929 ) 553 , * 36 * ( 1930 ) 383 , * 39 * ( 1932 ) 616 .
h. jallouli : thesis ( orsay 1997 ) . | we show that the 3d reductions of the bethe - salpeter equation have the same bound state spectrum as the original equation , with the possible exception of some solutions for which the corresponding 3d wave function vanishes .
the abnormal solutions of the bethe - salpeter equation ( corresponding to excitations in the relative time - energy degree of freedom ) , when they exist , are recovered in the 3d reductions via a complicated dependence of the final potential on the total energy .
we know however that the one - body ( or one high mass ) limit of some 3d reductions of the exact bethe - salpeter equation leads to a compact 3d equation ( by a mutual cancellation of the ladder and crossed graph contributions ) , which does not exhibit this kind of dependence on the total energy anymore .
we conclude that the exact bethe - salpeter equation has no abnormal solution at this limit , or has only solutions for which our 3d wave function vanishes .
this is in contrast with the results of the ladder approximation , where no such cancellation occurs .
we draw the same conclusions for the static model , which we obtain by letting the mass of the lighter particle go also to infinity .
these results support wick s conjecture that the abnormal solutions are a spurious consequence of the ladder approximation .
pacs 11.10.qr relativistic wave equations .
pacs 11.10.st bound and unstable states ; bethe - salpeter equations .
pacs 12.20.ds specific calculations and limits of quantum electrodynamics . + + keywords : bethe - salpeter equations .
salpeter s equation .
breit s equation .
relativistic bound states .
relativistic wave equations .
+ + |
recently it was realized that a strongly correlated 1d system , namely a chiral luttinger liquid ( @xmath3ll ) , exists in the edges of fractional quantum hall ( fqh ) liquids @xcite . because of their chiral nature , _
i.e. _ , the excitations in a given branch move only in one direction , spatially separated branches can not interfere and cause localization of states .
in contrast , non - chiral luttinger liquids are extremely sensitive to the presence of even the smallest amounts of impurity in the sample , for in 1d all states are localized , and long enough wires will behave as insulators . the characteristic property of ( chiral )
luttinger liquids is that the tunneling conductance between the edge states has a power law dependence on the temperature @xmath4 , where @xmath5 depends on the filling factor @xmath6 of the fqh state , taking the values @xmath7 or @xmath8 depending on the tunneling geometry .
by experimentally studying the tunneling between edge states in the fqh regime using a point contact geometry , milliken , umbach , and webb @xcite found this type of power law dependence of the tunneling conductance on the temperature .
their finding is consistent with the theoretical prediction @xmath9 for the @xmath10 fqh state @xcite .
the experimental confirmation of the luttinger liquid behavior in tunneling between edge states has boosted theoretical interest in further studies of properties of the conductance @xcite .
an exact solution for the conductance has been obtained using the thermodynamic bethe ansatz , and an exact duality between the @xmath5 and @xmath11 cases has been shown @xcite in the context of the tunneling current , as suggested in ref .
. the rich behavior of tunneling in chiral luttinger liquids extends well beyond transport measurements alone .
one should expect , based on experience in non - interacting systems , that the noise spectrum contains information not attainable , in the most general case , from just transport measurements . in general , the shape of the noise spectrum is determined by the dynamical properties of the system , which in turn contain information about the excited states . even for non - interacting electronic systems ,
non - trivial structures appear in the noise @xcite , the simplest example being the suppression of classical shot noise due to quantum statistics . in chiral luttinger liquids ,
the tunneling particles sometimes carry fractional charge and fractional statistics , and thus such strongly correlated 1d systems also provide the natural experimental realization for the study of features that arise in the noise spectrum for generalized quantum statistics . in reference
@xcite , the noise spectrum of the tunneling current between edge states directly at the point contact was calculated perturbatively .
to low orders in the tunneling amplitude , we found that there was a singularity at @xmath12 ; for small @xmath0 the noise spectrum has the form @xmath13 , where @xmath14 is the zero - frequency shot noise and @xmath15 .
the slope @xmath16 of the @xmath17 singularity has a strong non - linear dependence on the applied voltage @xmath18 ( @xmath19 ) , which is another signature of luttinger liquid behavior ( to be contrasted with the case of non - interacting electrons , @xmath20 , where the slope is independent of @xmath18 ) .
the exponent @xmath5 characterizes the luttinger liquid behavior .
this low - frequency part of the spectrum is the one more easily accessible experimentally .
secondly , there is another singularity at @xmath21 where @xmath22 is the josephson frequency of the electron ( @xmath23 ) or quasiparticle ( @xmath24 ) that tunnels through the point contact .
the shape of this singularity depends on @xmath5 and goes as @xmath25 .
measurements of the location of this singularity would give the value @xmath2 of the charge of the carriers of the current , which would be yet another way of observing fractional charge from noise measurements .
the method originally suggested is to measure the shot noise , which for small tunneling amplitude is related to the tunneling current @xmath26 by @xmath27 .
lastly , for @xmath28 , we found that the singularities at both @xmath12 and @xmath21 should persist to all orders in perturbation theory .
these results present a puzzle which we describe below and address in this paper .
the case of @xmath29 corresponds to a single quantum hall droplet with a constriction . in this case
, quasiparticles can tunnel across the constriction , from one edge to the other ( see fig . 1a ) .
these quasiparticles have fractional charge @xmath2 , given by @xmath30 . if the constriction is made narrower , the tunneling amplitude will increase . as the constriction is further narrowed ,
eventually the droplet will break into two disconnected pieces , and now only electrons will be able to tunnel from one edge to the other ( see fig .
their tunneling should once again behave like tunneling in a chiral luttinger liquid , but with new exponent @xmath31 and charge equal to @xmath32 .
this is the physical picture behind the duality seen in reference @xcite ; as the tunneling amplitude is increased ( or the voltage is decreased ) @xmath5 goes to @xmath11 . similarly ,
if we start with the two quantum hall droplets with exponent @xmath33 and increase the tunneling amplitude of the electron , eventually we will obtain the single droplet picture with exponent @xmath34 . in light of this duality and the results of reference @xcite , the following question arises .
if we start with the two disconnected droplets then we expect the singularity in the noise to occur at multiples of @xmath35 , the josephson frequency for the electron .
as the tunneling amplitude is increased , at some point we expect the singularity at the josephson frequency for the quasiparticle , @xmath22 , to appear . however , according to the perturbative calculations , to all orders in the electron tunneling amplitude the quasiparticle singularity does not appear . in this paper
we will address the question of what happens to this quasiparticle singularity and show how these two seemingly contradictory statements are resolved in the special case of @xmath36 and @xmath37 . in short , our results suggest that the singularity at the quasiparticle josephson frequency @xmath1 is destroyed by non - perturbtive effects . in some sense , the singularity obtained by perturbative calculations is smeared for finite tunneling strength .
this question is of interest because the location of the singularities tells us which particles are tunneling and , as mentioned above , should give a way to measure the fractional charge of the quasiparticles .
the results in this paper suggest what may be happening at other values of @xmath5 also , and we are currently working on this question .
one of the tools we use in this paper is the landauer - buttiker scattering approach .
the geometry is illustrated in figure 2 .
the choice of the landauer - buttiker approach is justified for a number of reasons .
the chiral nature of the system under study naturally poses the problem in terms of incoming and outgoing scattering states to and from the point contact region .
the incoming branches should be in equilibrium with their respective reservoirs of departure , and should be insensitive to the tunneling of charges in the tunneling region shown in fig .
this is so because information on tunneling events can not propagate in the direction opposite to the incoming branch chirality .
also , the landauer - buttiker approach and the chiral nature of the system suggest naturally a four - terminal geometry for experimental measurements , probing voltage fluctuations in the two incoming and two outgoing branchs .
the tunneling takes place in the point contact , or scattering region , which is not directly accessible by the probing leads .
auto - correlations of current / voltage fluctuations measured in the four terminals , as well as cross - correlations between different terminals , are the experimental probes that should allow the remote measurement of the tunneling events and noise spectrum .
the paper is organized as follows . in section
ii we briefly review the bosonization scheme for chiral luttinger liquids . in section iii
we obtain the noise spectrum perturbatively for the four terminal geometry , using the keldysh non - equilibrium formalism .
we show that only the noise spectrum for the outgoing branches is affected by the tunneling , whereas the incoming branches are completely insensitive to the charge transfer between the edges .
this is consistent with the landauer - buttiker picture and the chirality of the system . the noise spectrum for the incoming branches can thus serve as reference level for the measurement of the excess noise on the outgoing branches due to tunneling .
the noise spectrum obtained contains interesting structures both at low and high frequencies .
the tunneling excess noise vanishes for frequencies above the josephson frequency @xmath1 .
the issue of how the singularity moves from the quasiparticle frequency to the electron frequency is resolved in section iv , where we use the landauer - buttiker approach to solve exactly for the noise spectrum in the case of @xmath38 , for which the problem can be cast as a free fermion problem .
we show that the singularity at the quasiparticle frequency is smeared for finite tunneling and is not a true singularity , whereas the singularity at the electron frequency survives for all non - zero coupling . in section
v we discuss the duality when @xmath38 goes to @xmath39 , which we show is not exact in the naive sense for the case of noise , in contrast to the case of conductance .
we find that the noise spectrum of the current correlations on a single branch ( auto - correlations ) satisfies the duality relation , while current correlations between distinct branches ( cross - correlations ) do not satisfy the naive duality relation .
we show that the correct dual lagrangian to the @xmath38 theory is the @xmath40 theory plus a neutral density - density coupling , which has the same dimension as the tunneling operator .
the effect of the neutral coupling appears in the noise , but not in the conductance .
in this section we shall briefly review the bosonization scheme for edge states in the fqh effect ( for a thorough review , see ref .
the lagrangian we will use is better cast in this bosonic language .
the right and left moving excitations along the edges can be described by boson fields @xmath41 .
right and left moving electron and quasiparticle operators on the edges of a fqh liquid can be written as @xmath42 , where @xmath5 is related to the fqh bulk state .
for example , for a laughlin state with filling fraction @xmath43 we have @xmath44 for electrons and @xmath45 for quasiparticles carrying fractional charge @xmath46 .
the @xmath41 fields satisfy the equal - time commutation relations @xmath47=\pm i\pi \ { \rm sign}(x - y)\ { } .\ ] ] the dynamics of @xmath41
is described by @xmath48 where @xmath49 is the velocity of edge excitations ( which we will set to 1 ) .
density operators can be defined in terms of the @xmath41 through @xmath50 . here , for convenience , we have set the unit charge in the definition of the density to be the electron charge @xmath32 , so that @xmath51 and @xmath52 .
one can verify that @xmath53=\sqrt{\nu g}\ \psi^\dagger_{r , l}(t , y)\delta ( x - y)$ ] , so that indeed the cases @xmath8 and @xmath7 correspond to electron and quasiparticle charged operators , respectively .
the tunneling operators from right to left moving branches and vice - versa can be written as @xmath54 and @xmath55 .
thus we can write , in terms of @xmath56 , the following total lagrangian density : @xmath57- \ \gamma \delta(x)\ e^{i\sqrt{g}\phi(t,0)}+h.c.\ \ , \label{l}\ ] ] with @xmath58 satisfying @xmath59=4\pi i\delta(x - y)$ ] . in the following
we will set the edge velocity @xmath60 .
a voltage difference between the two edges of the qh liquid can be easily introduced in the model by letting @xmath61 , where @xmath62 , with @xmath23 for electron tunneling and @xmath63 for quasiparticle tunneling . in the following sections we will study non - equilibrium noise in chiral luttinger liquids described by the model above .
in this section we treat the tunneling between edge states perturbatively , and obtain the noise spectrum for the current / voltage fluctuations at the four leads as shown in fig .
2 . in the figure
we separate the branches into their right and left moving components , as well as incident and scattered ones .
right and left branches are incoming or outgoing depending on their position relative to the scatterer : @xmath64 both the currents and the densities at the four terminals can be related to the fields @xmath41 .
the densities are simply given by @xmath50 .
voltage measurements probe these densities .
the currents at the four terminals can be trivially related to the densities at those same terminals through the continuity equation for @xmath65 .
the currents are given by @xmath66 , with positive currents flowing to the right . by choosing the convention that positive currents flow in the direction of the arrows in fig .
2 , we can write new currents @xmath67 .
it then becomes transparent that there is a tight relationship between current and voltage in the chiral branches .
for example , measuring the noise in either the current or the voltage yields information about the other .
this kind of relationship between voltage and current noise was obtained in ref . .
we will thus focus on the calculation of density - density correlations , for these will give us information on both current and voltage noise .
we will label the densities at the four terminals shown in fig . 2 by @xmath68 , @xmath69 .
in terms of the right and left moving fields we have : @xmath70 where @xmath71 , @xmath72 . the noise spectrum of the density fluctuations in terminals @xmath73 is obtained from the correlations between the densities @xmath74 : @xmath75 these quantities are calculated perturbatively in appendix a , using the techniques in ref .
the components with @xmath76 are very sensitive to phases which depend on the position of the probes @xmath77 and @xmath78 .
these phases cancel in the case of auto - correlations , _
i.e. _ , when @xmath79 .
the quantities @xmath80 , which correspond to the noise spectrum obtained entirely from one of the four probes for @xmath81=1 to 4 , are thus the most robust measurements of fluctuations , because when they are extracted away from the junction they are independent of the position @xmath77 where they are taken . to second order in perturbation theory , @xmath82 is given by : @xmath83 where @xmath84 using the perturbative result to order @xmath85 for the tunneling current @xmath86 @xcite , @xmath87 can be wrtitten as @xmath88 notice that the effects of tunneling are contained in @xmath87 , and only appear in the outgoing branches , terminals @xmath89 .
the incoming branches are insensitive to the tunneling between edges , due to the chiral nature of the system .
information about what goes on in the junction can not propagate in the direction opposite to the chirality of the branch , and therefore the noise in the incoming branches is independent of the tunneling of charged particles between edges .
this result of chirality is clear within the landauer - buttiker scattering approach .
another physical consequence closely related to this is the fact that the average voltage along the branches remains constant outside the scattering region . the second point to notice from eq .
( [ noise2ndorder ] ) is that to order @xmath85 the noise in the outgoing branches that is in excess to the noise in the incoming branches has a singularity at the josephson frequency @xmath90 , vanishing for @xmath91 , as illustrated in fig .
the non - equilibrium voltage @xmath18 determines the frequency scale @xmath1 , up to which there is structure in the excess noise due to tunneling .
such vanishing of the excess noise spectrum past a frequency set by the non - equilibrium voltage should be familiar to readers accustomed to noise in non - interacting systems ( @xmath20 ) , in which the excess noise goes to zero linearly at the josephson frequency @xcite .
this point will be illustrated further in the next section , when we will have at hand the exact solution for the noise spectrum in the case of @xmath38 .
the strong coupling limit of the solution for @xmath38 also gives us the solution for @xmath40 , which we shall use for comparison purposes .
the last , and most important , point about this high frequency singularity in the noise spectrum is in regard to the connection between the two dual pictures illustrated in fig .
1 . in ref . @xcite it was pointed out that the singularity at the josephson frequency remained to all orders in perturbation theory
however , the perturbative expansion for the geometries in figs .
1a and 1b yields two distinct frequencies , namely the quasiparticle frequency @xmath92 when quasiparticles are the tunneling charges ( fig .
1a ) , and the electron frequency @xmath93 , when electrons are the tunneling current carriers ( fig . 1b ) .
these configurations are connected in the sense that one is the strong coupling limit of the other , and thus there should be a non - perturbative mechanism by which the singularity moves from one place to the other .
this was the clearest open question in ref .
@xcite , and which we can answer by focusing on the exactly solvable case of @xmath38 .
another exactly solvable point is the trivial case @xmath20 , which unfortunately can not be used to address this issue of the singularity in the noise spectrum because in this case the two frequencies @xmath94 and @xmath95 coincide . before answering the question about the high frequency singularity
, we will close this section with the implications of tunneling between edge states to the low frequency noise measured in the four terminal geometry . in ref .
@xcite , a correction to the low frequency shot noise spectrum was found , which corresponded to an @xmath17 singularity , or a cusp , in the noise spectrum .
this correction was found to order @xmath96 , while to order @xmath85 the low frequency corrections to the flat shot noise started as @xmath97 . in the four terminal geometry proposed in this paper ,
what is probed is not the tunneling current in the junction area ( as in ref .
@xcite ) , but its consequences in the current / voltage in the four terminals away from the scattering region . the four terminal measurement , as seen from eq .
( [ noise2ndorder ] ) , does have a correction @xmath98 to order @xmath85 . for @xmath99
we have , for example , @xmath100\ .
\nonumber\end{aligned}\ ] ] one recovers the classical shot noise expression for @xmath101 .
notice that , since these results are valid only to order @xmath85 , there is no correction to the classical shot noise expression for @xmath101 .
corrections appear at order @xmath96 ( see ref .
@xcite ) . also notice that the non - zero @xmath0 corrections to the shot noise depend on whether @xmath5 is larger or smaller then @xmath102 . for @xmath103 , the difference between the outgoing and incoming spectra ( the @xmath104 above , for example ) decreases with @xmath0 , whereas for @xmath105 it increases .
in this section we will use the landauer - buttiker scattering approach to obtain an exact solution for the noise when @xmath36 . in this approach ,
we use the quantum equations of motion derived from the hamiltonian to solve for the scattering states .
these scattering states describe free left movers and right movers that are incident on the impurity and then are reflected or scattered by the impurity .
the solutions for these states can be used to calculate the conductance and the noise in the various branches .
the advantage of focussing on @xmath36 is that for this value of @xmath5 the system can be described by free fermions @xcite , making it straightforward to solve for the scattering states .
however , already at @xmath36 , we expect to see singularities in the noise at @xmath22 , corresponding to quasiparticle tunneling .
as the tunneling amplitude @xmath106 increases ( or @xmath18 decreases ) , we expect to obtain the dual picture at @xmath107 , with electrons tunneling and a singularity at @xmath108 .
thus the full solution at @xmath36 will show us what happens to the quasiparticle singularity as @xmath106 is increased .
the hope is that the qualitative behavior of these results will also apply for other values of @xmath5 . when @xmath109 , the hamiltonian for the system is given by @xmath110 where @xmath111 are the free hamiltonians for the right and left moving fields , and @xmath112 , with @xmath113 .
the hamiltonian can be recast in terms of new chiral fields @xmath114 : @xmath115 the densities of the new fields @xmath116 are related to the densities @xmath117 by @xmath118 .
notice that the @xmath119 fields are decoupled in eq.([hpm ] ) , and the hamiltonian for @xmath120 is simply the free @xmath121 .
the hamiltonian for @xmath122 can be fermionized by defining @xmath123 .
one can check that @xmath124 defined as such satisfies the proper commutation relations @xmath125 . in terms of the fermionic fields @xmath126 ,
the hamiltonian @xmath127 is : @xmath128 \eta(x ) + \sqrt{2\pi}\delta(x)\left[\gamma \eta(x ) + \gamma^ * \eta^\dagger(x)\right ] { \big \ } } , \label{heta}\ ] ] where we absorbed the oscillating phases @xmath129 into a redefinition of the chemical potential .
the hamiltonian above contains terms linear in the fermionic fields @xmath124 and @xmath130 , which prevent a direct calculation of the commutators that would give us the equations of motion for the fields .
this problem can be circumvented by redefining the fermionic fields to be @xmath131 , with @xmath132 and @xmath133 , as in ref . @xcite .
more formally , such a transformation can be constructed from the proper handling of the zero modes of the bosonic fields @xmath58 @xcite , and one can identify @xmath134 with @xmath135 , the fermion counting operator commonly used to switch from periodic to anti - periodic boundary conditions in fermionic conformal field theories .
the hamiltonian we will use in the exact solution of the noise spectrum for the @xmath38 case is the one written in terms of the @xmath136 fields and @xmath134 : @xmath137 \psi(x ) + \sqrt{2\pi}\delta(x)\left[\gamma \psi(x)f + \gamma^ * f \psi^\dagger(x ) \right ] { \big \ } } , \label{hpsi } % \ ] ] where the non - vanishing equal - time commutation relations between @xmath138 and @xmath134 are @xmath139 the density @xmath140 can be written in terms of the fields @xmath141 and @xmath142 as @xmath143 , so that all correlations between @xmath140 s can be derived from the correlations of the fermions .
the fermionic model is solved using the equations of motion obtained by commuting the operators @xmath144 and @xmath134 with the hamiltonian : @xmath145 = ( i\partial_x + \omega_0 )
\psi(x ) + \sqrt{2\pi}\gamma^ * f\delta(x ) , \label{eompsi}\\ % % -i\partial_t \psi^\dagger(x ) & = & [ h , \psi^\dagger(x ) ] = ( i\partial_x - \omega_0 ) \psi^\dagger(x ) - \sqrt{2\pi}\gamma f\delta(x ) , \label{eompsid}\end{aligned}\ ] ] and @xmath146 = 2\sqrt{2\pi}\left[\gamma \psi(0 ) - \gamma^*\psi^\dagger(0)\right ] . \label{eomf}\ ] ] according to these equations , for @xmath147 , the field @xmath141 satisfies the free equation of motion for a rightmover with energy shifted by @xmath148 : @xmath149 at @xmath150 , it picks up a discontinuity because of the impurity . in order to preserve unitarity and obtain the proper commutation relations in the solutions of @xmath141 , in equation ( [ eomf ] )
the field @xmath151 must be given by @xmath152 .
with this definition , it is straightforward to solve the equations of motion .
the solutions are given by @xmath153 and @xmath154 where @xmath155 and @xmath156 given the commutation relation for @xmath141 , the @xmath157 satisfy the following commutation relation : @xmath158 these solutions can be interpreted as having an incident particle at energy @xmath0 that scatters into a particle with energy @xmath0 and a hole with energy @xmath159 ( see fig .
both the particle and hole scattering involve an energy dependent phase shift .
the reservoir is located to the left of the impurity , for some @xmath160 . to obtain the scattering state @xmath161
, we assume that the states leaving the reservoir are in equilibrium with the reservoir , which has energy @xmath148 .
thus , for @xmath162 , at zero temperature all the states with @xmath163 are filled .
this means that @xmath164 and @xmath165 using the commutation relations for @xmath166 in equation ( [ acom ] ) , we then find that @xmath167 and @xmath168 where @xmath169 in this paper , we will just concentrate on the case when @xmath170 .
however , we can obtain the finite temperature results by replacing @xmath171 with @xmath172 we can now use the solutions for @xmath141 and the scattering state to solve for the noise in both incoming and outgoing channels .
our calculations will closely follow those by buttiker in reference @xcite .
the noise is given by @xmath173 where we take only the connected part of the correlation function , and @xmath174 and @xmath175 are positive or negative depending on whether the current is evaluated in the incoming or outgoing channel .
we will begin by calculating the noise when @xmath176 . in this case , both of the currents are evaluated on the same side of the impurity . because of the time translational invariance of the correlators , the expression for the noise simplifies to @xmath177 to find the noise in the incoming channel , we must evaluate the expectation value @xmath178 with @xmath179 . using the solutions ( [ psixminus ] ) and ( [ psidxminus ] ) for @xmath141 and @xmath142
, we find @xmath180 this expectation value , and the resulting integrals for @xmath181 are evaluated in appendix c , with the result , @xmath182 if we want to calculate the noise in one of the two original @xmath183 and @xmath184 incoming branches , we must use the relations @xmath185 then the density - density correlations can be evaluated as follows : @xmath186 where the last equality follows from the fact that @xmath187 and @xmath140 are decoupled .
recall that @xmath187 is a free field , so that the contribution to the noise from @xmath187 is simply @xmath188 .
we find that the noise in each of the two incoming @xmath183 and @xmath184 branches is given by @xmath189 just as we found in the perturbative calculation with @xmath190 in eqs .
( [ s11seciii ] ) to ( [ s2seciii ] ) . using this scattering approach ,
it is clear that for these two incoming probes the noise is the same as for a free system , because in these two channels the densities have not yet reached the impurity .
next , we will calculate the noise in the outgoing current .
this time we must evaluate the correlator @xmath191 with @xmath192 . according to equations ( [ psixplus ] ) and ( [ psidxplus ] ) ,
this is equal to @xmath193 when we expand the @xmath194 s in terms of the @xmath166 s , we will obtain two different types of processes ( see fig .
5 ) . in the first , at time @xmath195 one particle is created while another is destroyed , and then at time @xmath196 the first particle is destroyed and another is created . in terms of the original tunneling picture , this describes the process where both at time @xmath196 and at time @xmath195 one quasiparticle tunnels from the left branch to the right branch and another tunnels in the opposite direction . in the second process , at time @xmath195 two particles are created and then at time @xmath196 they are destroyed ( or _ vice versa _ ) . in the original tunneling picture , this corresponds to two quasiparticles tunneling in one direction at time @xmath195 and two quasiparticles tunneling in the opposite direction at time @xmath196 . as shown in appendix c , this second process is responsible for the electron singularity at @xmath197 . in appendix c , the expectation values in eq.([ipipcalc ] ) and the integrals for @xmath198 are evaluated .
we find that the noise on the outgoing side of the impurity is @xmath199
\bigg\}. \label{spp}\end{aligned}\ ] ] to compare with our perturbative calculation for the noise in the original four probe geometry , we again make use of equation ( [ rhorhoii ] ) . thus , the noise in the two outgoing branches is related to @xmath198 as follows : @xmath200 the first striking feature to note in equation ( [ spp ] ) is that the noise due to the tunneling vanishes identically for @xmath201 .
this means that whenever @xmath17 is larger than the electron frequency , the noise shows no sign of the impurity ; it is the same as for the incoming branch .
this is also what happens for the free electron case , with @xmath20 .
to second order in perturbation theory , this is indeed the case for any @xmath5 , as seen in the previous section .
the strength of the results presented here is that for _ any value _ of the coupling @xmath106 the noise vanishes above the electron frequency when @xmath202 and @xmath203 .
( the last case , @xmath40 , is obtained by resorting to the strong coupling limit of the @xmath38 case . )
it is not clear whether this will happen for the other values of @xmath5 beyond second order in perturbation theory .
next , we can expand @xmath198 out for small and large @xmath204 to compare with the perturbative results .
as @xmath204 goes to zero , the noise becomes @xmath205 this agrees with the perturbative result for @xmath36 .
we note that the quasiparticle singularity arises because we took the @xmath206 limit of the arctangents .
in addition , because this step function is already zero for @xmath207 , the electron singularity at @xmath208 drops out .
thus , to this order we only have the quasiparticle singularity .
however , for any finite value of @xmath204 the quasiparticle singularity becomes smoothed out and the electron singularity appears . as we shall see later , though , the `` smoothed out '' quasiparticle singularity is still a more distinctive feature in the plots of the full noise than is the electron singularity .
next , for @xmath209 , the noise becomes @xmath210 if we make the identification that @xmath211 , the tunneling amplitude for @xmath36 , is related to @xmath212 , the tunneling amplitude for @xmath40 , by @xmath213 then this answer agrees with the perturbative result for @xmath40 .
( to make the comparison , we must recall that the @xmath148 in this equation corresponds to the josephson frequency for the quasiparticle , whereas the @xmath90 in the perturbative calculation eqs .
( [ s11seciii ] ) to ( [ s2seciii ] ) is the josephson frequency for the electron , which is twice as large . ) in addition , the expansion in @xmath214 of the scattering solution only contains powers of @xmath215 , and at every order in @xmath216 the electron singularity at @xmath208 remains .
these two properties also agree with the perturbative results found in @xcite .
lastly , we can make use of the scaling properties of the noise to write @xmath217 as a function only of @xmath218 and @xmath219 . the noise is then given by @xmath220 \nonumber\\ & \ & \ \ \ + { 1\over 2|\tilde \omega| } \left[\ln\left(1 + ( |\tilde \omega| - |\tilde \omega_0|)^2\right ) -\ln\left(1 + \tilde \omega_0 ^ 2\right)\right ] \bigg\}. \label{sscaled}\end{aligned}\ ] ] in figure 6a , the excess noise @xmath221 is plotted against @xmath222 for different values of @xmath223 . as @xmath224 becomes large , the excess noise approaches the step function in eq .
( [ sttsmallg ] ) . recall that @xmath225 , so this limit is equal to the weak coupling limit with @xmath206 . to see the strong coupling limit , in fig .
6b we plot the excess noise divided by @xmath226 ( in order to fit in the same scale ) .
as @xmath227 becomes small , this clearly has the cubic behavior in eq .
( [ sttscatlg ] ) .
finally , the full noise , divided by @xmath227 , is plotted in fig .
the cubic singularity at @xmath228 decays too quickly to appear in the full noise .
however , for @xmath229 and @xmath230 , there is clearly a `` blip '' in the plot of the noise , which shows the `` smoothed out '' quasiparticle singularity . for completeness
, we will conclude this section by giving the result for the noise @xmath231 between incoming and outgoing currents . by comparing this with the perturbative calculations of the cross - correlations
, we will see to what extent the duality symmetry holds .
in addition , once we have @xmath231 , @xmath198 and @xmath181 , we can calculate the noise in the hall current and the tunneling current .
the hall current is the total current running down the sample , given by @xmath232 and the tunneling current is the current that tunnels across the sample , which is given by @xmath233 .
the expression for the @xmath231 noise is @xmath234 where @xmath179 and @xmath192 .
again , we can expand the @xmath235 and @xmath236 in terms of the @xmath157 s and @xmath237 s in the solution for @xmath141 .
after evaluating the expectation values and performing the integrals over @xmath238 and @xmath196 , we find @xmath239 \nonumber \\ & + & i|\gamma|^2 { \rm sign}(\omega)\bigg [ 2\ln\left(\omega_0 ^ 2 + ( 4\pi|\gamma|^2)^2 \right ) - \ln\left((\omega + \omega_0)^2 + ( 4\pi |\gamma|^2)^2 \right ) \nonumber \\ & \ & \ \ \ \ \ \ - \ln\left((\omega - \omega_0)^2 + ( 4\pi |\gamma|^2)^2 \right)\bigg ] \bigg\ } e^{i\omega(x_+ - x_-)}. \label{spm}\end{aligned}\ ] ] we can again use equation ( [ rhorhoii ] ) to obtain the expression for the cross - correlations of the currents in the original four reservoirs . we find , for example , @xmath240 and @xmath241 the other cross - correlations , namely @xmath242 and @xmath243 , can be calculated similarly . for small @xmath204 ,
the noise is @xmath244 e^{i\omega(x_+ - x_- ) } , \label{stosg}\end{aligned}\ ] ] and when @xmath204 is large , the noise becomes @xmath245\bigg\ } e^{i\omega(x_+ - x_-)}.\end{aligned}\ ] ] in the following section , we will compare these results with the perturbative calculation .
we will find that for @xmath36 they agree , but they differ for @xmath40 . in section
v , we will also discuss this apparent breakdown of the duality transformation .
as we have seen in the previous sections , we expect this system to exhibit a duality symmetry . in this section
, we will first describe this duality symmetry more fully , and then compare the results from the perturbative and scattering calculation to see how consistent they are with this symmetry . for @xmath36 ,
the original picture of this system is a single quantum hall droplet with `` filling fraction '' @xmath246 .
quasiparticles can tunnel from one branch to the other , and they have charge @xmath247 , tunneling amplitude @xmath248 , and josephson frequency @xmath249 .
the lagrangian describing this system can be written as @xmath250 -\gamma_q e^{-i\omega_0 t } \delta(x ) e^{i\sqrt g \phi(t,0 ) } + h.c.,\ ] ] with @xmath36 and @xmath251 .
if we use the four probe geometry to study this system , then eq .
( [ rhonumdef ] ) gives the relation between the densities in the four probes , @xmath252 , @xmath253 , @xmath254 , and @xmath255 , and the densities of the leftmovers and rightmovers .
they are shown in fig . 7a . once @xmath106 is increased ( or @xmath18
is decreased ) , the droplet should split into two .
each of the two new droplets is still characterized by filling fraction @xmath6 .
however , now only electrons can tunnel across the gap from one branch to the other . for @xmath36 ,
the electron is made up of two quasiparticles , so the tunneling operator for the electron should be @xmath256 the charge is @xmath32 , and the josephson frequency is @xmath257 .
thus , when @xmath248 becomes large , this system can also be described by the lagrangian density @xmath250 -\gamma_e e^{-i\tilde\omega_0 t } \delta(x ) e^{i\sqrt { \tilde g } \phi(t,0 ) } + h.c.,\ ] ] where @xmath37 . however , in this geometry with the two droplets , we must be careful when we write the densities in the four probes in terms of the left - moving and right - moving densities . according to fig .
7b , this relation is given by @xmath258 with these identifications , @xmath259 in the four probe geometry equals @xmath260 in the two - droplet geometry , and similarly , @xmath261 is given by @xmath262 , where @xmath263 and @xmath264 .
also , we see that the hall current in the single droplet , @xmath265 is dual to the tunneling current in the two droplets , @xmath266 , because both are equivalent to @xmath267 we will first verify that the scattering and perturbative calculations agree for @xmath36 . we have already found that when the noise is evaluated on only one side of the junction , then both the scattering and perturbative results agree .
if one probe is in an incoming channel and the other probe is in an outgoing channel , then according to appendix a the perturbative result for the noise is @xmath268\bigg]\bigg\},\end{aligned}\ ] ] where we have set @xmath269 in eq .
( [ srrpert ] ) . on comparing this with the expansion of the scattering calculation for small @xmath106 in eq .
( [ stosg ] ) , we find that also in this case the scattering and perturbative results agree .
next , to check the duality transformation , we must compare the scattering calculation as @xmath270 with the perturbative calculation at @xmath40 .
again we found that if both probes are in the same branch then the two calculations agree .
this is rather remarkable , because when @xmath40 the system can be sensitive to short distance behavior , which means that it could depend on the detailed structure of the junction and on how it is regulated .
however , here we found that the weak - coupling perturbative calculation and the strong - coupling limit of the scattering calculation are the same , even though they treat the junction very differently . we conclude that , at least to the order in perturbation theory that we have calculated , the noise extracted from a single channel is not affected by the short - distance properties of the impurity . to complete the comparison
, we need the results for the noise between the incoming and outgoing channels . using eq .
( [ gtgoh ] ) to relate the quasiparticle tunneling amplitude to the electron tunneling amplitude , we find that the expansion for @xmath271 of the scattering calculation becomes @xmath272 \bigg\}e^{i\omega(x_+ - x_-)}.\ ] ] this must be compared with the perturbative calculation of @xmath273 . to obtain this perturbative result , we set @xmath246 , @xmath40 and replace @xmath148 by @xmath274 in eq .
( [ srrpert ] ) . then the perturbative calculation of the noise across the junction yields @xmath275 \bigg\}e^{i\omega(x_+ - x_-)}.\ ] ] we first note that this expression for the noise contains a linear divergence in the cutoff @xmath276 .
thus this perturbative calculation is regulator dependent , which is not surprising because the tunneling operator at @xmath40 has dimension @xmath203 and should be an irrelevant operator . in spite of this , both calculations do agree to order @xmath0 ( which is all that the derivations of the duality transformation in reference @xcite would predict ) ; it is only the higher order terms in @xmath0 that disagree .
this suggests that we are on the right track with the perturbative calculation , but we just need to add in the appropriate counter terms . to see which counter term we should add , we begin by recalling that we used the most relevant tunneling operator to describe the system .
however , for @xmath40 the operators @xmath277 and @xmath278 are just as relevant as the tunneling operator , so we must consider their effects also .
in fact , @xmath279 also encourages tunneling because it tries to equalize the density of right movers and left movers .
another way to look at it is that we can not have quasiparticles tunneling between the droplets , but density fluctuations on one side may affect the other side . in appendix b
, we found that when the interaction @xmath280 is included in the lagrangian , it gives the following contribution to the noise @xmath281 \bigg\ } e^{i\omega(x_1-x_2)}. \label{srhorho}\ ] ] first , we note that the density - density coupling does not affect the noise evaluated on only one side of the impurity ( i.e. when @xmath282 . ) according to equations ( [ s11seciii ] ) to ( [ s2seciii ] ) , ( [ sttscatlg ] ) , and ( [ gtgoh ] ) , this is necessary for the scattering and the perturbative calculations to agree .
it is reasonable that the noise evaluated on only one side of the junction should be less affected by the counter terms and the regulator than the noise between probes on either side of the junction , because even though in both cases all the measurements are done far from the junction , in the second case the information must travel from one side of the junction to the other .
second , we note that when @xmath283 , equation ( [ srhorho ] ) contains the linear term in @xmath284 , which also appears in the scattering calculation , but not in the original perturbative calculation . we find that the only density - density interaction that gives the same linear term as in the scattering result for _ all _ of the cross - correlations is the one given in equation ( [ lintdef ] ) , with @xmath285 .
when we add the density - density term with this choice for @xmath284 to the original perturbative calculation , we obtain @xmath286 \nonumber \\ & + & 2i|\gamma_e|^2\omega^2{\rm sign}(\omega)\left(\pi``\delta(0 ) '' - { 1\over 3\delta}\right ) \bigg\}e^{i\omega(x_+ - x_-)}. \label{pernoise}\end{aligned}\ ] ] thus , ( except for the divergent part ) , this perturbative result agrees with the scattering result . to cancel the divergent part
, we must regulate the delta - function properly and adjust the counterterm accordingly .
then the two results will agree in the limit as @xmath287 and @xmath288 .
another approach , which may be more appropriate , is to `` smooth out '' the density - density interaction .
this is accomplished by replacing the interaction in equation ( [ lintdef ] ) by the following expression @xmath289 where @xmath290 as @xmath291 .
this new interaction does not change the finite part of eq .
( [ pernoise ] ) , and the function @xmath134 can be chosen so that the divergence cancels . as a result , even though the duality symmetry is not exactly obeyed for the cross - correlations , it is possible to add in counter terms to bring the strong - coupling limit of one picture into agreement with the weak - coupling limit of the dual picture . to summarize , to the order in @xmath106 we have calculated , the action for @xmath36 is dual to the renormalized action for @xmath40 , given by @xmath250 -\gamma_e e^{-i2\omega_0 t } \delta(x ) e^{\sqrt { 2 } \phi(t,0 ) } + h.c .
+ 4\pi^2\gamma_e \delta(x)\left(\rho_r(t,0)-\rho_l(t,0)\right)^2;\ ] ] and if we only want to calculate the noise in one particular channel , then it is not necessary to include the @xmath292 interaction to obtain the dual picture .
as explained above , this action can be interpreted as containing two different terms that induce or encourage tunneling .
we can also use the relation @xmath293 to write the action as @xmath294 -\gamma_e e^{-i2\omega_0 t } \delta(x ) e^{\sqrt { \tilde 2 } \phi(t,0 ) } + h.c . , \ ] ] where @xmath295 is the `` renormalized '' velocity . in this case
, the velocity remains the same everywhere but right at the junction .
if , instead , we use eq .
( [ lintf ] ) for the density interaction , then the velocity is renormalized in a region around the junction .
in this paper we studied the four terminal tunneling noise spectrum for chiral luttinger liquids characterized by an exponent @xmath5 .
perturbative results are obtained for arbitrary @xmath5 .
perturbative calculations for quasiparticle tunneling reveal a singularity at the quasiparticle josephson frequency @xmath296 , while perturbative calculations for electron tunneling only produce a singularity at the electron josephson frequency @xmath108 .
this appears to be inconsistent with the duality picture that quasiparticle and electron tunneling describe the same tunneling junction in two different limits . to understand how the quasiparticle tunneling picture can smoothly connect to the electron tunneling picture , we calculated the exact noise spectrum for @xmath38 ( or @xmath40 due to duality ) .
we find that the singularity at the quasiparticle josephson frequency @xmath297 is smeared for finite tunneling and is not a true singularity , while the singularity at the electron josephson frequency @xmath108 survives in the exact result . from the exact result we also find that the noise spectrum of the current correlations on a single branch ( auto - correlations ) satisfies the duality relation , while current correlations between distinct branches ( cross - correlations ) do not satisfy the naive duality relation .
* acknowledgements * we would like to thank akira furusaki for helpful discussions .
this work is supported by nsf grants dmr-9400334 ( ccc ) and dmr-9411574 ( xgw ) .
xgw acknowledges the support from the a.p .
sloan foundation .
d. f. is currently a bunting fellow sponsored by the office of naval research and is also supported in part by funds provided by the u. s. department of energy ( d.o.e . ) under cooperative agreement # de - ac02 - 76er03069 and by national science foundation grant phy9218167 .
in order to obtain the noise spectrum of density - density correlations on given leads , we start by writing the correlations between density operators as follows : @xmath298 where @xmath299 take the values @xmath300 for @xmath183 moving branches and @xmath301 for @xmath184 moving ones . such compressed notation makes it simpler to identify incoming and outgoing branches in a unified way for both left and right movers : @xmath302 , for example , is the density in an incoming or outgoing branch if @xmath303 or @xmath304 , respectively .
the densities are related to the fields @xmath41 through @xmath305 , so that we can write @xmath306 where it is convenient to use @xmath307 the last correlation function is easy to calculate perturbatively using @xmath308 where @xmath309 is the unperturbed ground state , and @xmath310 is the ordering along the keldysh contour ( fig .
the scattering operator @xmath311 takes the initial state , evolves it from @xmath312 to @xmath313 and back to @xmath312 .
the use of the keldysh contour is necessary in the treatment of non- equilibrium problems , such as the one we have in hand
. a more detailed description of the method in the context treated here can be found in ref .
@xcite . in order to proceed we expand @xmath311 to second order in perturbation theory . in terms of the coulomb gas of ref .
@xcite , we have an insertion of two charges of opposite sign : latexmath:[\ ] ] finally , the noise between the currents on either side of the impurity can be calculated similarly , so we will omit the details here .
x. g. wen , phys .
b * 41 * , 12838 ( 1990 ) .
x. g. wen , phys .
b * 44 * , 5708 ( 1991 ) . c. l. kane and matthew p. a.
fisher , phys .
68 * , 1220 ( 1992 ) ; phys .
b * 46 * , 15233 ( 1992 ) ; phys . rev
* 72 * , 724 ( 1994 ) .
akira furusaki and naoto nagaosa , phys .
b * 47 * , 3827 ( 1993 ) ; phys . rev .
b * 47 * , 4631 ( 1993 ) . c. de c. chamon and x. g. wen , phys . rev .
70 * , 2605 ( 1993 ) .
f. p. milliken , c. p. umbach and r. a. webb , preprint .
k. moon , h. yi , c. l. kane , s. m. girvin , and matthew p. a. fisher , phys . rev . lett . * 71 * , 4381 ( 1993 ) .
f. guinea , g. gomez - santos , m. sasseti , and u. ueda , cond - mat/9411130 . c. h. mak , reinhold egger , maura sasseti , and ulrich weiss , preprint .
r. egger , m. sasseti and u. weiss , cond - mat/9504040 .
p. fendley , a. w. w. ludwig , and h. saleur , phys .
* 74 * , 3005 ( 1995 ) .
p. fendley , a. w. w. ludwig , and h. saleur , cond - mat/9503172 .
a. schmid , phys .
* 51 * ( 1983 ) 1506 ; m. p. a. fisher and w. zwerger , phys .
b * 32 * ( 1985 ) 6190 ; c. g. callan and d. e. freed , nucl . phys .
b * 374 * ( 1992 ) , 543 . c. l. kane and matthew p. a. fisher , phys . rev . lett . * 72 * , 724 ( 1994 ) .
c. de c. chamon , d. e. freed , and x. g. wen , phys .
b * 51 * , 2363 ( 1995 ) .
p. fendley , a. w. w. ludwig , and h. saleur , cond - mat/9505031 .
rolf landauer , phys .
16427 ( 1993 ) and references therein . m. buttiker , phys .
12485 ( 1992 ) and references therein . g. b. lesovik , pisma zh .
. fiz . * 49 * , 513 ( 1989 ) [ jetp lett . * 49 * , 592 ( 1989 ) ] .
xiao - gang wen , intl .
j. of mod .
b * 6 * , 1711 ( 1992 ) . s .-
eric yang , solid state comm . *
81 * , 375 ( 1992 ) .
f. guinea , phys .
b * 32 * , 7518 ( 1985 ) .
k. a. matveev , phys .
b * 51 * , 1743 ( 1995 ) ; a. furusaki and k. a. matveev , cond - mat/9505035 .
r. floreanini and r. jackiw , phys .
* 59 * , 1873 ( 1987 ) .
m. b. green , j. h. schwartz and e. witten , _ superstring theory _
, volume i , cambridge university press , cambridge , 1988 .
geometries for tunneling between edge states . by adjusting the gate voltage @xmath444
one can obtain either a simply connected qh droplet ( a ) , or two disconnected qh droplets ( b ) .
for the geometry in ( a ) both electrons and quasiparticles ( carrying fractional charge ) can tunnel from one edge to the other , whereas for the tunneling geometry in ( b ) only electrons can tunnel .
the luttinger liquid behavior is characterized by the exponent @xmath7 in ( a ) , and @xmath8 in ( b ) .
the tunneling current @xmath26 depends on the applied voltage between the right and left edges , and by increasing this voltage one can also cross over from the geometry ( b ) to the geometry ( a ) .
four terminal geometry for the measurement of tunneling between edge states .
the terminals 1 and 2 correspond to branches that are incoming to the scatterer , while terminals 3 and 4 correspond to outgoing ones .
the arrows indicate the direction of propagation for a given branch .
the incoming branches are in equilibrium with their reservoirs of origin , while the outgoing ones do get effected by the scatterer . by measuring fluctuations in the voltages / currents at the four terminals ( @xmath445 , @xmath69 ) , the auto - correlation spectra @xmath446 , with @xmath79 , and the cross - correlation spectra @xmath446 , with @xmath447 , can be obtained .
these voltage / current fluctuations contain information on the fluctuations of the tunneling current .
plots of the excess noise of outgoing branches ( probes 3 and 4 of fig .
2 ) calculated to second order in perturbation theory ( equation ( [ noise2ndorder ] ) ) .
the excess noise is normalized to the zero- frequency shot noise level , and the frequency @xmath0 to the josephson frequency @xmath90 ( @xmath448 _ vs. _ @xmath449 ) .
different singularities are obtained at @xmath450 for different values of @xmath5 : @xmath451 , @xmath452 , @xmath453 , @xmath454 , and @xmath203 .
one should keep in mind that , although the singularities all occur at @xmath450 , the value of @xmath90 depends on the charge of the current carrier , which in turn also depends on @xmath5 . figure 4 . a particle ( plane wave ) incoming from the left ( @xmath455 ) with energy @xmath0 scatters off the impurity at @xmath456 into a superposition of a particle at energy @xmath0 and a hole at energy @xmath159 on the right side of the impurity ( @xmath457 ) . in the case where the incoming state is a filled fermi sea up to the energy @xmath148 ,
the scattered state on the right side of the impurity will be completely filled up to energy @xmath458 , and partially filled between @xmath458 and @xmath148 .
it is this partially filled energy range from @xmath458 and @xmath148 which is responsible for the non - equilibrium properties of the system .
the tunneling processes @xmath459 ( a ) and @xmath414 ( b ) . in the process
@xmath459 , both at time 0 and @xmath196 , a quasiparticle tunnels from the left branch to the right branch , and another quasiparticle tunnels in the opposite direction . in the @xmath414 process , at time 0 two quasiparticles tunnel from , say , the left to the right branch , and at time @xmath196 the two quasiparticles tunnel back in the opposite direction .
the process @xmath414 is responsible for the singularity at the electron frequency @xmath460 .
plots for the renormalized noise @xmath461 _ vs. _ @xmath222 .
@xmath461 , @xmath462 and @xmath463 are the renormalized noise and frequencies , using the coupling constant as the scaling factor ( @xmath464 , @xmath218 and @xmath465 ) . in ( a )
the excess noise @xmath466 is plotted for large values of @xmath224 , which illustrate the weak coupling ( @xmath206 ) limit .
the rescaled excess noise @xmath467 is plotted in ( b ) .
it shows the strong coupling limit ( @xmath209 ) as @xmath468 .
the full noise @xmath469 is plotted in figure ( c ) .
for the larger values of @xmath148 , notice that the singularity at @xmath470 is hidden in the full noise .
meanwhile , some reminiscent signs of the quasiparticle singularity appear near @xmath471 .
the association of the four densities @xmath68 ( @xmath69 ) to the left and right moving branches for the dual pictures corresponding to ( a ) @xmath7 and ( b ) @xmath8 ( compare to figs .
1a and 1b ) .
notice that @xmath254 and @xmath255 change chirality under duality , and that the space coordinates ( the @xmath472 and @xmath473 axis ) should also be redefined under the duality transformation .
an insertion of an operator @xmath474 corresponds to the insertion of a charge @xmath475 on the contour at time @xmath196 .
similarly , an insertion of an operator @xmath476 corresponds to an insertion of a charge @xmath477 at time @xmath196 .
the time @xmath196 is ordered along the contour shown , and there is a distinction between charges placed on the top and bottom branches . in the illustration , we consider the particular case when the @xmath477 charge is inserted on the top contour , and the @xmath475 charge is inserted on the bottom contour . | we study non - equilibrium noise in chiral luttinger liquids using the landauer - buttiker scattering approach , obtaining the current / voltage noise spectrum for a four - terminal measurement scheme .
experimental consequences of the tunneling of charges are present in the four - terminal measurement of both the low - frequency shot noise ( @xmath0 near 0 ) , and the high - frequency josephson noise ( @xmath0 near @xmath1 ) . within perturbation theory ,
an algebraic singularity is present ( to all orders ) at the josephson frequency @xmath1 , whose position depends on the charge @xmath2 of the tunneling particles , either electrons or fractionally charged quasiparticles .
we show in a non - perturbative calculation for an exactly solvable point that the singularity at the quasiparticle frequency exists only in the limit of vanishing coupling , whereas the singularity at the electron frequency is present for all coupling strengths .
the vanishing coupling limit corresponds to perfectly quantized hall conductance in the case of quasiparticle tunneling between edge states in the fractional quantum hall regime , and thus tunneling destroys the singularity at the quasiparticle frequency concomitantly with the quantized current . |
the theory of brownian motion and random walks represents a cornerstone of statistical mechanics for providing an underlying physical picture and a microscopic explanation of many diffusion processes .
the most famous examples are probably those provided by the theory of brownian motion of einstein @xcite and smoluchowski @xcite . since then the application range of random walk theory has widened in many different fields , extending to e.g. biology , ecology , and anthropology . the standard drift - diffusion equation , @xmath1,\ ] ] where @xmath2 is the particle position vector , defines the dynamics of the brownian particle through the drift velocity field @xmath3 and the diffusion coefficient @xmath4 .
in principle @xmath3 and @xmath4 can be derived from the knowledge of the microscopic dynamics of the underlying random motion . in the well known problem of the overdamped brownian motion the velocity field turns out to be given by @xmath5 , where @xmath6 is the friction coefficient , @xmath7 the external force acting on the brownian particle , and @xmath8 the corresponding external potential .
the diffusion coefficient @xmath4 and the friction coefficient @xmath6 are related to each other and to the temperature @xmath9 of the environment through the fluctuation - dissipation theorem , @xmath10 @xcite .
besides the mentioned works by einstein @xcite and smoluchowski @xcite various derivations of the drift - diffusion equation have been proposed , starting from a microscopic picture of the corresponding random walk : the interpretation of the parameters in the resulting drift - diffusion equation depends on the specific problem under study see also van kampen s derivation of the diffusion equation from a general discrete one - step processes @xcite .
in fact , since their very beginning diffusion problems have been appearing in very different disciplines , as exemplified by the first random walk model introduced in the social sciences by louis bachelier in 1900 with the goal of modeling asset price dynamics @xcite . in many diffusion problems the underlying dynamics of the random walkers is intrinsically different from that of a brownian particle . in place of an actual drift force , one often finds an `` attractiveness function '' @xmath11 that modulates the otherwise isotropic and homogeneous probability that a random walker jumps at time @xmath12 from its current position to the generic position @xmath13 .
an example is chemotactic diffusion where @xmath11 represents the concentration of some substance affecting the drift motion of the random walker , that can be e.g. a cell , a bacterium or an ant .
even if the attractiveness function @xmath11 is not per se a potential , it is shown below that its effect is to generate a term analogous to an external force in the drift - diffusion equation .
diffusion processes taking places in such situations , in which the drift motion can be traced back to some attracting factor rather than to a ( known ) force , represent the subject of the present paper .
the basic difference is that in the latter case the velocity field , e.g. the potential gradient in the case of brownian motion , provides by definition the actual flux of particles or the average velocity of the brownian particle .
instead , the attractiveness function @xmath11 embodies the effects of some external factor on the preferences of the random walker and therefore _ vanishes in the absence of diffusion_. in the type of problems considered here , the random walker _ compares _ the ( known ) values of some attractiveness measure @xmath11 in the accessible positions before making the next jump , choosing the jump direction in a probabilistic way depending on the result of the comparison made .
the goal of the present paper is to study the general drift - diffusion equation for such type of systems directly in terms of the attractiveness @xmath11 . to this aim
we proceed through two steps : we start from a general random walk in a @xmath0-dimensional space where jumps are defined by a jump probability @xmath14 from position @xmath15 at time @xmath16 to position @xmath17 at time @xmath12 .
the mathematical preliminaries are laid out in sec .
. then we introduce an attractiveness function @xmath18 , assuming that the probability that the jump takes place along a certain direction is proportional to the local value of the attractiveness function .
the corresponding drift - diffusion equation is derived in sec .
[ attraction ] .
it is shown that to this aim one can simply replace the conditional jump probability @xmath14 with the corresponding _ conditional expectation _
@xmath19 obtained by weighting @xmath14 with the attracting function @xmath18 see sec .
[ derivation ] .
the probabilistic meaning of the attracting function @xmath18 is further discussed in sec .
[ stat ] .
the motivation for the present investigation is that the theoretical framework outlined above and the corresponding diffusion equation obtained well suit the description of many diffusion processes in which the appearance of a drift motion can be traced back to a non - constant function @xmath18 measuring a local preference or attractiveness .
the equation obtained is in fact simple , yet general enough , to be useful for applications to diffusion processes in physics and in interdisciplinary fields .
some applications to chemotaxis and social dynamics are considered in sec .
[ applications ] .
possible developments and further applications are discussed in section [ conclusions ] .
it is possible to give a general derivation of the drift - diffusion equation from a random walk model starting from an arbitrary form of the joint step - length and waiting - time distribution of the random walker @xcite .
here we adopt a similar procedure by first considering a generic jumping probability @xmath20 that describes the motion of the random walker in the absence of the attracting factor .
thereafter the effect of the attracting factor on this random motion is studied and the corresponding drift - diffusion equation is derived with the kramers - moyal procedure . here and in the following we consider the general case of a @xmath0-dimensional space , in which each point @xmath17 is specified in terms of its @xmath0 coordinates , @xmath21 . if @xmath22 is the probability that a random walker at location @xmath15 at time @xmath12 hops to location @xmath17 where it will be located at time @xmath23 , the random walker density @xmath24 at @xmath23 can be obtained from the density @xmath25 at time @xmath12 as @xmath26 the normalization condition @xmath27 is assumed to hold for any @xmath15 and @xmath28 . to have a simple well behaved continuum limit , where the dynamics can be approximated by a diffusion equation
, we impose the condition of _ locality _ : there exists a characteristic length scale @xmath29 , so that jumps with distances exceeding this scale are very improbable .
formally we require for the probability of long jumps to be at least exponentially suppressed : @xmath30 time arguments will be omitted in the following , i.e. @xmath31 .
it is shown in appendix [ app : km_bound ] , that this form of locality defined by eq .
( [ eq : locality_cond ] ) is sufficient for the kramers - moyal coefficients @xmath32 to be at least of order @xmath33 . from the kramers - moyal procedure
we obtain that the behavior of the random walkers at large scales is described by a drift - diffusion like equation of the form ( see app . [ app : kme_derivation ] ) @xmath34 + \mathcal{o}\left(\epsilon^{3}/\tau\right).\end{aligned}\ ] ] thus in the continuum limit , i.e. when @xmath35 with @xmath36 fixed , the higher order terms vanish and the diffusion equation is obtained in its canonical form . above @xmath37 , @xmath38 and summation over repeated indices is assumed .
@xmath39 , @xmath40 are the drift and the diffusion terms corresponding to @xmath41 : @xmath42 we stress that the characteristic length @xmath29 of the steps and a fixed time interval @xmath43 between the steps are the only restrictions imposed the motion of the random walker .
thus our set - up is fairly general as , at the macroscopic level , it can be applied to any drift - diffusion equation . in the following
we assume the existence of an external inhomogeneous effect , due to the local landscape attractiveness , on top of the random motion assumed above and study its effects on the drift and diffusion of the brownian particle .
the attracting factor is represented by the function @xmath18 , that is assumed to modulate the diffusion so that the transition probability becomes proportional to it , i.e. , @xmath44 since @xmath18 is not a probability but only modulates it , it has to be normalized in order to obtain the jump probability : the normalization conditions require that the driftless hopping probability @xmath45 is replaced by the conditional expectation value @xcite weighted by the attracting function @xmath18 , @xmath46 this new hopping probability @xmath47 still obeys the normalization condition of eq .
( [ pnorm ] ) and reduces to the unperturbed form @xmath41 when @xmath48 is a constant .
we stress that the normalization factor in the denominator of eq .
( [ p1 ] ) is a @xmath48-weighted sum of the hopping probabilities @xmath49 over all the possible _ arrival _ positions @xmath50 , starting from the same position @xmath15 , see fig .
( [ fig_2d ] ) . in a typical situation
the quantity @xmath18 can be directly related to some observable quantities , for example to pheromone density in the case of chemotactic diffusion discussed below .
a drift term is now induced by the attracting factor @xmath51 since according to eq .
( [ p1 ] ) @xmath52 . following the kramers - moyal procedure
( see app .
[ app : kme_derivation ] ) , the expansion of @xmath48 around @xmath17 yields @xmath53 thus , in the hypothesis that @xmath54 and @xmath18 vary slowly in space ( at scales comparable to @xmath29 ) and neglecting higher order terms in @xmath29 ( taking the continuum limit ) , the kramers - moyal procedure yields an additional drift term @xmath55 corresponding to the effective modulated transition probability , while negligibly the modifying the diffusion term . in conclusion ,
the following drift - diffusion equation is obtained : @xmath56 .\end{aligned}\ ] ] the drift - diffusion equation can also be recast as @xmath57 .\end{aligned}\ ] ] it follows , that the absence of the background drift term , i.e. when @xmath58 , the equilibrium distribution is given by @xmath59 if we assume a time independent @xmath48 . however , in many practical applications @xmath48 itself can be a dynamical quantity and thus the assumption of time independence might not hold at all or it could hold only approximately , e.g. when @xmath48 changes slowly enough for the density @xmath60 to remain close to the equilibrium . finally we note that the drift term in is proportional to the the diffusion tensor @xmath61 , thus rendering drift caused by an attracting factor @xmath48 conceptually different from a drift attributed to an external force . especially , in the absence of diffusion , i.e. when @xmath62 , the drift will vanish too in this respect see also section [ stat ] .
scheme of a diffusion process with constant step @xmath63 in ( a ) one and
( b ) two dimensions ( rayleigh problem ) . before jumping from the initial position @xmath15 to the arrival position @xmath17 , the random walker probes the environment , i.e. the values of the function @xmath51 in all the neighboring locations @xmath50 at a distance @xmath63 from @xmath15 .
the resulting probability @xmath47 to jump from @xmath15 to @xmath17 is the conditional expectation ( [ p1 ] ) .
, title="fig:",width=264 ] scheme of a diffusion process with constant step @xmath63 in ( a ) one and ( b ) two dimensions ( rayleigh problem ) . before jumping from the initial position @xmath15 to the arrival position @xmath17 , the random walker probes the environment , i.e. the values of the function @xmath51 in all the neighboring locations @xmath50 at a distance @xmath63 from @xmath15 .
the resulting probability @xmath47 to jump from @xmath15 to @xmath17 is the conditional expectation ( [ p1 ] ) . , title="fig:",width=264 ] for illustrative purposes let us consider the common example , where , in the absence of an attracting factor , the random walker would follow a homogeneous and isotropic transition probability of the form @xmath64 in this case the motion of the random walker consists of consecutive steps of constant length @xmath65 made at regular time intervals @xmath43 but with no preferred direction
. in two dimensions this coincides with the rayleigh problem that is illustrated in fig . [ fig_2d ] .
the diffusion constant related to ( [ eq : free_dif_p ] ) reads @xmath66 where @xmath0 denotes the dimension of the underlying space . in all , in the absence of an attracting factor , the random walker is described by the most common form of the diffusion equation , @xmath67 . in the presence of an attracting factor
, the random walker can choose to jump to any point on the circle in fig .
[ fig_2d ] , but before making a choice he probes the environment and adjusts his choice so that locations with an higher value of the attracting factor are preferred as a result drift is created .
however , the motion of the random walker is still driven by diffusion that remains unmodified by the attracting factor .
the resulting drift - diffusion equation can be obtained from eq . .
it reads @xmath68 this equation is formally equivalent to the drift - diffusion equation ( [ eq-0 ] ) with the velocity field given by @xmath69 , corresponding to an effective external potential @xmath70 the last relation could also be obtained from the equilibrium conditions where @xmath71 , but on the other hand @xmath72 . at a general level , eq .
( [ p1 ] ) describes a modulated jump probability that can have different origins . to have a concrete example in mind we can think of some ecological dispersal problem in which an individual checks e.g. the fitness of the local landscape before taking the next move
. in other problems of chemotactic diffusion , the cell or a bacterium that represents the random walker must be able somehow to monitor the surrounding environment in order to get information about the concentration of some chemical substance .
in all these cases there is a common relevant physical principle underlying the various interpretations of the conditional probability ( [ p1 ] ) : namely , the elementary jump process described by eq .
( [ p1 ] ) can be given a probabilistic interpretation in which a random walker first _ compares _
some property of the neighboring zones and thereafter chooses the direction of the next step ( randomly but with probabilities based on the comparison made ) .
the operative meaning of the conditional expectation ( [ p1 ] ) has been studied by othmer and stevens @xcite .
their comparison of various 1d and 2d lattice models of chemotactic diffusion clearly shows that the normalization implied by the conditional expectation ( [ p1 ] ) is the key - element for obtaining a weber - fechner - like logarithmic dependence of the chemotactic potential analogous to the potential @xmath73 of brownian motion on the relevant chemical concentration , corresponding here to the attractiveness function @xmath51 .
notice that even in those cases in which the random walker performs a comparison based on the ratio @xmath74 , expressing the attractiveness of the final position @xmath17 _ relative _ to that of the starting location @xmath15 rather than on the arrival values @xmath51 alone one still obtains the same expression as in eq . : @xmath75 due to the cancellation of the @xmath76 factors . in the typical situations considered in this paper the drift - diffusion equation , eq .
( [ dd1 ] ) , is obtained from the microscopic random walk dynamics with a given attractiveness function @xmath51 . before proceeding ,
it is useful to make a simple consistency check of the results obtained thus far by considering the complementary situation , i.e. , starting from the drift - diffusion equation ( [ eq-0 ] ) for an overdamped brownian particle and asking what is the form of the corresponding attractiveness function @xmath51 .
comparison with eq .
( [ dd1 ] ) shows that the external potential felt by the random walker is proportional to the logarithm of the attractiveness , i.e. , @xmath77 , where @xmath78 is some constant .
the definition of the attracting function @xmath48 implies that the transition probability from a starting point @xmath15 to an arrival point @xmath17 is proportional to @xmath79/\alpha \}$ ] , which coincides with the boltzmann - gibbs factor ( with @xmath80 representing temperature ) , used e.g. in the monte carlo method @xcite .
the same procedure can be useful when one starts from an assigned drift - diffusion equation and searches a corresponding microscopic random walk model defined in terms of an attracting function @xmath51 .
an example of non - isotropic diffusion is that driven by chemotaxis , a mechanism determining many types of biological motion based on the concentration of a certain chemical .
the motion of units is usually biased toward higher values of the concentration of a certain chemical , referred to as pheromone .
many instances of this kind of motion are known in different forms of life , with units represented by animals , insects , bacteria , and cells and the pheromone by the substance they or other sources produce ; for example , leukocyte cells in the blood move towards a bacterial inflammation following the positive gradient of some chemical caused by the infection ; slime mold _ dictyostelium discoideum _ , a single - cell amoebae , moves towards regions of high concentrations of the cyclic - amp produced by the amoebae themselves @xcite . a standard and simple model of chemotactic dynamics
was put forward by keller and segel @xcite , @xmath81 - \nabla [ n \chi(a ) \nabla a ] , \\
\label{ks2 } \frac{da}{dt } & = & g(a , n ) + \nabla \cdot [ d_a \nabla a ] \ , .\end{aligned}\ ] ] here @xmath54 is the density of diffusing biological units and @xmath82 is a dimensionless quantity proportional to the concentration of pheromone .
the function @xmath83 in eq .
( [ ks1 ] ) defines the type of population dynamics characterizing the units and @xmath84 in eq .
( [ ks2 ] ) the production of pheromone possibly due to the units themselves , as in the mentioned example of the _ dictyostelium discoideum _ , as well as its decay .
the second terms on the _ rhs _ of eqs .
( [ ks1 ] ) and ( [ ks2 ] ) describe the diffusion process undergone by the units and the pheromone , with diffusion coefficients @xmath4 and @xmath85 respectively . finally , one has to assign the form of the last term on the _ rhs _ of eq .
( [ ks1 ] ) responsible for the drift motion of the units due to the chemotactic substance , i.e. @xmath86 , usually defined as the `` chemotactic coefficient '' or `` chemotactic sensitivity '' . interpreting the drift term as that of a brownian particle acted upon by the effective force @xmath87
, one can set @xmath88 and rewrite the effective force simply as @xmath89 then eq . ( [ ks1 ] ) becomes the drift - diffusion equation of an overdamped brownian particle in an external potential @xmath90 . following this analogy
we will refer to the quantity @xmath91 as the `` chemotactic potential '' . by comparison between eqs .
( [ ks1 ] ) and ( [ dd1 ] ) and using eq .
( [ xa ] ) one can see that the attractiveness function is in this case given by @xmath92 where @xmath93 is some constant . according to this picture ,
units diffuse in space with a diffusion probability modulation represented by eq .
( [ k - x ] ) . in their original paper ,
keller and segel showed that in order for eqs.([ks1 ] ) and ( [ ks2 ] ) to reproduce the waves observed experimentally in microbe diffusion processes the chemotactic coefficient must have a singularity of order one or larger , that is , if one chooses a simple power law form @xmath94 , where @xmath95 is a constant and @xmath96 .
keller and segel chose @xmath97 and the least singular of all the possible powers , @xmath98 , corresponding to a logarithmic chemotactic potential , @xmath99 in this case eq .
( [ k - x ] ) shows that the attractiveness function is simply a power of the pheromone concentration .
in fact such a dependence is found e.g. in e.coli at least in a relevant part of the parameter ranges @xcite .
different laws and forms of the drift - diffusion equation can be obtained by choosing correspondingly different forms of the attractiveness function . in some studies
chemotactic potentials different from the one with a minimal exponent value @xmath100 @xcite , corresponding to the weber - fechner ( or log ) law , have been proposed .
the value @xmath101 can be justified on the base of experimental evidence @xcite moreover the applicability of the weber - fechner law seems to depend on the parameter range of the system under study .
the case @xmath102 corresponds to a chemotactic coefficient and potential @xmath103 referred to as the receptor law .
besides the decrease of the effect of pheromone concentration at very low concentrations @xmath104 , this form of @xmath91 avoids the infinite response that would be otherwise obtained in the case of a logarithmic chemotactic potential when @xmath98 for @xmath105 , since , on the other hand , here @xmath106 . the chemotactic potential ( [ b2 ] ) is associated to the attractiveness function @xmath107 \}$ ] that consistently exhibits the ineffectiveness of pheromone at very low concentrations as well as the saturation effect at high concentrations .
the function @xmath108 may also represent the relative difficulty or easiness in following a certain path in a 2d diffusion process across a surface .
diffusion of ions , small and large molecules on the surface of biomolecules provide relevant examples of such inhomogeneous diffusion . in this case
the kramers - moyal coefficients @xmath61 are directly related to the entropic forces due to the boundaries .
other examples come from the studies of dispersal of animals and humans across a geographical landscape .
there are many factors related to the physical geography which could be well described in terms of a modulation of diffusivity . in all these cases there
is usually no net systematic drift due to some sort of forces pushing individuals along certain directions , i.e. , it is difficult to formulate an analogue of the external potential acting on a brownian particle . on the other hand ,
it is natural to introduce ( and possibly measure experimentally ) the inhomogeneous probability that a certain path rather than another one is followed and express such a probability in terms of a local attracting factor .
for instance , paths that follow sea and river shores are known to represent corridors characterized by a higher mobility , while the same shores or mountains interrupting a path are locally associated to lower effective diffusivities @xcite . in the dispersal of human or animal groups ,
diffusion may become non - linear , since the presence of other individuals will condition the diffusion process of each single individual deciding in which direction to make the next diffusive step .
for instance , in the well known example of bacteria diffusion , a non - linear diffusion equation @xcite of the form @xmath109 \ , , \ ] ] describes an inhomogeneous diffusion depending on the density @xmath54 itself in which individuals escape regions of higher density faster , if @xmath110 is a growing function of the density @xmath60 .
the diffusion equation can be put into the form of eq . , @xmath111 with @xmath112 a standard form of the density - dependent diffusion coefficient is given by @xmath113 , where @xmath114 , @xmath115 , and @xmath116 are suitable positive parameters @xcite .
interestingly , in this case the corresponding attracting factor , @xmath117 resembles a stretched boltzmann - gibbs shape and reduces to the canonical form for @xmath118 .
it is worth mentioning other types of diffusion processes taking place across geographical landscapes , in which the probability that the diffusion process develops along a certain direction can be best related to some local factor rather than to an external force producing a drift . in this set of problems one can find e.g. processes of epidemic diffusion , opinion dynamics , and ( technological ) innovation spreading .
social dynamics of opinion spreading offers a good example .
many - agents social dynamics models or continuous opinion spreading models such as the model by borghesi and bouchaud , introduced to study political elections @xcite , can provide a detailed description of the system state evolution and the corresponding modulated diffusion process .
the example of the spreading of agriculture during the neolithic age is an interesting example of technological innovation spreading , since it took place across a wide area but under the influence of various factors , such as the same underlying geography affecting human dispersal or the geophysical properties of the land related to its suitability for agricultural purposes see ref .
@xcite for a recent investigation .
studies dealing specifically with technological spreading have to take into account the landscape inhomogeneities .
for instance hgerstrand s model of innovation diffusion @xcite describes the level of adoption of a new innovation in terms of a mean information field evolving in space and time according to a reaction - diffusion - like equation .
there are various factors making such a process inhomogeneous , possibly affecting directly the innovation diffusion process , see e.g. refs .
in this paper we have proposed a framework providing a simple description of a diffusion process influenced by some local factor represented by an assigned attracting function @xmath51 .
this framework may be most suitable to study problems of random motion in which the random walkers are able to compare the possible directions and to use the gathered information to make a ( probabilistic ) decision about the direction of the next step .
the information about the positions is encoded in the attracting function @xmath51 so that locations with a higher value of the attracting function are chosen with an higher probability .
we have discussed the example of chemotactic diffusion , also mentioning how the framework proposed easily lends itself to the study of other ecological and dispersal models where the influence of the spatial landscape can be described by a local fitness @xmath18 .
for instance , models of dispersal and competition between species , see ref .
@xcite , or related models where the local density of resource plays a key role @xcite , are perfect candidates for this type of study .
also the mathematical modeling of human dispersal , usually done in terms of reaction - diffusion equations ( for a recent study see ref .
@xcite ) , rarely takes into account the heterogeneity of the local landscape . in these and similar situations
the drift - diffusion equation framework proposed here may provide a suitable tool for carrying out a more realistic modeling of the actual diffusion process .
this work was supported by the institutional research funding iut ( iut39 - 1 and iut23 - 6 ) of the estonian ministry of education and research .
the fourier transformation of the conditional probability gives the moment generating function ( time arguments are omitted ) : @xmath119 where @xmath120 is the @xmath121-th moment for jumping from @xmath17 to @xmath15 .
this corresponds to the following expansion @xmath122 plugging this into the master equation @xmath123 gives the ( formally exact ) expansion @xmath124 where @xmath125 in @xmath0-dimensions : @xmath126 where @xmath127
we prove that the requirement of locality ( [ eq : locality_cond ] ) @xmath128 if @xmath129 guarantees that @xmath32 is at least of order @xmath33 : @xmath130 where @xmath131 is a constant .
this is sufficient to guarantee that the higher terms in @xmath29 vanish in the continuum limit , i.e. when @xmath35 with @xmath36 fixed . | we study a simple model of a random walker in @xmath0 dimensions moving in the presence of a local heterogeneous attracting factor expressed in terms of an assigned space - dependent `` attractiveness function '' , a situation frequently encountered in the study of various diffusion problems . the corresponding drift - diffusion equation and the explicit expressions for the velocity field and the diffusion coefficient are obtained and discussed .
we consider some examples of applications of the results obtained to chemotactic diffusion processes and social dynamics . |
the cdm paradigm , within which the majority of matter in the universe is weakly interacting , has enjoyed great success in explaining astrophysical and cosmological data .
the inclusion of cdm in the now standard @xmath10cdm cosmological model makes clear predictions for , among other things , the mass function of gravitationally collapsed structures ( e.g. @xcite ) , the ratio of baryonic mass to cdm in galaxy clusters ( e.g. @xcite ) , and the distribution of mass within the structures that form hierarchically in the universe ( e.g. @xcite ) .
observations testing the predicted mass function and cluster baryon fraction have largely validated cdm ( e.g. @xcite ) .
observational tests of the distribution of mass within bound structures is in some ways more challenging , requiring spatially resolved measurements of the gravitational potential of objects whose mass is predominately dark
. nevertheless , an extensive body of work now exists , using primarily x - ray and gravitational lensing observations of galaxy clusters @xcite . in this work
, we revisit the subject using _ chandra _ x - ray observations of a sample of massive , highly dynamically relaxed clusters .
although not representative of the cluster population at large , these systems are an ideal laboratory for deriving three - dimensional mass profiles based on observations of the intracluster medium ( icm ) because departures from hydrostatic equilibrium and systematics due to projection are minimized . both simulations and direct calibration using weak gravitational lensing indicate that the overall bias in _ chandra _ x - ray mass estimates for this sample is small ( @xmath11 per cent ; @xcite ) .
the specific features of the cluster mass distribution that we consider here are ( 1 ) the concentration parameter of cluster mass profiles , and its dependence on mass and redshift ; and ( 2 ) the shape of mass profiles , in particular departures from the baseline model defined by @xcite ( hereafter nfw ) .
the selection of the sample of massive , relaxed clusters employed here is detailed in the first paper in this series ( @xcite ; hereafter paper ) .
papers , and @xcite employ _ chandra _ data for this sample to respectively constrain cosmological parameters , through gas - mass fraction measurements ; scaling relations and thermodynamics of the icm ; and the average bias of the x - ray mass determinations .
section [ sec : data ] reviews our procedure for constraining cluster mass profiles from x - ray data , for which complete details are available in papers and , and introduces the specific mass models employed in this work . in section [ sec : results ] , we review the specific predictions of cdm for cluster mass profile concentrations and shapes , present our results , and compare them to others in the literature .
section [ sec : conclusions ] summarizes our conclusions .
we assume a concordance flat @xmath10cdm cosmology with @xmath12 and @xmath13 throughout . unless otherwise noted , quoted parameter uncertainties correspond to the maximum - likelihood ( i.e. shortest ) interval enclosing 68.3 per cent posterior probability , and best - fitting values are the posterior modes .
this work employs the sample of 40 massive , dynamically relaxed galaxy clusters identified in papers and ( table [ tab : thetable ] ) . for our purposes ,
dynamical relaxation was defined quantitatively in terms of x - ray image features , specifically the sharpness of the surface brightness peak and the alignment and symmetry of a series of standardized cluster isophotes ; in addition all clusters in the sample must have an emission - weighted temperature of @xmath14kev outside of their cores .
the procedure for cleaning and fitting the _ chandra _ data for these clusters is described in detail in paper ; in paper the data reduction was updated to use a newer version of the _ chandra _ calibration files ( specifically caldb 4.6.2 ) , and we use that version of the data reduction here .
a complete list of the specific obsids included in the analysis can also be found in that work .
note that , while the data for each of these 40 clusters can constrain the 2-parameter nfw mass model , the 3-parameter models discussed below can be constrained for only a subset of entire sample ( see sections [ sec : gnfw ] and [ sec : einasto ] ) . [ cols="<,^,>,>,>,<,>,<,>,<,^,^ " , ] as in our previous papers , the spectral analysis of the x - ray data here involves a deprojection , assuming spherical symmetry and hydrostatic equilibrium , in order to infer the 3-dimensional properties of the hot gas and gravitating matter in the clusters .
we consider three parametric models for the total mass profiles of clusters : the standard nfw profile , whose results were previously presented in papers and ; a generalized nfw ( gnfw ) profile , in which the power - law slope of the density profile at small radii is a free parameter ; and the einasto profile @xcite , which provides a more accurate description of simulated halos than the other two options , particularly at small radii @xcite . up to normalizing factors ,
the density profiles associated with these models are @xmath15 \right\ } , \nonumber\end{aligned}\ ] ] where @xmath16 and @xmath17 are scale radii .
figure [ fig : demo ] shows how the shapes of the gnfw and einasto models differ from the nfw profile .
note that , while the nfw model is a special case of the gnfw profile with @xmath18 , no value of @xmath19 makes the einasto profile completely equivalent to nfw , although values of @xmath20@xmath21 can closely approximate it over the radial range probed by our data ( roughly @xmath22 , in terms of the nfw scale radius ) .
the spectral analysis itself uses xspec , within which we have modified the clmass model @xcite to include the gnfw and einasto mass profiles in addition to nfw .
this model treats the cluster icm as a series of concentric , isothermal spherical shells of gas , in hydrostatic equilibrium with the gravitational potential given by one of the above parametric mass profiles .
we simultaneously model absorption from galactic hydrogen and , where necessary , the presence of foreground emission from the galactic halo and/or the local bubble .
complete details of these aspects of the analysis can be found in paper . for every cluster ,
we identify and exclude from the analysis a central region , typically 50kpc in radius , where visible features in the icm indicate that the assumptions of spherical symmetry and hydrostatic equilibrium are not met .
this limits the radial range over which we can draw conclusions about the shape of the mass profiles , in particular excluding the central galaxies of each cluster , where the contribution of cold gas and stars to the total mass may be significant ( e.g. @xcite ) .
the outer radius used in our analysis is set by the need to measure temperature profiles without background modelling dominating the uncertainties ( see paper ) .
the specific radial range for each cluster over which the mass profiles are constrained is listed in table [ tab : thetable ] . for each cluster and each mass model considered ,
our spectral analysis produces a markov chain of the parameters of interest , for which the density of samples is proportional to the joint posterior density of the parameters . rather than summarize these results by , e.g. ,
a mean and standard deviation , we continue to use the full set of samples when constraining properties of the ensemble of clusters ( e.g. the concentration - mass relation ) in section [ sec : results ] . in detail ,
the samples for each cluster are importance weighted and then directly numerically integrated over to provide a likelihood for the hierarchical model being fitted .
this is especially important for constraining the distributions of the gnfw and einasto shape parameters , as the posterior distributions of these parameters for individual clusters are frequently poorly constrained and highly asymmetric ( table [ tab : thetable ] ) .
we additionally calculate central cooling luminosities ( @xmath23 ) for each cluster in the sample .
these we define as the bolometric luminosity emitted within the radius where the cooling time equals 5gyr .
the cooling time is calculated as @xmath24 where @xmath25 is the icm temperature ; @xmath26 , @xmath27 and @xmath28 are the number densities of electrons , protons and all particles ; @xmath29 is the metallicity ; and @xmath30 is the cooling function .
these calculations are based on spectral analyses of each cluster that do not assume equilibrium or a particular underlying mass profile , using instead the projct model in xspec , as presented in paper . like total luminosity
, @xmath23 displays an overall trend with mass ; fitting a power - law plus log - normal intrinsic scatter model , we find @xmath31 with a log - normal scatter of @xmath32 .
this section presents our results and discusses them in the context of cdm predictions and other measurements in the literature . in section [ sec : cm ] , we assume the nfw mass model and investigate the mass and redshift dependence of cluster concentrations .
sections [ sec : gnfw ] and [ sec : einasto ] respectively consider the gnfw and einasto generalizations to the nfw profile . the concentration of mass within an nfw halo can be described by the parameter @xmath33 where @xmath34 is the radius within which the mean enclosed density is 200 times the critical density of the universe at a cluster s redshift .
numerous dark - matter - only simulations have shown that the concentration of halos is expected to vary with both redshift and mass on average , and to have an intrinsic scatter of @xmath35 per cent driven primarily by differences in accretion history @xcite .
early work suggested a scaling relation similar to @xmath36 @xcite .
more recent work revealed that the dependence on both mass and redshift is reduced for the most massive ( i.e. , cluster - scale ) halos ( e.g. @xcite ) , largely because these halos are preferentially recently formed ( e.g. @xcite ) .
@xcite have shown that selecting the most dynamically relaxed halos from simulations recovers a power - law with mass out to the largest masses .
our current understanding , at least on the basis of n - body simulations , is that concentrations for individual clusters grow steadily as they accrete through minor mergers , but that sufficiently disruptive mergers can `` reset '' the concentration parameter to lower values ( @xmath37 ) , with the overall trends largely reflecting differences in the typical accretion histories of halos as a function of mass and redshift @xcite . by construction
, our sample contains only the most dynamically relaxed , massive clusters at any redshift .
we therefore expect their accretion histories at the time of observation to be similar , and in particular that the time since the last major merger is significant in dynamical terms . as a consequence
, we do not necessarily expect the redshift dependence found in simulations for all halos to apply to our sample , since it must partially reflect evolution in the rate of major mergers that will not be present in our sample .
it therefore makes sense to consider models where both the dependence on mass and redshift are free parameters , @xmath38 our model also includes as a free parameter a log - normal intrinsic scatter in concentration , @xmath39 , which is assumed to have the same value at all masses and redshifts .
the concentrations and masses of our clusters , derived from fitting the nfw profile , are shown in the left panel of figure [ fig : cm ] .
note that the measurement correlation between @xmath40 and @xmath41 is typically strong , as illustrated by a joint constraint ellipse for the case of rxj1347.5@xmath421145 , and that this correlation is accounted for by our fitting method ( section [ sec : data ] ) .
fitting the model of equation [ eq : cm ] , we find @xmath43 and @xmath44 .
a dependence on mass , consistent with the expectation from simulations , is detected at @xmath45 significance , while the redshift dependence is consistent with zero .
we find an intrinsic scatter of @xmath46 , somewhat smaller that the @xmath47 scatter found for all halos in simulations .
the results are consistent with the notion that our sample consists of highly relaxed clusters , for which we expect a power - law with mass to hold even at the highest masses , with a smaller intrinsic scatter than the full population .
this is reinforced by the fact that our concentrations are on average greater than those in simulations at the same masses .
selecting relatively relaxed halos from simulations based on the absence of substructure and the ratio of kinetic to potential energy , @xcite find median concentrations of 45 in the mass range considered here , consistent with our observed clusters being typically more relaxed than the halos they selected . as relaxed , 34 times as many as we would expect to select from a mass - limited sample based on the criteria used to construct our cluster data set ( paper ) . ]
these authors also find a clear correlation between halo concentration and the time since `` formation '' , operationally defined as the time when the halo reaches half of its final mass ( see also @xcite ) .
taking their @xmath48 and @xmath49 results as a guide , our measured concentrations in the range 58 place the clusters formation times approximately between redshifts 12 .
our selection of the most relaxed systems is also plausibly responsible for a reduced redshift dependence ( @xmath50 consistent with zero ) compared with all clusters in simulations .
the right panel of figure [ fig : cm ] shows the residuals from the best - fitting concentration
mass relation as a function of central cooling luminosity ( as a proxy for the importance of non - gravitational physics ) , with the latter also expressed as a residual from the best - fitting @xmath23mass relation ( equation [ eq : lcool - scaling ] ) . while the two sets of residuals are not strongly correlated
, it may be worth noting the two clusters with large positive residuals in both concentration and cooling luminosity , rxj1347.5@xmath421145 and spt - clj2344@xmath424242 .
if the relatively efficient cooling at work in the centers of these clusters has an effect on the concentration of the dark matter as well as the gas , boosting the value of @xmath41 , this would not be represented in the dark - matter - only simulations used to predict the c
m relation .
( the closest analogs of these systems in simulations may , of course , still have greater than average concentration . )
concentration mass estimates based on x - ray data , mostly for low - redshift clusters , show broad agreement with the mass dependence predicted from n - body simulations @xcite .
more recently , measurements based on weak gravitational lensing have become possible .
weak lensing shear alone can constrain concentrations only for stacks or ensembles of clusters , but individual cluster constraints are possible in combination with strong lensing or magnification .
our own analysis of weak lensing data for a subset of the clusters used in this work finds an ensemble average concentration of @xmath51 ( paper ) , consistent with our x - ray results here within the statistical uncertainties .
results using strong and weak lensing by x - ray selected clusters @xcite and from stacked analyses of optically selected clusters @xcite are in broadly good agreement with our results .
however , an arguably better comparison is with the lensing analyses of the clash sample of clusters ( @xmath52 ) , which mostly consists of relaxed or nearly relaxed clusters ( by our definition ) . using clash strong+weak lensing , @xcite found @xmath53 and @xmath54 , consistent with our results ; incorporating magnification
, @xcite found @xmath55 while fixing @xmath56 .
we conclude that our concentration mass estimates for relaxed clusters are in good agreement with simulations and previous measurements , given the expectation that we have selected dynamically similar , and , in particular , the most dynamically relaxed , clusters at all redshifts .
direct tests of that selection on simulated cluster images would not be straightforward , given the inability of current hydrodynamic simulations to reproduce the cores of relaxed clusters .
however , the sensitivity of our results to a few strongly cooling clusters motivates doing so in future work , as well as expanding the relaxed cluster sample if possible .
we next turn to constraints on the inner slope of the cluster mass profiles , as parametrized by the gnfw model .
while a profile of the einasto form is now known to provide a better description of simulated clusters over a large range in radii , the gnfw model still represents a simple modification of the baseline nfw model , and one that is arguably easier to interpret in light of the limited radial range of our data .
n - body simulations generally place values of the inner slope of the gnfw profile , @xmath57 ( equation [ eq : profmodels ] ) , in the range 1.01.5 , with @xmath18 corresponding to the nfw model @xcite .
unfortunately , due to its greater complexity compared to the nfw profile , the gnfw model can not be fully constrained for all of the clusters in our sample using the current data . in practice ,
the scale radius parameter becomes unconstrained , taking on very large values , while the inner slope becomes relatively steep , taking values roughly appropriate for the average of the profile in the observed region ( i.e. , @xmath58 ) . a straightforward way to eliminate these cases
is to adopt the outer radius where the data can constrain the mass profile , @xmath59 , as an upper limit for the scale radius ( when fitting the nfw profile , all scale radii are comfortably within the measured range ) .
those clusters for which the 95.4 per cent confidence interval for @xmath60 includes values @xmath61 are removed from this portion of the analysis .
we have checked that making this cut based on the 68.3 per cent confidence interval for @xmath60 does not change our results apart from tightening the constraints slightly ( because fewer clusters are excluded ) .
thus , while this procedure at some level introduces the prior that the mass profiles not be too far from nfw , our results are not very sensitive to it .
the left panel of figure [ fig : gnfw ] shows these individual constraints ( see also table [ tab : thetable ] ) . using the 20 clusters that pass this check
, we fit a model for the average value of @xmath57 and its cluster - to - cluster scatter , which we assume to be described by a gaussian distribution with mean @xmath62 and width @xmath63 .
the joint constraints on these parameters are shown in the right panel of figure [ fig : gnfw ] , and are consistent with the nfw value of @xmath18 .
the marginalized constraints are @xmath64 and @xmath65 .
we emphasize that the data used to fit this model do not extend all the way into the cluster centers , although they do include radii down to @xmath66 ( see section [ sec : data ] ) .
our analysis indicates that , on average , the nfw model provides a good description of the most relaxed clusters down to small radii .
however , it also shows significant intrinsic scatter in @xmath57 , at the @xmath35 per cent level , even among the most relaxed clusters .
earlier work by @xcite and @xcite using x - ray observations reached similar conclusions on the mean value of @xmath57 , respectively finding @xmath67 and @xmath68 ( both 95 per cent confidence intervals ) .
neither of these authors fit for intrinsic scatter explicitly , but their results also pointed to its presence , to the extent that the scatter in individual - cluster constraints on @xmath57 was not consistent with measurement uncertainties alone .
more recently , @xcite also found inner mass profile slopes of massive clusters to be consistent with cdm - only predictions , using a combination of strong and weak gravitational lensing constraints , at radii where stars do not dominate the mass budget .
in detail , the cluster mass profile shapes in simulations have been found to be better described by a continuously changing power - law , such as the einasto form ( equation [ eq : profmodels ] ) , than by the nfw or gnfw models ( e.g. @xcite ) . values of the shape parameter , @xmath19 , seen in simulations are mass and redshift dependent ; for massive clusters at @xmath69 , average values of @xmath470.3 are typical @xcite . while we do not expect our data to be capable of distinguishing the gnfw and einasto models , due to the limited radial range covered ,
it is nonetheless interesting to see how values of @xmath19 measured from the x - ray data compare to these expectations .
similarly to the gnfw case , degeneracies between model parameters prevent us from constraining the einasto model for every cluster in our sample .
we follow the same procedure as in section [ sec : gnfw ] to remove from the analysis clusters where the scale radius of the einasto model is not constrained . unlike the gnfw case , the constraints on @xmath19 for several of the remaining clusters are consistent with either the lower or upper bound of the uniform prior we adopt in the analysis , @xmath70 ( left panel of figure [ fig : ein ] ) .
while this makes our results dependent on the width of this prior , in practice the effect is small ; for example , our results below are nearly identical to those we would obtain with a tighter prior of @xmath71 . as in the previous section
, we fit a gaussian model for the intrinsic distribution of einasto shape parameters , parametrized by @xmath72 and @xmath73 .
the results are shown in the right panel of figure [ fig : ein ] ; the marginalized constraints are @xmath74 and @xmath75 .
although the intrinsic scatter is relatively large , @xmath76 per cent , the mean value of the shape parameter is within the range expected for the most massive galaxy clusters at low redshifts @xcite .
note that , for this value of @xmath19 , the density profile is relatively similar to nfw over the range probed ( figure [ fig : demo ] ) ; hence , this result is consistent with the value @xmath77 arrived at when fitting the gnfw model . from their analysis of x - ray data for 11 clusters
, @xcite found @xmath78 at 95 per cent confidence , with a best - fitting value of @xmath47 .
this is somewhat smaller than our result , although we note that their cluster sample is also lower in mass on average ( @xmath79 ) , such that these results are not necessarily in conflict .
we present constraints on mass profile models for massive , relaxed galaxy clusters based on an analysis of x - ray data from _
chandra_. this analysis assumes hydrostatic equilibrium between the icm and the gravitational potential , and includes data only at cluster radii where we are confident that both departures from equilibrium and systematics due to background modeling are minimal . assuming the nfw mass profile model , the measured concentration mass relation has a power - law slope with mass of @xmath0 , consistent with cdm simulations of cosmological structure formation .
the measured relation is consistent with being constant as a function of redshift , a feature that is not seen in simulations ; however , this is plausibly the result of our selection of the most dynamically relaxed clusters at a given redshift .
simulations including gas physics , on which the same selection procedure is replicated , would be required to perform a completely fair comparison .
we detect an intrinsic scatter of @xmath3 in the concentration mass relation .
two clear high - concentration outliers from the mean relation also have the largest central cooling luminosities in the sample , suggesting a role for baryonic physics in the scatter .
however , the remaining clusters do not support a simple trend between concentration and cooling luminosity at fixed mass .
when fitting a gnfw mass profile , where the slope at small radii is a free parameter , we find an average value that is consistent with the nfw model , @xmath80 .
however , there is significant cluster - to - cluster scatter ( @xmath7 ) , even within the relaxed sample of clusters analyzed here . for the einasto profile
, we similarly measure a mean value , @xmath81 , which is consistent with cdm predictions for clusters of the mass studied here . in this case
, we measure a larger fractional intrinsic scatter , @xmath9 .
overall , our results confirm that the mass distribution within the most massive halos is in agreement with cdm predictions .
there are clear opportunities to improve this analysis in the future , both by finding additional highly relaxed clusters and obtaining deeper x - ray data ( especially to better constrain 3-parameter models like the gnfw and einasto models ) .
although it is beyond the scope of this work , x - ray data for dynamically relaxed clusters at intermediate radii can potentially be combined with weak lensing ( at larger radii , @xmath82 ) , and strong lensing or stellar velocity dispersions ( at small radii , @xmath83 ) , in the vein of e.g. @xcite .
while these additional data sets have their own challenges in the form of projection effects and velocity anisotropy , a careful combination for an appropriately chosen cluster sample potentially provides the most complete possible view of the mass distribution within clusters .
we acknowledge support from the u.s .
department of energy under contract number de - ac02 - 76sf00515 ; from the national aeronautics and space administration ( nasa ) through _ chandra _ award numbers go8 - 9118x and tm1 - 12010x , 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 - 03060 ; and from nasa under grant number nnx15ae12 g . | this is the fifth in a series of papers studying the astrophysics and cosmology of massive , dynamically relaxed galaxy clusters .
our sample comprises 40 clusters identified as being dynamically relaxed and hot in papers and of this series .
here we use constraints on cluster mass profiles from x - ray data to test some of the basic predictions of cosmological structure formation in the cold dark matter ( cdm ) paradigm .
we present constraints on the concentration
mass relation for massive clusters , finding a power - law mass dependence with a slope of @xmath0 , in agreement with cdm predictions . for this relaxed sample ,
the relation is consistent with a constant as a function of redshift ( power - law slope with @xmath1 of @xmath2 ) , with an intrinsic scatter of @xmath3 .
we investigate the shape of cluster mass profiles over the radial range probed by the data ( typically @xmath4kpc1mpc ) , and test for departures from the simple navarro , frenk & white ( nfw ) form , for which the logarithmic slope of the density profile tends to @xmath5 at small radii .
specifically , we consider as alternatives the generalized nfw ( gnfw ) and einasto parametrizations .
for the gnfw model , we find an average value of ( minus ) the logarithmic inner slope of @xmath6 , with an intrinsic scatter of @xmath7 , while in the einasto case we constrain the average shape parameter to be @xmath8 with an intrinsic scatter of @xmath9 .
our results are thus consistent with the simple nfw model on average , but we clearly detect the presence of intrinsic , cluster - to - cluster scatter about the average .
[ firstpage ] dark matter galaxies : clusters : general x - rays : galaxies : clusters |
the @xmath0 expansion of qcd @xcite , where @xmath1 is the number of color charges , has emerged as one of the principal tools for studying low - energy hadronic processes and hadron static observables . in the simplest construction , baryons
are assembled from a collection of @xmath1 quarks , each one transforming under the su@xmath2 fundamental representation , such that the aggregate forms a color singlet @xcite .
while analyzing baryons using the @xmath1 quarks spin - flavor and combinatoric properties has led to a large variety of interesting results far too numerous to review here @xcite , the most compelling physical picture for describing the dynamical properties of large @xmath1 baryons ( particularly scattering amplitudes ) is the chiral soliton approach @xcite originally motivated by the skyrme model @xcite .
one of the most intriguing properties of these studies is the emergence of model - independent linear relations among meson - baryon scattering amplitudes @xcite , whose origin gradually became understood as connected to the large @xmath1 limit @xcite .
in fact , the existence of such relations can be traced back to the contracted su(4 ) spin - flavor symmetry that emerges in the single - baryon sector as @xmath3 , which in turn is obtained by demanding consistent order - by - order unitarity in meson - baryon scattering processes @xcite ; the fact that the former can be derived from the latter was first demonstrated in refs .
@xcite .
the large @xmath1 scattering method @xcite explains the existence of baryon resonance multiplets that share similarities ( but are not identical to ) those appearing in large @xmath1 quark models ; in particular , one can study resonances with arbitrarily large widths @xcite , exotics @xcite , and three - flavor resonances @xcite .
in addition , the means by which the spurious @xmath4 states may be removed @xcite has been explored , as well as the nature of the competing chiral and large @xmath1 limits @xcite , and results for multipion processes @xcite . an essential ingredient to obtaining results useful for @xmath5 phenomenology is understanding how to include @xmath0 corrections .
the original large @xmath1 scattering amplitude relations were noted long ago to satisfy the @xmath6-channel isospin - spin exchange constraint @xmath7 @xcite . using operator techniques , ref .
@xcite showed that static pion - baryon couplings with @xmath8 are suppressed by a relative factor of @xmath9 , and the same techniques were used to show @xcite that nucleon - nucleon interactions with @xmath10 are suppressed by @xmath11 . much more recently ,
the same techniques were used to show @xcite the analogous result for @xmath6-channel pion - baryon scattering amplitudes .
the expression of these pion - baryon constraints in terms of @xmath12-channel observables is explored in refs .
@xcite . in this paper
we modify the approach for describing meson - baryon scattering amplitudes in the @xmath0 expansion , as it was applied to the case of pion photoproduction @xmath13 @xcite , to provide a model - independent expansion for electromagnetic multipole amplitudes of the related pion electroproduction process @xmath14 , which at its essence is the virtual photon process @xmath15 .
studies of @xmath16 using the more traditional `` operator '' approach to large @xmath1 baryons also appear in the literature @xcite .
the fundamentals of hadronic electroproduction are reviewed in ref .
the photon squared four - momentum @xmath17 generalizes from zero in the photoproduction case to a nonzero value for electroproduction , joining the center - of - momentum ( c.m . )
energy @xmath18 of the @xmath19 system as an independent kinematic variable for all amplitudes .
furthermore , virtual photons possess not only the familiar electric and magnetic transverse multipoles , but scalar and longitudinal multipoles as well , leading to a much richer set of experimental possibilities .
even so , the analysis of electroproduction amplitudes in the @xmath0 expansion is almost identical to that for the case of photoproduction . in this paper
we derive expressions for pion electroproduction amplitudes at leading order ( lo ) and next - to - leading order ( nlo ) in the @xmath0 expansion and examine our findings using results of the maid 2007 partial wave analysis from universitt mainz @xcite .
our purpose in this paper is to present only the results of a strict @xmath0 analysis .
while the @xmath0 expansion relates combinations of distinct amplitudes and predicts the magnitude of these differences at each value of @xmath17 and @xmath18an infinite number of testable predictions it does _ not _ predict the shapes of their @xmath17 or @xmath18 dependences . obtaining such predictions
would require imposing dynamical assumptions that lie beyond the raw mandates of the @xmath0 expansion .
one may impose calculations using , for example , chiral perturbation theory , specific quark models , or generalized parton distributions on top of the amplitude predictions of this paper ( granted that they have been generalized to allow for @xmath1 to be arbitrary ) to obtain predictions for the detailed @xmath17 and @xmath18 amplitude shapes , but such modifications lie outside the intentionally limited scope of this paper .
the only explicit @xmath0 dynamical effects we discuss below arise due to the displaced pole positions of baryon resonances in different channels . this paper is organized as follows : in sec .
[ relns ] we derive the linear electroproduction scattering amplitude relations . in sec .
[ results ] we confront the relations with the extensive results of the maid 2007 partial wave analysis , and comment upon the quality of the comparisons .
we summarize briefly in sec .
[ concl ] .
virtually all results stated in this section are a direct reprise of those for the photoproduction case , ref .
@xcite ; the differences are particularly noted .
the results for either process are obtained by starting with general meson - baryon scattering processes of the form @xmath20 , where @xmath21 are mesons and @xmath22 are baryons , each pair carrying fixed strangeness .
since the amplitude relations @xcite for such processes depend upon the mesons only through their quantum numbers , the same results may be used for electroproduction , for which @xmath23 is a virtual photon ( or technically , a meson interpolating field with the quantum numbers of a virtual photon ) .
the master scattering formula for the observable scattering amplitudes @xmath24 reads @xcite @xmath25([r_i][r_f][s_i][s_f ] [ \tilde{k}_i][\tilde{k}_f])^{1/2 } \left\ { \begin{array}{ccc } l_i & i_i & \tilde{k}_i \\ s_i & r_i & s_i \\
j & i & k \\ \end{array } \right\ } \left\ { \begin{array}{ccc } l_f & i_f & \tilde{k}_f \\
s_f & r_f & s_f \\
j & i & k \\
\end{array } \right\ } \tau_{k\tilde{k}_i\tilde{k}_fl_il_f } \ , , \nonumber \\\end{aligned}\ ] ] where subscripts @xmath26 and @xmath27 label initial- and final - state quantities , respectively . here
, @xmath28 indicates the baryon spin ( which equals its isospin for both @xmath29 and @xmath30 ) , @xmath12 and @xmath26 label the spin and isospin of the meson ( or photon ) , respectively , @xmath31 labels the orbital angular momentum of the meson ( or photon ) relative to the baryon target , @xmath32 labels the total spin of the system , and @xmath33 is a hybrid quantity that provides good quantum numbers @xmath34 and @xmath35 in the large @xmath1 limit . the overall conserved quantum numbers @xmath36 , @xmath37 , and @xmath38 arise from the total isospin @xmath39 , the total angular momentum @xmath40 , and the so - called grand spin @xmath41 ( so that @xmath38 is also a good quantum number in the large @xmath1 limit ) .
the braces are conventional @xmath42 symbols , and the multiplicity @xmath43 of each su(2 ) representation @xmath44 is denoted by @xmath45 $ ] .
the sums run over all values of @xmath38 , @xmath34 , and @xmath35 consistent with the nonvanishing of the @xmath42 symbols , _
i.e. _ , each row and column satisfies the triangle rule .
beyond all the group - theoretical factors lie the _ reduced amplitudes
_ @xmath46 , which are the undetermined finite dynamical quantities in the large @xmath1 scattering amplitude approach , analogous to reduced matrix elements in the wigner - eckart theorem ; their precise calculation would be tantamount to solving qcd exactly at leading order in the @xmath0 expansion .
virtual photons can carry either spin one or zero ( the latter in distinction to real photons ) ; however , the angular momentum @xmath47 labeling each electromagnetic multipole is comprised of the combination of the photon intrinsic spin with its orbital angular momentum relative to the target @xcite ( in this case , the initial baryon ) . in terms of eq .
( [ master ] ) , one effectively handles photon angular momentum by setting @xmath48 and @xmath49 .
the photon isospin content is more complicated ; photons carry both isovector and isoscalar contributions , which couple to baryons through operators carrying different @xmath1 counting . since this coupling occurs through the photon polarization vector @xmath50 , and since transverse , longitudinal , and scalar polarizations ( the latter two arising only for virtual photons ) all have nonvanishing components in spatial ( @xmath51 ) directions , the relevant operators representing photon couplings to baryons are @xmath52 and @xmath53 where @xmath54 and @xmath46 are pauli matrices in spin and isospin , respectively , and @xmath55 sums over the @xmath1 quark fields @xmath56 in the baryon . for the ground - state baryons ( _ e.g. _ , @xmath29 and @xmath30 ) ,
matrix elements of the isoscalar operator @xmath57 are of course @xmath58 ( since they are just components of the total baryon angular momentum ) , but those of the isovector operator @xmath59 are @xmath60 due to the collective contribution of the @xmath1 quarks . however , eq . ( [ master ] ) does not incorporate this constraint , and therefore it must be put in by hand : the full version of eq .
( [ master ] ) for electroproduction is the sum of a lo isovector ( @xmath61 ) piece and an isoscalar ( @xmath62 ) piece carrying an explicit @xmath0 suppression factor , as expressed below in eqs .
( [ ivec_lo])([ivec_nlo ] ) .
the master amplitude expression eq .
( [ master ] ) applied to pion electroproduction off a nucleon target has @xmath63 and @xmath64 as mentioned above , as well as @xmath65 , @xmath66 , @xmath67 ( but retaining for the moment the explicit symbol @xmath68 ) , and @xmath69 .
the triangle rules then force @xmath70 , and we relabel @xmath71 , which is the pion partial wave .
one finds @xmath72 \left\ { \begin{array}{ccc } j & \ell & \frac{1}{2 } \\ i_{\gamma } & i & k \\
\end{array } \right\ } \left\ { \begin{array}{ccc } j & l & \frac{1}{2 } \\
i_f & i & k \\
\end{array } \right\ } \tau^{\lambda}_{k\ell
l } \ , .\ ] ] amplitudes for specific charge channels are obtained by attaching the appropriate isospin clebsch - gordan coefficients to eq .
( [ s6 ] ) . labeling the isospin third component of the incoming nucleon by @xmath73 and that of the outgoing pion by @xmath74 , one has @xmath75
\left\ { \begin{array}{ccc } j & \ell & \frac{1}{2 } \\
i_{\gamma } & i & k \\
\end{array } \right\ } \left\ { \begin{array}{ccc } j & l & \frac{1}{2 } \\
i_f & i & k \\
\end{array } \right\ } \tau^{\lambda}_{k\ell l } \nonumber \\ & \times & \left ( \begin{array}{cc|c } i_f & \frac{1}{2 } & i \\
\nu & m_i-\nu & m_i \\ \end{array } \right ) \left ( \begin{array}{cc|c } i_{\gamma } & \frac{1}{2 } & i \\ 0 & m_i & m_i \\
\end{array } \right ) \ , .\end{aligned}\ ] ] the label @xmath76 indicates the type of multipole amplitude : @xmath77 odd gives electric ( e ) , longitudinal ( l ) , and scalar ( s ) multipoles ( the latter two being absent from photoproduction ) , and @xmath78 even gives magnetic ( m ) multipoles . the l and s multipoles are linearly dependent due to current conservation @xcite , so we choose in this analysis to eliminate l multipoles in favor of s multipoles .
we now exploit the result that amplitudes with @xmath79 are suppressed by a relative factor @xmath11 .
the first step is to rewrite the @xmath42 symbols in eq .
( [ master ] ) in terms of @xmath6-channel quantum numbers using the well - known su(2 ) relation known as the biedenharn - elliot sum rule @xcite . in terms of modified @xmath80 symbols ( called @xmath81 $ ] symbols in ref .
@xcite ) , @xmath82[b][c][d])^{1/4 } } \left [ \begin{array}{ccc } a & b & e \\
c & d & f \\
\end{array } \right ] \
, , \ ] ] one obtains @xmath83 } { 2\sqrt{[i_f][i_\gamma][l][\ell ] } } \left [ \begin{array}{ccc } i_f & l & k \\ \ell & i_\gamma & { \cal j } \end{array } \right ] \left [ \begin{array}{ccc } i_f & \frac{1}{2 } & i \\ \frac{1}{2 } & i_\gamma & { \cal j } \end{array } \right ] \left [ \begin{array}{ccc } l & \frac{1}{2 } & j \\ \frac{1}{2 } & \ell & { \cal j } \end{array } \right ] \ , \label{biedell } \nonumber \\ & & \end{aligned}\ ] ] where the quantum number @xmath84 clearly adopts the role of both @xmath85 ( as seen from the second @xmath81 $ ] symbol ) and @xmath86 ( as seen from the third ) . setting at last @xmath67 and defining the @xmath6-channel reduced amplitudes by @xmath87 } { \sqrt{[1][i_{\gamma}][l][\ell ] } } \sum_{k}[k ] \left [ \begin{array}{ccc } 1 & l & k \\
\ell & i_{\gamma } & \mathcal { j } \\ \end{array } \right]\tau^{\lambda}_{k\ell l } \ , , \ ] ] one obtains the multipoles for the isovector case ( @xmath88 ) : @xmath89 \left [ \begin{array}{ccc } l & \frac{1}{2 } & j \\
\frac{1}{2 } & \ell & \mathcal { j } \\
\end{array } \right ] \tau^{t\lambda1}_{\mathcal { j}\ell l } \ , , \ ] ] and those for isoscalar case ( @xmath90 ) : @xmath91^{1/4 } } \left [ \begin{array}{ccc } l & \frac{1}{2 } & j \\ \frac{1}{2 } & \ell & 1 \\ \end{array } \right ] \tau^{t\lambda0}_{1\ell l } \ , .\ ] ] note the explicit @xmath0 suppression factor in the isoscalar expression . in order to achieve relations at a consistent order in the @xmath0 expansion
, one must also include the independent nlo isovector amplitudes , which have @xmath92 .
generalizing eq .
( [ ivec_lo ] ) gives @xmath93 \ !
\left [ \begin{array}{ccc } l & \frac{1}{2 } & j \\ \frac{1}{2 } & \ell & x+1 \\
\end{array } \right ] \tau^{t\lambda(+)}_{x\ell l } \ !
+ \sum_{y } \left [ \begin{array}{ccc } 1 & \frac{1}{2 } & i \\
\frac{1}{2 } & 1 & y \\
\end{array } \right ] \ !
\left [ \begin{array}{ccc } l & \frac{1}{2 } & j \\ \frac{1}{2 } & \ell & y-1 \\
\end{array } \right ] \tau^{t\lambda(-)}_{y\ell l } \right\ } , \nonumber \\ & & \end{aligned}\ ] ] where @xmath94 in the first sum and @xmath95 in the second represent @xmath85 , and the amplitudes @xmath96 are independent of those at leading order . the total multipole amplitude , including lo and nlo terms to consistent order in @xmath0 , is therefore the sum of eqs .
( [ ivec_lo ] ) , ( [ iscalar ] ) , and ( [ ivec_nlo ] ) : @xmath97 by including all values of @xmath98 , @xmath94 , and @xmath95 allowed by the triangle rules in the [ 6j ] symbols and simplifying , one obtains the expression @xmath99\tau^{t\lambda 1}_{1\ell l } \right
. \nonumber \\ & & \left .
+ \frac{1}{n_c } \left ( \left[\begin{array}{ccc } l & \frac{1}{2 } & j \\ \frac{1}{2 } & \ell & 1 \\ \end{array}\right ] \tau^{t\lambda(+)}_{0\ell l } + \sqrt{\frac{2}{3 } } \left ( \delta_{i,\frac{1}{2}}- \frac 1 2 \delta_{i,\frac{3}{2 } } \right ) \delta_{\ell , l }
\tau^{t\lambda(-)}_{1 l l } \right ) \right\ } \nonumber\\ & & \left.+\frac{1}{n_c}\frac{\delta_{i,\frac{1}{2}}}{[1]^{1/4 } } \left [ \begin{array}{ccc } l & \frac{1}{2 } & j \\
\frac{1}{2 } & \ell & 1 \\ \end{array } \right ] \tau^{t\lambda 0}_{1\ell l } \right ] \ , , \end{aligned}\ ] ] which is identical in form to the expression found for photoproduction @xcite .
pion electroproduction possesses four charged channels : @xmath100 , @xmath101 , @xmath102 and @xmath103 . assuming isospin invariance , only three of these are independent @xcite , and so we choose to eliminate the channel @xmath103 for which all the particles are neutral .
since the initial- and final - state nucleons carry only spin 1/2 , parity invariance constrains electric and scalar multipole amplitudes to @xmath104 , while magnetic multipole amplitudes satisfy @xmath105 .
the spinless pion and spin-1/2 nucleon combine to give possible angular momenta @xmath106 ; it is convenient to express this information using the combination @xmath107 , which equals @xmath108 for @xmath106 , respectively , and write the amplitudes as @xmath109 . using the conventions @xcite employed by maid , the multipole amplitude @xmath109 ( where @xmath76 = e , s , or m ) is proportional to @xmath110 ; for example , @xmath111 .
the proportionality factor depends upon particle energies , @xmath76 , @xmath31 , and @xmath107 , but is not important for this analysis since in each relation presented below it cancels out as a common factor . using eq .
( [ full ] ) , one obtains very similar relations for electric and scalar multipoles : @xmath112 @xmath113 and @xmath114 the presence of an @xmath115 relation for s but not e amplitudes in eq .
( [ e1 ] ) reflects the existence of c0 but not e0 electromagnetic multipoles .
referring to eq .
( [ full ] ) , one obtains four relations at lo because one has six observable amplitudes , arising through three charged channels each with two allowed values of @xmath107 , but only two reduced amplitudes , @xmath116 . however , no relations survive at nlo because four new amplitudes ( @xmath117 and @xmath118 ) appear at this order .
equation ( [ e3 ] ) implies the vanishing of the e and s multipole amplitudes at lo for the process @xmath119 , which means that they are expected to be , on average , a factor of about @xmath120 smaller than the charged amplitudes ; but this is a rather qualitative statement [ in particular , eq .
( [ e3 ] ) does not test the equality of measurable amplitudes ] , and we omit numerical analysis of such amplitudes below . using eq .
( [ full ] ) for the magnetic multipole amplitudes , one again has six observable amplitudes expressed at lo in terms of only two reduced amplitudes ( @xmath121 and @xmath122 ) , leading to four relations : @xmath123 @xmath124 only three new reduced amplitudes ( @xmath125 , @xmath126 , and @xmath127 ) appear at nlo , meaning that one special combination holds at both lo and nlo : @xmath128+o(n^{-2}_c ) \ ; ( l\geq1 ) \ , . \nonumber \\\end{aligned}\ ] ] one expects this relation to improve generically upon the predictions of the lo relations by a factor of about @xmath129 .
note that several of the lo relations , specifically those of the form @xmath130 [ which are eqs .
( [ e1])([e2 ] ) and the first and third of relations in eq .
( [ m2 ] ) ] follow directly from isospin symmetry and the lo dominance of the isovector amplitude , as may be checked simply from the clebsch - gordan coefficients in eq .
( [ ivec_lo ] ) .
all of the electroproduction multipole relations presented here also appear for the photoproduction , of course excepting the scalar multipole relations .
our results consist of a comparison of the relations eqs .
( [ e1])([e2 ] ) and ( [ m1])([m3 ] ) to the experimental data as obtained from the maid 2007 partial - wave analysis @xcite .
the exceptionally large volume of available information is presented as concisely as possible : each allowed partial wave [ @xmath31 can be arbitrarily large , and eqs .
( [ e1])([m3 ] ) provide relations between such amplitudes , but data is typically compiled only up to @xmath131 is a complex - valued amplitude and depends upon two independent dynamical variables : the photon virtuality @xmath132 and the @xmath19 c.m.energy @xmath18 . in each plot
@xmath18 ranges from threshold to 2 gev , and @xmath133 ranges from 05 gev@xmath134 . in all figures we present both real and imaginary parts of the left - hand and right - hand sides ( l.h.s . and r.h.s . ) of each multipole amplitude combination appearing in eqs .
( [ e1])([e2 ] ) and ( [ m1])([m3 ] ) ( except for @xmath135 and @xmath136 imaginary parts , given as zero by maid ) , and then also plot scale - independent amplitude ratios according to the prescription @xmath137 for a lo relation , this quantity is dimensionless and predicted to be of order @xmath0 , while for nlo relations the prediction is @xmath138 .
absolute values appear in the denominator to avoid physically uninteresting behavior when l.h.s . and r.h.s .
are both near zero and happen to be equal and opposite .
maximally poor relationships ( @xmath139 or vice versa ) are manifested by ratio @xmath140 . on the other hand , relations that truly hold to @xmath141
might be expected to lie between @xmath142 ; however , this conclusion neglects the order - unity coefficient implicit in @xmath141 .
we choose as a useful metric to distinguish between @xmath141 and @xmath58 quantities their geometric mean , @xmath143 for @xmath129 .
similarly , we take the largest @xmath138 effects to be @xmath144 ; nevertheless , the reader is free to choose their own figure of merit upon viewing the plots .
figure [ e - plot ] tests the e multipole relations in eq .
( [ e1 ] ) , which compare amplitudes with @xmath145 , denoted as @xmath146 by maid .
figure [ s - plot ] does the same for @xmath147 multipoles .
the @xmath148 amplitudes for e and s [ relations in eq .
( [ e2 ] ) for multipoles @xmath149 and @xmath150 appear in figs .
[ e+plot ] and [ s+plot ] , respectively .
the m multipole relations eq . ( [ m1 ] ) containing a @xmath151 and relating @xmath152 amplitudes ( @xmath153 ) appear in fig .
[ m1plot ] .
all of the relations presented in these first figures are lo in the @xmath0 expansion .
figure [ m2plot ] tests the m multipole relations containing @xmath154 , first for the @xmath155 multipoles alone [ the lo relation represented by the first expression in eq .
( [ m2 ] ) ] and then for the nlo relation between @xmath153 multipoles given in eq .
( [ m3 ] ) .
we remind the reader that the relations explored in figs .
[ e - plot][s+plot ] and the lo comparisons of fig .
[ m2plot ] are dominated by the isovector component of the photon .
first observe that the limit @xmath156 corresponds to scattering with an on - shell photon , _
i.e. _ , real pion photoproduction .
the projection of each e and m multipole on the @xmath133 axis gives amplitude curves obtained in ref .
@xcite and analyzed there in terms of the multipole relations eqs .
( [ e1])([e2 ] ) and ( [ m1])([m3 ] ) restricted to @xmath157 .
the most prominent features in the amplitudes are of course the baryon resonances , which are most apparent through large enhancements of the imaginary parts at values of c.m .
energy @xmath18 equal to a resonance mass , but also noticeable from points where the real parts vanish .
the falloff of each amplitude with increasing @xmath133 may be interpreted as the canonical behavior of an @xmath158 electromagnetic transition form factor , with the on - shell photon ( @xmath157 ) providing the least disruptive probe of the initial nucleon and hence the most efficient probe for producing a resonance .
one finds , however , interesting exceptions to this reasoning ; in the imaginary part of the @xmath159 amplitudes ( fig . [ s - plot ] ) , one sees that the @xmath160 produced in @xmath161 ( l.h.s . )
peaks at @xmath162 , while that in @xmath163 ( r.h.s . )
peaks at @xmath164 gev@xmath134 . the resonant behavior accounts for the largest source of discrepancies of the @xmath0 relations .
often one finds resonant behavior on both sides of a given relation , but with greatly differing residues at the peak ; this is the case for the @xmath165 peak in each @xmath166 amplitude of fig .
[ e - plot ] or each @xmath167 amplitude of fig .
[ s - plot ] .
even the sign of the @xmath168 residue is different between the two @xmath169 amplitudes in fig .
[ e+plot ] , while the @xmath168 appears to be absent altogether from the r.h.s.plot for @xmath170 in fig .
[ s+plot ] although the amplitudes otherwise appear very similar .
these are interesting anomalies arising even in amplitudes for which the value of @xmath37 on both sides of the amplitude relation are the same , so that the same resonance ( or more accurately , different members of an isomultiplet ) appears on both sides . for fig .
[ m1plot ] and the nlo relation of fig .
[ m2plot ] , @xmath37 differs on the two sides of the equation , meaning that distinct resonances appear . in the @xmath171 amplitudes of fig . [ m1plot ] , for example , the @xmath172 resonance appearing in the @xmath173 ( l.h.s .
) amplitude is broad and shallow , while the @xmath174 forms a huge peak in the @xmath175 ( r.h.s . )
amplitude , with just a hint of the @xmath172 s true large-@xmath1 partner @xcite , the @xmath176 , perhaps just visible .
in fact , the @xmath174 should be eliminated from a large @xmath1 analysis of baryon resonances , because it is actually a degenerate partner to the nucleon whose width vanishes as @xmath177 , unlike the true resonances whose widths remain finite in this limit @xcite .
it is only due to the numerical accident that the chiral limit ( @xmath178 ) is more closely achieved in our @xmath129 universe than is the @xmath179 limit @xcite that the @xmath174 in our universe decays to @xmath180 .
however , true resonant `` shifted degenerate '' pairs such as @xmath172 and @xmath176 are particularly interesting because their masses differ by an amount that is @xmath138 relative to their average @xcite , about 100150 mev .
another example of this effect appears in the @xmath181 amplitudes in fig .
[ m1plot ] , with the @xmath160 [ and perhaps also the @xmath182 or @xmath183 appearing in the @xmath184 ( l.h.s . ) and @xmath185 with a much smaller residue appearing in the @xmath186 ( r.h.s . ) .
in fact , precisely the pair of @xmath160 and @xmath185 are considered as degenerate resonances in ref .
@xcite , where it is shown that the on - resonance couplings for the different channels behave exactly as expected from the @xmath0 expansion for @xmath129 . another `` shifted degenerate '' pair in fig .
[ m1plot ] appears to be @xmath165 in the @xmath187 ( l.h.s . )
amplitude and @xmath188 in the @xmath189 ( r.h.s . ) amplitude ; the @xmath188 is not apparent in the plots of ref .
@xcite , which employs the older maid 2003 analysis rather than the maid 2007 variant used here . in the case of the nlo relations eq . ( [ m3 ] ) considered in fig .
[ m2plot ] , the lo terms are @xmath155 and the nlo terms are @xmath190 , so entirely new resonances can appear at nlo .
a curious example occurs for @xmath115 ; the broad peak appearing for @xmath161 ( l.h.s . ) , probably due to the @xmath172 , is unmatched by a contribution in @xmath191 in the @xmath173 amplitudes , and moreover , the @xmath168 appears prominently in the @xmath175 nlo corrections .
on the other hand , the nlo contributions sometimes indicate a shifted degenerate pair ; this effect occurs for the already - mentioned @xmath160 and @xmath185 pair occurring at @xmath192 @xcite . in regions away from resonances and for amplitudes
not exhibiting any obvious resonances , the agreement tends to be rather better .
this is certainly true for the @xmath193 and 5 amplitudes , but also for larger values of @xmath133 and for values of @xmath18 close to threshold as well as values at least 50100 mev from resonant peaks . of the highest quality
are the isovector - dominated relations in figs .
[ e - plot][s+plot ] , for which the same value of @xmath37 appearing on either side of the equation means that resonances appearing on the two sides of the relation belong to an isomultiplet . on the other hand , the benefits of the nlo correction in fig .
[ m2plot ] remain ambiguous ; for example , some improvement appears in the real part of the @xmath115 relation , but the quality degrades in some kinematic regions for @xmath192 , and scarcely any change due to the nlo terms is noticeable for @xmath194 .
the @xmath0 expansion , which has provided so much qualitative and semiquantitative guidance to understanding baryons in general , and the baryon resonance spectrum in particular , appears in the case of electroproduction to produce more ambiguous results .
on one hand , it gives a natural explanation for the dominance of isovector over isoscalar amplitudes , and it provides a definite set of linear relations between multipole amplitudes that are expected to hold at all values of c.m .
energy @xmath18 and photon virtuality @xmath133 . for values of @xmath18
not near resonance masses as well as for larger values of @xmath133 , the agreement tends to be in accordance with the expectations of the @xmath0 expansion . even in the resonant region
, one often sees evidence of the `` shifted degenerate '' resonances carrying different quantum numbers [ such as @xmath160 and @xmath185 ] that have related couplings .
however , just as many cases exist in which the amplitudes in the resonant region do not entirely conform to naive @xmath0 expectations , both at leading and subleading order .
the specific reason that a given large @xmath1 relation for pion electroproduction works surprisingly well or surprisingly poorly in the resonant region remains a challenge for future research .
= 0.45 = 0.45 + = 0.45 = 0.45 + = 0.45 = 0.45 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 = 0.44 = 0.44 + = 0.44 = 0.44 + = 0.44 = 0.44 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 = 0.44 = 0.44 + = 0.44 = 0.44 + = 0.44 = 0.44 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 = 0.44 = 0.44 + = 0.44 = 0.44 + = 0.44 = 0.44 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 = 0.44 = 0.44 + = 0.44 = 0.44 + = 0.44 = 0.44 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 = 0.41 + = 0.41 = 0.41 + = 0.41 = 0.41 + = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48 + = 0.48 + = 0.48 = 0.48 + = 0.48 = 0.48
this work was supported by the nsf under grant no .
phy-0757394 . g. t hooft , nucl .
b * 72 * , 461 ( 1974 ) .
e. witten , nucl .
b * 160 * , 57 ( 1979 ) .
dashen , e.e .
jenkins , and a.v .
manohar , phys .
d * 49 * , 4713 ( 1994 ) [ erratum - ibid .
d * 51 * , 2489 ( 1995 ) ] [ arxiv : hep - ph/9310379 ] ; phys . rev .
d * 51 * , 3697 ( 1995 ) [ arxiv : hep - ph/9411234 ] ; j. dai , r.f .
dashen , e.e .
jenkins , and a.v .
manohar , phys .
d * 53 * , 273 ( 1996 ) [ arxiv : hep - ph/9506273 ] ; e.e .
jenkins and a.v .
manohar , phys .
b * 335 * , 452 ( 1994 ) [ arxiv : hep - ph/9405431 ] ; e.e .
jenkins and r.f .
lebed , phys .
d * 52 * , 282 ( 1995 ) [ arxiv : hep - ph/9502227 ] ; e.e .
jenkins , phys .
d * 54 * , 4515 ( 1996 ) [ arxiv : hep - ph/9603449 ] ; r. flores - mendieta , e.e .
jenkins , and a.v .
manohar , phys .
d * 58 * , 094028 ( 1998 ) [ arxiv : hep - ph/9805416 ] ; e.e .
jenkins , x.d .
ji , and a.v .
manohar , phys .
lett . * 89 * , 242001 ( 2002 ) [ arxiv : hep - ph/0207092 ] ; m.a .
luty and j. march - russell , nucl .
b * 426 * , 71 ( 1994 ) [ arxiv : hep - ph/9310369 ] ; m.a .
luty , j. march - russell , and m.j .
white , phys .
d * 51 * , 2332 ( 1995 ) [ arxiv : hep - ph/9405272 ] ; c. carone , h. georgi , and s. osofsky , phys .
b * 322 * , 227 ( 1994 ) [ arxiv : hep - ph/9310365 ] ; c.d .
carone , h. georgi , l. kaplan , and d. morin , phys .
d * 50 * , 5793 ( 1994 ) [ arxiv : hep - ph/9406227 ] ; d. pirjol and t.m .
yan , phys .
d * 56 * , 5483 ( 1997 ) [ arxiv : hep - ph/9701291 ] ; phys .
d * 57 * , 5434 ( 1998 ) [ arxiv : hep - ph/9711201 ] ; c.e .
carlson , c.d .
carone , j.l .
goity , and r.f .
lebed , phys .
b * 438 * , 327 ( 1998 ) [ arxiv : hep - ph/9807334 ] ; phys .
d * 59 * , 114008 ( 1999 ) [ arxiv : hep - ph/9812440 ] ; c.e .
carlson and c.d .
carone , phys .
b * 484 * , 260 ( 2000 ) [ arxiv : hep - ph/0005144 ] ; c.l .
schat , j.l .
goity , and n.n .
scoccola , phys .
* 88 * , 102002 ( 2002 ) [ arxiv : hep - ph/0111082 ] ; j.l .
goity , c.l .
schat , and n.n .
scoccola , phys .
d * 66 * , 114014 ( 2002 ) [ arxiv : hep - ph/0209174 ] ; d. pirjol and c. schat , phys .
d * 67 * , 096009 ( 2003 ) [ arxiv : hep - ph/0301187 ] ; phys .
d * 71 * , 036004 ( 2005 ) [ arxiv : hep - ph/0408293 ] ; j.l .
goity , c. schat , and n.n .
scoccola , phys .
b * 564 * , 83 ( 2003 ) [ arxiv : hep - ph/0304167 ] ; j.l .
goity , c. schat and n. scoccola , phys .
d * 71 * , 034016 ( 2005 ) [ arxiv : hep - ph/0411092 ] ; j.l .
goity and n.n .
scoccola , phys .
d * 72 * , 034024 ( 2005 ) [ arxiv : hep - ph/0504101 ] ; r.f .
lebed and d.r .
martin , phys .
d * 70 * , 016008 ( 2004 ) [ arxiv : hep - ph/0404160 ] ; n. matagne and f. stancu , phys .
d * 71 * , 014010 ( 2005 ) [ arxiv : hep - ph/0409261 ] ; phys . lett .
b * 631 * , 7 ( 2005 ) [ arxiv : hep - ph/0505118 ] ; phys .
d * 74 * , 034014 ( 2006 ) [ arxiv : hep - ph/0604122 ] ; nucl .
a * 811 * , 291 ( 2008 ) [ arxiv : hep - ph/0610099 ] ; phys .
d * 73 * , 114025 ( 2006 ) [ arxiv : hep - ph/0603032 ] ; phys .
d * 77 * , 054026 ( 2008 ) [ arxiv:0801.3575 [ hep - ph ] ] ; c. semay , f. buisseret , n. matagne , and f. stancu , phys .
d * 75 * , 096001 ( 2007 ) [ arxiv : hep - ph/0702075 ] .
adkins , c.r .
nappi , and e. witten , nucl .
b * 228 * , 552 ( 1983 ) ; g.s .
adkins and c.r .
nappi , nucl .
b * 249 * , 507 ( 1985 ) .
skyrme , proc .
soc . lond .
a * 260 * , 127 ( 1961 ) .
a. hayashi , g. eckart , g. holzwarth , and h. walliser , phys .
b * 147 * , 5 ( 1984 ) ; m.p .
mattis and m. karliner , phys .
d * 31 * , 2833 ( 1985 ) .
m.p . mattis and m.e .
peskin , phys .
d * 32 * , 58 ( 1985 ) ; m.p .
mattis , phys .
lett . * 56 * , 1103 ( 1986 ) .
m.p . mattis and m. mukerjee ,
lett . * 61 * , 1344 ( 1988 ) .
gervais and b. sakita , phys .
lett . * 52 * , 87 ( 1984 ) ; phys .
d * 30 * , 1795 ( 1984 ) .
dashen and a.v .
manohar , phys .
b * 315 * , 425 ( 1993 ) [ arxiv : hep - ph/9307241 ] ; * 315 * , 438 ( 1993 ) [ arxiv : hep - ph/9307242 ] .
dashen , e.e .
jenkins , and a.v .
manohar , phys .
d * 49 * , 4713 ( 1994 ) [ erratum - ibid .
d * 51 * , 2489 ( 1995 ) ] [ arxiv : hep - ph/9310379 ] .
cohen and r.f .
lebed , phys .
91 * , 012001 ( 2003 ) [ arxiv : hep - ph/0301167 ] ; phys .
d * 67 * , 096008 ( 2003 ) [ arxiv : hep - ph/0301219 ] ; phys .
d * 68 * , 056003 ( 2003 ) [ arxiv : hep - ph/0306102 ] .
cohen , d.c .
dakin , a. nellore , and r.f .
lebed , phys .
d * 69 * , 056001 ( 2004 ) [ arxiv : hep - ph/0310120 ] .
cohen and r.f .
lebed , phys .
b * 578 * , 150 ( 2004 ) [ arxiv : hep - ph/0309150 ] .
cohen and r.f .
lebed , phys .
d * 70 * , 096015 ( 2004 ) [ arxiv : hep - ph/0408342 ] ; phys .
b * 619 * , 115 ( 2005 ) [ arxiv : hep - ph/0504200 ] ; phys .
d * 72 * , 056001 ( 2005 ) [ arxiv : hep - ph/0507267 ] ; r.f .
lebed , phys .
b * 639 * , 68 ( 2006 ) [ arxiv : hep - ph/0603150 ] .
cohen and r.f .
lebed , phys .
d * 74 * , 036001 ( 2006 ) [ arxiv : hep - ph/0604175 ] .
cohen and r.f .
lebed , phys .
d * 74 * , 056006 ( 2006 ) [ arxiv : hep - ph/0608038 ] .
kwee and r.f .
lebed , phys . rev .
d * 75 * , 016002 ( 2007 ) [ arxiv : hep - ph/0611035 ] .
dashen , e.e .
jenkins , and a.v .
manohar , phys .
d * 51 * , 3697 ( 1995 ) [ arxiv : hep - ph/9411234 ] .
kaplan and m.j .
savage , phys .
b * 365 * , 244 ( 1996 ) [ arxiv : hep - ph/9509371 ] ; d.b .
kaplan and a.v .
manohar , phys .
c * 56 * , 76 ( 1997 ) [ arxiv : nucl - th/9612021 ] .
cohen , d.c .
dakin , a. nellore , and r.f .
lebed , phys .
d * 70 * , 056004 ( 2004 ) [ arxiv : hep - ph/0403125 ] .
kwee and r.f .
lebed , jhep * 0710 * , 046 ( 2007 ) [ arxiv : hep - ph/0612267 ] ; j. phys .
a * 41 * , 015206 ( 2008 ) [ arxiv:0711.4340 [ hep - ph ] ] .
cohen , d.c .
dakin , r.f .
lebed , and d.r .
martin , phys .
d * 71 * , 076010 ( 2005 ) [ arxiv : hep - ph/0412294 ] .
carlson and c.d .
carone , phys .
d * 58 * , 053005 ( 1998 ) [ arxiv : hep - ph/9804304 ] ; n.n .
scoccola , j.l .
goity , and n. matagne , phys .
b * 663 * , 222 ( 2008 ) [ arxiv:0711.4203 [ hep - ph ] ] ; j.l .
goity and n.n .
scoccola , phys . rev .
lett . * 99 * , 062002 ( 2007 ) [ arxiv : hep - ph/0701244 ] .
d. drechsel and l. tiator , j. phys .
g * 18 * , 449 ( 1992 ) .
d. drechsel , s.s .
kamalov , and l. tiator , eur .
j. a * 34 * , 69 ( 2007 ) [ arxiv:0710.0306 [ nucl - th ] ] ; the maid analysis is available at http://www.kph.uni - mainz.de / maid/. | we derive expressions for pion electroproduction amplitudes in the @xmath0 expansion of qcd , and obtain from them linear relations between the electromagnetic multipole amplitudes that hold at all energies .
the leading - order relations in @xmath0 compare favorably with available data ( especially away from resonances ) , but the next - to - leading - order relations tend to provide only small or no improvement .
= 1 255=255 by 60 255 by-60255 by |
the celebrated nonlinear schrdinger ( nls ) equation @xmath0 is considered to be one of the fundamental integrable equations admitting an @xmath1soliton solution .
it is proved integrable via the inverse scattering transform @xcite .
the nls equation has applications in a wide variety of physical systems such as water waves @xcite , plasma physics @xcite and nonlinear optics @xcite .
this equation can be used to model short soliton pulses in optical fibres @xcite .
the basic phenomena are described by the nonlinear schrdinger equation , but as the pulses get shorter various additional effects become important . in light of this fact , kodama and hasegawa @xcite developed a suitable higher - order nls equation to take these additional effects .
their equation takes the form @xmath2=0,\end{aligned}\ ] ] where the @xmath3 , @xmath4 are real constants , @xmath5 is a real spectral parameter and @xmath6 is a complex valued function .
the first three terms ( setting @xmath7 and @xmath8 ) form the standard nonlinear schrdinger equation . in general , the kodama - hasegawa higher - order nls equation is not integrable unless some restrictions are imposed on the real constants @xmath4 @xmath9 . with appropriate choices of these real constants , the inverse scattering transform can be applied to verify integrability of the resulting equation .
it is known until now , along with the nls equation itself , there are four cases in which integrability can be proven via inverse scattering transform .
these are the chen - lee - liu @xcite derivative nls equation @xmath10 , the kaup - newell @xcite derivative nls equation @xmath11 , the hirota @xcite nls equation @xmath12 and the sasa - satsuma @xcite nls equation @xmath13 .
sasa and satsuma @xcite consider the case where @xmath14 and @xmath15 , that is @xmath16=0.\end{aligned}\ ] ] sasa and satsuma @xcite introduce variable transformations @xmath17 then , setting @xmath18 , the equation is reduced to a complex modified kdv - type equation @xmath19 which is an equivalent version of .
the equation is commonly known as the sasa - satsuma ( ss ) equation , and we will denote it as such from now on .
the ss equation is one of the integrable extensions of the nls equation .
the integrability of this equation has been widely studied with various methods such as the inverse scattering scheme @xcite , the hirota s bilinear approach @xcite and the _ darboux - like _ transformations @xcite . in the present paper ,
we present a systematic approach to the construction of by means of a standard binary darboux transformation ( bdt ) and written in terms of quasideterminants @xcite .
quasideterminants have various nice properties which play important roles in constructing exact solutions of integrable systems @xcite . in this paper , we establish for the first time a standard bdt for the ss equation . our solutions for the ss equation are written in terms of quasigrammians rather than determinants .
it should be emphasised that these quasigrammian solutions arise naturally from the binary darboux transformation we present here .
furthermore , we present periodic , one - soliton , two - solitons and breather solutions for the ss equation . this paper is organized as follows . in section [ qdet ]
below , we give a brief review on quasideterminants . in section [ evss ]
, we construct a @xmath20 eigenfunction and corresponding constant @xmath21 square matrix for the eigenvalue problems of the ss equation via a symmetry matrix . in section [ bdtss ] , we state a standard binary darboux theorem for the sasa - satsuma system . in section [ bdtrss ] , we review the reduced binary darboux transformations for the ss equation , which can be considered as a dimensional reduction from @xmath22 to @xmath23 dimensions . in sections [ qgss ] , we present the quasigrammian solutions of the ss equation by using the binary darboux transformation . here
, the quasigrammians are written in terms of solutions of linear eigenvalue problems . in section [ psss ] , periodic , one - soliton , two - solitons and breather solutions of the ss equation
are given for both zero and non - zero seed solutions .
the conclusion is given in the final section [ conss ] .
in this short section we recall some of the key elementary properties of quasideterminants .
the reader is referred to the original papers @xcite for a more detailed and general treatment .
quasideterminants were introduced by gelfand and retakh in @xcite as a natural generalisation of the determinant to matrices with noncommutative entries .
many equivalent definitions of quasideterminants exist , one such being a recursive definition involving inverse minors .
let @xmath24 be an @xmath25 matrix with entries over an , in general non - commutative , ring @xmath26 .
then the quasideterminants of @xmath27 for @xmath28 are defined by @xmath29 where @xmath30 is the row vector obtained from @xmath31 row of @xmath27 with the @xmath32 element removed , @xmath33 is the column vector obtained from @xmath32 column of @xmath27 with the @xmath31 element removed and @xmath34 is the @xmath35 submatrix obtained by deleting the @xmath31 row and the @xmath32 column from @xmath27 .
quasideterminants can also be denoted by boxing the entry about which the expansion is made @xmath36 if the entries in @xmath27 happen to commute , then the quasideterminant @xmath37 can be expressed as a ratio of determinants @xmath38 in this paper , we will consider only quasideterminants that are expanded about a term in the last column , most usually the last entry .
for example considering a block matrix @xmath39 , where @xmath40 is an invertible ( square ) matrix over @xmath26 of arbitrary size and @xmath41 , @xmath42 are column and row vectors over @xmath26 of compatible lengths , respectively , and @xmath43 , the quasideterminant of @xmath27 is expanded about @xmath44 is @xmath45 later we will use the following invariance of quasideterminants which follows immediately from their definition . let @xmath46 and @xmath47 be invertible matrices of the same dimensions as @xmath40
. then @xmath48
the lax pair @xcite for the sasa - satsuma equation is given by @xmath49 where @xmath50 , @xmath51 , @xmath52 and @xmath53 are @xmath54 matrices such that @xmath55 and @xmath56 here @xmath57 is a spectral parameter and asterisk denotes complex conjugate .
it can be seen that the potential matrix @xmath51 in has two symmetry properties @xcite .
one is that it is skew - hermitian : @xmath58 .
the other one is that @xmath59 , where @xmath60 let @xmath61 be a vector eigenfunction for - for eigenvalue @xmath57 . using the second symmetry ,
it is easy to see that @xmath62 is another eigenfunction , for eigenvalue @xmath63 . using these vector eigenfunctions
we can define a @xmath20 matrix eigenfunction @xmath64 with @xmath21 eigenvalue @xmath65 @xmath66 satisfying @xmath67
let us consider the linear operators @xmath68 where @xmath69 , where @xmath26 is a ring , in general non - commutative .
the standard approach to darboux transformations @xcite involves a gauge operator @xmath70 , where @xmath71 is a solution to a linear system @xmath72 where @xmath73 is any eigenfunction of @xmath74 and @xmath27 , has the property of leaving the above linear problems invariant : @xmath75 where @xmath76 is an eigenfunction of @xmath77 and @xmath78 and the linear operators @xmath79 and @xmath80 have the same forms as @xmath74 and @xmath27 : @xmath81 let @xmath82 and @xmath83 be two standard darboux transformations map two linear operators @xmath74 and @xmath84 onto a common linear operator @xmath77 such that @xmath85 then one may define a binary darboux transformation @xcite @xmath86 such that @xmath87 . in order to define @xmath83
one needs an eigenfunction of @xmath84 .
this problem can be got around by using the formal adjoint operator @xmath88 constructed according to the rule @xmath89 , where @xmath90 denotes the hermitian conjugate of @xmath91 . if @xmath92 and @xmath93 , we derive the eigenfunction @xmath94 as @xmath95 where the eigenfunction potential @xmath96 is defined such that @xmath97 .
we can then construct the binary darboux transformation explicitly as @xmath98 with adjoint @xmath99 it should be pointed out that the above formulas obtained for @xmath100 and @xmath101 , the transformation makes sense for any @xmath102 matrices @xmath64 and @xmath103 such that @xmath104 , not just square @xmath105 matrices . here , only the eigenfunction potential @xmath106 must be an invertible square matrix .
suppose @xmath107}=\phi$ ] is a general eigenfunction of the operators @xmath108}=l$ ] , @xmath109}=m$ ] and @xmath110}=\psi$ ] a general eigenfunction of the adjoint lax operators @xmath111}=l^{\dagger}$ ] , @xmath112}=m^{\dagger}$ ] , where @xmath74 and @xmath27 are given by .
we then define the binary darboux transformations of the eigenfunctions @xmath113 and @xmath114 as @xmath115}&=&b_{\theta,\rho}\left(\phi_{[1]}\right)=\phi_{[1]}-\theta_{[1]}\omega\left(\theta_{[1]},\rho_{[1]}\right)^{-1}\omega\left(\phi_{[1]},\rho_{[1]}\right),\\ \psi_{[2]}&=&b_{\theta,\rho}^{-\dagger}\left(\psi_{[1]}\right)=\psi_{[1]}-\rho_{[1]}\omega\left(\theta_{[1]},\rho_{[1]}\right)^{-\dagger}\omega\left(\theta_{[1]},\psi_{[1]}\right)^{\dagger},\end{aligned}\ ] ] with @xmath116}=\phi_{[2]}\vert_{\phi\rightarrow \theta_{2}},\hspace{0.5cm}\rho_{[2]}=\psi_{[2]}\vert_{\psi\rightarrow \rho_{2}}.\end{aligned}\ ] ] after @xmath117 iterations , the @xmath118th binary darboux transformation is given by @xmath119}&=&b_{\theta,\rho}\left(\phi_{[n]}\right)=\phi_{[n]}-\theta_{[n]}\omega\left(\theta_{[n]},\rho_{[n]}\right)^{-1}\omega\left(\phi_{[n]},\rho_{[n]}\right),\\ \psi_{[n+1]}&=&b_{\theta,\rho}^{-\dagger}\left(\psi_{[n]}\right)=\psi_{[n]}-\rho_{[n]}\omega\left(\theta_{[n]},\rho_{[n]}\right)^{-\dagger}\omega\left(\theta_{[n]},\psi_{[n]}\right)^{\dagger},\end{aligned}\ ] ] with @xmath120}=\phi_{[n]}\vert_{\phi\rightarrow \theta_{n}},\hspace{0.5cm}\rho_{[n]}=\psi_{[n]}\vert_{\psi\rightarrow \rho_{n}},\end{aligned}\ ] ] where @xmath121 and @xmath122 . using the notation @xmath123 and @xmath124 , we have @xmath119}= \begin{vmatrix}\omega(\theta , p ) & \omega(\phi , p)\\\theta & \boxed{\phi}\end{vmatrix},\hspace{0.2 cm } \psi_{[n+1]}=\begin{vmatrix}\omega(\theta , p)^{\dagger } & \omega(\theta,\psi)^{\dagger}\\p & \boxed{\psi}\end{vmatrix},\end{aligned}\ ] ] and @xmath125},\psi_{[n+1]})= \begin{vmatrix}\omega(\theta , p ) & \omega(\phi , p)\\ \omega(\theta , \psi ) & \boxed{\omega(\phi , \psi)}\end{vmatrix}.\end{aligned}\ ] ] here
, we describe a reduction of the binary darboux transformation from @xmath22 to @xmath23 dimensions .
we choose to eliminate the @xmath126-dependence by employing a ` separation of variables ' technique .
the reader is referred to the paper @xcite for a more detailed treatment .
we make the anstze @xmath127 where @xmath128 are constant scalars and @xmath129 are @xmath130 constant matrices and the superscript @xmath131 denotes reduced functions , independent of @xmath126 .
hence in the dimensional reduction we obtain @xmath132 and @xmath133 and so the operator @xmath74 in becomes @xmath134 where @xmath135 is a matrix eigenfunction of @xmath136 such that @xmath137 , with @xmath57 replaced by the matrix @xmath65 , that is , @xmath138 it follows that the @xmath139dependence of the potential @xmath96 can also be made explicit by setting @xmath140 and @xmath141 .
then the dimensionally reduced binary darboux transformations are written as @xmath142 where @xmath143 is an algebraic potential satisfying the following conditions @xmath144 the transformed operators @xmath145}^{r}=b_{\theta^{r},\rho^{r}}l_{[n]}^{r}b_{\theta^{r},\rho^{r}}^{-1},\\ m_{[n+1]}^{r}=b_{\theta^{r},\rho^{r}}m_{[n]}^{r}b_{\theta^{r},\rho^{r}}^{-1}\end{aligned}\ ] ] have generic eigenfunctions and adjoint eigenfunctions @xmath146}^{r}&=&\phi_{[n]}^{r}-\theta_{[n]}^{r}\omega^{r}\left(\theta_{[n]}^{r},\rho_{[n]}^{r}\right)^{-1}\omega^{r}\left(\phi_{[n]}^{r},\rho_{[n]}^{r}\right),\\ \psi_{[n+1]}^{r}&=&\psi_{[n]}^{r}-\rho_{[n]}^{r}\omega^{r}\left(\theta_{[n]}^{r},\rho_{[n]}^{r}\right)^{-\dagger}\omega^{r}\left(\theta_{[n]}^{r},\psi_{[n]}^{r}\right)^{\dagger},\end{aligned}\ ] ] with @xmath147}=\phi^{r}_{[n]}\vert_{\phi^{r}\rightarrow \theta_{n}},\hspace{0.5cm}\rho^{r}_{[n]}=\psi^{r}_{[n]}\vert_{\psi^{r}\rightarrow \rho^{r}_{n}}.\end{aligned}\ ] ] from now on , for notational simplicity , we omit the superscript @xmath131 .
in this section we determine the effect of the bdt @xmath148 on the operator @xmath149 given by with @xmath64 an eigenfunction of @xmath74 and @xmath103 an eigenfunction of @xmath88 .
corresponding results hold for the operator @xmath27 given by and its corresponding adjoint @xmath150 . in the operator @xmath74 , the matrix coefficients @xmath50 and @xmath51 are both skew - hermitian .
this leads us to observe the relation @xmath151 in which we can let @xmath152 and @xmath153 .
the operator @xmath74 is transformed to a new operator @xmath84 such that @xmath154 where @xmath155 .
we find that @xmath156.\end{aligned}\ ] ] for notational convenience , we introduce a @xmath157 matrix @xmath158 such that @xmath159 $ ] , and hence @xmath160 from , since @xmath159 $ ] , it follows that @xmath161 in which the relation is now @xmath162 with @xmath163 , where the eigenfunction @xmath64 and the diagonal constant matrix @xmath65 are given by .
let @xmath164}=p$ ] , @xmath165}=\hat{p}$ ] , @xmath166}=\theta_{1}=\theta$ ] and @xmath167 so that @xmath168 .
then , the solution can be rewritten as @xmath169}=p_{[1]}-\theta_{[1 ] } \omega(\theta_{[1]},\theta_{[1]})^{-1}\theta_{[1]}^{\dagger}.\end{aligned}\ ] ] after @xmath118 repeated applications of the reduced binary darboux transformation @xmath170 , we have @xmath171}=p_{[n]}-\theta_{[n ] } \omega(\theta_{[n]},\theta_{[n]})^{-1}\theta_{[n]}^{\dagger}\end{aligned}\ ] ] with @xmath172}=\phi_{[n]}\vert_{\phi\rightarrow \theta_{n}}$ ] , where the general eigenfunction @xmath173}$ ] is given by as @xmath174}=\phi_{[n]}-\theta_{[n]}\omega(\theta_{[n]},\theta_{[n]})^{-1}\omega(\phi_{[n]},\theta_{[n]}).\end{aligned}\ ] ] let @xmath175 be a particular set of eigenfunctions of the linear operators @xmath74 , @xmath27 given by @xmath176 , and define @xmath123 for the @xmath20 matrices @xmath177 @xmath178 such that @xmath179 we express @xmath180}$ ] and @xmath181}$ ] in quasi - grammian forms as @xmath182}= p+ \begin{vmatrix } \omega(\theta,\theta ) & \theta^{\dagger}\\ \theta & \boxed{\begin{matrix}0&0 & 0\\0&0&0\\0&0&0\end{matrix } } \end{vmatrix},\hspace{0.2 cm } \phi_{[n+1]}=\begin{vmatrix}\omega(\theta,\theta ) & \omega(\phi,\theta)\\ \theta & \boxed{\phi}\end{vmatrix}.\end{aligned}\ ] ] in order to express the quasi - grammian solution @xmath180}$ ] explicitly in terms of the variable @xmath183 , we define the @xmath184 matrix @xmath185 as @xmath186 where @xmath187 , @xmath188 and @xmath189 denote the row vectors @xmath190 , @xmath191 and @xmath192 respectively .
thus , we obtain @xmath193}=p+ \begin{pmatrix } \begin{vmatrix}\omega(\theta,\theta ) & \phi_{1}^{\dagger}\\ \phi_{1 } & \boxed{0}\end{vmatrix } & \begin{vmatrix}\omega(\theta,\theta ) & \phi_{2}^{\dagger}\\ \phi_{1 } & \boxed{0}\end{vmatrix } & \begin{vmatrix}\omega(\theta,\theta ) & \phi_{3}^{\dagger}\\ \phi_{1 } & \boxed{0}\end{vmatrix}\\\\ \begin{vmatrix}\omega(\theta,\theta ) & \phi_{1}^{\dagger}\\ \phi_{2 } & \boxed{0}\end{vmatrix } & \begin{vmatrix}\omega(\theta,\theta ) & \phi_{2}^{\dagger}\\ \phi_{2 } & \boxed{0}\end{vmatrix } & \begin{vmatrix}\omega(\theta,\theta ) & \phi_{3}^{\dagger}\\ \phi_{2 } & \boxed{0}\end{vmatrix}\\\\ \begin{vmatrix}\omega(\theta,\theta ) & \phi_{1}^{\dagger}\\ \phi_{3 } & \boxed{0}\end{vmatrix } & \begin{vmatrix}\omega(\theta,\theta ) & \phi_{2}^{\dagger}\\ \phi_{3 } & \boxed{0}\end{vmatrix } & \begin{vmatrix}\omega(\theta,\theta ) & \phi_{3}^{\dagger}\\ \phi_{3 } & \boxed{0}\end{vmatrix } \end{pmatrix}.\end{aligned}\ ] ] by substituting into , we have quasi - grammian expressions for @xmath183 and @xmath194 , namely @xmath195}=u+2i\begin{vmatrix}\omega(\theta,\theta ) & \phi_{3}^{\dagger}\\ \phi_{1 } & \boxed{0}\end{vmatrix},\label{ssqgr}\\ = u+2i\begin{vmatrix}\omega(\theta,\theta ) & \phi_{2}^{\dagger}\\ \phi_{3 } & \boxed{0}\end{vmatrix},\label{ssqgr2}\\ u_{[n+1]}^{*}=u^{*}+2i\begin{vmatrix}\omega(\theta,\theta ) & \phi_{1}^{\dagger}\\ \phi_{3 } & \boxed{0}\end{vmatrix},\label{ssqgr3}\\ = u^{*}+2i\begin{vmatrix}\omega(\theta,\theta ) & \phi_{3}^{\dagger}\\ \phi_{2 } & \boxed{0}\end{vmatrix}.\label{ssqgr4}\end{aligned}\ ] ] these form a quasi - grammian solution @xmath183 of the sasa - satsuma equation and its complex conjugate , however , it is necessary to show that these four expressions are consistent .
that is , that the expressions on the right hand sides of and are equal and that the pairs are indeed complex conjugates .
the proof of this is given in section [ prf ] . in the expressions , the potential @xmath196 is a @xmath197 matrix satisfying the relation @xmath198 where @xmath65 is @xmath197 constant matrix such that @xmath199 . solving this relation for @xmath196
, we obtain the explicit expression @xmath200 where @xmath201 is @xmath21 potential satisfying the relation @xmath202 where @xmath203 and @xmath204 .
it follows from this relation that the potential @xmath96 can be written explicitly as @xmath205 where @xmath206 and @xmath207 are the scalar functions @xmath208 here we observe that @xmath209 and @xmath210 are such that @xmath211 and @xmath212 , for @xmath213 .
then the @xmath214 potentials @xmath201 satisfy the symmetry condition @xmath215 and the @xmath216 matrix potential @xmath196 , as given by , is skew - adjoint , @xmath217 now it is readily seen that @xmath218 using , and so , , and similarly , , are indeed complex conjugate .
it remains to prove that and are consistent , i.e. that @xmath219 note first that this each side in this equation represents a scalar and so the right hand side can also be written as @xmath220 where @xmath221 denotes the matrix transpose .
now let @xmath46 be the @xmath216 permutation matrix @xmath222 pre(post)-multiplying any matrix with @xmath223 ( rows ) columns by @xmath46 has the effect of interchanging its @xmath224th and @xmath225th ( rows ) columns for @xmath226 .
hence @xmath227 and @xmath228 now using the quasideterminant invariance , we get @xmath229 this completes the proof .
in order to construct particular solutions for the sasa - satsuma equation , we consider the quasi - grammian solution given by @xmath230}=u+2i\begin{vmatrix}\omega(\theta,\theta ) & \phi_{3}^{\dagger}\\ \phi_{1 } & \boxed{0}\end{vmatrix},\end{aligned}\ ] ] where @xmath187 and @xmath189 denote the first and third rows repectively of a @xmath231 matrix eigenfunction @xmath185 , and the potential @xmath196 is the @xmath232 matrix with entries defined by and .
let us consider the spectral problem @xmath233 with eigenvalue @xmath234 @xmath235 , where @xmath236 and @xmath74 and @xmath27 are given by - so that @xmath237 where @xmath50 , @xmath51 , @xmath52 and @xmath53 are given by - .
+ @xmath238 + in this case @xmath239 is a solution of the spectral problem @xmath233 with eigenvalue @xmath240 .
thus , from , we derive the following explicit solution @xmath241}=u+2i \begin{vmatrix } f_{11 } & -g_{11}^ { * } & \phi_{3}^{*}\\ g_{11 } & f_{11 } & \phi_{3}\\ \phi_{1 } & \phi_{2}^ { * } & \boxed{0 } \end{vmatrix},\end{aligned}\ ] ] where @xmath242 for @xmath243 , the above eigenvalue problems - transform into the first - order linear system @xmath244 which has solution @xmath245 , where @xmath246 + @xmath238 then , the solution becomes @xmath247}=2i \begin{vmatrix } f & -g^ { * } & \psi_{1}^{*}\\ g & f & \psi_{1}\\ \phi_{1 } & \phi_{1}^ { * } & \boxed{0 } \end{vmatrix}\end{aligned}\ ] ] which can be written as @xmath247}=-2i\left(\frac{\lambda_{1}+\lambda_{1}^{*}}{\lambda_{1}-\lambda_{1}^{*}}\right)\frac{g+g^{*}}{f^{2}+|g|^{2}},\end{aligned}\ ] ] where @xmath248 here the eigenfunction @xmath245 in which @xmath249 by setting @xmath250 and @xmath251 , we have @xmath247}=-2\frac{\xi(g+g^{*})}{\eta(f^{2}+|g|^{2})},\end{aligned}\ ] ] in which @xmath252 where @xmath253t\right)$ ] and @xmath254t\right)$ ] .
then we obtain a breather solution @xmath255}=8\xi\eta\hspace{0.1cm}\frac{\xi\left[3\cosh(2\alpha)+\sinh(2\alpha)\right]\cos(2\beta ) -\eta\left[\cosh(2\alpha)+3\sinh(2\alpha)\right]\sin(2\beta ) } { \xi^{2}\left[3\cosh(2\alpha)+\sinh(2\alpha)\right]^{2}+8\eta^{2}\sin^{2}(2\beta)}.\end{aligned}\ ] ] the solution is plotted in the figure [ fig:1 ]
. breather solution @xmath256}$ ] of the ss equation with the choice of parameters @xmath257 . ] for @xmath258 , @xmath259 is a solution of the sasa - satsuma equation , where we let @xmath260 to be a real constant .
we use this as a seed solution for application of binary darboux transformations . substituting @xmath259 into the linear system - and then solving for the eigenfunction @xmath261
, we obtain @xmath262}+c_{3}e^{-i\sqrt{\lambda_{j}^{2}+2k^{2}}\left[x+4\left(\lambda_{j}^{2}-k^{2}\right)t\right]},\\ \phi_{3j-1}&=&\frac{1}{2}c_{1}e^{-i\lambda_{j}\left(x+4\lambda_{j}^{2}t\right)}+c_{2}e^{i\sqrt{\lambda_{j}^{2}+2k^{2}}\left[x+4\left(\lambda_{j}^{2}-k^{2}\right)t\right]}+c_{3}e^{-i\sqrt{\lambda_{j}^{2}+2k^{2}}\left[x+4\left(\lambda_{j}^{2}-k^{2}\right)t\right]},\\ \phi_{3j}&=&k_{1}e^{i\sqrt{\lambda_{j}^{2}+2k^{2}}\left[x+4\left(\lambda_{j}^{2}-k^{2}\right)t\right]}+k_{2}e^{-i\sqrt{\lambda_{j}^{2}+2k^{2}}\left[x+4\left(\lambda_{j}^{2}-k^{2}\right)t\right]},\end{aligned}\ ] ] where @xmath263 , @xmath264 and @xmath265 , @xmath266 , @xmath267 are arbitrary constants .
+ + @xmath238 the solution can be written as @xmath268}=k-2i\frac{\left(\phi_{1}\phi_{3}^{*}+\phi_{2}^{*}\phi_{3}\right)f_{11}+\phi_{1}\phi_{3}g_{11}^{*}-\phi_{2}^{*}\phi_{3}^{*}g_{11}}{f_{11}^{2}+\left|g_{11}\right|^{2}},\end{aligned}\ ] ] where @xmath269 and @xmath270 are given by - . substituting @xmath271 and @xmath272 into
, we obtain @xmath273}=k+4\mu\frac{(|\phi_{1}|^{2}-|\phi_{2}|^{2})\left(\phi_{1}\phi_{3}^{*}-\phi_{2}^{*}\phi_{3}\right)}{(|\phi_{1}|^{2}-|\phi_{2}|^{2})^{2}+2\left|\phi_{1}\phi_{3}^{*}-\phi_{2}^{*}\phi_{3}\right|^{2}},\end{aligned}\ ] ] where @xmath274}+c_{3}e^{-id\left[x-4\left(k^{2}+\mu^{2}\right)t\right]},\\ \phi_{2}&=&\phi_{1}+c_{1}e^{\mu\left(x-4\mu^{2}t\right)},\\ \phi_{3}&=&-\frac{c_{2}}{k}(\mu - id ) e^{id\left[x-4\left(k^{2}+\mu^{2}\right)t\right]}-\frac{c_{3}}{k}(\mu+id)e^{-id\left[x-4\left(k^{2}+\mu^{2}\right)t\right]},\end{aligned}\ ] ] in which @xmath275 . here , for simplicity , we have chosen @xmath276 such that @xmath277 . for the case @xmath279 ,
can be written as @xmath280}=k-2k\mu \frac{c^{2}k e^{2i\alpha}+{c^{*}}^{2 } k^{*}e^{-2i\alpha}+2\mu|c|^{2}-2ks(ce^{i\alpha}+c^ { * } e^{-i\alpha})e^{-\beta}}{\mu(c^{2}k e^{2i\alpha}+{c^{*}}^{2 } k^{*}e^{-2i\alpha } ) + 4k^{2}|c|^{2}+2k^{2}|s|^{2 } e^{-2\beta}},\end{aligned}\ ] ] where @xmath281 $ ] , @xmath282 , @xmath283 , @xmath284 and @xmath285 . by choosing @xmath286 so that @xmath287 , we obtain a periodic solution @xmath247}=-k+2\frac{k(2k^{2}-\mu^{2})}{\mu\left[\mu\cos(2\alpha)+\sqrt{2k^{2}-\mu^{2}}\sin(2\alpha)\right]+2k^{2}}.\end{aligned}\ ] ] this solution is plotted in the figure [ fig:2 ] .
periodic solution @xmath256}$ ] of the ss equation with the choice of parameters @xmath288 . ]
if we choose @xmath289 then @xmath290 and @xmath291 .
the solution can be written in the following form @xmath247}=k-2k\mu \frac{\mu\cos(2\alpha)+d\sin(2\alpha)+\mu-4id[\cosh\beta -\sinh\beta]\cos\alpha } { \mu[\mu\cos(2\alpha)+d\sin(2\alpha)]+2k^{2}+4d^{2}[\cosh(2\beta ) -\sinh(2\beta)]}.\end{aligned}\ ] ] this solution is plotted in the figure [ fig:3 ] .
solution @xmath292}|^{2}$ ] of the ss equation with the choice of parameters @xmath293 . ] for the case @xmath295 , can be written as @xmath296}=k-2\mu k \frac{\mu ( e^{2\gamma}+ e^{-2\gamma})-e ( e^{2\gamma}-e^{-2\gamma})+2\mu+4e h_{1}^{-1}h_{2}(e^{\gamma}+e^{-\gamma})e^{-\beta}}{\mu^{2 } ( e^{2\gamma}+ e^{-2\gamma})-\mu e ( e^{2\gamma}-e^{-2\gamma})+4 k^{2}-8 e^{2}h_{1}^{-2}h_{2}^{2 } e^{-2\beta}},\end{aligned}\ ] ] where @xmath297 , @xmath298 $ ] , @xmath282 and @xmath299 , @xmath300 such that @xmath301 and @xmath302 . by choosing @xmath286 so that @xmath303 , we obtain @xmath247}=-k-2 \frac{k(\mu^{2}-2k^{2})}{\mu^{2}\cosh(2\gamma)-\mu\sqrt{\mu^{2}-2k^{2}}\sinh(2\gamma)+2k^{2}}.\end{aligned}\ ] ] this is one - soliton solution .
this solution is plotted in the figure [ fig:4 ] .
one - soliton solution @xmath292}|^{2}$ ] of the ss equation with the choice of parameters @xmath304 . ] if we choose @xmath305 so that @xmath306
. then the solution can be written as @xmath247}=-k-2ke \frac{e ( 1 - 4 e^{-2\beta})+4 i \mu \cosh \gamma e^{-\beta}}{\mu^{2}\cosh(2\gamma)-\mu e\sinh(2\gamma)+2k^{2}+4 e^{2 } e^{-2\beta}}.\end{aligned}\ ] ] this is two - soliton solution .
this solution is plotted in the figure [ fig:5 ] .
two - soliton solution @xmath292}|^{2}$ ] of the ss equation with the choice of parameters @xmath307 . ]
in this paper , we have presented a standard binary darboux transformation for the ss equation and using this we have constructed a wide family of solutions in quasigrammian form . these quasigrammians are expressed in terms of solutions of the linear partial differential equations given by - . moreover ,
periodic , one - soliton , two - solitons and breather solutions for zero and non - zero seeds have been given as particular examples for the ss equation .
examples of these solutions are plotted in the figures [ fig:1][fig:5 ] for particular choices of parameters .
one should notice that we have chosen @xmath259 , where @xmath260 is a real constant , as a seed solution of the ss equation .
this is the simplest non - zero seed .
however , one might also choose the seed @xmath308t\right)}$ ] , where @xmath309 , to construct various rich explicit solutions including those we present here .
furthermore , in the present paper , we only consider the case @xmath310 for constructing explicit solutions .
it can be obtained more explicit solutions by considering other cases such as @xmath311 .
finally , it should be pointed out that the binary darboux transformation technique is a universal instrument that allows us to construct exact solutions for other integrable systems .
d. j. benney and a. c. newell , the propagation of nonlinear wave envelopes , _ j. math . phys_. * 46 * , 133139 ( 1967 ) .
d. j. benney and g. j. roskes , wave instabilities , _ stud . appl . math_. * 48 * , 377385 ( 1969 ) .
h. h. chen , y. c. lee and c. s. liu , integrability of nonlinear hamiltonian systems by inverse scattering method , _ physica scripta _ * 20 * : 490492 ( 1979 ) .
g. darboux , comptes rendus de lacadmie des sciences * 94 * , 14569 ( 1882 ) .
i. gelfand and v. retakh , determinants of the matrices over noncomutative rings , funct .
app . * 25 * , 91102 ( 1991 ) .
i. gelfand , s. gelfand , v. retakh and r. l. wilson , quasideterminants , adv . math . * 193 * , 56141 ( 2005 ) . c. gilson , j. hietarinta , j. nimmo and y. ohta , sasa - satsuma higher - order nonlinear schrdinger equation and its bilinearization and multisoliton solutions , phys .
e , * 68 * 016614 ( 2003 ) .
s. ghosh , a. kundu and s. nandy , soliton solutions , liouville integrability and gauge equivalence of sasa satsuma equation , j. math .
phys * 40 * , 1993 ( 1999 ) . b. haider and m. hassan , quasi - grammian solutions of the generalized coupled dispersionless integrable system , sigma * 8 * 084 ( 2012 ) .
m. hamanaka , noncommutative solitons and quasideterminants , _ phys . scr_.
* 89 * 038006 ( 2014 ) .
a. hasegawa and f. tappert , transmission of stationary nonlinear optical pulses in dispersive dielectric fibres i. anomalous dispersion , _ appl .
phys . lett_.
* 23 * 142 ( 1973 ) .
a. hasegawa and f. tappert , transmission of stationary nonlinear optical pulses in dispersive dielectric fibres ii .
normal dispersion , _ appl .
phys . lett_.
* 23 * 171 ( 1973 ) .
r. hirota , exact envelope - soliton solutions of a nonlinear wave equation ,
_ j. math .
phys_. * 14 * : 805809 ( 1973 ) .
d. j. kaup and a. c. newell , an exact solution for a derivative nonlinear schrdinger equation , _ j. math
_ * 19*(4 ) : 798801 ( 1978 ) .
d.j . kaup and j. yang , the inverse scattering transform and squared eigenfunctions for a degenerate @xmath54 operator , inverse problems * 25 * 105010 ( 2009 ) .
y. kivshar and g. agrawal , _ optical solitons : from fibers to photonic crystals _ , academic press ( 2003 )
. y. kodama , optical solitons in a monomode fiber , _ j. stat .
* 39 * 5/6 ( 1985 ) .
y. kodama and a. hasegawa , nonlinear pulse propagation in a monomode dielectric guide , _ ieee j. quantum electron .
_ * qe-23 * 510 ( 1987 ) .
li and j.j.c .
nimmo , darboux transformations for a twisted derivation and quasideterminant solutions to the super kdv equation , _ proc .
r. soc . a _ * 466 * , 24712493 ( 2009 ) .
v. b. matveev , darboux transformation and explicit solutions of the kadomtcev - petviaschvily equation , depending on functional parameters , lett .
* 3 * , 21316 ( 1979 ) .
v.b . matveev and m.a .
salle , darboux transformations and solitons , springer series in nonlinear dynamics , springer - verlag , berlin ( 1991 ) .
nimmo and h. yilmaz , on darboux transformations for the derivative nonlinear schrdinger equation , journal of nonlinear mathematical physics *
21*(2 ) ( 2014 ) 278293 .
j. j. c. nimmo , c. r. gilson and y. ohta , applications of darboux transformations to the self - dual yang - mills equations , theor .
phys . * 122 * 23946 ( 2000 ) .
n. sasa and j. satsuma , new type of soliton solutions for a higher - order nonlinear schrdinger equation , j. phys .
japan * 60 * 40917 ( 1991 ) .
t. xu , d. wang , m. li and h. liang , soliton and breather solutions of the sasa
satsuma equation via the darboux transformation , phys .
* 89 * 075207 ( 2014 ) .
j. yang and d.j .
kaup , squared eigenfunctions for the sasa
satsuma equation , j. math . phys . *
50 * 023504 ( 2009 ) .
h. yilmaz , exact solutions of the gerdjikov - ivanov equation using darboux transformations , journal of nonlinear mathematical physics *
22*(1 ) ( 2015 ) . v. e. zakharov and a. b. shabat , exact theory of two - dimensional self - focusing and one - dimensional self - modulation of waves in nonlinear media , _ zh . eksp .
fiz_. * 61 * 118134 ( 1971 ) [ _ sov .
jetp _ * 34 * 6269 ( 1972 ) ] .
v. e. zakharov , stability of periodic waves of finite amplitude on the surface of a deep fluid , _ sov .
mech . tech_. * 4*,190194 ( 1968 ) .
v. e. zakharov , collapse of langmuir waves , _ sov .
. jetp _ * 35 * , 908914 ( 1972 ) . | the sasa - satsuma equation is an integrable higher - order nonlinear schrdinger equation .
higher - order and multicomponent generalisations of the nonlinear schrdinger equation are important in various applications , e.g. , in optics .
one of these equations is the sasa - satsuma equation .
we present the binary darboux transformations for the sasa - satsuma equation and then construct its quasigrammians solutions by iterating its binary darboux transformations .
periodic , one - soliton , two - solitons and breather solutions are given as explicit examples . |
in recent years , many efforts have been made to understand the effect of breakup of weakly bound nuclei during the fusion reaction in very asymmetric reactions where the capture cross section is equal to the complete fusion cross section @xcite .
the light radioactive nuclei , especially halo nuclei , such as @xmath3he , @xmath4b , @xmath5be , and the stable nuclei @xmath6li and @xmath7be are weakly bounded , hence there is a chance of the breakup in the colliding process . by performing a comparison of fusion data with theoretical predictions which do not take into account the dynamic breakup plus transfer channel effects , it has been shown @xcite , that for energies from about @xmath8 to @xmath9 ( @xmath10 is the height of the coulomb barrier ) complete fusion in the reactions @xmath6li+@xmath11pb,@xmath12bi and @xmath7be+@xmath13y,@xmath14sn,@xmath11pb,@xmath12bi is suppressed by about 30% .
however , the @xmath7be+@xmath15sm data is out of the systematics , showing a much smaller suppression of about 15% . the total fusion ( incomplete fusion + sequential complete fusion + complete fusion ) cross section for the same projectiles on targets of any mass , including @xmath7be + @xmath16al,@xmath17zn , does not seem to be affected by the dynamic breakup and transfer effects @xcite . as the charge of the target decreases , one expects that the coulomb breakup becomes weaker , and consequently the complete fusion suppression and incomplete fusion probability decrease .
the lack of a clear systematic behavior of the complete fusion suppression as a function of the target charge was explained in ref .
@xcite by different effects of the transfer channels on the complete fusion and by some problems with the experimental data analysis . in the present article
we try to reveal a systematic behavior of the complete fusion suppression as a function of the target charge @xmath18 and colliding energy @xmath19 by using the quantum diffusion approach @xcite and by comparing the calculated capture cross sections in the absence of breakup with the experimental complete and total fusion cross sections .
the effects of deformation and neutron transfer on the complete fusion are taken into consideration .
in the quantum diffusion approach @xcite the collision of nuclei is described with a single relevant collective variable : the relative distance between the colliding nuclei .
this approach takes into account the fluctuation and dissipation effects in collisions of heavy ions which model the coupling with various channels ( for example , coupling of the relative motion with low - lying collective modes such as dynamical quadrupole and octupole modes of the target and projectile nuclei @xcite ) .
we like to mention that many quantum - mechanical and non - markovian effects accompanying the passage through the potential barrier are considered in our formalism @xcite .
the nuclear deformation effects are taken into account through the dependence of the nucleus - nucleus potential on the deformations and mutual orientations of the colliding nuclei . to calculate the nucleus - nucleus interaction potential @xmath20 , we use the procedure presented in refs . @xcite . for the nuclear part of the nucleus - nucleus potential , a double - folding formalism with a skyrme - type density - dependent effective nucleon - nucleon interaction
is used . within this approach
many heavy - ion capture reactions with stable and radioactive beams at energies above and well below the coulomb barrier have been successfully described @xcite .
one should note that other diffusion models , which include the quantum statistical effects , were also proposed in @xcite .
we assume that the sub - barrier capture mainly depends on the optimal one - neutron ( @xmath21 ) or two - neutron ( @xmath22 ) transfer with a positive @xmath23-value .
our assumption is that , just before the projectile is captured by the target - nucleus ( just before the crossing of the coulomb barrier ) which is a slow process , the transfer occurs that can lead to the population of the first excited collective state in the recipient nucleus @xcite .
so , the motion to the @xmath24 equilibrium starts in the system before the capture because it is energetically favorable in the dinuclear system in the vicinity of the coulomb barrier .
for the reactions under consideration , the average change of mass asymmetry is connected to the one- or two - neutron transfer . since after the transfer the mass numbers , the isotopic composition and the deformation parameters of the interacting nuclei , and , correspondingly , the height @xmath25 and shape of the coulomb barrier are changed , one can expect an enhancement or suppression of the capture . when the isotopic dependence of the nucleus - nucleus interaction potential is weak and the deformations of the interacting nuclei
after the transfer have not changed , there is no effect of the neutron transfer on the capture cross section .
this scenario was verified in the description of many reactions in ref .
all calculated results are obtained with the same set of parameters as in ref .
we use the friction coefficient in the relative distance coordinate which is close to that calculated within the mean field approaches @xcite .
the absolute values of the quadrupole deformation parameters @xmath26 of even - even deformed nuclei are taken from ref .
@xcite . for the nuclei deformed in the ground state ,
the @xmath26 in the first excited collective state is similar to the @xmath26 in the ground state . for the quadruple deformation parameter of an odd nucleus ,
we choose the maximal value of the deformation parameters of neighboring even - even nuclei . for the double magic and neighboring nuclei , in the ground state we set @xmath27 .
there are uncertainties in the definition of the values of @xmath26 in light - mass nuclei .
however , these uncertainties weakly influence the capture cross sections in the asymmetric reactions treated . in the calculations for light nuclei we use @xmath26 from ref .
@xcite . in figs .
1 - 13 we compare the calculated @xmath28 capture cross sections with the experimental @xmath29 complete and total fusion cross sections in the reactions induced by projectiles @xmath7be , @xmath30b , @xmath0li , @xmath31he @xcite .
the difference between the capture cross section and the complete fusion cross section can be ascribed to the breakup effect . comparing @xmath28 and @xmath29
, one can estimate the breakup probability @xmath32 if at some energy @xmath33 , the values of @xmath28 was normalized so to have @xmath34 at any energy .
note that @xmath35 contains the contribution from two processes : the direct fusion of the projectile with the target ( @xmath36 ) , and the breakup of the projectile followed by the fusion of the two projectile fragments with the target ( @xmath37 ) .
a more adequate estimate of the breakup probability would then be : @xmath38 which leads to larger values of @xmath39 than the expression employed by us .
however , the ratio between @xmath36 and @xmath37 can not be measured experimentally , but can be estimated with the approach suggested in refs .
the parameters of the potential are taken to fit the height of the coulomb barrier obtained in our calculations .
the parameters of the breakup function @xcite are set to describe the value of @xmath29 . as shown in ref .
@xcite and in our calculations , in the @xmath4be+@xmath11pb reaction the fraction of @xmath37 in @xmath29 does not exceed few percents at @xmath404 mev .
this fraction rapidly increases and reaches about 1220% , depending on the reaction , at @xmath4110 mev .
because we are mainly interested in the energies near and below the barrier , the estimated @xmath37 does not exceed 20% of @xmath29 at @xmath4210 mev .
the results for @xmath39 are presented taking @xmath36 into account in eq .
( 2 ) . as seen in figs .
14 and 15 , at energies above the coulomb barriers the values of @xmath39 vary from 0 to 84% . in the reactions
@xmath43be+@xmath15sm,@xmath11pb,@xmath12bi the value of @xmath39 increases with charge number of the target at @xmath44 mev .
this was also noted in ref .
however , the reactions @xmath43be+@xmath13y,@xmath14sn are out of this systematics . in the reactions @xmath3li+@xmath15sm,@xmath45pt,@xmath12bi the value of @xmath39 decreases with increasing charge number of the target at @xmath44 mev .
while in the reactions @xmath43be+@xmath13y,@xmath15sm,@xmath11pb,@xmath12bi the value of @xmath39 has a minimum at @xmath46 and a maximum at @xmath47 mev , in the @xmath43be+@xmath14sn reaction the value of @xmath39 steady decreases with energy . in the reactions @xmath3li+@xmath15sm,@xmath45pt,@xmath12bi , @xmath48li+@xmath11pb,@xmath12bi , and
@xmath43li+@xmath11pb there is maximum of @xmath39 at @xmath49 mev .
however , in the reactions @xmath3li+@xmath11pb and @xmath48li+@xmath50ho @xmath39 has a minima @xmath51 mev and no maxima at @xmath46 . for @xmath43be ,
the breakup threshold is slightly larger than for @xmath3li .
therefore , we can not explain a larger breakup probability at smaller @xmath52 in the case of @xmath43be . in figs . 1 - 13 we also show the calculated capture cross sections normalized by some factors to obtain a rather good agreement between the experimental and theoretical results
. these average normalization factors are 0.7 , 0.75 , 0.9 , 0.64 , 0.7 , 1 , 0.9 for the reactions @xmath43be+@xmath12bi,@xmath11pb,@xmath15sm,@xmath14sn,@xmath13y,@xmath17zn,@xmath16al , respectively , 0.52 , 0.5 , 0.5 , 0.42 , 0.65 for the reactions @xmath3li+@xmath12bi,@xmath11pb,@xmath45pt,@xmath15sm,@xmath17zn , respectively , 0.6 , 0.7 , 0.65 , 0.6 , 0.65 , 0.75 for the reactions @xmath48li+@xmath12bi,@xmath53au,@xmath50ho,@xmath54tb,@xmath17zn,@xmath16al , respectively . for the reactions @xmath43li+@xmath11pb ( fig .
10 ) , @xmath3he+@xmath12bi ( fig .
11 ) , @xmath3he+@xmath17zn ( fig .
11 ) , @xmath55he+@xmath53au ( fig .
12 ) , @xmath56he+@xmath53au ( fig .
12 ) , @xmath5b+@xmath12bi ( fig .
13 ) and @xmath5b+@xmath54 tb ( fig .
13 ) , a satisfactory agreement between experimental fusion data and capture cross sections can be reached with average normalization factors 0.6 , 0.68 , 0.4 , 0.8 , 0.7 , 0.82 , and 0.95 , respectively .
note that these average normalization factors do not depend on @xmath19 . with the @xmath7be projectile
we obtain the complete fusion suppressions similar to those reported in refs . @xcite . for lighter targets ,
when the coulomb breakup becomes weaker , one expects that the suppression of complete fusion becomes smaller than for heavy targets .
an expected behavior for complete fusion suppression is that fusion probability increases with decreasing @xmath18 .
however , one can observe deviations from this rule . in the reactions
@xmath7be+@xmath14sn,@xmath13y the data show quite larger complete fusion suppression ( 3036)% . for the reactions induced by a @xmath55li projectile , one can see that the fusion suppression is nearly independent of @xmath18 .
the replacement of @xmath57li by @xmath55li in the reactions in fig .
9 almost does not change the experimental @xcite and calculated data .
the coulomb fields for very light systems @xmath43be+@xmath16al , and @xmath3he,@xmath48li+@xmath17zn are not strong enough to produce an appreciable breakup .
it is not realistic that the fusion suppression in the @xmath43be+@xmath17zn reaction is smaller than the one in the @xmath43be+@xmath16al reaction or the suppressions of fusion coincide in the reactions @xmath58he+@xmath17zn ( fig .
11 ) with stable and exotic projectiles .
note that the experimental data for the reactions @xmath6li+@xmath16al,@xmath17zn and @xmath7be+@xmath16al,@xmath59zn are for the total fusion .
in general , the total fusion does not seem to be affected by breakup @xcite .
so , there is a lack of a systematic behavior of the complete fusion suppression for the systems treated .
the possible explanation of it is that there are probably some problems with the data analysis which were earlier noted in refs .
@xcite from the point of view of a universal fusion function representation .
it could be also that at energies near the coulomb barrier the characteristic time of the breakup is larger than the characteristic time of the capture process and influences the complete fusion .
for the reactions @xmath6li+@xmath11pb , the characteristic times of the prompt and delayed breakup were studied recently in ref .
@xcite .
the large positive @xmath60-value in the @xmath43li+@xmath11pb reaction @xcite gives a possibility of a two - neutron transfer before the capture .
however , the capture cross sections calculated with and without neutron transfer are very close to each other because the effect of neutron transfer is rather weak in asymmetric reactions @xcite . the calculated capture
cross sections normalized by a factor of 0.6 are shown by the dotted line in the lower part of fig .
10 . in the upper part of fig .
10 , the predicted capture cross sections for the reaction @xmath5li+@xmath11pb are shown .
the lack of a clear systematic behavior of the complete fusion suppression as a function of the target charge requires new additional experimental and theoretical studies .
the quasielastic backscattering has been used @xcite as an alternative to investigate fusion ( capture ) barrier distributions , since this process is complementary to fusion . since the quasielastic experiment is usually not as complex as the capture ( fusion ) and breakup measurements , they are well suited to survey the breakup probability .
there is a direct relationship between the capture , the quasielastic scattering and the breakup processes , since any loss from the quasielastic and breakup channel contributes directly to capture ( the conservation of the reaction flux ) : @xmath61 where @xmath62 is the reflection quasielastic probability , @xmath63 is the breakup ( reflection ) probability , and @xmath64 is the capture ( transmission ) probability .
the quasielastic scattering is the sum of all direct reactions , which include elastic , inelastic , and transfer processes .
equation ( 3 ) can be rewritten as @xmath65 where @xmath66 and @xmath67 are the quasielastic and capture probabilities , respectively , in the absence of the breakup process . from these expressions
we obtain the useful formulas @xmath68 using eqs .
( 3 ) and ( 5 ) , we obtain the relationship between breakup and quasielastic processes : @xmath69=1-p_{qe}(e_{\rm c.m.},j)[1 + 1/a ] \nonumber\\ & = & 1-p_{qe}(e_{\rm c.m.},j)/p_{qe}^{nbu}(e_{\rm c.m.},j).\end{aligned}\ ] ] the last equation is one of important results of the present paper .
analogously one can find other expression @xmath70 which relates the breakup and capture processes . the reflection quasielastic probability @xmath71 for bombarding energy @xmath19 and angular momentum @xmath72
is given by the ratio of the quasielastic differential cross section @xmath73 and rutherford differential cross section @xmath74 at 180 degrees @xcite . employing eqs .
( 6 ) , ( 8) , and the experimental quasielastic backscattering data with toughly and weakly bound isotopes - projectiles and the same compound nucleus , one can extract the breakup probability of the exotic nucleus .
for example , using eq .
( 6 ) at @xmath72 and the experimental @xmath75[@xmath76he+@xmath11pb ] of the @xmath76he+@xmath11pb reaction with toughly bound nuclei ( without breakup ) and @xmath62[@xmath55he+@xmath77pb ] of the @xmath55he+@xmath77pb reaction with weakly bound projectile ( with breakup ) , and taking into consideration @xmath10(@xmath76he+@xmath11pb)@xmath78(@xmath55he+@xmath77pb ) for the very asymmetric systems , one can extract the breakup probability of the @xmath55he : @xmath79}{p_{qe}^{nbu}(e_{\rm c.m.},j=0)[^{4}he+^{208}pb]}.\end{aligned}\ ] ] comparing the experimental quasielastic backscattering cross sections in the presence and absence of breakup data in the reaction pairs @xmath55he+@xmath80zn and @xmath76he+@xmath81zn , @xmath55he+@xmath82sn and @xmath76he+@xmath14sn , @xmath55he+@xmath83u and @xmath76he+@xmath84u , @xmath56he+@xmath85pb and @xmath76he+@xmath11pb , @xmath7be+@xmath11pb and @xmath86be+@xmath87pb , @xmath5be+@xmath77pb and @xmath86be+@xmath87pb , @xmath56b+@xmath11pb and @xmath86b+@xmath77pb , @xmath56b+@xmath87pb and @xmath5b+@xmath85pb , @xmath7b+@xmath11pb and @xmath5b+@xmath77pb , @xmath88c+@xmath87pb and @xmath89c+@xmath11pb , @xmath90f+@xmath77pb and @xmath91f+@xmath11pb leading to the same corresponding compound nuclei , one can analyze the role of the breakup channels in the reactions with the light weakly bound projectiles @xmath2he , @xmath1be , @xmath92b , @xmath88c , and @xmath90f at near and below barrier energies .
one concludes that the quasielastic technique could be a very important tool in breakup research .
we propose to extract the breakup probability directly from the quasielastic cross sections of systems mentioned above .
comparing the calculated capture cross sections in the absence of breakup data and experimental complete fusion data , we analyzed the role of the breakup channels in the reactions with the light projectiles @xmath7be , @xmath0li and @xmath2he at near - barrier energies . within the quantum diffusion approach
the neutron transfer and deformation effects were taken into account . analyzing the extracted breakup probabilities
, we showed that there are no systematic trends of breakup in the reactions studied .
moreover , for some system with larger ( smaller ) @xmath18 we found the contribution of breakup to be smaller ( larger ) . almost for all reactions considered we obtained a satisfactory agreement between calculated
capture cross section and experimental fusion data , if the calculated capture cross section or the experimental fusion data are renormalized by some average factor which does not depend on the bombarding energy .
note that our conclusions coincide with those of refs .
@xcite , where the universal fusion function formalism was applied for the analysis of experimental data .
one needs to measure directly the breakup process in different systems , especially light ones , to understand the role of the coulomb breakup in the complete fusion process .
the other important subject to be investigated both experimental and theoretically is the characteristic time of the breakup .
the first steps in these directions were done in refs .
@xcite .
as shown , one no needs to measure directly the breakup process in different systems , especially light ones , to understand the role of the breakup in the capture ( complete fusion ) process . employing the experimental quasielastic backscattering data with weakly and toughly bound isotopes of light nucleus and eq . ( 6 ) , the dependence of breakup probability on @xmath19 can be extracted for the systems suggested . analyzing the extracted breakup probabilities , one can indirectly study the trends of breakup in the different reactions at energies near and below coulomb barrier .
we thank p.r.s . gomes for fruitful discussions and useful suggestions . c.a .
bertulani , epj web conf .
* 17 * , 15001 ( 2011 ) ; p.r.s .
gomes , j. lubian , and l.f .
canto , epj web conf .
* 17 * , 11003 ( 2011 ) , p.r.s .
gomes _ et al .
_ , epj web conf .
( 2012 ) in print .
canto , p.r.s .
gomes , r. donangelo , and m.s .
hussein , phys .
rep . * 424 * , 1 ( 2006 ) .
n. keeley , r. raabe , n. alamanos , and j.l .
sida , prog .
. phys . * 59 * , 579 ( 2007 ) .
canto , p.r.s .
gomes , j. lubian , l.c .
chamon , and e. crema , j. phys .
g * 36 * , 015109 ( 2009 ) .
canto , p.r.s .
gomes , j. lubian , l.c .
chamon , and e. crema , nucl . phys . *
a821 * , 51 ( 2009 ) .
gomes , j. lubian , and l.f .
canto , phys .
c * 79 * , 027606 ( 2009 ) .
r. rafiei _ et al .
c * 81 * , 024601 ( 2010 ) .
de faria , r. lichtenthler , k.c.c .
pires , a.m. moro , a. lpine - szily , v. guimares , d.r .
mendes , jr .
, a. arazi , a. barioni , v. morcelle , and m.c .
morais , phys .
c * 82 * , 034602 ( 2010 ) ; p. mohr , p.n .
de faria , r. lichtenthler , k.c.c .
pires , v. guimares , a. lpine - szily , d.r .
mendes , jr .
, a. arazi , a. barioni , v. morcelle , and m.c .
morais , phys .
c * 82 * , 044606 ( 2010 ) ; a. lpine - szily _ et al .
_ , epj web conf .
( 2012 ) in print .
gomes , r. linares , j. lubian , c.c .
lopes , e.n .
cardozo , b.h.f .
pereira , and i. padron , phys .
c * 84 * , 014615 ( 2011 ) .
sargsyan , g.g .
adamian , n.v .
antonenko , and w. scheid , eur .
j. a * 45 * , 125 ( 2010 ) ; v.v .
sargsyan , g.g .
adamian , n.v .
antonenko , w. scheid , and h.q .
zhang , eur .
j. a * 47 * , 38 ( 2011 ) ; j. of phys .
. ser . * 282 * , 012001 ( 2011 ) ; epj web conf .
* 17 * , 04003 ( 2011 ) ; v.v .
sargsyan , g.g .
adamian , n.v .
antonenko , w. scheid , c.j .
lin , and h.q .
zhang , phys .
c * 85 * , 017603 ( 2012 ) ; _ ibid _ * 85 * , 037602 ( 2012 ) .
sargsyan , g.g .
adamian , n.v .
antonenko , w. scheid , and h.q .
zhang , phys .
c * 84 * , 064614 ( 2011 ) ; _ ibid _ * 85 * , 024616 ( 2012 ) ; _ ibid _ * 86 * , 014602 ( 2012 ) ; _ ibid _ * 86 * , 034614 ( 2012 ) . s. ayik , b. yilmaz , and d. lacroix , phys .
c * 81 * , 034605 ( 2010 ) .
zubov , v.v .
sargsyan , g.g .
adamian , and n.v .
antonenko , phys .
c * 84 * , 044320 ( 2011 ) ; a.s .
zubov , v.v .
sargsyan , g.g .
adamian , n.v .
antonenko , and w. scheid , phys .
c * 81 * , 024607 ( 2010 ) ; phys .
c * 82 * , 034610 ( 2010 ) ; r.a .
kuzyakin , v.v .
sargsyan , g.g .
adamian , n.v .
antonenko , e.e .
saperstein , and s.v .
tolokonnikov , phys .
c * 85 * , 034612 ( 2012 ) .
h. hofmann , phys .
* 284 * , 137 ( 1997 ) ; s. ayik , b. yilmaz , a. gokalp , o. yilmaz , and n. takigawa , phys . rev .
c * 71 * , 054611 ( 2005 ) ; v.v .
sargsyan , z. kanokov , g.g .
adamian , and n.v .
antonenko , part . nucl . * 41 * , 175 ( 2010 ) ; g. hupin and d. lacroix , phys .
c * 81 * , 014609 ( 2010 ) .
s. szilner _
et al . _ ,
c * 76 * , 024604 ( 2007 ) ; s. szilner _
c * 84 * , 014325 ( 2011 ) ; l. corradi _ et al .
c * 84 * , 034603 ( 2011 ) .
adamian , a.k .
nasirov , n.v .
antonenko , and r.v .
jolos , phys .
nucl . * 25 * , 583 ( 1994 ) ; k. washiyama , d. lacroix , and s. ayik , phys .
c * 79 * , 024609 ( 2009 ) ; s. ayik , k. washiyama , and d. lacroix , phys . rev .
c * 79 * , 054606 ( 2009 ) .
s. raman , c.w .
nestor , jr , and p. tikkanen , at .
data nucl .
data tables * 78 * , 1 ( 2001 ) .
y. kanada - enyo , h. horiuchi , and a. ono , phys .
c * 52 * , 628 ( 1995 ) ; a dote , h. horiuchi , and y. kanada - enyo , phys .
c * 56 * , 1844 ( 1997 ) .
gasques , d.j .
hinde , m. dasgupta , a. mukherjee , and r.g .
thomas , phys .
c * 79 * , 034605 ( 2009 ) .
a. mukherjee _ et al .
b * 636 * , 91 ( 2006 ) . c. signorini _ et al .
a735 * , 329 ( 2004 ) ; m. dasgupta , d.j .
hinde , s.l .
sheehy , and b. bouriquet , phys .
c * 81 * , 024608 ( 2010 ) . m. dasgupta _ et al .
lett . * 82 * , 1395 ( 1999 ) .
gomes _ et al .
c * 73 * , 064606 ( 2006 ) .
parkar _ et al .
c * 82 * , 054601 ( 2010 ) .
palshetkar _ et al .
c * 82 * , 044608 ( 2010 ) .
gomes _ et al .
_ , phys . rev .
c * 71 * , 034608 ( 2005 ) .
marti _ et al .
c * 71 * , 027602 ( 2005 ) .
i. padron
_ et al . _ ,
c * 66 * , 044608 ( 2002 ) .
thakur _ et al .
_ , epj web conf . *
17 * , 16017 ( 2011 ) . m. dasgupta _ et al .
c * 70 * , 024606 ( 2004 ) .
wu _ et al .
c * 68 * , 044605 ( 2003 ) .
a. shrivastava _ et al .
rev . lett . * 103 * , 232702 ( 2009 ) .
rath _ et al .
c * 79 * , 051601(r ) ( 2009 ) .
d. torresi _ et al .
_ , epj web conf .
* 17 * , 16018 ( 2011 ) .
v. tripathi _ et al .
lett . * 88 * , 172701 ( 2002 ) .
kolata _ et al .
j. a * 13 * , 117 - 121 ( 2002 ) .
m. fisichella _ et al .
_ , epj web conf . *
17 * , 16003 ( 2011 ) .
w. loveland _ et al .
c * 74 * , 064609 ( 2006 ) ; w. loveland , epj web conf .
* 17 * , 02003 ( 2011 ) .
a. lemasson _ et al .
lett . * 103 * , 232701 ( 2009 ) ; a. lemasson epj web conf .
* 17 * , 01003 ( 2011 ) .
a.m. vinodkumar _ et al .
c * 80 * , 054609 ( 2009 ) .
a. diaz - torres , j. phys .
g * 37 * , 075109 ( 2010 ) .
a. diaz - torres , comp .
. comm . * 182 * , 1100 ( 2011 ) .
luong _ et al .
b * 695 * , 105 ( 2011 ) ; epj web conf .
* 17 * , 03002 ( 2011 ) .
gomes , private communication ( 2012 ) .
h. timmers , j.r .
leigh , m. dasgupta , d.j .
hinde , r.c .
lemmon , j.c .
mein , c.r .
morton , j.o .
newton , and n. rowley , nucl .
a584 * , 190 ( 1995 ) .
zhang , f. yang , c. lin , z. liu , and y. hu , phys .
c * 57 * , r1047 ( 1998 ) ; inter .
workshop on _ nuclear reactions and beyond , wold scientific ( 2000 ) , p.95 _ ; f. yang , c.j .
lin , x.k .
wu , h.q .
zhang , c.l .
zhang , p.zhou , and z.h .
liu , phys .
c * 77 * , 014601 ( 2008 ) .
sonzogni , j.d .
bierman , m.p.kelly , j.p.lestone , j.f.liang , and r. vandenbosch , phys .
c * 57 * , 722 ( 1998 ) .
s. sinha , m.r .
pahlavani , r. varma , r.k .
choudhury , b.k .
nayak , and a. saxena , phys .
c * 64 * , 024607 ( 2001 ) .
e. piasecki _ et al .
c * 65 * , 054611 ( 2002 ) ; e. piasecki _ et al .
c * 80 * , 054613 ( 2009 ) ; e. piasecki _ et al .
c * 85 * , 054604 ( 2012 ) ; _ ibid _ * 85 * , 054608 ( 2012 ) . | comparing the capture cross sections calculated without the breakup effect and experimental complete fusion cross sections , the breakup was analyzed in reactions with weakly bound projectiles @xmath0li , @xmath1be , and @xmath2he .
a trend of a systematic behavior for the complete fusion suppression as a function of the target charge and bombarding energy is not achieved .
the quasielastic backscattering is suggested to be an useful tool to study the behavior of the breakup probability in reactions with weakly bound projectiles . |
let @xmath0 be an orientable surface of genus @xmath1 with @xmath2 punctures . in this paper , we consider the surfaces of negative euler characteristic with at least one puncture .
a marked hyperbolic structure on @xmath0 is pair @xmath3 where @xmath4 is a complete hyperbolic metric on a surface @xmath5 and @xmath6 is a homeomorphism .
two marked hyperbolic structures @xmath7 and @xmath8 are called equivalent if there is an isometry in the isotopy class of @xmath9 . for simplicity , we usually denote a marked hyperbolic metric by @xmath4 instead of the pair @xmath3 .
the teichmller space @xmath10 is defined as the space of equivalent classes of marked hyperbolic structures on @xmath0 .
it is well known that @xmath10 , equipped with the natural topology is homeomorphic to a ball in @xmath11 .
given an open convex domain @xmath12 , hilbert defined a natural metric on @xmath13 , now called the hilbert metric , such that the straight line segments are geodesic segments under this metric . in @xcite ,
papadopoulos raised a problem : realize teichmller space as a bounded convex set somewhere and study the hilbert metric on it . in this paper
, we will study this problem . for the case of closed surfaces , yamada @xcite constructed a space which he called the teichmller - coxeter complex within which the original teichmller space sits as an open convex but unbounded subset , based on the weil - petersson completion of the teichmller space . then in @xcite , after introducing a new variational characterization of the hilbert metric , he defined the wei - petersson hilbert metric on the teichmller space , where the background geometry is the one induced by the weil - petersson geometry instead of the euclidean geometry . for the case of punctured surfaces ,
hamenstdt @xcite provided a parametrization of the teichmller space into @xmath14 by length functions such that the image of @xmath10 is the interior of a finite - sided polyhedron ( see [ sec : geometri para ] ) .
therefore the hilbert metric is well defined on @xmath10 .
hamenstdt s parametrization depends on the choice of a preferred triangulation @xmath15 of @xmath0 . more precisely ,
fix a puncture of @xmath0 and denote it as @xmath16 , let @xmath17 be a set of bi - infinite simple curves on @xmath0 such that for any marked hyperbolic metric @xmath4 the two ends of @xmath18 , @xmath19 , go into the puncture @xmath16 and such that @xmath20 consists of @xmath21 ideal triangles and @xmath22 once punctured discs .
such a set @xmath15 is called a _ preferred triangulation _ of @xmath0 .
there are countably many choices of preferred triangulations . in this paper
, we study the hilbert metric @xmath23 on @xmath10 based on hamentdt s parametrization . before stating our main result
, we briefly explain a deformation of hyperbolic metric introduced by thurston in @xcite , namely , the _ earthquake_. let @xmath24 be a simple closed curve on @xmath0 , and @xmath25 be a marked hyperbolic metric .
denote by @xmath26 the geodesic representative of @xmath24 on @xmath4 .
cutting @xmath4 along @xmath26 and twisting to the left about distance @xmath27 , we obtain a new marked hyperbolic metric , denoted by @xmath28 .
note that the notion of `` left '' twist depend only on the orientation of @xmath4 ( no orientation of @xmath26 is necessary ) .
thurston extended this construction to any measured geodesic lamination .
he proved the following result , one of whose proof can be found in @xcite .
[ prop : earthquake ] there is a ( unique ) continuous map @xmath29 , associating an element @xmath30 to @xmath31 , such that @xmath32 for all @xmath33 and all @xmath34 , and such that when @xmath24 is a simple closed geodesic , @xmath28 is obtained from @xmath4 by the earthquake defined above .
the metric @xmath28 defined in proposition [ prop : earthquake ] is said to be obtained from @xmath4 by a _ ( left ) earthquake of amplitude t along the measured geodesic lamination @xmath24 _ , and the orbits @xmath35 , @xmath36 and @xmath37 are called the _ earthquake line directed by @xmath24 and starting at @xmath4 _ , the _ anti - earthquake ray directed by @xmath24 and starting at @xmath4 _ , and the _ earthquake ray directed by @xmath24 and starting at @xmath4 _ , respectively .
recall that for a metric space @xmath38 , an unbounded path @xmath39 is called an _ almost - geodesic _ if for any @xmath40 , there exists @xmath41 , such that @xmath42 for any @xmath43 now we state our main result .
[ thm : main - theorem ] after reparametrization , every ( anti-)earthquake ray is an almost - geodesic in @xmath44 . in fact , the image of an earthquake ray under hamenstdt s parametrization eventually looks like a projective line ( see [ sec : earthquake ] ) .
this paper is organized as the following . in section 2 ,
we recall some basic properties of the hilbert metric and express the hilbert metric @xmath45 on the teichmller space based on hamenstdt s parametrization . in section 3
, we prove our main theorem . in section 4
, we study the dependence of the hilbert metric @xmath23 on the choice of the preferred triangulation @xmath15 .
we will show that a sphere @xmath46 centered at @xmath47 of radius @xmath48 with respect to @xmath45 for a preferred triangulation @xmath15 is again a sphere up to an additive constant with respect to @xmath49 for another preferred triangulation @xmath50 , provided that @xmath50 can be obtained from @xmath15 by a _ diagonal - flip _
( to be defined in [ sec : triangulation ] ) .
but the additive constant depends on the center point @xmath51 . in section 5 , we study the actions of the mapping class group on @xmath44 .
it is well known that when the teichmller space is endowed with the teichmller metric , the thurston metric or the weil - petersson metric , the mapping class group acts by isometries . but
here , we will show that the action of a positive dehn twist is not isometric ( see corollary [ cor : nonisometry ] ) .
instead , it is an almost isometry on an unbounded subset of @xmath44 ( see corollary [ cor : dehn - twist ] )
. * acknowledgements .
* i would like to thank my advisor lixin liu for his careful reading of this manuscript and many useful suggestions .
i would like to thank athanas papadopoulos for useful conversations during his visit at sun yat - sen university ( zhongshan university ) .
i also thank the referee for numerous comments and suggestions .
there are two versions of the hilbert s metric .
the first version is the original one due to hilbert which is defined on a bounded convex domain @xmath52 ( see fig.[fig : hilbert ] ) in @xmath53 .
let @xmath54 be two points in the interior of @xmath52 , the line passing through @xmath55 intersects the boundary @xmath56 at two points @xmath57 , where @xmath58 lies between @xmath59 and @xmath60 .
then the hilbert metric is defined as : @xmath61}=\frac{1}{2}\log{\frac{|a - y||b - x|}{|a - x||b - y|}},\ ] ] where @xmath62 $ ] represents the cross - ratio of @xmath63 .
the second version is due to birkhoff which is defined on the cone @xmath64 over a bounded convex domain @xmath52 ( see @xcite , @xcite and @xcite for more details about this version ) .
recall that a cone @xmath64 is called _ pointed _ if @xmath65 .
let @xmath64 be a closed , pointed ( convex ) cone over a convex bounded domain @xmath52 in @xmath53 .
given two nonzero vectors @xmath58 and @xmath60 in @xmath64 ( see fig.[fig : birk - hilbert ] ) , the _ birkhoff s version of the hilbert metric _ , denoted as @xmath66 is defined as : @xmath67 where @xmath68 @xmath69 denote by @xmath70 the cone point of @xmath71 , and suppose that the line @xmath72 passing through @xmath73 intersects the boundary @xmath74 at @xmath75 . to calculate @xmath76 explicitly ,
we distinguish two cases .
the first case is that the points @xmath77 are collinear . in this case
@xmath78 , hence @xmath79 .
the second case is that the points @xmath80 are not collinear .
we draw an auxiliary line @xmath81 from @xmath58 which is parallel to the line @xmath82 and intersects the line @xmath83 at @xmath84
. then @xmath85 similarly we get @xmath86 hence @xmath87.\ ] ] by the property of cross - ratio , we have @xmath88 for any @xmath89 .
it is clear that @xmath66 is not a metric on @xmath71 since it does not separate @xmath58 and @xmath90 for any @xmath33 .
in fact , @xmath66 is a metric on the projective space @xmath91 .
yamada @xcite gave an alternate definition of the hilbert metric by supporting hyperplanes .
recall that a convex set @xmath52 can be represented as @xmath92 where @xmath93 is the half space bounded by a supporting hyperplane @xmath94 of @xmath52 at the boundary point b , containing the convex set @xmath52 .
let @xmath95 be the set of all the supporting hyperplanes of @xmath52 ( see fig.[fig : yamada - hilbert ] ) .
yamada showed that the hilbert metric can be represented as : @xmath96 we briefly explain yamada s idea here .
let the line @xmath72 intersects @xmath52 at @xmath75 in the order @xmath97 .
let @xmath94 be a supporting hyperplane of @xmath52 at @xmath98 , and let @xmath99 be a hyperplane of @xmath52 at an arbitrary point @xmath100 . denote by @xmath101 the intersection point between @xmath72 and @xmath99 .
it is clear that @xmath102 similarly , we have @xmath103 since a cone over a bounded convex set is again a convex set , yamada s idea also applies to the birkhoff s version of the hilbert metric .
therefore @xmath104\neq[y],\ ] ] where @xmath105=\{\lambda x\in\mathcal c : \lambda>0\},$ ] and @xmath106 is the set of supporting hyperplanes of the cone @xmath71 .
* remark*it is easy to see that these two versions of hilbert metric coincide on the convex bounded domain @xmath52 , i.e. @xmath107 for any @xmath108 . in this paper
, we adopt the birkhoff s version of the hilbert metric . given a reference metric @xmath51 , a _
geodesic lamination _
@xmath109 is a closed subset of @xmath0 consisting of mutually disjoint simple geodesics which are called _ leaves _ of this geodesic lamination .
a _ transverse invariant measure
_ @xmath110 of a geodesic lamination @xmath109 is a radon measure defined on every arc @xmath111 transverse to the support of @xmath109 such that @xmath110 is invariant with respect to any homotopy of @xmath111 relative to the leaves of @xmath109 .
a _ measured geodesic lamination _ is a lamination @xmath109 endowed with a transverse invariant measure @xmath110 .
the simplest example of a measured geodesic lamination is a simple closed geodesic , where the transverse invariant measure is the dirac measure .
each measured geodesic lamination @xmath110 induces a functional on the space @xmath112 of isotopy classes of nontrival simple closed curves on @xmath0 , which assigns @xmath113}\int _ { \tilde \gamma}d\mu$ ] to any @xmath114\in \mathcal{s}$ ] . the amount @xmath113}\int _ { \tilde \gamma}d\mu$ ]
is called the _ intersection number _ of @xmath110 with @xmath114 $ ] and is denoted by @xmath115)$ ] .
two measured geodesic laminations @xmath116 are called _ equivalent _ if @xmath115)=i(\mu',[\gamma])$ ] for any @xmath114\in \mathcal s$ ] .
denote by @xmath117 the space of equivalent classes of measured geodesic laminations on @xmath0 , and equip @xmath118 with the weak topology of the functional space over @xmath119 . with this topology ,
the set of weighted simple closed curves , @xmath120 , is dense in @xmath117 .
the thurston boundary @xmath121 of the teichmller space @xmath10 consists of the projective classes of measured geodesic laminations and is homeomorphic to the unit sphere @xmath122 ( for more details about measured geodesic laminations we refer to @xcite and @xcite ) .
hamenstdt @xcite gave a geometric parametrization of the teichmller space @xmath123 .
we briefly recall this parametrization .
let @xmath124 and let @xmath25 .
fix one of the punctures of @xmath4 and denote it by @xmath16 . as we explained in the introduction ,
a preferred triangulation @xmath15 is a set of @xmath125 mutually disjoint simple geodesics @xmath126 on @xmath5 with the two ends of each simple geodesic going into the puncture @xmath16 and which decompose @xmath4 into @xmath21 ideal triangles and @xmath22 once - punctured discs .
the space of measured lamination @xmath127 on @xmath4 can be parameterized by the @xmath125-tuple @xmath128 , where @xmath110 is a measured geodesic lamination with compact support and @xmath129 represents the intersection number of @xmath110 with @xmath18 .
let @xmath130 be the set of all @xmath125-tuples@xmath131 of nonnegative real numbers with the following properties : \(1 ) @xmath132 if the geodesics @xmath133 are the sides of an ideal triangle on @xmath5 .
\(2 ) there is at least one ideal triangle on @xmath5 with sides @xmath133 such that @xmath134 . in particular , @xmath130 is a cone with vertex at the origin over the boundary of a convex finite - sided polyhedron @xmath135 in the sphere @xmath122 .
and @xmath130 is homeomorphic to @xmath136 .
[ mf ] the map @xmath137 is a homeomorphism of @xmath127 onto @xmath130 .
since for any marked hyperbolic metric @xmath138 , the length of @xmath18 is infinite for any @xmath19 , we need a little modification to parameterize the teichmller space @xmath123 . recall that every puncture of @xmath5 admits a standard cusp neighbourhood which is isometric to a cylinder @xmath139 equipped with the metric @xmath140 ( @xcite ) .
the @xmath141-coordinate in this representation is called the height .
let @xmath142 be an ideal triangle on the upper half plane ( see fig .
[ fig : triangle ] ) with sides @xmath143 .
each corner of @xmath142 can be foliated by horocycles .
extend these foliations until they fill all but in the center bounded by three horocycles @xmath144 .
we call @xmath145 the _ midpoint _ of @xmath18 with respect to the ideal triangle @xmath146 for @xmath147
. choose the height @xmath148 small enough such that @xmath149 is much smaller than the length of the horocycle @xmath150 ( note that the three horocycles @xmath151 have the same length ) .
every geodesic going into the cusp meets the horocycles @xmath141=constant orthogonally .
hence each choice of a height @xmath148 cuts from @xmath18 a unique compact arc of finite length since both ends of the geodesic @xmath18 go into the cusp for any @xmath19 .
denote by @xmath152 the length of this subarc of @xmath18 for a marked hyperbolic metric @xmath4 for any @xmath19 .
[ thm : teich ] let @xmath153 be the canonical projection .
the map @xmath154 is a diffeomorphism of @xmath10 onto a hypersurface in @xmath155 .
and the map @xmath156 is a diffeomorphism of @xmath10 onto the interior of a finite - sided closed convex polyhedron @xmath135 in @xmath157 which extends to a homeomorphism of @xmath158 onto @xmath135 .
, width=264 ] now we give the hilbert metric on the teichmller space .
let @xmath159 be the hyperplane in @xmath160 defined by @xmath161 where the indices @xmath162 satisfy the condition that the corresponding simple geodesics @xmath163 bound an ideal triangle on the surface .
let @xmath95 be the set of all such hyperplanes .
by theorem [ mf ] and theorem [ thm : teich ] , we know that the image of @xmath164 under the map @xmath165 is a polyhedron @xmath135 . following yamada s idea , we define the hilbert metric on @xmath10 by ( [ eq : yamada ] ) : @xmath166 for all @xmath167 , where @xmath168 represents the euclidean distance . from ( [ pi ] ) , we have @xmath169 hence @xmath170 now we give some explanations for ( [ eq : def ] ) .
since the surface @xmath0 is orientable , we fix an orientation . denote by @xmath171 the ideal triangle with three sides @xmath172 such that @xmath173 appears in the counterclockwise order . obviously @xmath174 .
let @xmath175 be the collection of all such triangles corresponding to @xmath15 .
the two ends of each of @xmath173 go into the cusp @xmath16 and the intersection of the horocycle @xmath176 with the ideal triangle @xmath171 consists of three components , denoted as @xmath177 ( the red lines in fig .
[ fig : triangle ] ) .
it is clear that the sum of the lengths @xmath178 over all these @xmath21 ideal triangles is less than the length of the horocycle @xmath176 .
denote by @xmath179 the midpoint of @xmath18 with respect to the ideal triangle @xmath171 .
@xmath145 together with @xmath180 cuts @xmath18 into four connected components , two of which are compact , denoted by @xmath181 and @xmath182 with respect to the counterclockwise order ( see fig . [ fig : triangle ] ) .
it follows that @xmath183 hence @xmath184 now ( [ eq : def ] ) can be rewritten as @xmath185 and the length of the horocycle @xmath186 can be expressed in terms of @xmath187 and @xmath188 as @xmath189 .
therefore @xmath190 we summarize our discussions above as the following proposition .
[ prop : hilbert ] with the notations above , the hilbert metric @xmath45 on the teichmller space @xmath10 can be expressed as the following two forms .
* for any @xmath167 , @xmath191 where , @xmath192 for any @xmath25 and any ideal triangle @xmath193 .
* for any @xmath167 , @xmath194 where @xmath195 for any ideal triangle @xmath193 .
* remark*the reason that we do not consider the funk metric is that the funk metric is not projectively invariant while the hilbert metric is .
if we choose a fixed homogenous coordinate for @xmath14 , we can define a funk metric on @xmath10 .
for any @xmath25 , @xmath196 , denote by @xmath197 the length of the geodesic representative of @xmath198 on @xmath4 . in @xcite , kerckhoff proved that for any simple closed curve @xmath198 , @xmath199 is a convex function of @xmath27 along the earthquake line @xmath200 .
based on this result , bonahon proved @xcite that each ( anti- ) earthquake ray converges to a unique point in the thurston boundary @xmath121 .
[ lem : ke ] for any @xmath25 , @xmath201 and @xmath196 , with @xmath202 , then @xmath203 where @xmath204 represents the angle at each intersection point between ( the geodesic representatives of ) @xmath198 and @xmath24 on @xmath28 measured counter - clockwise from @xmath198 to @xmath24 .
moreover , as t tends to @xmath205 ( resp .
@xmath206 ) , the function @xmath207 is strictly increasing ( resp . decreasing ) .
[ lem : bon ] for every @xmath25 and every @xmath34 , @xmath208 by applying similar arguments as in @xcite and @xcite , we get similar results for @xmath209 ( note that @xmath209 is the length of a particular compact subarc of @xmath18 ) .
[ lem : derivative ] for any @xmath25 , @xmath201 and @xmath210 , @xmath211 where @xmath204 represents the angle at each intersection point between ( the geodesic representatives of ) @xmath18 and @xmath24 on @xmath28 measured counter - clockwise from @xmath18 to @xmath24 .
moreover , as t tends to @xmath205 ( resp .
@xmath206 ) , the function @xmath207 is strictly increasing ( resp . decreasing ) .
the proof is exactly the same as that of lemma [ lem : ke ] in @xcite .
[ lem : eta - convergence ] for every @xmath25 and every @xmath34 , @xmath212 the proof is exactly the same as that of lemma [ lem : bon ] in @xcite .
also , it follows from lemma [ lem : estimate ] .
in fact , we can prove a stronger result .
[ lem : estimate ] for every @xmath25 and every @xmath34 , @xmath213 where @xmath149 represents the length of the horocycle centered at the puncture @xmath16 with height @xmath148 .
moreover , there is a constant @xmath214 $ ] such that @xmath215=c^+_i,\ \lim_{t\to -\infty } [ l_{\eta_i}(\mathcal e^t_\alpha x)-|t| i(\alpha , \eta_i)]=c^-_i.\ ] ] by symmetry , it suffices to prove the lemma for @xmath216 .
we start with the case that @xmath24 is a simple closed geodesic .
let @xmath217 .
we will show that , @xmath218 from lemma [ lem : derivative ] , @xmath219 as a consequence , we get our first inequality immediately , @xmath220 by the monotonicity of @xmath221 , for any positive integer @xmath222 , we have @xmath223.\ ] ] therefore , to prove @xmath224 for all @xmath216 , it suffices to prove @xmath225 for all @xmath226 .
is obtained from @xmath198 by two times dehn twist along @xmath24 .
, width=453 ] recall that @xmath227 is obtained from @xmath4 by @xmath222 times dehn twist along @xmath24 .
denote by @xmath228 the @xmath222 times dehn twist along @xmath24 .
hence @xmath229 .
denote by @xmath230 the horocycle centered at the puncture @xmath16 with height @xmath148 , and denote by @xmath231 , @xmath232 the geodesic representatives of @xmath18 , @xmath233 on @xmath4 , respectively .
it is clear that @xmath231 and @xmath232 have @xmath234 intersection points , one of which is the puncture @xmath16 .
the remaining @xmath235 intersection points cut the restrictions of @xmath231 and @xmath232 on @xmath236 into @xmath237 segments @xmath238 , where @xmath239 and @xmath240 represents the cusp neighbourhood of @xmath16 with boundary horocycle @xmath230 . on the other hand , @xmath231 and @xmath232 cut the horocycle @xmath230 into four segments
@xmath241 two of these four segments , say @xmath242 , together with above mentioned @xmath237 segments @xmath238 consist @xmath234 simple closed curves , and all of them are homotopic to @xmath24 ( see fig .
[ fig : closed - curves ] ) .
more precisely , let @xmath243 , @xmath244 , @xmath245 , @xmath246 , @xmath247 .
@xmath248 are simple closed curves , and all of them are isotopic to @xmath24 .
it follows that @xmath249 which means that @xmath250 this completes the proof of ( [ eq : simple ] ) .
now we extend these estimates to an arbitrary measured geodesic lamination @xmath110 . recall that @xmath120 is dense in @xmath117 .
there exists @xmath251 such that @xmath252 , as @xmath253 .
it follows from proposition [ prop : earthquake ] that for any given @xmath254 , @xmath255 as @xmath253 .
hence for any given @xmath256 , @xmath257 as @xmath253 . from ( [ eq : simple ] ) , we have @xmath258 let @xmath222 tend to infinity , we have @xmath259 the existence of the limit @xmath260 $ ] follows from the boundedness and the monotonicity of @xmath261 .
based on the estimates in lemma [ lem : estimate ] , we can describe the coordinates of an earthquake line @xmath262 .
denote by @xmath263 the hyperplane @xmath264 and denote by @xmath265 the projective map @xmath266 for a measured geodesic lamination @xmath110 , let @xmath267 .
it is clear that @xmath268 moreover , it follows from lemma [ lem : estimate ] that for large enough @xmath269 , @xmath270 in other words , the images of the earthquake rays @xmath271 and @xmath272 in @xmath263 look like straight line segments when @xmath273 is large enough .
now we prove the main theorem of this paper .
after reparametrization , every ( anti-)earthquake ray is an almost - geodesic in @xmath44 . by symmetry , it suffices to prove the theorem for every earthquake ray .
let @xmath274 be an earthquake ray directed by @xmath110 and starting at @xmath4 .
set @xmath275 .
let @xmath276 .
it follows from lemma [ lem : estimate ] that @xmath277 and that for any @xmath40 there is a constant @xmath278 depending on @xmath279 , @xmath4 , @xmath15 and @xmath110 such that @xmath280 moreover , if @xmath281 for some @xmath282 , @xmath283 .
next we divide the set of ideal triangles corresponding to the preferred triangulation @xmath15 into three types : * * type * a consists of ideal triangles @xmath171 which do not intersect @xmath110 , i.e. @xmath284 ; * * type * b consists of ideal triangles @xmath171 whose two sides , say @xmath285 , intersect @xmath24 and whose remaining side does not intersect @xmath110 , i.e. @xmath286 and @xmath287 ; * * type * c consists of ideal triangles @xmath171 with each side intersecting @xmath110 , i.e. @xmath288 . for each triangle @xmath171 , set @xmath289 now we discuss case by case . * type * a. in this case , @xmath284 . @xmath290
* type * b. in this case , @xmath286 and @xmath287 .
recall that for any ideal triangle @xmath291 ( see proposition [ prop : hilbert ] ) , @xmath292 hence for any @xmath40 , there is a constant @xmath293 depending on @xmath279 , @xmath15 , @xmath4 and @xmath110 such that @xmath294 , we have @xmath295 and @xmath296 moreover , there is @xmath297 depending on @xmath279 , @xmath15 , @xmath4 and @xmath110 such that @xmath298 , we have @xmath299 and @xmath300 where @xmath301 and @xmath302 are defined by ( [ eq : d - ijk ] ) . *
type * c. in this case , @xmath288 .
we distinguish two subcases .
if @xmath303 , it follow from lemma [ lem : estimate ] that there is a constant @xmath304 depending on @xmath279 , @xmath15 , @xmath4 and @xmath110 such that @xmath305 , we have @xmath306 if @xmath307 , there is a constant @xmath308 depending on @xmath279 , @xmath15 , @xmath4 and @xmath110 such that @xmath309 , we have @xmath310 therefore , for any @xmath311 , @xmath312 and @xmath313 where @xmath314 , and @xmath315 . now , reparametrize the earthquake ray @xmath316 as @xmath317 by setting @xmath318 .
it follows from ( [ eq : dist - st ] ) and ( [ eq : dist-0s ] ) that for any @xmath40 there is a @xmath41 depending on @xmath279 , @xmath15 , @xmath4 and @xmath110 such that @xmath319 and @xmath320 therefore @xmath317 is an almost geodesic in @xmath321 .
let @xmath323 be a proper geodesic metric space , which is endowed with the topology induced by the metric @xmath324 .
we embed @xmath323 into @xmath325 , the space of continuous real - valued functions on @xmath4 endowed with the topology of uniform convergence on bounded sets , by the map below : @xmath326,\end{aligned}\ ] ] where @xmath327 is a base point .
the _ horofunction boundary _ of @xmath323 is defined to be @xmath328 where @xmath329 represents the closure of @xmath330 in @xmath325 .
the horofunction boundary is independent of the base point , i.e. @xmath331 is homeomorphic to @xmath332 for @xmath333 .
a function in @xmath334 is called a _
horofunction_. rieffel ( @xcite ) observed that every almost geodesic converges to a unique point in the horofunction boundary of @xmath38 .
therefore , we have the following corollary .
[ cor : convergence ] every ( anti- ) earthquake ray converges to a unique point in the horofunction boundary of @xmath321 . *
remark*as we mentioned in the beginning of this section , bonahon proved @xcite that each ( anti- ) earthquake ray converges to a unique point in the thurston boundary @xmath121 .
it follows from theorem [ thm : teich ] that each ( anti- ) earthquake ray converges to a point in the euclidean boundary of the polytope @xmath135 . in general , let @xmath13 be a bounded convex domain in the euclidean space .
foertsch - karlsson ( @xcite ) proved that every geodesic under the hilbert metric converges to a boundary point in @xmath335 in the euclidean sense .
walsh ( @xcite ) proved that every sequence converging to a point in the horofunction boundary of the hilbert geometry converges to a point in the euclidean boundary @xmath335 .
hence every almost geodesic under the hilbert metric converges to a boundary point in @xmath335 .
in this section , we investigate the dependence of the hilbert metric @xmath23 on the choice of triangulation @xmath15 .
first of all , we define a basic operation for a triangulation , namely , _
diagonal - flip_. ( -0.5,0)(0.5,0 ) node[pos=0.5,above ] diagonal - flip ; ( -0.5,0)(0.5,0 ) node[pos=0.5,above ] @xmath337 ; let @xmath338 be an ideal quadrilateral with sides @xmath339 and ideal vertices ( corners ) @xmath340 . let @xmath341 be a horocycle around the corner @xmath342 whose length is smaller than @xmath149 , @xmath343 , where @xmath148 is chosen in section [ sec : geometri para ] .
let @xmath24 be a diagonal geodesic connecting @xmath344 which triangulates @xmath345 into two triangles @xmath346 and @xmath347 , and @xmath348 the `` midpoint '' of @xmath339 with respect to these two triangles .
further , let @xmath349 and @xmath350 be the midpoints of @xmath24 with respect to @xmath346 and @xmath347 respectively
. denote by @xmath351 and @xmath352 the distances from @xmath145 to @xmath353 and from @xmath145 to @xmath354 respectively .
let @xmath198 be the other diagonal geodesic , and the notations @xmath355 , @xmath356 , @xmath357 and @xmath358 are defined similarly .
( see fig .
[ fig : flip ] . ) with notations described above , we say that @xmath198 is obtained from @xmath24 by a _ diagonal - flip _ with respect to @xmath359 .
a triangulation @xmath50 is said to be obtained from @xmath15 by a diagonal - flip if @xmath360 and that @xmath361 can be obtained from @xmath24 by a diagonal - flip with respect to some quadrilateral whose four sides are contained in @xmath360 .
we define the _ shearing of @xmath359 with respect to @xmath24 _ , denoted by @xmath362 , in the following way .
the absolute value of @xmath362 is defined to be the distance between @xmath363 and @xmath364 , and the sign of @xmath362 is defined to `` + '' if @xmath364 sits on the left side of @xmath363 observed from @xmath346 , otherwise the sign is defined to be `` - '' .
* remark*the `` left '' notation in the definition of the shearing depends only on the orientation of the ideal quadrilateral .
the lemma below describes some basic properties of a diagonal - flip .
[ lem : flip ] 1 . for the shearing of the ideal quadrilateral @xmath359 along @xmath198 , @xmath365 .
2 . suppose @xmath366 , then @xmath367 and @xmath368 here we adopt the upper half - plane model for the hyperbolic geometry ( see fig .
[ fig : shearing ] ) .
since the map @xmath369 is an isometry of the hyperbolic metric , we suppose that the euclidean coordinates of @xmath370 are @xmath371 respectively .
\(1 ) the shearing of @xmath359 along @xmath24 is @xmath372 . to calculate the shearing of @xmath359 along @xmath198 , we perform the fractional linear map @xmath373 on the upper half plane which sends @xmath374 to @xmath375 respectively .
it is clear in this case that the shearing of @xmath376 along @xmath377 is @xmath378 , which means the shearing of @xmath359 along @xmath198 is also @xmath378 since @xmath379 is an orientation preserving isometry .
\(2 ) by assumption @xmath366 , hence @xmath380 .
it is clear that @xmath381 .
the coordinates of @xmath382 are @xmath383 and @xmath384 respectively .
in addition , the coordinate of the intersection point between @xmath385 and @xmath386 is @xmath387 .
therefore @xmath388 $ ] , hence @xmath389 $ ] .
similarly we get @xmath390 $ ] and @xmath391 $ ] . for the remaining inequalities , note that @xmath392 and @xmath393 , hence @xmath394 $ ] , and @xmath395 $ ] .
similarly we get @xmath396 $ ] , and @xmath397 $ ] .
the estimates from the lemma above describe a close relationship between the changes of the lengths of preferred arcs and the shearing along each simple geodesic @xmath18 after a diagonal - flip operation .
this relationship provides a clue for investigating the effect of the triangulation @xmath15 on the hilbert metric @xmath23 , i.e. investigating the relationship between @xmath398 and @xmath399 for two different triangulations @xmath15 , @xmath50 .
if we fix a point @xmath47 , the sphere @xmath400 centered at @xmath51 of radius @xmath48 with respect to @xmath45 is an almost - sphere , i.e. a sphere up to an additive constant , with respect to @xmath49 , where @xmath50 can be obtained from @xmath15 by a diagonal - flip .
[ prop : diagonal - sphere ] fix @xmath47 .
let @xmath401 be two preferred triangulations of @xmath0 such that one can be obtained from the other by a diagonal - flip , then there is a constant @xmath402 depending on @xmath403 such that @xmath404 set @xmath405 , @xmath406 .
first , we prove an inequality which holds for any preferred triangulation @xmath15 , @xmath407 indeed , there are two simple geodesics @xmath408 such that @xmath18 , @xmath409 , @xmath24 bound an ideal triangle . recall that @xmath410 then @xmath411 hence @xmath412 the first inequality is obvious .
now the proposition follows immediately from proposition [ prop : o - dist ] and ( [ eq : ob ] ) . [ prop : o - dist ] fix @xmath47 , then there is a constant @xmath413 depending on @xmath414 and @xmath148 such that @xmath415 set @xmath416 . by ( [ eq : def2 ] ) ,
@xmath417 note that @xmath418 + [ { l_{\eta_j}(x)-l_{\eta_k}(x)+l_{\eta_i}(x)}]\ ] ] for any ideal triangle @xmath171 , and @xmath419}\geq 2\rho_0,\text { for any } \delta_{ijk}.\ ] ] therefore @xmath420\\ & = & \frac{1}{2}\sup_{\eta\in\gamma}\log l_{\eta}(x)+\frac{1}{2}\sup_{\eta\in\gamma}\log l_{\eta}(x_0)-\log \rho_0 .
\end{aligned}\ ] ] next we deal with the inverse inequality . without loss of generality , we assume that @xmath421 and that @xmath143 bounds an ideal triangle . by ( [ eq : sum ] ) , at least one of @xmath422 and @xmath423 is not less than @xmath424 . then @xmath425\\ & = & \frac{1}{2}\sup_{\eta\in\gamma}\log l_{\eta}(x)-\frac{1}{2}[\log 2+\sup_{\eta\in\gamma}l_{\eta}(x_0 ) ] .
\end{aligned}\ ] ] set @xmath426 , the proposition follows .
the mapping class group @xmath427 consists of the isotopy classes of orientation - preserving self homeomorphisms of @xmath0 . in this section ,
we study the actions of mapping class group @xmath427 on the metric space @xmath44 .
the action is defined as following . for @xmath428 and @xmath429 ,
@xmath430 is defined as the marked hyperbolic surface @xmath431 .
denote by @xmath432 the subgroup of @xmath427 consisting of elements that fix each puncture individually .
it is well known that @xmath432 can be generated by finitely many dehn twists about nonseparating simple closed curves , where a nonseparating simple closed curve @xmath24 is a closed curve such that @xmath433 is connected ( see ( * ? ? ? * chap .
the lemma below describes the changes of @xmath152 under a dehn twist .
[ lem : dehn - twist ] assume that @xmath434 is an ideal triangle on @xmath25 with three sides @xmath435 , and that @xmath1 is a positive dehn twist of @xmath0 along an essential simple closed curve @xmath24 .
denote by @xmath436 , @xmath437 , the images of @xmath434 , @xmath435 , respectively , under the action of @xmath1 .
then , there is a constant @xmath438 depending on the length @xmath439 , the reference height @xmath148 and the isotopy classes of @xmath24,@xmath143 such that note that @xmath441 @xmath442 on the other hand , from ( [ eq : rho ] ) , @xmath443 @xmath444 hence @xmath445l_{\alpha}(x ) } { l_{\eta_{i+1}}(x)+l_{\eta_{i-1}}(x)-l_{\eta_i}(x)}\\ & \leq & 1+\frac{[i(\eta_{i+1},\alpha)+i(\eta_i,\alpha ) + i(\eta_{i-1},\alpha)]l_{\alpha}(x)}{2\rho_0}. \end{aligned}\ ] ] interchange @xmath18 with @xmath446 , we get the inverse inequality @xmath447l_{\alpha}(x)}{2\rho_0}.\ ] ] [ cor : dehn - twist ] let @xmath15 be a preferred triangulation of @xmath0 .
let @xmath1 be a positive dehn twist of @xmath0 along an essential simple closed curve @xmath24 .
set @xmath448 .
then there is a constant @xmath449 depending on @xmath15 , the isotopy class of @xmath24 , the reference height @xmath148 and @xmath450 such that @xmath451 [ prop : aymptotic ] let @xmath15 be a preferred triangulation , and @xmath428 be a positive dehn twist about a simple closed curve @xmath24 . for any given @xmath454 ,
there is a positive number @xmath455 depending on @xmath456 such that @xmath457 moreover , for any @xmath25 , @xmath458 [ cor : nonisometry ] let @xmath15 be a preferred triangulation , and @xmath428 be a positive dehn twist about a simple closed curve @xmath24 .
then the action of @xmath1 on @xmath10 is not isometric . in particular , @xmath460 is not @xmath427 invariant .
fathi a. , laudenbach f. and ponaru v. , thurston s work on surfaces .
translated from the 1979 french original by djun m. kim and dan margalit .
mathematical notes , 48 .
princeton university press , princeton , nj , 2012 .
foertsch t. and karlsson a. , hilbert metrics and minkowski norms .
j. geom . 83(1 - 2 ) , 22 - 31 .
hamenstdt u. , length functions and parameterizations of teichmller space for surfaces with cusps .
( 2003 ) , no . 1 , 75 - 88 .
lemmens b. , nussbaum r. , birkhoff s version of hilbert s metric and its applications in analysis . handbook of hilbert geometry , 275303 , irma lect .
phys . , 22 , eur .
soc . , zrich , 2014 .
seppl m. , sorvali t. , parametrization of teichmller spaces by geodesic length functions.holomorphic functions and moduli , vol .
ii ( berkeley , ca , 1986 ) , 267284 , math .
publ . , 11 , springer , new york , 1988 .
thurston , w. p. earthquakes in 2-dimensional hyperbolic geometry .
fundamentals of hyperbolic geometry : selected expositions , 267289 , london math .
lecture note ser .
, 328 , cambridge univ . press , cambridge , 2006 | hamenstdt gave a parametrization of the teichmller space of punctured surfaces such that the image under this parametrization is the interior of a polytope .
in this paper , we study the hilbert metric on the teichmller space of punctured surfaces based on this parametrization .
we prove that every earthquake ray is an almost geodesic under the hilbert metric .
ams mathematics subject classification ( 2010 ) : primary 51f99 ; secondary 57n16 ; 52b70 .
keywords : teichmller space , hilbert metric , earthquake , mapping class group .
= 1 |
in several areas of applied science researchers are interested in studying the time @xmath2 to take a given action , based on sequentially observed random variables ( rv s ) @xmath11 , as well as in the associated partial sums @xmath12 .
the waiting time @xmath2 and the corresponding random sum @xmath4 are usually referred to as _ stopping time _ and _ stopped sum _ respectively . stopping
time problems arise in many diverse scientific areas such as sequential analysis , quality control , mathematical finance , operations research , biology , actuarial science , etc . for a gentle introduction to the theory of stopping times and
stopped sums , the interested reader is referred to karlin and taylor ( 1975 ) . for a more thorough investigation of the theory of stopped
random walks we refer to gut ( 2009 ) .
when studying the distribution of @xmath2 in a sequence of independent and identically distributed ( iid ) trials , the stopped sum @xmath4 also provides useful information about the nature of the statistical experiment . the pioneering work of abraham wald ( 1945 ) in the area of sequential analysis established powerful identities that relate the distributional properties of @xmath2 and @xmath4 .
these identities are usually referred to as _
wald s ( fundamental ) identity _ and _ wald s _ ( _ first _ ) _ equation _ and they are , respectively , given by @xmath13where @xmath14 , and @xmath15 in a recent article antzoulakos and boutsikas ( 2007 ) established a particular relation between the distributions of @xmath2 and @xmath16 more specifically , they considered the waiting time @xmath17 until the @xmath18th occurrence of a pattern @xmath19 in a sequence of binary trials @xmath1 and the total number of successes @xmath20 observed until that time , and established a direct method to obtain the joint probability generating function ( pgf ) of @xmath21 from the pgf of @xmath17 only . in this paper we extend the aforementioned result for any distribution of the @xmath22 s and any stopping time @xmath2 , determining the joint distribution of @xmath8 from the distribution of @xmath2 or @xmath4**. * * the organization of the paper is as follows : in section 2 we state the main identities that connect the distributions of @xmath2 and @xmath23 along with the required theoretical backup . an important part of our work is comprised of the applications that are presented in section 3 . these applications , not only serve as an illustration of the applicability of the results of section 2 , but they also have an interest on their own . in the first one we consider the first exit time @xmath2 from an interval @xmath24 ( @xmath25 or @xmath26 ) of a random walk @xmath27 @xmath28 with exponentially distributed up and down steps@xmath29 by identifying the distribution of @xmath4 we extract an exact formula for the pgf of the boundary crossing time @xmath30 in the second application we consider a sequence @xmath31 @xmath28 of measurements taken from samples corresponding to lots of products from a manufacturing process ( e.g. number of defective items in each sample ) .
denoting by @xmath2 the waiting time until the sampling level of the inspection changes using a @xmath9-run switching rule associated with @xmath22 s , we obtain the joint pgf of @xmath2 and @xmath4 ( @xmath4 denotes the total number of defective items observed until switching ) by exploiting the fact that @xmath2 follows a geometric distribution of order @xmath9 .
finally , we demonstrate how the joint distribution of @xmath2 and @xmath4 can be useful in the estimation of the probability @xmath10 of an item being defective , by employing an em algorithm .
let @xmath32 be a sequence of distributions on @xmath33 such that @xmath34 @xmath28 for every @xmath35 in an interval @xmath36 containing zero . we can always construct a sequence of independent rv s @xmath1 on a probability space @xmath37 such that @xmath38 . moreover , if @xmath39 denotes the _ exponentially tilted _
@xmath40 i.e. @xmath41})/\mathbb{e}(e^{wz_{i}}),~w\in \mathbf{r},$ ] we can always change @xmath42 to a new probability measure @xmath43 on @xmath44 under which @xmath1 are still independent but now @xmath45 .
a formal construction of the probability space @xmath46 is given in the appendix .
we shall write @xmath47 for the expected value with respect to the measure @xmath48 we shall also use the notation @xmath49 it is easy to see that , in the special case when @xmath1 possess the same density @xmath50 with respect to @xmath42 , their density @xmath51 with respect to @xmath43 is given by @xmath52 * * remark .
( * * _ _ the derivative _ _ @xmath53 _ _ on _ _
@xmath54**)**. define @xmath55 to be the minimal @xmath56-algebra generated by @xmath57 .
the sequence @xmath58 is a nondecreasing sequence of @xmath56-algebras in @xmath59 the radon - nikodym derivative of @xmath43 with respect to @xmath42 when both are restricted to @xmath54 is @xmath60 ( that is , @xmath61 ) and hence @xmath62for every @xmath54-measurable random variable @xmath63 .
it is worth mentioning that , even though @xmath42 and @xmath43 are equivalent on every @xmath54 , they are mutually singular on @xmath64 when @xmath65 and @xmath1 are identically distributed ( that is , there exist disjoint sets @xmath66 in @xmath67 such that @xmath68 and @xmath69 ) .
this can be easily seen since there exists a set @xmath70 @xmath71 such that @xmath72 while ( invoking the strong law of large numbers ) @xmath73}$ ] converges to @xmath74 on some @xmath75 with @xmath68 and to @xmath76 on some @xmath77 with @xmath69 .
since @xmath72 we have that @xmath78 thus @xmath79 ( see e.g. theorem 35.8 in billingsley ( 1986 ) ) even though @xmath80 for every @xmath81 therefore , in general , there does not exist a radon - nikodym derivative of @xmath43 with respect to @xmath42 on @xmath67 and hence @xmath43 can not be constructed on @xmath67 from @xmath42 through a radon - nikodym derivative .
this fact does not induce any problem since we have guaranteed the existence of @xmath43 via the kolmogorov existence theorem ( see appendix ) .
let now @xmath2 be a stopping time associated with the sequence @xmath1 , i.e. the set @xmath82=\ { \omega \in \omega : t(\omega ) = n\}$ ] belongs to @xmath83 for every @xmath84 and let @xmath85 the next result relates the distributions of @xmath2 and @xmath4 .
let @xmath2 be a stopping time associated with the sequence @xmath1 , and let @xmath63 be a random variable such that @xmath86}$ ] is @xmath54-measurable .
then@xmath87})=\mathbb{\tilde{e}}_{w}(y\prod_{i=1}^{t}\mathbb{e}(e^{wz_{i}})i_{[t<\infty ] } ) \label{eq11}\]]for all real @xmath35 such that the above expectations exist .
if @xmath88}$ ] then @xmath89}$ ] a.s . and
@xmath90})<\infty , $ ] which , by the dominated convergence theorem ( dct ) , implies that @xmath91 .
thus,@xmath87})=\mathbb{e}(\lim_{k\rightarrow \infty } g_{k})=\lim_{k\rightarrow \infty } \mathbb{e}(g_{k})=\tsum \limits_{n=1}^{\infty } \mathbb{e}(yi_{[t = n]}e^{ws_{n}}).\]]by theorems assumptions , the r.v .
@xmath92}$ ] is @xmath54-measurable and hence ( see ( [ eq10 ] ) above ) @xmath93})=\mathbb{e}(yi_{[t = n]}e^{ws_{n}})/\tprod_{i=1}^{n}\mathbb{e}(e^{wz_{i}})$ ] .
therefore,@xmath87})=\tsum \limits_{n=1}^{\infty } \mathbb{\tilde{e}}_{w}(yi_{[t = n]}\tprod_{i=1}^{n}\mathbb{e}(e^{wz_{i}}))=\tsum\limits_{n=1}^{\infty } \mathbb{\tilde{e}}_{w}(yi_{[t = n]}\tprod_{i=1}^{t}\mathbb{e}(e^{wz_{i}}))\]]which , invoking again the dct , leads to ( [ eq11 ] ) provided that @xmath94})<\infty .$ ] the above result can be considered as a version of wald s likelihood ratio identity ( wlri , see e.g. siegmund ( 1985 ) , or lai ( 2004 ) ) . in the sequel we focus on a special use of equation ( [ eq11 ] ) .
our aim is to generalize the following result of antzoulakos and boutsikas ( 2007 ) : if @xmath95 , .. is a sequence of iid binary rv s ( trials ) with @xmath96 and @xmath2 denotes the waiting time ( i.e. the number of trials ) until a certain pattern @xmath19 occurs in @xmath95 , .. then , the joint pgf of @xmath8 follows from the pgf of @xmath2 through the relation@xmath97for all @xmath98 in a neighborhood of 0 , where the expectation @xmath99 is considered under @xmath43 such that @xmath100 .
the above identity , reveals that , when the distribution of @xmath2 is known then the joint distribution of @xmath8 is also known . in other words ,
the distribution of @xmath2 uniquely determines the joint distribution of @xmath8 and consequently the distribution of @xmath16 a generalization of ( [ eq12 ] ) could refer to any distribution for the @xmath22 s and any stopping time @xmath30 in addition , an inverse form of ( eq12 ) could also be very useful implying that the distribution of @xmath4 uniquely determines the joint distribution of @xmath101 as it is shown in the next two corollaries , generalizations of this form can be easily derived from equation ( [ eq11 ] ) .
[ cor1]if @xmath102 then @xmath103for all real @xmath104 such that the above expectations exist . in particular
, @xmath105 it follows from ( [ eq11 ] ) by letting @xmath1 be a sequence of iid rv s and by setting @xmath106 ( note that @xmath107}=u^{n}i_{[t = n]}$ ] is @xmath54-measurable ) .
[ cor2]if there exists a real function @xmath108 such that @xmath109 and @xmath110 is a probability measure with @xmath111 , then @xmath112for all real @xmath113 such that the above expectations exist .
in particular , @xmath114 by setting @xmath115 we have that the rv @xmath92}=u^{n}e^{(x - w_{u})(z_{1}+ ... +z_{n})}i_{[t = n]}$ ] is @xmath54-measurable
. therefore by employing ( [ eq11 ] ) with respect to the measures @xmath42 and @xmath110 we get @xmath116which readily leads to ( [ eq14 ] ) since @xmath117 it is worth mentioning that , more generally , we can similarly get from ( eq11 ) that,@xmath118})=\mathbb{\tilde{e}}_{w}(y(u\mathbb{e}(e^{wz}))^{t}i_{[t<\infty ] } ) \label{eq20a}\]]and @xmath119})=\mathbb{\tilde{e}}_{w_{u}}(ye^{(x - w_{u})s_{t}}i_{[t<\infty ] } ) , \label{eq20}\]]where @xmath63 is a rv such that @xmath92}$ ] is @xmath54-measurable .
the above corollaries imply that , under appropriate conditions , the distribution of @xmath4 uniquely determines the distribution of the stopping time @xmath2 and vice versa .
two applications illustrating this fact are presented in the following section .
let @xmath1 be a sequence of ( non - degenerate ) iid rv s representing the consecutive jumps of a random walk @xmath120 , that is , @xmath121 define also the following stopping time@xmath122for some @xmath123 .
obviously , @xmath2 expresses the steps of the random walkuntil it exits the set @xmath5 it can be easily verified that @xmath124 ( e.g. see karlin and taylor ( 1975 ) , p.264 ) and thus @xmath2 is finite a.s
. probabilities regarding the first passage , or boundary crossing times arise in a variety of contexts in applied probability and statistics , such as sequential analysis , ruin theory , queueing theory , stochastic finance etc .
usually , it is of interest to evaluate the probability @xmath125 , the distribution of @xmath2 and @xmath126 . in order to illustrate the applicability of identities ( [ eq13 ] ) and ( eq14 ) we consider first the case when the jumps @xmath22 are exponentially distributed ( negative or positive with probabilities @xmath10 and @xmath127 respectively ) and deduce explicit formulae for the pgf @xmath128 the joint gf @xmath129
, the conditional pgf @xmath130 and the expected values @xmath131 and @xmath132 we also consider the case @xmath26 corresponding to a random walk with only an upper barrier , which requires a different treatment ( in this case @xmath2 is not always a.s .
finite ) . *
( a ) * denote by @xmath133 the exponential distribution with parameter @xmath134 .
for @xmath135 , let @xmath136where @xmath137 and @xmath138 are two sequences of iid rv s such that @xmath139 @xmath140 .
it follows that the pdf @xmath50 of each @xmath22 is the mixture , @xmath141 where @xmath142 and moment generating function ( mgf ) given by@xmath143initially , we find the probability @xmath144 via wald s identity by using a standard technique ( see e.g. karlin and taylor ( 1975 ) , p.265 ) .
it can be verified that for @xmath145 we have that @xmath146 and therefore from ( [ eq13 ] ) ( or from ( [ eq26 ] ) ) we get @xmath147 hence , it follows that@xmath148and by solving with respect to @xmath144 we get @xmath149invoking the memoryless property of the exponential distribution we have that@xmath150and combining the above we deduce that for @xmath151 @xmath152for @xmath153 ( i.e. the case where @xmath154 ) we can take @xmath155 in the above formula and subsequently deduce that @xmath156 combining ( [ eq18 ] ) and ( [ eq19 ] ) we deduce the following proposition .
[ prop1]let @xmath175 be a random walk with step distribution @xmath176 where @xmath177 if @xmath2 denotes the time until the random walk exits @xmath178 then the probability generating function of @xmath2 is given by @xmath179where @xmath180 note that , for the special case @xmath181 the above generating function can also be derived by employing results established by khan ( 2008 ) .
apart * * * * from * * * * its theoretical interest , the above formula can also be used for the numerical determination of the distribution of @xmath2 for given values of the parameters @xmath182 and @xmath183 since@xmath184 in practice , this can be easily accomplished by the use of appropriate mathematical software ( e.g. using the function ` seriescoefficient ` of wolfram mathematica ) . in figure 1 the distribution of @xmath2 has been pictured for two sets of values of the parameters .
the height of the bars represent the probabilities @xmath185 while the small dots show the corresponding probabilities estimated by monte carlo simulation after @xmath186 iterations .
( @xmath187 @xmath188 @xmath189 @xmath190 @xmath191 and @xmath192),title="fig : " ] ( @xmath187 @xmath188 @xmath189 @xmath190 @xmath191 and @xmath192),title="fig : " ] an explicit formula for @xmath131 can be easily derived by differentiating @xmath193 given in proposition [ prop1 ] , with respect to @xmath194 and taking @xmath195 .
the details are left to the reader .
employing corollary [ cor2 ] , we can derive the joint gf of @xmath2 and @xmath4 , which yields @xmath196where @xmath197 and @xmath108 are given above .
moreover , for the pgf of the conditional distribution of @xmath2 , given that the random walk crossed the upper boundary , we observe that ( [ eq20 ] ) with @xmath198},~x=0,$ ] leads to@xmath199})=\mathbb{e}_{w_{u}}\left ( e^{-w_{u}s_{t}}i_{[s_{t}\geq b]}\right ) \\ & = & \mathbb{\tilde{e}}_{w_{u}}\left ( e^{-w_{u}s_{t}}|s_{t}\geq b\right ) \mathbb{\tilde{p}}_{w_{u}}(s_{t}\geq b).\end{aligned}\]]therefore we deduce the following result . [
prop2]let @xmath175 be a random walk with step distribution @xmath176 where @xmath200 if @xmath2 denotes the time until the random walk exits @xmath178 then the conditional pgf of @xmath2 given that @xmath201 , is @xmath202where @xmath203 and @xmath204 are as in ( [ eq27]),([eq16 ] ) and ( [ eq19 ] ) respectively .
proposition [ prop2 ] along with ( [ eq25 ] ) can be used for the calculation of the conditional probabilities @xmath205 . in figure 2 , which was constructed similarly to figure 1 , we have plotted the conditional distribution of @xmath2 for two sets of values of the parameters@xmath29 of @xmath206 given that @xmath207 ( @xmath208 @xmath209 @xmath210 @xmath211 @xmath212 and @xmath213),title="fig : " ] of @xmath206 given that @xmath207 ( @xmath208 @xmath209 @xmath210 @xmath211 @xmath212 and @xmath213),title="fig : " ] finally , it is worth mentioning that when the @xmath22 s follow a laplace distribution ( i.e. , @xmath214the pgf of @xmath2 takes on the simple form@xmath215where @xmath216 also , @xmath129 now simplifies to@xmath217while the conditional pgf of @xmath2 now reads@xmath218finally , by differentiating @xmath193 and @xmath219 , with respect to @xmath194 , taking @xmath220 and after some algebraic manipulations , we may also easily * * * * derive explicit formulae for @xmath131 , @xmath221 and @xmath219 . *
( b ) * we consider again the random walk @xmath1 discussed in ( a ) with @xmath26 ( i.e. now there exists only an upper barrier ) , that is @xmath2 denotes the waiting time ( steps ) until the random walk crosses @xmath222 .
exploiting the results of section 2 , we find the probability @xmath223 and the conditional pgf of @xmath2 given that @xmath224 . in this case , @xmath225 only when the mean step @xmath226 is positive .
we conveniently observe that the mean step under the probability measure @xmath110 is always positive , that is , @xmath227for all @xmath162 this can be justified as follows : note first that @xmath108 is strictly decreasing for @xmath159 $ ] with @xmath228 and @xmath229 it suffices to show that @xmath230 where @xmath231 the function @xmath232 is strictly increasing in @xmath233 $ ] ( @xmath234 for @xmath235)$ ] .
we examine the following three cases : 1 .
if @xmath236 , then @xmath237 and hence @xmath238 2 . if @xmath239 then @xmath240 and hence @xmath241 3 . if @xmath242 then directly , @xmath243 therefore , @xmath244 and from relation ( [ eq20 ] ) we deduce that@xmath245
} ) & = & \mathbb{\tilde{e}}_{w_{u}}(e^{-w_{u}s_{t}}i_{[t<\infty ] } ) = \mathbb{\tilde{e}}_{w_{u}}(e^{-w_{u}s_{t}}|t<\infty ) \mathbb{\tilde{p}}_{w_{u}}(t<\infty ) \\ & = & \mathbb{\tilde{e}}_{w_{u}}(e^{-w_{u}s_{t}}|t<\infty ) = \frac{(\theta _ { 1}-w_{u})e^{-w_{u}b}}{\theta _ { 1}},u\in ( 0,1).\end{aligned}\]]letting @xmath195 we get that @xmath246 since @xmath247})=\mathbb{e}(u^{t}|t<\infty ) \mathbb{p}(t<\infty ) $ ] we readily deduce the following proposition . [ prop3]let @xmath175 be a random walk with step distribution @xmath176 where @xmath177 if @xmath2 denotes the time until the random walk crosses @xmath248 then the conditional pgf of @xmath2 given that @xmath224 is @xmath249where @xmath250is as in ( [ eq27 ] ) .
moreover,@xmath251 by employing proposition [ prop3 ] we can easily compute the conditional probabilities @xmath252 through ( [ eq25 ] ) . in figure 3 the conditional probabilities @xmath253 have been plotted for two sets of values of the parameters . of @xmath206
given that @xmath224 ( @xmath254 @xmath255 @xmath210 @xmath256 and @xmath257),title="fig : " ] of @xmath206 given that @xmath224 ( @xmath254 @xmath255 @xmath210 @xmath256 and @xmath257),title="fig : " ] in the first case we have that @xmath258 while in the second case @xmath225 .
in the current paragraph we present an application in acceptance sampling which is a major component of the field of statistical process control . in acceptance sampling
we frequently deal with sampling systems / plans that have at least two sampling levels controlled by switching rules that are based on run and scan statistics .
two examples of such systems are the continuous sampling plans ( see , for example , schilling and neubauer ( 2009 ) ) and the military standard 105e ( see , for example , montgomery ( 2005 ) ) . in acceptance sampling for attributes
we take samples of fixed size corresponding to consecutive lots of items from a manufacturing process and we record the number @xmath31 @xmath135 of non - conforming ( defective ) items in the @xmath7-th sample . let @xmath259 be the acceptance number of the normal sampling level , that is a lot is rejected if the corresponding sample contains more than @xmath259 non - conforming items . assume that a switch in a more tightened ( reduced ) sampling level is instituted when each one of @xmath9-consecutive samples have more than ( less than or equal ) @xmath259 non - conforming items .
we denote by @xmath2 the waiting time ( i.e. number of lots ) until the sampling level of the inspection changes .
our aim is to obtain the joint pgf of @xmath2 and @xmath4 by exploiting the fact that @xmath2 follows a known distribution .
the study of the random variable @xmath4 is crucial , especially under a rectifying inspection program . in the sequel we deal with a sampling system that begins under the normalsampling level and a switch
is permitted only to the tightened one .
more specifically , suppose that the size of the samples is fixed and equal to @xmath260 and that the probability of an item being defective is equal to @xmath261 therefore , each @xmath31 @xmath135 follows a binomial distribution with parameters @xmath262 @xmath263 the number @xmath2 of inspected lots until the tightened sampling level is instituted can be expressed as@xmath264the stopped sum @xmath265 expresses the total number of defective items found until switching to the tightened sampling level . since @xmath22 s are discrete rv s we can conveniently set @xmath266 in corollary [ cor1 ] to get the following relation for the joint pgf of @xmath267@xmath268where @xmath269 .
the distribution of the @xmath22 s under the probability measure @xmath270 is @xmath271therefore , under @xmath272 @xmath22 follows a binomial distribution , with parameters @xmath260 and @xmath273the stopping time @xmath2 can be considered as the first time a success run of length @xmath9 occurs in a sequence of independent trials with success probability @xmath274 hence , @xmath224 and the distribution of @xmath2 is known as the geometric distribution of order @xmath9 ( see , for example , philippou et al .
( 1983 ) or balakrishnan and koutras ( 2002 ) ) with pgf given by , @xmath275 . \label{eq24}\ ] ] under the probability measure @xmath270 we have@xmath276and thus , @xmath277 is given by ( [ eq24 ] ) , by replacing @xmath278 with @xmath279 taking into account this observation , equality ( [ eq23 ] ) leads to the following formula for the joint pgf of @xmath267 @xmath280for all @xmath159 $ ] and @xmath281 $ ] guaranteeing that @xmath282 $ ] and @xmath283 as required by ( [ eq24 ] ) and ( [ eq28 ] ) .
the pgf @xmath284 follows readily from the above by setting @xmath285 the distribution of @xmath23 which has support @xmath286 can be numerically evaluated for specific values of the parameters @xmath287 and @xmath9 as described after formula ( [ eq25 ] ) . using this procedure we calculate @xmath288for
two sets of the parameters and the results are shown in figure 4 .
( @xmath289 @xmath290 @xmath291 @xmath292 and @xmath293),title="fig : " ] ( @xmath289 @xmath290 @xmath291 @xmath292 and @xmath293),title="fig : " ] it should also be mentioned , that since @xmath4 is a positive integer - valued rv , the generating function @xmath294 of the tail probabilities can be easily determined via the formula @xmath295 the tail probabilities of the distribution of @xmath4 can be used inpractice for the determination of the parameters of the above mentioned sampling plan . for various combinations of @xmath259 and @xmath9 it would be interesting to know the probability that the total number of defective items until switching exceeds a certain threshold .
for example , consider the case where @xmath296 and @xmath297 for @xmath298 , equation ( [ eq29 ] ) provides the pgf of @xmath4 from which , by differentiation , we get that @xmath299 ( note that @xmath300 can also be evaluated via wald s first equation ) . moreover ,
using @xmath301 , we can compute the percentile points of the distribution of @xmath4 , which provide complete knowledge about the performance of the sampling plan , in terms of the total number of defective items found until the switching .
since in that case the median of the distribution of @xmath4 is 100 , we deduce that there is a probability lower than 50% that the total number of defective items will exceed 100 until switching .
it is worth mentioning that the above procedure could easily be expressed in a more general setting .
for example , if the measurements @xmath31 @xmath135 from the inspected lots follow a general distribution with cdf @xmath302 ( continuous , discrete or mixed ) and a switching sampling level occurs at time @xmath2 according to some stopping rule ( e.g. a @xmath303 scan rule ) , then following the methodology described above we can similarly determine the joint generating function of @xmath8 provided that the pgf of @xmath2 is known ( e.g. is a geometric distribution of order @xmath304 see balakrishnan and koutras ( 2002 ) ) . in this respect
we state without proof the following proposition .
let @xmath305 be a sequence of iid measurements following a distribution @xmath302 and let @xmath2 be the waiting time ( i.e. number of @xmath22 s ) until a switching sampling level occurs based on the @xmath303 scan switching rule : @xmath9 out of @xmath306 consecutive @xmath22 s belong to a specific measurable set @xmath307 if @xmath308 denotes the pgf of the geometric distribution of order @xmath303 with success probability @xmath309 then@xmath310for all @xmath104 such that @xmath311 and @xmath312 the interested reader who wishes to study the general sampling system which permits a switch from the normal sampling level to the tightened or to the reduced sampling level may consult ebneshahrashoob and sobel ( 1990 ) for the pgf of the associated waiting time rv @xmath30 in this last subsection we present an interesting application of the formula of @xmath313 obtained above ( cf .
( [ eq29 ] ) ) , regarding the estimation of the probability @xmath10 of an item being defective .
assume that @xmath314 independent inspections are conducted according to the @xmath9-run switching rule described above and let @xmath315 be the waiting time ( i.e. number of lots ) until the sampling level of the @xmath7-th inspection changes , @xmath316 . denote also by @xmath317 the total number of defective items found until switching to the tightened sampling level has occurred in the @xmath7-th inspection , @xmath318 we are interested in estimating @xmath10 when only the sample values @xmath319 of the @xmath314 aforementioned waiting times are available .
since the likelihood function @xmath320does not have a convenient form in order to directly find the mle of @xmath321 we will show how we can alternatively employ an em algorithm , considering @xmath322 as missing values ( latent variables ) .
the likelihood function @xmath323 now has the simple form @xmath324since @xmath325 is not available , we can find the mle of @xmath10 by iteratively applying the following two steps ( em algorithm ; cf .
dempster et al . ( 1977 ) ) : * ( e - step ) : * given @xmath326 and the estimate of @xmath10 at the @xmath327-th step , say @xmath328 , compute the conditional expected value of the log likelihood function , @xmath329the expected value @xmath330 can be calculated by @xmath331which can be derived from equation ( [ eq29 ] ) .
more specifically , the sum @xmath332 is the coefficient of the @xmath18th order term in the power series expansion of @xmath333 ( where @xmath334 and @xmath335 can be derived from the series expansion of @xmath336 * ( m - step ) : * find the parameter @xmath337 that maximizes @xmath338 , i.e.,@xmath339 the above two steps are repeated until we achieve the desired accuracy in the estimate @xmath340 of @xmath10 ( e.g. @xmath341 where @xmath342 ) . from the above procedure we can also get an estimate , @xmath343 of the unobserved variable @xmath344 the observed fisher information , which can be exploited for establishing approximate confidence intervals for @xmath10 , takes on the form @xmath345 as an example of the above estimation procedure , suppose that a @xmath9-run switching rule is employed for @xmath346 inspections with @xmath347 and the resulted waiting times are:@xmath348(actually , these are simulated values with @xmath349 ) . by employing the em algorithm we obtain @xmath350 @xmath351 while the estimates of @xmath352 are@xmath353 the estimated standard error of @xmath340 is @xmath354 ( cf .
( [ eq34 ] ) ) and the approximate @xmath355 confidence interval for @xmath10 is @xmath356 .
* acknownledgement * the work of athanasios c. rakitzis is supported by the state scholarship foundation of greece . | let @xmath0be a stopping time associated with a sequence of independent random variables @xmath1 . by applying a suitable change in the probability measure we present relations between the moment or probability generating functions of the stopping time @xmath2 and the stopped sum @xmath3 .
these relations imply that , when the distribution of @xmath4 is known , then the distribution of @xmath2 is also known and vice versa .
applications are offered in order to illustrate the applicability of the main results , which also have independent interest . in the first one we consider a random walk with exponentially distributed up and down steps and derive the distribution of its first exit time from an interval @xmath5 in the second application we consider a series of samples from a manufacturing process and we let @xmath6 , denoting the number of non - conforming products in the @xmath7-th sample .
we derive the joint distribution of the random vector @xmath8 , where @xmath2 is the waiting time until the sampling level of the inspection changes based on a @xmath9-run switching rule .
finally , we demonstrate how the joint distribution of @xmath8 can be used for the estimation of the probability @xmath10 of an item being defective , by employing an em algorithm . *
key words and phrases * : stopping time , stopped sum , exponentially tilted probability measure , random walk , first exit time , boundary crossing probabilities , acceptance sampling , @xmath9-run switching rule , sooner waiting time distribution , joint generating function , em algorithm . *
ams 2000 subject classification * : [ cols= " < , < " , ] |
a stationary spherically symmetric wormhole may be defined as a handle or tunnel in a multiply - connected spacetime joining widely separated regions of the same spacetime or of different spacetimes @xcite .
interest in wormholes was renewed by the realization that much of our universe is pervaded by a dynamic dark energy that causes the universe to accelerate @xcite : @xmath0 is positive in the friedmann equation @xmath1 . in the equation of state ( eos )
@xmath2 , a value of @xmath3 is required for an accelerated expansion . the range of values @xmath4 is usually referred to as _ quintessence _ and the range @xmath5 as _ phantom energy_. in the latter case we get @xmath6 , in violation of the weak energy condition , considered to be a primary prerequisite for the existence of wormholes @xcite .
the special case @xmath7 corresponds to einstein s cosmological constant .
this value is sometimes called the _ cosmological constant barrier _ or the _ phantom divide_. the quintessence and phantom energy models taken together as a _ joint model _ and dubbed _ quintom _ for short @xcite suggests a single eos , @xmath2 , to cover all cases .
it is shown in ref .
@xcite , however , that @xmath8 can not cross the phantom divide , that is , in the traditional scalar field model , the eos can not cross the cosmological constant barrier .
( for further discussion , see refs .
the simplest quintom model involves two scalar fields , @xmath9 and @xmath10 , one quintessence - like and one phantom - like @xcite : @xmath11 so instead of @xmath8 in the eos , we will use @xmath12 and @xmath13 , respectively .
one reason for studying quintom dark energy is the bouncing universe @xcite , which provides a possible solution to the big - bang singularity .
an extension to the braneworld scenario is discussed in ref .
@xcite .
the purpose of this paper is to study various wormhole spacetimes that are supported by quintom dark energy .
we discuss the structure of such wormholes by means of an embedding diagram , as well as the junction to an external schwarzschild spacetime . both isotropic and anisotropic pressures are considered .
it is shown that in the isotropic models , the phantom - energy condition @xmath14 implies that @xmath15 , while in the anisotropic case , @xmath16 . as a final comment , since quintessence satisfies the weak and null energy conditions , it is not feasible to model a wormhole using the quintessence field .
so when considered as separate fields , only phantom energy can support a wormhole structure .
in the present study the metric for a static spherically symmetric wormhole spacetime is taken as @xmath17 where @xmath18 and @xmath19 are functions of the radial coordinate @xmath20 .
let us now consider a quintom model , which , as noted earlier , contains a quintessence - like field and a phantom - like field .
so we assume that the einstein field equations can be written as @xmath21 where @xmath22 is the energy - momentum tensor of the quintessence - like field and is characterized by a free parameter @xmath12 , which is ordinarily restricted by the condition @xmath23 .
( recall that this condition is required for an accelerated expansion . ) according to kiselev @xcite , the components of this tensor need to satisfy the conditions of additivity and linearity . taking into account the different signatures used in the line elements , the components can be stated as follows : @xmath24 @xmath25
the energy - momentum tensor compatible with spherically symmetry is @xmath26 as already noted , a phantom - like field is characterized by the equation of state @xmath27 where @xmath28 is the radial pressure and @xmath14 . in the discussion below
, @xmath29 is the lateral pressure .
the einstein field equations in the orthonormal frame are stated next : @xmath30 + \frac{1}{r^{2 } } = 8\pi g
( \rho + \rho_q ) , \label{eq3}\ ] ] @xmath31 -\frac { 1}{r^{2 } } = 8\pi g ( p_r - \rho_q ) , \label{eq3}\ ] ] @xmath32\ ] ] @xmath33 @xmath34
for our first model we assume a constant redshift function , @xmath35 referred to as the zero - tidal - force solution in ref .
the absence of tidal forces automatically satisfies a key traversability criterion .
our first assumption in the present model is an isotropic pressure : @xmath36 adding eqs .
( 7 ) and ( 8) and using eqs . ( 6 ) and ( 10 ) , we get @xmath37 = 8\pi g(\omega_{ph}+1 ) \rho.\ ] ] multiplying eq .
( 8) by @xmath38 and adding to eq .
( 9 ) leads to @xmath39 where @xmath40 the above equation yields @xmath41 where @xmath42 is an integration constant .
we rewrite the metric in the morris - thorne canonical form @xcite , @xmath43 $ ] , where the shape function is given by @xmath44 using eqs .
( 7 ) and ( 8) , one gets the following forms for @xmath45 and @xmath46 : @xmath47 and @xmath48 observe that @xmath49 , since @xmath50 , while @xmath51 implies that @xmath52 .
it follows that @xmath53 .
the assumption @xmath54 implies the absence of a horizon .
also , we would like the wormhole spacetime to be asymptotically flat , that is , @xmath55 as @xmath56 . to this end
, we require that @xmath57 . from eq . ( 13 ) we deduce that @xmath58 since @xmath14 , it now follows that @xmath15 , thereby having crossed the phantom divide .
this result is hardly surprising , given the nature of phantom wormholes .
on the other hand , the quintessence condition @xmath59 is still going to occur , namely in the anisotropic case , discussed next . in the case of an anisotropic pressure ,
the radial and lateral pressures are no longer equal . in an earlier paper on phantom - energy wormholes ,
zaslavskii @xcite proposed the form @xmath60 , @xmath61 , for the lateral pressure . in this manner
we obtain simple linear relationships between pressure and energy - density , but with @xmath28 not equal to @xmath29 .
observe first that eq .
( 11 ) remains the same . after multiplying eq .
( 8) by @xmath38 and adding to eq .
( 9 ) , we get , analogously , @xmath62 where @xmath63 otherwise eqs .
( 14)-(17 ) retain their form .
as before , we want @xmath64 .
so from eq .
( 19 ) , we have @xmath65 and @xmath66 since @xmath61 and @xmath14 , it follows that @xmath67 which is the condition for quintessence . in other words , in the anisotropic model , @xmath12 does not have to cross the phantom divide .
_ remark : _ the parameter @xmath68 could be negative . for example , if @xmath69 , we return to @xmath15 .
in this section we let @xmath70 . returning to the shape function @xmath71 , to meet the condition @xmath72
, we must have @xmath73 .
so the radius of the throat is @xmath74 and @xmath75 since @xmath76 , it now follows that @xmath77 , thereby satisfying the flare - out condition . from ref .
@xcite , we therefore obtain the exoticity condition " @xmath78 ^ 2}= \frac{a}{2r_0 ^ 2}>0,\ ] ] which shows that the weak energy condition has been violated .
we already checked the asymptotic flatness , so that our solution describes a static traversable wormhole supported by quintom dark energy .
as discussed in ref .
@xcite , one can picture the spacial shape of a wormhole by rotating the profile curve @xmath79 about the @xmath80axis .
this curve is defined by @xmath81 for example , choosing @xmath82 , we find that @xmath83.\ ] ] the profile curve is shown in figure 1 and the embedding diagram in figure 2 .
the proper distance @xmath84 from the throat to a point outside is given by axis . ]
@xmath85 for @xmath82 , @xmath86 ( see figure 3 . ) . ]
it is customary to join the interior solution of a wormhole to an exterior schwarzschild solution at some @xmath87 .
to do so , we demand that @xmath88 and @xmath89 be continuous at @xmath87 : @xmath90 and @xmath91 @xmath92 and @xmath93 are already continuous @xcite .
so @xmath94 and @xmath95 .
this , in turn , implies that @xmath96 .
hence , the matching occurs at @xmath97 the interior metric @xmath98 is given by @xmath99dt^2 + \frac{dr^2 } { 1-d/ r^{a}}\\ + r^2(d\theta^{2 } + \text{sin}^{2}\theta d\phi^{2 } ) \end{gathered}\ ] ] and the exterior metric @xmath100 by @xmath101dt^2 + \frac{dr^2}{1-d/ a^{a-1}r}\\ + r^2(d\theta^{2 } + \text{sin}^{2}\theta d\phi^{2 } ) .
\end{gathered}\ ] ]
returning to the isotropic case @xmath102 , let us eliminate @xmath45 and @xmath46 in eqs .
( 7 ) - ( 9 ) to obtain the following master equation : @xmath103 + + + where @xmath104 and @xmath105 as in eq .
( 13 ) , @xmath106 now we choose the shape function @xmath107 in such a way that the right - hand side of eq .
( 28 ) is zero . for this specific choice ,
one gets the following solution : @xmath108 where @xmath42 is an integration constant .
fortunately , this form is the same as in model 1 .
so the physical characteristics , such as the profile curve , the embedding diagram , and the proper radial distance , remain the same .
but the redshift function and stress - energy components are different . by making the proper substitutions , one gets from eq .
( 28 ) @xmath109 = -\frac{(\nu^{\prime})^{2}}{2},\ ] ] where @xmath110 solving this equation , we get @xmath111 ^ 2,\ ] ] where @xmath112 is an integration constant .
we have used the condition @xmath113 . starting with @xmath114 and simplifying
, one can readily deduce that @xmath115 ; so once again , @xmath15 , thereby having crossed the phantom divide .
this is consistent with the isotropic case discussed in subsection 3.1 .
( there is a similar consistency with the anisotropic case , as shown at the end of the present section . ) finally , we get the following forms for @xmath45 and @xmath46 : @xmath116\end{gathered}\ ] ] and @xmath117 @xmath118.\ ] ] @xmath34 in the anisotropic case @xmath60 , @xmath61 , eliminating @xmath45 and @xmath46 in eqs .
( 7)-(9 ) yields @xmath119 where @xmath120 and @xmath121 here @xmath122 is defined in eq .
( 19 ) . in eq .
( 33 ) , @xmath123 is replaced by @xmath122 and @xmath124 by @xmath125 from the condition @xmath126 , we deduce that @xmath127 since @xmath14 and @xmath61 , we obtain , once again , @xmath128
the combination of quintessence and phantom energy in a joint model is referred to as quintom dark energy .
the quintessence - like field is characterized by a free parameter @xmath12 with the restriction @xmath59 . for the corresponding free parameter @xmath13 in the phantom - like field ,
the condition is @xmath14 .
we have proposed in this paper that traversable wormholes may be supported by quintom dark energy .
two models were considered .
the first model , a constant redshift function , leads to the determination of the shape function @xmath129 , which meets the flare - out conditions .
the resulting spacetime is asymptotically flat .
this was followed by a brief discussion of the wormhole structure , including an embedding diagram , proper distance , and a junction to an external schwarzschild spacetime . for the second , more general model , it is possible to use the same shape function but with a different redshift function and stress - energy components . in each of these models ,
both isotropic and anisotropic pressures were considered . in the isotropic case
, the phantom - energy condition @xmath14 implies that @xmath15 , and in the anisotropic case , @xmath14 implies that @xmath59 .
morris and k.s .
thorne , am .
56 , 395 ( 1988 ) .
riess,_et al . _ ,
j. 116 , 1009 ( 1998 ) .
perlmutter,_et al . _ ,
j. 517 , 565 ( 1999 ) .
zaslavskii , phys .
d 72 , 061303 ( 2005 ) .
lobo , phys . rev .
d 71 , 084011 ( 2005 ) .
kuhfittig , class .
quantum grav .
23 , 5853 ( 2006 ) ; f. rahaman,_et al .
_ , phys . lett .
b 633 , 161 ( 2006 ) ; f. rahaman , _ et al . _ ,
scripta 76 , 56 ( 2007 ) ; f. rahaman , _ et al .
39 , 145 ( 2007 ) .
b. feng , x. wang , and x. zhang , phys .
b 607 , 35 ( 2005 ) .
xia , y .- f .
qiu , g .- b .
zhao , and x .- m .
zhang , arxiv : astro - ph/0703202 .
zhao , j .- q .
xia , m. li , b. feng , and x. zhang , phys .
d 72 , 123515 ( 2005 ) . r.r .
caldwell and m. doran , phys .
d 72 , 043527 ( 2005 ) .
a. vikman , phys . rev .
d 71 , 023515 ( 2005 ) .
w. hu , phys . rev .
d 71 , 047301 ( 2005 ) .
m. kunz and d. sapone , phys . rev .
d 74 , 123503 ( 2006 ) .
cai , t. qiu , r. brandenberger , y .- s .
piao , and x. zhang , jcap 0803 , 013 ( 2008 ) .
guo , y .- s .
piao , x. zhang , and y .- z .
zhang , phys .
b 608 , 177 ( 2005 ) . y .-
cai , t. qiu , y .- s .
piao , m. li , and x. zhang , jhep 0710 , 071 ( 2007 ) .
w. zhao and y. zhang , phys . rev .
d 73 , 123509 ( 2006 ) .
p. moyassari and m.r .
setare , phys . lett .
b 674 , 237 ( 2009 ) .
kiselev , class .
20 , 1187 ( 2003 ) .
lemos , f.s.n .
lobo , and s.q .
de oliveira , phys .
d 68 , 064004 ( 2003 ) . | the combination of quintessence and phantom energy in a joint model is referred to as quintom dark energy .
this paper discusses traversable wormholes supported by such quintom matter .
two particular solutions are explored , a constant redshift function and a specific shape function .
both isotropic and anisotropic pressures are considered . |
there are many 6d superconformal field theories ( scfts ) engineered from string theory .
the first examples were the maximally supersymmetric @xmath8 scfts of @xmath9 types , discovered in @xcite from considering iib string theory compactified on k3 .
the @xmath10-type theory governs the low energy dynamics of parallel and overlapping @xmath11 m5-branes @xcite .
vast number of @xmath12 scfts were found from f - theory compactified on elliptic calabi - yau 3-folds as well as from branes at low energy @xcite .
the classification of 6d @xmath0 scfts has been recently explored in @xcite .
the tensor branch of each 6d @xmath0 scft can be described in terms of tensor multiplets , hypermultiplets , and vector multiplets .
when there exists a gauge symmetry , the gauge anomaly should vanish .
each tensor multiplet can have self - dual strings .
they are the sources for the 2-form tensor field @xmath13 of the tensor multiplet , whose field strength @xmath14 respects the self - duality @xmath15 .
their tensions are proportional to the vev of a tensor multiplet scalar .
one interesting example is the self - dual strings of the 6d @xmath16 @xmath10 scft , which can be described by m2-branes connecting any two m5-branes of @xmath11 m5-branes @xcite .
they are the so - called ` m - strings ' which have been studied in various literatures such as @xcite .
another typical example is the self - dual strings of the 6d @xmath0 scft with @xmath3 global symmetry , being induced from m2-branes connecting m5-branes to an @xmath3 boundary wall of m - theory on @xmath17 @xcite .
these so - called ` e - strings ' have been studied in , e.g. , @xcite .
we shall often use the terminologies ` m - string theory ' and ` e - string theory ' to refer the @xmath16 @xmath10 scft and the @xmath0 scft with @xmath3 flavor symmetry , respectively .
there are a lot of 6d @xmath0 scfts with gauge symmetries . when a gauge group @xmath18 is semi - simple , @xcite search for all possible gauge theories being free of gauge anomaly .
if one considers some low - rank simple groups , i.e. , @xmath19 and @xmath1 , the allowed matter contents are given by @xmath20 apart from the minimal @xmath1 case with @xmath21 , the above theories are obtained by enhancing m - string and e - string theories with gauge symmetries and extra matters .
one can consult @xcite for further possible enhancements of m - string and e - string theories found from f - theory construction . in this paper
, we shall focus on the following sequences of 6d scfts : @xmath22 which are connected by higgs mechanism .
our main interest in this work is to study the physics of self - dual strings in these theories by constructing their 2d gauge theory description . for each sequence of 6d scfts
, we shall find a class of 2d gauge theories which provides a uniform description of self - dual strings .
`` decorations '' of the basic 6d scfts and self - dual strings , such as m - strings and e - strings , by 6d gauge symmetries and matters are important to fully understand the recent classification @xcite better . and also
, we shall discover new gauge theory description of e - strings themselves , which will exhibit some strong coupling properties of these strings more transparently than @xcite . recall that e - string theory is engineered from m5-branes probing an @xmath3 wall , which is also called m9-brane .
the m9-brane worldvolume has 4 transverse directions to m5-branes , rotated by the @xmath23 symmetry which can be decomposed into @xmath24 .
while all @xmath25 charged states decouple from e - string theory on a single m5-brane , the @xmath25 refinement turns out crucial in case of multiple m5-branes .
our new description of e - strings fully realizes both @xmath25 and @xmath26 in contrast to @xcite which only see @xmath27 , enabling us to study the refined spectrum of multiple m5-branes probing an m9-brane in section [ sec : multiple - m5-branes ] .
let us explain a useful bottom - up approach for finding the 2d gauge theories for self - dual strings .
when a 6d scft has a gauge symmetry , the equation of motion for @xmath13 is given by @xcite @xmath28 where @xmath29 is the field strength of a gauge field , and @xmath30 is a positive constant which depends on a theory @xcite .
self - dual strings are therefore regarded as instanton soliton strings in effective yang - mills description of 6d scfts in the tensor branch .
these strings carry nonzero integer charges of @xmath31 we shall consider self - dual instantons with @xmath32 in this paper .
this integral is taken over the @xmath33 space transverse to the string worldsheet .
if one applies the moduli space approximation to describe the low energy excitations on these strings , the dynamics of self - dual strings are described by non - linear sigma models , whose target space is given by the instanton moduli space @xmath34 .
the space @xmath34 has singular locus at which the instanton size shrinks to zero .
this small instanton singularity reflects the uv incompleteness of the 6d effective yang - mills theory .
although it is difficult to deal with the singular target space , the adhm construction @xcite often tells us the sensible uv completions of the non - linear sigma models . with the guidance from the adhm construction ,
one may obtain 2d @xmath4 adhm gauge theories which are weakly coupled in the uv regime .
the instanton moduli space @xmath34 reappears in the higgs branch of the adhm gauge theory @xcite away from the points where the instanton scale becomes zero .
one main problem that we discuss in section [ sec : strings - from - branes ] is that the adhm construction sometimes leads us to 2d gauge theories that suffer from gauge anomaly .
while there is no such issue for self - dual strings in the sequence to m - strings , this happens when we consider 6d scfts in the higgsing chain to e - string theory .
we find anomaly - free gauge theories by suitably modifying the naive adhm gauge theories , which include the same 6d instanton zero modes .
these anomaly - free theories have the desired non - linear sigma models on @xmath34 away from the small instanton singularity , but provide a uv completion at the singularity .
one important feature is that some global symmetries which we expect on the self - dual strings are partly broken in these gauge theories .
namely , the 6d symmetries @xmath35 are sometimes partially present in the 2d gauge theories .
the elliptic genera of these consistent 2d theories capture the bps excitations on self - dual strings and remain valid in the strong coupling low energy limit where the 6d global symmetries should get manifest .
the calculation of the elliptic genera and showing the global symmetry enhancement in the index functions are done in sections [ sec : ir - dynamics - symmetry ] and [ sec : multiple - m5-branes ] .
the outline of this paper is as follows . in section [ sec : strings - from - branes ] , we construct 2d anomaly - free gauge theories for self - dual strings in 6d scfts listed at and .
we check that they show the ir symmetry enhancements by computing the gauge theory elliptic genera in section [ sec : ir - dynamics - symmetry ] .
section [ sec : multiple - m5-branes ] is devoted to study e - string theories with higher dimensional tensor branches , i.e. for multiple m5-branes probing an m9-brane .
we obtain the fully flavored spectrum which refines the result of @xcite .
concluding remarks are given in section [ sec : conclusion ] .
in this section , we construct the anomaly - free @xmath5 gauge theories for @xmath36 self - dual strings in the following sequences of 6d gauge theories : @xmath37 it is tightly related to finding a suitable uv completion of the non - linear sigma model with the singular target space . as this procedure
has not been clearly understood , we shall refer to the string theory realization of 6d scfts and self - dual strings as our guide for a possible uv completion .
the higgsable theories in are readily realized in type iia string theory .
suppose that there exist @xmath38 overlapping d6-branes along the @xmath39 directions .
two parallel ns5-branes along the @xmath40 directions are located on these d6 branes and separated in the @xmath41 direction .
there are @xmath36 overlapping d2-branes along the @xmath42 directions connecting two ns5-branes .
these branes are mutually bps and so preserve some supersymmetry .
figure [ fig : m - string - higgsing - sequence ] shows the brane set - up . the 7d @xmath43 susy gauge theory on @xmath38 d6-branes implies that the low energy physics between two ns5-branes is described by the 6d @xmath0 @xmath44 gauge theory with @xmath45 fundamental hypermultiplets , which is coupled to the tensor multiplet from the relative motion of two ns5-branes .
the reduction mechanism of the 6d gauge group from @xmath43 to @xmath44 is explained in @xcite .
the 6d @xmath46 gauge symmetry with charged matters was anomalous , as only hypermultiplets could contribute to the @xmath46 gauge anomaly .
one can however obtain a sensible theory through the ` linear hypermultiplet ' having four scalars @xmath47 and @xmath48 , which correspond to the position coordinates of ns5-branes along @xmath49 and m - circle directions .
the @xmath47 appear as parameters in the d - term and the @xmath48 appears as @xmath50 .
the kinetic terms for linear hypermultiplet scalars are given by @xcite @xmath51 where the @xmath46 gauge invariance requires @xmath48 to transform as @xmath52 under @xmath53 .
this shows that the @xmath46 gauge field is massive and the @xmath46 gauge symmetry is spontaneously broken .
the low energy physics is therefore governed by the @xmath44 gauge theory with @xmath45 fundamental hypermultiplets .
this 6d theory is anomaly - free .
instanton strings of the 6d @xmath44 gauge theory are also self - dual strings of the tensor multiplet .
these self - dual strings are realized as @xmath36 d2-brane segments connecting two ns5-branes , lying along the @xmath42 directions and on top of the @xmath38 d6-branes .
the separation of two ns5-brane is taken to be very short , so the worldvolume theory of @xmath36 d2-branes becomes effectively given by the 2d supersymmetric yang - mills theory .
the d6- and ns5-branes preserve @xmath54 global symmetry , which is further decomposed to @xmath55 due to the presence of d2-branes . here
@xmath56 is the r - symmetry of 6d @xmath0 susy .
we now examine the preserved supercharges in this system .
let us denote the doublet indices of @xmath57 symmetry by @xmath58 .
32 supercharges of type iia string theory are decomposed into @xmath59 and @xmath60 , in which the two @xmath61 subscripts denote eigenvalues of @xmath62 and @xmath63 respectively .
introduction of d2 , d6 , ns5-branes imposes the susy projectors @xmath64 , @xmath65 , @xmath66 , so that 4 supercharges @xmath67 are preserved .
since @xmath67 have a definite 2d chirality , they form the @xmath68 susy whose r - symmetry corresponds to @xmath69 .
we summarize the field contents of 2d @xmath5 gauge theory in figure [ fig:2d - m - string - quiver ] , which are determined from massless modes of open strings connecting d2-d2 and d2-d6 .
these match with the adhm construction . in @xmath5 quiver diagrams , we draw a solid line to represent a @xmath5 hypermultiplet ( either twisted or untwisted ) and a dashed line to represent a @xmath5 fermi multiplet .
we make a few comments on the adhm gauge theory .
first , the adhm gauge theory is originally intended for the @xmath43 instantons instead of @xmath44 .
it involves extra degrees of freedom , regularizing singular @xmath46 instantons @xcite .
second , @xmath4 chiral susys are preserved so that one should check if the adhm gauge theory is sensible by looking at a gauge anomaly . in the @xmath5 susy
, fermions are right - moving if they are in hypermultiplets , and left - moving if they are in vector or fermi multiplets . using the normalization @xmath70 ,
the gauge anomaly is proportional to @xmath71 which vanishes at @xmath72 .
we also need to consider a mixed anomaly between gauge and global symmetries .
let us denote the abelian generators of @xmath73 , @xmath74 , and @xmath75 by @xmath18 , @xmath29 , and @xmath76 .
gauge - global mixed anomalies are given as follows .
@xmath77 these imply that the 2d quantum theory at @xmath78 preserves the @xmath46 combination @xmath79 of @xmath80 only .
third , there exists the fake @xmath46 symmetry generated by @xmath81 , which rotates no fields in the theory . in the above brane
set - up , the higgs mechanism can be regarded as a process of decreasing the number of d6-branes .
if @xmath38 reaches to be @xmath82 , the 6d scft no longer has a gauge symmetry .
the m - theory uplift of the brane configuration is simply two m5-branes supported at the origin of taub - nut . in this case
, the adhm gauge theory turns out to be the @xmath5 gauge theory description for m - strings introduced in @xcite .
a further higgsing removes the last d6-brane , leaving @xmath36 d2-branes suspended between two ns5-branes .
it is also uplifted to m - theory as a pair of parallel m5-branes .
the susy gauge theory on d2-branes then becomes 2d @xmath83 @xmath84 gauge theory again describing m - strings . in order to study the spectrum of m - string theory , we prefer the @xmath5 description to the @xmath83 description .
it is because only the first one realizes the @xmath23 global symmetry of m - string theory , which rotates 4 transverse directions to m5-branes .
let us decompose it into @xmath85 . at @xmath86
, @xmath26 appears as a part of @xmath5 r - symmetry , and also @xmath25 appears as a @xmath87 global symmetry . at @xmath88 , only @xmath27
is realized as a subgroup of @xmath23 r - symmetry in @xmath5 susy .
we now move to the higgsing sequence to the e - string theory .
the 6d @xmath1 and @xmath2 scfts in have 6 more fundamental hypermultiplets than those 6d scfts in .
as a 6d fundamental hypermultiplet induces a @xmath5 fermi multiplet in the 2d adhm construction , one may naively attempt to describe @xmath36 self - dual strings by adding 6 more fermi multiplets to figure [ fig:2d - m - string - quiver ] .
however , this trial immediately fails because a gauge anomaly vanishes only if @xmath72 .
instead , we make the following observations : * @xmath1 fundamental hypermultiplets are not distinguishable from @xmath89 hypermultiplets .
they induce the same instanton zero mode
. nevertheless , they supply different degrees of freedom which are supported at the small instanton singularity in the instanton moduli space . * @xmath2 antisymmetric hypermultiplets decouple from the rest of perturbative dynamics
. however , inclusion of antisymmetric hypermultiplets changes the way of resolving the singularity in the instanton moduli space , by providing extra degrees of freedom supported at the singular locus .
here we take first the empirical approach in order to find a sensible uv completion of the 2d system .
introduction of a single 6d @xmath44 antisymmetric hypermultiplet supplies a single set of the following 2d @xmath5 supermultiplets @xcite : [ cols="^,^,^,^,^,^",options="header " , ] where the @xmath49 and @xmath90 directions were taken to be @xmath91 with a radius @xmath92 . in the 11d limit @xmath93 , e - string theory has the global @xmath94 symmetry which rotates four transverse directions . on the other hand ,
the iia system in @xcite was obtained by the @xmath95 limit which reduces the @xmath96 circle .
note that the @xmath3 wilson line is required to reach the iia configuration , replacing the m9 by the o8@xmath97 and 8 d8 branes . even though the @xmath96 direction would be decompactified in the infrared regime of 2d gauge theory through the following relation @xmath98
, one could not capture the @xmath94 flavored spectrum due to a missing cartan in the iia set - up .
in fact , @xcite computed the @xmath99 flavored spectrum where @xmath99 is identified as the diagonal subgroup @xmath100 . for a single m5-brane
, it does not matter because 5d @xmath101 instanton calculus , which we explain below , shows that the bps spectrum itself does not include @xmath25 charged states @xcite .
@xmath25 charged states begin to appear at @xmath102 .
5d @xmath103 instanton calculus tells us that the bps spectrum would be a product of @xmath104 @xmath101 partition functions if not refined with the @xmath25 global symmetry @xcite , which happens because many bps states cancel to 0 in the elliptic genus without @xmath25 refinement .
this calls for a new description in which the @xmath94 symmetry is fully visible .
we can instead take the @xmath49 and @xmath90 directions to be taub - nut .
m5-branes are supported at the origin of taub - nut , which locally looks like @xmath105 .
the iia brane set - up in figure [ fig : multiple - ns5-brane ] is obtained in the limit which reduces the taub - nut circle .
again , it is required to turn on the @xmath3 wilson line along a reduced circle , breaking @xmath3 to @xmath106 symmetry .
even though both iia systems in @xcite and figure [ fig : multiple - ns5-brane ] reach to the same e - string theory in the strong coupling regime , the iia system in figure [ fig : multiple - ns5-brane ] is more favorable because it realize both @xmath25 and @xmath26 cartans as a part of flavor symmetry and the r - symmetry of 6d @xmath0 susy .
it makes our 2d gauge theory , introduced in figure [ fig : e - string - multiple - quiver ] , more useful when we study the fully refined spectrum of e - string theory .
let us explain how 5d @xmath103 instanton calculus describes e - string theory .
if one chooses one of @xmath107 directions , shared by m5 and m9-branes , to be a circle @xmath108 while keeping @xmath49 and @xmath90 directions as non - compact @xmath105 , the iia brane system can be obtained by reducing @xmath108 along which the @xmath3 wilson line is turned on .
m5-branes are mapped to d4-branes .
m9-brane is replaced by 8 d8-branes on top of the o8@xmath97 plane , carrying the @xmath106 symmetry .
the low energy dynamics of @xmath104 d4-branes is described by 5d @xmath103 gauge theory .
it has @xmath109 fundamental and @xmath82 antisymmetric hypermultiplets .
a remarkable fact is that certain 5d gauge theories essentially capture 6d physics by incorporating non - perturbative objects called instantons @xcite .
recall that the low energy dynamics of 5d sym instantons are uv - completed as d0-branes on d4-branes in string theory .
as d0-branes are kaluza - klein momenta along the m - theory circle , one can interpret 5d sym instantons as kaluza - klein momenta of a circle compactified 6d scft .
this relation has been confirmed in certain examples .
for example , the instanton partition function of 5d @xmath110 sym displays the tensor branch spectrum of 6d m - string theory @xcite . along the same line , the instanton partition function of 5d @xmath111 @xmath101 gauge theory agrees with the tensor branch spectrum of 6d e - string theory on a single m5-brane @xcite .
one can surely extend this relation to e - string theory on @xmath104 m5-branes , obtaining 5d @xmath103 gauge theory on @xmath104 d4-branes .
the precise relation between the 5d @xmath103 instanton partition function and the 6d tensor branch spectrum is explained in @xcite .
we shall refer to @xcite when we compare the elliptic genera of the gauge theory in figure [ fig : e - string - multiple - quiver ] with the 5d @xmath103 instanton partition function .
below we are going to compute the elliptic genera of 2d gauge theories in figure [ fig:2d - quiver ] and figure [ fig : e - string - multiple - quiver ] .
the point of this computation is verifying the @xmath3 symmetry enhancement .
for the latter case , we content ourselves with the sample computation at @xmath112 which will be compared with the 5d @xmath113 instanton partition function @xcite .
[ [ one - m5-brane ] ] one m5-brane + + + + + + + + + + + + at @xmath114 , we bring the @xmath115 and @xmath116 elliptic genera computed in section [ sec : ir - dynamics - symmetry ] .
one sets @xmath117 using the fake @xmath46 symmetry , then the equation imposes @xmath118 .
taking advantage of the @xmath119 branching rules , e.g. , @xmath120 the bps spectrum at @xmath115 can be organized into @xmath3 representations as follows .
@xmath121 \nonumber\end{aligned}\ ] ] similarly at @xmath116 , we obtain @xmath122 .
\nonumber \ ] ] which shows the @xmath3 flavor symmetry .
note that the @xmath123 fugacity @xmath124 automatically disappears , as expected from the 5d @xmath101 instanton calculus @xcite .
these results agree with the elliptic genera computed from the 2d @xmath125 gauge theory in figure [ fig : higgsed - ok](b ) @xcite .
[ [ two - m5-branes ] ] two m5-branes + + + + + + + + + + + + + we compute the elliptic genus at @xmath126 , which can be written as @xmath127 @xmath128 are chemical potentials for global symmetries @xmath129 .
we also denote a @xmath130 chemical potential by @xmath131 , related to @xmath132 parameters @xmath133 by @xmath134 one first sets @xmath135 using the fictitious @xmath46 symmetry , then locks various @xmath46 s to others according to .
after all , two independent @xmath46 s can remain which have no mixed anomaly .
this imposes @xmath136 we write down all contributing jk - residues , with the choice of @xmath137 , in the following .
* @xmath138 .
@xmath139 * @xmath140 for @xmath141 ( @xmath142 ) @xmath143 where the sign factor @xmath144 is defined to be @xmath145 and @xmath146 .
* @xmath147 .
@xmath148 the condition is translated into @xmath149 and @xmath150 .
we also rearrange nine independent variables @xmath151 , @xmath152 , @xmath153 , @xmath131 , @xmath154 into @xmath155 in order that the @xmath156 global symmetry becomes manifest . @xmath157 and @xmath158 are chemical potentials for @xmath25 and @xmath46 symmetries , where the @xmath46 appears in the @xmath159 symmetry enhancement from @xmath160 .
we denote the corresponding fugacities by @xmath161 and @xmath162 .
organizing the spectrum , we use the @xmath163 branching rules , e.g. , @xmath164 after the @xmath159 flavor symmetry enhancement , we observe that the @xmath159 further enhances to @xmath3 .
if one utilizes the @xmath119 branching rules , @xmath165-expansion of the elliptic genus can be written in terms of @xmath3 characters . @xmath166\end{aligned}\ ] ] we checked that it agrees with 5d @xmath113 instanton partition function @xcite .
in this paper , we studied those 6d @xmath1 and @xmath2 scfts which can be higgsed to m - string theory and e - string theory .
we constructed the anomaly - free 2d @xmath5 gauge theories for self - dual strings , inspired from the adhm construction .
this issue is closely related to finding a proper uv completion of the non - linear sigma model which governs the low energy dynamics of instanton string solitons in 6d effective sym .
our 2d consistent gauge theories capture the global symmetries of self - dual strings in the ir limit , implying that they provide a uv description for 6d instanton strings .
similar to the e - string case , it would be interesting to check our results for @xmath2 and @xmath1 independently , for example , by considering the instanton partition function of the dual 5d theories obtained after the circle compactification @xcite .
our analysis can be applied to other 6d scfts and little string theories , which contains 6d supersymmetric gauge theory as effective descriptions in tensor branch .
the 2d elliptic genus calculation can be an interesting starting point for understanding the index function of the 6d @xmath0 scfts on @xmath167 .
we hope to come back to these issues in near future .
the work of sk is supported in part by the national research foundation of korea ( nrf ) grant nrf-2015r1a2a2a01003124 .
the work of kl is supported in part by the national research foundation of korea ( nrf ) grants no .
2006 - 0093850 .
e. witten , `` some comments on string dynamics , '' in _ future perspectives in string theory .
proceedings , conference , strings95 , los angeles , usa , march 13 - 18,1995_. 1995 .
http://arxiv.org/abs/hep-th/9507121[arxiv:hep-th/9507121 [ hep - th ] ] .
a. strominger , `` open p - branes , '' http://dx.doi.org/10.1016/0370-2693(96)00712-5[_phys . lett .
_ * b383 * ( 1996 ) 4447 ] , http://arxiv.org/abs/hep-th/9512059[arxiv:hep-th/9512059 [ hep - th ] ] .
d. r. morrison and c. vafa , `` compactifications of f theory on calabi - yau threefolds . 1
, '' http://dx.doi.org/10.1016/0550-3213(96)00242-8 [ _ nucl .
phys . _ * b473 * ( 1996 ) 7492 ] , http://arxiv.org/abs/hep-th/9602114[arxiv:hep-th/9602114 [ hep - th ] ] .
n. seiberg and e. witten , `` comments on string dynamics in six - dimensions , '' http://dx.doi.org/10.1016/0550-3213(96)00189-7[_nucl .
phys . _ * b471 * ( 1996 ) 121134 ] , http://arxiv.org/abs/hep-th/9603003[arxiv:hep-th/9603003 [ hep - th ] ] .
e. witten , `` phase transitions in m theory and f theory , '' http://dx.doi.org/10.1016/0550-3213(96)00212-x[_nucl .
* b471 * ( 1996 ) 195216 ] , http://arxiv.org/abs/hep-th/9603150[arxiv:hep-th/9603150 [ hep - th ] ] .
d. r. morrison and c. vafa , `` compactifications of f theory on calabi - yau threefolds . 2 .
, '' http://dx.doi.org/10.1016/0550-3213(96)00369-0 [ _ nucl
. phys . _ * b476 * ( 1996 ) 437469 ] , http://arxiv.org/abs/hep-th/9603161[arxiv:hep-th/9603161 [ hep - th ] ] . j. d. blum and k. a. intriligator , `` new phases of string theory and 6-d rg fixed points via branes at orbifold singularities , '' http://dx.doi.org/10.1016/s0550-3213(97)00449-5[_nucl . phys . _ * b506 * ( 1997 ) 199222 ] , http://arxiv.org/abs/hep-th/9705044[arxiv:hep-th/9705044 [ hep - th ] ] . i. brunner and a. karch , `` branes at orbifolds versus hanany witten in six - dimensions , '' http://dx.doi.org/10.1088/1126-6708/1998/03/003[_jhep_ * 03 * ( 1998 ) 003 ] , http://arxiv.org/abs/hep-th/9712143[arxiv:hep-th/9712143 [ hep - th ] ] .
a. hanany and a. zaffaroni , `` branes and six - dimensional supersymmetric theories , '' http://dx.doi.org/10.1016/s0550-3213(98)00355-1 [ _ nucl . phys .
_ * b529 * ( 1998 ) 180206 ] , http://arxiv.org/abs/hep-th/9712145[arxiv:hep-th/9712145 [ hep - th ] ] .
m. del zotto , j. j. heckman , a. tomasiello , and c. vafa , `` 6d conformal matter , '' http://dx.doi.org/10.1007/jhep02(2015)054[_jhep_ * 02 * ( 2015 ) 054 ] , http://arxiv.org/abs/1407.6359[arxiv:1407.6359 [ hep - th ] ] .
d. r. morrison and w. taylor , `` classifying bases for 6d f - theory models , '' http://dx.doi.org/10.2478/s11534-012-0065-4[_central eur .
j. phys . _
* 10 * ( 2012 ) 10721088 ] , http://arxiv.org/abs/1201.1943[arxiv:1201.1943 [ hep - th ] ] .
j. j. heckman , d. r. morrison , and c. vafa , `` on the classification of 6d scfts and generalized ade orbifolds , '' http://dx.doi.org/10.1007/jhep06(2015)017 , 10.1007/jhep05(2014)028[_jhep _ * 05 * ( 2014 ) 028 ] , http://arxiv.org/abs/1312.5746[arxiv:1312.5746 [ hep - th ] ] .
[ erratum : jhep06,017(2015 ) ] . j. j. heckman , d. r. morrison , t. rudelius , and c. vafa , `` atomic classification of 6d scfts , '' http://dx.doi.org/10.1002/prop.201500024[_fortsch .
* 63 * ( 2015 ) 468530 ] , http://arxiv.org/abs/1502.05405[arxiv:1502.05405 [ hep - th ] ] .
l. bhardwaj , `` classification of 6d n=(1,0 ) gauge theories , '' http://arxiv.org/abs/1502.06594[arxiv:1502.06594 [ hep - th ] ] .
a. klemm , w. lerche , p. mayr , c. vafa , and n. p. warner , `` selfdual strings and n=2 supersymmetric field theory , '' http://dx.doi.org/10.1016/0550-3213(96)00353-7[_nucl .
phys . _ * b477 * ( 1996 ) 746766 ] , http://arxiv.org/abs/hep-th/9604034[arxiv:hep-th/9604034 [ hep - th ] ] . j. h. schwarz , `` selfdual superstring in six - dimensions , '' http://arxiv.org/abs/hep-th/9604171[arxiv:hep-th/9604171 [ hep - th ] ] .
b. haghighat , a. iqbal , c. kozcaz , g. lockhart , and c. vafa , `` m - strings , '' _
arxiv.org _ no . 2 , ( may , 2013 ) 779842 , http://arxiv.org/abs/1305.6322v2[1305.6322v2 ] .
o. j. ganor and a. hanany , `` small e(8 ) instantons and tensionless noncritical strings , '' http://dx.doi.org/10.1016/0550-3213(96)00243-x[_nucl .
phys . _ * b474 * ( 1996 ) 122140 ] , http://arxiv.org/abs/hep-th/9602120[arxiv:hep-th/9602120 [ hep - th ] ] .
a. klemm , p. mayr , and c. vafa , `` bps states of exceptional non - critical strings , ''
_ arxiv.org _
( july , 1996 ) , http://arxiv.org/abs/hep-th/9607139v2[hep-th/9607139v2 ] .
j. a. minahan , d. nemeschansky , c. vafa , and n. p. warner , `` e strings and n=4 topological yang - mills theories , '' http://dx.doi.org/10.1016/s0550-3213(98)00426-x[_nucl .
_ * b527 * ( 1998 ) 581623 ] , http://arxiv.org/abs/hep-th/9802168[arxiv:hep-th/9802168 [ hep - th ] ] .
t. eguchi and k. sakai , `` seiberg - witten curve for the e string theory , '' http://dx.doi.org/10.1088/1126-6708/2002/05/058[_jhep_ * 05 * ( 2002 ) 058 ] , http://arxiv.org/abs/hep-th/0203025[arxiv:hep-th/0203025 [ hep - th ] ] .
j. kim , s. kim , k. lee , j. park , and c. vafa , `` elliptic genus of e - strings , '' http://arxiv.org/abs/1411.2324[arxiv:1411.2324 [ hep - th ] ] .
u. h. danielsson , g. ferretti , j. kalkkinen , and p. stjernberg , `` notes on supersymmetric gauge theories in five - dimensions and six - dimensions , '' http://dx.doi.org/10.1016/s0370-2693(97)00645-x[_phys .
* b405 * ( 1997 ) 265270 ] , http://arxiv.org/abs/hep-th/9703098[arxiv:hep-th/9703098 [ hep - th ] ] .
m. bershadsky and c. vafa , `` global anomalies and geometric engineering of critical theories in six - dimensions , '' http://arxiv.org/abs/hep-th/9703167[arxiv:hep-th/9703167 [ hep - th ] ] .
a. gadde , b. haghighat , j. kim , s. kim , g. lockhart , and c. vafa , `` 6d string chains , '' http://arxiv.org/abs/1504.04614[arxiv:1504.04614 [ hep - th ] ] .
m. berkooz , r. g. leigh , j. polchinski , j. h. schwarz , n. seiberg , and e. witten , `` anomalies , dualities , and topology of d = 6 n=1 superstring vacua , '' http://dx.doi.org/10.1016/0550-3213(96)00339-2[_nucl .
phys . _ * b475 * ( 1996 ) 115148 ] , http://arxiv.org/abs/hep-th/9605184[arxiv:hep-th/9605184 [ hep - th ] ] . n. seiberg , `` nontrivial fixed points of the renormalization group in six - dimensions , '' http://dx.doi.org/10.1016/s0370-2693(96)01424-4[_phys .
* b390 * ( 1997 ) 169171 ] , http://arxiv.org/abs/hep-th/9609161[arxiv:hep-th/9609161 [ hep - th ] ] .
m. f. atiyah , n. j. hitchin , v. g. drinfeld , and yu .
i. manin , `` construction of instantons , '' http://dx.doi.org/10.1016/0375-9601(78)90141-x[_phys .
lett . _ * a65 * ( 1978 ) 185187 ] .
e. witten , `` sigma models and the adhm construction of instantons , '' http://dx.doi.org/10.1016/0393-0440(94)00047-8[_j .
phys . _ * 15 * ( 1995 ) 215226 ] , http://arxiv.org/abs/hep-th/9410052[arxiv:hep-th/9410052 [ hep - th ] ] .
m. r. douglas and g. w. moore , `` d - branes , quivers , and ale instantons , '' http://arxiv.org/abs/hep-th/9603167[arxiv:hep-th/9603167 [ hep - th ] ] .
kim , s. kim , e. koh , k. lee , and s. lee , `` on instantons as kaluza - klein modes of m5-branes , '' http://dx.doi.org/10.1007/jhep12(2011)031 [ _ jhep _ * 12 * ( 2011 ) 031 ] , http://arxiv.org/abs/1110.2175[arxiv:1110.2175 [ hep - th ] ] . s. shadchin , _ on certain aspects of string theory / gauge theory correspondence_. phd thesis , orsay , lpt , 2005 .
http://arxiv.org/abs/hep-th/0502180[arxiv:hep-th/0502180 [ hep - th ] ] .
c. hwang , j. kim , s. kim , and j. park , `` general instanton counting and 5d scft , '' http://dx.doi.org/10.1007/jhep07(2015)063[_jhep_ * 07 * ( 2015 ) 063 ] , http://arxiv.org/abs/1406.6793[arxiv:1406.6793 [ hep - th ] ] .
f. benini , r. eager , k. hori , and y. tachikawa , `` elliptic genera of 2d @xmath168 = 2 gauge theories , '' http://dx.doi.org/10.1007/s00220-014-2210-y[_commun .
phys . _ * 333 * no . 3 , ( 2015 )
12411286 ] , http://arxiv.org/abs/1308.4896[arxiv:1308.4896 [ hep - th ] ] .
d. tong , `` the holographic dual of @xmath169 , '' http://dx.doi.org/10.1007/jhep04(2014)193[_jhep_ * 04 * ( 2014 ) 193 ] , http://arxiv.org/abs/1402.5135[arxiv:1402.5135 [ hep - th ] ] .
f. benini , r. eager , k. hori , and y. tachikawa , `` elliptic genera of two - dimensional n=2 gauge theories with rank - one gauge groups , '' http://dx.doi.org/10.1007/s11005-013-0673-y[_lett .
* 104 * ( 2014 ) 465493 ] , http://arxiv.org/abs/1305.0533[arxiv:1305.0533 [ hep - th ] ] .
s. aspinwall , `` point - like instantons and the spin ( 32 ) / z(2 ) heterotic string , '' http://dx.doi.org/10.1016/s0550-3213(97)00232-0[_nucl .
* b496 * ( 1997 ) 149176 ] , http://arxiv.org/abs/hep-th/9612108[arxiv:hep-th/9612108 [ hep - th ] ] .
kim , s .- s .
kim , and k. lee , `` 5-dim superconformal index with enhanced en global symmetry , '' http://dx.doi.org/10.1007/jhep10(2012)142 [ _ jhep _ * 10 * ( 2012 ) 142 ] , http://arxiv.org/abs/1206.6781[arxiv:1206.6781 [ hep - th ] ] .
m. r. douglas , `` on d=5 super yang - mills theory and ( 2,0 ) theory , '' http://dx.doi.org/10.1007/jhep02(2011)011[_jhep_ * 02 * ( 2011 ) 011 ] , http://arxiv.org/abs/1012.2880[arxiv:1012.2880 [ hep - th ] ] .
n. lambert , c. papageorgakis , and m. schmidt - sommerfeld , `` m5-branes , d4-branes and quantum 5d super - yang - mills , '' http://dx.doi.org/10.1007/jhep01(2011)083[_jhep_ * 01 * ( 2011 ) 083 ] , http://arxiv.org/abs/1012.2882[arxiv:1012.2882 [ hep - th ] ] .
h. hayashi , s .- s .
kim , k. lee , m. taki , and f. yagi , `` a new 5d description of 6d d - type minimal conformal matter , '' http://dx.doi.org/10.1007/jhep08(2015)097[_jhep_ * 08 * ( 2015 ) 097 ] , http://arxiv.org/abs/1505.04439[arxiv:1505.04439 [ hep - th ] ] .
k. yonekura , `` instanton operators and symmetry enhancement in 5d supersymmetric quiver gauge theories , '' http://dx.doi.org/10.1007/jhep07(2015)167[_jhep_ * 07 * ( 2015 ) 167 ] , http://arxiv.org/abs/1505.04743[arxiv:1505.04743 [ hep - th ] ] . | we explore 6d @xmath0 superconformal field theories with @xmath1 and @xmath2 gauge symmetries which cascade after higgsing to the e - string theory on a single m5 near an @xmath3 wall .
specifically , we study the 2d @xmath4 gauge theories which describe self - dual strings of these 6d theories .
the self - dual strings can be also viewed as instanton string solitons of 6d yang - mills theories .
we find the 2d anomaly - free gauge theories for self - dual strings , amending the naive adhm gauge theories which are anomalous , and calculate their elliptic genera .
while these 2d theories respect the flavor symmetry of each 6d scft only partially , their elliptic genera manifest the symmetry fully as these functions as bps index are invariant in strongly coupled ir limit .
our consistent 2d @xmath5 gauge theories also provide new insights on the non - linear sigma models for the instanton strings , providing novel uv completions of the small instanton singularities .
finally , we construct new 2d quiver gauge theories for the self - dual strings in 6d e - string theory for multiple m5-branes probing the @xmath3 wall , and find their fully refined elliptic genera . addtoresetequationsection + + * higgsing towards e - strings * joonho kim@xmath6 , seok kim@xmath7 , kimyeong lee@xmath6 _ @xmath6school of physics , korea institute for advanced study , seoul 130 - 722 , korea .
_ + _ @xmath7department of physics and astronomy & center for theoretical physics , + seoul national university , seoul 151 - 747 , korea . _
+ e - mails : [email protected] , [email protected] , [email protected] |
the radiative decays of @xmath6 meson into a single photon and a pair of neutral pseudoscalar mesons are valuable sources of information in hadron physics
. these days can provide insight into the structure and the properties of low - mass scalar resonances .
in particular , the radiative decay @xmath0 where the scalar resonances @xmath1 and @xmath2 are known to provide the dominant contribution to the amplitude allows for a direct measurement of the couplings of the @xmath7 system to @xmath8 and @xmath9 scalar resonances and thus it can yield information on the properties of these scalar resonances .
moreover , the study of the radiative decay process @xmath10 is important because it provides a background to the reaction @xmath11 .
this latter process was proposed as a way to study cp violation and measuring the ratio @xmath12 @xcite . since this involves seeking for very small effects , if the branching ratio of the decay @xmath0 is of the order of @xmath13 , or more precisely if @xmath14 , then this background decay @xmath0 will limit the scope of cp violation measurements in @xmath11 decay at da@xmath15ne .
therefore , the study of the reaction @xmath10 and the calculation of the branching ratio @xmath16 is crutial for the measurements of cp violation and small cp violating parameters in @xmath11 decay .
the decay @xmath0 was first considered by achasov et al .
@xcite using the one - loop mechanism where the decay proceeds through the chain of reactions as @xmath17 .
they noted the negative interference between the contributions of @xmath18 and @xmath19 resonances , and they obtained the value @xmath20 for the branching ratio of the @xmath0 decay for some set of @xmath18 and @xmath19 masses and the values of the coupling constants @xmath21 and @xmath22 .
the radiative @xmath6-meson decays , among other radiative decay processes of the type @xmath23 , where v and p belong to the lowest multiplets of vector ( v ) and pseudoscalar ( p ) mesons , were studied by bramon et al .
@xcite in the framework of vector meson dominance ( vmd ) mechanism using standard lagrangians obeying su(3)-symmetry . in this framework
, the radiative @xmath0 reaction proceeds through the decay chains @xmath24 where the intermediate vector mesons are @xmath25 and @xmath26 .
they obtained the branching ratio for the decay as @xmath27 .
bramon et al .
@xcite later considered the radiative vector meson decays in chiral perturbation theory ( chpt ) using chiral effective lagrangians enlarged to include on - shell vector mesons , and they calculated the branching ratios for various decays of type @xmath28 at the one - loop level , including both @xmath29 and @xmath7 intermediate loops . in this approach the decay @xmath0 proceeds through charged - kaon loops , and they obtained the contribution of charged - kaon loops to the branching ratio as @xmath30 .
they concluded that due to smallness of the vmd contribution , this decay is dominated by charged - kaon loops .
an additional contribution to the decay of the type @xmath28 is provided by the amplitude involving scalar mesons as an intermediate state .
these radiative decays were studied by marko et al .
@xcite in the framework of unitarized chiral perturbation theory .
they used the techniques of chiral unitary theory developed earlier to include the final state interactions of two pseudoscalars by summing the pseudoscalar loops through the bethe - salpeter equation .
a review is given by oller et al .
@xcite . in this approach
the scalar resonances are generated dynamically by utilizing the one - loop pseudoscalar amplitudes .
the reaction @xmath0 was studied by oller @xcite within this framework .
the amplitude of this radiative decay has contributions coming from the scalar mesons @xmath31 and @xmath32 , and oller obtained for the branching ratio the value @xmath33 .
the radiative vector meson decays @xmath28 were also studied using linear sigma model ( l@xmath34 m ) , which is a @xmath35 chiral model that incorporates the pseudoscalar and scalar meson nonets @xcite .
the decay @xmath0 was studied within the framework of l@xmath34 m by escribano @xcite who obtained the result @xmath36 . in this work , we attempt to calculate the branching ratio @xmath16 using a phenomenological approach employed earlier in the studies of radiative @xmath37 @xcite and @xmath38 @xcite decays . in our calculation
we use the kaon - loop model @xcite where the initial vector meson @xmath6 decays into a pair of charged kaons , which after the emission of a photon couple to the neutral kaon pair @xmath39 through the scalar resonances @xmath19 and @xmath18 .
moreover , we consider the vertices involving the scalar mesons as point like , therefore the effects of the structure are reflected in the coupling constants .
the mechanism of the radiative decay process @xmath10 in the kaon - loop model is provided by the reactions @xmath40 where the last reaction proceeds by a two - step mechanism with the charged - kaon loop coupling to the final @xmath41 state through the scalar resonance @xmath31 or @xmath32 . in fig .
1 we show the processes contributing to the @xmath0 amplitude diagramatically where the diagram in fig .
1(c ) results from the minimal coupling for gauge invariance .
we do not make any assumption about the structure of the scalar mesons @xmath42 or @xmath19 .
we note that the @xmath6 meson and the scalar mesons @xmath18 and @xmath19 both couple strongly to the @xmath43 system , we therefore describe the @xmath44 and @xmath45 vertices phenomenologically by effective lagrangians .
the @xmath44 vertex in the diagrams shown in fig . 1 is described by the phenomenological lagrangian @xmath46 we utilize the experimental value for the branching ratio @xmath47 @xcite and determine the coupling constant @xmath48 as @xmath49 .
the @xmath50 vertex , where s denotes the scalar meson @xmath18 or @xmath19 , is described by the phenomenological lagrangian @xmath51 which is usually considered to define the coupling constant @xmath52 @xcite .
similarly , we describe the @xmath53 vertex by the effective lagrangian @xmath54 furthermore , isotopic spin invariance implies that the coupling constants @xmath55 and @xmath56 are related by the equations @xmath57 and @xmath58 @xcite . in our approach
we assume isotopic spin invariance and use the coupling constants @xmath55 and @xmath56 satisfying this requirement .
these coupling constants have been determined by theoretical calculations and from experimental analysis . in our phenomenological calculation
we use the values of these coupling constants which are determined from the experimental studies of the radiative decay processes @xmath59 and @xmath60 .
we then obtain the amplitude for the radiative decay reaction @xmath0 following from the diagrams shown in fig . 1 as @xmath61 i(a , b ) \nonumber \\ & & ~~\times { \cal m}(k^+k^-\rightarrow k^0\overline{k}^0)\end{aligned}\ ] ] where @xmath62 and @xmath63 are the polarizations and four - momenta of the @xmath6 meson and the photon , respectively . the invariant function i(a , b ) has been calculated in different contexts @xcite , and it is given by @xmath64 + \frac{a}{(a - b)^{2}}\left [ g\left(\frac{1}{b}\right)-g\left(\frac{1}{a}\right)\right ] \label{6}\end{aligned}\ ] ] @xmath65^{2}~ , & ~~x>\frac{1}{4 } \\ \frac{1}{4}\left [ \ln ( \frac{\eta_{+}}{\eta_{-}})-i\pi\right ] ^{2}~ , & ~~x<\frac{1}{4 } \end{array } \right .
\nonumber \\ & & \nonumber \\ & & g(x)=\left \ { \begin{array}{rr } ( 4x-1)^{\frac{1}{2 } } \arcsin(\frac{1}{2\sqrt{x}})~ , & ~~ x>\frac{1}{4 } \\
\frac{1}{2}(1 - 4x)^{\frac{1}{2}}\left [ \ln ( \frac{\eta_{+}}{\eta_{-}})-i\pi \right ]
~ , & ~~ x<\frac{1}{4 } \end{array } \right .
\nonumber \\ & & \nonumber \\ & & \eta_{\pm}=\frac{1}{2x}\left [ 1\pm(1 - 4x)^{\frac{1}{2}}\right ] ~,\end{aligned}\ ] ] where @xmath66 and @xmath67 with @xmath68 being the invariant mass of the final @xmath41 system given by @xmath69 .
the amplitude @xmath70 contains the scalar @xmath18 and @xmath19 resonances and in the approach we adopted it is given by @xmath71 since the scalar resonances @xmath18 and @xmath19 are unstable and they have a finite lifetime we use breit - wigner propagators with an energy dependent width for these resonances .
we therefore in the scalar meson propagator make the replacement @xmath72 where @xmath73 and @xmath74\left[1-\frac{(m_{\pi^0}-m_\eta)^2}{q^2}\right ] } } { \sqrt{\left[1-\frac{(m_{\pi^0}+m_\eta)^2}{m_{a_0}^2}\right]\left[1-\frac{(m_{\pi^0}-m_\eta)^2}{m_{a_0}^2}\right]}}~ \theta(\sqrt{q^2}-(m_{\pi^0}+m_\eta))~~,\end{aligned}\ ] ] and we use the experimental values for the widths @xmath75 and @xmath76 @xcite in the above expressions . then the differential decay probability for the radiative decay @xmath0 for an unpolarized @xmath6 meson at rest
is given as @xmath77 where e@xmath78 and e@xmath79 are the photon and @xmath80 meson energies respectively .
we perform an average over the spin states of @xmath6 meson and a sum over the polarization states of the photon .
the decay width @xmath0 is then obtained by integration @xmath81 where the minimum photon energy is e@xmath82 and the maximum photon energy is given as @xmath83 .
the maximum and minimum values for the energy e@xmath79 of @xmath80 meson are given by @xmath84
the branching ratios @xmath85 and @xmath86 are listed as @xmath87 and @xmath88 in the particle physics data tables @xcite .
achasov @xcite showed in detail that the existing data on radiative @xmath6 meson decays support the charged - kaon loop mechanism for @xmath85 and @xmath89 radiative decay reactions . from the above branching ratios within the framework of charged - kaon loop mechanism for these decays we determine the coupling constants @xmath90 as @xmath91 gev and @xmath92 gev . using these values for the coupling constants and the values of the relevant masses taken from the particle physics data tables @xcite ,
if we include the contribution of @xmath18 resonance only in the decay mechanism of the radiative decay @xmath0 , we obtain the result for the branching ratio as @xmath93 . on the other hand ,
if the contribution of @xmath19 resonance is considered only the branching ratio is @xmath94 .
since both @xmath18 and @xmath19 resonances make a contribution to the decay @xmath0 , when considering their contribution to the decay rate we have to note that the amplitudes involving @xmath18 and @xmath19 resonances interfere destructively due to isotopic spin invariance as reflected in the relations between the coupling constants @xmath57 and @xmath58 .
thus , if we consider that the interference between the contributions of @xmath18 and @xmath19 resonances is destructive , we obtain for the branching ratio the value @xmath95 , which is somewhat higher than the previous estimations .
however , it is small enough , we therefore conclude that this reaction will not provide a significant background to the measurements of @xmath96 decay for testing cp violation . in fig .
2 , we plot the distribution @xmath97 for the radiative decay @xmath0 in the phenomenological approach that we adopted , where we also indicate the contributions coming from @xmath18 resonance , @xmath19 resonances , and the contribution resulting from the destructive interference of these mechanisms for the coupling constants used above .
furthermore , in fig .
3 , we show the photon spectrum @xmath98 for the process @xmath0 for the same amplitudes as in fig .
2 . we note , however , that our result for the branching ratio @xmath16 depends sensitively on the values of the coupling constants @xmath99 and the masses @xmath100 of the scalar resonances @xmath18 and @xmath19 .
we , therefore , repeat our calculation using the values of the coupling constants @xmath52 and the masses @xmath100 determined by different experimental groups by performing fits to their data of the radiative @xmath37 and @xmath60 decays .
snd collaboration @xcite obtained the values @xmath101 mev , @xmath102 , @xmath103 mev , @xmath104 from their fit to snd data considering @xmath18 and @xmath34 mixing .
if we use these values , we obtain the results @xmath105 and @xmath106 and @xmath107 .
they also performed a fit to their data without @xmath18 and @xmath34 mixing in which case they obtained @xmath108 mev , @xmath109 , @xmath110 mev and @xmath111 . using these values in our calculation
give the branching ratios @xmath112 , @xmath113 and @xmath114 .
achasov et al .
@xcite also calculated the branching ratio for the decay @xmath0 using the above sets of parameters describing the snd data @xcite , and for the first set they obtained @xmath115 while for the second set their result was @xmath116 .
kloe collaboration at the da@xmath15ne collider also studied the decays @xmath37 and @xmath60 @xcite .
they performed two different fits in order to measure the parameters of the scalar states . in the first fit
, they included the contribution of a possible broad scalar @xmath34 state as well as the intermediate @xmath31 state interfering destructively in the analysis of the data of the @xmath37 reaction . in the second fit only the contribution of the intermediate @xmath31 state was considered . in the first case they obtained @xmath117 mev , @xmath118 and in the second case @xmath119 mev ,
they also reported the values @xmath121 mev , @xmath122 .
if we use the values of the coupling constants @xmath99 and the masses @xmath100 that the kloe collaboration obtained in the first fit , we obtain the results @xmath123 and @xmath124 and @xmath125 . on the other hand , using the values they obtained in their second fit without considering the contribution of intermediate @xmath34 state in our calculation results in the branching ratios @xmath126 and @xmath127 and @xmath128 . however , it was pointed out by achasov et al .
@xcite that the kloe data also allow @xmath129 @xcite and @xmath130 @xcite .
if we employ these values of the coupling constants with the values of the masses @xmath131 , @xmath132 and in these analyses @xcite in our calculation we obtained the results @xmath133 , @xmath134 , and @xmath135 . we can
, therefore , conclude that within the formalism we consider in this work the branching ratio @xmath136 of the radiative decay @xmath10 is small enough so that this decay will not cause any serious background problem for the studies of cp violation in the @xmath11 decay .
other effects such as structure @xcite and the finite widths of scalars @xcite on the radiative @xmath6 decays have also been investigated .
oller @xcite showed that the inclusion of these contribution does not change the conclusions of the charged - kaon loop model .
99 i. dunietz , j. hauser , and j. l. rosner , phys .
d * 35 * , 2166 ( 1987 ) .
n. n. achasov and v. n. ivanchenko , nucl.phys .
b * 315 * , 465 ( 1989 ) .
a. bramon , a. grau , and g. pancheri , phys .
b * 283 * , 416 ( 1992 ) .
a. bramon , a. grau , and g. pancheri , phys .
b * 289 * , 97 ( 1992 ) .
e. marco , s. hirenzaki , e. oset , and h. toki , phys .
b * 470 * , 20 ( 1999 ) .
j. a. oller , e. oset , and a. ramos , prog .
. phys . * 45 * , 157 ( 2000 ) .
j. a. oller , phys .
b * 426 * , 7 ( 1998 ) .
j. schechter and y. ueda , phys .
d * 3 * , 2874 ( 1971 ) .
r. escribano , hep - ph/0606314 .
a. gokalp and o. yilmaz , phys .
d * 64 * , 053017 ( 2001 ) .
a. gokalp , a. kucukarslan , s. solmaz , and o. yilmaz , j. phys .
g * 28 * , 2783 ( 2002 ) ; err .
g * 28 * , 3021 ( 2002 ) .
particle data group , s. eidelman et al .
b * 592 * , 1 ( 2004 ) .
f. e. close , n. isgur , and s. kumana , nucl .
b * 389 * , 513 ( 1993 ) .
j. lucio and j. pestiau , phys .
d * 42 * , 3253 ( 1990 ) ; d * 43 * , 2447(e ) ( 1991 ) .
n. n. achasov , nucl .
a * 728 * , 425 ( 2003 ) .
m. n. achasov et al . , phys .
b * 479 * , 53 ( 2000 ) .
m. n. achasov et al . , phys .
b * 485 * , 349 ( 2000 ) .
n. n. achasov and v. v. gubin , phys .
d * 64 * , 094016 ( 2001 ) .
a. aloisio , et al .
, phys . lett .
b * 536 * , 209 ( 2002 ) .
a. aloisio , et al .
b * 537 * , 21 ( 2002 ) .
n. n. achasov and a. v. kiselev , phys .
d * 68 * , 014006 ( 2003 ) .
n. n. achasov and a. v. kiselev , phys .
d * 73 * , 054029 ( 2006 ) .
s. kalashnikova , e. a. kudryavtsev , a. v. nefediev , c. hanhart , and j. haidenbauer , eur .
j. a * 24 * , 437 ( 2005 ) .
j. a. oller , nucl .
a * 714 * , 161 ( 2003 ) . | we study the radiative @xmath0 decay within a phenomenological framework by considering the contributions of the @xmath1 and @xmath2 scalar resonances .
we calculate the branching ratio @xmath3 by employing the coupling constants @xmath4 and @xmath5 as determined by different experimental groups .
= 10000 + |
fundamental aspects of phase transitions in complex systems can be revealed by the analysis of its microcanonical thermostatistics @xcite , which is characterised by the well known entropy @xmath3 , where @xmath4 denotes the density of states of a system with energy @xmath5 . in particular , the analysis of inflection points of the microcanonical inverse temperature @xmath6 plays an important role in the identification of stable , unstable and metastable regions in the phase diagram @xcite , providing alternative insights to the usual canonical analysis .
also , free - energy profiles can be obtained from the caloric curves @xmath7 vs @xmath5 , from where one can easily evaluate the values of barrier heights and latent heats . in this way , the microcanonical thermostatistics analysis has been incorporated in many studies in the literature , _
e.g. _ refs .
@xcite to name a few .
nevertheless , any analysis must rely on data obtained from efficient exploration of the configurational space .
it is well known that numerical simulations performed with monte carlo ( mc ) and molecular dynamics ( md ) methods pose limitations to the achievement of reliable data sampling @xcite .
such limitations are related to the critical slowing down @xcite , which is observed in studies of continuous phase transitions , and to the entrapment in local minima , in the case of systems with rugged energy landscapes . in both cases ,
the configurational space is poorly explored in a reasonable computational simulation time , which may produce biased physical averages . to overcome the trapping problem ,
it has been suggested that configurations must be sampled using algorithms based on generalised ensembles , where the updates are performed with non - boltzmann statistical weights @xmath8 .
for instance , the multicanonical algorithm ( muca ) @xcite , the extended gaussian ensemble ( ege ) @xcite , tsallis statistical weight @xcite , and replica exchange method ( rem ) @xcite either use a series of boltzmann weights or any convenient generalised sampling weight @xcite .
muca simulations sample configurations with a weight @xmath9 in such a way that the energy distribution is uniform , @xmath10 .
thus , a precise determination of @xmath9 is equivalent to obtain an estimate for the density of states @xmath4 , _
i.e. _ @xmath11 . the weights @xmath12 $ ] follows from the parameterization of the entropy @xmath13 , where @xmath14 and @xmath15 are the so - called multicanonical parameters .
the iterative procedure to obtain the muca parameters is described in detail in references @xcite , and can be read as , @xmath16e_{m } , \label{eq : an } \\\ ] ] where @xmath17 stands for the @xmath17th multicanonical simulation .
the recursion steps require the discretisation of the energy for continuous energy models .
therefore , it is convenient to define @xmath18 , where @xmath19 is the binsize , @xmath20 is an integer , and @xmath21 is a constant that defines a reference energy .
all the energies @xmath5 in the interval @xmath22 contribute to the histogram @xmath23 .
methods to improve sampling based on simulations at different temperatures have been proposed to either be conducted in parallel ( rem ) or as a random walk between different temperatures . in rem ,
@xmath24 non - interacting replicas of the system are simultaneously simulated by the usual mc or md algorithms , and from time to time , pairs of replicas at neighboring temperatures are exchanged with a transition probability . from the data produced by simulations performed at a single temperature @xmath25 or at a set of temperatures @xmath26 , with @xmath27 ,
it is necessary to employ a reweighing scheme to evaluate physical averages at a given temperature @xmath28 .
reweighting techniques @xcite use data from either a single histogram or multiple histograms obtained from mc or md simulations .
recently , a simple method called statistical weighted histogram analysis method ( st - wham ) @xcite has been proposed as an iteration - free procedure to obtain an estimate for the microcanonical inverse temperature . in this method
the usual wham equations @xcite are converted into a weighted average of the individual densities of states obtained from simulations carried out with different sampling weights @xmath8 . from energy histograms produced by multiple simulations , st - wham yields a statistical temperature @xmath29 , which is an estimate of the inverse microcanonical temperature @xmath30 .
interestingly , there is a numerical procedure based on the multicanonical recursion relations ( [ eq : an ] ) and ( [ eq : bn ] ) , which is called st - wham - muca @xcite , that can be used to replace the direct integration in order to evaluate the entropy @xmath1 .
although both st - wham and st - wham - muca have the advantage of _ a posteriori _ discretisation of energies , their naive implementations may lead to biased evaluations of physical quantities for continuous energy models just like all the aforementioned algorithms . as described in section
ii , the estimates @xmath31 for inverse microcanonical temperature @xmath0 depends on the derivatives of the energy histograms @xmath2 ( see eq .
( [ eq : st - wham2 ] ) ) . here
, we analyse how the estimates @xmath31 are energy binning dependent and , in section iii , we present two alternative approaches that avoids the need for energy binning to evaluate the microcanonical caloric curve for continuous energy models : ( i ) a proposal by berg and harris @xcite , which involves an empirical cumulative distribution ( cdf ) and uses discrete fourier series ; and ( ii ) a bayesian approach @xcite to model this cdf and assumes that the thermodynamic phase transitions are well described by the coexistence of two phases . a comparative analysis between these approaches
is made in order to characterise @xmath0 for two systems that undergo first - order - like phase transitions : the folding transition of a coarse - grained protein model for ubiquitin and the aggregation transition of two heteropolymers that follows a fibonacci sequence .
these examples allow us to describe the statistical and systematic errors involved in the numerical calculations of @xmath2 and @xmath32 , which are presented in section iv .
conclusions on this comparative analysis are presented in section v.
the st - wham @xcite yields a direct estimate of the inverse microcanonical temperature @xmath33 by considering the statistical inverse temperature @xmath34 where @xmath35 @xmath36 , and @xmath37 .
it is preferable to write eq .
( [ eq : st - wham1 ] ) as @xmath38 note that @xmath39 for simulations with the canonical weight . with the set of estimates @xmath40 , muca recurrence relations ( [ eq :
an ] ) and ( [ eq : bn ] ) can be applied to obtain estimates @xmath41 for the microcanonical entropy @xmath42 , @xmath43 this st - wham - muca algorithm is quite simple if one has @xmath40 .
we can numerically evaluate the derivatives in eq .
( [ eq : st - wham2 ] ) in a naive way , where the derivatives @xmath44 at energies @xmath45 follow from a linear regression around this point .
for instance , we use a linear regression with @xmath46 points ; selecting @xmath47 points means that the derivative at @xmath45 is calculated with the values of @xmath48 , where @xmath49 .
we chose a value for @xmath47 according to the energy binsize @xmath19 . consequently , we calculate the derivatives in the energy range @xmath50 . in this method , it is more convenient to directly calculate the derivative of @xmath51 than the derivative of @xmath52 .
we calculate the linear regression with a subroutine easily adapted from the linear fit subroutine in @xcite .
another approach can be devised by considering an algorithm based on the cumulative distribution function ( cdf ) @xcite .
the advantage of such approach is that it avoids histogramming when describing probability densities @xmath53 , dismissing the need for any _ ad hoc _ energy discretisation .
the method defines an estimator @xmath54 for the cdf @xmath55 , where the function @xmath54 is an empirical cumulative distribution function ( ecdf ) for the probability density @xmath53 .
the algorithm sorts the energy time series of length ndat in an ascending order @xmath56 , so any outliers can be eliminated by constructing a restricted ecdf @xmath57 in the range between two meaningful points @xmath58 and @xmath59 ( in general one takes @xmath60 and @xmath61 ) .
the basic idea is to propose an approximating function @xmath62 to describe @xmath57 , from where the difference function is defined , @xmath63 this function can be expanded in fourier series , @xmath64 which gives the fourier coefficients @xcite , @xmath65 here we use the same criteria of @xcite where the maximum number mmax of coefficients @xmath66 is obtained by imposing the two - sided kolmogorov test @xcite .
notably , eq . ( [ eq : di ] ) provides all the information that one needs to obtain a parameter - free energy probability density @xmath53 .
the reason is because eq .
( [ eq : re - def ] ) yields a smooth estimate of the cdf given mmax coefficients in the fourier expansion , @xmath67 even if one assumes a linear _ ansatz _ for @xmath62 ( see @xcite ) . in turn
, it becomes easy to obtain estimates for the probability density @xmath68 by differentiation , which consists in a smooth estimation of @xmath2 .
now , let us go back to the numerical differentiation in eq .
( [ eq : st - wham2 ] ) . to obtain parameter - free derivatives
@xmath69 , we just need to differentiate @xmath57 again .
this corresponds to the following numerical calculation , @xmath70 unlike the linear regression method , where we calculate the derivative of @xmath71 , here is more convenient to compute the derivative of @xmath2 directly .
we have included in appendix the function derpd(x ) , which can be used to estimate this derivative from the fourier coefficients @xmath66 .
one can easily incorporate this function in the program cdf_pd subroutine provided by berg and harris @xcite .
next we present our approach to implement a bayesian analysis . to this end , input data ( denoted by @xmath72 ) is the empirical cumulative distribution and not the histograms that are dependent on the energy binsize @xmath19 .
thus , the aim is to describe the empirical cumulative distribution by a given model ( called model m ) and using the bayesian analysis @xcite , extract values for its parameters ( denoted by @xmath73 ) from the _ posterior _ probability density function ( pdf ) @xmath74 . here
, @xmath75 denotes possible parameters related to the experimental conditions , which is the temperature in our numerical experiment , and @xmath76 refers to a set of observed experimental points @xmath77 . from the _
posterior _ pdf , we extract the desired probability distributions for each parameter , @xmath78 which correspond to the marginalized distributions .
these marginalized distributions are depicted with histograms in fig .
[ fig:6 ] for the ubiquitin . for each protein ,
our _ experimental _ data @xmath72 is produced by dividing the observed energy range into 35 points @xmath77 and _ experimental _ errors are extracted from the jackknife procedure .
we anticipate here that these _ experimental _ data are displayed in figures [ fig:4 ] and [ fig:10 ] for the folding - unfolding transition of ubiquitin and for the aggregation of two fibonacci sequences , respectively . within a bayesian approach ,
the set of observations @xmath79 is just a set among possible experimental outcomes @xmath72 at the points @xmath77 .
thus , given a fixed model m and a set of experimental values @xmath80 , parameters estimation within bayesian approach is obtained as follows @xcite , @xmath81 where @xmath82 is the _ prior _ distribution for a fixed model m. since our aim is to characterise @xmath32 in a first - order phase transition region , we assume that the energy distribution @xmath53 ( _ i.e. _ a continuous estimate for @xmath2 in eq .
( [ eq : st - wham2 ] ) ) is well described by a double - peak function .
this function characterises the coexistent phases at temperatures close enough to the transition temperature @xmath83 , whose shape is a consequence of the free - energy barrier between the ordered and disordered phases .
therefore , the expected energy distribution @xmath53 can be written as a normalized sum of two gaussian distributions with peaks centered at energies @xmath84 and @xmath85 , corresponding to the disordered and ordered phases , respectively @xcite .
if we assume the model m as the cumulative distribution function of @xmath53 , and recalling that for a gaussian distribution the cdf is the error function , we arrive to the following modelling , @xmath86 .
\label{eq : fe}\ ] ] this is a 5-parameter model , which we call model i in our analysis , with @xmath87 . now , to obtain the _ posterior _ probability of the parameter set @xmath73 , given data set @xmath76 , from the _ likelihood _
@xmath88 times the initial _ prior _ probabilities @xmath89 for our model i , we use a markov chain monte carlo . to this end
, we consider the metropolis algorithm with the proposal function @xmath90 ^ 2/2 \sigma_i^2 } \
, , \ ] ] where @xmath91 are the errors in our computational experiment .
the values @xmath92 are calculated with the accepted values for the parameters @xmath73 of model i.
in the following , we present results where we evaluate the performance of the different approaches in obtaining continuous estimates for @xmath2 and , consequently , the estimate @xmath32 for microcanonical inverse temperature . the comparative analysis is made with data obtained from both mc and md simulations , where we employ two simplified off - lattice protein models , _
i.e. _ all atoms in the polypeptide chain are replaced by beads located at the @xmath93-atom position , and present a continuous energy spectrum . for ubiquitin obtained with linear regression ( @xmath94 ) and cdf method with data generated from independent md simulations at @xmath95 .
statistical errors for the cdf method were calculated with 80 jackknife bins.,scaledwidth=45.0% ] for ubiquitin .
error bars are calculated with 20 jackknife bins and are shown only for the cdf method .
inset : comparative behaviour of @xmath96 produced with the cdf method and bayesian approach assuming model i.,scaledwidth=95.0% ] for ubiquitin .
error bars are calculated with 20 jackknife bins and are shown only for the cdf method .
inset : comparative behaviour of @xmath96 produced with the cdf method and bayesian approach assuming model i.,scaledwidth=95.0% ] ubiquitin is a 76 residue small protein ( pdb code 1ubq ) for which evidences suggest a two - state folding mechanism .
it had received considerable attention for what concerns solvent response and temperature dependence in the secondary structure formation @xcite and stretching experiments @xcite . to study the performance of the numerical approaches in evaluating @xmath0 through the folding - unfolding transition , we perform simulations with a coarse - grained version for ubiquitin but with a rather detailed potential .
we use the structure - based model implemented in smog@ctbp web server @xcite to perform md simulations with gromacs package .
the analyses were made from @xmath97 measurements during 250 ns of simulation after a period of thermalization .
measurements were obtained from 7 independent simulations at temperatures @xmath98 with @xmath99 and @xmath100 .
first , we evaluate the performance of the approaches considering the data obtained from a single md trajectory produced at @xmath101 , which is close to the transition temperature @xmath102 evaluated with statistics obtained with 7 independent md simulations .
figure [ fig:1 ] displays the estimates for the caloric curve using the linear regression to evaluate derivatives of @xmath103 from data obtained at @xmath104 .
this figure shows how noisy the caloric curve can be as a function of the energy binsize @xmath105 and 1.3 , keeping the number of points @xmath46 fixed .
figure [ fig:2 ] compares what seems a good energy discretisation with the cdf method .
error bars based on 20 jackknife bins are presented only for the cdf method .
now , fig .
[ fig:3 ] compares both methods when we consider the statistics obtained with the entire set of temperatures @xmath106 , @xmath107 . to present error bars for the cdf method
, we have increased the number of jackknife bins to 80 because this larger statistics .
nice agreement between both methods are obtained if we choose @xmath19 conveniently .
figure [ fig:4 ] displays what we call input experimental points to perform the bayesian analysis .
this _ experimental _ data was obtained from md simulations at @xmath104 .
the dashed line corresponds to the cumulative distribution obtained with model i. the parameters of this model that fit the input data are displayed in table 1 .
we include values calculated from means of the marginalized distributions , and global modes that presented the smallest @xmath108/d.o.f of the model . to verify how the proposed model describes the energy distribution
when compared with the cdf method and naive histogramming , we show in fig .
[ fig:5 ] the results obtained from these methods for the statistics collected at @xmath104 .
we realize that model i does not recover properly the region between the peaks of the energy distributions @xmath2 obtained with previous numerical methods .
the inset in this figure compares the calculations of @xmath96 following from the cdf and bayesian approaches . in fig .
[ fig:6 ] we display the marginalized distributions of @xmath73 , including the correlation between the parameters @xmath109 and @xmath58 to illustrate their interdependence .
we have always used flat _
priors _ over appropriated ranges to obtain the parameter distributions . and
@xmath109 with data obtained for ubiquitin at @xmath104.,scaledwidth=48.0% ] figure [ fig:7 ] shows the caloric curve for the aggregation of two chains of monomers described by the ab model , a coarse - grained off - lattice protein model that replaces the all - atom potentials by simpler physical interactions .
this model reduces the protein to a chain of monomers of two types : the hydrophobic ( @xmath110 ) and the polar or hydrophilic ( @xmath111 ) types , located at the @xmath93 atoms .
the ab model has also been used in studies to explore aggregation phenomenon @xcite , where more than one chain is included in the system , and in studies of microcanonical thermostatistics of heteropolymers @xcite . here , we consider a simple system which consists of two heteropolymers defined by fibonacci sequences with @xmath112 monomers , _ i.e. _ abbabbababbab . the statistics for this system amounts to @xmath113 sweeps per replica produced with rem .
attempts to exchange the replicas occur every @xmath114 sweeps , using a scheme that alternates attempts between even replicas and their neighbors , and odd replicas and their neighbors .
although one has several manners to define the temperature set @xcite , here we determine it using an arithmetic progression @xmath115 .
we consider @xmath116 replicas choosing @xmath117 , and @xmath118 . for the aggregation of two fibonacci sequences with the linear regression ( @xmath119 ) and cdf method for data obtained at @xmath120 .
statistical errors for the cdf method were calculated with 20 jackknife bins.,scaledwidth=95.0% ] for the aggregation of two fibonacci sequences with the linear regression ( @xmath119 ) and cdf method for data obtained at @xmath120 .
statistical errors for the cdf method were calculated with 20 jackknife bins.,scaledwidth=95.0% ] for the aggregation of two fibonacci sequences with the linear regression ( @xmath119 ) and cdf method for energy time series obtained from rem at @xmath121.,scaledwidth=45.0% ] figure [ fig:7 ] depicts the caloric curves @xmath31 obtained with energy binsizes @xmath122 , and @xmath123 .
the value @xmath46 was used to calculate the derivatives of @xmath51 .
different values of @xmath19 allows one to verify the systematic errors associated with @xmath19 and the use of linear regression . for a fixed @xmath47 , small values of @xmath19 ( or @xmath124 )
tend to reproduce the statistical fluctuations of the data . on the other hand , a value as large as @xmath125 eliminates the physical information associated with the small s - loop at @xmath126 . for comparison ,
we include in fig .
[ fig:8 ] the @xmath31 estimates from the cdf method .
these comparisons are based on a single time series selected from data produced with rem at the inverse temperature @xmath127 .
it is important to reliably calculate the caloric curve when assessing the canonical critical temperature and the latent heat of the transition .
we can figure out how important the fluctuations observed in these curves @xmath31 are by calculating the statistical errors associated with the cdf method .
figure [ fig:8 ] displays the statistical error bars calculated with 20 jackknife bins for the dataset with @xmath113 measurements obtained with rem at @xmath120 . both methods present comparable statistical error bars .
this figure shows that the small s - loop at @xmath126 does not result from the statistical fluctuations in our dataset .
therefore , any smooth estimate of this curve at @xmath126 , obtained for example with @xmath128 ( @xmath129 ) , would hide physical information . for the aggregation of two fibonacci sequences .
error bars are calculated with 20 jackknife bins are shown for the cdf method .
inset : comparative behaviour of @xmath96 as predicted with the cdf method and bayesian analysis assuming model i.,scaledwidth=95.0% ] for the aggregation of two fibonacci sequences .
error bars are calculated with 20 jackknife bins are shown for the cdf method .
inset : comparative behaviour of @xmath96 as predicted with the cdf method and bayesian analysis assuming model i.,scaledwidth=95.0% ] in fig .
[ fig:9 ] we include data produced with @xmath130 replicas together .
comparison of the results presented in this figure with those obtained by multicanonical simulations in refs .
@xcite reveals that they are in excellent agreement : even the small s - loop near @xmath131 appears in our case .
computation of the inverse of aggregation temperature @xmath132 also resulted in a quantitative agreement .
larger error bars appear around @xmath133 , probably because the lack of temperatures in the set used to exchange the replicas .
the averaged estimates from both derivative calculation methods agreed for a particular choice of @xmath134 and @xmath46 points in the linear regression method .
now , fig .
[ fig:10 ] shows the _ experimental _ data we have used to perform the bayesian analysis .
again , we initially keep the analysis restricted to a single energy time series , in this case produced at @xmath120 .
we have included the bayesian estimates ( dashed line ) of the cumulative function assuming the model i. the parameters for this model are presented in table i. to verify how well this model reproduces @xmath53 , we include in fig .
[ fig:11 ] the energy distributions obtained with a naive histogramming procedure , cdf method and bayesian analysis . clearly , bayesian analysis does not reproduce the expected behaviour between both peaks .
this leads to a biased estimation of @xmath96 as can be seem in the inset of this figure .
moreover , the approach with model i does not reproduce the small s - loop observed at @xmath126 ( data not shown ) , in contrast to the estimates with a convenient choice of @xmath19 or the cdf method . as a consequence , we again do not follow further analysis with the bayesian approach including all statistics from the 7 replicas .
we presented two alternative approaches for the estimation of the energy probability distributions avoiding the need of energy binning in order to obtain the microcanonical thermostatistics analysis from st - wham .
numerical comparisons between the approaches were presented and we showed the statistical and systematic errors that can arise when one evaluates the microcanonical inverse temperature for two continuous energy models that exhibits first - order like phase transitions .
our results indicate that the cdf method yields reliable estimates for both @xmath2 and @xmath0 when compared with the linear regression method , and far more successfully than the bayesian approach with model i. unlike the linear regression method , the cdf method avoids the undesirable task of choosing the energy binsize @xmath19 for a careful evaluation of the microcanonical temperature , as highlighted in the analysis of the aggregation transition of the two fibonacci sequences . in this case
, the caloric curve presents a quite unusual behaviour with two s - loops , indicating two transitions .
in particular , we showed that the use of large values for @xmath19 in linear regression method would hide physical information about the small s - loop at @xmath126 . on the other hand
, the small s - loop could not be detected with the bayesian analysis ( data not shown ) as a consequence of the poor evaluation of @xmath2 and its derivative ( fig .
[ fig:11 ] ) .
the reason is because the model i consists of only 5 parameters .
this is a low number compared to the cdf method , which allows a variable number of parameters ( _ i.e. _ fourier coefficients ) defined depending on the demand of the ecdf .
for instance , the cdf method needs 13 fourier coefficients to obtain probability densities for the ubiquitin data and this number goes up to 74 for fibonacci sequences .
furthermore , model i was constructed on the hypothesis that the coexistence of two thermodynamic bulk phases can be well described by two gaussian distributions .
actually , this is an approximation because the energies in between the two peaks , describing mixed phase configurations , are not properly take into account in the two gaussian model .
we thank bernd berg for carefully reading our manuscript .
this work has been supported by the brazilian agencies fapesp , cnpq and rusp ( university of so paulo ) .
derivative of the probability density : .... function derpd(x ) include ' .. / .. /libs / fortran / implicit.sta ' include ' .. / ..
/libs / fortran / constants.par ' parameter(nmax=100 ) common /cdfprob/ xmin , xrange , dn(nmax),m !
m number of fourier coefficients derpd = zero do j=1,m aux = j*pi / xrange aux = aux*aux derpd = derpd - dn(j)*aux*sin(j*pi / xrange*(x - xmin ) ) enddo return end .... d.h.e .
gross , microcanonical thermodynamics : phase transitions in small systems , lecture notes in physics 66 ( world scientific , singapore , 2001 ) .
gross , and j.f .
kenney , j. chem .
( 2005 ) 224111 .
frigori , l.g .
rizzi , and n.a .
alves , j. phys .
( 2010 ) 012018 .
s. schnabel , d.t .
seaton , d.p .
landau , and m. bachmann , phys .
e 84 ( 2011 ) 011127 .
rocha , s. schnabel , d.p .
landau , and m. bachmann , phys .
e 90 ( 2014 ) 022601 .
frigori , l.g .
rizzi , and n.a .
alves , j. chem .
( 2013 ) 015102 . c. junghans , m. bachmann , and w. janke , phys .
97 ( 2006 ) 218103 . c. junghans , m. bachmann , h. arkin , and w. janke , j. chem phys .
128 ( 2008 ) 085103 .
m. mddel , w. janke , and m. bachmann , phys .
phys . 12 ( 2010 ) 11548 .
t. bereau , m. bachmann , and m. deserno , j. am .
( 2010 ) 13129 .
church , c.e .
ferry and a. e. van giessen , j. chem .
( 2012 ) 245102 .
cho , j. kim , j. lee , t. keyes , j.e .
straub , and k.s .
kim , phys .
( 2014 ) 157802 .
a. mitsutake , y. sugita , and y. okamoto , biopolymers 60 ( 2001 ) 96 .
landau , and k. binder , a guide to monte carlo simulations in statistics physics . cambridge university press , 2000 .
berg , 2000 .
monte carlo methods , fields institute communications vol .
26 , edited by n. madras ( american mathematical society , providence , ri , 2000 ) , p. 1 - 24 .
berg , comput .
commun . 153
( 2003 ) 397 .
johal , and a. planes , e. vives , phys .
e 68 ( 2003 ) 056113 .
t. neuhaus , and j.s .
hager , phys .
e 74 ( 2006 ) 036702 .
frigori , l.g .
rizzi , and n.a .
alves , eur .
j. b 75 ( 2010 ) 311 .
i. andricioaei , and j.e .
straub , phys . rev .
e 53 ( 1996 ) r3055 .
hansmann , and y. okamoto , phys .
e 56 ( 1997 ) 2228 .
k. hukushima , and k. nemoto , j. phys .
japan 65 ( 1996 ) 1604 .
j. kim , t. keyes , and j.e .
straub , j. chem .
( 2010 ) 224107 .
a.m. ferrenberg , and r.h .
swendsen , phys .
61 ( 1988 ) 2635 .
[ erratum : 63 ( 1989 ) 1658 ] a.m. ferrenberg , and r.h .
swendsen , phys .
63 ( 1989 ) 1195 .
alves , b.a .
berg , and s. sanielevici , nucl .
b 376 ( 1992 ) 218 .
j. kim , t. keyes , and j.e .
straub , j. chem .
( 2011 ) 061103 .
rizzi , and n.a .
alves , j. chem .
( 2011 ) 141101 .
berg , and r.c .
harris , comput .
phys . commun . 179
( 2008 ) 443 .
a. caldwell , d. kollr , and k. krninger , comput .
( 2009 ) 2197 .
_ , _ numerical recipes _ ( cambridge university press , london , 1986 ) .
j. lee , and j.m .
kosterlitz , phys .
b 43 ( 1991 ) 3265 .
j. jacob , b. krantz , r.s .
dothager , p. thiyagarajan , and t.r .
sosnick , j. mol .
( 2004 ) 369 .
chung , and a. tokmakoff , proteins 72 ( 2008 ) 474 .
dastidar , and c. mukhopadhyay , phys .
e 72 ( 2005 ) 051928 .
sulkowska , a. kloczkowski , t.z .
sen , m. cieplak , r.l .
jernigan , proteins 71 ( 2008 ) 45 .
a. das , c. mukhopadhyay , proteins 75 ( 2009 ) 1024 .
a. imparato , a. pelizzola , phys .
( 2008 ) 158104 .
noel , p.c .
whitford , k.y .
sanbonmatsu , and j.n .
onuchic , nucl .
acids res .
38 ( 2010 ) w657 .
fiore , j. chem .
( 2011 ) 114107 .
cclllllc + [ -0.35 cm ]
+ [ -0.35 cm ] examples & & @xmath84 & @xmath109 & @xmath85 & @xmath135 & @xmath58 & @xmath108/d.o.f .
+ + [ -0.15 cm ] ubiquitin & mean & 250.36(37 ) & 22.73(22 ) & 102.64(27 ) & 22.31(15 ) & 0.4474(31 ) & 0.131 + & global mode & 250.51 & 22.64 & 102.55 & 22.27 & 0.4474 & 0.130 + fibonacci & mean & 0.797(15 ) & 0.7786(79 ) & -8.128(37 ) & 1.968(25 ) & 0.4838(35 ) & 0.350 + sequences & global mode & 0.803 & 0.7754 & -8.147 & 1.958 & 0.4841 & 0.349 + +
[ -0.35 cm ] | microcanonical thermostatistics analysis has become an important tool to reveal essential aspects of phase transitions in complex systems . an efficient way to estimate the microcanonical inverse temperature @xmath0 and the microcanonical entropy @xmath1
is achieved with the statistical temperature weighted histogram analysis method ( st - wham ) .
the strength of this method lies on its flexibility , as it can be used to analyse data produced by algorithms with generalised sampling weights .
however , for any sampling weight , st - wham requires the calculation of derivatives of energy histograms @xmath2 , which leads to non - trivial and tedious binning tasks for models with continuous energy spectrum such as those for biomolecular and colloidal systems . here ,
we discuss two alternative methods that avoid the need for such energy binning to obtain continuous estimates for @xmath2 in order to evaluate @xmath0 by using st - wham : ( i ) a series expansion to estimate probability densities from the empirical cumulative distribution function ( cdf ) , and ( ii ) a bayesian approach to model this cdf . comparison with a simple linear regression method is also carried out .
the performance of these approaches is evaluated considering coarse - grained protein models for folding and peptide aggregation . |
let us prove ( [ fec ] ) .
its lhs implies on @xmath163 to find how the first term of the integrand of ( [ bfec ] ) behaves in the limit of r@xmath164 , we need to find @xmath165 from @xmath166 and from @xmath167 the difference between the derivatives of @xmath15 and of @xmath168 tends to zero in the limit of @xmath169 so , it is irrelevant if we use one or the other in the definition of k ; we use the simplest one , @xmath170 and @xmath171 to find : @xmath172 then , the first term in the integrand of ( [ bfec ] ) contains @xmath173.\eta.[r,\rho{\hbox{\large a}}+v(1+{\hbox{\large a}}.r)]{\big\lbrace}(\delta\tau)^{2}\eta_{\mu\nu}+\delta\tau v_{\mu}r_{\nu}-r_{\mu}r_{\nu}{\big\rbrace}}{\rho^{7}},\ ] ] and , using the notation of ( [ agbc ] ) , @xmath174\longrightarrow a_{2}=b_{2}=[\hbox{\large a},v]+{\cal o}(r)\longrightarrow a=4\ ] ] @xmath175 according to ( [ nr ] ) , this would produce a divergent result if @xmath176 but @xmath43 and its limit are traceless ( @xmath177 , and so , @xmath178 therefore , the indeterminacy @xmath50 remains and a new application of the lhospital rule is demanded .
then , from ( [ np ] ) , for @xmath179 , we have @xmath180 and then @xmath181 the terms containing @xmath182 will still give a null contribution ( @xmath43 is traceless ) . with @xmath183 @xmath184.\eta.[v,\hbox{\large a}].(9v\hbox{\large a}+6\hbox{\large a}v){\big\rbrace}\equiv0,\ ] ] and then , @xmath185 the last term in the integrand of ( [ bfec ] ) is related to ( [ p]-[p0 ] ) .
so , we can write @xmath186 which is null because @xmath187.\eta.[r , v]\rbrace.\eta.(r - v\rho)}{\rho^{6}}=0;\ ] ] and @xmath188 which is also null because @xmath189.\eta.[r,\rho\;\hbox{\large a}+v\hbox{\large a}.r].(r-\rho\;v)\rbrace}= % % following line can not be broken before 80 char % % following line can not be broken before 80 char \lim_{\rho\to0}\frac{1}{\rho^{5}}{r.\eta.\lbrace{[r , v].\eta.[r , v]\rbrace}.r\hbox{\large a}.r}=0,\ ] ] as they can be easily verified .
finally , @xmath190 as a consequence of ( [ c ] ) .
10 f. rohrlich;classical charged particles "
. reading , mass .
d. jackson classical electrodynamics",2nd ed . ,
14 , john wiley & sons , new york , ny(1975 ) .
e.g.p.rowe , phys .
d , 12 , 1576(1975 ) and 18 , 3639(1978 ) ; nuovo cim , b73 , 226(1983 ) .
a. lozada , j. math .
phys . , 30,1713(1989 ) .
j. c. eliezer , rev .
phys . , 19,148(1947 ) . c. itzkyson , j. zuber quantum field theory " , p.300 , 485 , mcgraw - hill , singapore ( 1983 ) .
m. m. de souza , the lorentz - dirac equation and the structure of spacetime " .
hep - th/9505169 . c. teitelboim
; d. villaroel ; ch.g.van weert , rev . del nuovo cim . ,
vol 3 , n.9,(1980 ) .
m. m. de souza , the lorentz - dirac equation and the physical meaning of the maxwell s fields " .
hep - th/9510160 .
m. m. de souza , on the origin of mass and the electroweak mass spectrum without higgs " in hadrons physics 94 , 265 - 268 , ed.v .
herscovitz et al .
world scientific ( 1994 ) . c. teitelboim phys .
d 1 , 1572(1970 ) .
synge : ann . mat .
pura appl . , 84 , 241 ( 1975 ) .
r. tabensky : phys .
d 13 , 267 ( 1976 ) . | the problems of classical electrodynamics with the electron equation of motion and with non - integrable singularity of its self - field stress tensor are well known .
they are consequences , we show , of neglecting terms that are null off the charge world line but that gives a non null contribution on its world line .
the self - field stress tensor of a point classical electron is integrable , there is no causality violation and no conflict with energy conservation in its equation of motion , and there is no need of any kind of renormalization nor of any change in the maxwell s theory for this .
+ ( this is part of the paper hep - th/9510160 , stripped , for simplicity , of its non - minkowskian geometrization of causality and of its discussion about the physical meaning of the maxwell - faraday concept of field ) . three unsolved problems @xcite make of classical electrodynamics of a point electron a non - consistent theory : 1 ) the field singularity in the lienard - wiechert solution ; 2 ) the non - integrable singularities of its stress tensor ; 3 ) the bizarre , causality violating behaviour of solutions of the lorentz - dirac equation @xcite .
+ the lorentz - dirac equation , @xmath0 is the greatest paradox of classical field theory as it can not simultaneously preserve both the causality and the energy conservation , although there is nothing in the premisses for its derivation that justify such violations .
the presence of the schott term , @xmath1 , is the cause of all pathological features of ( [ lordireq ] ) , like microscopic non - causality , runaway solutions , preacceleration , and other bizarre effects @xcite . on the other hand
the presence of this term is necessary for the maintenance of energy - momentum conservation ; without it it would be required a null radiance for an accelerated charge , as @xmath2 .
the argument , although correct , that such causality violations are not observable because they are outside the scope of classical physics @xcite and are blurred @xcite by quantum - mechanics effects is not enough compelling , because these same problems remain in a quantum formalism , just disguised in other apparently distinct problems @xcite .
+ the presence of non - integrable singularities in the electron self - field stress tensor is also a major problem .
previous attempts on taming these singularities have relied on modifications of the maxwell s theory with ad hoc addition of extra- terms to the field stress tensor on the electron world - line ( see for example the reviews @xcite ) ; it is particularly interesting that , as we will show here , instead of adding anything we should actually not drop out some null terms .
is their contribution ( not null , in an appropriate limit ) that avoid the infinities .
the same problem happens in the derivations of the electron equation of motion : they are done with incomplete field expressions that do not contain these terms that are null only off the particle world - line .
the schott term in the lorentz - dirac equation is a consequence of this ; it does not appear in the equation when the full field expression is correctly used .
+ the solution to the first problem , which is in essence the origin of the second and third problems , involves a discussion of the meaning of the maxwell - faraday concept of field and requires an entirely new formalism based on an anticipated recognition , yet in a classical context , of the discrete and localized character of the electromagnetic interaction , which is proper of a quantum formalism .
the solution to the second and to the third problems is easier to describe and just requires the knowledge of the model of spacetime geometry behind this new formalism .
we will present here a simplified minkowskian version of it .
the complete riemannian version and its relevance to quantum field theory and to the question of the origin and meaning of mass , is too long to be included in this letter .
it is being discussed elsewhere @xcite .
the immediate difference in the outcomes of these two versions appears in the equation of motion for a point electron , @xmath3 the last term on the lhs corresponds to the energy associated to the riemannian curvature of the electron manifold and it is not present in its minkowskian counterpart ( [ ffinally ] ) .
this is not the lorentz - dirac equation as it does not contain the troublemaker third - order term .
so , it does not violate causality , and it is , nonetheless , compatible with 4-momentum conservation , as we discuss below .
+ equation ( [ lordireq ] ) can be obtained from energy - momentum conservation in the lienard - wiechert solution @xcite , @xmath4 in the limit of @xmath5 .
v is the tangent to the particle world - line @xmath6 , parameterized by its proper time @xmath7 , ( @xmath8 , and @xmath9 ) .
it is expressed in terms of retarded coordinates @xcite , by which any spacetime point x is constrained with a particle world - line point @xmath10 by @xmath11 and @xmath12 where @xmath13 , where @xmath14 is the minkowski metric tensor , ( with signature + 2 ) .
@xmath15 is the spatial distance between the point x where the electromagnetic field is observed and the point @xmath10 , position of the charge , in the charge rest frame at its retarded time .
@xmath16 implies on @xmath17 .
k , defined by @xmath18 is a null 4-vector , @xmath19 , and represents a light - cone generator , or the electromagnetic wave - front 4-vector .
+ the retarded maxwell field , @xmath20}$ ] , is given by @xmath21=\frac{1}{\rho}[k_{\mu},\hbox{\large a}_{\nu}]+\frac{\hbox{\large a}_{k}}{\rho}[k_{\mu},v_{\nu}]+\frac{[k_{\mu},v_{\nu}]}{\rho^{2}},\ ] ] where , for notational simplicity , we are using @xmath22 @xmath23:=ab - ba\;,$ ] @xmath24 and we are making the electron charge and the speed of light equal to 1 : @xmath25 .
+ using ( [ fl ] ) in @xmath26 for finding the electron self - field stress tensor , @xmath27.\eta.[k,\rho\hbox{\large a}+v(1+\rho{\hbox{\large a}}_k)]-\frac{\eta}{4}[k,\rho\hbox{\large a}+v(1+\rho{\hbox{\large a}}_k)]^{2},\ ] ] or @xmath28 , with @xmath29.\eta.[k,\hbox{\large a}+v{\hbox{\large a}}_k]-\frac{\eta}{4}[k,\hbox{\large a}+v{\hbox{\large a}}_{k}]^{2},\ ] ] @xmath30.\eta.[k,\hbox{\large a}+v{\hbox{\large a}}_k]+[k,\hbox{\large a}+v{\hbox{\large a}}_k].\eta.[f , v]-\frac{\eta}{2}tr[k , v].\eta.[k,\hbox{\large a}],\ ] ] @xmath31.\eta.[k , v]-\frac{\eta}{2}[k , v]^{2}.\ ] ] it is worth to explicitly write ( [ t]-[t4 ] ) for @xmath32 ( then @xmath19 ) and make some comments .
@xmath33 @xmath34 @xmath35 @xmath36 , although singular at @xmath37 , is nonetheless integrable . by
that it is meant that @xmath38 exists @xcite , while @xmath39 and @xmath40 are not integrable ; they generate , respectively , the problematic schott term in the lde and a divergent expression , the electron bound 4-momentum @xcite .
the most update prescription @xcite is to redefine @xmath39 and @xmath40 at the electron world - line in order to make them integrable , but without changing them at @xmath32 , so to preserve the standard results of classical electrodynamics .
this is possible with the use of distribution theory , but it is always an introduction of something strange to the system and in an _ ad hoc _ way .
the most unsatisfactory aspect of this procedure is that it regularizes the above integral but leaves an unexplained and unphysical discontinuity in the flux of 4-momentum from the charge world - line : @xmath41 + we observe that @xmath42 which is important in the identification of @xmath36 with the radiated part of @xmath43 , and that @xmath44 the important difference among the sets of equations ( [ t2]-[t4 ] ) and ( [ t2e]-[t4e ] ) is that while the equations in the first one are complete , in the sense that they keep the terms proportional to @xmath45 , which are null ( as @xmath19 ) , in the last set of equations they have been dropped off . but
these @xmath45-terms , even with @xmath46 should not be dropped from the above equations , since they are necessary for producing the correct limits when @xmath47 or @xmath48 as @xmath49 , in this limit we have a @xmath50-type of indeterminacy , which can be raised with the lhospital rule and @xmath51 this results in @xmath52 and@xmath53 a feynmann diagram , see the figure 1 , helps in the understanding of these two results . in the limit of @xmath54 or at @xmath55
there are 3 distinct velocities : k , the photon 4-velocity , and @xmath56 and @xmath57 , the electron initial and final 4-velocities .
this is the reason for this indeterminacy at @xmath55 . at @xmath58
there is only @xmath57 , and only @xmath56 at @xmath59 or , back to the usual picture , @xmath60 , in general .
we must observe that the classical picture of a continuous interaction can not resolve this indeterminacy .
the lesson one should learn from this is that even in a classical context , is necessary to take into account the discrete and localized ( quantum ) character of the fundamental electromagnetic interaction in order to have a clear and consistent physical picture .
this is the viewpoint adopted in @xcite , where the concept of a classical photon is introduced .
it fulfils a requirement of an equal - foot treatment to the electron and to its self electromagnetic field , which is more in accordance with the experimental data that show both as equally fundamental physical objects of nature .
+ to find the limit of something when @xmath5 will be done so many times in this letter that it is better to do it in a more systematic way .
we want to find @xmath61 where @xmath62 is a homogeneous function of r , @xmath63 .
then , we have to apply the lhospital rule consecutively until the indeterminacy is resolved .
as @xmath64 , the denominator of ( [ lr ] ) at @xmath65 will be different of zero only after the @xmath66-application of the lhospital rule , and then , its value will be @xmath67 + if p is the smallest integer such that @xmath68 where @xmath69 , then @xmath70 * example 1 : @xmath71 * example 2 : @xmath72}{\rho}=\frac{[r_{\mu},\hbox{\large a}_{\nu}]}{\rho^{2}}\;\;$ ] @xmath73 @xmath74}{\rho}$ ] diverge * example 3 : @xmath75}=-\frac{\hbox{\large a}.r}{\rho^{3}}r^{[\mu}v^{\nu]}\longrightarrow p=4>n=3 $ ] @xmath76}=0 $ ] * example 4 @xmath77}{\rho^{2}}=\frac{[r_{\mu},v_{\nu}]}{\rho^{2}}\;\ ; \;\longrightarrow \;\;p=2<n=3\longrightarrow\lim_{r\to0}\frac{[k_{\mu},v_{\nu}]}{\rho^{2}}$ ] diverge the second term of the rhs of ( [ fl ] ) does not contribute to the electron self - field at @xmath78 but the first and the third terms diverge , as expected , although they produce integrable contributions to @xmath79 let us find the integral of the electron self - field stress tensor at the electron causality - line : @xmath80 or @xmath81 in terms of retarded coordinates @xcite @xmath82 where @xmath83 is the element of solid angle in the charge rest - frame .
+ but , we will first prove a useful result to be used with ( [ nr ] ) , when the numerator has the form @xmath84 where a and b represent two possibly distinct homogeneous functions of r , and the subindices indicate the order of @xmath85 .
then @xmath86 so , for using ( [ nr ] ) with ( [ np ] ) , we just have to find the @xmath7-derivatives of a and b that produce the first non- null term at the limit of @xmath87 .
+ applying ( [ nr ] ) and ( [ np ] ) for finding @xmath88 we just have to consider the first term of the rhs of ( [ t ] ) ; as the second one is the trace of the first , its behaviour under this limit can be inferred .
* example 5 @xmath89.\eta.[k,\rho\hbox{\large a}+v(1+\rho\;{\hbox{\large a}}_k)]}{\rho^{4}}=\ ] ] @xmath90.\eta.[r,\rho\hbox{\large a}+v(1+\hbox{\large a}.r)]}{\rho^{4}}\ ] ] + so , @xmath91\longrightarrow a_{2}=b_{2}=[\hbox{\large a},v]+{\cal o}(r).\;\;$ ] therefore , according to ( [ np ] ) , for producing a non null @xmath92 , a and p must be given by @xmath93.\eta.[\hbox{\large a},v]+{\cal o}(r).\ ] ] then , we conclude from ( [ nr ] ) , that + @xmath89.\eta.[k,\rho\hbox{\large a}+v(1+\rho{\hbox{\large a}}_k)]}{\rho^{4}}=\frac{1}{4}[\hbox{\large a},v].\eta.[\hbox{\large a},v]\ ] ] we have , therefore , from ( [ t ] ) , that @xmath94.\eta.[\hbox{\large a},v]-\frac{\eta}{4}[\hbox{\large a},v]^{2}\ ] ] the flux of 4-momentum irradiated from the electron , which is the meaning of ( [ te ] ) , is finite and depends only on its instantaneous velocity and acceleration .
it is interesting that ( [ te ] ) comes entirely from the velocity term , @xmath95}{\rho^{2}}$ ] , as we can see from the following example .
* example 6 @xmath96.\eta.[k , v]}{\rho^{4}}=\lim_{\rho\to0}\frac{[r , v].\eta.[r , v]}{\rho^{4}}=\frac{1}{4}[\hbox{\large a},v].\eta.[\hbox{\large a},v]\ ] ] as @xmath97+[r,\dot { \hbox{\large a}}]\longrightarrow p - a = a=2\longrightarrow n_{4}=6[\hbox{\large a},v]+{\cal o}(r)\;\;$ ] and @xmath98 so , @xmath99 the contribution from the other two terms just cancel to zero , @xmath100 as can be easily verified .
we must realize that the discontinuity , @xmath101 still remains .
it is a consequence of the problem 1 ; its solution requires an understanding of the physical meaning of the maxwell - faraday concept of field in this context .
it is being discussed elsewhere @xcite .
the electron equation of motion let us now discuss the third problem .
the electron equation of motion , can be obtained from @xmath102 where @xmath103 is the total electron energy - momentum tensor , which includes the contribution from the electron kinetic energy , from its interaction with external fields and from its self field .
let us move directly to the part that will produce novel results : @xmath104 where @xmath105 is the external forces acting on the electron , and the last term represents the impulse carried out by the emitted electromagnetic field in the bhabha tube surrounding the electron world - line , defined by the heaviside function , @xmath106 .
using the divergence theorem , we have that the last term of the rhs of ( [ eqm ] ) is transformed into @xmath107 this term represents the flux of 4-momentum through the cylindrical hypersurface @xmath108 let us denote it by @xmath109 : @xmath110 as @xmath111 and @xmath112 we can write @xmath113 with @xmath114 @xmath115 @xmath116 @xmath117 and @xmath118 are both null . in order to show this we need to apply ( [ nr ] ) for a n(r ) with a generic form @xmath119 , where a , b , and c are functions of r , such that @xmath120 .
then it is easy to show , from ( [ np ] ) , that @xmath121 * from ( [ t4 ] ) , the integrand of ( [ p2 ] ) , produces ( again , we do not need to consider the trace term ) @xmath96.\eta.[k , v].(k - v)}{\rho^{4}}=\lim_{\rho\to0}\frac{[r , v].\eta.[r , v].(r-\rho v)}{\rho^{5}},\ ] ] or , schematically @xmath122 with @xmath123,$ ] and @xmath124 then , @xmath125+{\cal o(r)},\;\;$ ] @xmath126 and we have , from ( [ agbc ] ) , the following restrictions on a and c for producing a @xmath127 + @xmath128 and @xmath129 or @xmath130 therefore , according to ( [ nr ] ) @xmath131 * from the second term of the integrand of ( [ p1 ] ) and from ( [ t3 ] ) we have @xmath96.\eta.[k,\hbox{\large a}+v{\hbox{\large a}}_k].(k - v)}{\rho^{3}}=\lim_{\rho\to0}\frac{[r , v].\eta.[r,\rho\;\hbox{\large a}+v\hbox{\large a}.r].(r-\rho\;v)}{\rho^{5}},\ ] ] and then , from ( [ agbc ] ) with @xmath132 @xmath133\;\;\longrightarrow b_{2}=[\hbox{\large a},v]+{\cal o(r)}\longrightarrow a=4.\ ] ] @xmath134\;\;\longrightarrow a_{1}=[{\hbox{\large a}},v]+{\cal o(r)}\longrightarrow p=5=n.\ ] ] this produces @xmath135.\eta.[\hbox{\large
a},v].{\hbox{\large a}}={\hbox{\large a}}^{2}{\hbox{\large a}},$ ] which is cancelled by the equal contribution from the trace term of ( [ t3 ] )
. therefore , @xmath136 * from the first term of the integrand of ( [ p1 ] ) and from ( [ t4 ] ) we have @xmath96.\eta.[k , v]k\rho{\hbox{\large a}}.k } { \rho^{4}}=\lim_{\rho\to0}\frac{[r , v].\eta.[r , v]r{\hbox{\large a}}.r } { \rho^{5}}.\ ] ] then , from ( [ agbc ] ) , with @xmath137 and @xmath138\;\;\longrightarrow a_{2}= b_{2}=[\hbox{\large a},v]+{{\cal o(r)}}\longrightarrow a=5 \hbox { and } \;p=7>n=5.\ ] ] therefore , from ( [ nr ] ) @xmath139 consequently @xmath140 and @xmath141 @xmath142 is distinguished from @xmath117 and @xmath118 for not being @xmath15-dependent .
therefore , it is not necessary to use the lhospital rule on its determination , which , by the way , is not affected by the limit of @xmath143 .
the physical meaning of this is that the flux of 4-momentum through the cylindrical surface @xmath144 comes entirely from the photon field .
+ it is convenient now to introduce the spacelike 4-vector n , defined by @xmath145 and such that @xmath146 and @xmath147 .
n satisfies @xmath148 @xmath149 as can be found , for example , in references [ 4,13 ] or in the appendix b of of reference [ 10 ] . from ( [ t3 ] ) and ( [ c1 ] ) , or from ( [ n1 ] ) , we have @xmath150 from ( [ t2e ] ) , we have @xmath151 which , with ( [ n1 ] ) and ( [ n2 ] ) , gives @xmath152 then , from ( [ tn ] ) , ( [ t3f ] ) and ( [ p0f ] ) , we have @xmath153 the larmor term .
+ finally , from ( [ eqm ] ) , ( [ gauss ] ) , ( [ p ] ) and ( [ larmor ] ) , we can write the electron equation of motion , obtained from the lienard - wiechert solution , as @xmath154 the external force provides the work for changing the charge velocity and for the energy dissipated by the radiation .
it is non - linear , like the lorentz - dirac equation , but it does not contain the schott term , the responsible for its spuriously behaving solutions . this is good since it signals that there will be no problem with causality violation .
but the schott term in the lorentz - dirac equation has also the role of giving the guaranty of energy conservation , which is obviously missing in ( [ finally ] ) . assuming that the external force is of electromagnetic origin , ( @xmath155 the contraction of v with eq .
( [ finally ] ) would require a contradictory @xmath156 .
but this is just an evidence that ( [ finally ] ) can not be regarded as a fundamental equation .
it would be better represented as @xmath157 with @xmath158 it is just an effective or average result , in the sense that the contributions from the electron self field must be calculated , as in ( [ eqm ] ) , by the electromagnetic energy - momentum content of a spacetime volume containing the charge world - line , in the limit of @xmath5 : @xmath159 observe that in the last term , v , the speed of the electron , is replaced by k the speed of the electromagnetic interaction ; only in the limit of @xmath87 is that @xmath160 we have to repeat the same steps from ( [ eqm ] ) to ( [ finally ] ) in order to calculate this last term and to prove ( done in the appendix ) that it is null : @xmath161 so , there is no contradiction anymore .
besides , it throws some light on the physical meaning of @xmath162 , which is discussed in @xcite . |
the organizers chose it . i m not sure what ` formal theory '
is supposed to cover , but they made the useful suggestion that my job is to summarize the parallel talks .
there were 21 talks , covering a good range of interesting topics . with just a little squeezing the talks fall into four broad categories : perturbative methods in gauge theory and gravity , applications of ads / cft duality , vacuum energy , and string models of particle physics and cosmology .
for reasons of space , i will cite the parallel talks rather than the original papers ; please see those talks for detailed references .
feynman graphs are central to quantum field theory , but it is now a familiar idea , going back more than twenty years to parke and taylor , that even in perturbation theory there can be better ways to organize things . in particular there are major simplifications of the amplitudes that are not evident in the graphical expansion .
for example , the @xmath0-gluon amplitude with maximally helicity violating ( mhv ) polarization would be given by a sum of thousands of terms from feynman graphs , yet the final answer collapses into a single term .
and progress continues , as reviewed by britto @xcite . for the next - to - mhv amplitude ,
the original spinor helicity methods give an expression that fits on a single slide .
with twistor methods , this is reduced to a couple of lines , and with the newest residue methods one obtains an expression nearly as compact as for the mhv amplitudes .
these techniques have wide applications , to standard model processes , to model systems like @xmath1 supersymmetry , and to ( supergravity ) .
the standard model applications were discussed in zanderighi s plenary talk @xcite , so i will focus on the more formal directions .
a natural question is whether these ongoing simplifications are a sign of some deeper structure , like a symmetry . in maldacena s
talk @xcite , it was shown that the answer is ` yes ' , at least in part . to explain what he found , i need to recall the 1974 observation of t hooft , that to understand qcd
we should think first about the limit where the number of colors @xmath2 is taken to be large , with @xmath3 held fixed . in this limit ,
only the planar feynman graphs survive .
one still has an infinite number of graphs , each quite complicated , so this does not immediately give a solution ( unlike some simpler large-@xmath4 systems ) .
nevertheless , t hooft has given us a tantalizing and inspiring goal , to solve the theory in this limit .
one reason one might hope for progress is that in this limit the problem is reduced from 3 + 1 dimensions to 1 + 1 dimensions , since we can draw all of the graphs on a piece of paper , without crossings .
theories in 1 + 1 dimensions are often simpler . for example , bethe was able to write down the exact many - body wavefunctions for certain spin systems in one spatial dimension : they reduce to the product of free wavefunctions , times a factor that depends only on the ordering of the particles along the line .
it is important that not all systems in 1 + 1 dimensions have a wavefunction that can be expressed in this way .
rather , it is related to the presence of a symmetry , of a somewhat unusual type . for most of the familiar symmetries in physics
, the conserved charge is the integral of a local current .
the bethe ansatz charges are given as the double ( or multiple ) integral of a product of local currents , times a function again of the order of the currents along the line .
such charges are evidently present in the planar limit of the @xmath1 theory : this is known as ` integrability , ' and has been the subject of major research activity in the past few years .
it has been mysterious how this is related to the amplitude simplifications , but now maldacena reports a connection : one of the observed properties of the scattering amplitudes , ` dual conformal symmetry , ' turns out to be a symmetry of the bethe ansatz type .
this may be the tip of a large and interesting iceberg . as further progress toward realizing t hooft s goal for @xmath1 , bern @xcite discussed a conjecture for the exact mhv amplitudes : the all - loop amplitudes seem to have a simple iterative structure .
for the 2-to-2 amplitudes , this seems to correctly match on to the strong coupling calculation from the dual string theory , but for the 2-to-4 amplitudes there is a discrepancy to be understood .
newton s constant @xmath5 has units of mass - squared , like fermi s constant @xmath6 , and so the dimensionless coupling @xmath7 grows with energy : the theory is nonrenormalizable .
two natural resolutions suggest themselves .
the first is that gravity is only an effective theory , and that new physics comes in at high energy that gives finite amplitudes . in the case of the weak interaction ,
this is what happens .
the divergences of the fermi theory were an important clue : they point almost uniquely to spontaneously broken non - abelian gauge theory . similarly in string theory new physics enters at high energy .
a second possibility is that the divergences are an artifact of perturbation theory and cancel in the full theory .
the talk by ward @xcite presented evidence that this happens , through a resummation of perturbation theory that suppresses the graviton propagator at large momentum .
this is a technical subject that is difficult to assess , but i will raise one issue that comes to mind .
attempts to cut off gravity covariantly often lead to difficulties because the ward identity relates the 3-point vertex to the inverse propagator .
if the propagator is suppressed , the vertex is correspondingly enhanced so there is no net effect in loops .
this might also happen in attempts to resum consistently .
the talk by bern @xcite presented a third possibility , at least for the most supersymmetric gravity theory , @xmath8 .
the supersymmetry guarantees that there are no divergences at one or two loops , and it was expected by many supergravity experts that the first divergence would appear at three loops .
the calculation has now been done , and there is no three - loop divergence .
there are cancellations that may not be due to supersymmetry .
this raises the possibility that these cancellations continue to higher orders , and that the amplitudes are finite order by order .
this would be a remarkable result , whose physical interpretation is not at all clear .
the dimensionless coupling still blows up at high energy , so one is working with a perturbation theory that breaks down , and a nonperturbative completion is needed . the cancellation does nt fit into wilson s way of looking at field theory : a uv fixed point would show up as a cancellation of divergences between different orders rather than their absence order - by - order .
but it is not at all clear to what order supersymmetry provides protection : bern quotes various arguments , which predict that the first divergences might arise at three , five , six , seven , eight , or even nine loops!. however , there is not a complete superspace formulation of the theory .
] so it is not clear how far one must calculate to be sure that the finiteness is not just a consequence of supersymmetry that will eventually fail at some order .
we discussed t hooft s idea that large-@xmath2 qcd might be solvable , but he also went further and proposed that the solution might involve rewriting qcd as a string theory ! roughly speaking , at strong coupling
one might imagine that the planar feynman graphs become dense with propagators and vertices , forming a continuous two - dimensional surface which is the world - sheet of the string .
the obvious problem is that attempts to quantize strings directly lead to gravity and extra dimensions , not to qcd .
this was resolved by maldacena in a surprising way : by putting the strings into anti - de sitter spacetime , the boundary conditions freeze gravity , and reduce the kinematics to four dimensions .
so far , this does not apply to qcd .
maldacena s original duality was for the same @xmath1 theory that we discussed earlier , and indeed the string description has been useful in understanding some of the behaviors seen in perturbation theory .
the @xmath1 theory is conformally invariant and so has no masses it is a theory of ` unparticles , ' in the current jargon .
one can then deform toward progressively less symmetric duals : the conformal invariance is useful to get oriented but is not central to the duality .
for this reason we use the more general term ` gauge / gravity duality ' ( or ` field / string duality ' ) rather than the more specific ads / cft .
so far this principle has not been extended in a precise way to asymptotically free theories like qcd . however , the ability to solve strongly coupled quantum field theories , to have a tool completely complementary to perturbation theory , is a terrific thing .
it is like the harmonic oscillator for strongly coupled gauge theories , and has had surprising even to real qcd .
this subject was beautifully reviewed in the plenary talk by rajagopal @xcite , so i will mention just a few highlights .
one of the surprises from rhic is that the guark - gluon plasma is strongly - coupled , behaving like a liquid rather than a gas .
we have no good tools from quantum field theory to study strongly coupled non - equilibrium physics , so the solvable @xmath1 theory provides at least a reality check , a model system to compare with .
in fact it seems much better than that , in that its extra symmetries ( supersymmetry , conformal invariance ) have little consequence in the finite temperature plasma , and moreover there seem to be some universal properties of strongly coupled gauge theories that one can learn from .
two important numbers are the free energy , which one learns is only corrected by about 25% as compared to the free theory , and the shear viscosity to entropy ratio , where one learns that the remarkably small value of @xmath9 is attained in many strongly coupled systems .
there were conjectures that this might represent a universal lower bound , but in the talk by myers @xcite it was shown that the model - dependent first correction can have either sign .
thus it is a sort of ideal value , but not a bound .
another example of the value of an exactly solvable system has been in the development of relativistic hydrodynamics .
this was reviewed in the talk by starinets @xcite .
rajagopal gave a long list of other areas where the dual theory has given valuable insights @xcite .
i refer you to his talk for the details , and will focus on the conceptual question , is rhic really making black holes ? " the obvious answer is , `` of course not , it s just quarks and gluons , '' but i would argue that the truth is more subtle .
a duality means that we have a quantum theory that has more than one classical limit . in the present case , one limit is gauge theory and the other is string theory .
such a situation erases distinctions that we make with our classical experience .
the familiar example is wave / particle duality , but now we have the same thing happening between black holes and the quark - gluon plasma . of course ,
with wave / particle duality there are situations where things are clearly wave - like or clearly particle - like , but we know that we can interpolate between these . with gauge / gravity duality ,
to interpolate we have to imagine varying the parameters of the theory . in a sense
we can do this , because in randall - sundrum models one has physics with both gauge and gravitational descriptions , and depending on the parameters the physics is closer to one or to the other .
but even in qcd , the surprise has been how successful the gravitational description is . for the shear viscosity , and other properties ,
the best calculations have not had anything to do with quarks and gluons , but instead come from the propagation of gravitational fields near a black hole !
there are many condensed matter systems with ` quantum critical points , ' strongly coupled phases that have universal low energy behaviors but no simple description .
for example , it is believed that the high temperature superconductors exhibit this phenomenon .
there has recently been some attention to the idea that there might be model systems with gravitational duals , where one might be able to do calculations and gain insight that can not be done in any other way .
some condensed matter systems have effectively relativistic behavior at the critical point , but others are nonrelativistic , with space and time scaling differently . the talk by kachru @xcite described a class of dual systems with this property .
it was good to see several talks about vacuum energy , because this is such an important problem .
the vacuum is highly nontrivial full of higgs fields , color fluctuations , zero point energies and gravity is a universal force .
so why is the vacuum energy not huge ?
this question has been around since pauli , but since the discovery of the cosmic acceleration we have two new puzzles : why is the vacuum energy not exactly zero , and why is its magnitude so similar to the matter density in the universe in the current era ( cosmic coincidence ) .
these new questions get a lot of attention , but the original problem of 60 or 120 orders of magnitude is still there , and this is what i will focus on .
many many solutions have been proposed over the years , but it is very difficult to get around the basic problem : the vacuum is full of stuff , and everything gravitates . a useful litmus test for proposed solutions is to ask what cancels the electron zero point energy ( zpe ) . at the scale of the electron mass
we know that quantum field theory works to high precision , and the modes from this scale and below generate a vacuum energy 30 orders of magnitude larger than that observed .
this is large enough , as pauli noted , that the universe `` would not even reach out to the moon '' @xcite .
one might imagine that for some reason zero point energy does not gravitate , but any such idea will conflict with the great success of quantum field theory . for example , in the standard model the top quark zpe is an important contribution to the higgs potential and the higgs mass .
if this did not gravitate it would feed through to the rest of the standard model and spoil the universality of gravity , conflicting strongly both with theory and experiment . as another example
, the gravitation of the vacuum polarization contribution to atomic energy levels , which is a shift in the zero point energy due to the field of the atom , can be seen to high precision in the eotvos experiment .
the talk by milton @xcite showed that the gravitation of casimir energy is consistent with renormalization .
if the vacuum energy were exactly zero , it would be natural to look for a symmetry explanation .
supersymmetry can set the vacuum energy to zero .
symmetry under scaling of lengths and energies ( or conformal transformations ) could also have this effect .
but if these are symmetries of nature they must be broken , since the electron has nonzero mass and is not degenerate with the selectron .
thus these ideas do not pass the litmus test , because the zero point energy depends on the electron mass and so , even if it vanishes in a symmetric state , it is induced by the breaking .
two of the parallel talks explored symmetry solutions to the cosmological constant problem @xcite .
mannheim proposed a solution based on broken conformal symmetry , but i believe that in the end the net contribution of the electron zpe is exactly the same is it would be without this symmetry .
nevzorov revisits the interesting no - scale structure , where the vacuum energy in certain supergravity theories is zero even after supersymmetry breaking .
however , this is only a classical result , and it does not survive loop corrections such as the zpe .
so these two attempts based on symmetry seem to run into the usual problems .
besides symmetries , the other broad class of ideas is adjustment mechanisms .
in addition to the known physics there is an ` adjustment sector , ' whose vacuum energy is variable in some way so as to cancel that from the rest of the theory .
the adjustment sector must have a large number of stable states ( vacua ) with different values of the energy density . but how can such an adjustment sector arise from fundamental physics ?
in fact , string theory seems to provide a realization of this : by varying the shape of the compact dimensions , and the configurations of branes and fluxes , there is the potential to have an enormous number of long - lived vacua , with finely spaced values of the vacuum energy .
as discussed above , we have a clue from cosmology that the fundamental theory must have many vacua , with adjustable vacuum energy . as model builders
, it would be nice if only the vacuum energy adjusted while all the other observable parameters descended from the underlying theory in a fixed way .
unfortunately there is no particular reason that nature should be so nice to us , and this does not seem to be what happens in string theory : the other constants also vary with the geometry of the compact dimensions .
this of course makes it challenging to connect the underlying theory to observation , but the parallel session speakers were undaunted .
even very large numbers of vacua need not be an obstacle : sometimes , as in statistical mechanics , one can make predictions precisely because the number of available states is very large .
the challenge for now is to map out this space of states , and figure out the rules by which it is populated .
gmeiner @xcite considered a large class of vacua based on intersecting d6-branes and o6-planes .
this is a broad framework for constructing standard model - like vacua , and general features may emerge that depend on the brane configuration but are insensitive to other details of the compactification .
he presents statistics for various characteristics such as the number of generations .
in contrast to earlier work along the same lines , the number three does not seem to be disfavored .
ratz @xcite explored heterotic orbifolds with ` local grand unification , ' where the grand unified ( gut ) symmetry is broken locally in the compact dimensions . in this way the successful predictions of guts can survive , such as the unification of couplings , while some of the problems are absent , notably the color triplet partners of the higgs that would induce rapid proton decay
this was a very optimistic talk : putting in only a small number of ingredients , such as a unification structure and @xmath10 symmetry , the authors found interesting results for the flavor structure , neutrino see - saw , supersymmetry breaking , and the @xmath11-problem .
one of the big questions for the lhc is whether we can find more evidence for guts .
the prediction of @xmath12 is striking , but with just one number it may still be an accident .
camara @xcite , krefl @xcite , and weigand @xcite all discussed the technology of the calculation of string loop and instanton corrections , which may be important for supersymmetry breaking and mediation , the @xmath11 term , and neutrino and other fermion masses .
supersymmetry breaking and its mediation is one place where observable physics may connect closely to the short distance theory , and in particular to the geometry of the compact dimensions .
seiberg @xcite discussed gauge mediation in a model - independent way .
there are several generic predictions : sfermion degeneracy and absence of flavor changing neutral currents , general mass relations @xmath13 , small @xmath14 terms , problems with @xmath11 and @xmath15 , and the gravitino as the lightest supersymmetric partner .
other predictions , such as gaugino and mass sfermion relations , are model dependent .
weigand @xcite presented a string theory realization of gauge mediation .
early cosmology is another area where observation may connect to short distance physics and to the geometry of string compactification . in string models ,
the potential energy responsible for inflation is closely tied to the dynamics of moduli stabilization .
this sector is generally independent of the sector responsible for standard model , though they may couple through supersymmetry breaking . just as the standard model has many kinds of vacuum energy , so does the moduli sector : branes , fluxes , curvature , curvature , zpe , instantons , @xmath16 effects . only in the past few years
have these been have these been under enough control to build detailed models .
slow - roll inflation requires a potential which is a nearly flat function of a scalar field .
one natural mechanism is to have an extra brane - antibrane pair , where the inflaton is the separation
. models of this type were discussed by haack @xcite and mcallister @xcite .
another possible potential comes from shape deformations of the compact dimensions ( moduli ) , as discussed by quevedo @xcite in large - dimension models .
the phenomenology of inflation centers on two key cosmic microwave background ( cmb ) parameters , the ratio @xmath17 of tensor to scalar perturbations and the spectral slope @xmath18 .
the default harrison - zeldovich model has a scale invariant spectrum , @xmath19 , with no tensor modes , @xmath20 .
the original slow - roll inflation models made the distinctive predictions of nonzero @xmath17 and @xmath18 slightly red , though the inflationary model - building industry can now populate much more of the parameter space . the first detailed string model , kklmmt , gave @xmath20 and @xmath21 @xcite . the various models discussed at this meeting are shown on fig . 1 .
there is an important distinction between models with @xmath17 zero or nonzero .
models with @xmath20 generally involve relatively small changes in the brane positions or moduli , possibly involving only a small region of the compact space , while @xmath22 requires large changes that generally involve the whole space
. there will be many new cmb experiments in next five to ten years planck , spider , clover , quiet , bicep , ebex , polarbear , ... and the error bars in both directions will be several times smaller . while this ca nt falsify all inflation models , or all string inflation models , it will falsify most of them , and hopefully focus attention on certain kinds of model and compactification
. ultimately it would be valuable to have more than two parameters .
possible additional observables are nongaussianities and cosmic strings .
there have been suggestions that the data give some evidence for these , but at no more than a two - sigma level .
judging from the parallel talks , ` formal theory ' covers the development of technical tools needed to describe nature , and moreover the application of these tools to a broad range of physics .
the central technical tool , quantum field theory , is around eighty years old , and it is clear that we are still learning new things . our new tool , string theory , is not quite half as old .
we have learned some important things about it , notably all the ways in which it is entangled with quantum field theory , and we still have a lot more to learn .
j. polchinski , `` the cosmological constant and the string landscape , '' in _ 23rd solvay conference in physics : the quantum structure of space and time , _ eds .
d. gross , m. henneaux , a. sevrin .
hackensack , world scientific , 2007 .
[ arxiv : hep - th/0603249 ] .
see n. straumann , `` the mystery of the cosmic vacuum energy density and the accelerated expansion of the universe , '' eur .
j. phys .
* 20 * , 419 ( 1999 ) [ arxiv : astro - ph/9908342 ] .
k. milton , these proceedings .
s. kachru , r. kallosh , a. linde , j. m. maldacena , l. p. mcallister and s. p. trivedi , `` towards inflation in string theory , '' jcap * 0310 * , 013 ( 2003 ) [ arxiv : hep - th/0308055 ] .
e. komatsu _ et al .
_ [ wmap collaboration ] , `` five - year wilkinson microwave anisotropy probe ( wmap ) observations : cosmological interpretation , '' arxiv:0803.0547 [ astro - ph ] . | this is a summary talk covering recent progress in perturbative methods for gauge theories and gravity , applications of ads / cft duality , vacuum energy , and string theory models of particle physics and cosmology . |
the security of coherent state continuous variable quantum key distribution ( cv - qkd ) @xcite is fundamentally based on the inability of an eavesdropper to perfectly distinguish between non - orthogonal quantum states @xcite . in this paper , we look at how much information a potential eavesdropper can gain when trying to distinguish between two pure coherent states as opposed to distinguishing between two mixed coherent states .
this is of particular interest in cv - qkd protocols , such as post - selection @xcite ) , where it is important to determine if an eavesdropper obtained more information in the case of distinguishing between two pure coherent states or distinguishing between two mixed states .
in our analysis , we will use the _ probability of error _ ( @xmath0 ) measure to distinguish between quantum states .
we point out that one could potentially consider other distinguishability measures such as the kolmogorov distance , the bhattacharyya coefficient and the shannon distinguishability ( for a review of these measures see @xcite ) .
however , as we shall see the probability of error measure has a number of useful properties and can be directly calculated for the quantum states we consider in our analysis .
we consider the distinguishability between two general quantum states that are described by the two density matrices @xmath1 and @xmath2 .
it was originally shown by helstrom @xcite that the probability of error between these two density matrices is minimized by performing an optimal positive operator - valued measure ( povm ) @xcite . in this case , the probability of error for the distinguishing between two general quantum states can be expressed as @xcite @xmath3 $ ] and @xmath4 are the eigenvalues of the matrix @xmath5 .
we note that when the two states are indistinguishable the probability of error is @xmath6 . on the other hand , in the case
when the two states are completely distinguishable the probability of error is @xmath7 . and
@xmath8 ) and ( b ) two mixed coherent states ( @xmath9 and @xmath10 ) . here the dotted lines and shadings in ( b ) indicate which of the two coherent states are mixed.,width=453 ]
we now look at distinguishing between two coherent states using the previously defined probability of error .
a coherent state is defined as @xmath11 where @xmath12 is the displacement operator .
it is also a minimum uncertainty state and an eigenstate of the annihilation operator @xmath13 , i.e. @xmath14 where @xmath15 is the amplitude of the electromagnetic wave @xcite .
any two coherent states @xmath16 and @xmath17 are always non - orthogonal and only approach orthogonality ( i.e. @xmath18 ) when @xmath19 where the magnitude is @xmath20 . in the following analysis
we will define a coherent state displace in the amplitude and phase quadratures @xcite , by an amount @xmath21 and @xmath22 respectively , as @xmath23 .
consequently , we can write the density operators of two pure coherent states @xmath24 and @xmath25 that we consider here as @xmath26 in our analysis we also consider two mixed coherent states , i.e. an equally weighted mixture of coherent states mirrored in the phase quadrature and with both mixtures having the same amplitude component .
the density operators corresponding to these two mixed states , @xmath27 and @xmath28 , are defined as @xmath29a ) and fig .
( [ two_mixed_vs_two_pure_discrete]b ) give a two - dimensional phase space illustration of the two pure coherent states and the two mixed coherent states defined by eq .
( 2 ) and eq .
( 3 ) respectively . according to eq . ( [ error_probability ] )
we need to determine the eigenvalues of @xmath30 for both the two pure states and two mixed states . to do this we write @xmath31 in its matrix representation which can be expanded in terms of the orthogonal fock or number states @xmath32 defined as @xcite @xmath33 where @xmath34 is the creation operator of a harmonic oscillator and @xmath35 .
for example , the coherent state @xmath36 written in terms of this fock basis is @xmath37 once @xmath31 is written in matrix form we can then numerically determine its eigenvalues . in this fock state expansion the inner product of an arbitrary coherent state with a fock state is given by @xmath38 where @xmath32 and @xmath39 are fock states . calculating the general matrix coefficients for the case of the two pure coherent states we obtain @xmath40\end{aligned}\ ] ] similarly for the two mixed state case we find @xmath41\end{aligned}\ ] ]
numerically we calculate the eigenvalues of eq .
( [ matrices_pure ] ) and eq .
( [ matrices_mixed ] ) up to certain values of @xmath42 and @xmath43 .
then according to eq .
( [ error_probability ] ) this will give us the probability of error in distinguishing between two quantum states .
now having numerically calculated @xmath0 we would like to interpret this in terms of the information gained from using the distinguishing measure .
in the context of cv - qkd it is important to determine how much shannon information an eavesdropper can obtain by distinguishing between two ( pure or mixed ) quantum states . the information obtained by distinguishing between two states can be calculated using the shannon information formula for a binary symmetric channel @xcite @xmath44 figure ( [ 3d_plot ] ) shows the difference between the shannon information obtained by distinguishing between two coherent states @xmath45 compared with distinguishing between two mixed states @xmath46 .
this information difference is defined as @xmath47 .
figure ( [ 3d_plot ] ) plots @xmath48 in terms of the amplitude and phase quadrature displacements of the pure and mixed states as defined in eq . ( 3 ) and eq .
( 4 ) , respectively . here
we have expanded up to @xmath49 fock states , i.e. @xmath50 in our numerical analysis .
there are two main features of this plot .
firstly , we notice that , given our distinguishability measure and initial configuration of coherent states in phase space , two mixed states never give more information than two pure state , i.e. @xmath51 .
secondly , there is a flat region where the information gain is zero , i.e. the information from distinguishing between two mixed states is the same as that of two pure states .
this means as we move the states further and further apart in the amplitude quadrature ( while keeping the phase quadrature fixed ) , the probability of error tends to zero and hence an information gain of zero .
the same result occurs when the amplitude quadrature is fixed while varying the phase quadrature .
this is somewhat surprising because the more separated two mixed states become the more indistinguishable they are and consequently less information can be extracted .
but in this case what it is telling us is that at some point the two mixed states start behaving " like two pure states . .
in this case the two mixed states never give more information than two pure states.,width=302 ]
in conclusion , we have shown that a continuous variable quantum key distribution protocol where an eavesdropper needs to distinguish between two pure coherent states , rather than two mixed coherent states ( where the various mixtures have the same amplitude component ) , the eavesdropper will never get more information from the two mixed coherent states .
we showed this using the probability of error as the distinguishability measure along with the shannon information formula . | we use the probability of error as a measure of distinguishability between two pure and two mixed symmetric coherent states in the context of continuous variable quantum cryptography .
we show that the two mixed symmetric coherent states ( in which the various components have the same real part ) never give an eavesdropper more information than two pure coherent states . |
the massive x - ray binary cygnus x-1 is a seminal target in the study of gas dynamics in the vicinity of a stellar mass black hole .
its x - ray luminosity and energetic jets @xcite are powered by gas accretion from the nearby companion star hd 226868 ( o9.7 iab ; @xcite ) in a spectroscopic binary with a 5.6 day orbital period .
there are several ways in which mass transfer from the supergiant to the black hole may occur in this system @xcite .
the o - supergiant , like other massive and luminous stars , has a strong radiatively driven wind that may be partially accreted through the gravitational force of the black hole .
the supergiant is large and is probably close to filling its critical roche surface @xcite , so a gas stream through the inner l1 point may also be present .
the actual gas flow in the direction of the black hole is probably intermediate between a spherically symmetric wind and a roche lobe overflow stream , and there is evidence that the flow is best described as a focused wind @xcite .
the gas ions responsible for accelerating the wind may become ionized in the presence of a strong x - ray source , leading to a lower velocity , `` stalled '' wind @xcite . in situations of very high x - ray flux , photoionization may extend so close to the supergiant s photosphere that the wind never reaches the stellar escape velocity and thus ceases to become an x - ray accretion source @xcite .
however , such a high x - ray flux may heat the outer gas layers to temperatures where the thermal velocities exceed the escape velocity to create a thermal wind that may fuel black hole accretion .
important clues about the mass transfer process come from the temporal variations of the observed x - ray flux .
cyg x-1 is generally observed in either a low flux / hard spectrum state , with an x - ray spectrum that is relatively flat , or a high flux / soft spectrum state with a steeper power - law spectrum @xcite .
the gamma - ray portion of the spectrum is also elevated during the high / soft state @xcite .
the high / soft state usually lasts for periods of days to months , and the fraction of time observed in the high / soft state has increased from @xmath2 in 19962000 to @xmath3 since early 2000 @xcite .
this increase may be related to an overall increase in the supergiant s radius in the period from 1997 to 2003 2004 that is suggested by changes in the long term optical light curve @xcite .
the system sometimes experiences so - called failed - state transitions , when it starts to increase in flux , but then stops at an intermediate state and returns to the low / hard state .
all these transitions probably reflect changes in the inner truncation radius of the accretion disk surrounding the black hole that are caused by a variable accretion rate ( largest when the system is in the high / soft state ; @xcite ) .
thus , the temporal variations in the x - ray state offer us the means to compare the black hole accretion processes with observational signatures related to mass transfer . the h@xmath0 emission formed in the high density gas at the base of the stellar wind is an important diagnostic of the mass loss rate in massive stars @xcite .
the h@xmath0 emission variations in hd 226868 over the last few years are documented in independent spectroscopic investigations by @xcite and @xcite .
both of these studies concluded that the h@xmath0 emission appears strongest when the system is in the low / hard x - ray state , while a range of weak to moderate emission strengths are observed during the high / soft states .
this is a surprising result , since taken at face value , strong emission is associated with a large wind mass loss rate , and the simplest expectation that the x - ray accretion flux increases with mass loss rate is , in fact , not observed .
there are several possible explanations : ( 1 ) a denser wind may be more opaque to x - rays .
however , this seems unlikely because the observed inverse relation between h@xmath0 emission strength and x - ray flux is observed at all orbital phases , not just when the supergiant and its wind are in the foreground .
( 2 ) the x - ray source may photoionize and heat the gas responsible for the h@xmath0 emission , so that a larger x - ray flux leads to a decrease in h@xmath0 strength .
this clearly occurs at some level , but both @xcite and @xcite argue that portions of the wind shaded from the ionizing flux also display significant temporal variations .
( 3 ) changes in the x - ray flux will lead to variations in the ionized volume of gas surrounding the black hole , and consequently , the total acceleration of the wind in the direction towards the black hole will vary with the distance traveled before the atoms responsible for line - driving are ionized .
thus , a stronger , denser wind might reach a faster speed before ionization , and since the bondi - hoyle accretion rate varies as @xmath4 , the gas captured by the black hole ( and the associated x - ray flux ) declines .
this last process can be tested through direct study of the degree of wind ionization observed in the ultraviolet p cygni lines formed in the supersonic part of the wind outflow .
when the system is observed with the ionization region in the foreground , the absorption cores of these p cygni lines will be truncated at a blueshift corresponding to the highest projected speed before encountering the ionization zone , the so - called hatchett - mccray effect @xcite .
the binary is so faint in the ultraviolet that high dispersion spectra were very difficult to obtain with the _ international ultraviolet explorer ( iue ) _
satellite @xcite , but a good series of observations were obtained with _
iue _ at lower spectral resolution that clearly indicate the weakening of the wind lines when the black hole is in front @xcite .
most of these spectra were obtained in the low / hard x - ray state ( 5 ) , when according to the varying wind strength model the mass loss rate is higher and the wind is less ionized and faster .
we embarked on a new program of high s / n , high dispersion uv spectroscopy to test this hypothesis with the _ hubble space telescope _ space telescope imaging spectrograph ( stis ) .
we obtained observations at the two orbital conjunction phases in both 2002 and 2003 .
these sets of observations were both made during the rare high / soft x - ray state , and planned observations during the low / hard state were unfortunately scuttled by the stis electronics failure in 2004 .
however , we can rebin the high quality _ hst _ spectra made during the high / soft state to the lower resolution of the _ iue _ archival spectra ( mostly low / hard state ) in order to test whether or not the wind ionization state does in fact differ significantly between states .
we describe a program of supporting optical spectroscopy we have obtained to check the orbital phase ( 2 ) and wind strength ( 3 ) at the times of the _ hst _ observations .
we compare the h@xmath0 measurements with the contemporaneous x - ray light curve recorded with the _ rossi x - ray timing explorer _ all - sky monitor instrument @xcite and confirm our earlier result showing how h@xmath0 tends to strengthen as the x - ray flux declines ( 3 ) .
we then describe the observed variations in the main uv wind lines and present a simple `` shadow wind '' model for the profiles ( 4 ) .
we compare the variations observed in the _ hst _ and _ iue _ spectra in 5 , and then reassess the question of the mass transfer process in 6 .
we will discuss the photospheric features in the uv spectra in a forthcoming paper ( caballero nieves et al .
, in preparation ) .
the _ hst _ stis spectra were obtained with the first - order g140 m grating in a series of subexposures at different grating tilts in order to record the uv spectrum over the full available range ( 1150 to 1740 ) .
two full sets were made near each orbital conjunction phase in runs in both 2002 and 2003 .
all the spectra were reduced using the idl software developed at nasa goddard space flight center for the stis instrument definition team .
the spectra were rebinned on a @xmath5 , heliocentric wavelength scale to a spectral resolution of @xmath6 and rectified to a pseudo - continuum based upon the flux in relatively line - free regions .
the space observations were supported by contemporaneous , ground - based observations of the red spectrum in the vicinity of h@xmath0 .
these 125 spectra were made with the university of toronto david dunlap observatory 1.88 m telescope , noao kitt peak national observatory 0.9 m coud feed telescope , herzberg institute of astrophysics dominion astrophysical observatory 1.85 m telescope , and the academy of sciences of the czech republic astronomical institute ondejov observatory 2 m telescope .
we also obtained a smaller set of 22 blue spectra that record the variations in the @xmath7 line that probably forms in the focused part of the wind @xcite .
a summary of all these observing runs is given in table 1 that lists run number , range of heliocentric julian dates of observation , spectral range recorded , spectral resolving power ( @xmath8 , where @xmath9 fwhm of the line spread function ) , number of spectra obtained , and details about the telescope , spectrograph , and detector .
the spectra were reduced and transformed to rectified flux on a uniform heliocentric wavelength scale ( as described in @xcite ) .
most of the red spectra also record the @xmath10 absorption line , and we decided to measure the stellar radial velocity from this line to check on the orbital ephemeris at the time of the _ hst _ observations .
the radial velocities were measured in the same way as outlined by @xcite by fitting a gaussian to the central line core . for the sake of completeness
, we also measured the radial velocity of the supergiant from the _ hst _ uv spectra using a cross - correlation method @xcite with an _ iue _ spectrum of hd 34078 as the reference template .
the results are presented in table 2 ( given in full in the electronic version ) that lists the heliocentric julian date of mid - observation , orbital phase , radial velocity and its associated error , observed minus calculated velocity residual , h@xmath0 equivalent width , and the corresponding run number from table 1 .
we note that independent measurements of the @xmath10 line in the ondejov spectra using the korel package @xcite led to fully consistent results .
we computed orbital elements from these velocities ( omitting those from _ hst _ ) using the non - linear , least - squares fitting method of @xcite .
we made a circular orbital fit with the period fixed at the value obtained by @xcite from data spanning a 26 yr interval .
our results are compared to those from @xcite and from @xcite in table 3 , and they are consistent with these earlier studies .
the current epoch for the time of supergiant inferior conjunction @xmath11(ic ) occurs @xmath12 days later than the prediction from @xcite , and we will adopt this revised epoch for the definition of orbital phase throughout the paper .
we omitted the _ hst _
measurements from the orbital solution because of concerns about possible systematic differences in the velocities derived from the uv lines and @xmath10 , but the @xmath13 residuals for the _ hst _ measurements given in table 2 show that the uv measurements are in reasonable agreement with the velocity curve derived from @xmath10 .
the optical red spectra were obtained with the primary goal of monitoring the gas density at the base of the stellar wind of the supergiant .
the h@xmath0 observations of hd 226868 are summarized in two panels in figure 1 according to the x - ray state at the time of observation ( see fig . 2 below ) .
the top portions show plots of the profiles arranged by orbital phase while the lower grayscale images show the spectral flux interpolated in radial velocity and orbital phase .
the white line in the lower image shows the radial velocity curve of the supergiant ( table 3 ) .
these figures show that most of the emission / absorption complex appears to follow the orbit of the supergiant as expected for an origin in the supergiant wind .
we measured the h@xmath0 equivalent width in the same way as before @xcite by making a numerical integration over a 40 range centered on h@xmath0 , and these measurements are listed in column 6 of table 2 .
we estimate that the typical measurement error is @xmath14 ( depending mainly on the s / n ratio of the individual spectrum ) .
we show the time evolution of the h@xmath0 equivalent width for a total of 240 measurements from @xcite and the new observations in the top panel of figure 2 .
the lower panel of this figure shows the daily average soft x - ray flux over the same interval from the all - sky monitor on the _ rossi x - ray timing explorer _ @xcite .
these flux measurements are the quick - look results provided by the _
rxte_/asm team .
the two arrows in the top panel indicate the times of the two _ hst _ observing runs , which took place when cyg x-1 was in the high / soft state .
the new measurements confirm the trends described by @xcite and @xcite that the h@xmath0 emission tends to be stronger when the soft x - ray flux declines and that there is a considerable range in emission strength when the soft x - ray flux is large ( see fig . 1 ) .
figures 3 and 4 show a detailed view of the time evolution of the h@xmath0 emission strength and x - ray flux for the week surrounding the _ hst _ runs in 2002 and 2003 .
the bottom panels in these figures show the x - ray fluxes in both the low energy ( 1.5 3 kev ; @xmath15 signs ) and higher energy ( 5 12 kev ; @xmath16 signs ) bands for the individual , 90 s exposure , dwell measurements . unfortunately ,
both the h@xmath0 and x - ray measurements are not exactly coincident in time with the _ hst _ observations , so we have made a time interpolation between the closest available measurements to estimate the h@xmath0 emission and x - ray flux levels at the times of the _ hst _ observations ( summarized in table 4 ) .
all four _ hst _ observations occurred when the h@xmath0 emission was weak and the soft x - ray flux was uniformly strong .
the other important optical emission line in the spectrum of hd 226868 is the @xmath7 feature . @xcite and @xcite found that this emission probably forms between the supergiant and black hole in a higher density and slower region of the wind , the focused wind predicted by @xcite .
we obtained a limited number of spectra of the @xmath7 feature ( runs 8 and 9 in table 1 ) at the times of the two _ hst _ observation sets , and these observations show profile variations with orbital phase that are quite similar to those seen previously @xcite .
we present in table 5 the observed equivalent width of the @xmath7 emission / absorption complex ( made by numerical integration between 4677 and 4695 ) for spectra from these two runs and from a third run ( # 10 in table 1 ) that occurred when the system had returned to the x - ray low / hard state .
there was no significant difference in the amount of @xmath7 emission present between the times of the two x - ray states .
we compare in table 6 the averages of these measurements with earlier orbital phase averages @xcite and time averages @xcite of the emission strength that correspond mainly to times when the system was in the low / hard x - ray state .
note that @xcite report only the net emission equivalent width after subtraction of the photospheric absorption profile of the similar star hd 149038 , and we estimate that this procedure increased the emission strength by @xmath17 according to the plot of the spectrum of hd 149038 given in figure 1 of @xcite .
we find that the @xmath7 emission has remained more or less constant in strength over the decades of observation and between the x - ray states , and this suggests the tidal gas stream towards the black hole experiences much less variation than does the global wind as observed in the h@xmath0 emission line .
our primary interest here is how the x - ray flux ionizes portions of the supergiant s wind and how our line of sight through the ionized zones changes with orbital phase .
we show in figures 5 , 6 , and 7 the changes observed between conjunctions in the major uv wind lines of @xmath18 , @xmath19 , and @xmath20 . in a companion paper , @xcite show the variations observed in several other weaker lines .
the top panel in each of these figures shows the variations in 2002 between the phases with the black hole behind ( @xmath21 ; _ solid line _ ) and in the foreground of the supergiant ( @xmath22 ; _ dotted line _ ) , and the bottom panel shows the same for the 2003 run .
the spectra are plotted as a function of radial velocity for the shorter wavelength component in the frame of the supergiant ( according to the orbital solution in table 3 ) .
all three of these transitions display a large reduction in the extent and depth of the blueshifted absorption component when the black hole is in the foreground ( the hatchett - mccray effect ) .
these changes reflect the x - ray photoionization and resulting superionization of these ions in the wind gas seen projected against the supergiant .
we also find some evidence of the associated reduction in the strength of the red emission component due to the loss of these gas ions in the outflow away from our line of sight when the black hole is in the background .
the variations between conjunctions appear to be almost identical for the observations in 2002 and 2003 , which is probably due to the very similar x - ray fluxes that existed at those times of observation ( 3 , table 4 ) .
the shapes of the wind profiles near orbital phase @xmath22 suggest that the p cygni absorption troughs have almost entirely disappeared , or that the wind ionization extends all the way towards the exposed photosphere of the supergiant .
we made some simple calculations of the appearance of wind profiles that arise only from the gas hidden from the x - ray source , the so - called shadow wind @xcite that is illustrated diagrammatically in figure 8 .
here we assume that the normal supergiant wind is confined to the region where the line of sight to the black hole is blocked by the supergiant .
however , we expect that in reality this shadow region is partially exposed to x - rays by wind scattering ( probably comparable to the scattering along our line of sight through the wind , i.e. , a few percent of the unobscured x - ray flux ; @xcite ) and by the coriolis deflection that will bring the shadowed gas into regions of x - ray illumination @xcite .
the line synthesis calculations are based on a modification of the sobolev - exact integration method @xcite that was developed by @xcite .
the simplifying assumption in this model is that the observer lies along the axis joining the stars at the two conjunctions ( or that the orbital inclination is @xmath23 and the orbital phases are @xmath24 ) .
however , since the actual inclination is smaller ( @xmath25 ; @xcite ) , our line of sight at the conjunctions will include somewhat different portions of the occulted and unocculted wind ( fig .
8) , so these models are first approximations of the predicted variations for a shadow wind . in a companion paper @xcite ,
we present a more complete calculation based upon a realistic orbital inclination and the method outlined by @xcite .
the line synthesis is based upon a set of adopted parameters and two fitting parameters .
most of the adopted parameters come from the study of cyg x-1 and similar x - ray binaries by @xcite , and in particular , we assume a wind velocity law exponent of @xmath26 ( eq . 2 in @xcite ) , a semimajor axis equivalent to @xmath27 ( where @xmath28 is the radius of the supergiant ; @xcite ) , and a characteristic turbulent velocity in the wind equal to @xmath29 @xcite where @xmath30 is the terminal velocity in the undisturbed wind .
the photospheric components corresponding to the wind transitions were assumed to be gaussian in shape with parameters set by gaussian fits of the photospheric profiles in the non - lte , line blanketed model spectra of @xcite ( for @xmath31 k , @xmath32 , @xmath33 km s@xmath34 , a linear limb darkening coefficient of @xmath35 , and a spectral resolving power of 10000 ) .
the final two parameters , the wind terminal velocity @xmath30 and the integrated optical depth through the shadow wind of the blue component of the transition @xmath36 , were fit by trial and error in order to match the entire set of the three wind lines at each conjunction .
note that the value of @xmath30 is set mainly by fits of the profiles at orbital phase @xmath37 when the undisturbed wind is in the foreground and projected against the star .
the best match was made with @xmath38 km s@xmath34 and @xmath39 , 4 , and 10 for the , , features , respectively . the shadow wind model profiles are presented in figures 9 , 10 , and 11 for these three wind lines .
the top panel in these figures shows the model ( _ diamonds _ ) and average observed spectra ( _ solid line _ ) for @xmath37 while the lower panel illustrates the same for phase @xmath40 ( black hole in the foreground ) .
the model profiles were renormalized in flux in each case to match the pseudo - continuum beyond the line wings in the observed spectra .
each panel also contains the predicted photospheric spectrum for the region from the models of @xcite .
these representative fits successfully reproduce many of the profile characteristics , especially when line blending from the other photospheric lines is taken into account .
this suggests that the geometry of the photoionization zone is probably not too different from that assumed for a shadow wind ( at least for the high / soft x - ray state ) .
however , we caution again that our fitting parameters are based upon an axial viewing orientation while our actual view is more oblique . for example , according to the geometry sketched in figure 8 ( for @xmath41 ) , the fastest moving part of the shadow wind that is projected against the supergiant at @xmath37 occurs at a radial distance from the center of the supergiant of @xmath42 where the wind has not yet reached terminal velocity .
thus , our fit value of @xmath38 km s@xmath34 is probably well below the actual value ( which is probably closer to 1600 km s@xmath34 ) . the low orbital inclination and the subsequent limited projection of the shadow wind region against the disk of the photosphere results in p cygni absorption troughs that are unusually weak for the spectra of supergiants like hd 226868 .
we show in figures 12 and 13 montages of the and wind features in four other o9.7 iab supergiants as seen in high dispersion spectra from the _ iue _ archive
. the mean spectrum of hd 226868 at orbital phase @xmath37 ( shown at the top of both figures ) shows that the wind profiles have a lower optical depth and attain a smaller blue - shifted velocity because only a small portion of the shadow wind is projected against the disk of the supergiant at its inferior conjunction ( see fig .
our original goal was to obtain another set of stis spectra when cyg x-1 was in the low / hard x - ray state in order to determine how the wind ionization conditions change with x - ray state . with the loss of stis ,
such a comparison is not possible at present .
however , the low dispersion fuv spectra of hd 226868 made with _
iue _ were in most cases made when cyg x-1 was in the low / hard state .
thus , we can investigate differences in the wind ionization properties between x - ray states through a comparison of the _ hst _ high / soft state spectra with the _ iue _ low / hard state spectra .
we collected 30 , low dispersion , short wavelength prime camera spectra of hd 226868 from the _ iue _ archive , and transformed these to rectified flux versions on a uniform wavelength grid .
next , we smoothed , rectified , and rebinned the _ hst _ stis spectra onto the same _ iue _ wavelength grid so that the line blended structures would appear the same as they do in the _ iue _ spectra .
we then measured the effective absorption strength by determining the mean flux across a spectral range that extends over the full range of the apparent wind feature ( as done by @xcite ) .
this average flux will reflect both the changing p cygni absorption and the other line blends ( including interstellar components ) , but since the latter are generally constant in time , the average flux will serve to show the relative variations in the wind absorption ( low flux when the p cygni trough is deep and high flux when the trough weakens ) .
the average flux measurements for both the _ iue _ and rebinned _ hst _ spectra are given in table 7 , which lists the heliocentric julian date of mid - exposure , orbital phase ( from table 3 ) , the mean rectified flux across the and wind lines , the telescope of origin , the x - ray state at the time of the observations , and a code for references discussing the contemporary x - ray fluxes .
note that we did not measure the transition in the _ iue _ spectra because these spectra are poorly exposed at the short wavelength end .
the _ iue _
average flux measurements have a typical error of @xmath43 based upon the scatter in the results from closely separated pairs of spectra .
the average fluxes across the wind lines are plotted as a function of orbital phase in figures 14 and 15 for and , respectively ( see a similar depiction in fig . 1 from @xcite ) .
different symbols show these measurements for the different x - ray states as observed with _
iue _ and _ hst_. we suspect that despite our efforts to rectify the _ iue _ and _ hst _ spectra in the same way , there are probably still some systematic differences since the mean fluxes for the _ hst _ spectra appear to be somewhat lower than those for the _ iue _ spectra .
nevertheless , the amplitude of line strength variation appears to be more or less the same in each of the _ iue _ low / hard state , _ iue _ high / soft state , and _ hst _ high / soft state spectra .
this result indicates that the hatchett - mccray effect ( and the amount of wind photoionization it represents ) occurs at about the same level in both x - ray states .
the x - ray accretion flux of cyg x-1 is fueled by mass transfer from the supergiant .
we argue in this section that the mass transfer process is dominated by a wind focused along the axis joining the stars .
however , the accretion of this gas by the black hole may be influenced by the strength of the radiatively driven shadow wind that is directed away from the black hole .
we begin by reviewing the most pertinent observational results from this investigation , and then we consider the interplay between the dynamics of the wind outflow and the x - ray accretion flux . first , the _ hst _ stis spectra of hd 226868 that we obtained at two epochs when the system was in the high / soft state show dramatic variations in the wind line strength that result from a superionization of the gas atoms illuminated by the x - ray flux .
shadow wind models , in which the wind ions only exist in the region where x - rays are blocked by the supergiant , make a reasonably good match to the observed profile variations , so we suspect that x - ray photoionization dominates much of the zone between the black hole and the facing hemisphere of the supergiant .
a similar degree of wind ionization probably also exists in the x - ray low / hard state since similar orbital variations in wind line strength are found in _
iue _ low dispersion spectra made during the x - ray low / hard state .
second , the _ hst _ spectra suggest that stellar wind gas emanating from parts of the photosphere facing the x - ray source attains only a small velocity before becoming photoionized .
for example , the highest optical depth wind feature , @xmath20 , shows only a very modest p cygni absorption core at phase @xmath40 ( see fig .
11 ) that extends blueward no more than about @xmath44 km s@xmath34 ( and it is possible that this small component results from a minor part of shadow wind projected against supergiant at @xmath40 ; see fig . 8) .
this very low wind speed is probably less than the stellar escape velocity ( @xmath45 km s@xmath34 near the poles ) .
these results from the uv wind lines indicate that very little mass loss is occurring by a radiatively driven wind for surface regions that are exposed to the x - ray source .
the fact that the wind features appear similarly weak in _ iue _ spectra obtained in the low / hard state suggests that a spherical , radiatively driven wind from the hemisphere of the supergiant facing the black hole is probably always weak or absent , and thus , accretion from a spherically symmetric wind must play a minor role in feeding the black hole in cyg x-1 . on the other hand
, we found that the emission equivalent width of the @xmath7 line is consistently strong between x - ray states and over the available record of observation .
the orbital phase variations of this spectral feature @xcite are successfully matched by models of emission from an enhanced density and slower gas outflow region between the stars that is expected for a focused stellar wind @xcite .
thus , while x - ray ionization reduces the wind outflow away from the axis joining the stars , the x - ray flux is apparently insufficient to stop the outflow in the denser gas of the focused wind ( the result of both tidal and radiative forces ) .
consequently , it is this focused wind component that is probably the primary means of mass transfer in cyg x-1 .
we caution that the relative constancy of the @xmath7 emission flux does not necessarily imply that the mass loss rate in the focused wind is also steady .
for example , the increased x - ray photoionization during the high / soft state may lead to an increase in @xmath7 emission ( see fig . 2 in @xcite ) , so that a lower mass loss rate but higher ionization fraction might result in the same amount of observed emission .
however , the presence of the @xmath7 emission in both x - ray states indicates that focused wind mass loss always occurs at some level .
finally , we confirm that the h@xmath0 p cygni line forms mainly in the base of the stellar wind of the supergiant since we observe that h@xmath0 follows the orbital velocity curve of the supergiant ( fig . 1 ) .
the new observations are consistent with earlier results @xcite in demonstrating that the h@xmath0 emission strength is generally weaker in the high / soft x - ray state .
photoionization and heating may extend down to atmospheric levels where the gas densities are sufficient to create h@xmath0 emission , so that the reduction in h@xmath0 strength in the x - ray high / soft state may partially result from photoionization related processes .
however , @xcite showed that h@xmath0 emission variability was present in those doppler shifted parts of the profile corresponding to the x - ray shadow hemisphere of the supergiant ( see their fig .
15 ) , so part of the h@xmath0 variations must be related to gas density variations at the base of the stellar wind . thus , the observed h@xmath0 variations suggest that the high / soft x - ray state occurs when the global , radiatively driven part of the wind is weaker . long term , quasi - cyclic variations in wind strength are apparently common among hot supergiants @xcite .
@xcite and @xcite suggested that the variations in x - ray state are caused by changes in wind velocity due to changes in supergiant mass loss rate . during times when the supergiant s wind is denser and the mass loss rate is higher , the photoionization region would be more restricted to the region closer to the black hole .
consequently , a radiatively driven wind could accelerate to a higher speed before stalling when the gas enters the ionization zone , and thus , the faster wind would result in a lower black hole accretion rate and x - ray luminosity ( creating the low / hard x - ray state ) .
conversely , if the wind mass loss rate drops , then the x - ray ionization zone will expand , the maximum wind velocity towards the black hole will decline , and the net accretion rate will increase ( perhaps creating the high / soft state ; @xcite ) . this creates a positive feedback mechanism that may continue until the wind is ionized all the way down to the stellar photosphere facing the x - ray source @xcite .
if this scenario is correct , then we expect that the outflow velocities in the direction of the black hole ( as measured in blue extent of the p cygni lines at phase @xmath40 ) will be larger than the supergiant escape velocity during the x - ray low / hard state . the superb quality _
hst_/stis spectra indicate that outflow velocities are too low to launch the wind during the x - ray high / soft state . moreover , the low resolution _ iue _
spectra from the low / hard state appear to show a very similar pattern of the loss of the p cygni absorption at @xmath40 ( fig . 12 and 13 ) , indicating that significant ionization zones still exist in the low / hard state .
taken at face value , these _ iue _ results suggest that the spherical component of wind outflow towards the black hole is weak and slow in both x - ray states , so a wind speed modulation is probably not the explanation for the accretion variations associated with the x - ray states .
we will require new , high quality , uv spectroscopy of cyg x-1 during the low / hard state in order to make a definitive test of this idea
. the radiatively driven wind of the supergiant leads to effective mass loss only in the x - ray shadowed hemisphere and in the focused wind between the stars in cyg x-1 .
the outflow in the shadow wind region will experience a coriolis deflection , so that the trailing regions of the shadow wind will eventually enter the zone of x - ray illumination @xcite .
once photoionized , this gas will stall with the loss of the important ions for radiative acceleration , and some of this slower gas may extend around the orbital plane to the vicinity of the black hole .
although this deflected wind gas is probably not a major accretion source @xcite , it may affect the accretion dynamics of the focused wind .
for example , when the shadow wind mass loss rate is high ( times of strong h@xmath0 emission ) , the resulting stalled wind component will create a higher ambient gas density on the leading side of the zone surrounding the black hole .
the focused wind flow will make a trajectory towards the following side of the black hole , and while gas passing closer to the black hole will merge into an accretion disk , gas further out will tend to move past the black hole before turning into the outskirts of the disk . the presence of the stalled gas on the leading side may deflect away this outer , lower density part of the flow and effectively inhibit gas accretion from the focused wind .
the subsequent reduction in gas accretion by the black hole may correspond to the conditions required to produce the low / hard x - ray state , while conversely a reduction in the stalled gas from the shadow wind may promote mass accretion and produce the high / soft state @xcite .
clearly , new hydrodynamical simulations are needed to test whether the stalled wind component is sufficient to alter the accretion of gas from the focused wind and create the environments needed for the x - ray transitions .
we thank the staffs of the david dunlap observatory , kitt peak national observatory , dominion astrophysical observatory , ondejov observatory , and the space telescope science institute ( stsci ) for their support in obtaining these observations .
the kpno spectra supporting the second _ hst _ run were obtained with the assistance of participants in the noao teacher leaders in research based science education program , including joan kadaras ( westford academy , westford , ma ) , steve harness ( kingsburg joint union high school , kingsburg , ca ) , elba sepulveda ( croem , mayaguez , pr ) , and dwight taylor ( goldenview middle school , anchorage , ak ) .
we also thank saku vrtilek and bram boroson for helpful comments , and we are especially grateful to an anonymous referee whose report was pivotal to our discussion of the results .
support for _ hst _ proposal number go-9840 was provided by nasa through a grant from the space telescope science institute , which is operated by the association of universities for research in astronomy , incorporated , under nasa contract nas5 - 26555 .
the x - ray results were provided by the asm / rxte teams at mit and at the rxte sof and gof at nasa s gsfc .
the _ iue _
data presented in this paper were obtained from the multimission archive at the space telescope science institute ( mast ) .
support for mast for non - hst data is provided by the nasa office of space science via grant nag5 - 7584 and by other grants and contracts .
bolton s research is partially supported by a natural sciences and engineering research council of canada ( nserc ) discovery grant .
hadrava s research is funded under grant projects gar 202/06/0041 and lc06014 .
herrero thanks the spanish mec for support under project aya 2007 - 67456-c02 - 01 .
this work was also supported by the national science foundation under grants ast-0205297 , ast-0506573 , and ast-0606861 .
institutional support has been provided from the gsu college of arts and sciences and from the research program enhancement fund of the board of regents of the university system of georgia , administered through the gsu office of the vice president for research .
we are grateful for all this support .
blondin , j. m. 1994 , , 435 , 756 blondin , j. m. , kallman , t. r. , fryxell , b. a. , & taam , r. e. 1990 , , 356 , 591 boroson , b. , kallman , t. , mccray , r. , vrtilek , s. d. , & raymond , j. 1999 , , 519 , 191 brocksopp , c. , fender , r. p. , larionov , v. , lyuty , v. m. , tarasov , a. e. , pooley , g. g. , paciesas , w. s. , & roche , p. 1999
, , 309 , 1063 brocksopp , c. , fender , r. p. , & pooley , g. g. 2002 , , 336 , 699 davis , r. , & hartmann , l. 1983 , , 270 , 671 day , c. s. r. , & stevens , i. r. 1993 , , 403 , 322 done , c. 2002 , roy .
london phil .
a , 360 , issue 1798 , 1967 friend , d. b. , & castor , j. i. 1982 , , 261 , 293 gallo , e. , fender , r. , kaiser , c. , russell , d. , morganti , r. , oosterloo , t. , & heinz , s. 2005 , , 436 , 819 gies , d. r. , & bolton , c. t. 1986a , , 304 , 371 gies , d. r. , & bolton , c. t. 1986b , , 304 , 389 gies , d. r. , et al .
2003 , , 583 , 424 groenewegen , m. a. t. , lamers , h. j. g. l. m. , & pauldrach , a. w. a. 1989 , , 221 , 78 hadrava , p. 2007
, proceedings of ragtime 8/9 : workshops on black holes and neutron stars , ed .
s. hledik & z. stuchlik ( opava , czech republic : silesian univ . ) , 73 ( arxiv:0710.0758v1 ) hatchett , s. , & mccray , r. 1977 , , 211 , 552 herrero , a. , kudritzki , r. p. , gabler , r. , vilchez , j. m. , & gabler , a. 1995 , , 297 , 556 ho , c. , & arons , j. 1987 , , 316 , 283 kaper , l. 1998 , in boulder - munich ii : properties of hot , luminous stars ( asp conf .
series , vol .
131 ) , ed .
i. d. howarth ( san francisco : asp ) , 427 karitskaya , e. a. , et al .
2006a , ibvs , 5678 , 1 karitskaya , e. a. , et al .
2006b , astron . &
trans . , 24 , 383 kemp , j. c. , barbour , m. s. , & mcbirney , r. e. 1981 , , 244 , l73 lamers , h. j. g. l. m. , cerruti - sola , m. , & perinotto , m. 1987 , , 314 , 726 lanz , t. , & hubeny , i. 2003 , , 146 , 417 levine , a. m. , bradt , h. , cui , w. , jernigan , j. g. , morgan , e. h. , remillard , r. , shirey , r. e. , & smith , d. a. 1996 , , 469 , l33 ling , j. c. , mahoney , w. a. , wheaton , w. a. , jacobson , a. s. , & kaluzienski , l. 1983 , , 275 , 307 markova , n. , puls , j. , scuderi , s. , & markov , h. 2005 , , 440 , 1133 mcclintock , j. e. , & remillard , r. a. 2006 , in compact stellar x - ray sources , ed . w. h. g. lewin & m. van der klis ( cambridge , uk : cambridge univ . press ) , 157 mcconnell , m. l. , et al .
2002 , , 572 , 984 miller , j. m. , wojdowski , p. , schulz , n. s. , marshall , h. l. , fabian , a. c. , remillard , r. a. , wijnands , r. , & lewin , w. h. g. 2005 , , 620 , 398 morbey , c. l. , & brosterhus , e. b. 1974 , , 86 , 455 ninkov , z. , walker , g. a. h. , & yang , s. 1987 , , 321 , 438 oda , m. 1980 , iau circ .
, 3502 , 2 ogawara , y. , mitsuda , k. , masai , k. , vallerga , j. v. , cominsky , l. r. , grunsfeld , j. m. , kruper , j. s. , & ricker , g. r. 1982 , nature , 295 , 675 penny , l. r. , gies , d. r. , & bagnuolo , w. g. , jr . 1999 , , 518 , 450 perotti , f. , et al . 1986 , , 300 , 297 priedhorsky , w. c. , terrell , j. , & holt , s. s. 1983 , , 270 , 233 puls , j. , et al .
1996 , , 305 , 171 shaposhnikov , n. , & titarchuk , l. 2006 , , 643 , 1098 stevens , i. r. 1991 , , 379 , 310 tarasov , a. e. , brocksopp , c. , & lyuty , v. m. 2003 , , 402 , 237 treves , a. , et al .
1980 , , 242 , 1114 van loon , j. th .
, kaper , l. , & hammerschlag - hensberge , g. 2001 , , 375 , 498 vrtilek , s. d. , boroson , b. s. , hunacek , a. , gies , d. , & bolton , c. t. 2008 , , in press walborn , n. r. 1973 , , 179 , l123 wen , l. , cui , w. , levine , a. m. , & bradt , h. v. 1999 , , 525 , 968 wilms , j. , nowak , m. a. , pottschmidt , k. , pooley , g. g. , & fritz , s. 2006 , , 447 , 245 cccccl 1&2419.8 2828.7 & 6510 6710 & 8900 & 70 & ddo/1.88m / cass . , 1800 g mm@xmath34/thomson @xmath46 + 2&2448.7 2453.9 & 6530 6710 & 7400 & 15 & kpno/0.9m / coud , b ( order 2)/ti5 + 3&2825.8 2828.9 & 6320 8970 & 2720 & 6 & kpno/0.9m / coud , rc400 ( order 1)/f3 kb + 4&2826.8 2827.9 & 6353 6756 & 6900 & 10 & dao/1.85m / cass . , 21121r / site-2 + 5&2730.7 2904.5 & 6258 6770 & 10900 & 24 & ondejov/2m / coud , 700mm / site @xmath47 + 6&2450.3 2453.3 & 1150 1740 & 14500 & 4 & hst/2.4m / stis , g140m / fuv - mama + 7&2825.7 2827.8 & 1150 1740 & 14500 & 4 & hst/2.4m / stis , g140m / fuv - mama + 8&2448.7 2454.0 & 4624 4740 & 13000 & 14 & kpno/0.9m / coud , b ( order 3)/ti5 + 9&2824.9 2826.8 & 3759 5086 & 2990 & 3 & kpno/0.9m / coud , rc400 ( order 2)/f3 kb + 10&2912.8 2915.8 & 4182 4942 & 5700 & 5 & kpno/4m / rc spec . , bl380 ( order 2)/t2 kb + lcccccc 2419.847 & 0.115 & 45.4 & 3.1 & 2.2 & @xmath480.38 & 1 + 2426.848 & 0.365 & 55.6 & 4.5 & 6.1 & @xmath480.54 & 1 + 2427.841 & 0.543 & @xmath4822.9 & 3.2 & 1.6 & @xmath480.75 & 1 + 2429.605 & 0.858 & @xmath4865.3 & 3.1 & @xmath483.3 & @xmath480.02 & 1 + 2443.803 & 0.393 & 42.4 & 3.2 & 2.1 & @xmath480.10 & 1 + 2445.846 & 0.758 & @xmath4888.6 & 3.1 & @xmath4810.6 & @xmath480.07 & 1 + 2446.804 & 0.929 & @xmath4842.0 & 3.1 & @xmath485.4 & @xmath480.10 & 1 + 2448.694 & 0.267 & 68.4 & 1.3 & 1.0 & 0.01 & 2 + 2448.850 & 0.294 & 67.5 & 3.1 & 2.5 & 0.10 & 1 + 2448.954 & 0.313 & 66.5 & 1.1 & 4.3 & @xmath480.22 & 2 + 2449.683 & 0.443 & 18.2 & 1.8 & @xmath482.2 & @xmath480.04 & 2 + 2449.790 & 0.462 & 7.7 & 3.3 & @xmath484.3 & @xmath480.15 & 2 + 2449.932 & 0.488 & @xmath488.9 & 1.7 & @xmath489.5 & @xmath480.48 & 2 + 2450.272 & 0.548 & @xmath4827.5 & 3.0 & @xmath480.6 & & 6 + 2450.335 & 0.560 & @xmath4831.6 & 3.0 & 0.2 & & 6 + 2450.637 & 0.614 & @xmath4845.7 & 3.2 & 7.2 & @xmath480.43 & 1 + 2450.667 & 0.619 & @xmath4849.2 & 1.8 & 5.5 & @xmath480.60 & 2 + 2450.715 & 0.628 & @xmath4860.3 & 3.5 & @xmath482.8 & @xmath480.56 & 1 + 2450.760 & 0.636 & @xmath4860.2 & 1.3 & @xmath480.2 & @xmath480.72 & 2 + 2450.792 & 0.641 & @xmath4853.6 & 3.1 & 8.1 & @xmath480.38 & 1 + 2450.812 & 0.645 & @xmath4866.3 & 3.1 & @xmath483.6 & @xmath480.83 & 1 + 2450.912 & 0.663 & @xmath4866.6 & 1.6 & 0.8 & @xmath480.74 & 2 + 2451.603 & 0.786 & @xmath4879.6 & 3.1 & @xmath483.4 & @xmath480.69 & 1 + 2451.642 & 0.793 & @xmath4882.9 & 3.1 & @xmath487.4 & @xmath480.33 & 1 + 2451.642 & 0.793 & @xmath4882.3 & 3.1 & @xmath486.9 & @xmath480.35 & 1 + 2451.686 & 0.801 & @xmath4876.8 & 3.2 & @xmath482.5 & @xmath480.54 & 1 + 2451.701 & 0.803 & @xmath4877.0 & 1.6 & @xmath483.0 & @xmath480.45 & 2 + 2451.709 & 0.805 & @xmath4879.7 & 3.2 & @xmath486.0 & @xmath480.73 & 1 + 2451.737 & 0.810 & @xmath4876.0 & 3.1 & @xmath483.1 & @xmath480.41 & 1 + 2452.711 & 0.984 & @xmath4812.9 & 3.1 & @xmath480.4 & 0.11 & 1 + 2452.733 & 0.988 & @xmath4810.9 & 3.1 & @xmath480.2 & @xmath480.04 & 1 + 2452.734 & 0.988 & @xmath489.1 & 1.6 & 1.4 & 0.04 & 2 + 2452.772 & 0.995 & @xmath485.5 & 3.2 & 2.0 & 0.13 & 1 + 2452.795 & 0.999 & @xmath485.3 & 3.2 & 0.3 & @xmath480.34 & 1 + 2452.836 & 0.006 & @xmath481.2 & 1.3 & 1.1 & 0.16 & 2 + 2452.962 & 0.029 & 12.4 & 1.6 & 4.4 & 0.10 & 2 + 2453.210 & 0.073 & 34.0 & 3.0 & 6.9 & & 6 + 2453.272 & 0.084 & 34.5 & 3.0 & 2.8 & & 6 + 2453.664 & 0.154 & 59.2 & 1.5 & 4.2 & @xmath480.21 & 2 + 2453.744 & 0.168 & 57.4 & 3.1 & @xmath481.1 & @xmath480.23 & 1 + 2453.754 & 0.170 & 56.7 & 1.2 & @xmath482.1 & @xmath480.22 & 2 + 2453.766 & 0.172 & 53.3 & 3.1 & @xmath486.0 & @xmath480.22 & 1 + 2453.806 & 0.180 & 64.5 & 3.2 & 3.7 & @xmath480.18 & 1 + 2453.828 & 0.183 & 59.8 & 3.1 & @xmath481.8 & @xmath480.13 & 1 + 2453.845 & 0.186 & 53.1 & 3.1 & @xmath489.0 & @xmath480.66 & 1 + 2453.934 & 0.202 & 59.9 & 1.3 & @xmath484.7 & @xmath480.09 & 2 + 2454.635 & 0.327 & 57.3 & 3.1 & @xmath482.1 & @xmath480.09 & 1 + 2454.660 & 0.332 & 52.6 & 3.1 & @xmath485.8 & 0.09 & 1 + 2454.682 & 0.336 & 53.4 & 3.1 & @xmath484.0 & 0.12 & 1 + 2454.705 & 0.340 & 49.5 & 3.1 & @xmath487.0 & @xmath480.07 & 1 + 2454.728 & 0.344 & 44.7 & 3.1 & @xmath4810.8 & @xmath480.05 & 1 + 2454.750 & 0.348 & 44.8 & 3.1 & @xmath489.6 & @xmath480.17 & 1 + 2454.773 & 0.352 & 43.5 & 3.1 & @xmath489.8 & @xmath480.22 & 1 + 2454.796 & 0.356 & 39.6 & 3.1 & @xmath4812.6 & @xmath480.07 & 1 + 2454.818 & 0.360 & 41.2 & 3.1 & @xmath489.9 & @xmath480.04 & 1 + 2454.841 & 0.364 & 43.6 & 3.1 & @xmath486.2 & @xmath480.03 & 1 + 2454.859 & 0.368 & 44.2 & 3.1 & @xmath484.6 & @xmath480.55 & 1 + 2463.626 & 0.933 & @xmath4843.2 & 3.2 & @xmath488.3 & @xmath480.15 & 1 + 2479.857 & 0.832 & @xmath4864.9 & 3.1 & 3.8 & @xmath480.13 & 1 + 2519.667 & 0.941 & @xmath4831.2 & 3.1 & 0.5 & @xmath480.30 & 1 + 2524.692 & 0.838 & @xmath4870.8 & 3.1 & @xmath483.6 & @xmath480.28 & 1 + 2530.586 & 0.891 & @xmath4847.5 & 3.1 & 3.9 & @xmath480.13 & 1 + 2554.578 & 0.175 & 63.8 & 3.1 & 3.9 & @xmath480.08 & 1 + 2571.581 & 0.211 & 65.1 & 3.1 & @xmath480.6 & @xmath480.77 & 1 + 2576.659 & 0.118 & 47.9 & 3.1 & 3.8 & @xmath480.29 & 1 + 2578.510 & 0.449 & 12.8 & 3.1 & @xmath485.1 & @xmath480.39 & 1 + 2580.562 & 0.815 & @xmath4870.5 & 3.1 & 1.6 & @xmath480.73 & 1 + 2592.565 & 0.959 & @xmath4829.0 & 3.1 & @xmath485.1 & @xmath481.05 & 1 + 2611.473 & 0.335 & 57.2 & 3.1 & @xmath480.5 & @xmath480.57 & 1 + 2613.498 & 0.697 & @xmath4878.6 & 3.1 & @xmath484.5 & @xmath480.83 & 1 + 2619.458 & 0.761 & @xmath4879.6 & 3.1 & @xmath481.7 & @xmath480.85 & 1 + 2625.476 & 0.836 & @xmath4875.7 & 3.3 & @xmath488.0 & @xmath480.90 & 1 + 2626.457 & 0.011 & @xmath484.5 & 3.1 & @xmath484.4 & @xmath480.57 & 1 + 2638.476 & 0.157 & 60.8 & 3.1 & 5.0 & @xmath480.60 & 1 + 2646.474 & 0.585 & @xmath4831.6 & 3.1 & 10.8 & @xmath481.05 & 1 + 2709.867 & 0.906 & @xmath4844.6 & 3.3 & 1.2 & @xmath480.70 & 1 + 2730.654 & 0.618 & @xmath4840.0 & 5.0 & 14.4 & @xmath481.21 & 5 + 2730.755 & 0.636 & @xmath4864.4 & 4.2 & @xmath484.2 & @xmath481.82 & 1 + 2744.818 & 0.148 & 56.7 & 3.1 & 3.4 & @xmath481.25 & 1 + 2746.612 & 0.468 & 14.6 & 3.0 & 5.2 & @xmath481.34 & 5 + 2762.871 & 0.371 & 48.3 & 3.1 & 0.7 & @xmath480.64 & 1 + 2805.784 & 0.035 & 2.3 & 3.1 & @xmath488.3 & @xmath480.35 & 1 + 2813.681 & 0.445 & 18.4 & 3.1 & @xmath481.2 & @xmath480.53 & 1 + 2824.647 & 0.403 & & & & @xmath480.29 & 1 + 2825.702 & 0.592 & @xmath4831.8 & 3.0 & 13.0 & & 7 + 2825.765 & 0.603 & @xmath4841.1 & 3.0 & 7.9 & & 7 + 2825.839 & 0.616 & @xmath4844.3 & 4.6 & 9.4 & @xmath480.40 & 3 + 2825.844 & 0.617 & @xmath4843.3 & 10.7 & 10.7 & @xmath480.36 & 3 + 2826.817 & 0.791 & @xmath4865.7 & 3.1 & 10.1 & @xmath480.34 & 1 + 2826.832 & 0.793 & @xmath4849.9 & 8.0 & 25.5 & @xmath480.44 & 4 + 2826.860 & 0.798 & @xmath4869.2 & 5.4 & 5.6 & @xmath480.29 & 4 + 2826.885 & 0.803 & @xmath4869.4 & 5.2 & 4.7 & @xmath480.36 & 4 + 2826.909 & 0.807 & @xmath4867.3 & 5.4 & 6.2 & @xmath480.37 & 4 + 2826.933 & 0.811 & @xmath4862.6 & 5.2 & 10.1 & @xmath480.43 & 4 + 2827.702 & 0.949 & @xmath4827.9 & 3.0 & 0.3 & & 7 + 2827.765 & 0.960 & @xmath4826.4 & 3.0 & @xmath483.0 & & 7 + 2827.800 & 0.966 & @xmath4823.7 & 6.8 & @xmath483.2 & @xmath480.44 & 3 + 2827.812 & 0.968 & @xmath4823.7 & 7.2 & @xmath484.1 & @xmath480.47 & 3 + 2827.812 & 0.968 & @xmath4823.0 & 5.2 & @xmath483.5 & @xmath480.60 & 4 + 2827.836 & 0.973 & @xmath4821.7 & 5.2 & @xmath484.1 & @xmath480.55 & 4 + 2827.882 & 0.981 & @xmath4813.5 & 5.3 & 0.4 & @xmath480.68 & 4 + 2827.905 & 0.985 & @xmath4814.0 & 5.2 & @xmath482.0 & @xmath480.60 & 4 + 2827.929 & 0.989 & @xmath487.3 & 5.2 & 2.8 & @xmath480.79 & 4 + 2828.594 & 0.108 & 40.7 & 3.1 & 0.1 & @xmath480.16 & 1 + 2828.657 & 0.119 & 45.3 & 3.1 & 0.7 & 0.08 & 1 + 2828.680 & 0.123 & 45.6 & 3.1 & @xmath480.3 & @xmath480.09 & 1 + 2828.702 & 0.127 & 46.4 & 3.1 & @xmath480.8 & @xmath480.04 & 1 + 2828.725 & 0.131 & 45.4 & 3.1 & @xmath483.0 & @xmath480.13 & 1 + 2828.747 & 0.135 & 50.0 & 3.1 & 0.3 & @xmath480.03 & 1 + 2828.885 & 0.160 & 57.5 & 5.5 & 1.1 & @xmath480.20 & 3 + 2828.892 & 0.161 & 55.5 & 4.6 & @xmath481.3 & @xmath480.17 & 3 + 2835.440 & 0.330 & 61.4 & 2.9 & 2.7 & 0.00 & 5 + 2840.510 & 0.236 & & & & @xmath480.09 & 5 + 2846.494 & 0.305 & 64.8 & 2.9 & 1.2 & @xmath480.53 & 5 + 2857.443 & 0.260 & 71.3 & 2.9 & 3.6 & @xmath480.09 & 5 + 2857.471 & 0.265 & 69.7 & 2.9 & 2.1 & @xmath480.24 & 5 + 2859.510 & 0.629 & @xmath4872.4 & 2.9 & @xmath4814.4 & @xmath480.52 & 5 + 2860.535 & 0.812 & @xmath4872.9 & 2.9 & @xmath480.2 & @xmath480.48 & 5 + 2860.558 & 0.816 & @xmath4871.8 & 2.9 & 0.1 & @xmath480.47 & 5 + 2861.539 & 0.991 & @xmath487.9 & 2.9 & 1.3 & @xmath480.31 & 5 + 2862.547 & 0.171 & 61.6 & 3.2 & 2.5 & @xmath480.35 & 5 + 2862.572 & 0.176 & 58.2 & 2.9 & @xmath481.8 & @xmath480.35 & 5 + 2874.472 & 0.301 & 72.1 & 2.9 & 7.9 & @xmath480.82 & 5 + 2874.494 & 0.305 & 67.6 & 2.9 & 4.1 & @xmath480.76 & 5 + 2874.516 & 0.309 & 72.6 & 2.9 & 9.6 & @xmath480.70 & 5 + 2874.538 & 0.312 & 70.5 & 2.9 & 8.2 & @xmath480.66 & 5 + 2878.417 & 0.005 & 5.0 & 2.9 & 7.7 & @xmath480.68 & 5 + 2878.441 & 0.009 & 8.2 & 2.9 & 9.0 & @xmath480.54 & 5 + 2898.466 & 0.585 & @xmath4836.2 & 2.9 & 6.2 & @xmath481.14 & 5 + 2901.357 & 0.102 & 38.4 & 3.1 & 0.0 & @xmath481.26 & 5 + 2903.436 & 0.473 & 22.2 & 2.9 & 14.9 & @xmath480.96 & 5 + 2903.475 & 0.480 & 16.9 & 2.9 & 12.9 & @xmath480.98 & 5 + 2904.480 & 0.659 & @xmath4855.5 & 2.9 & 11.1 & @xmath481.12 & 5 lccc @xmath49 ( d ) & 5.599829 ( 16 ) & 5.599829 & 5.599829 + @xmath11(ic ) ( hjd-2,400,000 ) & 41,874.707 ( 9 ) & 51,730.449 ( 8) & 52,872.788 ( 9 ) + @xmath50 ( km s@xmath34 ) & 74.9 ( 6 ) & 75.6 ( 7 ) & 73.0 ( 7 ) + @xmath51 ( km s@xmath34 ) & & @xmath487.0 ( 5 ) & @xmath485.1 ( 5 ) + @xmath52 ( km s@xmath34 ) & & 5.3 & 5.4 + lccccc 52448.748 & 0.276 & @xmath53 & 0.03 & 8 & high / soft + 52448.905 & 0.304 & @xmath54 & 0.02 & 8 & high / soft + 52449.733 & 0.452 & @xmath55 & 0.03 & 8 & high / soft + 52449.885 & 0.479 & @xmath56 & 0.02 & 8 & high / soft + 52449.969 & 0.494 & @xmath57 & 0.03 & 8 & high / soft + 52450.713 & 0.627 & @xmath58 & 0.03 & 8 & high / soft + 52450.855 & 0.653 & @xmath59 & 0.02 & 8 & high / soft + 52450.958 & 0.671 & @xmath60 & 0.02 & 8 & high / soft + 52451.732 & 0.809 & @xmath61 & 0.07 & 8 & high / soft + 52452.796 & 0.999 & @xmath62 & 0.08 & 8 & high / soft + 52452.930 & 0.023 & @xmath63 & 0.02 & 8 & high / soft + 52453.709 & 0.162 & @xmath60 & 0.05 & 8 & high / soft + 52453.889 & 0.194 & @xmath62 & 0.03 & 8 & high / soft + 52453.965 & 0.208 & @xmath64 & 0.03 & 8 & high / soft + 52824.884 & 0.445 & @xmath65 & 0.06 & 9 & high / soft + 52824.891 & 0.447 & @xmath66 & 0.06 & 9 & high / soft + 52826.841 & 0.795 & @xmath67 & 0.04 & 9 & high / soft + 52912.769 & 0.140 & @xmath68 & 0.02 & 10 & low / hard + 52912.773 & 0.140 & @xmath60 & 0.02 & 10 & low / hard + 52913.768 & 0.318 & @xmath68 & 0.02 & 10 & low / hard + 52914.766 & 0.496 & @xmath66 & 0.02 & 10 & low / hard + 52915.761 & 0.674 & @xmath69 & 0.02 & 10 & low / hard + lccccc 19711981 & @xmath70 & 0.12 & @xcite + 19801984 & @xmath71 & 0.09 & @xcite + 19961998 & @xmath67 & 0.12 & @xcite + 20022003 & @xmath53 & 0.12 & this paper , high / soft state + 2003 & @xmath53 & 0.11 & this paper , low / hard state + lcccccl 43629.222 & 0.313 & 1.079 & 0.622 & iue & low & 1,2,3 + 43629.917 & 0.437 & 1.109 & 0.725 & iue & low & 1,2,3 + 43630.898 & 0.612 & 1.094 & 0.691 & iue & low & 1,2,3 + 43632.884 & 0.967 & 0.854 & 0.632 & iue & low & 1,2,3 + 43636.069 & 0.536 & 1.099 & 0.756 & iue & low & 1,2,3 + 43638.907 & 0.042 & 0.898 & 0.667 & iue & low & 1,2,3 + 43701.758 & 0.266 & 0.965 & 0.728 & iue & low & 1,2,3 + 43708.857 & 0.534 & 1.069 & 0.765 & iue & low & 1,2,3 + 43799.446 & 0.711 & 0.963 & 0.733 & iue & low & 1,2,3 + 43802.485 & 0.254 & 0.901 & 0.649 & iue & low & 1,2,3 + 43804.629 & 0.636 & 0.995 & 0.789 & iue & low & 1,2,3 + 43845.978 & 0.020 & 0.753 & 0.588 & iue & low & 1,2,3 + 43848.020 & 0.385 & 0.993 & 0.735 & iue & low & 1,2,3 + 43892.220 & 0.278 & 1.004 & 0.751 & iue & low & 1,2,3 + 43894.744 & 0.729 & 0.932 & 0.678 & iue & low & 1,2,3 + 43894.811 & 0.741 & 0.886 & 0.671 & iue & low & 1,2,3 + 43896.419 & 0.028 & 0.795 & 0.648 & iue & low & 1,2,3 + 44002.919 & 0.046 & 0.814 & 0.610 & iue & low & 1,2,3 + 44003.118 & 0.082 & 0.843 & 0.654 & iue & low & 1,2,3 + 44003.259 & 0.107 & 0.843 & 0.533 & iue & low & 1,2,3 + 44035.011 & 0.777 & 0.887 & 0.696 & iue & low & 1,2,3 + 44265.582 & 0.952 & 0.739 & 0.548 & iue & low & 1,4 + 44412.390 & 0.168 & 0.929 & 0.683 & iue & high & 5 + 44416.117 & 0.834 & 0.817 & 0.577 & iue & high & 5 + 44419.259 & 0.395 & 0.924 & 0.764 & iue & high & 5 + 44419.797 & 0.491 & 1.009 & 0.805 & iue & high & 5 + 44422.310 & 0.940 & 0.840 & 0.678 & iue & high & 5,6 + 44423.302 & 0.117 & 0.790 & 0.584 & iue & high & 5,6 + 44425.301 & 0.474 & 0.993 & 0.746 & iue & high & 5,6 + 44426.941 & 0.767 & 0.962 & 0.757 & iue & high & 5,6 + 52450.272 & 0.548 & 0.947 & 0.714 & hst & high & 7 + 52450.335 & 0.560 & 0.947 & 0.722 & hst & high & 7 + 52453.210 & 0.073 & 0.752 & 0.515 & hst & high & 7 + 52453.272 & 0.084 & 0.761 & 0.520 & hst & high & 7 + 52825.702 & 0.592 & 1.030 & 0.743 & hst & high & 7 + 52825.765 & 0.603 & 1.014 & 0.725 & hst & high & 7 + 52827.702 & 0.949 & 0.804 & 0.551 & hst & high & 7 + 52827.765 & 0.960 & 0.787 & 0.530 & hst & high & 7 + emission strength ( _ above _ ) and daily average soft x - ray flux ( _ below _ ) .
the two arrows in the upper panel indicate the times of the stis observations that occurred during the x - ray high / soft state .
the recent , more densely sampled observations show clearly how the h@xmath0 emission increases as the soft x - ray flux declines.,title="fig:",height=453 ] emission strength ( _ above _ ) and x - ray flux for each dwell observation ( _ below _ ) for the week surrounding the first _
hst _ run in 2002 .
the arrows in the top panel show the times of the stis observations for each orbital conjunction phase .
the symbols in the lower panel indicate the count rates in the 1.5 3 kev ( @xmath15 ) and 5 12 kev ( @xmath16 ) bands.,title="fig:",height=453 ] wind feature in both the 2002 ( _ top _ ) and 2003 ( _ bottom _ ) runs .
the spectra are plotted as a function of doppler shift for the blue component of the doublet in the rest frame of the supergiant ( which leads to small offsets in the positions of the narrow , interstellar lines that are stationary in the absolute frame ) .
vertical line segments indicate the rest wavelength positions of both components of the doublet.,title="fig:",height=453 ] assumed ) .
the dotted lines show the part of the line of sight that is projected onto the disk of the star at the two orbital phases for an observer at the correct direction with respect to the orbital plane .
the shadow wind region projected against the star in each case is the area bounded by the surface of the star and the dashed and dotted line for that phase .
cm marks the center of mass position.,title="fig:",height=453 ] wind feature ( plotted as a function of doppler shift for the blue component of the doublet in the rest frame of the supergiant ) .
the top panel shows the profiles for case where the black hole is in the background while the bottom profile illustrates the other conjunction case where the black hole is in the foreground ( photoionizing the wind gas seen projected against the supergiant ) .
the dotted line shows the predicted photospheric line spectrum ( offset for clarity ) from the models of @xcite .
note that the latter representation does not include the broad ly@xmath0 wings that depress the continuum in the observed spectrum towards the left hand side of the diagram.,title="fig:",height=453 ] as a function of radial velocity for the blue member of the doublet .
the spectrum of hd 226868 is plotted at the top as a thick line , and below are average spectra for other o9.7 iab supergiants ( offset in intensity for clarity ) . from top to bottom , these are hd 75222 , hd 152003 , hd 149038 , and hd 167264.,title="fig:",height=453 ] | we present results from _ hubble space telescope _ ultraviolet spectroscopy of the massive x - ray and black hole binary system , hd 226868 = cyg x-1 .
the spectra were obtained at both orbital conjunction phases in two separate runs in 2002 and 2003 when the system was in the x - ray high / soft state .
the uv stellar wind lines suffer large reductions in absorption strength when the black hole is in the foreground due to the x - ray ionization of the wind ions .
a comparison of the _ hst _ spectra with archival , low resolution spectra from the _ international ultraviolet explorer satellite _ shows that similar photoionization effects occur in both the x - ray high / soft and low / hard states .
we constructed model uv wind line profiles assuming that x - ray ionization occurs everywhere in the wind except the zone where the supergiant blocks the x - ray flux .
the good match between the observed and model profiles indicates that the wind ionization extends to near to the hemisphere of the supergiant facing the x - ray source .
we also present contemporaneous spectroscopy of the h@xmath0 emission that forms in the high density gas at the base of the supergiant s wind and the @xmath1 emission that originates in the dense , focused wind gas between the stars .
the h@xmath0 emission strength is generally lower in the high / soft state compared to the low / hard state , but the @xmath1 emission is relatively constant between x - ray states .
the results suggest that mass transfer in cyg x-1 is dominated by the focused wind flow that peaks along the axis joining the stars and that the stellar wind contribution from the remainder of the hemisphere facing the x - ray source is shut down by x - ray photoionization effects ( in both x - ray states ) .
the strong stellar wind from the shadowed side of the supergiant will stall when coriolis deflection brings the gas into the region of x - ray illumination .
this stalled gas component may be overtaken by the orbital motion of the black hole and act to inhibit accretion from the focused wind .
the variations in the strength of the shadow wind component may then lead to accretion rate changes that ultimately determine the x - ray state . |
cobalt compounds are of particular interest due to the possibility that co ions can exhibit different spin states and hence the occurrence of temperature - driven spin - state transitions .
the most prominent example is lacoo@xmath1 , which has been intensively studied and controversially debated for more than fifty years , see e.g.@xcite .
the co@xmath4 ions in lacoo@xmath1 feature a @xmath5 configuration which in principle can occur in three different spin states : a nonmagnetic low - spin ( ls ) ( @xmath6 , @xmath7 ) , an intermediate - spin ( is ) ( @xmath8 , @xmath9 ) and a high - spin ( hs ) state ( @xmath10 , @xmath11 ) .
it is generally agreed that the co@xmath4 ions in lacoo@xmath1 realize the ls state at low temperatures . above
approximately 25k a higher spin state , either is or hs , becomes thermally populated affecting various physical properties , e.g. the magnetic susceptibility @xmath12 or the thermal expansion @xmath13 , which both exhibit pronounced maxima in their temperature dependencies @xcite . the susceptibility is obviously affected because the excited spin state , either is or hs , induces a strong increase of the magnetization above 25k .
the thermal expansion is affected due to the different ionic radii of the smaller ls - co@xmath4 with empty and the larger is- or hs - co@xmath4 ions with partially filled @xmath14 orbitals .
the spin - state transition can be well described in a ls is scenario , i.e. the excited spin state is the is state , with a constant energy gap of @xmath15k @xcite .
however , more recent investigations show , that a ls / hs model including spin - orbit coupling is more reasonable @xcite .
a consequence of the latter model is a temperature - dependent energy gap @xmath16(t ) , which strongly increases with increasing temperature @xcite .
the spin - state transition in lacoo@xmath1 is strongly affected by both , heterovalent and isovalent doping on the la site .
the former possibility causes hole doping and chemical pressure and suppresses the spin - state transition due to the implementation of co@xmath17 ions @xcite . the latter one , i.e. chemical pressure without changing the co valence , is usually realized by introducing trivalent rare - earth ions @xmath18 . in @xmath0coo@xmath1
the spin - state transition is not suppressed but its onset is shifted to higher temperature .
the energy gap between the ls and the excited spin state increases from about @xmath15k for @xmath2 la to @xmath19k for @xmath2 eu @xcite .
moreover , due to the decreasing ionic radius of the lanthanide series , the structure changes from rhombohedral in lacoo@xmath1 to orthorhombic for @xmath20 , nd and eu . recently , it has been reported , that the low - temperature thermal conductivity of lacoo@xmath1 is also very anomalous @xcite .
this behaviour has been qualitatively attributed to the onset of the spin - state transition . however , a quantitative analysis of the anomalous thermal conductivity of lacoo@xmath1 has not been presented yet .
the aim of this paper is to develop a consistent picture of the influence of the spin - state transition on the thermal conductivity .
therefore , we measured the thermal conductivity @xmath21 of the series @xmath0coo@xmath1 with @xmath2 la , pr , nd and eu .
moreover , we studied the thermal expansion @xmath22 , which is a very sensitive probe to investigate spin - state and also crystal - field transitions and their coupling to the lattice . in the quantitative analysis of our data
, we will consider both models of the spin - state transition , which are favoured for lacoo@xmath1 in the literature , i.e. we will consider the ls is scenario with a constant @xmath16 and the more recently proposed spin - orbit coupled hs ( sochs ) model with a temperature dependent energy gap @xmath23 .
lrrrrcr sample & @xmath24 ( k ) & @xmath16(k ) & @xmath25 ( % ) & @xmath26(@xmath27v / k ) & @xmath28(mm@xmath29 ) & @xmath30(mm ) + lacoo@xmath1 ( s1 ) & 25 & 185 & 0.7 & -700 & @xmath31 & 3.7 + lacoo@xmath1 ( s2 ) & & & & 1000 & @xmath32 & 2.3 + lacoo@xmath1 ( s3 ) & & & & -600 & @xmath33 & 3.4 + lacoo@xmath1 ( s4 ) & & & & 1000 & @xmath34 & 2.6 + lacoo@xmath1 ( s5 ) & & & & -300 & @xmath35 & 4.7 + prcoo@xmath1 & 175 & 1200 & 2.8 & -400 & @xmath36 & 2.5 + ndcoo@xmath1 & 230 & 1700 & 4.8 & -400 & @xmath37 & 2.8 + eucoo@xmath1 & 400 & @xmath19 & @xmath38 & -500 & @xmath39 & 2.0 + all @xmath0coo@xmath1 crystals have been grown in a floating - zone image furnace .
we examine five different lacoo@xmath1 crystals identical to those used in @xcite , where details of the sample preparation and characterization are given . for eucoo@xmath1
this information can be found in @xcite .
the ndcoo@xmath1 and prcoo@xmath1single crystals have been grown in the same way as those of lacoo@xmath1and eucoo@xmath1 .
characteristic properties of all crystals are listed in .
the thermal conductivity measurements have been performed by a standard steady - state method using a differential chromel - au+0.07@xmath40fe - thermocouple @xcite .
the thermal expansion below @xmath41k was measured using a home - built high - resolution capacitance dilatometer @xcite , whereas the high - temperature measurements @xmath42k were performed using commercial inductive dilatometer ( tma 7 , perkin - elmer ) .
coo@xmath1 with @xmath2 la , pr , nd and eu . the solid line is the estimated phonon contribution of ndcoo@xmath1(see text ) .
( b ) anomalous part of the thermal expansion obtained by subtracting the phononic background . for clarity
the different data sets are offset by @xmath43/k with respect to each other .
the arrows signal the approximate onset temperature of the spin - state transition and the solid circles denote the metal - insulator transition temperature estimated from resistivity measurements for each compound .
the solid lines are fits of the respective schottky anomalies , see equations([schott ] ) and ( [ schott2 ] ) , assuming a constant energy gap @xmath16 between the different spin states of co@xmath4 .
the dashed line for lacoo@xmath1 is a similar fit using @xmath23 obtained from the magnetic susceptibility ( see text ) . ] in ( a ) we show the thermal - expansion coefficients @xmath44 of @xmath0coo@xmath1 .
the low - temperature results of lacoo@xmath1 and eucoo@xmath1 were already discussed in detail in @xcite .
the thermal expansion consists of a phononic part and a schottky contribution , caused by the thermal population of the @xmath45 orbitals of the co@xmath4 ions , i.e. @xmath46 .
the latter contribution causes the large maximum at @xmath47k in lacoo@xmath1 . to further analyze the data we subtract @xmath48 , which we estimate using a debye function with the debye temperature @xmath49k of lacoo@xmath1 @xcite
here , we used the same @xmath50 for the various cobaltates , but sample - dependent prefactors determined by scaling the debye function to the low - temperature data of each compound .
as an example , we show @xmath48 for ndcoo@xmath1 in ( a ) ; @xmath48 of eucoo@xmath1 ( prcoo@xmath1 ; not shown ) is slightly larger ( smaller ) . since
a clear separation of @xmath48 and @xmath51 is not possible for the low - temperature data of lacoo@xmath1 , we used @xmath48 of eucoo@xmath1also for lacoo@xmath1 .
the resulting @xmath51 of all crystals are shown in ( b ) .
we note that , in particular for prcoo@xmath1 and ndcoo@xmath1 the spin - state transition is seen much better in the thermal - expansion data than in the magnetic susceptibility .
the reason is that @xmath48 is rather small compared to the total thermal expansion , whereas @xmath12 is dominated by the large contribution of the @xmath3 moments of the pr@xmath4 and nd@xmath4 ions , respectively , which makes a further analysis rather uncertain .
the insulator - metal transitions occurring above about 500k , see e.g. @xcite also cause anomalies in @xmath22 .
the solid circles in ( b ) signal the transition temperatures @xmath52 , which have been determined from resistivity measurements on our crystals ( not shown ) . within the ls /
is scenario with a constant @xmath16 , @xmath53 is given by @xcite @xmath54 where @xmath55 is the degeneracy of the excited spin state and @xmath56 is a measure of the ionic radii difference of the co@xmath4 in the ls and the excited spin state .
moreover , @xmath25 is related to the uniaxial pressure dependence of @xmath57 via @xmath58 , where @xmath59 is boltzmann s constant , @xmath60@xmath61 the volume per formula unit and @xmath62 means uniaxial pressure along the direction of which @xmath13 is measured @xcite . to get the hydrostatic pressure dependence one has to use the volume expansion which in the case of the twinned @xmath0coo@xmath1 single crystals is given by @xmath63
. the corresponding fit of @xmath13 of lacoo@xmath1 ( already presented in @xcite ) is shown by the solid line in ( b ) .
the fit parameters are @xmath15k and @xmath64 giving a hydrostatic pressure dependence @xmath65 , in agreement with the increase of @xmath16 of @xmath66% obtained from measurements of @xmath12 under @xmath67gpa @xcite
. for the sochs model one has to consider the temperature dependence of @xmath16 @xcite , which can be calculated from the measured @xmath68 @xcite .
for lacoo@xmath1 , this yields a linear increase @xmath69 with @xmath70k and @xmath71 in the temperature range almost up to the mi transition . using this @xmath72 in the partition sum , equation([schott ] ) modifies to @xmath73 the corresponding fit of @xmath51 with @xmath74 as the only free parameter gives the dashed line in ( b ) .
obviously , both fits hardly differ because the modified @xmath75 is compensated by a larger value of @xmath76 .
we note , however , that this does _ not _ necessarily correspond to a larger pressure dependence of @xmath77 , because @xmath51 of equation([schott2 ] ) is determined by the pressure dependencies of @xmath77 and that of the slope @xmath78 ; see e.g.the discussions in @xcite . in view of the small differences between both fits for lacoo@xmath1 , and since there are no indications for a temperature - dependent @xmath16from other physical quantities , the fits of @xmath51 of the other @xmath0coo@xmath1 crystals have been done for constant energy gaps only . since equations([schott ] ) and ( [ schott2 ] ) are identical for @xmath79 , this analysis is not able to distinguish between both models .
the obtained values of @xmath16 and @xmath25 are given in .
as expected , @xmath16 strongly increases with decreasing radius of the @xmath18 ions and therefore the spin - state transition monotonically shifts to higher temperature . the onset temperature @xmath24 of the spin - state transition for each compound , @xmath80 , 175 , 230 and 400k for @xmath2 la , pr , nd and eu , respectively , are marked by arrows in ( b ) .
coo@xmath1 with @xmath81la , pr , nd and eu ; the solid line is a fit via equation([kphonon ] ) .
( b ) thermal conductivity of five different lacoo@xmath1 single crystals s1s5 ( note the different scale ) . the inset shows the maximum values of the thermal conductivity @xmath82 of these crystals vs. the saturation values of the magnetization @xmath83 measured at 2k .
as shown by the sold line , we find an approximate proportionality @xmath84 ( see text ) . ]
\(a ) displays the thermal - conductivity data of @xmath0coo@xmath1 .
as has been already found in previous studies on lacoo@xmath1@xcite , the overall shape and qualitative behaviour of @xmath21 is rather unusual .
first of all , @xmath85 is rather low in the whole temperature range and its temperature dependence clearly deviates from the typical behaviour expected for a conventional phononic heat conductor .
the thermal conductivity of lacoo@xmath1 exhibits a maximum around 2030k . towards higher temperatures
@xmath85 rapidly drops and features a minimum around 150k instead of the expected @xmath86-decrease ( see also @xcite ) .
qualitatively , this minimum can be traced back to the spin - state transition of lacoo@xmath1 , which sets in close to the maximum of @xmath85 .
the thermal population of the co @xmath45 levels induces a certain fraction of co@xmath4 ions with a larger ionic radius compared to the ls co@xmath4 ions .
these randomly distributed larger co@xmath4 ions do not only cause the huge anomaly in @xmath22 , but also lead to additional disorder in the lattice and therefore act as additional scattering centers for phonons .
hence , the thermal conductivity is additionally suppressed .
a quantitative description will be given below .
since the spin - state transition shifts to higher temperature when la@xmath4 is replaced by smaller @xmath18 ions , the above scenario suggests that the onset of the suppression of @xmath21 of @xmath0coo@xmath1systematically shifts towards higher temperature with decreasing radius of the @xmath18 ion .
let us first consider the data of eucoo@xmath1 , which is the compound with the smallest @xmath18 ion and the largest @xmath87k @xcite .
hence , the spin - state transition should not affect the thermal conductivity below room temperature and , indeed , the thermal conductivity of eucoo@xmath1 exhibits much larger values than in lacoo@xmath1 .
its low - temperature maximum exceeds that of lacoo@xmath1 by about one order of magnitude .
moreover , the temperature dependence of @xmath85 of eucoo@xmath1 meets the expectation of a purely phononic heat conductor . the thermal conductivity of ndcoo@xmath1 is lower than @xmath85 of eucoo@xmath1 in the entire temperature range and , at a first glance , its overall temperature dependence seems to be consistent with a conventional phononic picture , too . a closer inspection of the data yields , however , a slight slope change towards a somewhat steeper decrease of @xmath21 at @xmath88k , i.e. close to the onset temperature of the spin - state transition , see ( b ) .
this slope change supports the above proposed explanation that the spin - state transition causes additional scattering centers .
prcoo@xmath1 also exhibits an unusual temperature dependence of @xmath85 , which appears somewhat similar to that of lacoo@xmath1 . above the low - temperature maximum
, @xmath21 rapidly decreases to a small value of @xmath89w / km and remains almost temperature independent between about 80k and 175k . above this temperature a further decrease of @xmath85 sets in . since this decrease starts again at the onset of the spin - state transition ( see ( b ) ) , it is natural to attribute it to the additional scattering due to the presence of co@xmath4 ions in different spin states .
in contrast , the anomalous temperature dependence of @xmath21 at lower temperature is not related to the spin - state transition and , as will be discussed separately in , this anomalous feature is related to the crystal - field splitting of the @xmath3 shell of pr@xmath4 . summarizing the presented data
so far , we observe in all crystals with @xmath90k that this onset coincides with an additional suppression of @xmath85 .
since @xmath16 strongly increases with decreasing @xmath18 radius @xcite , @xmath24 also shifts towards higher temperature and the additional suppression of the thermal conductivity is less obvious , because at higher temperature the usual suppression of @xmath85 due to phonon - phonon umklapp scattering becomes more and more dominant .
however , the data in ( a ) also show , that @xmath85 is not only suppressed at temperatures _ above _
@xmath24 . instead , with increasing size of @xmath18 a monotonic decrease of @xmath85 in the _ entire _ temperature range is found ( here we have neglected the additional suppression of @xmath91 in prcoo@xmath1 ) .
if the spin - state transition would be the only effect causing additional disorder scattering , @xmath21 should approach the bare phononic value below @xmath24 .
taking eucoo@xmath1 with stable ls co@xmath4 ions up to above room temperature as a rough measure for this bare phononic heat conduction , it becomes clear that this expectation is not observed in our data .
in general , the magnitude of the low - temperature maximum of @xmath85 depends on the sample quality , which determines the maximum free path .
thus , the different values within the @xmath0coo@xmath1series could arise from different sample qualities . however , there is no indication for a systematic increase of the crystal quality with decreasing @xmath18 ion radius
. is also influenced by twinning .
all these single crystals are twinned , but it seems plausible that the average size of the twin domains increases from lacoo@xmath92 to eucoo@xmath92 , because the deviation from the cubic symmetry increases .
however , the observed correlation between @xmath82 and the magnetic impurity concentration for the different lacoo@xmath92 crystals , makes it rather unlikely to explain the different low - temperature maxima of @xmath85 in terms of different domain sizes . ] in order to study a possible influence of the crystal quality , we measured @xmath85 of five different lacoo@xmath1 single crystals .
as shown in ( b ) the high - temperature behaviour is essentially the same for all crystals , but the low - temperature maxima differ considerably .
nevertheless , all of them remain much lower than that of eucoo@xmath1 .
we relate the low - temperature maxima of @xmath85 to the defect concentrations by considering the low - temperature magnetization , since a really pure lacoo@xmath1 with all co@xmath4 ions in the ls state would be non - magnetic ( apart from core dia- and van vleck paramagnetism ) . at @xmath93k
we measured magnetization up to 14 tesla in order to reach its saturation @xmath94 , which is a measure of the content of magnetic impurities .
the inset of displays the low - temperature maxima of @xmath21 vs. @xmath94 , which indeed roughly follow a @xmath95 dependence as shown by the solid line .
the most likely source for magnetic impurities in lacoo@xmath1 is a weak oxygen nonstoichiometry , which changes the average valence from co@xmath4 towards co@xmath96 or co@xmath17 . due to the odd number of @xmath97 electrons , the latter are magnetic in all possible electronic configurations , but in most cases co@xmath96 and co@xmath17 realize @xmath98 and @xmath99 states with spin @xmath100 and @xmath101 , respectively .
experimental evidence for a weak charge - carrier doping in nominally pure lacoo@xmath1 stems from the large values of different signs of the room - temperature thermopower for the different lacoo@xmath1 crystals ; cf . and the discussion in @xcite . from magnetization data on sr - doped lacoo@xmath1it
is known that low doping concentrations induce so - called high - spin polarons with effective spin values up to @xmath102 @xcite .
such polarons are formed around the doping - induced co@xmath17 ions , which cause a spin - state transition in the adjacent co@xmath4 ions .
most probably , this is a transition from the ls to an is state , because the electronic configuration @xmath103 of co@xmath4 is favourable for the hopping of the @xmath45 electron for both kinds of neighbours , co@xmath17 with @xmath99 and co@xmath96 with @xmath98 .
other possible configurations often only allow a @xmath104 hopping or may cause the so - called spin - blockade effect @xcite .
the formation of high - spin polarons will depend on @xmath16 between the ls ground state and the excited spin state of co@xmath4 . since @xmath16 strongly increases with decreasing @xmath18 radius , the occurrence of such polarons in the @xmath0coo@xmath1 series should get more and more unlikely for smaller @xmath18 ions .
unfortunately , the presence of high - spin polarons in prcoo@xmath1 and ndcoo@xmath1 can not be studied by magnetization measurements because of the strong magnetism of the partially filled @xmath3 shells of pr@xmath4 and nd@xmath4 . for eucoo@xmath1 such
an analysis is again easier , because eu@xmath4 only exhibits van vleck paramagnetism and , in fact , the impurity contribution to the magnetic susceptibility in eucoo@xmath1 is one order of magnitude smaller than in lacoo@xmath1 @xcite .
moreover , the magnetic impurity contribution of the la@xmath105eu@xmath106coo@xmath1 series systematically decreases with increasing @xmath107 , i.e. with decreasing average ionic size on the la site and with increasing @xmath16 @xcite .
these observations strongly support the above idea that the systematic dependence of the low - temperature thermal conductivity of @xmath0coo@xmath1 on the ionic size of @xmath18 is related to phonon scattering by high - spin polarons , whose formation depend on a weak oxygen - nonstoichiometry and on @xmath16 .
for a quantitative analysis of the thermal conductivity we use an extended debye model @xcite @xmath108 here , @xmath50 denotes the debye temperature , @xmath109 the sound velocity , @xmath110 the phonon frequency , @xmath111 and @xmath112 the phonon relaxation time .
the scattering rates of different independent scattering mechanisms sum up to a total scattering rate @xmath113 the three terms on the right - hand side refer to the typical scattering rates for phonon heat transport in insulators , namely boundary ( @xmath114 ; @xmath115 mm is the characteristic length scale of the sample @xcite and was kept fixed for all crystals ) , point defect ( @xmath116 ) and phonon - phonon umklapp scattering ( @xmath117 ) . since
the lattice spacing is a natural lower limit for the mean - free path , @xmath21 is expected to approach a minimum value at higher temperatures . in order to model this , @xmath112 in equation([rates ] ) is replaced by @xmath118 } with the minimum mean - free path @xmath119 being of the order of the lattice constant ; compare the discussion in @xcite .
as shown by the solid line in , the thermal conductivity data of eucoo@xmath1 can be reproduced reasonably well by equation([kphonon ] ) . in order to reduce the number of adjustable parameters
, we used @xmath49k and @xmath120 m/s determined for lacoo@xmath1 @xcite .
the other parameters have been adjusted to fit the experimental data .
we obtain @xmath121s@xmath61 , @xmath122s@xmath29/k and @xmath123 ; @xmath119 is not reached in eucoo@xmath1 up to 300k . in the following analysis we consider the thermal conductivity of eucoo@xmath1 as the bare phononic heat conductivity that is not influenced by the spin - state transition . for the other members of the @xmath0coo@xmath1 series we introduce an additional scattering mechanism arising from the thermal occupation of the co@xmath4 @xmath14 shells , which causes additional lattice disorder due to the presence of smaller and larger co@xmath4 ions in the ls and the excited spin state , respectively . considering the co@xmath4 ions of different sizes as point - like defects , we assume an @xmath124 frequency dependence for this additional scattering rate ( cf.equation([rates ] ) ) @xmath125 where @xmath126 is an adjustable parameter describing the scattering strength and @xmath127 models the temperature dependence of the additional lattice disorder . the unknown functions @xmath128 for lacoo@xmath1 , prcoo@xmath1 and ndcoo@xmath1 can be calculated from the experimental data in the following way : we use equation([rates ] ) with all the parameters obtained from the fit of @xmath85 of eucoo@xmath1 kept fix and add @xmath129 . then the total @xmath130 is used in equation([kphonon ] ) and the values of @xmath128 at each temperature are determined point by point from the measured @xmath21 .
the resulting @xmath128 curves for all three @xmath0coo@xmath1 samples are shown by the open symbols in . in order to avoid additional complications arising from the aforementioned high - spin polarons ,
we restrict the analysis to the temperature range above @xmath131k .
for lacoo@xmath1 the resulting @xmath128 has a pronounced maximum around 100k , and strongly resembles the temperature dependence of the magnetic susceptibility of lacoo@xmath1 . here , we have used sample s2 which exhibits the highest value of the low - temperature maximum of @xmath85 , i.e. the sample which is least affected by impurity scattering .
we note , however , that in the considered temperature range above 25k the resulting @xmath128 is not very different for the other lacoo@xmath1 crystals . for prcoo@xmath1 ,
@xmath128 features a clear maximum followed by a slope change around 175k .
the obtained values of @xmath128 for ndcoo@xmath1 are much lower than those of lacoo@xmath1 and prcoo@xmath1 and there is no maximum , but again a clear slope change around @xmath132k .
thus , we find in all three @xmath0coo@xmath1 compounds either an increase of @xmath128 or a slope change towards larger values at temperatures close to the respective onset temperatures of the spin - state transition ; see . in order to model the additional scattering rate we use the so - called nordheim rule @xcite , which is a successful approach to describe @xmath85 in disordered mixed alloys .
there , the additional scattering rate is proportional to @xmath133 , with @xmath107 being the fraction of one of the constituents . in the present case
this corresponds to the fractions @xmath134 of co@xmath4ions in the higher and @xmath135 ) in the ls state .
thus , the temperature dependence of the additional scattering rate of equation([tauexp ] ) is expected to be given by ( see equations ( [ schott ] ) and ( [ schott2 ] ) ) @xmath136 , which are needed to suppress @xmath85 from a hypothetical bare phonon heat conductivity represented by @xmath85 of eucoo@xmath1 to the measured values of @xmath85 of lacoo@xmath1 , prcoo@xmath1 and ndcoo@xmath1(see text for details ) . in all three @xmath0coo@xmath1
an additional scattering sets in above the respective onset temperature @xmath24 ( marked by the arrow in each panel ) of the spin - state transition . the lines show the expected additional scattering rates due to the spin - state transition for a ls / is scenario ( solid ) and for the sochs model ( dashed ) ; see equations ( [ tauexp ] ) and ( [ equ_taudis ] ) . in panel (
a ) the scattering strengths @xmath126 for each model have been adjusted by fitting the lines to the experimental data in the temperature range below 150k .
these values have been kept fixed to calculate the lines in panels ( b ) and ( c ) , and , as shown by the solid symbols , the additional scattering above @xmath24 completely vanishes , when these lines are subtracted from the experimental @xmath128 . ]
the solid line in ( a ) corresponds to a calculation of @xmath128 via equation ( [ equ_taudis ] ) using the temperature - independent @xmath15k of the ls is scenario . here
, @xmath126 is the only adjustable parameter , and for @xmath137s@xmath61 this model yields a good description of the experimental data in the temperature range from 30k to 150k .
the dashed line shows the corresponding fit ( yielding @xmath138s@xmath61 ) for the same temperature range based on the sochs scenario with a temperature - dependent @xmath23 , which is in reasonable agreement with the data , too
. obviously , both models are not able to describe the decrease of @xmath127 above 150k .
this disagreement does , however , not contradict the models , because for @xmath139k the measured @xmath85 of lacoo@xmath1 is already close to its minimum value arising from the aforementioned lower limit of the mean - free path .
for @xmath140 of prcoo@xmath1 and ndcoo@xmath1 , we keep the values of @xmath126 fixed and use the gap values @xmath141k and @xmath142k obtained from the analysis of @xmath22 , cf . .
the corresponding curves , which for the different models only differ in the scattering strengths , are shown as solid and dashed lines in figs.[taurcoo3](b ) and ( c ) . when we subtract these lines from the experimental data , the slope changes around @xmath24completely vanish , as is shown by the solid symbols in figs.[taurcoo3](b ) and ( c ) .
this gives further evidence that the spin - state transition is indeed responsible for the suppression of the thermal conductivity of prcoo@xmath1 and ndcoo@xmath1 as it is the case for lacoo@xmath1 . with fits based on equation([kphonon ] ) assuming ( a ) the ls is or ( b ) the sochs model for the spin - state transition .
the dashed ( solid ) lines are obtained when a minimum mean - free @xmath119 is ( is not ) taken into account ( see text ) . ]
encouraged by this finding we tried to fit the thermal conductivity of lacoo@xmath1 via equation([kphonon ] ) including the additional scattering rate given by equations([tauexp ] ) and ( [ equ_taudis ] ) in equation([rates ] ) .
again , we restrict our analysis to @xmath143k and keep most of the fit parameters fixed to their values as obtained from the debye fit of the bare phononic @xmath85 of eucoo@xmath1 . from these parameters , we only readjust the prefactor @xmath144 of the point - defect scattering term in equation([rates ] ) , because the number of point defects varies in different crystals , and point - defect scattering also reduces @xmath85 at high temperatures .
in addition , the scattering strength @xmath126 is readjusted , but changes only little ( to @xmath145s@xmath61 and @xmath146s@xmath61 for the ls / is and the sochs model , respectively ) , and we consider the minimum mean - free path @xmath119 .
the results for the ls is as well as the sochs model are displayed in .
both models describe the thermal conductivity well over a large temperature range of about 200k , in particular , when the minimum mean - free path is taken into account . as was already seen in the quality of the fits based on the different models
hardly differs .
therefore , the analysis of the thermal conductivity of @xmath0coo@xmath1 does not allow to judge , whether the ls / is or the sochs scenario is preferable .
next we consider the additional suppression of @xmath85 of prcoo@xmath1 below @xmath24 , which is not related to the spin - state transition .
as can be seen in , the thermal expansion of this compound also exhibits an additional anomaly at lower temperatures .
this anomaly arises from the crystal - field splitting of the @xmath3 shell of the pr@xmath4 ions . in an orthorhombic crystal field
the @xmath61h@xmath147 multiplet of the free pr@xmath4 ion splits into nine singlets .
we are not aware of experimental investigations of the crystal - field splitting in prcoo@xmath1 , but it is known that the crystal - field levels of other orthorhombic pr@xmath28o@xmath1 are very similar @xcite . only the lower levels slightly depend on the exact orthorhombic distortion , which is mainly characterized by the @xmath28o@xmath28 canting angle .
the energy splitting between the crystal - field ground state and the first excited level in prcoo@xmath1 can be determined from a fit of the low - temperature thermal expansion by a schottky anomaly . as shown in ( a ) ,
such a fit for two non - degenerate levels ( setting @xmath148 and replacing @xmath16 by @xmath149 in equation ( [ schott ] ) ) describes the experimental data well and yields @xmath150k and @xmath151 % . for the higher - lying energy levels we use the values @xmath152k , @xmath153k , @xmath154 of prnio@xmath1 @xcite , which has almost the same @xmath28o@xmath28 canting angle as prcoo@xmath1 .
we note here that the thermal population of these higher - lying levels may contribute to @xmath22 at higher temperatures , but below about 50k the two - level system used is sufficient to model the experimental data .
the most likely source for the suppression of the thermal conductivity of prcoo@xmath1 in the intermediate temperature range is resonant scattering between different @xmath3 crystal - field levels , as it is also known from e.g. rare - earth garnets @xcite .
in such a resonant scattering process a phonon is absorbed by inducing a transition between different energy levels and then re - emitted in an arbitrary direction ( see e.g.@xcite ) , causing an additional thermal resistance .
a mechanism based on random volume changes as discussed above for the spin - state transition of co@xmath4 is unlikely because the @xmath3 orbitals are inner shells .
this difference is directly reflected in the magnitudes of the schottky peaks , which are mainly determined by the underlying volume change .
the low - temperature peak in @xmath13 of prcoo@xmath1 is e.g. one order of magnitude smaller than that of lacoo@xmath1 ; see and .
due to the spin - state transition of lacoo@xmath1 by a resonant scattering process completely failed . ] .
the thermal - expansion data ( symbols ) in panel ( a ) are well described by a schottky anomaly ( line ) of a two - level system with non - degenerate levels with energy splitting @xmath155k .
the lines f01 , f02 and f12 in panel ( b ) are fits for @xmath156k to the thermal conductivity data ( symbols ) including either the resonant scattering process @xmath157 , @xmath158 or @xmath159 , respectively . a sketch of the energy levels is also included @xcite . ] in order to model @xmath21 in the low-@xmath160 range we include the rate @xmath161 in equation([rates ] ) describing a direct resonant scattering process @xcite . here , @xmath162 is the energy difference between the two levels , and @xmath163 , @xmath164 denote their population determined by the partition sum . for the fit we keep again @xmath49k and @xmath165 m/s , while @xmath144 , @xmath166 , @xmath167 and @xmath0 are adjusted with respect to the data .
the fits have been restricted to @xmath168k , since other processes may become dominant at very low temperatures . in a first attempt , resonant scattering between @xmath169 and the first excited level @xmath170 is considered , yielding the fit labelled f01 in ( b ) .
this fit only reproduces the data for @xmath171k but totally misses the low - temperature maximum in @xmath21 .
basically the same result is achieved by considering the process @xmath158 ( curve f02 in ( b ) ) .
a much better description of the data is obtained , however , for the process @xmath159 ( curve f12 ; with parameters @xmath172s@xmath61 , @xmath173s@xmath29/k , @xmath174 and @xmath175s@xmath61 ) .
hence , the transition @xmath159 appears to be the dominant scattering channel of the phonons in prcoo@xmath1 at intermediate temperatures .
in this paper we have presented measurements of the thermal - expansion coefficient and the thermal conductivity of @xmath0coo@xmath1 , with trivalent @xmath2 la , pr , nd and eu . as a consequence of the chemical pressure due to the decreasing ionic radius from @xmath2 la to eu , the energy gap between the ls ground state and the first excited spin state of the co@xmath4 ions increases drastically and the temperature - induced spin - state transition of the co@xmath4ions is systematically shifted towards higher temperature . this strong pressure dependence is also reflected in pronounced schottky anomalies in the thermal - expansion coefficients , making this quantity a sensitive probe to analyze spin - state transitions .
the microscopic origin of the pressure dependence is the volume change due to the partial occupation of the @xmath45 levels of the co@xmath4 ions in the excited spin state in contrast to the empty @xmath45 levels of ls co@xmath4 ions .
another consequence of this volume difference is a strong suppression of the phonon heat transport , which sets in above the onset temperature of the spin - state transition , because the random occupation of the co sites with co@xmath4 ions in different spin states , i.e. of different sizes , causes additional lattice disorder . by considering this disorder as point - like defects
, our analysis clearly shows this effect in the temperature dependence of the corresponding scattering rates of lacoo@xmath1 , prcoo@xmath1 and ndcoo@xmath1 , where the spin - state transition sets in below room temperature , in contrast to eucoo@xmath1 which exhibits a stable ls state up to about 400k .
quantitatively , the temperature dependence of the additional scattering due to the spin - state transition can be described in a model based on the nordheim rule , which is usually applied to describe scattering in disordered mixed alloys . for the cobaltates
, this means that the temperature dependence of this additional scattering rate is given by the product of the occupation numbers of the ls and of the excited spin states .
moreover , we find that the scattering strength is constant for lacoo@xmath1 , prcoo@xmath1 and ndcoo@xmath1 .
our analysis does , however , not allow to distinguish between the different models discussed for the spin - state transition in lacoo@xmath1 , namely a ls is scenario with a temperature - independent energy gap or the more recently proposed model based on a spin - orbit coupled high - spin state with a temperature - dependent energy gap .
the above described model explains a damping of the phonon heat transport above the onset temperature of the spin - state transition .
experimentally , however , we find that as a function of increasing size of the @xmath18 ions the thermal conductivity decreases also in the temperature range below the respective onset of the spin - state transition . by studying five different lacoo@xmath1single crystals
, we could show that the low - temperature thermal conductivity inversely scales with the amount of magnetic impurities .
most likely , these impurities arise from a weak oxygen nonstoichiometry giving rise to a certain amount of co@xmath96 or co@xmath17 ions . since it is known that these magnetic co ions tend to form so - called high - spin polarons by inducing a spin - state transition in the neighbouring co@xmath4 ions , our data suggest that the phonon damping at low temperature arises from such a doping - induced spin - state transition .
the drastic increase of the energy gap to the excited spin state of co@xmath4 for decreasing @xmath18 ionic radii makes the formation of spin polarons more and more unlikely , and thus can naturally explain the systematic increase of the thermal conductivity in the temperature range below the temperature - induced spin - state transition .
additional low - temperature features have to be considered in prcoo@xmath1 , because the thermal expansion exhibits another schottky anomaly and there is an additional suppression of the thermal conductivity .
both effects are not related to the temperature - dependent spin - state transition of co@xmath4 , but stem from the crystal - field splitting of the ground - state multiplet of the @xmath3 shell of pr@xmath4 .
our analysis demonstrates that the schottky anomaly of the thermal expansion arises from a thermal occupation of the first excited crystal - field level , whereas the phonon damping is a consequence of resonant phonon scattering between the first two excited crystal - field levels .
we acknowledge useful discussions with m. haverkort and a. sologubenko .
this work was supported by the deutsche forschungsgemeinschaft through sonderforschungsbereich 608 . | we present a combined study of the thermal expansion and the thermal conductivity of the perovskite series @xmath0coo@xmath1 with @xmath2 la , nd , pr and eu . the well - known spin - state transition in lacoo@xmath1 is strongly affected by the exchange of the @xmath0 ions due to their different ionic radii , i.e. chemical pressure .
this can be monitored in detail by measurements of the thermal expansion , which is a highly sensitive probe for detecting spin - state transitions .
the co ions in the higher spin state act as additional scattering centers for phonons , therefore suppressing the phonon thermal conductivity . based on the analysis of the interplay between spin - state transition and heat transport , we present a quantitative model of the thermal conductivity for the entire series . in prcoo@xmath1 ,
an additional scattering effect is active at low temperatures .
this effect arises from the crystal field splitting of the @xmath3 multiplet , which allows for resonant scattering of phonons between the various @xmath3 levels . |
the disks around pre - main sequence stars are the sites of planet formation . models of disk evolution depend on many disk properties , including the total disk mass , the detailed density and temperature structure , and accretion characteristics .
determining these properties are key to constrain how , when and where in disks planets can form . since the main mass constituent in these disks is cold molecular hydrogen that can not be observed directly , observations of dust and trace gas molecules
are required to access the disk midplane properties .
the ionization fraction @xmath9 in disks is especially important because the magneto - rotational instability ( mri ) , believed to drive viscous accretion , requires ionization to couple the magnetic field to the gas @xcite .
insufficient ionization may suppress the mri and create a quiescent `` dead zone '' with important implications for processes such as dust settling , planetesimal formation , and planet migration @xcite . according to standard chemical models ,
disks are comprised of three characteristic layers ( e.g. * ? ? ?
* ; * ? ? ?
figure [ fig0 ] shows a schematic of the proposed layered structure focusing on ions .
the uppermost layer is exposed to ultraviolet radiation and x - rays that dissociate molecules and photoionize molecules and atoms to produce , e.g. c@xmath1 .
deeper into the disk , the radiation field is attenuated and molecules such as co become abundant . in this cooler molecular layer
, hco@xmath1 is predicted to be the most abundant ion ( e.g. * ? ? ?
* ) . at even lower heights , at temperatures below 20 k
, co begins to freeze out onto grains .
n@xmath0 can remain in the gas - phase at slightly lower temperatures than co @xcite , but in the deepest midplane layer , where temperatures drop below 16 k , the only molecules predicted to remain in the gas phase are h@xmath0 , h@xmath13 and their isotopologues .
indeed , h@xmath0d@xmath1 and d@xmath0h@xmath1 may provide the best , and possibly the only , observable tracers of the gas kinematics and the ionization fraction in the cold , dense midplane @xcite .
it is important to note that in addition to the vertical layers , each region has an inner truncation radius where the midplane temperature exceeds the co and n@xmath0 freeze - out temperatures , respectively .
ions such as h@xmath0d@xmath1 are therefore only expected in the outer disk , beyond the co snowline , whose radius will depend on the luminosity of the central star ( e.g. * ? ? ?
dm tau is a pre - main sequence star located @xmath15140 pc away from the sun , surrounded by a gas - rich keplerian circumstellar disk first discovered by @xcite . with a mass of 0.04 m@xmath16 ,
the disk is one of the more massive protoplanetary disks known @xcite .
it is a transition disk with an imaged central hole of radius 19 au @xcite .
the dm tau disk shows strong emission from species proposed to trace cold gas @xcite , in particular dco@xmath1 and n@xmath0h@xmath1 , and has previously been detected in a variety of molecular lines using single - dish telescopes @xcite .
it is therefore an especially good target to search for h@xmath0d@xmath1 emission .
@xcite reported a detection of the h@xmath0d@xmath1 @xmath2 line at 372 ghz at the 4.7 @xmath17 level toward dm tau using the caltech submillimeter observatory ( cso ) .
however , @xcite suggest that the dm tau result is `` at best a 2@xmath17 detection '' when analyzed with a proper model . to help to resolve these conflicting claims ,
we have used the submillimeter array to observe this h@xmath0d@xmath1 line toward dm tau with improved sensitivity .
we analyze these results , together with observations of other molecular ions , to provide new constraints on the ionization fraction in different regions of the disk .
the paper is organized as follows . in 2 , we describe the sma observations and complementary iram 30 meter telescope observations of molecular ions in the dm tau disk . in 3 , we present upper limits on h@xmath0d@xmath1 and n@xmath0h@xmath1 j=43 emission and detections of the co j=32 and h@xmath8co@xmath1 j=32 lines and other resolved observations of molecular ions from the literature . in 4 , we use the line intensities together with an empirically derived disk structure to constrain the midplane temperature ( 4.1 ) and to calculate upper limits on the @xmath18h@xmath0d@xmath1 column ( 4.2 ) , the total h@xmath0d@xmath1 column ( 4.3 ) , the total midplane ion column ( 4.4 ) , the midplane ionization fraction ( 4.5 ) , and the ionization fraction in higher disk layers ( 4.6 ) . in 5 , we compare the results with predictions from a generic disk chemistry model .
dm tau was observed with the submillimeter array on top of mauna kea , hawaii , on 2010 october 7 .
six of the array antennas were available for these observations , arranged in a compact configuration spanning baselines of 16 to 77 meters . using the dual receiver mode , we simultaneously targeted the @xmath20-h@xmath0d@xmath1 @xmath2 line at 372.4213 ghz and the n@xmath0h@xmath1 j=43 at 372.6725 ghz with the `` high frequency '' receiver , and the 345.796 ghz co j=32 line with the `` low frequency '' receiver .
table 1 summarizes the spectral line setup . at these frequencies ,
the synthesized beam sizes were @xmath21 and @xmath22 , respectively .
the observing conditions were generally very good , with atmospheric opacity @xmath23 measured at the nearby caltech submillimeter observatory , and stable atmospheric phase .
this opacity corresponds to a zenith transmission of @xmath24% at the h@xmath0d@xmath1 line frequency .
short observations of the two calibrators 3c120 and j0449 + 113 were interleaved with the observations of dm tau for gain calibration .
the bandpass response was calibrated with observations of uranus and the available bright quasars 3c273 , 3c454.3 and 3c84 .
observations of uranus and callisto provided the absolute scale for flux densities , via comparison to theoretical models for their emission , and the two calibrators agree within the expected uncertainties , but all reported data were calibrated with callisto .
the systematic uncertainty in the absolute flux scale is @xmath1510% .
the data were edited and calibrated with the idl - based mir software packagecqi / mircook.html ] .
continuum and spectral line images were generated and cleaned using miriad . to check that the flux calibration is accurate
, we compared the continuum fluxes of 0.22 jy at 0.87 mm and 0.30 jy at 0.81 mm with previously published values @xcite and found that they agree within the reported uncertainties . the h@xmath8co@xmath1 j=32 line at 260.255 ghz was observed toward dm tau on 2011 january 25 with the iram 30 meter telescope .
the observations were carried out with the emir 330 ghz receiver , with a beam size of @xmath1595 ( fwhm ) .
the receiver was connected to a unit of the autocorrelator with a spectral resolution of 320 khz and a bandwidth of 240 mhz , equivalent to an unsmoothed velocity resolution of @xmath150.4 km s@xmath4 .
typical system temperatures were 300 - 400 k. the observations were carried out using wobbler switching with a @xmath25 throw .
pointing was checked every @xmath152 hours on j0430 + 052 and j0316 + 413 with a typical accuracy of @xmath26 .
the main - beam brightness temperature was calculated from the antenna temperatures in _ pako _ using reported main beam and forward efficiencies ( b@xmath27 and f@xmath27 ) of 88% and 53% , respectively .
the data were reduced with the class program , part of the gildas software package .
linear baselines were determined from velocity ranges without emission features , and then subtracted from the spectra .
we did not detect either the h@xmath0d@xmath1 @xmath2 line or the n@xmath0h@xmath1 j=43 line .
figure [ fig1 ] shows imaging upper limits for these lines , along with moment maps for the co j=32 line and three lines from other molecular ions , hco@xmath1 j=32 , dco@xmath1 j=32 and n@xmath0h@xmath1 j=32 @xcite .
the velocity gradient due to disk rotation appears similar in all of the moment maps , even for the weak n@xmath0h@xmath1 and dco@xmath1 emission .
the apparent disk sizes are also similar within the uncertainties .
to estimate the disk emission region , we fit an elliptical gaussian to the co j=32 visibilities .
the resulting major and minor fwhm are 31 and 25 .
the effective disk size is defined as twice the fwhm of the co emission , which will contain 98% of the emission assuming a gaussian distribution .
this provides a convenient definition of the disk size , which is not model dependent , but it is important to note that the defined disk radius does not directly correspond to the outer edge of the disk when the disk is modeled as a power law , or to the critical radius when the disk is modeled with the similarity solution . with the adopted definition ,
the co 32 disk is 62@xmath2850 , corresponding to a disk radius of @xmath15400 au , which compares well with previous gas size estimates @xcite .
the fwhm of all ions are consistent with the co disk within the fit uncertainties .
the synthesized beam sizes across the disk major axis vary between 20 and 40 , corresponding to spatial scales of 280560 au .
the observations therefore only resolve the chemistry in the outer disk .
figure [ fig2 ] shows the spatially integrated spectra of the species imaged in figure [ fig1 ] .
the integration is carried out using an elliptical integration area the size of the co disk .
table [ tab2 ] lists the line fluxes integrated over the velocity range from 2 to 10 km s@xmath4 .
the hco@xmath1 j=32 has been previously observed toward dm tau using the iram 30 m telescope by @xcite , and the integrated intensity of 4.1[0.5 ] jy km s@xmath4 compares well with the sma value of 5.3[0.4 ] jy km s@xmath4 @xcite , considering the mutual calibration uncertainties , which are in addition to the rms in brackets . to convert the sma fluxes to intensities
, we adopt the co j=3 - 2 size scale for all of the emission lines and use the flux - to - intensity conversion formula : @xmath29 , where the constant factor is specific for the sma observations , @xmath30 is the intensity in k , @xmath31 the flux per beam ( here the 62@xmath2850 co disk ) in jy , @xmath32 the wavelength of the line in millimeters and @xmath33 and @xmath34 are the major and minor axes of the beam ( disk ) in arc seconds ( table [ tab2 ] ) . for the h@xmath0d@xmath1 and n@xmath0h@xmath1 j=4 - 3 lines , we calculate 2@xmath17 upper limits with the standard formula : @xmath35 , where @xmath36 is the number of channels across the co 32 line at fwhm .
the @xmath37-limit on the h@xmath0d@xmath1 line intensity is @xmath30.47 k km s@xmath4 .
averaged over the 22@xmath38 beam of the caltech submillimeter observatory , this 2@xmath17 upper limit corresponds to @xmath30.03 k km s@xmath4 , a factor of 2.5 below the claimed detection by @xcite .
this conclusively rules out the previous detection claim and entails that h@xmath0d@xmath1 is not yet detected in any protoplanetary disk .
this is consistent with the upper limits found by chapillon et al .
( subm . to a&a ) using the jcmt .
figure [ fig2 ] also includes the detection of h@xmath8co@xmath1 j=32 from the iram 30 meter telescope .
the reported intensity in table [ tab2 ] has been scaled to account for beam dilution by @xmath39 , i.e. the ratio of the iram 30 meter beam area over the co disk area .
the fact that the h@xmath8co@xmath1 line is detected indicates that the main isotopologue is optically thick , since if the emission ratio of hco@xmath40 h@xmath8co@xmath41 , the h@xmath8co@xmath1 line would have been buried in the noise .
we use the observations to constrain the ionization fraction in the disk in a four - step process .
first , the observed upper limit on the @xmath19-h@xmath0d@xmath1 1@xmath42 - 1@xmath43 line is used to constrain the @xmath19-h@xmath0d@xmath1 column density n(@xmath19-h@xmath0d@xmath1 ) for a range of plausible temperatures for the disk midplane .
second , we use recent model results on the temperature dependent @xmath5 ratio to convert n(@xmath19-h@xmath0d@xmath1 ) limits to total ( @xmath44 ) h@xmath0d@xmath1 limits , i.e. n(h@xmath0d@xmath1 ) .
third , we use calculations of the h@xmath13/h@xmath0d@xmath1/d@xmath0h@xmath1/d@xmath13 ratios for the appropriate range of temperatures to constrain the total column density n(@xmath45h@xmath46d@xmath47 ) of these ions in the midplane .
finally , we use the total column density of ions in the midplane together with a physical model of the dm tau disk and measurements of n(hco@xmath1 ) , n(dco@xmath1 ) , and n(n@xmath0h@xmath1 ) to constrain @xmath9 in the disk midplane , an intermediate layer where co has started to freeze out , and the upper warm molecular layer . before embarking on this series of steps
, we place observational constraints on the temperatures in the disk midplane with the aim to limit the range of temperatures to consider when deriving a midplane @xmath9 from the @xmath19-h@xmath0d@xmath1 1@xmath42 - 1@xmath43 . for this investigation
, we define the midplane as the region where all heavy elements are frozen out and h@xmath13 and its isotopologues are the dominant charge carriers . theoretically , the midplane temperature is limited by the freeze - out temperatures of co and n@xmath0 , since both react readily with h@xmath13 , resulting in hco@xmath1 and n@xmath0h@xmath1 @xcite . since n@xmath0 is maintained in the gas phase at slightly lower temperatures than co @xcite
, the n@xmath0h@xmath1 rotational temperature provides the strongest observational limit available on the midplane temperature . to estimate the n@xmath0h@xmath1 rotational temperature , we assume lte and co - spatial emission from the n@xmath0h@xmath1 32 and 43 lines
this is a reasonable approximation because chemical models predict that n@xmath0h@xmath1 should be abundant in a narrow temperature regime between where co begins to freeze - out and n@xmath0 freeze - out is not complete @xcite .
assuming optically thin emission , unity filling factor within the co 32 disk , and molecules thermalized at a single rotational temperature , @xmath48 from @xcite , where @xmath49 is the total number of molecules , @xmath50 is the temperature dependent partition function , @xmath51 the energy of the upper level in k , and @xmath52 the rotational temperature .
we obtain the partition function @xmath50 by second order interpolation of the listed values in cdms , @xmath53 is gathered directly , and @xmath51 is calculated from the listed @xmath54 values : @xmath55 , where @xmath56 is the line frequency in ghz .
then an upper limit on the n@xmath0h@xmath1 temperature can be calculated from @xmath57 using the integrated intensity of the j=32 line and the 2@xmath17 upper limit on the j=43 line , we obtain a 2@xmath17 upper limit on the n@xmath0h@xmath1 rotational temperature of @xmath326 k. this is consistent with the predictions that n@xmath0h@xmath1 should become abundant below the co freeze - out temperature of 20 k , but does not provide any stronger constraints .
it is also consistent with the n@xmath0h@xmath1 j=10 line reported by @xcite .
we estimate the line flux from their figure 1 to be @xmath30.1 jy km s@xmath4 .
using this limit and the integrated intensity of the j=32 line results in a lower limit on the n@xmath0h@xmath1 rotational temperature of 11 k. more sensitive observations of the n@xmath0h@xmath1 lines would be valuable to constrain this excitation temperature further . in the absence of more detailed information , we calculate lte column densities of midplane ions for a range of temperatures , from 10 to 20 k. the 20 k boundary is set by laboratory experiments on co and n@xmath0 freeze - out above this temperature a majority of co and n@xmath0 are maintained in the gas - phase quickly destroying any formed h@xmath13 @xcite .
the co snowline was recently reported to correspond to a freeze - out temperature of 19 k , consistent with predictions @xcite . in general , column densities
are calculated most accurately by constructing a self - consistent chemical - physical disk model and applying a radiative transfer code to predict line emission profiles that can be compared directly to observations . without a detection of the h@xmath0d@xmath1 line , however , it is difficult to constrain where the h@xmath0d@xmath1 emission arises in the disk .
we therefore opt for a classical lte approach ( e.g. * ? ? ?
* ) , calculated for the range of reasonable disk midplane temperatures ( see [ sec : midplane_temp ] ) to put limits on the disk averaged n(@xmath19-h@xmath0d@xmath1 ) .
a significant complication for the lte calculation is that @xmath20 ( @xmath19 ) and @xmath58 ( @xmath59 ) h@xmath0d@xmath1 can behave as separate species at low temperatures @xcite resulting in a non - thermal partitioning between the @xmath19 and @xmath59 states . to calculate the partition function for @xmath19-h@xmath0d@xmath1
, we use a two - level approximation , which assumes no transitions between the @xmath19 and @xmath59 h@xmath0d@xmath1 , and that transitions between the ground state and first excited state dominate within each `` species '' of h@xmath0d@xmath1 .
this approximation is expected to be valid at temperatures of 10 to 20 k , since e@xmath60 is 104 k for the first excited state of @xmath19-h@xmath0d@xmath1 .
the @xmath19-h@xmath0d@xmath1 partition function is calculated from @xmath61 where the degeneracies , @xmath62 and @xmath63 , are equal to 3 .
assuming lte ( equation 1 ) , excitation temperatures in the range 10 to 20 k and the partition function appropriate for these temperatures , n(@xmath19-h@xmath0d@xmath1 ) limits are calculated from the observed upper limit on the h@xmath0d@xmath1 j=1@xmath42 - 1@xmath43 line flux .
as shown in figure 3a , the limits on the disk average n(@xmath19-h@xmath0d@xmath1 ) range from @xmath64 @xmath7 ( with the highest value corresponding to the lowest temperature ) .
the limit on the total n(h@xmath0d@xmath1 ) , i.e. @xmath65 , depends on the @xmath19-h@xmath0d@xmath1 column density and the temperature dependent @xmath5 ratio .
this ratio has been modeled by @xcite from 4 to 20 k for a molecular hydrogen density of 10@xmath66 @xmath67 and an interstellar grain distribution .
the derived @xmath5 ratio is 0.50.1 for temperatures of 1020 k , with the lowest ratio at 1317 k. while the assumed grain size distribution and density are more appropriate for dense clouds than disk midplanes , the grain size distribution mainly affects the temperature profile , and @xcite find that the h@xmath0d@xmath1 @xmath5 is nearly constant with increasing density .
thus we expect the values calculated with these assumptions should be valid for disk midplane conditions .
figure 3b shows the result of applying the literature @xmath5 ratios to the previously calculated @xmath19-h@xmath0d@xmath1 column density limits to obtain limits on n(h@xmath0d@xmath1 ) as a function of temperature .
these limits range from @xmath68 to @xmath69 @xmath7 .
the presence of a peak value at 13 k is a direct consequence of the fact that the limit on n(@xmath19-h@xmath0d@xmath1 ) decreases with temperature , while the @xmath5 ratio has a minimum at 1317 k. the limit on the total column density of ions in the midplane , n(@xmath45 h@xmath46d@xmath70 ) , depends on n(h@xmath0d@xmath1 ) and the ratio of n(h@xmath0d@xmath1 ) to n(@xmath45 h@xmath46d@xmath70 ) . like the @xmath5 ratio
, this ratio is expected to vary with density and temperature , as well as other environmental properties such as grain size @xcite .
this ratio will also depend on whether or not co and n@xmath0 are frozen out completely , or if low abundances of these species can be maintained in the gas - phase through non - thermal desorption @xcite .
@xcite modeled the ratios of all deuterated h@xmath13 isotopologues for molecular hydrogen density 10@xmath71 @xmath67 and interstellar grains as a function of temperature .
as mentioned previously , the disk midplane density is likely higher than assumed in these models .
@xcite found an increasing importance of d@xmath13 with increasing density at 10 k , but it is unclear whether this effect is important also at higher temperatures .
another potential unknown is the abundance of h@xmath1 , though @xcite found that its contribution is negligible at the high densities of the disk midplane . with these caveats in mind
, we use the ratios derived by @xcite with maximum depletion , i.e. minimum amounts of co and n@xmath0 in the gas - phase , to calculate upper limits on n(@xmath45 h@xmath46d@xmath70 ) in the disk midplane as a function of assumed temperature .
figure 3c shows the results , with limits that range from @xmath72 to @xmath73 @xmath7 , with the peak value at 13 k. the modification of the shape compared to the h@xmath0d@xmath1 curve is due to the fact that the n(h@xmath0d@xmath1)/n(@xmath45 h@xmath46d@xmath70 ) ratio peaks near 13 k and then falls off quickly at higher temperatures . to derive limits on the midplane ion abundances
, we adopt the model of the dm tau disk density and temperature structure described by @xcite based on parametric fitting of the spectral energy distribution and resolved observations of millimeter - wave dust emission .
this model does not include any chemistry , but rather provides an observationally constrained physical structure , which can be used to derive the gas masses in different temperature layers in the disk .
this is needed to convert an ion column density into an ion abundance .
we apply a standard interstellar gas - to - dust ratio of 100 to convert the dust densities in this model to gas densities . to determine a midplane mass
, we define the disk midplane as all material @xmath7416 k , the temperature for complete n@xmath0 freeze - out in most astrophysical environments based on laboratory measurements @xcite , i.e. the region where it is most likely that h@xmath13 and its isotopologues become abundant . in this disk layer , the density weighted temperature is 11 k , and the disk averaged gas column density n@xmath75 is @xmath76 @xmath7 .
using this column density , and the upper limit on the ion column density in the midplane of @xmath77 @xmath7 from table 3 and fig .
4 , @xmath78 n@xmath79 . alternatively , if the onset of co freeze - out at 20 k defines the region where h@xmath0d@xmath1 is abundant rather than n@xmath0 freeze - out at 16 k , then the density averaged ion temperature is 13 k , and @xmath80 n@xmath79 .
table [ tab3 ] summarizes the midplane ( @xmath316 k ) ionization results and also shows how the ionization fractions compare if h@xmath0d@xmath1 is only assumed to be present in the midplane or in the entire t@xmath320 k region .
it is important to keep in mind that all of these calculations are based on model dependent values , and better limits on the midplane ionization abundance requires not only more sensitive @xmath18h@xmath0d@xmath1 observations , but also direct constraints on the @xmath5 ratio and the relative abundances of different isotopologues . in particular
the latter estimate may be off by an order of magnitude from different model predictions of the importance of d@xmath13 compared to h@xmath0d@xmath1 .
it is interesting to compare the estimates of ionization fraction in different layers of the disk , since this may impact e.g. mri activation and accretion flow .
the mass of each layer , defined by a temperature interval , can be estimated from the same dm tau physical model as was used in 4.5 . between 20 k , where co begins to freeze out , and
the co photodissociation region , hco@xmath1 is predicted to be the most abundant ion ( e.g. * ? ? ?
below 20 k , as co starts to freeze out , n@xmath0h@xmath1 becomes more abundant , since reactions with co are the main destruction pathway of n@xmath0h@xmath1 .
below 16 k , co and n@xmath0 are completely frozen out and h@xmath13 and its isotopologues become the main charge carriers . the dco@xmath1 abundance peaks between the region where hco@xmath1 dominates and the midplane @xcite and therefore is assumed to coexist with n@xmath0h@xmath1 in the intermediate layer . in this layer , dco@xmath1/hco@xmath1 exceeds unity and the hco@xmath1 contribution to the charge balance should be small . this model result is supported by comparing dco@xmath1 line intensities : @xcite detected dco@xmath1 21 at @xmath81 jy km s@xmath4 and the intensity ratio of the 32/21 lines implies a rotational temperature of 1123 k , assuming lte . for the midplane and the low - temperature intermediate layer ,
the emission regions of the ions are expected to be well confined , and the assumption of lte at a single rotational temperature should be a reasonable first approximation . for the warm molecular layer where hco@xmath1 dominates , the assumption of lte is not _ a priori _ a good approximation since this layer spans a temperature range from 20 to 100 k or higher .
however , the derived hco@xmath1 column density is not very sensitive within this range ; it varies by less than a factor of three when using eq .
1 over the full range of plausible temperatures and the lte approach retains utility . as discussed below , the outcome of this lte calculation compares well with a radiative transfer calculation of hco@xmath1 emission in the dm tau disk @xcite .
we estimate the disk averaged n(hco@xmath1 ) , n(n@xmath0h@xmath1 ) and n(dco@xmath1 ) using equation 1 , the observed intensities from table [ tab2 ] , and the mass weighted temperatures from the dm tau structural model of 38 k for the co dominated layer ( @xmath1020 k ) and 18 k for the intermediate layer where n(n@xmath0h@xmath1 ) and n(dco@xmath1 ) are most abundant ( 1620 k ) .
table 3 lists the results and figure [ fig4 ] shows the resulting column densities together with the h@xmath13 limit ( lte at 11 k ) , within their respective temperature regimes ( @xmath316 k , 1620 k , and @xmath1020 k ) , corresponding to the different disk layers .
note that n(hco@xmath1 ) is calculated from h@xmath8co@xmath1 assuming a standard @xmath82c/@xmath8c ratio of 70 , since the high opacity of the main isotopologue underestimates n(hco@xmath1 ) by a factor of seven .
the calculation results in a disk averaged dco@xmath1/hco@xmath1 ratio of @xmath150.05 .
this is comparable to a previous estimate of 0.035 toward the disk around tw hya @xcite .
both exceed the cosmic d / h ratio of @xmath1510@xmath83 by three orders of magnitude , demonstrating an efficient deuterium fractionation in these disks .
the derived ratio is an order of magnitude above the ratio found by @xcite , who used resolved hco@xmath1 1 - 0 and single - dish dco@xmath1 2 - 1 emission and a radiative transfer calculation rather than lte .
the derived disk averaged n(hco@xmath1 ) are almost identical over the inner 400 au of the disk : @xmath84 @xmath7 from their tabulated column density at 300 au and power law index , compared to @xmath85 @xmath7 calculated here assuming lte at 38 k. the different ratios must thus be due to different assumptions on the dco@xmath1 emission conditions .
the hco@xmath1 column density compares well with model predictions of the outer disk ( 100400 au ) .
it is for example consistent with the disk chemistry model by @xcite within a factor of 3 .
@xcite observed n@xmath0h@xmath1 10 with the plateau de bure interferometer and derived n@xmath0h@xmath1/hco@xmath1 of @xmath150.020.03 .
this ratio agrees well with the results in table 3 , which implies a disk - averaged ratio of 0.03 .
the disk averaged n(h@xmath0 ) within each temperature regime are 15 , 2.4 , and @xmath86 @xmath7 , and these values are used with the derived ion column densities to calculate disk averaged @xmath9 in each layer .
figure [ fig4 ] shows that the hco@xmath1 abundance and the midplane ion limits are comparable at @xmath87 n@xmath79 , while the abundance of dco@xmath1 + n@xmath0h@xmath1 is an order of magnitude lower at @xmath88 n@xmath79 .
overall , the derived ion abundances suggest a decreasing @xmath9 with depth in the disk , which might be expected from attenuation of ionizing radiation .
the assumption that dco@xmath1 and n@xmath0h@xmath1 alone are representative for the degree of ionization in the intermediate layer may be oversimplified , however .
the @xmath89 layer is a transition region where some hco@xmath1 may still be present and h@xmath13 may already have started to become important . a combination of better limits on the midplane ions as well as more detailed modeling of the intermediate layer with dco@xmath1 and n@xmath0h@xmath1 will be needed to confirm the depth dependent trend on @xmath9 .
in the previous section the disk averaged column densities and abundances were calculated based on observed line intensities and a density and temperature structure of dm tau from the literature . this particular structure model has not been used as a basis for chemical modeling . in this section
we therefore use predictions on the ion chemistry in generic disk models to compare our observational results with the current understanding of the chemical compositions in disks .
model predictions on the abundances of h@xmath0d@xmath1 and other ions in protoplanetary disks have been made by many authors ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) for generic t tauri disk structures .
in particular , the models of @xcite include abundances and column densities for all of the ions considered in this study , and therefore we focus on comparing the observational results with these predictions .
we examine specifically the disk averaged ionization fraction , the ionization fraction in different disk layers , and the relative abundances of different ions .
@xcite based their chemical calculations on the disk structure model of @xcite , representing a generic t tauri disk . the assumed accretion rate , disk mass etc are within a factor of a few of the values that have been derived for dm tau @xcite .
most disk chemistry predictions are therefore expected to be valid for dm tau , except for absolute ion column densities .
the chemical modeling is based on the umist database , updated with multiple deuterated species @xcite . within this model framework , @xcite investigated the importance of cosmic ray desorption and photodesorption on the disk chemical abundances , which has a large effect on the ion abundance ratios .
an important caveat when comparing observations and model predictions , is that @xcite presents column densities as a function of radius with no simple procedure to calculate disk averaged column densities .
observations can therefore only be compared with model predictions at specific radii . from the dm tau structure model
the observed emission is dominated by the outer disk ; beyond 50 au the column densities of the different temperature layers are almost constant , which entails that the larger emission region of the outer disk will dominate . in particular more than 2/3 of the ion emission
is expected to originate outside of 250 au if the ions are present across the co disk . in the models by @xcite the ion column densities are constant within a factor of 2 at 100400 au , except for the relative importance of different h@xmath13 isotopologues .
it is therefore useful to compare model predictions in this regime with the observed disk averaged values as long as this factor of two uncertainty is considered .
@xcite provides the most complete model results at 250 au and this radius is used in all comparisons between model and observations .
the main sources of ionizing radiation in disks are ultraviolet ( uv ) photons , cosmic rays , radioactive decay and x - rays ( e.g. * ? ? ?
* ) , and the ionization level therefore depends on a combination of the incident fluxes and the attenuation of external photons with dust and gas column density .
@xcite include both the interstellar radiation field and a stellar uv field that is 500 times stronger at 100 au . because of efficient absorption by dust and h@xmath0
, the uv ionization is mainly important in the disk atmosphere and upper molecular layer .
cosmic rays are assumed to penetrate unperturbed to the disk midplane resulting in an ionization rate of @xmath90 s@xmath4 .
radioactive decay also contributes to the midplane ionization , at a rate of @xmath91 s@xmath4 .
the effect of x - rays are not included in the model and the model may therefore underestimate @xmath9 in lower disk layers . in this model , the total ion column density ( h@xmath13 , h@xmath0d@xmath1 , d@xmath0h@xmath1 , d@xmath13 , hco@xmath1 , dco@xmath1 , n@xmath0h@xmath1 ) at 250 au is @xmath92 @xmath7 , corresponding to an average @xmath9 of @xmath93 n@xmath79 .
this value is consistent with the observed abundances and limits . in the related disk model of @xcite ,
the ion abundance depends on the assumed cosmic ray flux , as expected . in disks , including the model from @xcite the temperature increases both inward and upward in the disk , resulting in a vertically layered structure outside of an inner truncation radius . in the model ,
hco@xmath1 extends through the disk , while dco@xmath1 is only abundant outside of 50 au and h@xmath0d@xmath1 outside of 100 au . as mentioned above , beyond 100 au
the different ion column densities are basically constant , justifying the comparison between disk averaged observations and model predictions at 250 au . because all uv photons and probably some of the x - rays and cosmic rays are absorbed in the outer disk layers , the ionization fraction should decrease between the outer disk layers and the disk midplane . at 250
au , in the disk chemistry models ( fig . 3 in * ? ? ?
* ) the ion abundance in the upper molecular layer , where co is in the gas - phase and hco@xmath1 is the dominant ion , is a factor of 5 higher than in the midplane ( @xmath94 compared to @xmath95 n@xmath79 ) . in the intermediate region , where dco@xmath1 and n@xmath0h@xmath1 abundances peak , the ion abundance is about a factor of 23 below the co layer ion abundance .
this is qualitatively consistent with the results of the analysis of the dm tau observations ( see figure 4 ) .
_ hco@xmath1 and dco@xmath1 : _ the hco@xmath1 column density predicted by @xcite is @xmath96 @xmath7 at 250 au and is found to not depend on the different desorption mechanisms considered .
in contrast , the dco@xmath1 column density predictions vary by an order magnitude , resulting in a dco@xmath1/hco@xmath1 ratio of 0.11 .
the higher dco@xmath1/hco@xmath1 ratios appear with desorption through cosmic ray heating , which maintains co in the gas - phase in the disk midplane where the temperatures are low enough for efficient deuterium fractionation and enhanced dco@xmath1 production .
the observations indicate that the disk averaged dco@xmath1/hco@xmath1 abundance ratio is below 0.1 , which suggests that co is confined to the warm @xmath1020 k molecular layer .
the observational results are consistent with including uv photodesorption , which is efficient only in the warm molecular layer . _ n@xmath0h@xmath1 : _
the n@xmath0h@xmath1/hco@xmath1 abundance ratio is predicted to be in the range from @xmath97 to @xmath98 , but the models with the highest values also predict dco@xmath1/hco@xmath1 ratios of unity and can be excluded . compared to the value of 0.03 derived from the observations , the models are an order of magnitude too low .
the reason for this discrepancy is unknown , demonstrating that even the simplest nitrogen - based chemistry remains poorly constrained .
_ h@xmath0d@xmath1 : _
n(h@xmath0d@xmath1 ) is predicted to be 10@xmath82 @xmath7 , with midplane abundance @xmath1510@xmath99 n@xmath79 .
these values are well below the observational limits , and the reason is that d@xmath13 is the dominant carrier of charge in the midplane in these models , 2050 times more abundant than h@xmath0d@xmath1 . in the disk model by @xcite ,
the h@xmath0d@xmath1 column density is predicted to be significantly higher , 10@xmath8 @xmath7 , likely because multiple deuteration is not considered .
both models are thus consistent with the current limits , and better constraints on the chemical makeup of the midplane will require not only more sensitive observations of h@xmath0d@xmath1 , but also robust @xmath5 determinations , and additional constraints on the abundance ratios of the isotopolgues .
1 . a new sma observation of the h@xmath0d@xmath1 @xmath2 line toward the dm tau disk results in a disk - averaged intensity upper limit of @xmath30.47 k km s@xmath4 .
2 . assuming lte and a temperature dependent @xmath5 conversion factor , the observation corresponds to an h@xmath0d@xmath1 column density @xmath3421@xmath100 @xmath7 , for midplane temperatures of 1020 k. 3 . based on sma and iram 30 meter telescope observations of additional molecular ions toward dm tau and a previously published model of the disk structure
, we estimate ionization fractions of @xmath101 in the molecular layer where t@xmath1020 k ( based on h@xmath8co@xmath1 ) , @xmath102 in the 1620 k layer where n@xmath0h@xmath1 and dco@xmath1 are common , and @xmath103 in the midplane below 16 k where h@xmath13 isotopologues dominate .
the ion emission is expected to be dominated by the outer disk , beyond 250 au , because of the larger emission area .
4 . the estimated ionization fractions depend on adopted model predictions of @xmath5 ratios , deuteration levels and the dominating ions in different disk layers .
observational constraints on all three parameters together with deeper h@xmath0d@xmath1 observations are therefore key to calculating more accurate ionization fractions in disks .
the manuscript has benefitted from comments by an anonymous referee and from discussions with eric herbst and ted bergin .
the sma is a joint project between the smithsonian astrophysical observatory and the academia sinica institute of astronomy and astrophysics and is funded by the smithsonian institution and the academia sinica .
support for k. i. o. is provided by nasa through a hubble fellowship grant 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 .
we also acknowledge nasa origins of solar systems grant no .
nnx11ak63 lccccccc usb : s08 & 345.75345.85&256&0.36&co 32 & 345.7960&33 + usb : s08 & 372.37372.48&128&0.66&h@xmath0d@xmath1 1@xmath42 - 1@xmath43&372.4213&104 + usb : s11 & 372.63372.73&128&0.66&n@xmath0h@xmath1 43&372.6725&45 + lcccc co 32 & 16.7[0.2 ] & 0.87 & 5.43[0.06 ] & 216[22 ] + hco@xmath1 32 & 5.3[0.4 ] & 1.12 & 2.9[0.2 ] & 143[14 ] + dco@xmath1 32 & 0.7[0.1 ] & 1.39 & 0.62[0.08 ] & 104[10 ] + n@xmath0h@xmath1 32 & 1.0[0.1 ] & 1.07 & 0.52[0.06 ] & 143[14 ] + n@xmath0h@xmath1 43 & @xmath31.8 & 0.81 & @xmath30.47 & 303[30 ] + h@xmath0d@xmath1 1@xmath42 - 1@xmath43 & @xmath31.8 & 0.81 & @xmath30.47 & 303[30 ] + h@xmath8co@xmath1 32&0.70[0.08 ] & 1.15&0.22[0.03]@xmath104 + + @xmath104 after correction for beam dilution .
lcccc @xmath19-h@xmath0d@xmath1 & @xmath316/@xmath320 & 11@xmath104 / 13@xmath105&@xmath32.4@xmath104 / @xmath31.9@xmath105&@xmath30.08@xmath104 / @xmath30.06@xmath105 + h@xmath0d@xmath1 & @xmath316/@xmath320&11 / 13&@xmath310 / @xmath321 & @xmath30.35 / @xmath30.6 + @xmath45h@xmath106d@xmath107 & @xmath316/@xmath320&11 / 13 & @xmath340 / @xmath385 & @xmath32.7 / @xmath34.5 + n@xmath0h@xmath1 & 1620&18 & 0.31@xmath108 & @xmath109 + dco@xmath1 & 1620&18 & 0.36@xmath110 & @xmath111 + hco@xmath1 & @xmath1020&38 & 9.4@xmath112 & @xmath113 + + @xmath104 density weighted temperature in the midplane layer when t@xmath316 k. + @xmath105 density weighted temperature in the midplane layer when t@xmath320 k. | we present millimeter - wave observations of several molecular ions in the disk around the pre - main - sequence star dm tau and use these to investigate the ionization fraction in different regions of the disk .
new submillimeter array ( sma ) observations of h@xmath0d@xmath1 j=@xmath2 , n@xmath0h@xmath1 j=43 and co j=32 are presented .
h@xmath0d@xmath1 and n@xmath0h@xmath1
are not detected and using the co 32 disk size the observations result in an upper limit of @xmath30.47 k km s@xmath4 for both lines , a factor of 2.5 below previous single - dish h@xmath0d@xmath1 observations
. assuming lte , a disk midplane temperature of 1020 k and estimates of the h@xmath0d@xmath1 @xmath5 ratio , the observed limit corresponds to n@xmath6 @xmath7 .
we adopt a parametric model for the disk structure from the literature and use new iram 30 meter telescope observations of the h@xmath8co@xmath1 j=32 line and previously published sma observations of the n@xmath0h@xmath1 j=32 , hco@xmath1 j=32 and dco@xmath1 j=32 lines to constrain the ionization fraction , @xmath9 , in three temperature regions in the disk where theoretical considerations suggest different ions should dominate : ( 1 ) a warm , upper layer with t@xmath1020 k where co is in the gas - phase and hco@xmath1 is most abundant , where we estimate @xmath11 , ( 2 ) a cooler molecular layer with t = 1620 k where n@xmath0h@xmath1 and dco@xmath1 abundances are predicted to peak , with @xmath12 , and ( 3 ) the cold , dense midplane with t@xmath316 k where h@xmath13 and its deuterated isotopologues are the main carriers of positive charge , with @xmath14 .
while there are considerable uncertainties , these estimates are consistent with a decreasing ionization fraction into the deeper , colder , and denser disk layers .
stronger constraints on the ionization fraction in the disk midplane will require not only substantially more sensitive observations of the h@xmath0d@xmath1 @xmath2 line , but also robust determinations of the @xmath5 ratio , observations of d@xmath0h@xmath1 and stronger constraints on where n@xmath0 is present in the gas phase . |
the work has been partially supported by anpcyt and conicet of argentina , and by mineco ( spain ) grant .
fpa2013 - 47443-c2 - 1-p , gva - prometeoii/2014/066 and sev-2014 - 0398 .
y. b. zeldovich and m. y. khlopov , phys .
b * 79 * ( 1978 ) 239 .
c. t. hill , nucl .
b * 224 * ( 1983 ) 469 .
doi:10.1016/0550 - 3213(83)90386 - 3 v. vento , int .
j. mod .
phys . a * 23 * ( 2008 ) 4023 doi:10.1142/s0217751x08041669 [ arxiv:0709.0470 [ astro - ph ] ] .
l. n. epele , h. fanchiotti , c. a. garcia canal and v. vento , eur .
j. c * 56 * ( 2008 ) 87 doi:10.1140/epjc / s10052 - 008 - 0628 - 0 [ hep - ph/0701133 ] .
l. n. epele , h. fanchiotti , c. a. g. canal , v. a. mitsou and v. vento , eur .
j. plus * 127 * ( 2012 ) 60 doi:10.1140/epjp / i2012 - 12060 - 8 [ arxiv:1205.6120 [ hep - ph ] ] .
l. n. epele , h. fanchiotti , c. a. g. canal , v. a. mitsou and v. vento , arxiv:1607.05592 [ hep - ph ] .
k. a. milton , rept .
* 69 * ( 2006 ) 1637 doi:10.1088/0034 - 4885/69/6/r02 [ hep - ex/0602040 ] .
a. billoire , r. lacaze , a. morel and h. navelet , phys .
lett . * 78b * ( 1978 ) 140 .
doi:10.1016/0370 - 2693(78)90367 - 2 t. muta and t. niuya , prog .
* 68 * ( 1982 ) 1735 .
doi:10.1143/ptp.68.1735 a. czarnecki and s. g. karshenboim , hep - ph/9911410 .
l. patrizii and m. spurio , ann .
nucl . part .
* 65 * ( 2015 ) 279 doi:10.1146/annurev - nucl-102014 - 022137 [ arxiv:1510.07125 [ hep - ex ] ] .
b. acharya _ et al . _
[ moedal collaboration ] , int .
j. mod .
a * 29 * ( 2014 ) 1430050 doi:10.1142/s0217751x14300506 [ arxiv:1405.7662 [ hep - ph ] ] .
b. acharya _ et al .
_ [ moedal collaboration ] , jhep * 1608 * ( 2016 ) 067 doi:10.1007/jhep08(2016)067 [ arxiv:1604.06645 [ hep - ex ] ] .
b. acharya _ et al . _
[ moedal collaboration ] , phys .
* 118 * ( 2017 ) no.6 , 061801 doi:10.1103/physrevlett.118.061801 [ arxiv:1611.06817 [ hep - ex ] ] . | we show that due to the large coupling constant of the monopole - photon interaction the possibility that monopolium , and therefore monopole - antimonopole , decay into many photons must be contemplated experimentally .
* multiphoton annihilation of monopolium * h. fanchiotti@xmath0 , c.a .
garca canal@xmath0 and v. vento@xmath1 _ @xmath0iflp / conicet and departamento de fsica _ + _ universidad nacional de la plata , c.c.67 , 1900 , la plata , argentina _
+ _ @xmath1departamento de fsica terica - ific .
universidad de valencia - csic .
_ + _ e-46100 , burjassot ( valencia ) , spain _ inspired by an old idea of dirac and zeldovich @xcite we proposed that monopolium , a bound state of monopole anti - monopole @xcite could be easiaer to detect experimentally than free monopoles @xcite .
we have studied the annihilation of monopolium into two photons @xcite . in here
we are going to argue and explicitly show that monopolium might annihilate most preferably into many low energy photons .
the same type of arguments hint that monopole - antimonopole also annihilate most preferably into many low energy photons .
0.3 cm 0.3 cm let us assume that the monopole - photon coupling is described by some complicated effective vertex of the form @xmath2 , as shown in fig .
[ vertex ] where @xmath3 is the dressed monopole magnetic coupling constant , @xmath4 being the momentum transfer in the vertex , and @xmath5 a function containing all the non perturbative behavior associated with the large magnetic coupling of the monopole , which is normalized so that @xmath6 .
recalling the parapositronium @xcite results rewritten in terms of the effective vertex we get , @xmath7 where @xmath8 by dirac s quantization condition ( dqc ) @xcite . in fig .
[ multiphoton ] we show one of the @xmath9 contributions to the amplitude for a @xmath10 photon decay , and we note that this type of contributions in the above ratios are determined only by vertices and propagators . in the rest frame of the bound system the annihilation into many photons leads to an average momentum for each photon much smaller than the mass of monopolium and therefore much smaller than the mass of the monopole . in order to make an estimate of the above ratios we assume that in the propagators the monopole mass dominates over the momentum and therefore the calculation of the width , in units of monopole mass , depends exclusively on three factors : the number of diagrams @xmath11 , the photons symmetry factor @xmath12 and the phase space of outgoing massless particles , namely 1.0 cm 0.3 cm @xmath13 where @xmath14 is the monopolium mass and @xmath15 being @xmath10 the number of photons emitted . with all these approximations
we obtain @xmath16 since the photon momenta on the average are going to be small , we can further substitute @xmath17 .
this approximation is valid for dirac s original formulation .
some coupling schemes lead to small effective couplings at threshold @xcite . in these cases
the analysis has to be carried out off threshold . in fig .
[ ration2 ] we plot eq.([ratio ] ) for two different binding energies and note that we get bell shape distributions . for small binding energies ( @xmath18 ) @xmath19 with @xmath20 . for @xmath21 @xmath22 with @xmath23.thus
the multiplicity decreases as the binding energy increases .
we now study the dependence of multiplicity with the binding energy .
to do so we find the maxima of the ratio in eq.([ratio ] ) as a function of binding energy . the result is plotted in fig .
[ ratiobindingexact ] were we show the binding energy in units of monopole mass as a function of the average photon multiplicity in the annihilation .
the outcome is clear , large average multiplicities,10 - 12 photons , arise if the binding energy is small , @xmath24 , while smaller average multiplicities , 4 - 6 photons , occur for large bindings , @xmath21 . in the latter case
considerable rates extend up to multiplicities of 8 photons . 0.3 cm we can get an analytic result for small bindings , i.e. large @xmath25 , by obtaining the maximum of the logarithm of eq.([ratio ] ) using sterling s formula ( @xmath26 ) . the maxima are given by the zeros of @xmath27 which for large @xmath25 has the approximate solution @xmath28 and implies a multiplicity of @xmath29 .
present searches of low mass monopoles have been carried out by making use of their magnetic properties trapped in matter , by looking for tracks associated with their ionization properties and by studying two photon decays at colliders @xcite . at colliders like the tevatron or the lhc
the conservation of magnetic charge implies that either monopolium or a pair monopole - antimonopole could be produced .
monopolium , being chargeless , is difficult to detect except for its decay properties .
if monopole - antimonopole are produced close to threshold they might annihilate given their large interaction before reaching the detector .
in this experimental scenario we might not have been able to detect monopolium and/or monopoles because we have been looking into trapped monopoles , ionization remnants or the two photon channel , which according to our present investigation is most unlikely .
our analysis shows that a characteristic feature of monopolium decay is many photons arising from the annihilation vertex . at the monopolium threshold
, these photons will have on average much smaller energy than the energy of a typical collider process .
the same type of argument applies to monopole - antimonopole annihilation . to conclude we state , that as long as the exact dynamics of monopoles and their properties are unknown , large multiplicity of photon events might be the signal for the discovery of these elusive particles . |
one of the important results in the theory of derivations in banach bimodules is the theorem of j. r. ringrose on automatic continuity of every derivation acting on a @xmath1-algebra @xmath0 with values in a banach @xmath0-bimodule @xcite .
this theorem extends the well - known result that every derivation of a @xmath1-algebra @xmath0 is automatically norm continuous @xcite . in the case when @xmath0 is an @xmath2-algebra ( in particular , @xmath3-algebra ) , every derivation on @xmath0 is inner @xcite , @xcite .
significant examples of @xmath3-modules are non - commutative symmetric spaces of measurable operators affiliated with a von neumann algebra . at the present time
the theory of symmetric spaces is actively developed ( see , e.g. @xcite , @xcite ) and it gives useful applications both in the geometry of banach spaces and in the theory of unbounded operators .
every non - commutative symmetric space is a solid linear space in the @xmath4-algebra @xmath5 of all @xmath6-measurable operators affiliated with a von neumann algebra @xmath0 , where @xmath6 is a faithful normal semifinite trace on @xmath0 @xcite .
the algebra @xmath5 equipped with the natural topology @xmath7 of convergence in measure generated by the trace @xmath6 is a complete metrizable topological algebra . in its turn
the algebra @xmath5 represents a solid @xmath4-subalgebra of the @xmath4-algebra @xmath8 of all locally measurable operators , affiliated with a von neumann algebra @xmath0 @xcite , @xcite , in addition , the latter algebra @xmath8 with the natural topology @xmath9 of convergence locally in measure is a complete topological @xmath4-algebra @xcite . in @xcite there
are established significant results in description of derivations of these algebras . in particular
, it is proved that for every @xmath9-continuous derivation @xmath10 acting on the algebra @xmath8 there exists @xmath11 , such that @xmath12 $ ] for all @xmath13 , that is the derivation @xmath10 is inner .
one of the main corollary of this result provides a full description of derivations @xmath10 acting from @xmath0 into a banach @xmath0-bimodule @xmath14 of locally measurable operators affiliated with @xmath0 .
more precisely it is shown that any derivation an a von neumann algebra @xmath0 with values in a banach @xmath0-bimodule @xmath14 is inner @xcite . in the present paper
we establish the similar property of derivations in more general case when @xmath14 is a quasi - normed @xmath0-bimodule of locally measurable operators affiliated with @xmath0 .
the proof proceed in several stages .
firstly , in section 2 we establish the version of ringrose theorem for arbitrary quasi - normed @xmath1-bimodules without the assumption of their completeness .
after that in section 3 we supply the proof of the main result of the present paper ( theorem [ t_mbm ] ) showing that every derivation @xmath15 is necessarily inner , i.e. @xmath16 for some @xmath17 .
in particular , @xmath10 is a continuous derivation from @xmath18 into @xmath19 .
in addition , the operator @xmath20 may be chosen so that @xmath21 , where @xmath22 is the modulus of concavity of the quasi - norm @xmath23 .
we use terminology and notations from the von neumann algebra theory @xcite and the theory of locally measurable operators from @xcite .
let @xmath24 be the @xmath4-algebra of all bounded linear operators acting in a hilbert space @xmath25 , and let @xmath26 be the identity operator on @xmath25 .
let @xmath0 be a von neumann algebra acting on @xmath25 , and let @xmath27 be the centre of @xmath0 . denote by @xmath28 the lattice of all projections in @xmath0 and by @xmath29 the set of all finite projections in @xmath0 .
a densely - defined closed linear operator @xmath30 affiliated with @xmath0 is said to be _
measurable _ with respect to @xmath0 if there exists a sequence @xmath31 such that @xmath32 and @xmath33 for every @xmath34 , where @xmath35 is the domain of @xmath30 and @xmath36 is the set of all natural numbers .
the set @xmath37 of all measurable operators is a unital @xmath4-algebra over the field @xmath38 of complex numbers with respect to strong sum @xmath39 , strong product @xmath40 and the adjoint operation @xmath41 @xcite .
a densely - defined closed linear operator @xmath30 affiliated with @xmath0 is called _ locally measurable _ with respect to @xmath0 if there is a sequence @xmath42 such that @xmath43 and @xmath44 for all @xmath34 .
the set @xmath8 of all locally measurable operators ( with respect to @xmath0 ) is a unital @xmath4-algebra over the field @xmath38 with respect to the same algebraic operations as in @xmath37 , in addition @xmath0 and @xmath37 are @xmath4-subalgebras of @xmath8 . for every subset @xmath45 , the set of all self - adjoint ( respectively , positive ) operators in @xmath46 is denoted by @xmath47 ( respectively , @xmath48 ) .
the partial order in @xmath49 is defined by its cone @xmath50 and is denoted by @xmath51 .
let @xmath30 be a closed linear operator with the dense domain @xmath35 in @xmath25 , let @xmath52 be the polar decomposition of the operator @xmath30 , where @xmath53 and @xmath54 is a partial isometry in @xmath24 such that @xmath55 is the right support @xmath56 of @xmath30 .
it is known that @xmath13 ( respectively , @xmath57 ) if and only if @xmath58 ( respectively , @xmath59 ) and @xmath60 @xcite . for every @xmath13 the projection @xmath61 , where @xmath62 and @xmath56 are left and right supports of @xmath30 respectively ,
is called the support of @xmath30 . if @xmath63 , then the spectral family of projections @xmath64 for @xmath30 belongs to @xmath0 @xcite , in addition , @xmath65 .
denote by @xmath66 the @xmath1-norm in the von neumann algebra @xmath0 .
we need the following property of partial order in the algebra @xmath8 .
* proposition 6.1 ) [ mchext ] let @xmath0 be a von neumann algebra acting on the hilbert space @xmath25 , @xmath67 and @xmath68
. then @xmath69 for some @xmath70 , in particular , @xmath71 .
now , let us recall the definition of the local measure topology .
if @xmath0 is a commutative von neumann algebra , then @xmath0 is @xmath4-isomorphic to the @xmath4-algebra @xmath72 of all essentially bounded measurable complex - valued functions defined on a measure space @xmath73 with the measure @xmath74 satisfying the direct sum property ( we identify functions that are equal almost everywhere ) ( see e.g. ( * ? ? ? * ch .
1 ) ) . the direct sum property of a measure @xmath74 means that the boolean algebra @xmath75 of all projections of a commutative von neumann algebra @xmath72 is order complete , and for any non - zero @xmath76 there exists a non - zero projection @xmath77 such that @xmath78 , where @xmath79 , @xmath80 and @xmath81 , when @xmath82 . consider the @xmath4-algebra @xmath83 of all measurable almost everywhere finite complex - valued functions defined on @xmath73 ( functions that are equal almost everywhere are identified ) .
define on @xmath84 the local measure topology @xmath85 , that is , the hausdorff vector topology , whose base of neighbourhoods of zero is given by @xmath86 @xmath87 where @xmath88 , @xmath89 , @xmath90 .
the topology @xmath85 does not change if the measure @xmath74 is replaced with an equivalent measure @xcite .
now let @xmath0 be an arbitrary von neumann algebra and let @xmath91 be a @xmath4-isomorphism from @xmath27 onto the @xmath4-algebra @xmath72 . denote by @xmath92 the set of all measurable real - valued functions defined on @xmath73 and taking values in the extended half - line @xmath93 $ ] ( functions that are equal almost everywhere are identified ) .
let @xmath94 be a dimensional function on @xmath95 @xcite . for arbitrary scalars
@xmath96 and a set @xmath97 , @xmath90 , we set @xmath98 @xmath99 it was shown in @xcite that the system of sets @xmath100 defines a hausdorff vector topology @xmath9 on @xmath8 such that the sets of form a neighbourhood base of an operator @xmath13 .
the topology @xmath9 on @xmath8 is called the _ local measure topology_. it is known that @xmath101 is a complete topological @xmath4-algebra , and the topology @xmath9 does not depend on a choice of dimension function @xmath102 and on a choice of @xmath4-isomorphism @xmath91 ( see e.g. @xcite , @xcite ) .
let us mention the following important property of the topology @xmath9 .
* proposition 2.5 ) [ p2_2 ] the von neumann algebra @xmath0 is dense in @xmath103 . let @xmath0 be a semifinite von neumann algebra acting on the hilbert space @xmath25 , let @xmath6 be a faithful normal semifinite trace on @xmath0 .
an operator @xmath104 is called _ @xmath6-measurable _ if for any @xmath105 there exists a projection @xmath106 such that @xmath107 and @xmath108 .
the set @xmath5 of all @xmath6-measurable operators is a @xmath4-subalgebra of @xmath37 and @xmath109 .
if the trace @xmath6 is finite , then @xmath110 .
let @xmath7 be the _ measure topology _
@xcite on @xmath5 whose base of neighbourhoods of zero is given by @xmath111 the pair @xmath112 is a complete metrizable topological @xmath4-algebra . here
, the topology @xmath7 majorizes the topology @xmath9 on @xmath5 and , if @xmath6 is a finite trace , the topologies @xmath7 and @xmath9 coincide @xcite .
let @xmath0 be a @xmath1-algebra with identity @xmath26 and @xmath113 be an arbitrary bimodule over @xmath0 with bilinear mappings @xmath114 such that @xmath115 for all @xmath116 . by introduction of algebraic operation @xmath117 , the @xmath0-bimodule @xmath113 become a complex linear space .
recall that a real function @xmath118 on a complex linear space @xmath113 is called a _ quasi - norm _ on @xmath113 , if there exists a constant @xmath119 such that for all @xmath120 the following properties hold : @xmath121 .
@xmath122 ; @xmath123 .
@xmath124 ; @xmath125 . @xmath126 .
the couple @xmath127 is called a quasi - normed space and the least of all constants @xmath128 satisfying the inequality @xmath125 above is called the modulus of concavity of the quasi - norm @xmath118 and denoted by @xmath129 .
every quasi - normed space @xmath127 is a locally bounded hausdorff topological vector space ( see e.g. @xcite ) .
if an @xmath0-bimodule @xmath113 is equipped with a quasi - norm @xmath130 , then the couple @xmath131 is called a quasi - normed @xmath0-bimodule if for all @xmath132 the equality @xmath133 holds , where @xmath66 is the @xmath1-norm on @xmath0 .
let @xmath113 be an @xmath0-bimodule over a @xmath1-algebra @xmath0 .
a linear mapping @xmath134 is called a _ derivation _ , if @xmath135 for all @xmath136 .
a derivation @xmath134 is called _ inner _ , if there exists an element @xmath137 , such that @xmath138=dx - xd$ ] for all @xmath139 .
if @xmath131 is a quasi - normed @xmath0-bimodule , then , by every inner derivation @xmath134 is a continuous linear mapping from @xmath140 into @xmath131 . in the following theorem
we show that every derivation @xmath141 is continuous and strengthen the ringrose theorem by omitting the assumption of completeness of the space @xmath131 .
our proof uses the original proof of j.ringrose ( * ? ? ?
* theorem 2 ) . [ t_ring ]
let @xmath131 be a quasi - normed @xmath0-bimodule .
then every derivation @xmath141 is a continuous mapping from @xmath140 into @xmath131
. set @xmath142 exactly as in ( * ? ? ?
* theorem 2 ) we obtain that @xmath143 is a two - sided ideal in @xmath0 . since @xmath144 we have @xmath145 let us show that @xmath143 is a closed ideal in @xmath0 .
let @xmath146 and @xmath147 as @xmath148 .
then every @xmath149 is a continuous mapping from @xmath0 into @xmath113 and for every fixed @xmath150 we have @xmath151 the principle of uniform boundedness ( see e.g. ( * ? ? ? * theorem 2.8 ) ) implies that @xmath152 is continuous , that is @xmath153 .
thus , @xmath143 is a closed two - sided ideal in @xmath0 .
now , we show that the restriction @xmath154 is continuous .
suppose the contrary .
then there exists a sequence @xmath155 such that @xmath156 set @xmath157
. then @xmath158 by ( * ? ? ?
* lemma 1 ) there exists @xmath159 with @xmath160 such that @xmath161 .
hence , @xmath162 that is the mapping @xmath163 is unbounded .
that contradicts the inclusion @xmath164 .
thus , the restriction @xmath154 is a continuous mapping .
next , we claim that the quotient algebra @xmath165 is finite - dimensional
. assume the contrary and as in ( * ? ? ?
* theorem 2 ) choose a sequence @xmath166 such that @xmath167 since @xmath168 the mapping @xmath169 is unbounded , therefore there exist @xmath170 such that @xmath171 by setting @xmath172 we have @xmath173 and @xmath174 .
the inequality @xmath175 implies that @xmath176 now , using the inequality @xmath177 we obtain that @xmath178 that is a contradiction , thus the quotient algebra @xmath165 is finite - dimensional . therefore , @xmath0 is a direct sum @xmath179 , where @xmath180 is a finite - dimensional subspace in @xmath0 and @xmath143 is a closed ideal in @xmath140 .
since the restrictions @xmath154 and @xmath181 are continuous it follows that the derivation @xmath182 is continuous too .
in this section we establish the main result of the present paper , which give description of all derivations acting on a von neumann algebra @xmath0 with values in a quasi - normed @xmath0-bimodule @xmath14 of locally measurable operators affiliated with @xmath0 .
let @xmath0 be an arbitrary von neumann algebra .
a linear subspace @xmath14 of @xmath8 , is called an @xmath0-bimodule of locally measurable operators if @xmath183 whenever @xmath184 and @xmath185 .
it is clear that @xmath14 is a bimodule over the @xmath1-algebra @xmath0 in the sense of section 3 . if @xmath14 is a @xmath0-bimodule of locally measurable operators , @xmath186 and @xmath187 is the polar decomposition of operator @xmath30 then @xmath188 and @xmath189 .
in addition , by proposition [ mchext ] every @xmath0-bimodule satisfies the following condition @xmath190 if an @xmath0-bimodule of locally measurable operators @xmath14 is equipped with a quasi - norm @xmath191 , satisfying the inequality , then @xmath192 is called a _ quasi - normed @xmath0-bimodule of locally measurable operators_. examples of quasi - normed @xmath0-bimodules of locally measurable operators , which are not normed @xmath0-bimodules , are the noncommutative @xmath193-space @xmath194 associated with a faithful normal semifinite trace @xmath6 for @xmath195 .
recall that two projections @xmath199 are called equivalent ( notation : @xmath200 ) if there exists a partial isometry @xmath150 such that @xmath201 and @xmath202 . for projections
@xmath199 notation @xmath203 means that there exists a projection @xmath204 such that @xmath205 .
let @xmath222 be an arbitrary derivation , that is @xmath10 is a linear mapping such that @xmath135 for all @xmath136 . in (
* lemma 3.1 ) it is proved that @xmath223 and @xmath224 for all @xmath225 . in particular , @xmath226 and the restriction @xmath227 of the derivation @xmath10 to @xmath228
is a derivation on @xmath228 with values in @xmath229 . let @xmath10 be a derivation from @xmath0 with values in a @xmath0-bimodule @xmath14 of locally measurable operators .
let us define a mapping @xmath230 by setting @xmath231 , @xmath232 .
a direct verification shows that @xmath233 is also a derivation from @xmath0 with values in @xmath14 .
a derivation @xmath10 is said to be _ self - adjoint _ , if @xmath234 .
every derivation @xmath10 from @xmath0 with values in a @xmath0-bimodule @xmath14 of locally measurable operators can be represented in the form @xmath235 , where @xmath236 are self - adjoint derivations from @xmath0 with values in @xmath14 . for the proof of the assertion that every derivation @xmath237 is inner
we need the following significant theorem establishing that every @xmath9-continuous derivation acting on the @xmath4-algebra @xmath8 is inner . for application of the theorem [ t1 ] in our case ,
when @xmath10 is a derivation acting on @xmath0 with values in quasi - normed @xmath0-bimodule @xmath238 of locally measurable operators , we need @xmath9-continuity of @xmath10 . in the following proposition [ p4 ] using theorem [ t_ring ] we prove this property of derivation @xmath10 .
let @xmath6 be a faithful normal finite trace @xmath6 on the von neumann algebra @xmath0 .
in this case , the algebra @xmath0 is finite .
moreover , @xmath239 , @xmath240 and @xmath103 is an @xmath241-space @xcite .
since the trace @xmath6 is finite , we have that @xmath240 , and therefore it is sufficient to show that @xmath245 .
suppose the contrary .
passing to a subsequence , if necessary , we may choose @xmath246 such that @xmath247 for all @xmath34 .
set @xmath253 it is clear that @xmath254 . since @xmath255 the set @xmath256 is an orthocomplemented complete modular lattice , and therefore @xmath256 is a continuous geometry @xcite .
in particular , @xmath257 consequently , for all @xmath34 there exists an integer @xmath258 such that @xmath259 .
set @xmath260 .
the sequence of projections @xmath261 is increasing , moreover , @xmath262 and @xmath263 therefore , @xmath264 . by proposition [ pe4 ]
we have @xmath265 since the sequence @xmath266 is increasing , the last inequality implies that @xmath267 for all @xmath34 . using the convergence @xmath264
we obtain @xmath268 , that contradicts .
* lemma 6.8 ) [ l_e4 ] if @xmath0 is a von neumann algebra with a faithful normal finite trace @xmath6 , @xmath270 and @xmath244 , then there exists a subsequence @xmath271 such that @xmath272 , where @xmath273 and @xmath274 .
[ p4 ] let @xmath0 be a von neumann algebra with a faithful normal finite trace @xmath6 , let @xmath238 be a quasi - normed @xmath0-bimodule of locally measurable operators and let @xmath10 be a derivation on @xmath8 such that @xmath275 .
then the derivation @xmath10 is @xmath9-continuous . since the restriction @xmath281 of the derivation @xmath10 to the von neumann algebra @xmath0 is a derivation from @xmath0 into the quasi - normed @xmath0-bimodule @xmath238 , by theorem [ t_ring ] we have @xmath282 .
lemma [ l_e2 ] implies that @xmath283 .
* theorem 1 ) [ t_bs ] let @xmath0 be a von neumann algebra and @xmath287 .
then there exist a self - adjoint operator @xmath288 in the centre of the @xmath4-algebra @xmath8 and a family @xmath289 of unitary operators from @xmath0 such that @xmath290 let @xmath0 be a von neumann algebra and let @xmath238 be a quasi - normed @xmath0-bimodule of locally measurable operators .
then any derivation @xmath15 is inner , that is there exists an element @xmath20 such that @xmath138,x\in\mathcal{m}$ ] , in addition @xmath21 .
if @xmath291 or @xmath292 then @xmath293 may be chosen so that @xmath294 . choose pairwise orthogonal central projections @xmath297 such that @xmath298 + @xmath299 , @xmath300 is a properly infinite von neumann algebra and on every von neumann algebra @xmath301 exists a faithful normal finite trace . by (
* theorem 3.3 ) the derivation @xmath302 is @xmath303-continuous .
since the von neumann algebra @xmath301 is finite with a faithful normal finite trace and for the derivation @xmath304 the inclusion @xmath305 holds , proposition [ p4 ] implies that @xmath306 is also @xmath307-continuous for all @xmath308 .
therefore , by ( * ? ? ?
* corollary 2.8 ) , the derivation @xmath295 is @xmath9-continuous . by theorem [ t1 ]
the derivation @xmath295 is inner , that is there exists an element @xmath11 , such that @xmath309 $ ] for all @xmath13 .
it is clear that @xmath310=\overline{\delta}(\mathcal{m})=\delta(\mathcal{m})\subset \mathcal{e}$ ] .
let us show that we can choose @xmath20 such that @xmath138 $ ] for all @xmath139 .
by theorem [ t_bs ] and taking @xmath315 in ( [ e_bs ] ) we obtain that there exist self - adjoint operators @xmath316 in the centre of the @xmath4-algebra @xmath8 and unitary operators @xmath317 such that @xmath318|\geq |a_i - c_i|,\ i=1,2.\ ] ] since @xmath319\in\mathcal{e}$ ] and @xmath14 is @xmath0-bimodule of locally measurable operators we have that @xmath320 , @xmath321 ( see ) . therefore @xmath322 .
since @xmath316 are central elements from @xmath8 it follows that @xmath323=[d , x]$ ] for all @xmath139 .
now , suppose that @xmath291 . in this case ,
@xmath324=[d , x]-[d , x^*]^*=\delta(x)-(\delta(x^*))^*=\delta(x)-\delta^*(x)=0 $ ] for any @xmath139 .
consequently , the operator @xmath325 commutes with every elements from @xmath0 , and by proposition [ p2_2 ] , @xmath326 is a central element in the algebra @xmath8 .
therefore we may suggest that @xmath138,\ x\in\mathcal{m}$ ] , where @xmath327 . according to theorem [ t_bs ]
there exist @xmath328 from the centre of the algebra @xmath8 and a family @xmath289 of unitary operators from @xmath0 such that @xmath329|\geq ( 1-\varepsilon)|a - c|.\ ] ] for @xmath330 and @xmath315 we have @xmath331|=2|[-id , u_{1/2}]|=2[d , u_{1/2}]\in\mathcal{e}.\ ] ] consequently , @xmath332 , moreover , @xmath333=[ia , x]=[b , x]\ ] ] for all @xmath139 .
since @xmath334|=|[d , u_\varepsilon]|=|\delta(u_\varepsilon)|,\ ] ] it follows that @xmath335 for all @xmath105 , that implies the inequality @xmath336 .
now , suppose that @xmath339 and @xmath340 .
equality ( [ e1 ] ) implies that @xmath341 consequently , @xmath342 similarly , @xmath343 . since @xmath344 , @xmath345
, there exist @xmath346 , such that @xmath347,\ im(\delta)(x)=[d_2,x]$ ] for all @xmath139 and @xmath348 , @xmath321 .
taking @xmath349 , we have that @xmath20 , @xmath350+i[d_2,x]=[d , x]$ ] for all @xmath139 , in addition @xmath21 .
let @xmath0 be a semifinite von neumann algebra and let @xmath6 be a faithful normal semifinite trace on @xmath0 .
let @xmath5 be the @xmath4-algebra of all @xmath6-measurable operators affiliated with @xmath0 .
for each @xmath351 and @xmath352 we define the decreasing rearrangement ( or generalised singular value function ) by setting @xmath353 let @xmath14 be a linear subspace in @xmath5 equipped with a quasi - norm @xmath196 satisfying the following condition : @xmath354 in this case , the pair @xmath19 is called _ quasi - normed symmetric spaces of @xmath6-measurable operators_. it is easy to see that every quasi - normed symmetric space of measurable operators is a quasi - normed @xmath0-bimodule of locally measurable operators , and therefore theorem [ t_mbm ] implies the following [ c_last ] let @xmath19 be a quasi - normed symmetric spaces of @xmath6-measurable operators , affiliated with a semifinite von neumann algebra @xmath0 with a faithful semifinite normal trace @xmath6 .
then any derivation @xmath355 is continuous and there exists @xmath20 such that @xmath138 $ ] for all @xmath139 and @xmath356 p.g .
dodds , t.k .
dodds and b. de pagter , _ noncommutative banach function spaces _
z. , * 201 * ( 1989 ) , 583597 .
kadison and j.r .
ringrose , _ fundamentals of the theory of operator algebras ii _ , academic press , orlando , 1986 . n.j .
kalton , n.t .
peck , j.w .
roberts , _ an f - space sampler , _ london math .
lecture notes 89 .
cambridge univ . press .
1985 . n.j .
kalton , f.a .
sukochev , _ symmetric norms and spaces of operators .
_ j. reine angew .
, * 621 * ( 2008 ) , 81121 .
i.kaplansky , _ any orthocomplemented complete modular lattice is continuous geometry _ , ann
, * 61 * ( 1955 ) , 524542 .
d.olesen , _ derivations of @xmath2-algebras are inner _ , pacific j. math .
, * 53 * ( 1974 ) , no.1 , 555561 . j. r.ringrose , _ automatic continuity of derivations of operator algebras _ , j.london math
. soc . , * 5 * ( 1972 ) , no .
2 , 432438 . | we prove that every derivation acting on a von neumann algebra @xmath0 with values in a quasi - normed bimodule of locally measurable operators affiliated with @xmath0 is necessarily inner . |
classical cepheids are fundamental standard candles to determine both galactic and extragalactic distances .
they obey well defined optical and near infrared period luminosity and period luminosity colour relations .
thanks to the unprecedented spatial resolution of the hubble space telescope , cepheids can be used as primary distance indicators up to the virgo cluster ( @xmath7 25 mpc ) .
this means that they can also be adopted to provide robust absolute calibrations of secondary distance indicators , and in turn to estimate the value of the hubble constant ( see e.g. * ? ? ?
the accuracy reached in the determination of cosmological parameters using the cepheid distance scale relies on the knowledge of the physical mechanisms that govern their radial oscillations and on their dependence on intrinsic parameters ( luminosity , mass , radius , chemical composition ; see e.g. @xcite , and references therein ) .
in particular , the estimate of cepheid radii and of the effective temperature ( using mean intrinsic colours ) allows us to constrain their intrinsic luminosity .
furthermore , the cepheid mass can be estimated by adopting a period mass radius relation ( see e.g. * ? ? ? * and references therein ) , and in turn to provide independent constraints on the mass - luminosity relations .
the methods currently used to derive cepheid radii from photometric and spectroscopic data ( radial velocity ) are based on the classical baade wesselink technique @xcite . among them , the surface brightness technique @xcite and the cors method @xcite rely on solid physical bases .
the reader interested in a detailed discussion concerning the different baade wesselink methods is referred to @xcite .
the use of interferometric methods and the measurement of parallaxes provide a model - independent method for radius measurements .
however , this approach has only been applied to a very limited sample of bright nearby cepheids @xcite . on the the other hand , the baade
wesselink methods can be applied to large samples of cepheids and it has been recently extended to magellanic cepheids @xcite . however , the precision of baade wesselink distances strongly depends on the quality of photometric and radial velocity data . in a previous paper @xcite , we adapted the cors method to derive cepheid radii observed in the strmgren system .
however the model atmospheres adopted in the quoted investigation to estimate the surface brightness did not allow us to include cepheids with periods longer than 1213 days .
the cutoff in the long period range limited the accuracy of the period radius relation .
recently , @xcite provided new calibrations based on a new release of model atmospheres by @xcite for the walraven photometry of galactic cepheids originally collected by @xcite .
the walraven system includes five bands , ( v , b , l , u , w ) , designed to measure the features of hydrogen spectrum .
three of them have the central wavelength in the ultraviolet region ( see * ? ? ?
* ; * ? ? ?
* for more details about this photometric system ) .
in particular the central wavelength of the walraven filters are the following : @xmath8 . in this investigation
we show that this new calibration of the vbluw cepheid time series data by @xcite allows us to derive precise stellar parameters ( surface gravity and effective temperature ) at fixed chemical composition and to calibrate with high accuracy the surface brightness function , a fundamental ingredient for the application of the baade wesselink method .
this opportunity stimulated us to apply the cors version of the baade
wesselink technique as formulated by @xcite and @xcite , to derive reliable radii for a sample of cepheids having periods from a few days up to @xmath9 days and to obtain more accurate period radius and period luminosity relations .
the layout of the paper is the following : the original cors method is introduced in sec .
[ sec - cors ] .
the grids of atmosphere models and the calibration of the surface brightness in the walraven system are described in sec .
[ sec - cors - modif ] . in sec .
[ sec - data ] we describe the adopted photometric and spectroscopic data , while the estimate of the cepheids radii and of the period radius relation are discussed in secs .
[ sec - radii ] and [ sec - pr ] , respectively .
in sec . [ sec - pl ]
we discuss the derivation of the period
luminosity relation .
finally , sec .
[ sec - conclusion ] summarizes the main results of this investigation .
the original cors method @xcite is a variant of the classical baade wesselink technique @xcite and was developed to derive the radii of pulsating stars .
it relies on the surface brightness function : @xmath10 where @xmath11 is the apparent visual magnitude and @xmath12 is the angular diameter of the star . for a variable star eq .
( [ eq - sb ] ) can be written for each phase @xmath13 along the pulsational cycle . by differentiating the above equation with respect to the phase , by multiplying the result with a colour index @xmath14 , where @xmath15,@xmath16 are two generic photometric bands ) and by integrating over the entire cycle
, one can write : @xmath17 where @xmath18 , @xmath19 is the period , @xmath20 is the radial velocity and @xmath21 is the radial velocity projection factor .
the other two terms , @xmath22 and @xmath23 , are the following : @xmath24 the projection factor @xmath21 correlates radial and pulsation velocity , i.e. @xmath25 . the uncertainty affecting the estimates of
this parameter is still the main source of systematic errors in the baade wesselink method .
we shall discuss the @xmath21 value in sec .
[ sec - pfactor ] . by solving eq .
( [ eq - cors ] ) we estimate the value of the radius @xmath26 at an arbitrary phase @xmath27 , while the radius at any phase @xmath28 along the pulsational cycle can be evaluated by integrating the radial velocity curve between @xmath27 and @xmath28 . the term @xmath22 present in eq .
( [ eq - cors ] ) can be easily determined using colour and magnitude ( @xmath22 ) , while the term @xmath23 is not directly connected to observational data , but approximates the area of the loop performed by variable stars in the colour - colour plane .
typically the @xmath23 term has a small value ( @xmath29 , * ? ? ? * ) and in the original baade wesselink method it is neglected ( see * ? ? ?
* ) . however , the cepheid radii estimated by including the @xmath23 term in the eq .
( [ eq - cors ] ) are more accurate than those based on the original baade wesselink method @xcite . in the next section
we describe how to calibrate the surface brightness function to estimate the term @xmath23 from eq .
( [ eq - db ] ) .
in order to calibrate the surface brightness function , we assume the validity of the quasi static approximation ( qsa ) for the cepheid atmosphere ( see also * ? ? ? * ) . under this hypothesis ,
any photometric quantity can be expressed as a function of effective temperature and gravity @xmath30 , @xmath31 .
then , we can write : @xmath32 for the surface brightness , and : @xmath33 for two intrinsic colour indices .
if the determinant of the jacobian @xmath34 the last two equations can be inverted and we obtain : @xmath35 and we can also write the surface brightness as a function of the two colour indices : @xmath36 in general , the invertibility condition is not valid over the entire parameter space , since the same pair of colours trace different pairs of gravity and temperature ( @xmath37 ) . figure [ fig - griglia ] shows the colour degeneracy present in the grid of atmosphere model provided by @xcite ( _ http://wwwuser.oat.ts.astro.it/castelli/colors/vbluw.html_ ) , for the walraven two - colour diagram @xmath38 vs @xmath39 .
however , a local solution can be found , provided that an appropriate choice of the colours @xmath40 and @xmath41 and of the range in colour is made . by using the quoted colours @xmath38 and @xmath39 , we succeeded in inverting eq .
( [ eq - direct ] ) for the range of parameters typical of cepheids , i.e. @xmath42 dex and @xmath43 k. , @xmath39 colours : lines of constant temperature and constant gravity are plotted.,width=317 ] to verify that real cepheids are located inside the theoretical grid , we plotted the loops , obtained by fitting the observed colours of selected cepheids , in the quoted colour colour plane .
the result for three cepheids with short- , medium- and long - period , is shown in fig .
[ fig - griglia - loop ] .
a glance at the data plotted in this figure shows that the three loops performed by actual cepheids are located inside the limits of the predicted grid .
@xcite and @xcite pointed out that the flux in the @xmath44 band was measured with sufficient accuracy only for the brightest cepheids due to the intrinsic faintness of cepheid fluxes in the near - ultraviolet , but also because the telescope mirror reflectivity in this band was deteriorated at the epoch of observations .
therefore , we decided to perform a further test by using the pair of colours @xmath38 and @xmath45 , even if the curves of constant temperature and gravity are not orthogonal as in the plane with the @xmath39 colour .
accurate solutions can be obtained in the range @xmath46 k and @xmath47 dex .
however , the loops of cepheids with periods longer than 11 days fall outside the grid of models .
this means that the structural parameters of cepheids and the surface brightness function are less accurate using this colour colour plane .
nevertheless , the ratios of the cepheid radii with @xmath48 days , estimated using the two different colour colour planes , show that the differences are smaller than @xmath49 . on the basis of this evidence
, we decided to focus our attention on the @xmath50 plane ( see appendix [ appb ] for more details ) .
days ) , rz vel ( @xmath51 days ) and l car ( @xmath52 days).,width=317 ] the availability of a region across the theoretical grid covering the locus occupied by the cepheid colour colour loops and the validity of the invertibility condition shown in the above section provide us with an instrument to write effective temperature and gravity as a function of the two colour indices and , finally to calibrate the surface brightness through eq .
( [ eq - sb - colour ] ) . as a first step ,
we have fitted the theoretical grid shown in fig .
[ fig - griglia - loop ] by means of polynomials , in order to express temperature and gravity as a function of @xmath38 and @xmath39 colours , according to eq .
( [ eq - inverse ] ) ( see appendix [ appa ] for details about the fitted equations ) . the surfaces @xmath54 and @xmath55 are shown in fig .
[ fig - fit - t ] , [ fig - fit - g ] .
the rms around the fitted equations amount to 0.0033 dex and 0.15 dex for @xmath56 and @xmath57 , respectively . as a function of @xmath38 and @xmath39 , obtained from the polynomial fit of the atmosphere models ( crosses ) from @xcite.,width=340 ] , but for @xmath57.,width=340 ] following the walraven convention for magnitudes , the surface brightness can be derived in the following way : @xmath58 where @xmath59 is the bolometric correction calculated as a function of @xmath60 and @xmath61 through a polynomial fit ( see appendix [ appa ] ) . a plot of the fit of @xmath59 as a function of @xmath56 and @xmath57 is shown in fig .
[ fig - fit - bc ] .
, as a function of @xmath56 and @xmath57 , obtained from the polynomial fit of atmosphere models ( crosses ) from @xcite.,width=340 ] the procedure described here allows us to calculate the term @xmath62 in the complete cors formulation ( eq . [ eq - cors ] ) . in the previous section
we have described the procedure followed to derive the physical parameters of cepheids starting from the measured colours . in order to test the accuracy of the fitted relations ( eq . [ eq - inverse ] ) , we have performed a comparison of the effective temperature and gravity derived from the fitted relations for selected cepheids in our sample with the results obtained by other authors using an independent spectroscopic approach . as an example , in fig .
[ fig - test ] we have plotted the values of effective temperature and gravity obtained with our equations for the pulsational cycle of the cepheid u sgr . in the same figure we also overplotted the values of the physical parameters obtained spectroscopically by @xcite using the line depth ratios . we note that the curve of effective temperature follows the spectroscopic data with great accuracy .
the agreement is less satisfactory for the curve of gravity , probably as a consequence of the cited problems with the @xmath44 band .
similar results are obtained for another dozen of stars for which the comparison with spectroscopic data was possible .
we can conclude , that our approximation for the effective temperature is quite accurate , with mean relative differences of @xmath63 . on the other hand
, we can reproduce the effective gravity , @xmath57 , with mean relative differences of @xmath64 . however , the larger error affecting the estimate of the surface gravity is not a thorny problem in the estimate of cepheid radii , because the gravity enters only in the determination of the @xmath65b term , through the contribution of the bc to the surface brightness ( @xmath53 ) .
the @xmath65b term contributes at most with a 5% to the radius , therefore , we can conclude that even a significant uncertainty in our determination of the gravity has a minimal impact on the determination of cepheid radii .
days ) . the open blue circles with the associated error bars represent the values of the physical parameters obtained spectroscopically by @xcite.,width=317 ] as a further test of the accuracy of our approximation for the surface brightness @xmath53 , we have compared the angular diameter @xmath66 of l car ( which is the only single cepheid for which this comparison is possible ) obtained through interferometric measurements @xcite with those obtained from the surface brightness equation : @xmath67 where @xmath28 is the phase along the pulsational cycle and @xmath66 is given in mas ( see fig .
[ fig - test2 ] ) .
a glance at the figure shows that current solution shows a systematic difference with the interferometric measurements in the phase range between 0.2 and 0.45 .
this discrepancy is due to the limited accuracy of the calibration we are using to estimate the surface gravity .
typically , current calibration underestimates the gravity , in the above phase interval , and , in turn overestimates the radius .
the reader interested in a thorough analysis of this problem is referred to @xcite .
he suggested that the main culprit of the above discrepancy is the fact that we use grid of atmosphere models that neglect the microturbolence variation along the pulsation cycle .
furthermore , the problem in the @xmath44-band reflectivity , described above , also affects the surface gravity calibration .
however , it is worth mentioning that a detailed analysis of the residuals between current estimates and interferometric measurements indicate a mean difference and a dispersion of the order of 3% .
assuming that the interferometric measurements are minimally affected by systematic errors , we can use this average difference to estimate our systematic indetermination on @xmath53 through eq .
[ eq - sb ] , obtaining a value of @xmath68 .
recalling that @xmath53 enters only in the determination of the @xmath65b , which in turn is only @xmath69 5% of the radius , we found that the systematic uncertainty on our estimate of @xmath53 implies a 0.4% uncertainty on the radius .
( in mas ) of l car , obtained from the surface brightness calibration using eq .
[ eq - ang - diam ] ( black points ) , is plotted with the interferometric measurements obtained by @xcite ( red points ) and by @xcite ( blue points ) .
bottom panel : difference between the two curves the mean and standard deviation of such a difference is labelled.,width=317 ] the agreement between the physical parameters obtained with our procedure and those obtained independently from spectroscopy and interferometry seems a valuable test of the accuracy in the derivation of the surface brightness and of the @xmath23 term , thus supporting the soundness of our approach .
after the validation of the quoted approach , we applied it to real data in order to obtain the radii of a sample of classical galactic cepheids .
the photometric data and the radial velocities used in this work are described in the following .
the original sample we have used in this work includes 175 cepheids observed in the walraven vbluw photometric system during 1962 and 197071 at the leiden southern station ( south africa ) @xcite .
it is @xmath70 complete for all cepheids brighter than @xmath71 magnitude at minimum of the light curve and with declination lower than @xmath72 .
we have selected all the type i cepheids ( 163 stars ) included in this sample , but we have retained only those stars for which reliable radial velocity data were available in the literature ( see next section ) .
furthermore , we have rejected also those stars having the colour colour loop at least in part outside the grid of models .
this behaviour can be observed for cepheids having an extremely large loop or a loop with a peculiar orientation in the colour colour plane ( abnormal loop orientations can be caused by the presence of companions or by peculiar chemical abundances ) .
a complete list of the selected stars , with the photometric parameters used in this work , is shown in tab .
[ tab - stars ] . [ cols=">,^,^,^,<",options="header " , ] [ tab - coeff ]
figure [ fig - griglia - lu - loop ] shows the grid of models in the plane @xmath73 for the parameter range @xmath46 and @xmath47 . according to the procedure described in appendix [ appa ]
, we performed a fit of the grid using polynomial equations in the variable @xmath38 and @xmath45 . for brevity
, we do not give the explicit expressions of the fit , but we recall that the rms of residuals around the fit are 0.0023 and 0.14 for @xmath56 and @xmath57 respectively . in the same graphic the loops of the cepheids l car ( @xmath74 days ) and x pup ( @xmath75 days ) are also plotted .
since part of the loops falls outside of the grid , we are not able to accurately estimate the structural parameters of the two stars for all phases of the pulsational cycle .
the same behaviour is observed for most cepheids with period longer than 11 days .
it would be interesting to investigate about this finding in a subsequent work .
a possible reason may be that the l band covers an extremely line - rich region of the spectrum .
apart from a very large number of metallic lines , it includes the higher balmer lines and molecular lines of cn and ch .
for this reason , l may be more sensitive to dynamic effects ( emission components due to shockwaves , enhanced and time variable microturbulence ) not included in the physics at the base of models , hence more sensitive to breakdown of the quasi - hydrostatic equilibrium assumption .
the presence of molecular lines in l may also mean that the kurucz / castelli model atmospheres have shortcomings at the cool end of the grids that become particularly pronounced in l. using only the cepheids with loops fully contained on the grids , we have estimated their radii comparing them with old values obtained using the @xmath39 colour .
the ratios of their values , plotted in fig .
[ fig - ratios ] , show that the differences are smaller than @xmath49 , with mean ratio equal to 0.98 and a standard deviation of 0.03 .
we have also overplotted the new radii on the period radius relation ( eq .
[ eq - pr - xsct - db ] ) and they seem to lie on the fitted relation as shown in fig .
[ fig - overplot ] .
colour and the old values for a sample of singular cepheids with colour - colour loop fully contained on the @xmath73 grid of models .
the mean ratio and the standard deviation , indicated in red , are respectively @xmath76 and @xmath77.,width=317 ] | we present a new derivation of the cors baade wesselink method in the walraven photometric system .
we solved the complete baade wesselink equation by calibrating the surface brightness function with a recent grid of atmosphere models .
the new approach was adopted to estimate the mean radii of a sample of galactic cepheids for which are available precise light curves in the walraven bands .
current radii agree , within the errors , quite well with cepheid radii based on recent optical and near infrared interferometric measurements .
we also tested the impact of the projection factor on the period radius relation using two different values ( @xmath0 , @xmath1 ) that bracket the estimates available in the literature .
we found that the agreement of our period radius relation with similar empirical and theoretical period
radius relations in the recent literature , improves by changing the projection factor from @xmath0 to @xmath1 .
our period radius relation is @xmath2 , with a @xmath3 dex .
thanks to accurate estimates of the effective temperature of the selected cepheids , we also derived the period luminosity relation in the v band and we found @xmath4 with @xmath5 mag , for @xmath1 .
it agrees quite well with recent results in the literature , while the relation for @xmath0 deviates by more than @xmath6 .
we conclude that , even taking into account the intrinsic dispersion of the obtained period luminosity relations , that is roughly of the same order of magnitude as the effect of the projection factor , the results of this paper seem to favour the value @xmath1 .
[ firstpage ] cepheids baade wesselink radii period
radius relation
walraven photometry . |
as summarized by norberg @xcite , an insurance company which invests its wealth in a financial market is exposed to two kinds of risk . the first kind
, called insurance risk , is the traditional liability risk caused by insurance claims , and the second kind , called financial risk , is the asset risk related to risky investments .
the interplay of the two risks unavoidably leads to a complicated stochastic structure for the wealth process of the insurance company .
paulsen @xcite proposed a general continuous - time risk model in which the cash flow of premiums less claims is described as a semimartingale and the log price of the investment portfolio as another semimartingale .
since then the study of ruin in the presence of risky investments has experienced a vital development in modern risk theory ; some recent works include paulsen @xcite , klppelberg and kostadinova @xcite , heyde and wang @xcite , hult and lindskog @xcite , bankovsky _ et al . _
@xcite , and hao and tang @xcite . during this research ,
much attention has been paid to an important special case of paulsen s set - up , the so - called bivariate lvy - driven risk model , in which the two semimartingales are independent lvy processes fulfilling certain conditions so that insurance claims dominate financial uncertainties .
a well - known folklore says that risky investments may impair the insurer s solvency just as severely as do large claims ; see norberg @xcite , kalashnikov and norberg @xcite , frolova _ et al . _
@xcite , and pergamenshchikov and zeitouny @xcite . in this paper
, we describe the insurance business in a discrete - time risk model in which the two risks are quantified as concrete random variables .
this discrete - time risk model builds an efficient platform for investigating the interplay of the two risks .
the ruin probabilities of this model have been investigated by nyrhinen @xcite , tang and tsitsiashvili @xcite , collamore @xcite , and chen @xcite , among many others .
concretely , for each @xmath3 , denote by @xmath4 the insurer s net loss ( the total amount of claims less premiums ) within period @xmath5 and by @xmath6 the stochastic discount factor ( the reciprocal of the stochastic accumulation factor ) over the same time period .
then the random variables @xmath7 , @xmath8 , and @xmath9 , @xmath10 , represent the corresponding insurance risks and financial risks , respectively . in this framework , we consider the stochastic present values of aggregate net losses specified as @xmath11 and consider their maxima @xmath12 if @xmath13 , @xmath14 , form a sequence of independent and identically distributed ( i.i.d . )
random pairs fulfilling @xmath15 and @xmath16 , then , by lemma 1.7 of vervaat @xcite , @xmath17 converges almost surely ( a.s . ) as @xmath18 . in this case , denote by @xmath19 the a.s . limit .
clearly , @xmath20 is non - decreasing in @xmath5 and @xmath21 thus , if @xmath22 and @xmath23 , then @xmath20 also converges a.s . to a limit , denoted by @xmath24 , as @xmath18 .
we conduct risk analysis of the insurance business through studying the tail probabilities of @xmath17 and @xmath20 for @xmath25 .
the study of tail probabilities is of fundamental interest in insurance , finance , and , in particular , quantitative risk management .
moreover , the tail probability of @xmath20 with @xmath26 is immediately interpreted as the finite - time or infinite - time ruin probability . in most places of the paper ,
we restrict ourselves to the standard framework in which @xmath7 , @xmath8 , form a sequence of i.i.d .
random variables with generic random variable @xmath0 and common distribution @xmath27 on @xmath28 , @xmath9 , @xmath10 , form another sequence of i.i.d .
random variables with generic random variable @xmath2 and common distribution @xmath29 on @xmath30 , and the two sequences are mutually independent . under the assumption that the insurance risk @xmath0 has a regularly - varying tail dominating that of the financial risk @xmath2 , tang and tsitsiashvili @xcite obtained some precise asymptotic formulas for the finite - time and infinite - time ruin probabilities .
the dominating relationship between @xmath0 and @xmath2 holds true if we consider the classical black scholes market in which the log price of the investment portfolio is modelled as a brownian motion with drift and , hence , @xmath2 has a lognormal tail , lighter than every regularly - varying tail .
however , empirical data often reveal that the lognormal model significantly underestimates the financial risk .
it shows particularly poor performance in reflecting financial catastrophes such as the recent great recession since 2008 .
this intensifies the need to investigate the opposite case where the financial risk @xmath2 has a regularly - varying tail dominating that of the insurance risk @xmath0 . in this case , the stochastic quantities in ( [ s ] ) and ( [ m ] ) become much harder to tackle with the difficulty in studying the tail probability of the product of many independent regularly - varying random variables .
tang and tsitsiashvili @xcite gave two examples for this opposite case illustrating that , as anticipated , the finite - time ruin probability is mainly determined by the financial risk .
chen and xie @xcite also studied the finite - time ruin probability of this model and they obtained some related results applicable to the case with the same heavy - tailed insurance and financial risks . in this paper , under certain technical conditions ,
we give a unified treatment in the sense that no dominating relationship between the two risks is required .
that is to say , the obtained formulas hold uniformly for the cases in which the insurance risk @xmath0 is more heavy - tailed than , less heavy - tailed than , and equally heavy - tailed as the financial risk @xmath2 . in our main result , under the assumption that every convex combination of @xmath31 and @xmath29 is of strongly regular variation ( see definition [ srv ] below ) , we derive some precise asymptotic formulas for the tail probabilities of @xmath17 and @xmath20 for @xmath26 .
all the obtained formulas appear to be linear combinations of @xmath32 and @xmath33 .
hence , if one of @xmath32 and @xmath34 dominates the other , then this term remains in the formulas but the other term is negligible ; otherwise , both terms should simultaneously present .
these formulas are in line with the folklore quoted before , confirming that whichever one of the insurance and financial risks with a heavier tail plays a dominating role in leading to the insurer s insolvency . in the rest of this paper , section [ sec2 ] displays our results and some related discussions after introducing the assumptions , section [ sec3 ] prepares some necessary lemmas , and section [ sec4 ] proves the results .
throughout this paper , all limit relationships hold for @xmath35 unless otherwise stated . for two positive functions @xmath36 and @xmath37 , we write @xmath38 or @xmath39 if @xmath40 , write @xmath41 if both @xmath38 and @xmath42 , and write @xmath43 if both @xmath44 and @xmath45 . for a real number @xmath46
, we write @xmath47 and @xmath48 .
we restrict our discussions within the scope of regular variation .
a distribution @xmath49 on @xmath50 is said to be of regular variation if @xmath51 for all @xmath46 and the relation @xmath52 holds for some @xmath53 . in this case
, we write @xmath54 .
however , such a condition is too general to enable us to derive explicit asymptotic formulas for the tail probabilities of the quantities defined in ( [ s ] ) and ( [ m ] ) . to overcome this difficulty ,
our idea is to employ some existing results and techniques related to the well - developed concept of convolution equivalence .
a distribution @xmath55 on @xmath56 is said to be convolution equivalent if @xmath57 for all @xmath46 and the relations @xmath58 and @xmath59 hold for some @xmath60 , where @xmath61 stands for the @xmath62-fold convolution of @xmath55 .
more generally , a distribution @xmath55 on @xmath50 is still said to be convolution equivalent if @xmath63 is . in this case
, we write @xmath64 .
relation ( [ l(alpha ) ] ) itself defines a larger class denoted by @xmath65 .
it is known that the constant @xmath66 in relation ( [ s(alpha ) ] ) is equal to @xmath67 see cline @xcite and pakes @xcite .
we shall use the notation @xmath68 as above for the moment generating function of a distribution @xmath55 throughout the paper .
the class @xmath69 coincides with the well - known subexponential class .
examples and criteria for membership of the class @xmath70 for @xmath71 can be found in embrechts @xcite and cline @xcite .
note that the gamma distribution belongs to the class @xmath65 for some @xmath71 but does not belong to the class @xmath70 .
hence , the inclusion @xmath72 is proper .
recent works in risk theory using convolution equivalence include klppelberg _ et al . _
@xcite , doney and kyprianou @xcite , tang and wei @xcite , griffin and maller @xcite , griffin _ et al . _
@xcite , and griffin @xcite . for a distribution @xmath49 on @xmath50 ,
define @xmath73 which is still a proper distribution on @xmath50 .
actually , if @xmath74 is a real - valued random variable distributed as @xmath49 , then @xmath55 denotes the conditional distribution of @xmath75 on @xmath76 . for every @xmath77 ,
it is clear that @xmath78 if and only if @xmath79 .
we now introduce a proper subclass of the class @xmath80 .
[ srv]a distribution @xmath49 on @xmath50 is said to be of strongly regular variation if @xmath55 defined by ( [ u v ] ) belongs to the class @xmath81 for some @xmath60 . in this case
, we write @xmath82 .
examples and criteria for membership of the class @xmath83 can be given completely in parallel with those in embrechts @xcite and cline @xcite .
this distribution class turns out to be crucial for our purpose . clearly , if @xmath84 follows @xmath85 for some @xmath60 , then @xmath86 since the conditional distribution of @xmath75 on @xmath76 belongs to the class @xmath70 .
our standing assumption is as follows : [ assumption a]every convex combination of @xmath31 and @xmath29 , namely @xmath87 for @xmath88 , belongs to the class @xmath89
. some interesting special cases of assumption [ assumption a ] include : a. @xmath90 and @xmath91 ; or , symmetrically , @xmath92 and @xmath93 . b. @xmath90 , @xmath94 , and @xmath95 ; or , symmetrically , @xmath96 , @xmath97 , and @xmath98 . c. @xmath90 , @xmath99 , and the function @xmath100 is @xmath101-regularly varying ( that is to say , @xmath102 for every @xmath103 ) . for ( a ) and ( b ) , recall a fact that , if @xmath104 , @xmath105 , and @xmath106 , then @xmath107 and @xmath108 are equivalent ; see theorem 2.1(a ) of klppelberg @xcite and the sentences before it .
this fact can be restated as that , if @xmath109 , @xmath110 , and @xmath111 , then @xmath112 and @xmath113 are equivalent . by this fact , the verifications of ( a ) and ( b ) are straightforward . for ( c ) , by theorem 2.0.8 of bingham _ et al . _
@xcite , the relation @xmath102 holds uniformly on every compact @xmath114-set of @xmath30 .
then the verification can be done by using theorems 3.4 and 3.5 of cline @xcite . in this subsection , we assume that @xmath115 and @xmath116 are two independent sequences of i.i.d .
random variables with @xmath0 distributed as @xmath31 on @xmath50 and @xmath2 as @xmath29 on @xmath30 . under assumption
[ assumption a ] , by lemma [ c2 ] below ( with @xmath117 ) , we have @xmath118 note that both @xmath119 and @xmath120 are finite under assumption [ assumption a ] .
the moments of @xmath2 will appear frequently in the paper , so we introduce a shorthand @xmath121 for @xmath60 to help with the presentation . starting with this asymptotic formula and proceeding with induction , we shall show in our main result that the relations @xmath122 and @xmath123 hold for every @xmath124 , where the coefficients @xmath125 , @xmath126 , and @xmath127 are given by @xmath128 furthermore , we shall seek to extend relations ( [ r1 ] ) and ( [ r2 ] ) to @xmath129 . for this purpose , it is natural to assume @xmath130 ( which excludes the case @xmath131 ) to guarantee the finiteness of the constants @xmath132 , @xmath133 , and @xmath134 . note in passing that @xmath130 implies @xmath135 , which is an aforementioned requirement for @xmath19 and @xmath24 to be a.s .
straightforwardly , @xmath136 it is easy to see that @xmath137 actually , when @xmath138 we use the elementary inequality @xmath139 for any nonnegative sequence @xmath140 , and when @xmath141 we use minkowski s inequality . in order for @xmath142 to be a.s .
finite , we need another technical condition @xmath143 .
the finiteness of @xmath144 can be verified similarly to ( [ bound of moment ] ) . applying the dominated convergence theorem to the expressions for @xmath126 and @xmath127
, we obtain @xmath145 now we are ready to state our main result , whose proof is postponed to sections [ proof of ( a)][proof of ( c ) ] .
[ no1]let @xmath115 and @xmath146 be two independent sequences of i.i.d .
random variables with @xmath0 distributed as @xmath31 on @xmath50 and @xmath2 as @xmath29 on @xmath30 . under assumption
[ assumption a ] , we have the following : a. relations ( [ r1 ] ) and ( [ r2 ] ) hold for every @xmath124 ; b. if @xmath130 , then relation ( [ r1 ] ) holds for @xmath129 ; c. if @xmath130 and @xmath147 , then relation ( [ r2 ] ) holds for @xmath129 .
as we pointed out before , theorem [ no1 ] does not require a dominating relationship between @xmath32 and @xmath34 . even in assertions
( b ) and ( c ) where @xmath130 is assumed , there is not necessarily a dominating relationship between @xmath32 and @xmath34 , though the conditions on @xmath32 and @xmath34 become not exactly symmetric any more .
additionally , theorems 5.2(3 ) and 6.1 of tang and tsitsiashvili @xcite are two special cases of our theorem [ no1](a ) with @xmath148 and @xmath149 , respectively . since the famous work of kesten @xcite , the tail probabilities of @xmath150 and @xmath24 have been extensively investigated , mainly in the framework of random difference equations and most under so - called cramr s condition that @xmath151 holds for some @xmath71 .
traditional random difference equations appearing in the literature are often different from ones such as ( [ m1 ] ) and ( [ t recursive equation ] ) below associated to our model .
nevertheless , under our standard assumptions on @xmath152 and @xmath153 , these subtle differences are not essential and the existing results can easily be transformed to our framework .
we omit such details here . corresponding to our model , kesten s work @xcite shows an asymptotic formula of the form @xmath154 assuming , among others , that @xmath2 fulfills cramr s condition and @xmath0 fulfills a certain integrability condition involving @xmath2 .
kesten s constant @xmath155 , though positive , is generally unknown .
see enriquez _
@xcite for a probabilistic representation for this constant .
goldie @xcite studied the same problem but in a broader scope and he simplified kesten s argument .
note that cramr s condition is essentially used in these works . among few works on this topic beyond cramr
s condition we mention grey @xcite and goldie and grbel @xcite . for the case
where @xmath97 for some @xmath71 , @xmath156 for some @xmath157 , and @xmath158 , indicating that the insurance risk dominates the financial risk , grey s work @xcite shows a precise asymptotic formula similar to ours .
goldie and grbel @xcite interpreted the study in terms of perpetuities in insurance and finance and they derived some rough asymptotic formulas .
corresponding to our model , their results show that @xmath19 exhibits a light tail if @xmath0 is light tailed and @xmath159 , while @xmath19 must exhibit a heavy tail once @xmath160 , regardless of the tail behavior of @xmath0 , all being consistent with the consensus on this topic that risky investments are dangerous .
we also refer the reader to hult and samorodnitsky @xcite , collamore @xcite , blanchet and sigman @xcite , and hitczenko and wesoowski @xcite for recent interesting developments on the topic .
in contrast to these existing results , we do not require cramr s condition or a dominating relationship between @xmath32 and @xmath34 in theorem [ no1](b ) , ( c ) .
the coefficients @xmath133 and @xmath134 appearing in our formulas , though still generally unknown , assume transparent structures as given in ( [ bc ] ) , which enable one to easily conduct numerical estimates .
the condition @xmath130 in theorem [ no1](b ) , ( c ) is made mainly to ensure the finiteness of @xmath133 and @xmath134 .
however , it excludes some apparently simpler cases such as @xmath161 and classical random walks ( corresponding to @xmath162 ) .
the tail behavior of the maximum of a random walk with negative drift , especially with heavy - tailed increments , has been systematically investigated by many people ; see , for example , feller @xcite , veraverbeke @xcite , korshunov @xcite , borovkov @xcite , denisov _ et al . _
@xcite , and foss _ et al . _
@xcite , among many others .
the study of random walks hints that the tail probabilities of @xmath163 and @xmath24 behave essentially differently between the cases @xmath130 and @xmath151 .
actually , if @xmath164 , then all of @xmath125 , @xmath126 , and @xmath127 diverge to @xmath165 as @xmath18 , and theorem [ no1 ] leads to @xmath166 this fails to give precise asymptotic formulas for @xmath167 and @xmath168 , though still consistent with kesten and goldie s formula @xmath154 since @xmath169 . for this case
, intriguing questions include how to capture the precise asymptotics other than kesten and goldie s for @xmath170 and @xmath171 and how to connect the asymptotics for @xmath172 and @xmath173 as @xmath174 to kesten and goldie s formula @xmath154 .
the approach developed in the present paper seems not efficient to give a satisfactory answer to either of these questions
. admittedly , the standard complete independence assumptions on the two sequences @xmath175 and @xmath153 , though often appearing in the literature , are not of practical relevance .
however , theorem [ no1 ] offers new insights into the tail probabilities of the sums in ( [ s ] ) and their maxima in ( [ m ] ) , revealing the interplay between the insurance and financial risks .
furthermore , extensions that incorporate various dependence structures into the model are expected and usually without much difficulty .
we show in the next subsection a simple example for such extensions .
as done by chen @xcite , in this subsection we assume that @xmath176 is a sequence of i.i.d .
random pairs with @xmath177 following a farlie gumbel morgenstern ( fgm ) distribution @xmath178 , x\in\mathbb{r } , y>0,\ ] ] where @xmath31 on @xmath50 and @xmath29 on @xmath30 are two marginal distributions .
in view of the decomposition @xmath179 the fgm structure can easily be dissolved .
hereafter , for a random variable @xmath84 and its i.i.d .
copies @xmath180 and @xmath181 , denote by @xmath182 a random variable identically distributed as @xmath183 and independent of all other sources of randomness . under assumption [ assumption a ] , by ( [ pi ] ) and lemma [ c2 ] below
, we can conduct an induction procedure to obtain @xmath184 and @xmath185 for every @xmath124 , where @xmath186 additionally , under the conditions of theorem [ no1](b ) , ( c ) , letting @xmath187 leads to @xmath188 where @xmath0 and @xmath189 are independent of @xmath24 and @xmath19 .
it is easy to verify the finiteness of @xmath190 and @xmath191 .
we summarize the analysis above into the following corollary and will show a sketch of its proof in section [ proof of cor ] .
[ no2]let @xmath176 be a sequence of i.i.d .
random pairs with common fgm distribution ( [ fgm ] ) . under assumption
[ assumption a ] , we have the following : a. relations ( [ r3 ] ) and ( [ r4 ] ) hold for every @xmath124 ; b. if @xmath130 , then relation ( [ r3 ] ) holds for @xmath129 ; c. if @xmath130 and @xmath147 , then relation ( [ r4 ] ) holds for @xmath129 . as a sanity check , letting @xmath192 , the results in corollary [ no2 ] coincide with those in theorem [ no1 ] .
in this section , we prepare a series of lemmas , some of which are interesting in their own right .
we first recall some well - known properties of distributions of regular variation and convolution equivalence . if @xmath54 for some @xmath53 , then for every @xmath157 and every @xmath193 there is some constant @xmath194 such that potter s bounds
@xmath195 hold whenever @xmath196 and @xmath197 ; see theorem 1.5.6(iii ) of bingham _ et al .
since @xmath78 if and only if @xmath55 defined by ( [ u v ] ) belongs to @xmath65 , potter s bounds above can easily be restated in terms of a distribution @xmath198 as that , for every @xmath157 and every @xmath193 there is some constant @xmath194 such that the inequalities @xmath199 hold whenever @xmath196 and @xmath200 . by lemma 5.2 of pakes
@xcite , if @xmath201 then it holds for every @xmath202 that @xmath203 the first lemma below describes an elementary property of convolution equivalence .
[ l1]let @xmath204 be @xmath205 i.i.d .
real - valued random variables with common distribution @xmath206 for some @xmath60 .
then @xmath207 for every @xmath208 and @xmath209 , write @xmath210 by relation ( [ pakes ] ) , we have @xmath211 and @xmath212 where in the last step we used @xmath213 and the dominated convergence theorem . similarly , @xmath214 plugging these limits into ( [ 123 ] ) yields the desired result . hereafter , for @xmath205 distributions @xmath215 , denote by @xmath216 a convex combination of @xmath217 , where @xmath218 .
[ new1]let @xmath215 be @xmath205 distributions and let @xmath60 .
the following are equivalent : a. @xmath219 for every @xmath220 ; b. @xmath219 for some @xmath220 and the relation @xmath221 holds for every @xmath222 and every @xmath223 .
first prove that ( b ) implies ( a ) .
denote by @xmath224 this specific element in @xmath225 such that @xmath226 .
for every @xmath227 , it is easy to see that @xmath228 and that @xmath229 by ( [ assumption ] ) .
thus , @xmath230 follows from the closure of the class @xmath70 under weak equivalence as mentioned in the last paragraph of section [ a ] . for the other implication ,
we only need to use ( a ) to verify ( [ assumption ] ) . for arbitrarily fixed @xmath231 and every @xmath223 , each of the sums @xmath232 and
@xmath233 is proportional to a convolution - equivalent tail .
thus , @xmath234 by the arbitrariness of @xmath235 , relation ( [ assumption ] ) follows .
the following lemma shows the usefulness of convolution equivalence in dealing with the tail probability of the sum of independent random variables .
note that the lemma does not require any dominating relationship among the individual tails .
additionally , in view of lemma [ new1 ] , letting @xmath131 in lemma [ l2 ] retrieves theorem 1 of li and tang @xcite .
[ l2]let @xmath215 be @xmath205 distributions on @xmath236 and let @xmath60 . if @xmath230 for every @xmath227 , then @xmath237 and @xmath238 clearly , we only need to prove relation ( [ for n ] ) .
introduce @xmath5 independent random variables @xmath204 with distributions @xmath215 , respectively . for every @xmath208 and @xmath239 ,
@xmath240 according to whether or not there is exactly only one @xmath241 occurring in the union , we split the probability on the right - hand side into two parts as @xmath242 first we deal with @xmath243 . for a real vector @xmath244 ,
write @xmath245 , and for each @xmath246 , write @xmath247 we have @xmath248 since @xmath249 is proportional to a convolution - equivalent tail , by the dominated convergence theorem , @xmath250 hence , it follows from ( [ assumption ] ) and the dominated convergence theorem that @xmath251 next we turn to @xmath252 .
write @xmath253 , which has a convolution - equivalent tail proportional to @xmath254 , and let @xmath255 be i.i.d .
copies of @xmath256 . clearly , @xmath257 thus , by lemma [ l1 ] , @xmath258 plugging ( [ a ] ) and ( [ b ] ) into ( [ c ] ) yields the desired result . due to the connection between convolution equivalence and strongly regular variation
, we can restate lemmas [ new1 ] and [ l2 ] in terms of strongly regular variation .
actually , the next lemma shows an equivalent condition for assumption [ assumption a ] .
[ new2]let @xmath259 be @xmath205 distributions and let @xmath60 .
the following are equivalent : a. @xmath260 for every @xmath227 ; b. @xmath260 for some @xmath261 and the relation @xmath262 holds for every @xmath103 and every @xmath223 .
the lemma below expands the tail probability of the product of independent , nonnegative , and strongly regular random variables , forming an analogue of the well - known breiman s theorem in a different situation . for breiman
s theorem , see breiman @xcite and cline and samorodnitsky @xcite . [ c2]let @xmath263 be @xmath205 independent nonnegative random variables with distributions @xmath259 , respectively , and let @xmath60 .
if @xmath264 for every @xmath227 , then the distribution of @xmath265 belongs to the class @xmath266 and @xmath267 the next lemma shows kesten s bound for convolution tails without the usual requirement @xmath268 .
it improves lemma 5.3 of pakes @xcite for the case @xmath269 .
[ kesten]let @xmath55 be a distribution on @xmath50 .
if @xmath270 for some @xmath60 , then for every @xmath157 there is some constant @xmath271 such that the relation @xmath272 holds for all @xmath124 and all @xmath208 . when @xmath268 , the assertion has been given in lemma 5.3 of pakes @xcite
hence , we only need to consider @xmath273 ( for which @xmath71 must hold ) .
let @xmath274 be a sequence of i.i.d .
random variables with common distribution @xmath55 , and set @xmath275 . clearly , @xmath276 note that the distribution of @xmath277 still belongs to the class @xmath70 and @xmath278 .
hence , for every @xmath279 , by lemma 5.3 of pakes @xcite , there is some constant @xmath280 such that , for all @xmath124 and all @xmath208 , @xmath281 by ( [ potter s for l ] ) , there are some constants @xmath282 and @xmath194 such that , for all @xmath124 and all @xmath196 , @xmath283 plugging ( [ k2 ] ) into ( [ k1 ] ) and noticing that @xmath284 , we have , for all @xmath124 and all @xmath285 , @xmath286 for @xmath287 , we choose an integer @xmath288 .
then , for @xmath287 and @xmath289 , using the same derivations as in ( [ k1])([k3 ] ) , we obtain @xmath290 at last , for @xmath287 and @xmath291 , it is obvious that @xmath292 a combination of ( [ k3])([k5 ] ) gives that , for some constant @xmath271 and for all @xmath124 and all @xmath208 , @xmath293 by setting @xmath294 to be small enough such that @xmath295 , we complete the proof .
the following lemma will be crucial in proving theorem [ no1](b ) , ( c ) .
[ infinite]let @xmath115 be a sequence of ( arbitrarily dependent ) random variables with common distribution @xmath31 on @xmath50 , let @xmath116 be another sequence of i.i.d .
random variables with common distribution @xmath29 on @xmath56 , and let the two sequences be mutually independent .
assume that there is some distribution @xmath296 for @xmath71 such that @xmath297 assume also that @xmath130 .
then @xmath298 choose some large constant @xmath280 such that the inequality @xmath299 holds for all @xmath300 , and then introduce a nonnegative random variable @xmath301 with a distribution @xmath302 clearly , @xmath303 for all @xmath208 and @xmath304 for all large @xmath46 . the inequality @xmath305 for all
@xmath208 means that @xmath0 is stochastically not greater than @xmath301 , denoted by @xmath306 . moreover , since @xmath296 , there is some large but fixed constant @xmath307 such that @xmath308 .
for this fixed @xmath309 , define @xmath310 and then introduce another nonnegative random variable @xmath311 with a distribution @xmath312 clearly , @xmath313 , @xmath314 for all @xmath315 , and @xmath316 for all .
thus , @xmath317 . let @xmath318 , @xmath319 ,
copies of @xmath311 independent of @xmath301 .
choose some @xmath320 such that @xmath321 .
by lemma [ kesten ] , there is some constant @xmath282 such that , for all @xmath322 and all @xmath323 , @xmath324 noticeably , the derivation in ( [ inf1 ] ) tacitly requires that @xmath325 are positive .
nevertheless , in case @xmath326 assigns a mass at @xmath327 , the upper bound in ( [ inf1 ] ) is still correct and can easily be verified by conditioning on @xmath328 . by lemma [ c2 ] , @xmath329 moreover , by ( [ potter s ] ) , there is some constant @xmath194 such that , for all @xmath330 and @xmath331 , @xmath332 now we start to estimate the tail probability in ( [ bound ] ) . choosing some large @xmath5 such that @xmath333 .
clearly , for all @xmath330 , @xmath334 \\[-8pt ] & \leq & \biggl ( \sum_{i>\sqrt{x / x_{0}}}+\sum _ { n < i\leq\sqrt { x / x_{0 } } } \biggr ) \pr \biggl ( x^{\ast}\prod _ { j=1}^{i}y_{j}^{\ast } > \frac { x}{i^{2 } } \biggr ) \nonumber \\ & = & i_{1}(x)+i_{2}(n , x ) , \nonumber\end{aligned}\ ] ] where @xmath335 is understood as @xmath327 in case @xmath336
. first we deal with @xmath337 . by chebyshev s inequality ,
@xmath338 this means that @xmath337 converges to @xmath327 at least semi - exponentially fast since @xmath313 .
thus , @xmath339
next we deal with @xmath335 .
we further decompose it into three parts as @xmath340 \\[-8pt ] & & { } + \sum_{n <
i\leq\sqrt{x / x_{0}}}\pr \biggl ( x^{\ast } \prod _ { j=1}^{i}y_{j}^{\ast } > \frac{x}{i^{2}},\prod_{j=1}^{i}y_{j}^{\ast } \leq1 \biggr ) \nonumber \\ & = & i_{21}(n , x)+i_{22}(n , x)+i_{23}(n , x ) .
\nonumber\end{aligned}\ ] ] by conditioning on @xmath301 and then applying ( [ inf1])([inf22 ] ) , we obtain @xmath341 since @xmath342 , it follows that @xmath343 applying both ( [ inf1 ] ) and ( [ inf22 ] ) , we have @xmath344 which implies that @xmath345 similarly , applying ( [ inf22 ] ) twice , @xmath346 which , together with @xmath347 , gives that @xmath348 a combination of relations ( [ two terms])([inf25 ] ) completes the proof .
we first prove relation ( [ r1 ] ) .
it is easy to verify that @xmath349 where @xmath350 denotes equality in distribution ; see also theorem 2.1 of tang and tsitsiashvili @xcite .
we proceed with induction . for @xmath351
, it follows from lemma [ c2 ] that @xmath352 thus , relation ( [ r1 ] ) holds for @xmath351 .
now we assume by induction that relation ( [ r1 ] ) holds for @xmath353 and prove it for @xmath5 . by this induction assumption and assumption [ assumption a ] , we know that every convex combination of the distributions of @xmath4 and @xmath354 belongs to the class @xmath355 . applying lemma [ l2 ] with @xmath131 , we have @xmath356 which , together with assumption [ assumption a ] , implies that every convex combination of the distributions of @xmath357 and @xmath6 belongs to the class @xmath358 .
applying lemma [ c2 ] , we obtain @xmath359 therefore , relation ( [ r1 ] ) holds for @xmath5 .
next we turn to relation ( [ r2 ] ) .
introduce a sequence of random variables @xmath360 through the recursive equation @xmath361 equipped with @xmath362 .
it is easy to see that @xmath363 for @xmath124 .
then the proof of relation ( [ r2 ] ) can be done by using the recursive equation ( [ t recursive equation ] ) and going along the same lines as in the proof of relation ( [ r1 ] ) above .
note that @xmath125 and @xmath126 increasingly converge to the finite constants @xmath132 and @xmath133 . also recall lemma [ infinite ] .
hence , for every @xmath364 , there is some large integer @xmath365 such that both @xmath366 and @xmath367 hold .
now we start to deal with @xmath168 . on the one hand , for every @xmath157 , by theorem [ no1](a ) , relation ( [ t4 ] ) , and assumption
[ assumption a ] , in turn , we obtain @xmath368 \\[-8pt ] & \sim & ( 1-\varepsilon ) ^{-\alpha } \bigl ( a_{n_{0}}\overline { f}(x)+b_{n_{0}}\overline{g}(x ) \bigr ) + \delta \varepsilon^{-\alpha } \bigl ( \overline{f}(x)+\overline{g}(x ) \bigr ) \nonumber \\ & \leq & \bigl ( ( 1-\varepsilon ) ^{-\alpha}a_{\infty } + \delta \varepsilon^{-\alpha } \bigr ) \overline{f}(x)+ \bigl ( ( 1-\varepsilon ) ^{-\alpha}b_{\infty}+\delta\varepsilon^{-\alpha } \bigr ) \overline{g}(x ) .
\nonumber\end{aligned}\ ] ] on the other hand , by theorem [ no1](a ) and relation ( [ t3 ] ) , @xmath369 by the arbitrariness of @xmath294 and @xmath235 in ( [ t5 ] ) and ( [ t6 ] ) , we obtain relation ( [ r1 ] ) for @xmath129 .
first we establish an asymptotic upper bound for @xmath167 .
as in the proof of theorem [ no1](b ) , for every @xmath364 , suitably choose some large integer @xmath365 such that relations ( [ t3 ] ) , ( [ t4 ] ) , and the relation @xmath370 hold simultaneously . for every @xmath157 , by theorem [ no1](a ) , relation ( [ t4 ] )
, assumption [ assumption a ] , and relation ( [ t7 ] ) , in turn , we obtain @xmath371 since @xmath294 and @xmath235 are arbitrary positive constants , it follows that @xmath372 for the corresponding asymptotic lower bound , as analyzed in the proof of theorem [ no1](a ) , it suffices to prove that @xmath373 where @xmath374 is the weak limit of the sequence @xmath375 defined by ( [ t recursive equation ] ) .
we apply the method developed by grey @xcite to prove ( [ lower ] ) .
consider the stochastic difference equation @xmath376 which inherits a stochastic structure from ( [ t recursive equation ] ) .
note that the weak solution of ( [ stochastic difference equation ] ) exists and is unique .
furthermore , the limit distribution of @xmath377 is identical to this unique solution and , hence , it does not depend on the starting random variable @xmath378 .
see vervaat @xcite and goldie @xcite for these and related statements .
it is easy to check that @xmath379 ; see the proof of theorem 1 of grey @xcite for a similar argument . construct a new starting random variable @xmath380 independent of @xmath381 with tail @xmath382 starting with @xmath380 , the recursive equation ( [ t recursive equation ] )
generates the sequence @xmath383 correspondingly . comparing ( [ starting random variable ] ) with ( [ stochastic difference equation ] ) ,
we see that @xmath380 and , hence , every @xmath384 are stochastically not greater than @xmath374 ; namely , it holds for all @xmath385 and all @xmath386 that @xmath387 furthermore , it holds that @xmath388 where the last step is analogous to ( [ cited ] ) .
thus , by assumption [ assumption a ] , the distribution of @xmath380 belongs to the class @xmath358 .
then , by going along the same lines of the proof of theorem [ no1](a ) and using equation ( [ t recursive equation ] ) starting with @xmath380 , we obtain @xmath389 with @xmath390 since @xmath384 weakly converges to @xmath391 and @xmath130 , it is easy to see that @xmath392 and @xmath393 , with the latter subject to a straightforward application of the dominated convergence theorem . thus , substituting ( [ lower 2 ] ) into ( [ lower 1 ] ) and letting @xmath18 on the right - hand side of the resulting formula , we arrive at relation ( [ lower ] ) as desired .
clearly , the recursive equations ( [ m1 ] ) , ( [ t recursive equation ] ) , and the identity @xmath394 for @xmath202 still hold since @xmath395 is a sequence of i.i.d . random pairs . introduce four independent random variables @xmath396 , @xmath397 , @xmath398 , and @xmath399 with distributions @xmath31 , @xmath400 , @xmath29 , and @xmath401 , respectively , and let them be independent of @xmath395 . using decomposition ( [ pi ] )
, we have @xmath402 when @xmath351 , applying lemma [ c2 ] to each term on the right - hand side of ( [ decomposition ] ) gives @xmath403 then , as in the proof of theorem [ no1](a ) , proceeding with induction according to ( [ decomposition ] ) leads to ( [ r3 ] ) .
relation ( [ r4 ] ) can be derived similarly .
this proves corollary [ no2](a ) .
corollary [ no2](b ) , ( c ) can be verified by the similar ideas used in proving theorem [ no1](b ) , ( c ) .
the key ingredient is establishing a relation similar to ( [ bound ] ) .
write @xmath404 , @xmath405 , @xmath406 , and so on .
it follows from ( [ c1 ] ) that @xmath407 as in the proof of lemma [ infinite ] , we can construct independent random variables @xmath408 and @xmath311 both with tails equal to @xmath409 for all large @xmath46 such that @xmath410 , @xmath317 , and @xmath411 . for some large @xmath5 such that @xmath333 , we write @xmath412 then , going along the same lines of the rest of the proof of lemma [ infinite ] , we obtain @xmath413 which suffices for our purpose .
the authors are very grateful to the two reviewers and the associate editor for their insightful comments and constructive suggestions , which have helped significantly improve this work .
li s research was supported by the national natural science foundation of china ( grant no .
11201245 ) and the research fund for the doctoral program of higher education of china ( grant no .
20110031120003 ) .
tang s research was supported by a centers of actuarial excellence ( cae ) research grant ( 20132016 ) from the society of actuaries . | consider an insurance company exposed to a stochastic economic environment that contains two kinds of risk .
the first kind is the insurance risk caused by traditional insurance claims , and the second kind is the financial risk resulting from investments .
its wealth process is described in a standard discrete - time model in which , during each period , the insurance risk is quantified as a real - valued random variable @xmath0 equal to the total amount of claims less premiums , and the financial risk as a positive random variable @xmath1 equal to the reciprocal of the stochastic accumulation factor .
this risk model builds an efficient platform for investigating the interplay of the two kinds of risk .
we focus on the ruin probability and the tail probability of the aggregate risk amount . assuming that every convex combination of the distributions of @xmath0 and @xmath2 is of strongly regular variation
, we derive some precise asymptotic formulas for these probabilities with both finite and infinite time horizons , all in the form of linear combinations of the tail probabilities of @xmath0 and @xmath2 . our treatment is unified in the sense that no dominating relationship between @xmath0 and @xmath2 is required .
./style / arxiv - general.cfg |
these are notes from a minicourse given at a workshop in melbourne july 1115 2011 .
there is little pretension to originality ; the main novelty is firstly that we give a new ( and much shorter ) proof of coornaert s theorem on patterson sullivan measures for hyperbolic groups ( theorem [ measure_is_conformal ] ) , and secondly that we explain how to combine the results of calegari fujiwara in @xcite with that of pollicott sharp @xcite to prove central limit theorems for quite general classes of functions on hyperbolic groups ( corollary [ distance_clt ] and theorem [ holder_clt ] ) , crucially _ without _ the hypothesis that the markov graph encoding an automatic structure is ergodic .
a final section on random walks is much more cursory .
the fundamental idea in geometric group theory is to study groups as automorphisms of geometric spaces , and as a special case , to study the group itself ( with its canonical self - action ) as a geometric space .
this is accomplished most directly by means of the _ cayley graph _ construction .
let @xmath0 be a group and @xmath1 a ( usually finite ) generating set .
associated to @xmath0 and @xmath1 we can form the _ cayley graph _ @xmath2 .
this is a graph with vertex set @xmath0 , and with an edge from @xmath3 to @xmath4 for all @xmath5 and @xmath6 .
the action of @xmath0 on itself by ( left ) multiplication induces a properly discontinuous action of @xmath0 on @xmath2 by simplicial automorphisms .
if @xmath0 has no @xmath7-torsion , the action is free and properly discontinuous , and the quotient is a wedge of @xmath8 circles @xmath9 . in this case , if @xmath0 has a presentation @xmath10 we can think of @xmath2 as the covering space of @xmath9 corresponding to the subgroup of the free group @xmath11 normally generated by @xmath12 , and the action of @xmath0 on @xmath2 is the deck group of the covering . with generating set @xmath13 we assume the reader is familiar with the notion of a metric space , i.e. a space @xmath14 together with a symmetric non - negative real - valued function @xmath15 on @xmath16 which vanishes precisely on the diagonal , and which satisfies the triangle inequality @xmath17 for each triple @xmath18 .
a metric space is a _ path metric space _ if for each @xmath19 , the distance @xmath20 is equal to the infimum of the set of numbers @xmath21 for which there is a @xmath22-lipschitz map @xmath23 \to x$ ] sending @xmath24 to @xmath25 and @xmath21 to @xmath26 .
it is a _ geodesic _
metric space if it is a path metric space and if the infimum is achieved on some @xmath27 for each pair @xmath28 ; such a @xmath27 is called a geodesic .
finally , a metric space is _ proper _ if closed metric balls of bounded radius are _ compact _ ( equivalently , for each point @xmath25 the function @xmath29 is proper ) .
the graph @xmath2 can be canonically equipped with the structure of a _ geodesic metric space_. this is accomplished by making each edge isometric to the euclidean unit interval . if @xmath1 is finite , @xmath2 is _
proper_. note that @xmath0 itself inherits a subspace metric from @xmath2 , called the _
word metric_. we denote the word metric by @xmath30 , and define @xmath31 ( or just @xmath32 if @xmath1 is understood ) to be @xmath33 . observe that @xmath34 and that @xmath31 is the length of the shortest word in elements of @xmath1 and their inverses representing the element @xmath3 .
the most serious shortcoming of this construction is its dependence on the choice of a generating set @xmath1 .
different choices of generating set @xmath1 give rise to different spaces @xmath2 which are typically not even homeomorphic . the standard way to resolve this issue
is to coarsen the geometric category in which one works .
let @xmath35 and @xmath36 be metric spaces .
a map @xmath37 ( _ not _ assumed to be continuous ) is a _ quasi - isometric map _ if there are constants
@xmath38 so that @xmath39 for all @xmath40 .
it is said to be a _
quasi - isometry _ if further @xmath41 is a _ net _ in @xmath42 ; that is , if there is some @xmath12 so that @xmath42 is equal to the @xmath12-neighborhood of @xmath41 .
one also uses the terminology _ @xmath43 quasi - isometric map _ or _
@xmath43 quasi - isometry _ if the constants are specified . note that a @xmath44 quasi - isometric map is the same thing as a @xmath45 bilipschitz map . the best constant @xmath45 is called the _ multiplicative constant _ , and the best @xmath46 the _ additive constant _ of the map .
we denote the @xmath12-neighborhood of a set @xmath47 by @xmath48 .
hence a quasi - isometry is a quasi - isometric map for which @xmath49 for some @xmath12 .
it is much more common to use the terminology _ quasi - isometric embedding _ instead of quasi - isometric map as above ; we consider this terminology misleading , and therefore avoid it .
[ qi_equivalence ] quasi - isometry is an equivalence relation .
reflexivity and transitivity are obvious , so we must show symmetry . for
each @xmath50 choose @xmath51 with @xmath52 ( such an @xmath25 exists by definition ) and define @xmath53 .
observe @xmath54 by definition . then @xmath55 similarly , @xmath56 proving symmetry .
note that the compositions @xmath57 and @xmath58 as above move points a bounded distance .
one can define a category in which objects are equivalence classes of metric spaces under the equivalence relation generated by thickening ( i.e. isometric inclusion as a net in a bigger space ) , and morphisms are equivalence classes of quasi - isometric maps , where two maps are equivalent if their values on each point are a uniformly bounded distance apart . in this category , quasi - isometries are isomorphisms .
in particular , the set of quasi - isometries of a metric space @xmath14 , modulo maps that move points a bounded distance , is a _ group _ , denoted @xmath59 , which only depends on the quasi - isometry type of @xmath14 .
determining @xmath59 , even for very simple spaces , is typically extraordinarily difficult .
a metric space @xmath35 is quasi - isometric to a point if and only if it has bounded diameter .
a cayley graph @xmath2 ( for @xmath1 finite ) is quasi - isometric to a point if and only if @xmath0 is finite .
[ generating_sets_qi ] if @xmath1 and @xmath60 are two finite generating sets for a group @xmath0 then the identity map from @xmath0 to itself is a quasi - isometry ( in fact , a bilipschitz map ) of @xmath61 to @xmath62 . for , there are constants @xmath63 and @xmath64 so that @xmath65 for all @xmath6 , and @xmath66 for all @xmath67 , and therefore @xmath68 . because of this , the quasi - isometry class of @xmath61 is _ independent _ of the choice of finite generating set , and we can speak unambiguously of the quasi - isometry class of @xmath0 . the schwarz lemma connects the geometry of groups to the geometry of spaces they act on .
[ schwarz_lemma ] let @xmath0 act properly discontinuously and cocompactly by isometries on a proper geodesic metric space @xmath14 .
then @xmath0 is finitely generated by some set @xmath1 , and the orbit map @xmath69 sending @xmath3 to @xmath70 ( for any @xmath51 ) is a quasi - isometry from @xmath61 to @xmath14 .
since @xmath14 is proper and @xmath0 acts cocompactly there is an @xmath12 so that @xmath71 .
note that @xmath72 is a net , since every point of @xmath14 is contained in some translate @xmath73 and is therefore within distance @xmath12 of @xmath70 .
let @xmath74 .
since @xmath0 acts properly discontinuously , there are only finitely many @xmath3 in @xmath0 for which @xmath75 is nonempty ; let @xmath1 be the nontrivial elements of this set . now , if @xmath76 are arbitrary , let @xmath27 be a geodesic in @xmath14 from @xmath70 to @xmath77 .
parameterize @xmath27 by arclength , and for each integer @xmath78 let @xmath79 be such that @xmath80
. then @xmath81 and therefore @xmath82 which shows incidentally that @xmath1 generates @xmath0 .
conversely , if @xmath83 and @xmath79 is a sequence of elements with @xmath84 and @xmath85 and each @xmath86 , then there is a path @xmath87 from @xmath88 to @xmath89 of length at most @xmath90 , and the concatenation of these paths certifies that @xmath91 this completes the proof of the lemma .
if @xmath0 is a group and @xmath92 is a subgroup of finite index , then @xmath0 and @xmath92 are quasi - isometric ( for , both act properly discontinuously and cocompactly on @xmath2 ) .
two groups are said to be _
commensurable _ if they have isomorphic subgroups of finite index ; the same argument shows that commensurable groups are quasi - isometric .
any two regular trees of ( finite ) valence @xmath93 are quasi - isometric ; for , any such tree admits a cocompact action by a free group of finite rank , and any two free groups of finite rank are commensurable .
the set of ends of a geodesic metric space is a quasi - isometry invariant .
a famous theorem of stallings @xcite says that a finitely generated group with more than one end splits over a finite subgroup ; it follows that the property of splitting over a finite subgroup is a quasi - isometry invariant .
finiteness of the edge groups ( in a splitting ) is detected quasi - isometrically by the existence of _ separating _ compact subsets .
quasi - isometry can further detect the finiteness of the vertex groups , and in particular one observes that a group is quasi - isometric to a free group if and only if it is virtually free .
any two groups that act cocompactly and properly discontinuously on the same space @xmath14 are quasi - isometric .
for example , if @xmath94 are closed riemannian manifolds with isometric universal covers , then @xmath95 and @xmath96 are quasi - isometric .
it is easy to produce examples for which the groups in question are not commensurable ; for instance , a pair of closed hyperbolic @xmath97-manifolds @xmath98 ,
@xmath99 with different invariant trace fields ( see @xcite ) . in the geometric group theory literature , lemma [ schwarz_lemma ] is often called the `` milnor varc ( or varc - milnor ) lemma '' ; `` varc '' here is in fact the well - known mathematical physicist albert schwarz ; it is our view that the orthography `` varc '' tends to obscure this . actually , the content of this lemma was first observed by schwarz in the early 50 s and only rediscovered 15 years later by milnor at a time when the work of soviet mathematicians was not widely disseminated in the west . in a geodesic metric space
a _ geodesic triangle _ is just a union of three geodesics joining three points in pairs .
if the three points are @xmath100 we typically denote the ( oriented ) geodesics by @xmath101 , @xmath102 and @xmath103 respectively ; this notation obscures the possibility that the geodesics in question are not uniquely determined by their endpoints .
a geodesic metric space @xmath35 is _ @xmath104-hyperbolic _ if for any geodesic triangle , each side of the triangle is contained in the @xmath104-neighborhood of the union of the other two sides .
a metric space is _ hyperbolic _ if it is @xmath104-hyperbolic for some @xmath104 .
one sometimes says that _ geodesic triangles are @xmath104-thin_. -thin triangle ; the gray tubes have thickness @xmath104 . ]
a tree is @xmath24-hyperbolic .
hyperbolic space ( of any dimension ) is @xmath104-hyperbolic for a uniform @xmath104 .
if @xmath14 is a simply - connected complete riemannian manifold with curvature bounded above by some @xmath105 then @xmath14 is @xmath104-hyperbolic for some @xmath104 depending on @xmath45 .
a geodesic metric space @xmath14 is @xmath106 for some @xmath45 if triangles are _ thinner _ than comparison triangles in a space of constant curvature @xmath45 .
this means that if @xmath107 is a geodesic triangle in @xmath14 , and @xmath108 is a geodesic triangle in a complete simply connected riemannian manifold @xmath42 of constant curvature @xmath45 with edges of the same lengths , and @xmath109 is an isometry on each edge , then for any @xmath110 we have @xmath111 .
the initials @xmath112 stand for cartan
toponogov , who made substantial contributions to the theory of comparison geometry . from the definition ,
a @xmath106 space is @xmath104-hyperbolic whenever the complete simply connected riemannian @xmath7-manifold of constant curvature @xmath45 is @xmath104-hyperbolic .
hence a @xmath106 space is hyperbolic if @xmath105 .
[ exa : nearest_point ] nearest point projection to a convex subset of a @xmath106 space with @xmath113 is distance nonincreasing .
therefore the subspace metric and the path metric on a convex subset of a @xmath106 space agree , and such a subspace is itself @xmath106 .
thinness of triangles implies thinness of arbitrary polygons .
[ exa : quadrilateral ] let @xmath14 be @xmath104-hyperbolic and let @xmath114 be a geodesic quadrilateral .
then either there are points on @xmath115 and @xmath116 at distance @xmath117 or there are points on @xmath118 and @xmath119 at distance @xmath117 , or possibly both . the number of essentially distinct ways in which an @xmath120-gon can be thin is equal to the @xmath120th catalan number . by cutting up a polygon into triangles and examining the implications of @xmath104-thinness for each triangle
, one can reason about the geometry of complicated configurations in @xmath104-hyperbolic space .
[ nearest_point_projection ] let @xmath14 be @xmath104-hyperbolic , let @xmath27 be a geodesic segment / ray / line in @xmath14 , and let @xmath121 .
then there is a point @xmath122 on @xmath27 realizing the infimum of distance from @xmath123 to points on @xmath27 , and moreover for any two such points @xmath124 we have @xmath125 . the existence of some point realizing the infimum follows from the properness of @xmath126 , valid for any geodesic in any metric space .
let @xmath124 be two such points , and if @xmath127 let @xmath128 be the midpoint of the segment @xmath129 , so @xmath130 . without loss of generality
there is @xmath131 on @xmath132 with @xmath133 hence @xmath134 .
but then @xmath135 contrary to the fact that @xmath122 minimizes the distance from @xmath123 to points on @xmath27 . lemma [ nearest_point_projection ] says that there is an approximate `` nearest point projection '' map @xmath136 from @xmath14 to any geodesic @xmath27 ( compare with example [ exa : nearest_point ] ) .
this map is not continuous , but nearby points must map to nearby points , in the sense that @xmath137 .
we would now like to show that the property of being hyperbolic is preserved under quasi - isometry .
the problem is that the property of @xmath104-hyperbolicity is expressed in terms of geodesics , and quasi - isometries do not take geodesics to geodesics . a _ quasigeodesic _ segment / ray / line is the image of a segment / ray / line in @xmath138 under a quasi - isometric map . for infinite or semi - infinite intervals
this definition has content ; for finite intervals this definition has no content without specifying the constants involved .
hence we can talk about a @xmath43 quasigeodesic segment / ray / line .
[ morse_lemma ] let @xmath35 be a proper @xmath104-hyperbolic space .
then for any @xmath43 there is a constant @xmath139 ( depending in an explicit way on @xmath140 ) so that any @xmath43 quasigeodesic @xmath27 is within hausdorff distance @xmath139 of a genuine geodesic @xmath141 .
if @xmath27 has one or two endpoints , @xmath141 can be chosen to have the same endpoints .
if @xmath27 is noncompact , it can be approximated on compact subsets by finite segments @xmath87 .
if we prove the lemma for finite segments , then a subsequence of the @xmath142 , converging on compact subsets , will limit to @xmath141 with the desired properties ( here is where we use properness of @xmath14 ) .
so it suffices to prove the lemma for @xmath27 a segment . in this case
choose any @xmath141 with the same endpoints as @xmath27 .
we need to estimate the hausdorff distance from @xmath27 to @xmath141 .
fix some constant @xmath139 and suppose there are points @xmath143 on @xmath27 that are both distance @xmath139 from @xmath141 , but @xmath144 for all @xmath131 on @xmath27 between @xmath123 and @xmath145 .
choose @xmath146 a sequence of points on @xmath27 and @xmath147 a sequence of points on @xmath141 closest to the @xmath146 so that @xmath148 .
consider the quadrilateral @xmath149 .
by example [ exa : quadrilateral ] either there are close points on @xmath150 and @xmath151 , or close points on @xmath152 and @xmath153 ( or possibly both ) .
suppose there are points @xmath154 on @xmath152 and @xmath155 on @xmath153 with @xmath156 .
then any nearest point projections of @xmath154 and @xmath155 to @xmath141 must be at most distance @xmath157 apart .
but @xmath147 and @xmath158 are such nearest point projections , by definition , and satisfy @xmath148 .
so it must be instead that there are points @xmath154 on @xmath150 and @xmath159 on @xmath151 which are at most @xmath160 apart .
but this means that @xmath161 , so the length of @xmath27 between @xmath123 and @xmath145 is at least @xmath162 where @xmath124 are points on @xmath27 closest to @xmath143 . on the other hand , @xmath163 .
since @xmath27 is a @xmath43 quasigeodesic , if @xmath164 is big enough , we get a uniform bound on @xmath139 in terms of @xmath140 . the remaining case where @xmath164 is itself uniformly bounded but @xmath139 is unbounded quickly leads to a contradiction .
[ delta_hyperbolic_qi ] let @xmath42 be @xmath104-hyperbolic and let @xmath37 be a @xmath43 quasi - isometry .
then @xmath14 is @xmath165-hyperbolic for some @xmath104 .
hence the property of being hyperbolic is a quasi - isometry invariant .
let @xmath166 be a geodesic triangle in @xmath14 with vertices @xmath167 .
then the edges of @xmath168 are @xmath43 quasigeodesics in @xmath42 , and are therefore within hausdorff distance @xmath139 of geodesics with the same endpoints .
it follows that every point on @xmath169 is within distance @xmath170 of @xmath171 and therefore every point on @xmath115 is within distance @xmath172 of @xmath173 .
the morse lemma lets us promote quasigeodesics to ( nearby ) geodesics .
the next lemma says that quasigeodesity is a _ local _ condition .
a path @xmath27 in @xmath14 is a _
@xmath174-local geodesic _ if the subsegments of length @xmath175 are geodesics .
similarly , @xmath27 is a _
@xmath174-local @xmath43 quasigeodesic _ if the subsegments of length @xmath175 are @xmath43 quasigeodesics .
[ k_local_geodesics ] let @xmath14 be a @xmath104-hyperbolic geodesic space , and let @xmath176 .
then any @xmath174-local geodesic is @xmath43 quasigeodesic for @xmath43 depending explicitly on @xmath104 .
more generally , for any @xmath43 there is a @xmath174 and constants @xmath177 so that any @xmath174-local @xmath43 quasigeodesic is a @xmath177 quasigeodesic .
let @xmath27 be a @xmath174-local geodesic segment from @xmath123 to @xmath122 , and let @xmath141 be any geodesic from @xmath123 to @xmath122 .
let @xmath131 be a point on @xmath27 furthest from @xmath141 , and let @xmath131 be the midpoint of an arc @xmath178 of @xmath27 of length @xmath179 . by hypothesis
, @xmath178 is actually a geodesic .
let @xmath180 and @xmath181 be points on @xmath141 closest to @xmath182 and @xmath183 .
the point @xmath131 is within distance @xmath160 either of @xmath141 or of one of the sides @xmath184 or @xmath185 .
if the latter , we would get a path from @xmath131 to @xmath180 or @xmath181 shorter than the distance from @xmath182 or @xmath183 , contrary to the definition of @xmath131 .
hence the distance from @xmath131 to @xmath141 is at most @xmath160 , and therefore @xmath27 is contained in the @xmath160 neighborhood of @xmath141 .
now let @xmath186 take points on @xmath27 to closest points on @xmath141 .
since @xmath136 moves points at most @xmath160 , it is approximately continuous .
since @xmath27 is a @xmath174-local geodesic , the map @xmath136 is approximately monotone ; i.e. if @xmath146 are points on @xmath27 with @xmath187 moving monotonely from one end of @xmath27 to the other , then @xmath188 and the projections also move monotonely along @xmath27 . in particular , @xmath189 and @xmath136 is a quasi - isometry .
the constants involved evidently depend only on @xmath104 and @xmath174 , and the multiplicative constant evidently goes to @xmath22 as @xmath174 gets large .
the more general fact is proved similarly , by using lemma [ morse_lemma ] to promote local quasigeodesics to local geodesics , and then back to global quasigeodesics .
corollary [ delta_hyperbolic_qi ] justifies the following definition : a group @xmath0 is hyperbolic if @xmath2 is @xmath104-hyperbolic for some @xmath104 for some ( and hence for any ) finite generating set @xmath1 .
free groups are hyperbolic , since their cayley graphs ( with respect to a free generating set ) are trees which are @xmath24-hyperbolic .
virtually free groups , being precisely the groups quasi - isometric to trees , are hyperbolic .
a group quasi - isometric to a point or to @xmath138 is finite or virtually @xmath190 respectively ; such groups are called _ elementary hyperbolic groups _ ; all others are _
nonelementary_. fundamental groups of closed surfaces with negative euler characteristic are hyperbolic . by the uniformization theorem , each such surface can be given a hyperbolic metric , exhibiting @xmath191 as a cocompact group of isometries of the hyperbolic plane .
a _ kleinian group _ is a finitely generated discrete subgroup of the group of isometries of hyperbolic @xmath97-space .
a kleinian group @xmath0 is is _ convex cocompact _ if it acts cocompactly on the convex hull of its limit set ( in the sphere at infinity ) .
such a convex hull is @xmath192 , so a convex cocompact kleinian group is hyperbolic .
see e.g. @xcite for an introduction to kleinian groups .
[ invariant_quasiaxis ] let @xmath0 be hyperbolic .
then there are finitely many conjugacy classes of torsion elements ( and therefore a bound on the order of the torsion ) and there are constants @xmath43 so that for any nontorsion element @xmath3 there is a @xmath43 quasigeodesic @xmath27 invariant under @xmath3 on which @xmath3 acts as translation .
let @xmath5 be given .
consider the action of @xmath3 on the cayley graph @xmath2 .
the action is simplicial , so @xmath193 has no strict local minima in the interior of edges , and takes integer values at the vertices ( which correspond to elements of @xmath0 ) .
it follows that there is some @xmath194 for which @xmath195 is minimal , and we can take @xmath194 to be an element of @xmath0 ( i.e. a vertex ) . if @xmath196 then we can join @xmath194 to @xmath197 by a geodesic @xmath198 and let @xmath199 .
note that @xmath3 acts on @xmath27 by translation through distance @xmath174 ; since this is the minimum distance that @xmath3 moves points of @xmath0 , it follows that @xmath27 is a @xmath174-local geodesic ( and therefore a @xmath43 quasigeodesic by lemma [ k_local_geodesics ] ) .
note in this case that @xmath3 has infinite order .
otherwise there is @xmath194 moved a least distance by @xmath3 so that @xmath200 .
since @xmath0 acts cocompactly on itself , there are only finitely many conjugacy classes of elements that move some point any uniformly bounded distance , so if @xmath3 is torsion we are done .
if @xmath3 is not torsion , its orbits are proper , so for any @xmath60 there is an @xmath201 so that @xmath202 ; choose @xmath60 ( and @xmath201 ) much bigger than some fixed ( but big ) @xmath120 .
let @xmath27 be a geodesic from @xmath194 to @xmath203 .
then for any @xmath204 the geodesic @xmath205 has endpoints within distance @xmath206 of the endpoints of @xmath27 . on the other hand
, @xmath207 so @xmath27 contains a segment @xmath198 of length at least @xmath208 such that @xmath209 is contained in the @xmath160 neighborhood of @xmath27 for @xmath210 . to see this ,
consider the quadrilateral with successive vertices @xmath194 , @xmath203 , @xmath211 and @xmath212 .
two nonadjacent sides must contain points which are at most @xmath160 apart . since @xmath213
, the sides must be @xmath27 and @xmath205 .
we find @xmath198 and @xmath209 in the region where these two geodesics are close .
consequently , for any @xmath214 the sequence @xmath215 is a @xmath43 quasigeodesic for some uniform @xmath43 _ independent of @xmath120_. in particular there is a constant @xmath139 ( independent of @xmath120 ) so that @xmath216 for @xmath204 , and therefore the infinite sequence @xmath217 for @xmath218 is an @xmath219-local @xmath43 quasigeodesic .
since @xmath43 is fixed , if @xmath120 is big enough , this infinite sequence is an honest @xmath177 quasigeodesic invariant under @xmath3 , by lemma [ k_local_geodesics ] . here
@xmath177 depends only on @xmath104 and @xmath0 , and not on @xmath3 .
lemma [ invariant_quasiaxis ] can be weakened considerably , and it is frequently important to study actions which are not necessarily cocompact on @xmath104-hyperbolic spaces which are not necessarily proper .
the quasigeodesic @xmath27 invariant under @xmath3 is called a _
quasiaxis_. quasiaxes in @xmath104-hyperbolic spaces are ( approximately ) _ unique _ : let @xmath0 be hyperbolic , and let @xmath3 have infinite order .
let @xmath27 and @xmath220 be @xmath3-invariant @xmath43 quasigeodesics ( i.e. quasiaxes for @xmath3 )
. then @xmath27 and @xmath220 are a finite hausdorff distance apart , and this finite distance depends only on @xmath43 and @xmath104 . consequently the centralizer @xmath221 is virtually @xmath190 .
let @xmath222 and @xmath223 a closest point to @xmath123 .
since @xmath3 acts on both @xmath27 and @xmath220 cocompactly , there is a constant @xmath139 so that every point in @xmath27 or @xmath220 is within @xmath139 from some point in the orbit of @xmath123 or @xmath145 .
this implies that the hausdorff distance from @xmath27 to @xmath220 is at most @xmath224 ; in particular , this distance is finite .
pick two points on @xmath27 very far away from each other ; each is distance at most @xmath224 from @xmath220 , and therefore most of the geodesic between them is within distance @xmath160 of the geodesic between corresponding points on @xmath220 . but
@xmath27 and @xmath220 are themselves @xmath43 quasigeodesic , and therefore uniformly close to these geodesics .
hence some points on @xmath27 are within a uniformly bounded distance of @xmath220 , and therefore all points on @xmath27 are .
if @xmath194 commutes with @xmath3 , then @xmath194 must permute the quasiaxes of @xmath3 .
therefore @xmath194 takes points on any quasiaxis @xmath27 for @xmath3 to within a bounded distance of @xmath27 .
hence @xmath221 , thought of as a subset of @xmath0 , is quasiisometric to a quasiaxis ( that is to say , to @xmath138 ) , and is therefore virtually @xmath190 .
this shows that a hyperbolic group can not contain a copy of @xmath225 ( or , for that matter , the fundamental group of a klein bottle ) .
this is more subtle than it might seem ; @xmath225 _ can _ act freely and properly discontinuously by isometries on a proper @xmath104-hyperbolic space for example , as a parabolic subgroup of the isometries of @xmath226 .
if @xmath227 is a closed @xmath97-manifold , then @xmath228 is hyperbolic if and only if it does not contain any @xmath229 subgroup .
note that this includes the possibility that @xmath228 is elementary hyperbolic ( for instance , finite ) .
this follows from perelman s geometrization theorem @xcite .
if @xmath3 is an isometry of any metric space @xmath14 , the _ translation length _ of @xmath3 is the limit @xmath230 for some @xmath231 .
the triangle inequality implies that the limit exists and is independent of the choice of @xmath123 .
moreover , from the definition , @xmath232 and @xmath233 is a conjugacy invariant .
lemma [ invariant_quasiaxis ] implies that for @xmath0 acting on itself , @xmath234 if and only if @xmath3 has finite ( and therefore bounded ) order .
consequently a hyperbolic group can not contain a copy of a baumslag
solitar group ; i.e. a group of the form @xmath235 . for
, we have already shown hyperbolic groups do not contain @xmath225 , and this rules out the case @xmath236 , and if @xmath237 then for any isometric action of @xmath238 on a metric space , @xmath239 . by properness of @xmath2 and the morse lemma
, there is a constant @xmath201 so that for any @xmath5 the power @xmath240 has an invariant _ geodesic _ axis on which it acts by translation .
it follows that @xmath241 , and in fact @xmath242 ; this cute observation is due to gromov @xcite .
two geodesic rays @xmath243 in a metric space @xmath14 are _ asymptotic _ if they are a finite hausdorff distance apart .
the property of being asymptotic is an equivalence relation , and the set of equivalence classes is the _ gromov boundary _ , and denoted @xmath244 .
if @xmath14 is proper and @xmath104-hyperbolic , and @xmath25 is any basepoint , then every equivalence class contains a ray starting at @xmath25 . for
, if @xmath27 is a geodesic ray , and @xmath245 goes to infinity , then by properness , any collection of geodesics @xmath246 contains a subsequence which converges on compact subsets to a ray @xmath220 . by @xmath104-thinness
each of the triangles @xmath247 is contained in a uniformly bounded neighborhood of @xmath27 , so the same is true of @xmath220 ; in particular , @xmath220 is asymptotic to @xmath27 .
we give @xmath244 the topology of convergence on compact subsets of equivalence classes .
that is , @xmath248 if and only if every subsequence of the @xmath87 contains a further subsequence whose equivalence classes have representatives that converge on compact subsets to some representative of the equivalence class of @xmath27 . [ boundary_compact ]
let @xmath14 be a @xmath104-hyperbolic proper geodesic metric space .
then @xmath244 is compact .
if @xmath87 is any sequence of rays , and @xmath249 is an equivalent sequence starting at a basepoint @xmath25 , then by properness @xmath249 has a subsequence which converges on compact subsets .
in fact , we can define a ( compact ) topology on @xmath250 by saying that @xmath251 if and only if every subsequence of a sequence of geodesics @xmath252 contains a further subsequence which converges on compact subsets to a representative of @xmath27 . with this topology , @xmath253 is compact , @xmath244 is closed in @xmath253 , and the inclusion of @xmath14 into @xmath253 is a homeomorphism onto its image .
a bi - infinite geodesic @xmath27 determines two ( distinct ) points in @xmath244 ; we call these the _ endpoints _ of @xmath27 .
two geodesics with the same ( finite or infinite ) endpoints are hausdorff distance at most @xmath160 apart .
conversely , any two distinct points in @xmath244 are spanned by an infinite geodesic @xmath27 . for ,
if @xmath254 are two infinite rays ( starting at a basepoint @xmath25 for concreteness ) , and @xmath255 are points on @xmath254 respectively going to infinity , some point @xmath146 on any geodesic @xmath256 is within @xmath104 of both @xmath246 and @xmath257 , and if @xmath258 then @xmath259 and @xmath260 would be a finite hausdorff distance apart .
otherwise some subsequence of the @xmath146 converges to @xmath123 , and the geodesics @xmath256 converge on compact subsets to a ( nonempty ! )
bi - infinite geodesic @xmath27 through @xmath123 asymptotic to both @xmath259 and @xmath260 . evidently , geodesic triangles with some or all endpoints at infinity are @xmath165-thin for some @xmath165 depending only on @xmath104 ( one can take @xmath261 ) . by abuse of notation , in the sequel we will call a metric space @xmath104-hyperbolic if all geodesic triangles even those with some endpoints at infinity are @xmath104-thin .
let @xmath262 be hyperbolic geodesic metric spaces .
then any quasi - isometric map
@xmath263 extends uniquely to a _ continuous _ map @xmath264 .
in particular , the gromov boundary @xmath244 depends ( up to homeomorphism ) only on the quasi - isometry type of @xmath14 , and @xmath59 acts on @xmath244 by homeomorphisms . if @xmath0 is a hyperbolic group , we define @xmath265 to be the gromov boundary of some ( any ) @xmath2 .
if @xmath0 is free , @xmath265 is a cantor set .
if @xmath0 is a @xmath191 of a closed surface with negative euler characteristic , @xmath265 is a circle .
if @xmath0 is a convex cocompact kleinian group , @xmath265 is homeomorphic to the limit set .
for example , if @xmath0 is the fundamental group of a hyperbolic @xmath97-manifold with totally geodesic boundary , @xmath265 is a sierpinski carpet .
in fact , a theorem of kapovich
kleiner @xcite says that if @xmath0 is a hyperbolic group which does not split over a finite or virtually cyclic subgroup , and if @xmath265 is @xmath22-dimensional ( in the topological sense of dimension ) , then @xmath265 is homeomorphic to the circle , the sierpinski carpet , or the menger sponge . evidently , @xmath265 is empty if and only if @xmath0 is finite , and if @xmath265 is nonempty , it has at least two points , and has exactly two points if and only if @xmath0 is itself quasi - isometric to the geodesic joining these two points , which holds if and only if @xmath0 is virtually @xmath190 .
if @xmath5 has infinite order , a quasiaxis @xmath27 is asymptotic to two points @xmath266 . under ( positive ) powers of @xmath3 , points stay a constant distance from @xmath27 , and move towards one of the endpoints , say @xmath267 . as homeomorphisms from @xmath253 to itself , the elements @xmath268 with @xmath269 converge uniformly ( in the compact - open topology ) on @xmath270 to the constant map to @xmath267 .
we call @xmath267 the _ attracting _ endpoint and @xmath271 the _ repelling _ endpoint of @xmath3 ; the actions of @xmath3 on @xmath265 is sometimes expressed by saying that it has _ source - sink dynamics_. [ free_from_pingpong ] let @xmath76 be of infinite order , with quasiaxes @xmath27 and @xmath220 . if @xmath27 and @xmath220 share an endpoint ( without loss of generality the attracting endpoint of each ) and @xmath123 is close to both @xmath27 and @xmath220 , then there are @xmath272 for which @xmath273 is bounded .
since the action of @xmath0 on its cayley graph is properly discontinuously , it follows that there are distinct @xmath274 with @xmath275 so that @xmath276 for some positive @xmath277 .
in particular , in this case @xmath3 and @xmath194 together generate a virtual @xmath190 subgroup , and their quasiaxes have the same endpoints .
otherwise the endpoints are disjoint , and because of the source - sink dynamics , klein s pingpong argument implies that sufficiently large powers @xmath278 generate a ( nonabelian ) _ free _ subgroup of @xmath0 .
[ action_minimal ] suppose @xmath0 is nonelementary . then the action of @xmath0 on @xmath265 is minimal ; i.e. every orbit is dense .
consequently @xmath265 is infinite and perfect .
if @xmath0 is nonelementary , there are @xmath279 whose quasiaxes have distinct endpoints @xmath280 and @xmath281 respectively .
if @xmath282 is arbitrary , then either @xmath283 or @xmath284 ; it follows that every attracting / repelling point is in the closure of every orbit .
now let @xmath27 be a geodesic from @xmath271 to @xmath267 and let @xmath220 be a geodesic ray asymptotic to @xmath131 .
pick @xmath285 on @xmath27 and let @xmath79 be a sequence of elements with @xmath286 converging to @xmath131 . at most one component of @xmath287 can come close to the basepoint @xmath25 .
hence there is some subsequence so that either @xmath288 or @xmath289 , and therefore every point is in the closure of the orbit of some attracting / repelling point .
this proves the lemma .
another way to see the compactification @xmath244 is in terms of ( equivalence classes of ) _
let @xmath27 be a geodesic ray parameterized by length .
the _ horofunction _ ( also called the _ busemann function _ ) associated to @xmath27 is the limit @xmath290 the level sets of horofunctions are called _
horospheres_. this limit exists and is finite , by the triangle inequality .
moreover , it is @xmath22-lipschitz .
if @xmath27 and @xmath220 are asymptotic , then there is some constant @xmath291 so that @xmath292 .
if @xmath25 is the endpoint of @xmath27 , we let @xmath293 denote any horofunction of the form @xmath294 , and say that @xmath293 is _ centered at @xmath25_. here is another way to define @xmath294 without reference to @xmath27 . on any proper metric space , the set of @xmath22-lipschitz functions mod constants is compact ( in the topology of convergence on compact subsets ) . for any @xmath51
the function @xmath295 is @xmath22-lipschitz , and @xmath296 embeds @xmath14 in the space of @xmath22-lipschitz functions on @xmath14 mod constants .
the closure of this image defines a natural compactification of @xmath14 ; quotienting further by bounded functions gives @xmath253 . for each @xmath297
the preimage is the set of equivalence classes of functions @xmath293 . in this way we think of @xmath293 as a normalization of the function which measures `` distance to @xmath25 '' .
the space @xmath244 can be metrized following gromov ( see @xcite ) .
[ a_distance ] fix some basepoint @xmath25 and some constant @xmath298 .
the _ @xmath299-length _ of a rectifiable path @xmath27 in @xmath14 is the integral along @xmath27 of @xmath300 , and the _ @xmath299-distance _ from @xmath26 to @xmath301 , denoted @xmath302 , is the infimum of the @xmath299-lengths of paths between @xmath26 and @xmath301 .
a straightforward calculation shows that there is an @xmath303 so that for @xmath304 , the @xmath299-length defines a metric on @xmath253 .
in fact , any @xmath305 with @xmath306 will work .
if @xmath299 is too big , @xmath299-length still extends to a pseudo - metric on @xmath253 , but now distinct points of @xmath244 might be joined by a sequence of paths with @xmath299-length going to @xmath24 . increasing @xmath299 will decrease the hausdorff dimension of @xmath244 ; of course , the hausdorff dimension must always be at least as big as the topological dimension . in any case
, it follows that @xmath244 is metrizable .
the following lemma is useful to compare length and @xmath299-length .
[ a_comparison_lemma ] for @xmath304 there is a constant @xmath307 so that for all points @xmath308 there is an inequality @xmath309 where @xmath310 is the ordinary distance from the basepoint @xmath25 to the geodesic @xmath102 . for a proof , see @xcite .
the quantity @xmath310 is sometimes abbreviated by @xmath311 ( the basepoint @xmath25 is suppressed in this notation ) , and called the _
gromov product_. so we can also write @xmath312 .
because of this inequality , different choices of @xmath299 give rise to hlder equivalent metrics on @xmath244 .
if @xmath14 is a group @xmath0 , we take @xmath313 as the basepoint , by convention . with our notation , @xmath314 is ambiguous , since it depends on a choice of geodesic from @xmath26 to @xmath301 .
since we only care about @xmath311 up to a uniform additive constant , we ignore this issue .
one common normalization , adopted by gromov , is to use the formula @xmath315 .
these definitions are interchangeable for our purposes , as the ambiguity can always be absorbed into some unspecified constant .
a group @xmath0 acting by homeomorphisms on a compact metrizable space @xmath227 is said to be a _ convergence action _ if the induced action on the space @xmath316 of distinct ordered triples is properly discontinuous .
[ action_convergence ] the action of @xmath0 on @xmath265 is a convergence action .
moreover , the action on the space of distinct triples is cocompact .
if @xmath100 is a distinct triple of points in @xmath265 , there is a point @xmath123 within distance @xmath104 of all three geodesics @xmath317 ; moreover , the set of such points has uniformly bounded diameter in @xmath0 .
this defines an approximate map from distinct triples to points in @xmath0 .
since the action of @xmath0 on itself is cocompact , the same is true for the action on the space of distinct triples .
similarly , if the action of @xmath0 on the space of distinct triples were not properly discontinuous , we could find two bounded regions in @xmath0 and infinitely many @xmath79 in @xmath0 taking some point in one bounded region to some point in the other , which is absurd .
the converse is a famous theorem of bowditch : [ bowditch_theorem ] let @xmath0 act faithfully , properly discontinuously and cocompactly on the space of distinct triples of some perfect compact metrizable space @xmath227 .
then @xmath0 is hyperbolic and @xmath227 is @xmath0-equivariantly homeomorphic to @xmath265 .
the results in this section are due to coornaert @xcite , although because of our more narrow focus we are able to give somewhat different and shorter proofs . however by and large our proofs , like coornaert s , are obtained by directly generalizing ideas of sullivan @xcite in the context of kleinian groups .
let @xmath0 be a hyperbolic group , and let @xmath318 denote the set of elements of ( word ) length @xmath319 , with respect to some fixed generating set .
the _ critical exponent _
@xmath320 ( also called the _ volume entropy _ of @xmath0 ) is the quantity @xmath321 in other words , the exponential growth rate of @xmath0 . since every nonelementary hyperbolic group contains many free groups ( example [ free_from_pingpong ] ) , @xmath322 if and only if @xmath0 is elementary .
define the ( poincar ) _ zeta function _ by the formula @xmath323 then @xmath324 diverges if @xmath325 and converges if @xmath326 .
[ zeta_diverges ] the zeta function diverges at @xmath327 .
we will show in [ combing_section ] ( theorem [ cannon_theorem ] ) that for any hyperbolic group @xmath0 and any generating set @xmath1 there is a _ regular language _
@xmath328 consisting of geodesics , which evaluates bijectively to @xmath0 .
in particular , @xmath329 for any @xmath120 . in any regular language @xmath21 the generating function @xmath330 is _ rational _ ( theorem [ regular_language_generating_function ] ) ; i.e. it is the power series expansion of @xmath331 for some integral polynomials @xmath332 , and consequently @xmath333 for some real @xmath194 and non - negative integer @xmath174 , and constant @xmath139 . evidently , for @xmath21 as above , @xmath334 and the zeta function diverges at @xmath194 . for @xmath326
construct a probability measure @xmath335 on @xmath336 ( i.e. on @xmath337 ) supported in @xmath0 , by putting a dirac mass of size @xmath338 at each @xmath5 . as @xmath285 converges to @xmath194 from above
, this sequence of probability measures contains a subsequence which converges to a limit @xmath339 . since the zeta function diverges at @xmath194 , the limit @xmath339 is supported on @xmath340 .
this measure is called a _ patterson sullivan _ measure , by analogy with the work of patterson and sullivan @xcite on kleinian groups . for any @xmath3 ,
the pushforward of measure @xmath341 is defined by @xmath342 , and similarly for @xmath343 . for any @xmath344
there is an inequality @xmath345 . from the definition on @xmath335
, this implies that @xmath341 is absolutely continuous with respect to @xmath335 , and its radon nikodym derivative satisfies @xmath346 .
passing to a limit we deduce that @xmath347 .
the most important property of the measure @xmath339 is a refinement of this inequality , which can be expressed by saying that it is a so - called _ quasiconformal measure of dimension @xmath194_. the `` conformal '' structure on @xmath244 is defined using the @xmath299-distance for some fixed @xmath298 ( recall definition [ a_distance ] ) . for @xmath5
define @xmath348 by @xmath349 for some horofunction @xmath350 centered at @xmath26 .
a probability measure @xmath339 on @xmath244 is a _
quasiconformal measure of dimension @xmath351 _ if @xmath343 is absolutely continuous with respect to @xmath339 for every @xmath5 , and there is some constant @xmath139 independent of @xmath3 so that @xmath352 notice that the ambiguity in the choice of horofunction @xmath350 is absorbed into the definition of @xmath353 ( which only depends on @xmath350 mod constant functions ) and the constant @xmath139 .
the support of any quasiconformal measure is evidently closed and @xmath0-invariant , so by lemma [ action_minimal ] , it is all of @xmath265 . from the definition of the radon
nikodym derivative , @xmath339 is a quasiconformal measure of dimension @xmath351 if there is a constant @xmath139 so that for all @xmath26 we can find a neighborhood @xmath354 of @xmath26 in @xmath253 for which @xmath355 for all @xmath356 .
for some reason , coornaert chooses to work with _ pullbacks _ of measure @xmath357 instead of pushforward .
therefore the roles of @xmath3 and @xmath358 are generally interchanged between our discussion and coornaert s .
[ measure_is_conformal ] the measure @xmath339 is a quasiconformal measure of dimension @xmath351 where @xmath359 . evidently the support of @xmath339 is @xmath0-invariant , and is therefore equal to all of @xmath265 .
let @xmath360 , let @xmath350 be a horofunction centered at @xmath26 , and let @xmath5 . by @xmath104-thinness and the definition of a horofunction
, @xmath361 is close to @xmath362 for @xmath301 sufficiently close to @xmath26 . in particular
, there is a neighborhood @xmath354 of @xmath26 in @xmath253 so that @xmath363 for some @xmath139 , and for all @xmath301 in @xmath354 .
for each @xmath364 we have @xmath365 .
taking the limit as @xmath366 and defining @xmath351 by @xmath367 proves the theorem . to make use of this observation
, we introduce the idea of a _ shadow _ , following sullivan . for @xmath5 and @xmath12 a positive real number , the _ shadow _ @xmath368 is the set of @xmath369 such that every geodesic ray from @xmath313 to @xmath26 comes within distance @xmath12 of @xmath3 .
said another way , @xmath26 is in @xmath368 if @xmath3 comes within distance @xmath12 of any geodesic from @xmath313 to @xmath26 .
given @xmath370 , for any fixed @xmath120 the shadows @xmath368 with @xmath371 cover @xmath265 efficiently : [ uniform_covering_by_shadows ] fix @xmath12 .
then there is a constant @xmath201 so that for any @xmath369 and any @xmath120 there is at least @xmath22 and there are at most @xmath201 elements @xmath3 with @xmath371 and @xmath372 . if @xmath370 , if @xmath27 is any geodesic from @xmath313 to @xmath26 , and if @xmath3 is any point on @xmath27 , then @xmath373 .
conversely , if @xmath3 and @xmath194 are any two elements with @xmath374 and @xmath375 then @xmath376 .
sullivan s fundamental observation is that the action of @xmath358 on @xmath368 is uniformly close to being _ linear _ , in the sense that the derivative @xmath377 varies by a bounded multiplicative constant on @xmath368 : [ linear_on_shadow ] fix @xmath12 .
then there is a constant @xmath139 so that for any @xmath373 there is an inequality @xmath378 recall @xmath379 for some horofunction @xmath350 .
but by @xmath104-thinness and the definition of a shadow , there is a constant @xmath380 so that @xmath381 for any @xmath26 in @xmath368 . from this
one readily obtains a uniform estimate on the measure of a shadow : [ size_of_shadow ] fix @xmath12 .
then there is a constant @xmath139 so that for any @xmath382 there is an inequality @xmath383 let @xmath384 be the measure of the biggest atom of @xmath339 , and fix @xmath385 . by compactness of @xmath265
there is some @xmath46 so that every ball in @xmath265 of diameter @xmath386 has mass at most @xmath387 .
now , @xmath388 is the set of @xmath389 for which every geodesic ray from @xmath358 to @xmath26 comes within distance @xmath12 of @xmath313 . as @xmath390
, the diameter of @xmath391 goes to zero uniformly in @xmath3 ( this follows from the quasi - equivalence of @xmath302 and @xmath392 ; see lemma [ a_comparison_lemma ] ) .
consequently there is some @xmath393 so that for all @xmath394 the measure @xmath395 is between @xmath396 and @xmath22 , independent of @xmath3 .
now , by lemma [ linear_on_shadow ] and the definition of a quasiconformal measure , there is a constant @xmath63 so that @xmath397 taking reciprocals , and using the fact that @xmath398 completes the proof ( at the cost of adjusting constants ) . note that the argument shows that @xmath339 has no atoms , since any @xmath369 is contained in some shadow of measure @xmath399 for any @xmath120 . we deduce the following corollary .
[ coornaert_growth_rate ] let @xmath0 be a hyperbolic group .
then there is a constant @xmath139 so that @xmath400 for all @xmath120 .
the lower bound is proved in
[ combing_section ] , so we just need to prove the upper bound . for each @xmath3 with @xmath371 lemma [ size_of_shadow ] says @xmath401 . on the other hand
, lemma [ uniform_covering_by_shadows ] says that every point @xmath402 is in at most @xmath201 sets @xmath368 with @xmath371 , so @xmath403 corollary [ coornaert_growth_rate ] has important consequences that we will explore in [ combing_section ] .
a second corollary gives very precise metric and dynamical control over @xmath265 .
an action of a group @xmath0 on a space @xmath14 is said to be _ ergodic _ for some measure @xmath339 on @xmath14 if for any two subsets @xmath404 , @xmath405 of @xmath14 with @xmath406 there is some @xmath5 with @xmath407 .
[ ergodic_at_infinity ] let @xmath339 be a quasiconformal measure on @xmath265 of dimension @xmath351 .
then @xmath339 is quasi - equivalent to @xmath351-dimensional hausdorff measure ; i.e. there is a constant @xmath139 so that @xmath408 for any @xmath404 . in particular , the space @xmath265 has hausdorff dimension @xmath351 , and its @xmath351-dimensional hausdorff measure is finite and positive .
moreover , the action of @xmath0 on @xmath340 is ergodic for @xmath339 .
evidently , the second and third claims follow from the first ( if @xmath404 is a @xmath0-invariant subset of @xmath265 of positive @xmath339-measure , the restriction of @xmath339 to @xmath404 is a quasiconformal measure of dimension @xmath351 , and is therefore quasi - equivalent to @xmath409 and thence to @xmath339 . in particular , @xmath404 has full measure ) .
so it suffices to show that @xmath339 and @xmath409 are quasi - equivalent .
since @xmath410 it follows that every metric ball @xmath411 in @xmath265 can be sandwiched between two shadows @xmath412 with @xmath413 and @xmath414 . from lemma [ size_of_shadow ] we obtain @xmath415 . from this and the definition of hausdorff measure
, we will obtain the theorem .
if @xmath404 is any measurable set , cover @xmath404 by balls @xmath416 of radius @xmath417 .
then @xmath418 so letting @xmath419 we get @xmath420 .
the following proof of the reverse inequality was suggested to us by curt mcmullen . for any @xmath104 let @xmath45 be compact and @xmath421 open so that @xmath422 and both @xmath423 and @xmath424 are less than @xmath104 . by compactness
, there is an @xmath46 so that every ball of radius @xmath386 centered at a point in @xmath45 is contained in @xmath421 .
now inductively cover @xmath45 by balls @xmath425 of non - increasing radius @xmath417 in such a way that the center of each @xmath416 is not in @xmath426 for and @xmath427 .
then the balls with the same centers and half the radii are disjoint , so @xmath428 and therefore @xmath429 .
taking @xmath430 gives @xmath431 and we are done .
coornaert only gives the proof of the inequality @xmath432 in his paper , referring the reader to sullivan @xcite for the proof of @xmath433 .
however , there is a gap in sullivan s proof of the reverse inequality , of which the reader should be warned .
the approximate linearity of @xmath358 on @xmath368 has many other applications .
for example , see the proof of theorem 1 in @xcite .
on a riemannian manifold , a `` geodesic '' is just a smooth path that locally minimizes length ( really , energy ) .
a sufficiently long geodesic is typically not globally length minimizing , and the entire subject of morse theory is devoted to the difference . by contrast ,
one of the most important qualitative features of negative curvature is that ( quasi)-geodesity is a _ local _ property ( i.e. lemma [ k_local_geodesics ] ) .
this localness translates into an important combinatorial property , known technically as _ finiteness of cone types_. this is the basis of cannon s theory of hyperbolic groups , and for the more general theory of automatic groups and structures ( see @xcite for more details ) .
let @xmath1 be a finite set , and let @xmath434 denote the set of finite words in the alphabet @xmath1 .
an _ automaton _ is a finite directed graph @xmath166 with a distinguished initial vertex , and edges labeled by elements of @xmath1 in such a way that each vertex has at most one outgoing edge with a given label .
some subset of the vertices of @xmath166 are called _
accept states_. a word @xmath435 is @xmath434 determines a simplicial path in @xmath166 , by starting at the initial vertex , by reading the letters of @xmath435 ( from left to right ) one by one , and by moving along the corresponding edge of @xmath166 if it exists , or halting if not .
associated to @xmath166 there is a subset @xmath436 consisting of precisely those words that can be read in their entirety without halting , and for which the terminal vertex of the associated path ends at an accept state .
one says that @xmath21 is _ parameterized _ by ( paths in ) @xmath166 .
a subset @xmath436 is a _ regular language _ if there is a finite directed graph @xmath166 as above that parameterizes @xmath21 .
note that @xmath166 is not part of the data of a regular language , and for any given regular language there will be many graphs that parameterize it .
a language is _ prefix - closed _ if , whenever @xmath437 , every prefix of @xmath435 is also in @xmath21 ( the empty word is a prefix of every word ) .
if @xmath21 is prefix - closed and regular , there is a @xmath166 parameterizing @xmath21 for which every vertex is an accept state .
if @xmath166 is any graph that parameterizes @xmath21 , remove all non - accept vertices and the edges into and out of them .
[ regular_language_generating_function ] let @xmath21 be a regular language , and for each @xmath120 , let @xmath438 denote the set of elements of length @xmath120 , and @xmath439 the set of elements of length @xmath319 .
let @xmath440 and @xmath441 be ( formal ) generating functions for @xmath442 and @xmath443 respectively .
then @xmath444 and @xmath445 are _ rational _ ; i.e. they agree as power series expansions with some ratio of integral polynomials in @xmath446 .
note that @xmath447 so it suffices to prove the theorem for @xmath444 .
let @xmath166 parameterize @xmath21 , and let @xmath227 be the adjacency matrix of @xmath166 ; i.e. @xmath448 is equal to the number of edges from vertex @xmath449 to vertex @xmath450 .
let @xmath451 be the vector with a @xmath22 in the initial state , and @xmath24 elsewhere , and let @xmath452 be the vector with a @xmath22 in every accept state , and @xmath24 elsewhere
. then @xmath453 .
a formal power series @xmath454 is rational if and only if its coefficients satisfy a linear recurrence ; i.e. if there are constants @xmath455 ( not all zero ) so that @xmath456 for all @xmath457 . for , @xmath458 vanishes in degree @xmath459 , and is therefore a polynomial ( reversing this argument proves the converse ) . if @xmath460 is the characteristic polynomial of @xmath227 , then @xmath461 , and @xmath462 proving the theorem .
another way of expressing @xmath444 , more useful in some ways , is as follows .
let @xmath21 be a regular language .
then there is an integer @xmath351 so that for each value of @xmath120 mod @xmath351 either @xmath442 is eventually zero , or there are finitely many constants @xmath463 and polynomials @xmath146 so that @xmath464 for all sufficiently large @xmath120 . for a proof see e.g. @xcite thm .
in particular , either @xmath443 has polynomial growth , or @xmath465 for some real @xmath307 and integer @xmath174 , and constant @xmath139 .
let @xmath1 be a set .
a total order @xmath466 on @xmath1 extends to a unique lexicographic ( or dictionary ) order on @xmath434 as follows :
1 . the empty word precedes everything ; 2 .
if @xmath467 and @xmath468 are both nonempty and start with different letters @xmath469 then @xmath470 if and only if @xmath471 ; and 3 .
if @xmath470 and @xmath435 is arbitrary , then @xmath472 .
if @xmath0 is a group and @xmath1 is a generating set for @xmath0 , there are finitely many geodesic words representing any given element ; the lexicographically first geodesic is therefore a canonical representative for each element of @xmath3 , and determines a language @xmath436 that bijects with @xmath0 under evaluation .
we denote evaluation by overline , so if @xmath473 , we denote the corresponding element of @xmath0 by @xmath474 .
we similarly denote length of an element of @xmath434 by @xmath475 .
so we always have @xmath476 with equality if and only if @xmath467 is geodesic .
given @xmath5 the _ cone type _ of @xmath3 , denoted @xmath477 , is the set of @xmath478 for which some geodesic from @xmath313 to @xmath197 passes through @xmath3 .
for any @xmath120 , the _ @xmath120-level _ of @xmath3 is the set of @xmath194 in the ball @xmath479 such that @xmath480 .
cannon showed that the @xmath120 level ( for @xmath120 sufficiently large ) determines the cone type , and therefore that there are only _ finitely many cone types_. [ bounded_level_determines_cone_type ] the @xmath481 level of an element determines its cone type .
let @xmath3 and @xmath194 have the same @xmath481 level , and let @xmath482 be geodesics with @xmath483 and @xmath484 .
only @xmath313 has an empty @xmath481 level , so we may assume @xmath482 both have length @xmath485 .
we prove the lemma by induction .
suppose @xmath486 , @xmath487 and @xmath488 are geodesics , where @xmath6 .
we must show that @xmath489 is a geodesic .
suppose to the contrary that there is some @xmath490 where @xmath491 and @xmath492 .
then @xmath493 is in the @xmath481 level of @xmath194 , which agrees with the @xmath481 level of @xmath3 , and therefore @xmath494 .
but then concatenating a geodesic representative of @xmath495 with @xmath496 gives a shorter path to @xmath497 , certifying that @xmath488 is not geodesic , contrary to assumption .
@xmath3 at 70 107 @xmath194 at 303 82 @xmath496 at 250 160 @xmath496 at 14 185 @xmath498 at 343 245 @xmath497 at 107 270 from the @xmath481 level of @xmath194 to @xmath498 gives a shortcut from the @xmath481 level of @xmath3 to @xmath497 .
this figure is adapted from @xcite.,title="fig : " ] the following theorem is implicit in @xcite , though expressed there in somewhat different language .
[ cannon_theorem ] let @xmath0 be a hyperbolic group , and @xmath1 a symmetric generating set . fix a total order @xmath466 on @xmath1 .
then the language of lexicographically first geodesics is prefix - closed and regular . that this language is prefix - closed is obvious .
we show it is regular by describing an explicit parameterizing graph . as a warm up ,
we show first that the language of _ all _ geodesics is regular .
a parameterizing graph can be taken as follows .
the vertices ( all accept states ) are precisely the set of cone types , and there is an edge labeled @xmath285 from a cone type of the form @xmath477 to one of the form @xmath499 whenever @xmath500 . by the definition of cone types ,
this is well - defined . by lemma [ bounded_level_determines_cone_type ] , the number of cone types is finite , so this is a finite graph . by construction
, this graph exactly parameterizes the language of all geodesics
. now fix a total order @xmath466 on @xmath1 . for each @xmath5 , let @xmath501 be the lexicographically first geodesic from @xmath313 to @xmath3 . for each @xmath5 a _ competitor _ of @xmath3 is some @xmath194 with @xmath502 , with @xmath503 , and for which @xmath504 for all @xmath449 , where @xmath505 denotes the prefix of @xmath501 of length @xmath449 , and similarly for @xmath506 ( this is described by saying that _ @xmath507 synchronously fellow - travels @xmath501 _ ) . if there is some @xmath508 with @xmath509 and @xmath510 then by @xmath104-thinness and the definition of geodesics , @xmath507 synchronously fellow - travels @xmath501 .
it follows that for all @xmath5 and @xmath6 we have @xmath511 if and only if @xmath512 is a geodesic , and there is no competitor @xmath194 of @xmath3 and @xmath513 so that @xmath514 . given @xmath5
define @xmath515 to be the set of @xmath194 for which @xmath197 is a competitor of @xmath3 .
associated to @xmath3 is the list @xmath516 of pairs @xmath517 together with the cone type of @xmath3 itself .
note that the set of possible lists @xmath516 is _
finite_. we can now define a parameterizing graph by taking the vertices ( all accept states ) to be the possible lists @xmath516 , and there is an edge labeled @xmath285 from a list of the form @xmath516 to a list of the form @xmath518 if and only if @xmath500 , and there is no @xmath519 and @xmath513 with @xmath520 .
this is evidently a finite directed graph , which parameterizes the @xmath501 ; we must show it is well - defined .
first of all , @xmath521 if and only if one of the two following possibilities occurs : 1 . there is some @xmath522 and @xmath523 with @xmath524 ; or 2 .
there is some @xmath525 in @xmath526 with @xmath527 .
both of these possibilities depend only on @xmath221 , @xmath477 or @xmath528 for some @xmath529 , and not on @xmath3 itself .
second of all , if @xmath530 , then @xmath531 depends only on @xmath528 and @xmath180 in the first case , and on @xmath477 and @xmath180 in the second case .
this shows the graph is well - defined , and completes the proof of the theorem .
this completes the proof of lemma [ zeta_diverges ] , and the subsequent results in [ patterson_sullivan_subsection ] .
the reader will note that the results in this section do not depend on lemma [ zeta_diverges ] , so our reasoning has not been circular .
let @xmath0 be a group , and @xmath1 a generating set .
a _ combing _ for @xmath0 ( with respect to @xmath1 ) is a prefix - closed regular language @xmath328 which bijects with @xmath0 under evaluation , and satisfies @xmath532 for all @xmath533 ( i.e. @xmath21 is a language of geodesics ) .
theorem [ cannon_theorem ] says that every hyperbolic group admits a combing . if @xmath21 is a combing with respect to @xmath1 , the _
@xmath21-cone type _ of @xmath3 , denoted @xmath534 , is the set of @xmath478 for which the @xmath21-geodesic evaluating to @xmath197 contains a prefix ( which is also an @xmath21-geodesic ) evaluating to @xmath3 .
there is a graph @xmath166 parameterizing @xmath21 with one vertex for each @xmath21-cone type , and an edge from @xmath534 to @xmath535 labeled @xmath285 whenever @xmath536 .
the reader should be warned that many competing definitions of combing exist in the literature .
suppose @xmath21 is a combing of @xmath0 , and @xmath166 is a graph parameterizing @xmath21 , so that there is a ( length - preserving ) bijection between directed paths in @xmath166 starting at the initial vertex , and words of @xmath21 , by reading the edge labels of the path . if @xmath533 , we let @xmath537 denote the corresponding path in @xmath166 , and @xmath538 the successive vertices in @xmath166 visited by @xmath537 .
a function @xmath539 is _ weakly combable _ with respect to a combing @xmath21 if there is a graph @xmath166 parameterizing @xmath21 and a function @xmath540 from the vertices of @xmath166 to @xmath190 so that @xmath541 for all @xmath533 .
a function @xmath542 is _ combable _ if it is weakly combable with respect to some combing @xmath21 , and if there is a constant @xmath139 so that @xmath543 for all @xmath5 and @xmath6 ; and it is _ bicombable _ if it is combable , and further satisfies @xmath544
. it might be more natural to define a function @xmath540 on the _ edges _ of @xmath166 instead of its vertices ; however , associated to any directed graph @xmath166 there is another graph the _ line graph _ of @xmath166 whose vertices are the edges of @xmath166 , and whose edges are the composable pairs of edges of @xmath166 , and the line graph of @xmath166 parameterizes @xmath21 if @xmath166 does .
the property of being combable or bicombable does not depend on the choice of a generating set or a combing .
the proof proceeds along the same lines as theorem [ cannon_theorem ] .
the key point is that words in @xmath21 are ( uniformly ) quasigeodesic with respect to @xmath545 , and therefore stay within a bounded distance of words in @xmath546 with the same evaluation .
therefore an automaton reading the letters of an @xmath546-word can keep track of the states of a collection of automata simultaneously reading nearby @xmath21-words , and keeping track of how @xmath542 changes as one goes along .
see @xcite for details .
[ word_length_example ] word length in any generating set is bicombable .
in fact , if @xmath1 is a ( possibly unsymmetric ) set which generates @xmath0 as a semigroup , word length in @xmath1 is bicombable
. one can generalize word length by giving different generators ( and corresponding edges in the cayley graph ) different lengths ; providing the lengths are all integral and positive , the resulting ( geodesic ) word length is bicombable .
[ sum_example ] the sum or difference of two ( bi)combable functions is ( bi)combable .
the following definition is due to epstein
fujiwara @xcite ( also see @xcite ) .
let @xmath198 be a path in @xmath2 .
a _ copy _ of @xmath198 is a translate @xmath547 for some @xmath5 . given a path @xmath27 in @xmath548 ,
define @xmath549 to be the maximal number of disjoint copies of @xmath198 in @xmath27 , and for @xmath5 define the _ small counting function _ @xmath550 by the formula @xmath551 counting functions are bicombable .
in fact , we can add @xmath198 to @xmath1 as a ( semigroup ) generator , but insist that the ( directed ) edges labeled @xmath198 have length @xmath552 instead of @xmath22 ; this defines a new distance function @xmath553 which is bicombable ( by example [ word_length_example ] ) , and therefore so is the difference @xmath554 ( by example [ sum_example ] ) . many variations on this idea are possible ; for instance , the `` big '' counting functions @xmath555 which count _ all _ copies of @xmath198 in @xmath27 , not just the maximal number of disjoint copies . a directed graph @xmath166
is sometimes called a _ topological markov chain_. a topological markov chain can be promoted to a genuine ( stationary ) markov chain by assigning probabilities to each edge in such a way that the probabilities on the edges leaving each vertex sum to @xmath22 . recall that we write the adjacency matrix as @xmath227 ; we think of this as an endomorphism of the vector space @xmath354 spanned by the states of @xmath166 .
let @xmath556 denote the vector with all components equal to @xmath22 , and let @xmath557 denote the vector corresponding to the initial state .
two states in a topological markov chain are said to be _ communicating _ if there is a directed path from each to the other .
the property of being communicating is an equivalence relation .
we write @xmath558 for equivalence classes
@xmath63 if there is a directed path from some ( any ) vertex of @xmath63 to some ( any ) vertex of @xmath64 ; observe that @xmath559 is a partial order .
we call each equivalence class a _ component_. the induced ( directed ) subgraph associated to a component @xmath139 is itself a topological markov chain .
its adjacency matrix @xmath560 has the property that for any @xmath449 and @xmath450 there is an @xmath120 ( in fact , infinitely many @xmath120 ) so that @xmath561 is positive ; one says such a markov chain is _
irreducible_. if there is a fixed @xmath120 so that @xmath561 is positive for all @xmath449 , @xmath450 we say the markov chain is _ aperiodic _ ; this holds exactly when the gcd of the lengths of all loops in @xmath139 is @xmath22 . a markov chain ( on a finite state space ) which is both irreducible and aperiodic is _
ergodic_. [ perron_frobenius ] let @xmath227 be a real matrix with positive entries .
then there is a unique eigenvalue @xmath307 of biggest absolute value , and this eigenvalue is real and positive .
moreover , @xmath307 is a simple root of the characteristic polynomial , and it has a right ( left ) eigenvector with all components positive , unique up to scale .
finally , any other non - negative right ( left ) eigenvector is a multiple of the @xmath307 eigenvector .
since the entries of @xmath227 are positive , @xmath227 takes the positive orthant strictly inside itself .
the projectivization of the positive orthant is a simplex , and therefore @xmath227 takes this simplex strictly into its interior .
it follows that @xmath227 has a unique attracting fixed point in the interior this simplex ; this fixed point corresponds to the unique eigenvector @xmath468 ( up to scale ) with non - negative entries , and its entries are evidently all positive , and its associated eigenvalue @xmath307 is real and positive . if @xmath136 is any plane containing this unique positive eigenvector , the projectivization of @xmath136 is an @xmath562 ; since the eigenvector becomes an attracting fixed point in this @xmath562 , it is not the only fixed point .
this shows that @xmath307 is a simple eigenvalue ; a similar argument shows that @xmath563 is not an eigenvalue .
let @xmath564 be any other eigenvalue . if @xmath564 is real , then @xmath565 .
suppose @xmath564 is complex , acting as composition of a dilation with a rotation on some plane @xmath136 .
if @xmath566 then the restriction of @xmath227 to @xmath567 acts projectively like a rotation ; but this contradicts the fact that @xmath468 is a projective attracting fixed point . this proves the theorem .
if @xmath227 is non - negative , there is still a non - negative real eigenvector @xmath468 with a real positive eigenvalue @xmath307 , and every other eigenvalue @xmath564 satisfies @xmath568 . in this generality
, @xmath307 might have multiplicity @xmath569 , and the jordan block associated to @xmath307 might not be diagonal .
however if @xmath227 is _ irreducible _ , then @xmath307 has multiplicity @xmath22 , the eigenvector @xmath468 is strictly positive , and every other eigenvalue with absolute value @xmath307 is simple and of the form @xmath570 .
these facts can be proved similarly to the proof of lemma [ perron_frobenius ] now let @xmath0 be a hyperbolic group , @xmath21 a combing with respect to some generating set , and @xmath166 a graph parameterizing @xmath21 .
let @xmath571 be the quotient directed graph whose vertices are the components of @xmath166 .
note that @xmath571 contains no directed loops .
associated to each vertex of @xmath571 is an adjacency matrix @xmath560 which has a unique maximal real eigenvalue @xmath572 of multiplicity @xmath22 .
we let @xmath573 , and we call a component _ maximal _ if @xmath574
. the next lemma is crucial to what follows , and depends on coornaert s estimate of the growth function ( i.e. corollary [ coornaert_growth_rate ] ) .
[ maximal_components_parallel ] the maximal components do not occur in parallel ; that is , there is no directed path from any maximal component to a distinct maximal component .
since there is a directed path from the initial vertex to every other vertex , the number of paths of length @xmath120 is of the form @xmath575 for polynomials @xmath123 , @xmath122 and @xmath576 , where @xmath307 is as above .
moreover , the degree of @xmath123 is one less than the length of the biggest sequence of maximal components @xmath577 .
the number of paths of length @xmath120 is equal to the number of elements of @xmath0 of length @xmath120 , so corollary [ coornaert_growth_rate ] implies that the degree of @xmath123 is zero .
it follows that all but exponentially few paths @xmath27 of length @xmath120 in @xmath27 are entirely contained in one of the maximal components of @xmath166 , except for a prefix and a suffix of length @xmath578 .
consequently , the properties of a `` typical '' path in @xmath166 can be inferred from the properties of a `` typical '' path conditioned to lie in a single component . for any vector @xmath468 ,
the limits @xmath579 exist , and are the projections onto the left and right @xmath307-eigenspaces respectively .
heuristically , @xmath580 is the distribution of endpoints of long paths that start with distribution @xmath468 , and @xmath581 is the distribution of starting points of long paths that end with distribution @xmath468 .
recall that @xmath557 denotes the vector with a @xmath22 in the coordinate corresponding to the initial vertex and @xmath24s elsewhere , and @xmath556 denotes the vector with all coordinates equal to @xmath22 .
define a measure @xmath582 on the vertices of @xmath166 by @xmath583 , and scale @xmath582 to a probability measure @xmath564 .
define a matrix @xmath201 by @xmath584 if @xmath585 , and define @xmath586 otherwise .
[ n_preserves_mu ] the matrix @xmath201 is a stochastic matrix ( i.e. it is non - negative , and the rows sum to @xmath22 ) and preserves the measure @xmath564 . if @xmath587 then @xmath588 by fiat .
otherwise @xmath589 . to see that @xmath201 preserves @xmath582 ( and therefore @xmath564 ) , we calculate @xmath590 in words , @xmath591 is the probability that a point on a path will be in state @xmath449 , conditioned on having originated at the initial vertex in the distant past , and conditioned on having a distant future
. for each @xmath120 let @xmath592 denote the set of paths in @xmath166 of length @xmath120 starting at the initial vertex , and let @xmath593 denote the set of _ all _ paths in @xmath166 of length @xmath120 .
we can naturally identify @xmath594 with the vertices of @xmath166 .
restricting to an initial subpath defines an inverse system @xmath595 , and the inverse limit @xmath596 is the space of ( right ) infinite paths .
similarly define @xmath597 .
if we give each @xmath593 and @xmath592 the discrete topology , then @xmath596 and @xmath598 are cantor sets .
if @xmath599 and @xmath600 are two elements of @xmath596 , we define @xmath601 to be the first index at which @xmath25 and @xmath602 differ , and define a metric on @xmath596 by setting @xmath603 for some @xmath298 ( the notation @xmath604 is deliberately intended to suggest a resemblance to the gromov product ) .
if we like , we can define @xmath605 and metrize it ( as a compact space , in which each @xmath593 sits as a discrete subset ) in the same way .
similarly , give @xmath598 the induced metric , and define @xmath606 likewise .
the shift operator @xmath607 is defined by @xmath608 .
we define a probability measure @xmath564 on each @xmath593 by @xmath609 where @xmath564 and @xmath201 are the measure and stochastic matrix whose properties are given in lemma [ n_preserves_mu ] . by the definition of an inverse limit , there is a map @xmath610 for each @xmath120 which takes an infinite path to its initial subpath of length @xmath120 ; the preimages of subsets of the @xmath593 under such maps are a basis for the topology on @xmath596 , called _
cylinder sets_. the measures @xmath564 as above let us define a borel probability measure @xmath564 on @xmath596 by first defining it on cylinder sets ( note that the definitions of @xmath564 on different @xmath593 are compatible ) and extending it to all borel sets in the standard way ; lemma [ n_preserves_mu ] implies that @xmath564 is @xmath60-invariant ( i.e. @xmath611 for all measurable @xmath612 ) .
there is a bijection between @xmath592 and @xmath438 , and by evaluation with @xmath613 .
this map extends continuously to a map @xmath614 , by sending @xmath615 .
the map @xmath614 is surjective , lipschitz in the @xmath299-metric , and bounded - to - one .
that the map is lipschitz follows immediately from the definition , and the observation that @xmath616 for @xmath617 .
the restrictions @xmath618 are all bijections , so we just need to check that @xmath615 is surjective and bounded - to - one .
since @xmath619 is continuous , @xmath620 is compact and @xmath336 is hausdorff , the image is compact . since the image is dense ( because it contains @xmath613 for all @xmath120 ) , it is surjective .
finally , observe that if @xmath26 and @xmath621 are any two points in @xmath598 , and @xmath243 are the associated infinite geodesics in @xmath0 , then @xmath622 is a compact initial segment , since after they diverge they never meet again ( by the definition of a combing ) .
fix @xmath623 , let @xmath624 be a finite subset of @xmath625 , and let @xmath87 be the geodesic rays in @xmath0 corresponding to the @xmath624 .
for all but finitely many points @xmath123 on any @xmath87 , each @xmath626 intersects the ball @xmath627 _ disjointly _ from the others . in particular , the number of points in the preimage of any point in @xmath265 is bounded by the cardinality of a ball ( in @xmath0 ) of radius @xmath104 .
recall that in [ patterson_sullivan_subsection ] we defined probability measures @xmath335 on @xmath0 for each @xmath326 .
note that @xmath628 where @xmath307 is as above .
for each @xmath120 we define a probability measure on @xmath592 ( which , by abuse of notation , we call @xmath335 ) by @xmath629 for @xmath630 , and observe that the limit as @xmath631 from above ( which we denote @xmath632 ) exists and depends only on the cone type @xmath633 . since the patterson sullivan measure @xmath339 is supported in @xmath265 , the measures @xmath339 on each @xmath592 are compatible , thinking of each @xmath592 as a collection of cylinder sets in @xmath598 , and define a unique probability measure @xmath339 on @xmath598 which pushes forward under @xmath619 to @xmath339 on @xmath265 .
[ nu_to_mu ] the measure @xmath564 on @xmath596 is the limit @xmath634 .
we give the sketch of a proof . for any @xmath26 ,
let @xmath635 be the ( regular ) language of suffixes of words in @xmath21 with @xmath26 as a prefix , and let @xmath636 be the subset of @xmath635 of length @xmath120
. then @xmath637 .
if there is no path from the final state @xmath638 to a maximal component , the growth rate of @xmath635 is strictly less than that of @xmath21 , and @xmath639 .
otherwise both growth functions are eventually of the form @xmath640 plus something exponentially small compared to @xmath641 .
define measures @xmath642 on @xmath592 by @xmath643 .
then by considering the form of the growth functions of @xmath21 and @xmath635 , we see that there is a constant @xmath139 ( not depending on @xmath26 or @xmath120 ) so that @xmath644 . scaling @xmath642 to be a probability measure
, we can set @xmath645 .
the proof now follows from the definition of @xmath646 ; see @xcite lem . 4.19 for details .
let @xmath647 be a ( stationary ) irreducible markov chain on a finite state space , with stationary measure @xmath564 , and let @xmath648 be a real - valued function on the state space ( since this space is finite , there are no additional assumptions on @xmath648 ; in general we require @xmath648 to be integrable , and have finite variance ) .
define @xmath649 , and @xmath650 .
[ markov_clt ] with notation as above , there is some @xmath651 so that for any @xmath652 , @xmath653 equivalently , there is convergence in probability @xmath654 where @xmath655 denote the normal distribution with mean @xmath24 and standard deviation @xmath198 ( in case @xmath656 , we let @xmath655 denote a dirac mass centered at @xmath24 ) . now , each maximal component @xmath139 as above is a stationary irreducible markov chain , with stationary measure the conditional measure @xmath657 .
the measure @xmath564 on @xmath596 decomposes measurably into the union of ( shift - invariant ) subspaces @xmath658 , the subspace of ( right ) infinite sequences contained in the component @xmath139 . consequently ,
if @xmath542 is a combable function on @xmath0 , then for each maximal component @xmath139 , there are constants @xmath659 and @xmath660 , so that for @xmath564-a.e .
@xmath661 , the random variable @xmath662 converges in probability to @xmath663 . by lemma [ nu_to_mu ] , for @xmath339-a.e .
@xmath664 there is a unique @xmath139 so that @xmath665 for sufficiently big @xmath120 ; we say that @xmath26 is _ associated _ to the component @xmath139 .
let @xmath666 be the set of @xmath26 associated to a fixed @xmath139 .
for @xmath339-a.e .
@xmath667 we have convergence in probability @xmath668 ( one way to see this is to observe that this is a shift - invariant tail property of @xmath26 , and use lemma [ nu_to_mu ] ) . for combable functions ,
this is the end of the story .
it is certainly possible for the constants @xmath669 to vary from component to component .
but for bicombable @xmath542 we have the following key lemma : [ common_variance ] let @xmath542 be bicombable
. then there are constants @xmath670 so that @xmath671 and @xmath672 for all maximal components @xmath139 . call @xmath664 _ typical _ if there are constants @xmath673 and @xmath674 ( necessarily unique ) so that @xmath675 . for each @xmath139
we have seen that @xmath339-a.e .
@xmath667 is typical with @xmath676 and @xmath677 .
the map @xmath678 is finite - to - one , and takes the measure @xmath339 on @xmath598 to the patterson
sullivan measure @xmath339 on @xmath265 .
hence @xmath679 has positive measure for each @xmath139 .
let @xmath369 be typical , and let @xmath680 be the associated geodesic sequence of elements in @xmath0 converging to @xmath681 .
now let @xmath3 be arbitrary , let @xmath621 be any element of @xmath598 with @xmath682 , and let @xmath683 be the geodesic sequence of elements in @xmath0 associated to @xmath621 .
by @xmath104-thinness , @xmath684 is eventually approximately constant , and therefore bounded . since @xmath542 is bicombable , @xmath621 is typical , with @xmath685 and @xmath686 .
but the action of @xmath0 on @xmath339 is ergodic for @xmath339 , by corollary [ ergodic_at_infinity ] , and therefore for any @xmath687 there are typical @xmath667 , @xmath688 with @xmath689 and @xmath690 , and with @xmath691 for some @xmath3 .
this completes the proof .
let @xmath0 be hyperbolic , and let @xmath542 be bicombable .
then there are constants @xmath670 so that if @xmath692 denotes a random element of @xmath613 ( in the @xmath339 measure ) , there is convergence in probability @xmath693 .
note that @xmath404 and @xmath198 as above are _ algebraic _ , and one can estimate from above the degree of the field extension in which they lie from the complexity of @xmath166 . the uniform measure and the measure @xmath339 on @xmath613 are uniformly quasi - equivalent on a large scale , in the sense that there are constants @xmath12 and @xmath139 so that for any @xmath694 , there is an inequality @xmath695 it follows that if @xmath692 denotes a random element of @xmath613 ( in the uniform measure ) , the distribution @xmath696 has a tail that decays like @xmath697 .
since length with respect to one generating set is bicombable with respect to another , we obtain the following corollary : let @xmath0 be hyperbolic , and let @xmath1 and @xmath545 be two finite generating sets for @xmath0 .
there is an algebraic number @xmath698 so that if @xmath692 is a random element of @xmath1 of word length @xmath120 , then the distribution @xmath699 has a tail that decays like @xmath697 when @xmath120 is sufficiently large .
it is a slightly subtle point that @xmath700 , and the inequality is strict except for essentially trivial cases . to push these techniques
further , we must study classes of functions more general than combable functions , and invoke more sophisticated limit theorems .
there is a well - known framework to carry out such analysis , pioneered by ruelle , sinai , bowen , ratner , parry etc . ; @xcite is a standard reference .
the setup is as follows . for simplicity ,
let @xmath227 be a @xmath701 matrix with @xmath24@xmath22 entries for which there is a constant @xmath120 so that all the entries of @xmath702 are positive ( i.e. @xmath227 is the adjacency matrix of a topological markov chain with @xmath174 states which is irreducible and aperiodic ) .
let @xmath596 be the space of ( right ) infinite sequences @xmath599 satisfying @xmath703 for all @xmath120 , and let @xmath60 be the shift operator on @xmath596 . as before , we can metrize @xmath596 by @xmath603 for some fixed @xmath298 , and observe that the action of @xmath60 on @xmath596 is _
mixing_. this means that for all nonempty open sets @xmath704 there is @xmath201 so that @xmath705 is nonempty for all @xmath706 .
note that if @xmath227 is irreducible but not aperiodic , there is nevertheless a decomposition of @xmath596 into @xmath351 disjoint components which are cycled by @xmath60 , and such that @xmath707 is mixing on each component , where @xmath351 is the gcd of the periods of @xmath60-invariant sequences .
let @xmath708 be the space of @xmath60-invariant probability measures on @xmath596 .
this is a convex , compact subset of the space of all measures in the weak-@xmath709 topology .
it is not hard to show that the topological entropy @xmath194 of @xmath60 is equal to the supremum of the measure theoretic entropies @xmath710 , and that @xmath711 where @xmath307 is the perron frobenius eigenvalue of @xmath227 ; see e.g. @xcite .
the shift @xmath60 uniformly expands @xmath596 by a factor of @xmath299 , and therefore if a function on @xmath596 is sufficiently regular , it tends to be smoothed out by @xmath60 .
define @xmath712 by @xmath713 .
we would like the iterates @xmath714 to have a uniform modulus of continuity ; this is achieved precisely by insisting that @xmath648 be hlder continuous , that is , that there is some @xmath715 so that @xmath716 .
the set of functions @xmath648 on @xmath596 , hlder continuous of exponent @xmath715 , is a banach space with respect to the norm @xmath717 where @xmath718 is the least such @xmath139 so that @xmath719 .
we denote this banach space @xmath720 .
let @xmath648 be hlder continuous on @xmath596 .
the _ pressure _ of @xmath648 , denoted @xmath721 , is @xmath722 .
it turns out that the supremum is realized on some invariant measure @xmath723 of full support , known as the _ equilibrium state _ ( or _ gibbs state _ ) of @xmath648 .
that is , @xmath724 .
see e.g. @xcite ch . 1 for a proof of this theorem , and of theorem [ rpf_theorem ] below .
the _ ruelle transfer operator _
@xmath725 associated to @xmath648 is defined by the formula @xmath726 .
note that @xmath725 acts as a bounded linear operator on @xmath720 .
[ rpf_theorem ] the operator @xmath725 has a simple positive eigenvalue @xmath727 which is strictly maximal in modulus . the essential spectrum is contained in a ball whose radius is strictly less than @xmath727 , and the rest of the spectrum outside this ball is discrete and consists of genuine eigenvalues .
there is a strictly positive eigenfunction @xmath728 satisfying @xmath729 , and an `` eigen probability measure '' @xmath730 satisfying @xmath731 , and if we scale @xmath728 so that @xmath732 , then the equilibrium state @xmath723 is equal to @xmath733 .
@xmath730 can be thought of as a left eigenvector for @xmath725 , and @xmath728 as a right eigenvector .
when @xmath648 is identically zero , @xmath725 is basically just the matrix @xmath227 , and @xmath723 is basically just @xmath564 as constructed in [ markov_chain_subsection ] .
pollicott @xcite proved a complexified version of the rpf theorem , and showed that @xmath721 and @xmath728 are _ analytic _ on an open subset of the _ complex _ banach space @xmath734 which contains a neighborhood of @xmath735 ( i.e. of @xmath720 ) .
because of the simplicity and analyticity of the maximal eigenvector / value , one can study the _ derivatives _ of pressure . for simplicity ,
let @xmath736 .
then we can compute @xmath737 and a further differentiation gives @xmath738 where @xmath739 ( suitably normalized ) .
now , let @xmath740 .
then from the definition of the transfer operator , @xmath741 , and therefore one obtains @xmath742 if we set @xmath3 to be identically zero , then @xmath743 is just the equilibrium measure @xmath564 from before .
if we change @xmath648 by a constant @xmath744 to have mean @xmath24 , then the ergodic theorem shows @xmath745 @xmath564-a.e . and therefore @xmath746 it is usual to denote this limiting quantity by @xmath747 .
the analyticity of @xmath748 lets us control the higher moments of @xmath749 in a uniform manner , and therefore by applying fourier transform , one obtains a central limit theorem @xmath750 .
better estimates of the rate of convergence can be obtained by studying @xmath751 ; see @xcite .
this theorem can be combined with lemma [ common_variance ] to obtain a central limit theorem for certain functions on hyperbolic groups whose ( discrete ) derivatives along a combing satisfy a suitable hlder continuity property .
such functions arise naturally for groups acting cocompactly on @xmath106 spaces with @xmath105 , where one wants to compare the intrinsic geometry of the space with the `` coarse '' geometry of the group .
let @xmath752 be a complete @xmath106 geodesic metric space with @xmath105 , and let @xmath0 act cocompactly on @xmath752 by isometries .
pick a basepoint @xmath753 , and define a function @xmath754 on @xmath0 by @xmath755 . since @xmath0 is hyperbolic , if we fix a finite generating set @xmath1 we can choose a geodesic combing @xmath21 with respect to @xmath1 as above .
now , for any @xmath6 define @xmath756 .
it is straightforward to see from the @xmath106 property that there are constants @xmath139 and @xmath715 ( depending on @xmath45 and @xmath0 ) so that @xmath757 for all @xmath285 and all @xmath758 .
an element of @xmath759 corresponds to a path in @xmath166 .
reading the edge labels determines a word in the generators ( a suffix of some word in @xmath21 ) , and by evaluation , an element of @xmath0 .
let @xmath760 denote this evaluation map ( note that this is not injective ) .
we can define a function @xmath761 on @xmath759 by @xmath762 where @xmath763 is the label associated to the transition from @xmath764 to @xmath765 ( we could suggestively write @xmath766 ) .
evidently , @xmath761 extends to a hlder continuous function on @xmath253 .
furthermore , for each @xmath630 , we have @xmath767 . for each big component @xmath139 , it follows that @xmath339-a.e .
@xmath667 are @xmath669 typical ( for the function @xmath761 ) for some @xmath669 depending only on @xmath139 . since @xmath754 is lipschitz on @xmath0 in the left and right invariant metrics , the argument of lemma [ common_variance ] implies that @xmath669 are equal to some common values @xmath670 , and therefore we obtain the following corollary : [ distance_clt ] let @xmath752 be a complete @xmath106 geodesic metric space with @xmath105 , and let @xmath0 act cocompactly on @xmath752 by isometries . pick a basepoint @xmath753 , and a finite generating set @xmath1 for @xmath0
. then there are constants @xmath404 and @xmath198 so that if @xmath692 is a random element of @xmath613 ( in the @xmath339 measure ) , there is convergence in probability @xmath768 . evidently , the only properties of the function @xmath754 we use are that it is lipschitz in both the left- and right - invariant metrics , and satisfies a hlder estimate @xmath757 for all @xmath285 and all @xmath758 .
any such function on a hyperbolic group satisfies a central limit theorem analogous to corollary [ distance_clt ] . for the sake of completeness , therefore , we state this as a theorem : [ holder_clt ] let @xmath0 be a hyperbolic group , and @xmath1 a finite generating set for @xmath0 .
let @xmath754 be a real - valued function which is lipschitz in both the left- and right - invariant word metrics on @xmath0 , and satisfies @xmath757 for all @xmath285 in @xmath1 and all @xmath758
. then there are constants @xmath404 and @xmath198 so that if @xmath692 is a random element of @xmath613 ( in the @xmath339 measure ) , there is convergence in probability @xmath769 .
the idea of using the thermodynamic formalism to study the relationship between distance and word length in cocompact groups of isometries of hyperbolic space is due to pollicott sharp @xcite ; corollary [ distance_clt ] and theorem [ holder_clt ] above are simply the result of combining their work with @xcite and @xcite .
nevertheless , we believe they are new .
the main references for this section are kaimanovich @xcite and kaimanovich vershik @xcite .
the theory of random walks is a vast and deep subject , with connections to many different parts of mathematics .
therefore it is necessary at a few points to appeal to some standard ( but deep ) results in probability theory , whose proof lies outside the scope of this survey .
a basic reference for probability theory is @xcite .
we give more specialized references in the text where relevant .
this section is brief compared to the earlier sections , and is not meant to be comprehensive .
let @xmath0 be a group and let @xmath564 be a probability measure on @xmath0 .
we further assume that @xmath564 is _ nondegenerate _ ; i.e. that the support of @xmath564 generates @xmath0 as a semigroup .
an important example is the case where @xmath564 is the uniform measure on a symmetric finite generating set @xmath1 .
there are two ways to describe random walk on @xmath0 determined by @xmath564 : as a sequence of elements visited in the walk , or as a sequence of increments . in the first description ,
a random walk @xmath770 is a markov chain with state space @xmath0 , with initial state @xmath313 , and with transition probability @xmath771 . in the second description ,
a random walk @xmath772 is a sequence of random elements of @xmath0 ( the increments of the walk ) , independently distributed according to @xmath564 .
the two descriptions are related by taking @xmath773 .
we write this suggestively as @xmath774 and @xmath775 .
we use the notation @xmath776 for the product probability space , and @xmath777 for the probability space of infinite sequences with the measure @xmath778 on cylinder sets defined by @xmath779 with this notation , @xmath301 is a random element of @xmath776 and @xmath26 is a random element of @xmath777 .
the shift operator @xmath60 acts on @xmath780 by @xmath781 or @xmath782 .
it is measure preserving for @xmath783 but not for @xmath778 ; in fact , from the definition , the support of @xmath778 is contained in the set of sequences starting at @xmath313 .
the action of the shift @xmath60 on @xmath776 is ergodic . for
, if @xmath404 is a subset satisfying @xmath784 , then a sequence @xmath301 is in @xmath404 if and only if @xmath785 is in @xmath404 for sufficiently big @xmath120 .
this is a tail event for the sequence of independent random variables @xmath786 , so by kolmogorov s @xmath24@xmath22 law ( see @xcite thm .
1.1.2 ) @xmath404 has measure @xmath24 or @xmath22 .
let @xmath0 be a group and @xmath1 a finite generating set .
let @xmath564 be a probability measure on @xmath0 .
first moment _ of @xmath564 is @xmath787 ; if this is finite , we say @xmath564 has _
finite first moment_. [ drift_exists ] let @xmath564 be a probability measure on @xmath0 with finite first moment .
let @xmath788 be a random walk determined by @xmath564 .
then @xmath789 exists almost surely , and is independent of @xmath26 .
in fact , if @xmath790 denotes @xmath120-fold convolution ( i.e. the distribution of the random variable @xmath638 ) , then @xmath791 .
we set @xmath774 .
define @xmath792 .
then @xmath793 satisfies @xmath794 i.e. @xmath793 form a _
subadditive cocycle_. kingman s subadditive ergodic theorem ( see e.g. @xcite ) says that for any subadditive @xmath795 cocycle @xmath793 on a space with a @xmath60-invariant measure , the limit @xmath796 exists a.s . and is @xmath60-invariant . in our circumstance , finite first moment implies that @xmath797 ( and all the @xmath793 ) are in @xmath795 , so the theorem applies . since the action of @xmath60 on @xmath776 is ergodic , the limit is independent of @xmath301 .
the lemma follows . @xmath21 as above is called the _ drift _ of the random walk associated to @xmath564 .
since each @xmath798 we necessarily have @xmath799 . if @xmath800 and @xmath564 is symmetric ( i.e. @xmath801 for all @xmath3 ) with finite support , then @xmath802 .
we now focus our attention on the case of hyperbolic groups and simple random walk ( i.e. when @xmath564 is the uniform measure on a finite symmetric generating set ) .
[ positive_drift ] let @xmath0 be a nonelementary hyperbolic group , and let @xmath564 be a nondegenerate probability measure on @xmath0 with finite first moment
. then the drift @xmath21 of random walk with respect to @xmath564 is positive .
we give the idea of a proof .
let @xmath790 denote the @xmath120-fold convolution of @xmath564 as before .
the probability measures @xmath790 have a subsequence converging to a weak limit @xmath803 in @xmath336 .
clearly the support of @xmath803 is contained in @xmath265 ( a group for which @xmath804 for any @xmath3 and for @xmath564 nondegenerate is a finite group ) . to prove the lemma it suffices to show that for any @xmath139 , for sufficiently large enough @xmath120 there is an inequality @xmath805 for all @xmath806 , since then @xmath807 .
now , for each @xmath194 , if @xmath3 satisfies @xmath808 , then the closest point on the geodesic from @xmath809 to @xmath3 is within @xmath104 of some geodesic from @xmath313 to @xmath809 .
so as @xmath32 goes to infinity , the @xmath299-distance from @xmath809 to @xmath3 goes to @xmath24 .
hence for this inequality to fail to hold , almost half of the mass of @xmath810 must be concentrated near the _ antidiagonal _ ; i.e. the set of @xmath811 .
from this we can deduce that either the desired inequality is satisfied , or else most of the mass of @xmath790 must be concentrated near a single geodesic through @xmath313 .
taking @xmath812 , the support of @xmath803 must consist of exactly two points , and @xmath0 is seen to be elementary , contrary to hypothesis .
it is a theorem of guivarch ( see @xcite , thm .
8.14 ) that if @xmath0 is any group with a nondegenerate measure @xmath564 ( always with finite first moment ) for which the drift of random walk is zero , then @xmath0 is amenable . some care is required to parse this statement : on an amenable group some nondegenerate measures _ may _ have positive drift , but on a nonamenable group , _ every _ nondegenerate measure has positive drift .
a nonelementary hyperbolic group always contains many nonabelian free groups , and is therefore nonamenable ; this gives a more highbrow proof of lemma [ positive_drift ] .
[ regular_sequence ] let @xmath14 be a @xmath104-hyperbolic space .
the following two conditions are equivalent for a sequence @xmath813 in @xmath14 and a number @xmath814 : 1 .
@xmath815 and @xmath816 ; 2 .
there is a geodesic ray @xmath27 so that @xmath817 .
a sequence @xmath813 satisfying either condition is said to be _
regular_. that ( 2 ) implies ( 1 ) is obvious , so we show that ( 1 ) implies ( 2 ) . for simplicity , we use the notation @xmath818 .
the path obtained by concatenating geodesics from @xmath813 to @xmath819 has finite @xmath299-length , and therefore converges to some unique @xmath820 .
let @xmath821 ( resp .
@xmath822 ) be geodesic rays from the origin to @xmath813 ( resp .
@xmath823 ) and parameterize them by distance from the origin .
fix some positive @xmath46 , and let @xmath824 be such that for any two @xmath825 the geodesics @xmath821 and @xmath826 are within @xmath104 on the interval of length @xmath827 , and let @xmath828 so that @xmath829 for @xmath830 .
now , @xmath831 and therefore @xmath832 . on the other hand , @xmath833 .
consequently the sequence @xmath146 is a quasigeodesic , and therefore there is a constant @xmath834 so that @xmath835 for any @xmath706 .
since @xmath836 and @xmath822 are asymptotic , @xmath837 for sufficiently large @xmath120 , and therefore @xmath838 for sufficiently large @xmath120 . taking @xmath419 proves the lemma with @xmath839 .
together with lemma [ positive_drift ] this gives the following corollary : [ random_walk_near_geodesic ] let @xmath0 be a nonelementary hyperbolic group , and let @xmath564 be a nondegenerate probability measure on @xmath0 with finite first moment .
then there is @xmath814 so that for a.e .
random walk @xmath26 there is a unique geodesic ray @xmath840 with @xmath841 .
it suffices to show that if @xmath564 has finite first moment , then @xmath842 almost surely .
let @xmath774 , and for any @xmath843 let @xmath844 be the event that @xmath845 .
then the probability of @xmath844 is @xmath846 , and therefore @xmath847 therefore by the easy direction of the borel
cantelli lemma ( see e.g. @xcite 1.1.4 ) the probability that @xmath844 occurs infinitely often is zero . since this is true for every @xmath46 , we have @xmath848 almost surely .
define an equivalence relation @xmath849 on @xmath780 by @xmath850 if and only if there are integers @xmath851 so that @xmath852 .
the _ measurable envelope _ of @xmath849 is the smallest measurable equivalence relation generated by @xmath849 .
the quotient measure space @xmath853 of @xmath777 by the measurable envelope is called the _ poisson boundary _ of @xmath0 with respect to @xmath564 .
in other words , @xmath339-measurable functions on @xmath166 correspond precisely to @xmath60-invariant @xmath778-measurable functions on @xmath780 .
we let @xmath854 be the quotient map , so that @xmath855 . now ,
@xmath0 acts on @xmath780 on the left coordinatewise .
this action commutes with @xmath60 , and descends to an action on @xmath166 . since @xmath849 is @xmath60-invariant , @xmath856 ,
so @xmath857 ; i.e. the measure @xmath339 is _
@xmath564-stationary_. a _ @xmath564-boundary _ is a @xmath0-space with a @xmath564-stationary measure @xmath307 which is obtained as a @xmath60-equivariant ( measurable ) quotient of @xmath777 . any @xmath564-boundary factors through @xmath853 .
a @xmath564-boundary is _ @xmath564-maximal _ if the map from @xmath853 is a measurable isomorphism .
kaimanovich @xcite gave two very useful criteria for a @xmath564-boundary to be maximal .
let @xmath405 be a @xmath564-boundary , and for @xmath858 let @xmath859 be the image of @xmath26 under the ( @xmath0-equivariant ) quotient map @xmath860 . if there is a family of measurable maps @xmath861 such that @xmath778-a.e .
@xmath862 then @xmath405 is maximal .
together with corollary [ random_walk_near_geodesic ] , this gives the following important result : let @xmath0 be a nonelementary hyperbolic group , and let @xmath564 be a nondegenerate probability measure of finite first moment .
let @xmath863 take a random walk to its endpoint ( which exists @xmath778-a.e . ) , and let @xmath864 .
then @xmath865 is the poisson boundary of @xmath866 .
simply define @xmath867 to be the maps that take a point @xmath389 to @xmath868 where @xmath840 is a parameterized geodesic ray from @xmath313 to @xmath26 , and @xmath21 is the drift .
if @xmath648 is a function on @xmath0 , the operator @xmath869 ( convolution with @xmath564 ) is defined by @xmath870 .
a function @xmath648 on @xmath0 is _ @xmath564-harmonic _ ( or just harmonic if @xmath564 is understood ) if it is fixed by @xmath869 ; i.e. if it satisfies @xmath871 for all @xmath3 in @xmath0 .
in general we need to impose some condition on @xmath648 for @xmath872 to be defined .
if the support of @xmath564 is finite , then @xmath648 can be arbitrary , but if the support of @xmath564 is infinite , we usually ( but not always ! ) require @xmath648 to be in @xmath873 .
we let @xmath874 denote the banach space of bounded @xmath564-harmonic functions on @xmath0 . in probabilistic terms ,
if @xmath648 is harmonic and @xmath858 is a random walk , the random variables @xmath875 are a _ martingale _ ; i.e. the expected value of @xmath876 given @xmath877 is @xmath878 ( see e.g. @xcite 5.2 for an introduction to martingales ) .
there is an intimate relation between harmonic functions and poisson boundaries , expressed in the following proposition .
the banach spaces @xmath874 and @xmath879 are isometric .
given @xmath880 and @xmath858 , the random variables @xmath881 are a bounded martingale , and therefore by the martingale convergence theorem ( @xcite thm .
5.2.22 ) , converge a.s . to a well - defined limit .
evidently this limit is measurable and @xmath60-invariant , and therefore descends to a function on @xmath166 which we denote @xmath882 .
explicitly , @xmath883 .
conversely , given @xmath884 we define @xmath885 ( this expression is known as the _ poisson formula _ ) .
since @xmath339 is stationary , @xmath648 is harmonic .
the mean value property of harmonic functions implies that these maps are isometries , since a harmonic function achieves its maximum on the boundary .
note that the poisson formula is available for _ any _ @xmath564-boundary .
that is , if @xmath405 is a @xmath0-space with a @xmath564-stationary probability measure @xmath307 , and @xmath882 is any element of @xmath886 , then @xmath887 is a bounded harmonic function on @xmath0 .
if @xmath307 is not invariant , @xmath648 is typically nonconstant . the remainder of this section is devoted to some miscellaneous applications of random walks to hyperbolic and other groups .
there is a close resemblance between the measure @xmath339 and the patterson
sullivan measures constructed in [ patterson_sullivan_subsection ] .
this resemblance can be sharpened if one looks at a natural metric on @xmath0 adapted to the random walk , namely the so - called _ green metric_. let @xmath0 be a group and @xmath564 a probability measure on @xmath0 with finite first moment .
the _ green metric _ on @xmath0 is the metric for which the distance between @xmath3 and @xmath194 is @xmath888 of the probability that random walk starting at @xmath3 ever hits @xmath194 .
if @xmath564 is symmetric , so is the green metric , since random walks are time - reversible . note that the green metric is degenerate if random walk is recurrent . for simple random
walk , this occurs only if @xmath0 is finite , or is virtually @xmath190 or @xmath889 , by a classical result of varopoulos ( see @xcite ) . for nondegenerate measures with finite first moment on non - elementary hyperbolic groups , blachre and brofferio @xcite
show that the green metric and the word metric are quasi - isometric ( one needs to be somewhat careful : the green metric is _ not _ in general a geodesic metric ) .
[ harmonic_measure_dimension ] let @xmath0 be a non - elementary hyperbolic group , and for @xmath369 , let @xmath890 denote the ball of radius @xmath12 in the @xmath299-metric ( see definition [ a_distance ] ) .
let @xmath564 be a symmetric probability measure with finite first moment , and let @xmath339 be the associated harmonic measure on @xmath265 .
then for @xmath339-almost every @xmath369 , there is convergence @xmath891 where @xmath21 is the drift in the word metric , and @xmath892 is the drift in the green metric .
note that kingman s subadditive ergodic theorem implies that the drift @xmath892 with respect to the green metric is well - defined , essentially by the same argument as the proof of lemma [ drift_exists ] .
the classical _ harnack inequality _ relates the values of a positive harmonic function at two points . in its infinitesimal version
, it asserts an upper bound on the logarithmic derivative of a positive harmonic function .
let @xmath648 be a non - negative bounded harmonic function on @xmath893 , for simplicity .
the poisson formula says that @xmath894 where @xmath895 is the visual measure as seen from @xmath123 .
if @xmath339 is visual measure as seen from the origin , and @xmath3 is any isometry taking the origin to @xmath123 , then @xmath896 . to understand how @xmath648 varies as a function of @xmath123 therefore , it suffices to understand how @xmath895 varies as a function of @xmath123 .
if @xmath405 is an infinitesimal ball centered at some point @xmath26 in @xmath897 , then the visual size of @xmath405 grows like @xmath898 as one moves distance @xmath446 in the direction of @xmath26 .
hence : let @xmath648 be a non - negative bounded harmonic function on @xmath893
. then the logarithmic derivative of @xmath648 satisfies the inequality @xmath899 .
if @xmath648 is a non - negative harmonic function on a group @xmath0 , the analog of this inequality is @xmath900 for any @xmath5 and @xmath6 where @xmath351 is the dimension of @xmath339 , which can be determined from theorem [ harmonic_measure_dimension ] .
if @xmath1 is a closed surface of genus @xmath901 and @xmath902 is injective , then @xmath903 acts on @xmath0 by left translation , and there is an associated foliated bundle with fiber the ideal circle @xmath904 with its natural @xmath903 action .
we can build a harmonic connection for this circle bundle ; i.e. a choice of measure @xmath905 on the circle @xmath906 over each @xmath5 so that for any subset @xmath907 we have @xmath908 . since the circle is @xmath22-dimensional , these measures integrate to _ metrics _ on the circles @xmath906 for which the curvature is harmonic .
the harnack inequality then gives _ a priori _ bounds on this curvature , and one can deduce local compactness results for families of injective surface maps of variable genus . for stable minimal surfaces in hyperbolic @xmath97-manifolds , such _ a priori _ bounds were obtained by schoen @xcite and are an important tool in low - dimensional topology .
the idea of using harnack - type inequalities to obtain curvature bounds is due to thurston @xcite ( also see @xcite , example 4.6 ) .
a _ norm _ on a group is a non - negative function @xmath909 so that @xmath910 for all @xmath758 .
a functor from groups to norms is _ monotone _ if @xmath911 for any @xmath5 and @xmath912 .
if @xmath913 is a norm on @xmath0 , and @xmath564 is a probability measure with finite first moment , it makes sense to study the growth rate of @xmath913 under @xmath564-random walk on @xmath0 .
if @xmath0 is finitely generated , one can study the growth rate of @xmath913 under all simple random walks ; if they all have the same growth rate , this rate is an invariant of @xmath0 .
since @xmath564 random walk on @xmath0 pushes forward to @xmath914 random walk on @xmath915 , the growth rate of a monotone family of norms can not increase under a homomorphism ; thus if the growth rate of @xmath916 on @xmath0 is strictly smaller than the growth rate of @xmath917 on @xmath92 , there are strong constraints on the homomorphisms from @xmath0 to @xmath92 . as an example , consider the _ commutator length _ @xmath918 . for any group @xmath0 and any @xmath3 in the commutator subgroup @xmath919 $ ] , the commutator length @xmath920 is just the least number of commutators in @xmath0 whose product is @xmath3 ( for technical reasons , one usually studies a closely related quantity , namely the _ stable commutator length _ ; see e.g @xcite for an introduction ) .
one of the main theorems of @xcite is as follows : let @xmath0 be hyperbolic , and let @xmath564 be a nondegenerate symmetric probability measure with finite first moment whose support generates a nonelementary subgroup .
there is a constant @xmath139 so that if @xmath692 is obtained by random walk of length @xmath120 , conditioned to lie in @xmath919 $ ] , then @xmath921 with probability @xmath922 .
said another way , commutator length grows like @xmath923 under random walk in a hyperbolic group .
similar estimates on commutator length can be obtained for groups acting in a suitable way on ( not necessarily proper ) hyperbolic spaces ; the most important examples are mapping class groups and relatively hyperbolic groups . as a corollary ,
if @xmath92 is any finitely generated group , and commutator length in @xmath92 grows like @xmath924 for simple random walk ( with respect to some generating set ) , then there are no interesting homomorphisms from @xmath92 to any hyperbolic group @xmath0 , and no interesting actions of @xmath92 on certain hyperbolic complexes .
i would like to thank vadim kaimanovich , anders karlsson , joseph maher , curt mcmullen , richard sharp , and alden walker .
i would also like to thank the anonymous referee for some useful comments .
danny calegari was supported by nsf grant dms 1005246 .
99 s. blachre and s. brofferio , _ internal diffusion limited aggregation on discrete groups having exponential growth _ , probab .
theory relat .
fields * 137 * ( 2007 ) , 323-343 s. blachre , p. hassinsky and p. mathieu , _ harmonic measures versus quasiconformal measures for hyperbolic groups _ ,
preprint , arxiv:0806.3915 b. bowditch , _ a topological characterization of hyperbolic groups _ ,
ams * 11 * ( 1998 ) , no .
3 , 643667 r. bowen , _ equilibrium states and the ergodic theory of anosov diffeomorphisms _
, lecture notes in mathematics * 470 * , springer - verlag 1975 r. brooks , _ some remarks on bounded cohomology _ , riemann surfaces and related topics : proceedings of the 1978 stony brook conference ( suny stony brook ny 1978 ) ann . math . studies * 97 * , princeton university press , 1981 , 5363 d. calegari , _ foliations and the geometry of @xmath97-manifolds _ , oxford mathematical monographs .
oxford university press , oxford , 2007 .
d. calegari , _ scl _ , msj memoirs , * 20*. mathematical society of japan , tokyo , 2009 .
d. calegari and k. fujiwara , _ combable functions , quasimorphisms and the central limit theorem _
, ergodic theory dynam .
systems * 30 * ( 2010 ) , no . 5 , 13431369 d. calegari and j. maher , _ statistics and compression of scl _ , preprint , arxiv:1008.4952 j. cannon , _ the combinatorial structure of cocompact discrete hyperbolic groups _ , geom . ded . *
16 * ( 1984 ) , no .
2 , 123148 z. coelho and w. parry , _ central limit asymptotics for shifts of finite type _ , israel j. math .
* 69 * ( 1990 ) , no .
2 , 235249 m. coornaert , _ mesures de patterson sullivan sure le bord dun espace hyperbolique au sens de gromov _ , pac
. j. math .
* 159 * ( 1993 ) , no .
2 , 241270 m. coornaert , t. delzant and a. papadopoulos , _ gomtrie et thorie des groupes _ , les groupes hyperboliques de gromov , springer - verlag , berlin , 1990 . m. coornaert and a. papadopoulos , _ symbolic dynamics and hyperbolic groups _ , lecture notes in mathematics * 1539 * , springer - verlag 1993 d. epstein , j. cannon , d. holt , s. levy , m. paterson and w. thurston , _ word processing in groups _ ,
jones and bartlett , boston , ma , 1992 d. epstein and k. fujiwara , _ the second bounded cohomology of word - hyperbolic groups _
, topology * 36 * ( 1997 ) , 12751289 h. federer , _ geometric measure theory _ ,
. wiss . * 153 * springer - verlag new york , 1969 p. flajolet and r. sedgewick , _ analytic combinatorics _ , cambridge university press , cambridge , 2009 a. furman , _ random walks on groups and random transformations _ ,
handbook of dynamical systems , vol .
1a , 9311014 , north - holland , amsterdam , 2002 m. gromov , _ hyperbolic groups _ , essays in group theory , msri publ . * 8 * , 75263 , springer , new york , 1987 p. hall and c. heyde , _ martingale limit theory and its application _ , academic press , new york , 1980 v. kaimanovich , _ the poisson formula for groups with hyperbolic properties _ , annals of mathematics , * 152 * ( 2000 ) , 659692 v. kaimanovich and a. vershik , _ random walks on discrete groups : boundary and entropy _ , ann .
* 11 * ( 1983 ) , no .
3 , 457490 s. kakutani , _ random ergodic theorems and markoff processes with a stable distribution _ , proceedings of the second berkeley symposium on mathematical statistics and probability , 1950 , university of california press , berkeley ( 1951 ) , 247261 m. kapovich and b. kleiner , _ hyperbolic groups with low dimensional boundary _ , ann .
sci . cole norm .
( 4 ) * 33 * ( 2000 ) , no . 5 , 647669 h. kesten , _ full banach means on countable groups _ , math .
* 7 * ( 1959 ) , 146156 c. maclachlan and a. reid , _ the arithmetic of hyperbolic @xmath97-manifolds _ , springer - verlag gtm * 219*. springer - verlag , new york , 2003 . b. maskit , _ kleinian groups _ ,
287*. springer - verlag , berlin , 1988 .
p. nicholls , _ the ergodic theory of discrete groups _ ,
lms * 143 * , cambridge university press , cambridge , 1989 s. patterson , _ the limit set of a fuchsian group _
, acta . math .
* 136 * ( 1976 ) , 241273 g. perelman , _ the entropy formula for the ricci flow and its geometric applications _
, preprint arxiv : math/0211159 g. perelman , _ ricci flow with surgery on three - manifolds _ , preprint arxiv : math/0303109 j .- c .
picaud , _ cohomologie borne des surfaces et courants godsiques _ , bull .
france * 125 * ( 1997 ) , no .
1 , 115142 m. pollicott , _ a complex ruelle - perron - frobenius theorem and two counterexamples _ , ergodic theory dynam .
systems * 4 * ( 1984 ) , no .
1 , 135146 m. pollicott and r. sharp , _ comparison theorems and orbit counting in hyperbolic geometry _ , trans . ams * 350 * ( 1998 ) , no .
2 , 473499 d. ruelle , _ thermodynamic formalism _ , addison - wesley publishing co. , reading , mass . , 1978 r. sharp , _ local limit theorems for free groups _ , math . ann .
* 321 * ( 2001 ) , 889904 r. schoen , _ estimates for stable minimal surfaces in three - dimensional manifolds _ ,
seminar on minimal submanifolds , pp .
111126 , ann . of math
. stud . * 103 * , princeton univ .
press , princeton , n.j .
j. stallings , _ on torsion - free groups with infinitely many ends _
, ann . math .
( 2 ) * 88 * ( 1968 ) , 312334 m. steele , _
kingman s subadditive ergodic theorem _ , ann .
henri poincar * 25 * ( 1989 ) , no . 1 , 9398 d. stroock , _ probability theory , an analytic view _ , cambridge university press , cambridge , 1993 d. sullivan , _ on the ergodic theory at infinity of an arbitrary discrete group of hyperbolic motions _ , riemann surfaces and related topics : proceedings of the 1978 stony brook conference ( state univ .
new york , stony brook , n.y . , 1978 ) , pp .
465-496 , ann . of math .
stud . * 97 * , princeton univ . press , princeton , n.j . ,
d. sullivan , _ the density at infinity of a discrete group of hyperbolic motions _ , publ .
ihes * 50 * ( 1979 ) , 171202 w. thurston , _
@xmath97-manifolds , foliations and circles ii _ , preprint n. varopoulos , l. saloff - coste , and t. coulhon , _ analysis and geometry on groups _ , cambridge tracts in math . , * 100 * , cambridge university press , cambridge , 1992 .
w. woess _ random walks on infinite graphs and groups _ , cambridge university press , cambridge , 2000 | these notes are a self - contained introduction to the use of dynamical and probabilistic methods in the study of hyperbolic groups .
most of this material is standard ; however some of the proofs given are new , and some results are proved in greater generality than have appeared in the literature .
[ subsection ] [ theorem]lemma [ theorem]proposition [ theorem]corollary [ theorem]conjecture [ theorem]question [ theorem]problem [ theorem]definition [ theorem]construction [ theorem]notation [ theorem]remark [ theorem]example |
the loss of particle number as a good quantum number is a serious deficiency of the bcs and hfb ( hartree fock bogolyubov ) approximations in their applications to finite nuclei .
however , it can be restored with relative ease in a coherent - state representation @xcite in a manner that preserves much of the elegance and simplicity of the models . following early suggestions by mottelson @xcite , number conserving extensions of bcs and hfb theory
have been considered by many authors : cf .
, for example , the number - projected bcs approximation of dietrich , mang , and pradal @xcite , the projected quasi - particle model of lande , ottaviani and savoia @xcite , the antisymmetrized geminal power states of coleman _ et al . _
@xcite , the broken pair model of lorazo , gambhir , and others @xcite , talmi s generalized seniority scheme @xcite , and the coherent correlated pair method of vary and plastino @xcite .
a review of the subject has been given by allaart _
_ @xcite .
a common feature of these schemes is to approximate the ground state , for a system in which pair correlations dominate , by a state of the form @xmath0 where @xmath1 is a quasi - particle vacuum state and @xmath2 is a projector onto the space of states of particle number @xmath3 .
we show herein that such a projection can be applied analytically , when the wave functions are expressed in a coherent - state representation .
more importantly , projection from quasi - particle states gives a complete set of @xmath3-particle states which reflect many of the properties of the states from which they were projected .
thus , a potentially exact formalism emerges which is able to take advantage of the physical insights gained from the bcs and hfb approximations .
analytic number projection when combined with analytic methods for projecting out states of good angular momentum from deformed intrinsic states @xcite raises the possibility of developing good theories of nuclei which exhibit both rotational and superconducting properties
. it also suggests a way of extending the physically significant definition of generalized seniority to the classification of complete sets of orthonormal states . for simplicity ,
we restrict consideration in this paper to the bcs model . however , the methods extend to the general hb theory by the number projection methods given in ref . @xcite .
for each @xmath4-shell , there is an su(2 ) quasispin algebra spanned by the operators @xmath5 where @xmath6 are creation and annihilation operators for a nucleon ( e.g. , neutron ) which satisfy the fermion anticommutation relations @xmath7 note that we follow einstein s convention of using upper indices to label the components of a tensor contragredient to a tensor labeled by lower indices . with this convention ,
the annihilation operator @xmath8 with superscripts is the hermitian adjoint of the creation operator @xmath9 and the annihilation operator @xmath10 with subscripts is the @xmath11 component of a spherical tensor .
the contraction of a pair of upper and lower indices is then a scalar ; e.g. , @xmath12 is the number operator for the @xmath4-shell .
the quasispin operators satisfy the usual su(2 ) commutation relations @xmath13
= 2\hat s^j_0 \ , , \quad [ \hat s^j_0 , \hat s^j_\pm ] = \pm \hat s^j_\pm \ , , \ ] ] and the nucleon operators transform under su(2 ) as components of quasispin@xmath14 tensors : @xmath15 = a_{jm},&\quad & [ \hat s^j_+\ , , a_{jm}]=a^\dagger_{jm}\ , , \cr [ \hat s^j_0 , a^\dagger_{jm } ] = \half a^\dagger_{jm}\ , , & & [ \hat s^j_0 , a_{jm}]=- \half a_{jm}\ , . \cr\end{array}\ ] ] representations of a quasispin algebra are characterized by highest and lowest weights .
for example , the zero - particle state @xmath16 is a lowest weight state for all the quasispin algebras ; @xmath17 where @xmath18 .
such a state is said to have multishell seniority @xmath19 .
more generally , a state @xmath20 of @xmath21 particles that satisfies the equations @xmath22 is said to have seniority @xmath23 and multishell seniority @xmath24 , where @xmath25 .
the multishell seniority defines the values of the lowest weights and provides useful identifiers of the corresponding quasispin irreps .
a coherent state for the combined quasispin algebras is defined by @xmath26 where @xmath27 is a lowest weight state and @xmath28 provides whatever labels are needed , in addition to @xmath29 , to characterize the state . a coherent state is a quasispin transform of a lowest weight state .
the simplicity of the bcs and hfb theories stems from the fact that the quasi - particle operators of the bogolyubov valatin transformation are also quasispin transforms of nucleon creation and annihilation operators .
in particular , a coherent state of the seniority - zero irrep @xmath30 is the vacuum state of the quasi - particles defined , for real values of @xmath31 , by the bogolyubov valatin transformation @xmath32 with @xmath33 and @xmath34 .
this well - known result follows from the observation that @xmath35 although the transformation @xmath36 is not unitary , it is simply related to the unitary transformation @xmath37 for which @xmath38 the relationship is given by the expansion @xmath39 thus , the transformation @xmath40 is unitary , to within a factor @xmath41 , where @xmath42 and @xmath43 .
likewise , the bv transformation is seen as the unitary su(2 ) transformation @xmath44 in a vcs ( vector coherent state ) representation @xcite , a lowest weight state @xmath27 is regarded as an intrinsic state for an irrep of the combined quasi - spin algebras and represented by an intrinsic wave function @xmath45 .
an arbitrary state @xmath46 is then represented by a holomorphic wave function @xmath47 defined over a set of complex coordinates @xmath48 with values @xmath49 for example , if the zero - particle vacuum @xmath16 has intrinsic wave function @xmath50 , an arbitrary quasi - particle vacuum state @xmath51 has wave function @xmath52 with values @xmath53 where @xmath54 is the overlap @xmath55 with @xmath56 .
it follows that the norm of a quasi - particle vacuum state @xmath1 is the value of the function @xmath57 at @xmath58 . in the vcs representation , an element @xmath59 of an su(2 )
quasispin algebra is represented as a differential operator @xmath60 defined by @xmath61(z ) & = & \sum_{\nu\rho jm } \xi_{\nu\rho jm } \langle { \nu\rho jm}| e^{\hat s_-(z)}\hat x |\psi\rangle \nonumber\\ & = & \sum_{\nu\rho jm } \xi_{\nu\rho jm } \langle { \nu\rho jm}|\big(\hat x + [ \hat s_-(z),x ] \nonumber\\ & & \quad + \half [ \hat s_-(z),[\hat s_-(z),x]]\big ) e^{\hat s_-(z ) } |\psi\rangle \,.\end{aligned}\ ] ] evaluation of the right hand side of this expression by means of the identities @xmath62 gives the representation @xmath63 where @xmath64 is a diagonal operator on intrinsic wave functions with eigenvalues @xmath65 the number operator @xmath66 is represented @xmath67
let @xmath68 denote the @xmath69-pair state @xmath70^n |0\rangle = n!\ , \pi_{2n}\ , |x\rangle \,.\ ] ] states of good nucleon number can also be projected from excited quasi - particle states .
consider first the one quasi - particle state @xmath71 . to within a factor @xmath72 , the @xmath73-particle component of such a state is given by @xmath74 a^\dagger_{jm } |0\rangle = { 1\over n ! } \big [ s_+(x)\big]^n |jm\rangle \,,\ ] ] where @xmath75 .
this is a special case of a general result .
* claim 1 : * let @xmath76 denote some combination of @xmath77 nucleon creation operators @xmath78_{jm}\ ] ] which creates a state @xmath79 having the property that @xmath80 the subscript @xmath81 provides whatever labels are needed in addition to @xmath82 to specify the state .
let @xmath83 denote the corresponding quasi - particle operator in which each @xmath9 is replaced by @xmath84
. then @xmath85^n |\mu\beta jm\rangle \,.\ ] ] * proof : * the claim follows from the identity @xmath86 which implies that @xmath87 q.e.d .
it will be noted that to obtain complete sets of linearly independent states by number projection in this way , the states @xmath88 should include not only states that satisfy @xmath89 but also states such as @xmath90 for @xmath91 .
imposing the constraint @xmath92 for all such states is one way , but not the only way , of ensuring that no linear combination of such states is of the form @xmath93 .
the latter states are not needed because the @xmath94-particle states projected from @xmath95 are proportional to those projected from @xmath96 .
the above results lead to the following claim .
* claim 2 : * let @xmath97 denote a complete set of orthonormal states , as defined in claim 1 , that satisfy the constraint @xmath98 .
then the @xmath3-particle states number - projected from the corresponding quasi - particle states @xmath99 form a ( nonorthonormal ) basis for the @xmath3-particle nucleus .
the number - projected state @xmath68 has vcs wave function @xmath100 where @xmath101 and @xmath102 is the component of @xmath57 of degree @xmath69 in its argument .
it turns out , as shown in refs .
@xcite that the function @xmath103 is well - known in the theory of symmetric polynomials @xcite as a schur function ( or s - function ) ; such functions are the characters of fully antisymmetric irreps of the unitary groups and easy to derive .
the definition @xmath104^n |0\rangle\ ] ] implies that @xmath105(zx ) \,.\ ] ] where @xmath106 thus , the needed schur functions satisfy @xmath107 following the methods of ref .
@xcite , this recursion relation is solved by making a change of variables from @xmath108 to the symmetric power functions @xmath109 the recursion relation then becomes @xmath110 \phi^s_{n-1 } \,,\ ] ] where @xmath111 and has solutions @xmath112 the general expression is obtained by means of the identity @xmath113 derived in ref .
@xcite . with this identity ,
the recursion relation is given in the useful form @xmath114 and has explicit solution given by @xmath115 some properties of these functions are discussed in refs.@xcite .
it follows from claim 1 that the vcs wave function for a one quasi - particle state @xmath116 has values given by @xmath117 from the observation that @xmath118 it follows that @xmath119 where @xmath120 this function has values @xmath121 where @xmath122 is the set of quasi - spins with @xmath123 thus , by considering @xmath57 to be one of a set of functions parameterized by values @xmath124 of the quasispins , @xmath125 is seen to be the member of the set with @xmath126 replaced by @xmath127 .
this substitution will be denoted by an operator @xmath128 , i.e. , @xmath129 so that @xmath130 the adjustment of the coherent state wave function @xmath131 by the replacement @xmath132 to take account of the occupation of one state of a single - particle level is an expression of the well - known _ blocking effect_. the number - projected one quasi - particle state @xmath133^n a^\dagger_{jm } |0\rangle\ ] ] is now seen to have vcs wave function with @xmath134 where @xmath135 is the schur function @xmath136 similarly , vcs wave functions are derived for all other number - projected quasi - particle states .
for example , the seniority two states projected from a two quasi - particle state @xmath137_{jm } @xmath138_{jm } \phi^{j_1j_2}_n(zx ) \,,\ ] ] where @xmath139 is the schur function for the quasi - spin set @xmath140 with components @xmath141 the seniority - zero state @xmath142 has vcs wave function given by @xmath143\phi^s_n(\tilde x ) \,,\ ] ] where @xmath144 and @xmath145 .
thus , with the identity ( [ eq:4.47 ] ) , the vcs wave function for the state @xmath142 is given by @xmath146 as expected , this wave function becomes identical to @xmath147 when @xmath148 .
to illustrate the techniques , we consider the simple bcs hamiltonian @xmath149 with @xmath150 for given real values of @xmath151 , the quasi - particle operators are defined by @xmath152 such that @xmath153 and @xmath34 so that @xmath154 from standard bcs theory , the expectation of @xmath155 for a quasi - particle vacuum state is given by @xmath156 \phi^s(x^2 ) \,.\ ] ] with the above expressions for @xmath157 and @xmath158 , this expression becomes @xmath159 \phi^s(x^2 ) \nonumber\\ & = & \sum_j 4s_j \varepsilon_j x^2_j \phi^j ( x^2 ) -g\sum_j
2s_jx_j^4 \phi^{jj}(x^2 ) \nonumber\\ & & - g\sum_{ij } 4s_is_j x_ix_j \phi^{ij}(x^2 ) \,.\end{aligned}\ ] ] the energy of the @xmath69-pair state @xmath68 , @xmath160 is now obtained by picking out the components of @xmath161 and @xmath162 of degree @xmath69 in @xmath58 .
recalling that @xmath163 , we immediately obtain @xmath164 \ , .
\label{eq:5.74})\end{aligned}\ ] ] thus , the values of the @xmath165 coefficients can be fixed such that the energy @xmath166 is minimized and the variational equation @xmath167 is satisfied . as observed above
, the @xmath73-particle component of the one quasi - particle state @xmath168 is the state @xmath169 .
thus , we consider the energies of number - projected one quasi - particle states defined by @xmath170 now the coherent state representation of the number operator @xmath171 , when acting on a seniority zero state , is given by @xmath172 therefore @xmath173 when @xmath165 is assigned the value for which eqn .
( [ eq:3.56 ] ) is satisfied .
it follows that @xmath174 |n(x)\rangle \,.\ ] ] if we now ( temporarily ) regard @xmath165 as a variable parameter , we obtain from claim 1 the identity @xmath175 |n(x)\rangle & = & u_j \langle x| \alpha^{jm } [ h , a^\dagger_{jm } ] |n(x)\rangle \nonumber\\ & = & n ! p_nu_j \langle x| \alpha^{jm } [ h , a^\dagger_{jm } ] |x\rangle \ , , \label{eq:64}\end{aligned}\ ] ] where @xmath176 picks out the component of degree @xmath69 in @xmath58 in what follows it .
note that for a particular value of @xmath165 , the matrix element @xmath177 |x\rangle$ ] is a number .
however , by regarding @xmath165 first as a variable and evaluating the matrix element as a function of @xmath165 , it is meaningful to extract the component of this function of degree @xmath69 ; the rhs of eqn .
( [ eq:64 ] ) is then the value of this component at the assigned value of @xmath165 .
the quasi - particle vacuum properties of the state @xmath1 , which hold for arbitrary @xmath165 , can now be used to put the matrix element on the right into the standard equations - of - motion form @xcite @xmath178 |x\rangle = \langle x|\ { \alpha^{jm } , [ h , a^\dagger_{jm } ] \ }
|x\rangle \,.\ ] ] this matrix element is then evaluated by bcs methods to give @xmath179\ } |x\rangle = \big [ \varepsilon_j u_j^2 - g u_j^2 v_j^2 + gu_jv_j \sum_i 2s_i u_iv_i\big ] \phi^s(x^2 ) \,.\ ] ] the final step of number projection is now easy and gives the expression for the number - projected one quasi - particle energies @xmath180 this is an explicit and simple expression for @xmath181 which involves only the values of known ( schur ) functions at the specific @xmath165 for which @xmath182 is a minimum .
let @xmath183 be an orthonormal set of two - particle seniority - two states .
the matrix elements between the corresponding number - projected quasi - particles states are given by @xmath184 |x\rangle \big ) \,.&\end{aligned}\ ] ] thus , the matrix elements for seniority - two states are given by number projection of corresponding quasi - particle tamm dancoff expressions , as defined in the equations - of - motion formalism @xcite .
moreover , the projections are accomplished by expanding the unprojected expressions in terms of schur functions . after evaluating the overlaps @xmath185 and @xmath186
, one can construct an orthonormal set of seniority - zero states for a @xmath187-particle nucleus .
however , so that use can be made of claim 1 , it is convenient to start with operators @xmath188 that satisfy the condition @xmath189 the overlap @xmath190 is obtained from the value @xmath191 at @xmath192 of the coherent state wave function for the state @xmath193 ; an expression for this wave function is given by eqn .
( [ eq:3.64 ] ) .
the second overlap is given by @xmath194 the matrix element @xmath195 is evaluated starting from the observation that @xmath196 and that @xmath197 vanishes when @xmath192 because of eqn .
( [ eq:3.56 ] ) .
it follows that @xmath198 writing @xmath199 and defining @xmath200 we can use claim 1 to obtain @xmath201 then , with the identity @xmath202 we obtain @xmath203 |x\rangle \big ] \,.\end{aligned}\ ] ]
talmi @xcite has defined the sequence of @xmath204 states proportional to @xmath205^n |0\rangle ; $ ] @xmath206 as states of generalized seniority zero .
likewise , the sequences of @xmath207 states @xmath205^n |2 jm\rangle ; n = 1 , 2 , \ldots \}$ ] , where @xmath208 is a two - particle state that is annihilated by all the quasispin lowering operators ( i.e. , @xmath209 @xmath210 ) , are defined to be states of generalized seniority two .
states of higher generalized seniority are similarly defined .
thus , the concept of generalized seniority identifies subsets of states of the corresponding seniority ( more precisely summed multishell seniority @xmath211 ) .
the generalized seniority zero states , for example , exclude states that are generated by combinations of the @xmath212 operators that are not simply multiples of @xmath213 .
such states , sometimes described as broken @xmath214 pair states @xcite , are also described as generalized seniority - two or higher states , depending on how many @xmath213 pairs are broken .
the problem is that , while one can define two - particle @xmath204 parent states @xmath215 lowering operators , in parallel with lowest - weight quasispin states , the sequences of states @xmath216^n generalized seniority - zero states .
consequently , as talmi has emphasized , generalized seniority does not define a complete orthogonal scheme .
nevertheless , generalized seniority can be given a definition that does characterize a complete set of orthogonal states for each nucleus separately .
for example , one can define a generalized seniority - two creation operator @xmath217 for the @xmath187-particle nucleus such that @xmath218 with the identity @xmath219^n |0\rangle = n!p_n \langle z| \hat s_+(yp ) |x\rangle = n !
p_n \sum_j y_j { \partial\over \partial x_j } \phi^s ( zx ) \,,\ ] ] this equation reduces to the easy to solve equation for the @xmath220 coefficients @xmath221 states of higher generalized seniority may be defined similarly .
such a definition , has not been used to our knowledge probably because it would appear to destroy the simplicity and elegance of the concept .
however , the facility to carry out number - projection analytically suggests that this may no longer be a significant concern .
to check the above methods , they have been applied to a simple two - level model having hamiltonian @xmath222 with @xmath223 and @xmath224 ; the excitation energy @xmath225 then sets the unit of the energy scale . for simplicity
, we also set @xmath226 and considered a model nucleus with @xmath227 nucleons ; this corresponds to a situation in which , in the @xmath228 limit , the lower level is fully occupied and the upper level empty .
for such a two - level model , we can set @xmath229 because replacing these parameters values by the substitution @xmath230 , @xmath231 would only change the overall norm of the state @xmath232^n |0\rangle \to k^n [ \hat s^1_+ + x\hat s^2_+ ] ^n |0\rangle \,.\ ] ] the energy @xmath233 of the state , with @xmath227 , is then evaluated directly from eqn .
( [ eq:5.74 ] ) and the value of @xmath165 , for which it is a minimum , determined .
the structure of the state @xmath68 is revealed by expanding it on a basis of states labeled by the two - level quasispin quantum numbers @xmath234 the state @xmath235 is the state with m nucleon pairs in the upper level and @xmath236 pairs ( or @xmath11 hole pairs ) in the lower level .
the expansion @xmath237 in this basis is easily inferred from the identity @xmath238 which implies that the squares of the @xmath239 coefficients are given by the corresponding expansion of @xmath240 as a polynomial in @xmath241 ; @xmath242 the @xmath239 coefficients are shown , for @xmath243 and three values of @xmath244 , in comparison with exactly computed wave functions for the model in fig .
[ fig : wfns ] .
= 3.5 in for @xmath243 , the model corresponds to 14 nucleons in two @xmath245 levels .
results computed for other values of @xmath227 show that the number - projected wave functions rapidly become exact as @xmath246 . as an indication of the improvement given by number projection over unprojected bcs results
is given by comparing the ratio of the number of nucleons occupying the upper level to the number in the lower level .
the mean number in the upper level is given for the above wave functions by @xmath247 and the mean number for the lower level is @xmath248 fig .
[ fig : ratios ] shows the ratio @xmath249 for several values of the pairing interaction in comparison to the exact ratios and those of the bcs approximation .
= 3.5 in as @xmath250 , the occupancies given by the bcs approximation are found to approach those of an exact calculation .
however , for a finite number of particles , a considerable improvement is gained by number projection ( as expected from previous results ) .
the calculations that have been done , cf . for example refs .
@xcite , show clearly that number projection effects very substantial improvement to the accuracy of the bcs approximation in applications to finite nuclei .
moreover , the coherent state techniques introduced in refs . @xcite and @xcite were found to facilitate the practical steps of carrying out the number projection considerably .
these techniques have been developed in this paper in the hope that they will be taken up by others in calculations of the low - energy spectra of nuclei in which pair - correlations are dominant .
of particular interest are the low - energy spectra of single - shell nuclei with both pairing and quadrupole interactions where the objective will be explore the emergence of deformed rotational states .
other recent developments that may be usefully deployed in concert with number projection are : the identification of a range of seniority - conserving interactions @xcite , and analytic techniques to carry out angular momentum projection from a class of deformed intrinsic states @xcite . | as shown in previous work , number projection can be carried out analytically for states defined in a quasi particle scheme when the states are expressed in a coherent state representation .
the wave functions of number projected states are well known in the theory of orthogonal polynomials as schur functions .
moreover , the functions needed in pairing theory are a particularly simple class of schur functions that are easily constructed by means of recursion relations .
it is shown that complete sets of states can be projected from corresponding quasi
particle states and that such states retain many of the properties of the quasi particle states from which they derive .
it is also shown that number projection can be used to construct a complete set of orthogonal states classified by generalized seniority for any nucleus .
epsf.tex amssym.def |
we would like to thank the technical staff of the elsa and mami machine groups and of all the participating institutions of their invaluable contributions to the success of the experiment .
we acknowledge financial support from the deutsche forschungsgemeinschaft ( dfg ) within the sfb / tr16 and from the schweizerische nationalfond . the collaboration with st .
petersburg received funds from dfg and rfbr .
u.thoma thanks for an emmy noether grant from the dfg .
sarantsev acknowledges support from rssf .
yao _ et al .
_ , j. phys .
g * 33 * ( 2006 ) 1 .
meiner , j.w .
durso , nucl .
phys . a * 430 * ( 1984 ) 670 . c. hajduk , b. schwesinger , phys .
b * 140 * ( 1984 ) 172 .
s. capstick , n. isgur , phys .
d * 34 * ( 1986 ) 2809 .
u. lring _ et al .
_ , eur .
j. a * 10 * ( 2001 ) 395 .
u. lring _ et al .
_ , eur .
j. a * 10 * ( 2001 ) 447 .
glozman _ et al .
_ , phys .
d * 58 * ( 1998 ) 094030 .
t. burch _ et al .
_ , phys .
d * 74 * ( 2006 ) 014504 .
o. krehl _ et al .
_ , phys .
c * 62 * ( 2000 ) 025207 .
s. schneider _ et al .
_ , eur .
j. a * 28 * ( 2006 ) 107 .
li _ et al .
_ , phys .
d * 46 * ( 1992 ) 70 .
s. capstick , p.r .
page , phys .
d * 60 * ( 1999 ) 111501 .
t. nakano _ et al .
_ , phys .
* 91 * ( 2003 ) 012002 .
barmin _ et al .
_ , phys . atom
. nucl .
* 66 * ( 2003 ) 1715 [ yad .
* 66 * ( 2003 ) 1763 ] .
s. stepanyan _ et al .
_ , phys . rev
. lett .
* 91 * ( 2003 ) 252001 .
j. barth _ et al .
_ , phys .
b * 572 * ( 2003 ) 127 .
jaffe and f. wilczek , phys .
* 91 * ( 2003 ) 232003 .
d. diakonov _ et al .
_ , z. phys . a * 359 * ( 1997 ) 305
r. a. arndt _ et al .
_ , phys .
c * 69 * ( 2004 ) 035208 .
m. battaglieri _ et al .
_ , phys .
* 96 * ( 2006 ) 042001 .
s. niccolai _ et al .
_ , phys . rev
. lett .
* 97 * ( 2006 ) 032001 .
morsch , p. zupranski , phys .
c * 61 * ( 2000 ) 024002 .
h. p. morsch and p. zupranski , phys .
c * 71 * ( 2005 ) 065203
. j. patzold _ et al .
_ , phys .
c * 67 * ( 2003 ) 052202 .
j. ahrens _ et al .
news * 4 * ( 1994 ) 5 .
w. hillert , eur .
j. a * 28s1 * ( 2006 ) 139 .
_ , `` n * and @xmath54 * decays into @xmath55 '' , arxiv:0707.3592 .
m. kotulla _ et al .
b * 578 * ( 2004 ) 63 .
f. harter _ et al .
_ , phys .
b * 401 * ( 1997 ) 229 .
m. wolf _
et al . _ ,
j. a * 9 * ( 2000 ) 5 .
s.j . hall _
et al . _ ,
meth . a * 368 * ( 1996 ) 698 .
r. novotny , ieee trans . nucl
* 38 * ( 1991 ) 379 .
et al . _ , nucl .
instr . meth .
a * 346 * ( 1994 ) 168 .
m. kotulla _ et al .
* 89 * ( 2002 ) 272001 .
y. assafiri _ et al .
_ , phys .
* 90 * ( 2003 ) 222001 .
anisovich _ et al .
_ , eur .
phys . j. a * 24 * ( 2005 ) 111 .
anisovich and a.v .
sarantsev , eur .
phys . j. a * 30 * ( 2006 ) 427 .
anisovich _ et al .
_ , eur .
j. a * 25 * ( 2005 ) 427 .
sarantsev _ et al .
_ , eur .
j. a * 25 * ( 2005 ) 441 .
anisovich _ et al .
_ , baryon resonances and polarization transfer in hyperon photoproduction " , arxiv:0707.3596 .
et al . _ , `` evidence for two negative - parity @xmath56 excitations decaying into @xmath57 '' , in preparation .
o. bartholomy _ et al .
_ , phys .
* 94 * ( 2005 ) 012003
. h. van pee _
et al . _ ,
j. a * 31 * ( 2007 ) 61 .
o. bartalini _ et al .
_ , eur .
phys . j. a * 26 * ( 2005 ) 399 .
m. dugger _
et al . _ ,
c * 76 * ( 2007 ) 025211 .
r. a. arndt _ et al .
_ , phys .
c * 74 * ( 2006 ) 045205 .
s. prakhov _ et al .
_ , phys .
c * 69 * ( 2004 ) 045202 .
arndt _ et al .
_ , http://gwdac.phys.gwu.edu .
gomez tejedor and e. oset , nucl .
phys . a * 600 * ( 1996 ) 413 .
nacher _ et al .
_ , nucl .
a * 695 * ( 2001 ) 295 .
e. hernandez , e. oset and m. j. vicente vacas , phys .
c * 66 * , ( 2002 ) 065201 . | properties of the roper resonance , the first scalar excitation of the nucleon , are determined .
pole positions and residues of the @xmath0 partial wave are studied in a combined analysis of pion- and photo - induced reactions .
we find the roper pole at @xmath1mev and an elasticity of @xmath2 .
the largest decay coupling is found for the @xmath3 ( @xmath4-@xmath5-wave ) .
the analysis is based on new data on @xmath6 for photons in the energy range from the two - pion threshold to 820mev from taps at mainz and from 0.4 to 1.3gev from crystal barrel at bonn and includes further data from other experiments .
the partial wave analysis excludes the possibility that the roper resonance is split into two states with different partial decay widths .
+ _
pacs : 11.80.et , 13.30.-a , 13.40.-f , 13.60.le_ , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , and + the lowest - mass excitation of the nucleon , the roper @xmath7 resonance with spin and parity @xmath8 , and the second scalar nucleon excitation @xmath9 @xcite , remain to be the most enigmatic states in baryon spectroscopy . in the bag model @xcite and in the skyrme model
@xcite , the roper resonance was interpreted as surface oscillation , also called breathing mode . in quark models , two low - mass scalar excitations of the nucleon
are predicted . using a linear confining potential and one - gluon - exchange @xcite or instanton - induced interactions @xcite , a level ordering
is calculated in which the mass of the @xmath7 exceeds the mass of the negative - parity state @xmath10 by 80mev ; experiments find it @xmath11mev below .
the spacing between the two scalar excitations is predicted to be @xmath12mev @xcite while experiments find 270mev @xcite .
when one - gluon exchange interactions are replaced by exchanges of goldstone bosons , the @xmath7 mass can well be reproduced @xcite , the @xmath9 mass was not calculated .
lattice gauge calculations indicate that the first scalar excitation of the nucleon should be expected above @xmath10 @xcite .
compared to model and lattice predictions , the mass of the roper resonance is too small ; compared to other low - mass resonances , its width too large .
these problems would not occur if @xmath13 were the first radial scalar excitation of the proton .
the roper resonance can then be interpreted within a coupled - channel meson exchange model based on an effective chiral - symmetric lagrangian @xcite ; no genuine @xmath14 ( 3 quark ) resonance was needed to fit @xmath15 phase shifts and inelasticities , in agreement with @xcite .
motivated by the @xmath16 dependence of the roper helicity amplitude which would seem to suggest a hybrid nature @xcite , capstick and page @xcite calculated masses of baryonic hybrids .
their masses were , however , too large to interpret the roper resonance as a hybrid .
the @xmath17 , a baryon with positive strangeness , which may have been observed in low - statistics photo - production experiments @xcite , made the roper resonance @xcite and/or the @xmath13 @xcite to viable pentaquark candidates .
the existence of a very narrow @xmath0 state in the mass region 1650 - 1750 mev was investigated in @xcite . the fading evidence for @xmath17 @xcite makes this interpretation less attractive . morsch and zupranski @xcite suggested the roper mass region might house two resonances , one at 1390mev with a small elastic width and large coupling to @xmath18 , and a second one at higher mass around 1460mev with a large elastic width and small @xmath18 coupling .
the former resonance was found to be produced in @xmath15 scattering , and in @xmath19-proton scattering using an @xmath19 beam of @xmath20gev kinetic energy ; the latter resonance was suggested to be excited by @xmath21 .
the two resonances may have rather different wave functions @xcite .
studies of the reaction @xmath22 suggested that the low - energy tail of the roper resonance might decay to both @xmath3 and @xmath23 @xcite .
obviously , the @xmath0 partial wave is not sufficiently constrained by precision data and supports very different interpretations . in this letter
, we present data on @xmath24 of the a2-taps collaboration at the mainz microtron ( mami ) electron accelerator @xcite and of the cb - elsa collaboration at the bonn electron stretcher accelerator ( elsa ) @xcite .
the bonn set up and the analysis method was described briefly in the preceding paper @xcite . here
we give only a short summary of the taps setup , for more details see @xcite .
earlier data taken at mami @xcite have smaller statistics and are not discussed here .
the photon energy at mami covered the range 285820mev .
the photon energies were measured in the glasgow tagged photon facility @xcite with an average energy resolution of 2mev .
the taps detector @xcite consisted of six blocks each with 62 hexagonally shaped baf@xmath25 crystals arranged in an 8@xmath268 matrix and a forward wall with 138 baf@xmath25 crystals arranged in a 11@xmath2614 rectangle .
this setup covered @xmath2740% of @xmath28 .
the @xmath29 reaction channel was identified by constructing the 4-momenta of the two neutral pions from their @xmath30 decays ; proton detection was not required in the analysis .
the @xmath31 mesons were detected via their @xmath32 decay and identified by their invariant mass .
the mass of the missing particle was calculated from the four - momenta of the pions , and the beam energy @xmath33 using the mass of the target proton . the resulting distribution is shown in fig .
[ gg_vs_gg]a and demonstrates the unambiguous identification of the reaction @xmath34 . at incident beam energies above the @xmath35 production threshold of @xmath36 mev ,
a possible background from the @xmath37 decay with one undetected @xmath31 can be cleanly separated from the reaction of interest ( see fig . [
gg_vs_gg]a ) .
further details are given in @xcite . in fig .
[ gg_vs_gg]b the total cross section is displayed .
two peaks due to the second and third resonance region are observed , with peaks at @xmath38 and @xmath39mev .
there is good general agreement between the three data sets .
the graal data @xcite fall off at high masses more rapidly than the cb - elsa data . at low energies , the taps data fall below the cb - elsa data while the peak cross sections of all 3 experiments agree reasonably well .
the discrepancies show the difficulties of extracting total cross sections when the full phase space is not covered by the detector .
note that the extrapolation was done differently : the cb - elsa and the a2-taps collaborations used the result of this partial wave analysis ; the graal collaboration used a simulation based on @xmath40 and @xmath41 phase space . the inclusion of both , cb - elsa and taps data , provides an additional tool to estimate the systematic error of the experimental data . the fit curves in fig .
[ gg_vs_gg]b are discussed below .
[ cols="^,^,^ " , ] but are defined more precisely here .
the helicity coupling agrees with the pdg mean value but from the variety of different solutions we estimate a larger error .
note that our helicity amplitude is defined in the complex @xmath42 plane at the pole position of the roper .
the elastic width ( @xmath43 ) is compatible with previous findings ( 60 - 70% ) .
its decay fraction into @xmath23 ( @xmath44 ) is not in conflict with the pdg mean value ( 20 - 30% ) ; only the @xmath3 partial decay width deviates significantly from pdg .
we find @xmath45 while pdg gives 5 - 10% .
due to its larger phase space , decays into @xmath46 are more frequent than those into @xmath3 even though the latter decay mode provides the largest coupling . for a radial excitation ,
this is not unexpected : about 50% of all @xmath47 resonances decay into j/@xmath48 , more than 25% of @xmath49 resonances decay via @xmath50 @xcite .
the large value of @xmath51 might therefore support the interpretation of the roper resonance as radial excitation .
an alternative interpretation of the @xmath18 decay is offered by hernandez , oset and vicente - vacas @xcite who take into account the re - scattering of the two final state pions .
the authors of @xcite do not fit data ; instead they show that they can reproduce qualitatively the phenomenology of @xmath52 decays by rescattering thus avoiding the need to introduce a genuine @xmath53 amplitude . in this letter
, we have presented new data on photoproduction of two neutral pions in the energy range from the 2@xmath31 production threshold up to a photon energy of 820mev ( mainz ) and up to 1300mev ( bonn ) and reported results from a partial wave analysis of this and of related reactions .
the focus of this letter is the roper resonance .
we show that the data are incompatible with the conjecture that conflicting results on its properties could originate from the presence of two similar resonances and their interference , where both are in the @xmath0 wave and both fall into the 1300 to 1500mev mass region . due to the fact that the roper properties are over - constrained by the data , we can rule out this possibility .
the decay pattern is consistent with an interpretation of the roper resonance as first radial excitation of the nucleon . |
let @xmath17 be a prime number , @xmath0 a number field , and @xmath18 a finite set of places of @xmath0 .
let @xmath19 be the compositum of all extensions of @xmath0 ( in a fixed algebraic closure @xmath20 ) which are unramified outside @xmath18 , and put @xmath21 for its galois group .
these _ arithmetic fundamental groups _ play a very important role in number theory .
algebraic geometry provides the most fruitful known source of information concerning these groups .
namely , given a smooth projective variety @xmath22 , the @xmath17-adic tale cohomology groups of @xmath23 are finite - dimensional vector spaces over @xmath24 equipped with an action of @xmath25 where @xmath18 consists of the primes of bad reduction for @xmath22 together with the primes of @xmath0 of residue characteristic @xmath17 .
the richness of this action can be judged , for example , by the intimate relationships between algebraic geometry and the theory of automorphic forms which it mediates . for this and many other reasons
, it would be difficult to overstate the importance of these @xmath17-adic galois representations .
nonetheless , linear @xmath17-adic groups simply form too restrictive a class of groups to capture _ all _ galois - theoretic information , and some important conjectures in the subject , notably the fontaine - mazur conjecture @xcite ( to mention only one , see the discussion in section [ sec : conjectures ] ) , point specifically toward the kind of information inside arithmetic fundamental groups which can not be captured by finite - dimensional @xmath17-adic representations . in this work ,
we discuss a method for studying finitely ramified extensions of number fields via _ arithmetic dynamical systems on _ @xmath26 .
at least conjecturally , this method provides a vista on a part of @xmath25 invisible to @xmath17-adic representations .
we now sketch the construction , which is quite elementary .
let @xmath0 be a perfect field , and suppose @xmath27 $ ] is a polynomial of degree @xmath28 such that its derivative @xmath29 is not identically @xmath30 in @xmath31 $ ] . for each @xmath32 ,
let @xmath33 be the @xmath7-fold iterate of @xmath8 , i.e. @xmath34 and @xmath35 for @xmath32 .
let @xmath1 be a parameter for @xmath36 with function field @xmath2 .
we are interested in the tower of branched covers of @xmath37 given by @xmath38,\ ] ] as well as extensions of @xmath0 obtained by adjoining roots of its specializations at arbitrary @xmath39 .
the variable - separated polynomial @xmath12 is clearly absolutely irreducible ( since it is linear in @xmath1 ) and of degree @xmath40 in @xmath41 $ ] ; it is separable over @xmath9 by the assumption that @xmath29 is not identically 0 .
fix an algebraic closure @xmath42 of @xmath9 , and let @xmath20 be the algebraic closure of @xmath0 determined by this choice , i.e. the subfield of @xmath11 consisting of elements algebraic over @xmath0 . for @xmath32 ,
let @xmath43 be the set of roots in @xmath11 of @xmath12 ; it has cardinality @xmath40 .
we denote by @xmath44 the @xmath45-regular rooted tree whose vertex set is @xmath46 , and whose edges connect @xmath47 to @xmath48 exactly when @xmath49 ; its root ( at ground level ) is @xmath1 .
let us choose and fix an _ end _
@xmath50 of this tree ; in other words , we have @xmath51 and @xmath52 for @xmath13 . for each @xmath13 , we consider the field @xmath53/(\ph_n)$ ] and its galois closure @xmath54 over @xmath9 .
let @xmath55 be the integral closure of @xmath56 $ ] in @xmath57 .
corresponding to each @xmath58 , we may fix compatible _ specialization maps _ @xmath59 with image @xmath60 , a normal extension field of @xmath0 and put @xmath61 for the correspoding compatible system of roots of @xmath62 .
we denote by @xmath63 the image of the restriction of @xmath64 to @xmath65 .
we refer the reader to subsection [ subsec : specialization ] for more details , but we should emphasize here that @xmath62 is not necessarily irreducible over @xmath0 ; hence , although @xmath60 depends only on @xmath66 and @xmath67 , the isomorphism class of @xmath63 depends _ a priori _ on the choice of @xmath68 as well as on the choice of compatible @xmath64 .
also , the galois closure of @xmath69 is contained in , but possibly distinct from , @xmath60 . nonetheless ,
unless stated otherwise , @xmath68 and @xmath67 are arbitrary but fixed , and in this case we will usually not decorate @xmath63 with @xmath68 and occasionally we may write simply @xmath70 instead of @xmath71 . taking the compositum over all @xmath13
, we obtain the _ iterated extension _
@xmath72 attached to @xmath8 , with galois closure @xmath73 over @xmath9 .
similarly for each @xmath74 , we obtain a specialized iterated extension @xmath75 with galois closure over @xmath0 contained in @xmath76 .
we put @xmath77 for the _ iterated monodromy group _ of @xmath8 and for @xmath39 , we denote by @xmath78 its specialization at @xmath67 .
the group @xmath79 has a natural and faithful action on the tree @xmath44 , hence comes equipped with a rooted tree representation @xmath80 . for more on rooted trees and
iterated monodromy groups ( in a more general context , in fact ) , see nekrashevych @xcite as well as bartholdi - grigorchuk - nekrashevych @xcite .
since we are interested in finitely ramified towers ( meaning those where only finitely many places of the base field are ramified ) , we need to answer the following question : which polynomials @xmath8 have the property that the corresponding iterated tower @xmath81 , as well as all of its specialziations @xmath82 , are finitely ramified ?
we first recall some standard terminology from polynomial dynamics .
we say that @xmath83 is _ periodic _ for @xmath8 if @xmath84 for some @xmath13 .
moreover , @xmath85 is _ preperiodic _ for @xmath8 if for some @xmath86 , @xmath87 is periodic for @xmath8 ; equivalently , @xmath88 is preperiodic for @xmath8 means that @xmath89 , i.e. the orbit of @xmath88 under the iterates of @xmath8 , is a finite set .
we put @xmath90 for the set of affine _ ramification _ and _ branch _ points , respectively .
the elements of @xmath91 and @xmath92 are also the _ critical points _ , respectively _ critical values _ of @xmath8 .
the polynomial @xmath8 is called _ post - critically finite _ if every member of @xmath91 is a preperiodic point for @xmath8 .
in other words , @xmath8 is post - critically finite exactly when the _ post - critical set _
@xmath93 , i.e. the union of the orbits of critical points under the iterates of @xmath8 , is a finite set .
it has long been known that the post - critical set plays a crucial role in the dynamics of the polynomial .
indeed , the class of dynamical systems corresponding to post - critically finite polynomials is a well - studied one , having gained prominence following a celebrated theorem of thurston ; see , for example , douady - hubbard @xcite , bielefed - fisher - hubbard @xcite , as well as the papers by poirier @xcite , pilgrim @xcite , and pakovich @xcite ; the latter two concern the connection with actions of @xmath94 on certain finite trees .
our starting point is the following characterization of finitely ramified iterated extensions .
[ thm:1 ] the iterated tower of function fields @xmath15 is finitely ramified if and only if @xmath8 is post - critically finite .
if @xmath8 is post - critically finite , every specialization @xmath95 of this tower is finitely ramified .
the first assertion of the theorem is clear geometrically since @xmath96 . the second assertion , however , is not a formal consequence of the first , since any element of @xmath0 is a unit in @xmath9 ; the proof we give for it proceeds via proposition [ prop : recur ] , where we calculate a recurrence for the discriminant of @xmath12 ( valid for an arbitrary polynomial @xmath8 ) , giving a more precise version of the theorem .
the proof of proposition [ prop : recur ] rests on a riemann - hurwitz genus formula for polynomials @xcite .
now let us return to the case of a number field @xmath0 .
for each post - critically finite @xmath27 $ ] , and each @xmath97 , theorem [ thm:1 ] provides a surjection @xmath98 for an effectively determined finite set @xmath99 of places of @xmath0 ( see definition [ def : s ] and corollary [ cor : imr ] ) .
we call @xmath100 the _ iterated monodromy representation _ attached to @xmath8 and @xmath67 .
the study of automorphism groups of rooted trees is a relatively new and quite active topic in group theory ( see @xcite , @xcite , and @xcite ) .
the structure of non - abelian subgroups of these automorphism groups appears to be quite different from that of linear @xmath17-adic groups ( see the papers just cited as well as bux - perez @xcite ) .
the natural action of iterated monodromy groups on rooted trees leads us to the expectation that iterated monodromy representations @xmath100 attached to post - critically finite polynomials @xmath101 $ ] have the potential of revealing aspects of arithmetic fundamental groups which are not visible to @xmath17-adic representations ; see the discussion in section [ sec : conjectures ] as well as boston s preprint @xcite , where tree representations are suggested as the proper framework for studying finitely ramified tame extensions .
since all finitely ramified @xmath17-adic galois representations with infinite image are expected , by a conjecture of fontaine and mazur , to be wildly ( even deeply ) ramified at some prime of residue characteristic @xmath17 , an immediate question is what can be said about the presence of wild ramification in specialized iterated extensions @xmath95 .
experimentation leads to the expectation that generically the primes of residue characteristic dividing @xmath45 ramify deeply in @xmath95 .
for example , if @xmath102 with @xmath103 and @xmath104 , then for all @xmath105 , the extensions @xmath95 are deeply ramified at all @xmath17 dividing @xmath45 .
( see , however , questions [ ques:1 ] and [ ques:2 ] in [ sec : conjectures ] ) .
under an assumption of good reduction for @xmath8 , we prove a partial result toward this expectation , namely for integral @xmath67 , we estimate from below the power of @xmath17 dividing the discriminant of @xmath62 . to be precise , in
[ sec : wild ] , we will prove the following theorem .
[ thm : wild1 ] let @xmath0 be a number field .
suppose @xmath27 $ ] is post - critically finite , has degree divisible by @xmath17 , and has good reduction at a valuation @xmath47 of residue characteristic @xmath17 , i.e. @xmath8 has @xmath47-integral coefficients with @xmath47-unital leading coefficient .
then for any @xmath106 , @xmath107 assuming @xmath62 is @xmath0-irreducible for all @xmath7 , this estimate shows that the tower of rings @xmath108 $ ] , where @xmath109 is a compatible sequence of roots of @xmath62 , is wildly ramified at @xmath17 .
note that @xmath108 $ ] is an _
order _ inside the maximal order of @xmath110 ; it is the discriminant of the latter which is our primary interest , but the theorem estimates the discriminant of the former .
this is one sense in which the above theorem is only a partial answer to our question about wild ramification in iterated extensions . on the other hand , in section [ sec : quad ] , we illustrate with the tower corresponding to @xmath111 , the possibility that the orders @xmath112 $ ] ( for a large set of @xmath113 ) are maximal , giving examples of monogne number fields .
the organization of this article is as follows . in [ sec : prelim ] , we outline some preliminary facts regarding post - critically finite polynomials , including a classification of the very simplest examples for each degree , namely those that are _ critically fixed _ ( every critical point is fixed , also known as _ conservative _ ) and _ simply ramified _ ( every non - trivial ramification index is @xmath114 ) . in [ sec : pcf ] and [ sec : wild ] , we prove theorems [ thm:1 ] [ thm : wild1 ] ) , respectively . in [ sec : decomposition ] , we describe the decomposition of unramified primes in iterated towers in terms of simple properties of certain finite graphs . in [ sec : quad ] , we study the quadratic case in more detail , obtaining a recursion for writing down post - critically finite quadratic polynomials , which give number fields of independent interest ; we also discuss the example @xmath115 in detail , proving monogenicity of certain number fields .
finally , in [ sec : conjectures ] , we outline a number of questions and open problems .
* acknowledgments . *
work on this project began in june 2003 during visits by wa and cm to amherst supported partially by nsf grant dms-0226869 .
a few months later , the preprint @xcite of v. nekrashevych was posted on ` arxiv.org ` ; in it , iterated monodromy groups from a more general perspective are studied .
we learned from @xcite that the construction ( [ pn ] ) we had been studying had earlier been suggested by r. pink ( private communication ) .
a special case also occurs in boston s preprint @xcite .
cm thanks r. pink for a helpful conversation .
cm and fh would like to thank e. bayer - fluckiger for financial support of their visits to epfl , lausanne in 2004 where a portion of this work was completed .
we are grateful to r. benedetto and p. gunnells for useful discussions .
in this section , @xmath0 is a perfect field and @xmath117 $ ] is a polynomial of degree @xmath28 whose derivative @xmath29 is not identically @xmath30 .
we maintain all other notation introduced in 1 .
thinking of @xmath8 as a branched cover @xmath118 of degree @xmath45 , the singular fibers are those of cardinality less than @xmath45 . leaving aside @xmath119 which is totally ramified
, the points in a singular fiber ( the ramification points ) are exactly the critical points , i.e. the roots of @xmath29 : writing @xmath120 for any @xmath121 , we have @xmath122 hence @xmath123 divides @xmath124 if and only if @xmath125 . the critical values ( the images under @xmath8 of the critical points ) , are the points having a singular fiber , i.e. they are exactly the branch points . in algebraic language
, @xmath126 is in @xmath92 if and only if @xmath127 has a multiple root , which happens if and only if @xmath128 has @xmath129 as a root . in other words
, @xmath130 is a branch point if and only if the system @xmath131 has a common root @xmath132 , and these roots are the ramification points above @xmath130
. we could adopt the convention that @xmath91 and @xmath92 are `` multisets '' where each critical point or critical value occurs according to the multiplicity of the corresponding roots of @xmath29 , but to avoid confusion , we will be explicit about the multiplicities by writing @xmath133 and putting , for @xmath134 , @xmath135 [ lem : integral ] for each @xmath13 , @xmath12 is separable and absolutely irreducible over @xmath9 .
the ring @xmath136 $ ] is integrally closed ( in its fraction field @xmath137 ) .
all of this follows essentially from the fact that @xmath138 never vanishes .
the reader can easily check the absolute irreducibility of @xmath139 . for separability ,
assume that @xmath12 has a multiple root , @xmath140 say .
then @xmath140 is a root of @xmath141 since @xmath29 is not identically @xmath30 , neither is @xmath142 , and so @xmath140 is algebraic over @xmath0 , and then so is @xmath143 , a contradiction .
note that if @xmath144 , then @xmath12 is not separable over @xmath9 , for in that case every root of @xmath12 is vacuously a root of @xmath145 and is therefore a multiple root .
next , observe that @xmath146 = k[\xi_n]$ ] since @xmath143 .
since @xmath146=k[\xi_n]$ ] is finite as a @xmath56$]-module , it can not be a field ; so @xmath147 $ ] is isomorphic to @xmath31 $ ] .
since @xmath31 $ ] is normal , the same holds for @xmath147 $ ] . here
we wish to clarify the nature of the specialization maps @xmath148 associated with specializing @xmath1 to @xmath58 as well as the relationship between the iterated monodromy group @xmath79 and its specializations @xmath149 .
we do so by defining a notion of _ global specialization_. let @xmath150 be the integral closure of @xmath56 $ ] in @xmath10 .
by integrality ( and the going up theorem ) , the maximal ideal @xmath151 of @xmath56 $ ] extends to a prime ideal @xmath152 of @xmath150 such that @xmath153 = ( t - t_0)$ ] .
the ring @xmath154 is integral over @xmath0 , so is actually a field .
thus @xmath152 is maximal , and @xmath154 is algebraic over @xmath0 .
so there are embeddings @xmath155 .
fix one , and consider the associated map @xmath156 with kernel @xmath152 .
we call such a map a _ global specialization _ associated with @xmath67 .
the image of the global specialization , which is a field @xmath157 , is independent of the choice of global specialization @xmath158 .
now we can define the specializations @xmath159 and @xmath160 by restriction of the global specialization .
the field @xmath161 can be defined as the image of @xmath162 , and can be shown to be independent of the choice of global specialization ( associated with @xmath67 ) .
however , @xmath163 , the image of @xmath164 , depends on the global specialization @xmath158 as well as on the choice of @xmath140 . in this optic , the relationship between the groups @xmath165 and the group @xmath166 is elucidated as follows .
let @xmath167 be the _ decomposition group _ associated to @xmath152 ( consisting of the elements of @xmath168 fixing the chosen maximal ideal @xmath152 of @xmath150 ) .
then @xmath167 acts on @xmath169 , and therefore on @xmath157 .
thus we get a homomorphism @xmath170 . as usual , this is a surjection , and if @xmath67 is not in the post - critical set , then it is actually an isomorphism .
thus , for @xmath171 , @xmath172 is isomorphic to a subgroup @xmath167 of @xmath173 , hence it too has an action on the rooted tree @xmath44 .
two self - maps @xmath174 of @xmath37 defined over @xmath0 ( i.e. @xmath175 ) , are equivalent over @xmath0 ( or @xmath0-conjugate ) if there exists an automorphism @xmath176 of @xmath37 ( defined over @xmath0 ) such that the diagram @xmath177{e , t}{\gamma } \arrow{s , l}{\ph } \node[2]{\pp^1 } \arrow{s , r}{\psi } \\ \node{\pp^1 } \arrow[2]{e , t}{\gamma } \node[2]{\pp^1 } \end{diagram } \label{fgisom}\ ] ] commutes . in other words ,
@xmath8 and @xmath178 are equivalent over @xmath0 if and only if there exist @xmath179 satisfying @xmath180 such that @xmath181 where @xmath182 .
the equivalence ( or conjugacy ) class of @xmath8 , denoted @xmath183 $ ] , is a _ dynamical system _ on @xmath37 .
for @xmath184 , we say @xmath183 $ ] is _ arithmetic _ if there exists @xmath185 with @xmath183=[\psi]$ ] .
note that if @xmath27 $ ] is a polynomial map , the images of @xmath8 under _ affine _ transformations @xmath186 over @xmath0 form exactly the set of polynomial maps @xmath0-isomorphic to @xmath8 .
also , if @xmath176 takes @xmath33 to @xmath187 for @xmath188 , then it does so for all @xmath13 .
thus , the study of iterations of @xmath8 and @xmath178 coincide ( they simply take place in different coordinates ) precisely when @xmath8 and @xmath178 are conjugate . in particular , if @xmath183=[\psi]$ ] , then the iterated extensions @xmath10 and @xmath189 are isomorphic .
for a more detailed discussion , including the relationship between fields of moduli and fields of definition of dynamical systems on @xmath37 , we refer the reader to silverman @xcite .
when discussing the coefficients of a post - critically finite polynomial , it is often convenient to normalize by working with _
post - critically finite polynomials .
[ lemma1 ] every polynomial @xmath27 $ ] of degree @xmath103 is equivalent over some finite extension @xmath190 to a monic polynomial in @xmath191 $ ] .
furthermore , if @xmath178 and @xmath8 are two @xmath192-equivalent monic polynomials for some finite extension @xmath190 , then @xmath193 where @xmath194 is in @xmath192 and @xmath195 is a @xmath196st root of unity .
suppose @xmath197 is the leading term of @xmath8 .
if @xmath198 , then @xmath199
. so @xmath200 has leading term @xmath201 .
when we let @xmath202 be a root of @xmath203 , we find that @xmath204 is monic .
now let @xmath8 and @xmath178 be monic equivalent polynomials in @xmath191 $ ] . if @xmath205 , then @xmath206 has leading term @xmath207 .
thus if @xmath208 , then @xmath202 must be a @xmath196th root of unity .
post - critically finite polynomials can be classified in terms of certain combinatorial objects called _ hubbard trees _ , see @xcite and @xcite , as well as @xcite for their relationship , in the case of two critical values , to _ dessins denfant _ of genus 0 . instead of describing this classification
, in this subsection , we simply want to illustrate that post - critically finite polynomials are in plentiful supply by describing some of the most simplest families of examples . in order to avoid rationality questions , in this subsection
we assume that @xmath209 is algebraically closed . to write down examples
, we can make various simplifying assumptions ; for example we can limit the number of critical points ( or values ) .
if @xmath8 has only one critical point and this point is fixed , we see quickly that @xmath8 is conjugate to @xmath210 ; specializations of this map constitute the classical theory of `` pure '' extensions .
another family of examples is given by the _ chebyshev polynomials _ which have only two critical values ; we study the quadratic one @xmath115 in [ sec : quad ] .
more generally , polynomials with two critical values are called _ generalized chebyshev polynomials _ or more commonly _
shabat polynomials _ ; they have quite a rich structure , as can be seen from the readable survey of shabat - zvonkin @xcite . here
we make a different set of simplifying assumptions , and completely classify the resulting post - critically finite dynamical systems for each degree @xmath103 .
namely , we assume that the critical points are fixed and that all the ramification indices are two ; the latter condition is equivalent to requiring that the polynomial has @xmath211 critical points .
other than @xmath102 , this is the simplest family of post - critically finite polynomials .
it gives simple examples of post - critically finite polynomials not equivalent to any monic polynomial with _
_ coefficients .
a polynomial @xmath27 $ ] of degree @xmath103 is said to be _ critically fixed , simply ramified _
( cfsr ) if @xmath8 has @xmath211 critical points , each of which is a fixed point for @xmath8 . if @xmath0 does not have characteristic @xmath114 , the polynomial @xmath212 has exactly one critical point , @xmath213 , which is a fixed point
thus @xmath8 is a cfsr polynomial .
it is easy to see that @xmath8 is the unique such polynomial , up to equivalence , of degree @xmath114 .
let @xmath214 .
the polynomial @xmath215 has derivative @xmath216 .
thus @xmath8 has two critical points @xmath217 .
the fixed points of @xmath8 are @xmath30 and the two critical points , so @xmath8 is a cfsr polynomial .
this polynomial @xmath8 gives an example of a monic , post - critically finite polynomial which does not have integral coefficients .
is there a monic polynomial @xmath178 equivalent to @xmath8 with integer coefficients ? by lemma [ lemma1 ] we only need to consider polynomials of the form @xmath218 in the first case , @xmath219 let @xmath47 be a place ( valuation ) in @xmath220 above @xmath114 normalized so that @xmath221 .
we want to find @xmath194 so that the coefficients are integral .
so , @xmath222 . this implies @xmath223 .
thus the coefficient of @xmath224 is not @xmath114-integral .
a similar argument applies to the second case .
we conclude that there are no monic polynomials with integral coefficients equivalent to @xmath8 .
this gives an example of a post - critically finite polynomial not equivalent to any monic polynomial with integral coefficients .
we will now assume that @xmath0 is algebraically closed .
thus , up to equivalence , cfsr polynomials can be taken to be monic . in an effort to normalize further , consider the roots of the fixed point polynomial @xmath225 .
these include all @xmath211 critical points ( roots of @xmath29 ) , but the polynomial is of degree @xmath45 .
thus there is a @xmath45th root @xmath226 ; here , we allow @xmath226 to be one of the @xmath211 critical points if @xmath227 has a double root .
after conjugating by a translation @xmath176 , we can assume that @xmath228 .
in particular , @xmath229 .
this motivates the following .
a _ normalized cfsr polynomial _
$ ] is a monic polynomial with @xmath231 .
the above argument gives the following .
if @xmath0 is algebraically closed , then every cfsr polynomial is equivalent to a normalized cfsr polynomial .
next , we will show that over an algebraically closed field , there is , up to equivalence , a unique cfsr polynomial of each degree .
assume @xmath230 $ ] is a normalized cfsr polynomial of degree @xmath45 .
we rewrite @xmath231 as @xmath232 by differentiating this equation we get @xmath233 , so @xmath234 differentiating the first of these gives @xmath235 .
so if @xmath236 , @xmath237 .
thus @xmath238 continuing in this manner , we get that the @xmath7th derivative @xmath239 is @xmath240 if @xmath241 .
so , for @xmath241 .
@xmath242 in particular , if @xmath243 then @xmath244 this gives uniqueness .
existence follows from the fact that @xmath245 satisfies the equation @xmath246 so is a normalized cfsr polynomial .
[ prop : cfsr ] the polynomial @xmath247 is the unique normalized cfsr polynomial of degree @xmath45 . is it true that all post - critically finite polynomials over the complex numbers are equivalent to a monic polynomial with algebraic coefficients ?
the answer to this question is known to be positive not just for cfsr polynomials ( by proposition [ prop : cfsr ] ) but for all critically fixed polynomials , by a theorem of tischler @xcite ; see pakovich @xcite for more on critically fixed polynomials .
in this section , we prove theorem [ thm:1 ] . the main tool is a polynomial version of the riemann - hurwitz genus formula .
recall that the resultant of two polynomials @xmath248 and @xmath249 in @xmath31 $ ] satisfies @xmath250 where the product is over the roots @xmath251 of @xmath249 ( counted with multiplicity ) and @xmath252 is the coefficient of the highest power of @xmath253 appearing in @xmath249 .
( simon)[lem : simon ] suppose @xmath254 where @xmath255 $ ] are polynomials with resultant @xmath256 .
if @xmath257 , then @xmath258 a proof can be found in the thesis of d. simon @xcite , see proposition iv.2.2 as well as remarque on p. 92 .
now we take a degree @xmath45 polynomial @xmath259 $ ] with leading coefficient @xmath260 , and put @xmath261 where @xmath139 is defined by ( [ pn ] ) .
[ prop : recur ] suppose @xmath8 is a degree @xmath45 polynomial in @xmath31 $ ] with leading coefficient @xmath260 , and put @xmath262 . then , for @xmath13 , the discriminant @xmath263 $ ] satisfies the recurrences @xmath264 where @xmath265 are the multiplicities defined by ( [ mr ] ) and ( [ mbeta ] ) .
we apply lemma [ lem : simon ] with @xmath266 , @xmath267 ; note that @xmath268 .
we have @xmath269 thus , by lemma [ lem : simon ] , @xmath270 where the factor @xmath271 occurs because @xmath272 is the leading coefficient of @xmath273 when considered as a polynomial in @xmath253 .
we have @xmath274 this completes the proof .
we now have the tools for proving theorem [ thm:1 ] , but we first make a convenient definition . [ def : s ] for a post - critically finite polynomial @xmath275 $ ] of degree @xmath45 with leading coefficient @xmath260 , and @xmath39 , we let @xmath276 be the set of real infinite places of @xmath0 together with those finite ones which do not vanish on at least one of the following : @xmath45 , @xmath252 , and @xmath277 as @xmath278 runs over @xmath279 . after a simple change of variables , we may assume that @xmath275 $ ] .
by lemma [ lem : integral ] and proposition [ prop : recur ] , the primes that ramify in @xmath280 are precisely those corresponding to the postcritical set @xmath281 , i.e. @xmath282 . thus @xmath283 is finitely ramified if and only if @xmath8 is post - critically finite . now suppose @xmath8 is post - critically finite and fix @xmath39 . by proposition [ prop :
recur ] , for every @xmath13 , a place of @xmath284 which does not lie over @xmath276 is unramified in @xmath285 .
hence @xmath82 is finitely ramified and consequently , so is @xmath286 .
[ cor : imr ] let @xmath0 be a number field , @xmath275 $ ] a post - critically finite polynomial of degree @xmath103 . for @xmath287
, the action of @xmath288 on @xmath157 induces an iterated monodromy representation @xmath289 , where @xmath99 .
our aim here is to prove theorem [ thm : wild1 ] . throughout this section ,
we suppose @xmath0 is a characteristic 0 field equipped with an ultrametric valuation @xmath47 having valuation ring @xmath290 ; we assume that @xmath291 .
the residue field of @xmath0 with respect to @xmath47 , i.e. the reduction of @xmath292 modulo its maximal ideal @xmath293 , is denoted @xmath294 .
we assume that @xmath294 has positive characteristic @xmath295 .
a polynomial @xmath296 $ ] has _ good reduction _ at @xmath47 if @xmath297 in other words , @xmath8 has good reduction when it is @xmath47-integral with @xmath47-unital leading coefficient .
[ lem : w1 ] suppose @xmath192 is an algebraic extension of @xmath0 and fix an extension @xmath298 of @xmath47 to @xmath192 .
let @xmath27 $ ] be a polynomial of degree @xmath4 with good reduction at @xmath47 .
if @xmath299 has @xmath300 , then @xmath301 is not preperiodic for @xmath8 .
suppose @xmath302 has @xmath303 . since the leading coefficient of @xmath8 is a @xmath47-adic unit , there is a unique term in the sum @xmath304 with minimal valuation , namely @xmath305 .
since @xmath298 is ultrametric , we have @xmath306 . applying this principle to @xmath307
, we obtain @xmath308 .
thus , the set @xmath309 can not be finite since @xmath310 is not finite .
[ lemma : zero ] suppose @xmath27 $ ] is a polynomial of degree @xmath45 divisible by @xmath17 ( the residue characteristic of @xmath47 ) , has good reduction at @xmath47 , and is post - critically finite
. then the image of @xmath29 in @xmath311 $ ] is identically @xmath30 .
let @xmath192 be a splitting field for @xmath312 over @xmath0 , @xmath298 an extension of @xmath47 to @xmath192 with valuation ring @xmath313 and residue field @xmath314 .
by lemma [ lem : w1 ] , @xmath315 for every critical point @xmath316 .
we have @xmath317 .
since the critical points @xmath316 are @xmath298-integral , @xmath318 with @xmath319 $ ] . thus , @xmath312 is identically zero in @xmath320 $ ] and hence also in @xmath311 $ ] . for
a polynomial @xmath321 $ ] , let us write @xmath322 if @xmath323 is in @xmath324 $ ] but @xmath325 is not .
by lemma [ lemma : zero ] , @xmath326 . using the product rule for differentiation and induction on @xmath7
, we have @xmath327 .
we have @xmath328 since @xmath329 , @xmath330 divides @xmath331 .
this completes the proof . via the example @xmath111
, we will see in [ sec : quad ] that the bound given theorem [ thm : wild1 ] is sometimes met
. more generally , suppose @xmath104 , @xmath8 has degree @xmath332 and @xmath333 is such that for all @xmath32 , @xmath334 is galois of degree @xmath17 .
we also assume that @xmath17 is totally ramified in @xmath335 for all @xmath7 .
these criteria are met , for example , for the chebyshev polynomial of degree @xmath17 and @xmath336 ( giving the cyclotomic @xmath337-extension of @xmath338 ) . then @xmath339 . here
is how to see this . in @xmath340
, one knows that @xmath341 , the higher ramification group of ( lower ) index @xmath342 , is trivial once @xmath343 .
hence , @xmath344 the left hand side of the above inequality is equal to the valuation at the prime above @xmath17 of the different of @xmath340 .
one concludes by using the discriminant formula in a tower of extensions .
it is not difficult to write down polynomials @xmath345 $ ] such that there is no wild ramification in the iterated tower of @xmath8 . according to theorem [ thm : wild1 ] ,
if such a polynomial is monic , it will not be post - critically finite , so the resulting iterated tower of function fields will be infinitely ramified . here is a quadratic example .
let @xmath346 with @xmath347 .
then @xmath348 is _ odd _ for all @xmath349 ( for instance by proposition [ prop : recur ] ) .
however , @xmath8 is not post - cricially finite .
indeed , its only critical point is @xmath350 .
for @xmath351 the @xmath114-adic valuation of @xmath352 , @xmath353 is negative , hence by lemma [ lem : w1 ] , @xmath8 is not post - critically finite .
in this section , @xmath0 is a number field .
we now describe , in terms of certain graphs , how primes of @xmath0 not dividing the discriminant of @xmath62 ( assumed to be irreducible ) decompose when we adjoin a root of this polynomial .
a simple consequence of this description is that no finite prime of @xmath0 splits completely in @xmath82 .
we first set up some notation .
we assume @xmath275 $ ] is post - critically finite .
recall the notation from [ sec : intro ] regarding @xmath354 .
fixing @xmath106 , let us assume that @xmath62 is irreducible over @xmath0 for all @xmath13 and choose a coherent system @xmath109 of their roots so that @xmath355 .
for the rest of this section , we assume @xmath356 is a prime of @xmath357 which is not in @xmath276 ( see definition [ def : s ] ) . for such @xmath356 ,
the splitting of @xmath356 in the ring of integers of @xmath63 coincides with the splitting of @xmath356 in the ring @xmath108 $ ] ; the latter factorization mirrors exactly the factorization of the polynomial @xmath62 over the residue field @xmath358 .
for example , the primes of degree 1 in @xmath108 $ ] which lie over @xmath356 correspond to the roots of @xmath359 over @xmath360 , i.e. the points in @xmath360 whose image under the @xmath7th iterate of @xmath8 is the image @xmath361 of @xmath67 in @xmath360 .
a prime of degree @xmath362 lying over @xmath356 corresponds to a galois orbit of @xmath362 points defined over a degree @xmath362 extension of @xmath360 mapping to @xmath361 by @xmath33 .
such data is conveniently summarized in terms of certain directed graphs we now define . for @xmath363 ,
let @xmath364 be a degree @xmath362 extension of the residue field @xmath360 .
we denote by @xmath365 the following directed graph : the vertices are the elements of @xmath364 and the graph has a directed edge @xmath366 if and only if @xmath49 .
after we choose an ordering @xmath367 of the elements of @xmath364 , the adjacency matrix @xmath368 of @xmath365 has @xmath369 entry @xmath370 if @xmath371 and @xmath30 otherwise .
we write @xmath372 , @xmath373 for @xmath374 and @xmath375 . for calculations , it is useful to note that @xmath376 . in other words ,
the in - degree of a vertex @xmath47 in @xmath377 is the number of length @xmath7 paths on @xmath365 ending at @xmath47 .
for example , let @xmath378 be the vertex corresponding to the reduction of @xmath67 modulo @xmath356 . then the following quantities all coincide : * the number of degree 1 primes of @xmath108 $ ] over @xmath356 , * the in - degree of @xmath361 on @xmath379 , * the sum of the entries in the column of @xmath380 corresponding to @xmath361 , * the number of length @xmath7 paths on @xmath372 ending at @xmath361 .
note that , by ( c ) for example , there are at most @xmath381 degree @xmath370 primes of @xmath63 lying over @xmath356 , hence @xmath356 does not split completely in @xmath82 .
more generally , we can count the number of primes of any given degree over @xmath356 by taking into account the action of @xmath382 .
namely , the graph @xmath365 has the following additional structure : each vertex is `` colored '' , we will say _ weighted _ , by a positive divisor @xmath383 of @xmath362 where @xmath383 is the exact degree of that vertex over @xmath360 .
furthermore , every directed edge has the property that the weight of the initial vertex is a multiple of the weight of the terminal vertex .
also @xmath384 acts on the graph and the weight of a vertex equals the size of its orbit under this action . summarizing the discussion
, we have the following proposition describing prime decomposition in @xmath69 in terms of graphs .
[ prop : splitting ] suppose @xmath385 $ ] is post - critically finite and that @xmath386 is such that @xmath62 is irreducible over @xmath0 for all @xmath13 .
suppose @xmath387 is not in @xmath276 .
then , for @xmath363 , the number of degree @xmath362 primes of @xmath63 lying over @xmath356 is @xmath388 where @xmath389 is the number of paths of length @xmath362 on @xmath365 which start with a vertex of weight @xmath362 and end at @xmath361 , the weight 1 vertex corresponding to the image of @xmath67 in @xmath360 .
alternatively , one could take the quotient graph of @xmath365 by identifying vertices which are in the same orbit of @xmath384 , and give a vertex in the new graph the weight equal to the number of points identified .
then the degree @xmath362 primes of @xmath63 lying over @xmath356 are in bijective correspondence with the paths of length @xmath7 on the quotient graph starting with a vertex of weight @xmath362 and ending at @xmath361 .
we should note that as long as @xmath390 , the decomposition of @xmath356 in @xmath63 depends only on the residue of @xmath67 modulo @xmath356 . for a fixed pair @xmath391 and @xmath7 tending to infinity
, each graph @xmath392 has @xmath393 vertices and an equal number of edges , hence is one of a finite number of graphs . therefore , the sequence @xmath392 , @xmath394 is always eventually periodic . in fact
, it is relatively simple to describe exactly what happens to the sequence of graphs in our situation .
each connected component of @xmath365 consists of a unique cycle or `` loop '' with a number of `` arms '' emanating from it .
the minimal period of the sequence @xmath395 is the lowest common multiple of the length of the unique loop in each connected component of @xmath365 and the preperiod is the least common multiple of the length of the longest arm in each connected component of @xmath396 .
all of these facts are easily verified and left as amusing exercises for the reader .
a highly interesting question is whether one can capture the graph - theoretical description of prime decomposition in iterated extensions via appropriate zeta and @xmath397-functions . here , we settle for a typical example as an illustration .
[ ex : tame ] let @xmath398 $ ] .
let @xmath399 be a prime of norm 13 .
we map @xmath400 \rightarrow \f_\p \simeq \f_{13}$ ] by sending @xmath401 , and list the elements of @xmath402 as @xmath403 .
we write down the adjacency matrix @xmath373 and draw the graph for @xmath8 and @xmath404 .
@xmath405 the graph @xmath372 : @xmath406 & 8\ar@/^/[r ] & 7 \ar@/^/[d ] & \\ & & 5 \ar@/^/[u ] & \ar@/^/[l]6 } \qquad \qquad \xymatrix{3 \ar@/^/[r ] & 4\ar@/^/[r ] & 11 \ar@/_/[rd]&&\ar@/_0.5cm/[ll ] 9&\ar@/_/[l ] 1\\ 10 \ar@/_/[ur]&&&12 \ar@/_/[ur]&\ar@/^/[l]2 & } \ ] ] the graph @xmath407 : @xmath408 & 7\ar@(ur , dr ) [ ] \\ & 6 \ar@/^/[u ] } \qquad \qquad \xymatrix{8\ar@/^/[r ] & 5\ar@(ur , dr ) [ ] } \qquad \qquad \xymatrix{2\ar@/^/[r ] & 9 \ar@/^0.5cm/[rr]&&12 \ar@/^/[ld]&\ar@/_/[l ] 4\\ & & 11 \ar@/^/[lu ] & & \\ & 3 \ar@/^/[ru]&1\ar[u]&10 \ar@/_/[lu ] & } \ ] ] note that @xmath372 has two connected components , one with a loop of length 2 and the other with a loop of length 3 .
the longest arm in each component has length 2 .
the reader can check either by taking powers of the adjacency matrix or by drawing the graphs that @xmath372 occurs only once in the sequence @xmath379 , but starting with @xmath409 , the sequence has period @xmath410 . note that @xmath410 is the product of the lengths of the loops in the connected components of @xmath411 . with base field @xmath412
, the number of degree 1 places in @xmath137 over the prime @xmath413 for @xmath414 is the periodic sequence @xmath415 of period @xmath416 . as a check on the calculations
, we verified using gp - pari that with @xmath417 , the polynomials @xmath418 for @xmath419 , factor over @xmath402 into distinct irreducible factors of the following degrees : @xmath420
in this section , we make a few remarks and give some examples concerning quadratic polynomials . by applying automorphisms of @xmath37
, we bring each quadratic polynomial to a standard form @xmath421 .
we then write down recurrence conditions for post - criticality of @xmath8 .
the minimal number fields over which preperiodic points of prescribed preperiod @xmath383 and period @xmath7 for such quadratic polynomials are defined form an interesting family of number fields in their own right . put @xmath422.\ ] ] let @xmath423 .
it is the unique branch point for the cover of @xmath37 given by the polynomial @xmath424 , i.e. @xmath425 . theorem [ thm:1 ] now simplifies as follows : @xmath15 is finitely ramified if and only if @xmath426 is preperiodic for @xmath8 .
if @xmath427 is quadratic , we take @xmath428 , so that @xmath429 .
we then have @xmath430 is _ monic_. note that @xmath176 fixes @xmath30 .
since an isomorphism from @xmath8 to @xmath178 carries @xmath92 to @xmath431 , applying a @xmath0-automorphism taking @xmath426 to @xmath30 , we see that @xmath178 is conjugate to @xmath8 where @xmath432 we leave to the reader the exercise that for each quadratic @xmath433 $ ] , there is a unique @xmath121 such that @xmath178 is conjugate to @xmath123 .
note that via the automorphism @xmath434 , the maps @xmath435 and @xmath123 are @xmath0-isomorphic . now consider a normalized quadratic polynomial @xmath436 .
we have @xmath437 for @xmath32 , consider the recurrence @xmath438 with initial condition @xmath439 .
then @xmath8 is post - critically finite if and only if @xmath226 is a root of @xmath440 for some @xmath441 . *
exercises . *
i ) if @xmath442 and @xmath443 has periodic branch points , then @xmath444 .
\ii ) if @xmath445 $ ] has preperiodic branch point , then @xmath446 is an algebraic integer . in this subsection
, we turn to an example which was the starting point of this article .
while reading an article of lemmermeyer , we came across the classical fact that the cyclotomic @xmath447-extension of @xmath338 can be written as @xmath448 where @xmath449 . indeed , using the half - angle formula for cosines , one establishes easily that the nested square root expression given above evaluates @xmath450 .
what attracted our attention here was that in the resulting recurrence - tower , the number of ramified primes is finite ( indeed only @xmath114 ramifies , and it does so totally and deeply ) . since the @xmath451 are roots of the @xmath7th iterate equation @xmath452 , where @xmath111 , it was natural to wonder whether for every @xmath453 , @xmath454 cuts out a finitely ramified tower .
that this is so is guaranteed by theorem [ thm:1 ] since @xmath115 is post - critically finite .
indeed , it is the first member of the chebyshev family of post - critically finite polynomials . for more details
see proposition 5.6 of @xcite where the iterated monodromy group of any chebyshev polynomial of degree @xmath103 is shown to be infinite dihedral . for the rest of this subsection , let @xmath111 . here
we will verify that another property of the cyclotomic @xmath447-tower ( the specialization of the tower at @xmath455 ) holds for many values of @xmath456 , namely that a root of @xmath62 generates over @xmath352 the ring of integers of the number field it cuts out .
[ lem : eisen ] for @xmath457 , @xmath458 , the polynomial @xmath359 is irreducible over @xmath338 .
we note that @xmath459 , @xmath460 . if @xmath461 , we apply the einsenstein criterion to @xmath462 at the prime @xmath114 .
if @xmath463 , we use @xmath464 instead . if @xmath457 is congruent to @xmath465 modulo @xmath466 , and if @xmath467 and @xmath468 are square - free , then for @xmath13 , the stem field @xmath469/(\ph_n(x , t_0))$ ] of the polynomial @xmath470 is monogne , as @xmath471 . letting @xmath472 , by proposition [ prop :
recur ] , we have for @xmath13 , @xmath473 or @xmath474 since @xmath475 .
also , for @xmath188 , we have : @xmath476 .
we need to compare @xmath477 with the discriminant @xmath478 of the ring of integers of @xmath479 . for @xmath188 , we clearly have @xmath480 , since @xmath467 is square - free . for @xmath13 ,
let us now determine the ramification for each extension @xmath481 .
we first remark that @xmath482 , with @xmath483 .
next we observe that @xmath484 .
hence , for @xmath13 , in the extension @xmath481 , only the places dividing @xmath485 are allowed to ramify .
we first examine the tame ramification .
suppose @xmath486 is a prime divisor of @xmath487 .
then @xmath488 and so @xmath486 is split in @xmath489 .
let @xmath486 be an odd prime divisor of @xmath468 .
since @xmath490 , there exists a prime @xmath491 of @xmath479 lying over @xmath486 which is ramified in @xmath481 .
in fact , there are two primes over @xmath486 in @xmath284
. one of them is totally ramified in @xmath492 .
the other is unramified .
therefore , the valuation @xmath493 at the prime ideal @xmath491 of the different of the extension @xmath481 is precisely @xmath494 .
it remains to study the wild ramification . for @xmath13 ,
let us put @xmath495 we note that @xmath114 is ramified in @xmath489 and that @xmath496 is a uniformizer for the unique place @xmath497 of @xmath284 lying over @xmath114 .
we will proceed by induction .
suppose , for some @xmath13 , that @xmath114 is totally ramified in @xmath498 and that @xmath499 is a uniformizer of the unique place @xmath500 of @xmath479 lying over @xmath114 .
we claim that @xmath501 is not a square modulo @xmath502 . to see this ,
let us suppose that @xmath501 is a square modulo @xmath503 . since the residue field is @xmath504 , we get , in the case @xmath505 , @xmath506 with @xmath507 , which is impossible . thus , for @xmath463 , kummer theory tells us that @xmath481 is ramified at the unique place above @xmath114 . for @xmath461 ,
the argument is simpler , since , in that case , the valution of @xmath508 at @xmath499 is the same as that of @xmath451 , namely @xmath370 . by kummer theory ,
@xmath481 is ramified at the unique place above @xmath114 . in conclusion , @xmath509 is totally ramified at @xmath114 .
if @xmath510 , it is clear that @xmath511 is a uniformizer of the unique place of @xmath512 lying over @xmath114 .
the same holds for @xmath513 when @xmath463 ; note that @xmath514 .
this completes the induction step .
next , let us calculate conductors .
let @xmath515 be a generator of the galois group @xmath516 . then @xmath517 the valuation at @xmath518 of @xmath114 is @xmath519 .
hence , the element @xmath515 belongs to @xmath520 , but not to @xmath521 ( we are using the higher ramification groups in the lower numbering ) .
consequently , @xmath522 where @xmath523 is the valuation at @xmath518 and @xmath524 is the different of the extension @xmath481 .
now we are able to determine the discriminant of @xmath498 .
we have the recurrence formula @xmath525 which coincides up to sign with the recurrence ( [ recurd ] ) for @xmath477 .
we also have the coincidence of initial conditions , @xmath526 . since @xmath527 is a square ,
we conclude that @xmath480 for all @xmath7 , and so @xmath528 $ ] .
in this section , we discuss in a bit more detail conjectural and known properties of iterated monodromy representations , especially as compared with those of @xmath17-adic representations .
we also list a number of open problems .
let us first recall a conjecture of fontaine and mazur : if @xmath0 is a number field and @xmath18 is a finite set of places of @xmath0 none of which has residue characteristic @xmath17 , then all finite - dimensional @xmath17-adic representations of @xmath25 factor through a finite quotient ( see conj .
5a of @xcite as well as kisin - wortmann @xcite ) .
on the other hand , infinite tamely and finitely ramified extensions of number fields do exist ( and are in plentiful supply ) thanks to the criterion of golod and shafarevich , see e.g. roquette @xcite .
thus , at least for certain pairs @xmath529 , there is a sizeable portion of @xmath25 which is predicted to be invisible to finite - dimensional @xmath17-adic representations .
when @xmath18 contains all places above @xmath17 , it is also expected , by a conjecture of boston @xcite ( which we recall below ) , that @xmath17-adic representations do not capture all of @xmath25 .
suppose @xmath530 is a residual representation of @xmath25 . by mazur s theory of deformations
, there exists a universal ring @xmath531 ( local , noetherian and complete ) and a versal deformation @xmath532 such that @xmath533 is the restriction of @xmath534 .
let @xmath535 be the subfield of @xmath19 fixed by @xmath536 .
we put @xmath537 where @xmath538 is the maximal pro-@xmath17 extension of @xmath397 inside @xmath19 .
if @xmath18 contains all place above @xmath17 ( @xmath17 odd , or for @xmath17 even we assume @xmath0 is totally complex ) , then the cohomological dimension of @xmath539 is at most @xmath114 .
the purely group - theoretical conjecture b of boston @xcite concerning the rank - growth of subgroups of @xmath540 , then implies the _ non - injectivity conjecture _
( @xcite , p. 91 ) to the effect that @xmath534 is never injective . in a certain sense
, one expects that @xmath534 forgets a non - trivial part of @xmath539 .
how can one shed light on those sides of arithmetic fundamental groups which are apparently not illuminated by the theory @xmath17-adic representations ? as a counterpoint to the fontaine - mazur conjecture , a conjecture of boston
@xcite asserts that infinite tame quotients of @xmath541 possess faithful actions on rooted trees .
iterated monodromy groups are canonically equipped with such an action @xcite .
it is therefore natural to seek such representations via specializations of iterated towers of post - critically finite polynomials , in the wild case as well as in the tame case . in the wild case , it would be interesting to produce iterated monodromy representations whose image does not have any infinite @xmath17-adic analytic quotients . since very little is known about the structure of infinite tamely and finitely ramified extensions of number fields ,
the following question is of particular interest . 1 . for each @xmath13 , @xmath62 is irreducible over @xmath0 , ( i.e. @xmath544 is a field of degree @xmath40 over @xmath0 ) , 2 .
there is a finite set @xmath18 of places of @xmath0 such that @xmath545 is unramified outside @xmath18 for all @xmath13 , and such that 3 .
@xmath18 does not contain any primes dividing @xmath45 ? by theorem [ thm:1 ]
, it is possible to fulfill ii ) by taking @xmath8 to be a post - critically finite polynomial .
satisfying i ) is not too difficult either , since we can arrange a place of @xmath0 to ramify totally in @xmath479 ( essentially an eisenstein condition , see for example lemma [ lem : eisen ] ) .
condition iii ) asks that @xmath546 be _
tamely ramified_. it is not difficult to arrange i ) and iii ) simultaneously by imposing congruence conditions ( e.g. see example [ ex : tame ] ) .
however , satisfying all conditions together appears to be quite difficult .
a positive answer to question [ ques:1 ] would provide , for the first time , an explicit method for constructing an infinite tamely and finitely ramified extension of a number field .
currently the only method for producing such extensions is via the golod - shafarevich criterion . on the other hand , a negative answer would assert that an analogue of the fontaine - mazur conjecture holds for finitely ramified iterated extensions .
we should mention that for the function field of a curve over a finite field with a square number of elements , recursive constructions of garcia - stichenoth ( see @xcite for example ) for tamely and finitely ramified extensions exist ; that such constructions always arise from modular curves is a conjecture of elkies @xcite . recall that an algebraic extension @xmath397 over a number field @xmath0 is called _ asymptotically good _
if i ) @xmath547 is infinite , and ii ) for every sequence of distinct intermediate subfields of @xmath547 , the root discriminant over @xmath338 is the @xmath7th root of the absolute value of its discriminant ] remains bounded .
a more general and more concise version of question [ ques:1 ] is the following .
under the assumption of good reduction of the polynomial @xmath8 , the analogue of this question where we replace the number field discriminant with the polynomial discriminant , has a negative answer by theorem [ thm : wild1 ] .
namely , for a polynomial @xmath548 $ ] of degree @xmath28 , define its root discriminant by @xmath549 . an immediate consequence of theorem [ thm : wild1 ] is [ cor : wild1 ] if @xmath550 $ ] is post - critically finite , has degree divisible by @xmath17 , and has good reduction at @xmath17 , then for any @xmath457 , the sequence of polynomials @xmath551 is asymptotically bad in the sense that ( the @xmath17-part of ) @xmath552 tends to infinity with @xmath7 .
this result is in agreement with a conjecture of simon @xcite , to the effect that any infinite sequence of distinct polynomials over @xmath352 is asymptotically bad .
thus , to tackle questions [ ques:1 ] and [ ques:2 ] , one would very likely have to understand the index of the order @xmath108 $ ] in @xmath553 .
h. bass , m. otero - espinar , d. rockmore , and c. tresser , cyclic renormalization and automorphism groups of rooted trees .
lecture notes in mathematics , 1621 .
springer - verlag , berlin , 1996 .
mr1392694 ( 97k:58058 ) n. elkies , explicit modular towers , pages 23 - 32 in proceedings of the thirty - fifth annual allerton conference on communication , control and computing ( 1997 , t. basar , a. vardy , eds . ) , univ .
of illinois at urbana - champaign 1998 ( ` arxiv : math.nt/0103107 ` ) j .- m .
fontaine and b. mazur , geometric galois representations .
elliptic curves , modular forms , & fermat s last theorem ( hong kong , 1993 ) , 4178 , ser .
number theory , i , internat .
press , cambridge , ma , 1995 | let @xmath0 be a number field , @xmath1 a parameter , @xmath2 , and @xmath3 $ ] a polynomial of degree @xmath4 .
the polynomial @xmath5 $ ] , where @xmath6 is the @xmath7-fold iterate of @xmath8 , is absolutely irreducible over @xmath9 ; we compute a recursion for its discriminant .
let @xmath10 be the field obtained by adjoining to @xmath9 all roots ( in a fixed @xmath11 ) of @xmath12 for all @xmath13 ; its galois group @xmath14 is the iterated monodromy group of @xmath8 . the iterated extension @xmath10 is finitely ramified over @xmath9 if and only if @xmath8 is post - critically finite .
we show that , moreover , for post - critically finite @xmath8 , every specialization of @xmath15 at @xmath16 is finitely ramified over @xmath0 , pointing to the possibility of studying galois groups of number fields with restricted ramification via tree representations associated to iterated monodromy groups of post - critically finite polynomials .
we discuss the wildness of ramification in some of these representations , describe prime decomposition in terms of certain finite graphs , and also give some examples of monogne number fields that arise from the construction
. + august 12 , 2004 |
emission lines arising from transitions in fex have been widely detected in solar extreme - ultraviolet ( euv ) spectra ( see , for example , dere 1978 ; thomas & neupert 1994 ) . jordan ( 1965 ) first proposed the use of fex lines to determine the electron density in the solar corona , while jordan ( 1966 ) employed euv transitions of fex and other fe ions to derive both the electron density and coronal fe abundance .
since then , several authors have undertaken analyses of the solar euv spectrum of fex , including for example nussbaumer ( 1976 ) and bhatia & doschek ( 1995 ) . to date , the most complete study is probably that of del zanna , berrington & mason ( 2004 ) , which also provides an excellent review of previous work on fex .
the theoretical line ratios calculated by del zanna et al .
employ radiative rates generated with the superstructure code by either bhatia & doschek ( 1995 ) or themselves . for electron impact excitation rates , they use results for transitions among the lowest 31 fine - structure levels of fex calculated with the breit - pauli rmatrx code , either by pelan & berrington ( 2001 ) or once again by themselves .
recently , aggarwal & keenan ( 2004 , 2005 ) have used the fully relativistic grasp and dirac rmatrx codes to calculate radiative rates and electron impact excitation cross sections , respectively , for transitions among the energetically lowest 90 fine - structure levels of fex . in this paper
we use these results , plus additional atomic data presented here , to generate theoretical emission - line ratios for fex , and compare these with high resolution spectra from the solar extreme - ultraviolet research telescope and spectrograph ( serts ) .
our aims are threefold , namely to ( i ) assess the importance of blending in the serts observations , ( ii ) detect new fex emission lines , and ( iii ) identify the best fex line ratios for use as electron density diagnostics .
this work is of particular relevance due to the recent launch of the _ hinode _ mission , which has on board the euv imaging spectrometer ( eis ) , covering the 170211 and 246292 wavelength ranges ( culhane et al .
2007 ) , similar to the serts spectral coverage of 170225 and 235450 ( thomas & neupert 1994 ) .
it is clearly important that emission lines observed by the eis are fully assessed , and the best diagnostics identified .
serts provides the ideal testbed for this , due to its larger wavelength coverage , allowing more lines from the same species to be detected and compared with theoretical predictions .
furthermore , the best serts spectral resolution is about 0.03 [ full width at half - maximum ( fwhm ) ] , obtained for the 170225 wavelength range observed in second - order ( brosius , davila & thomas 1998a ) , which is a full factor of two better than the 0.0650.075 resolution available from eis ( young et al .
hence the serts data sets should allow emission features to be resolved and assessed which are blended in eis spectra .
indeed , this is illustrated by our previous work on fexiii ( keenan et al .
2007 ) , where we resolved the 203.79 and 203.83 features which are blended in eis observations ( young et al . 2007 ) .
serts has had a total of 10 successful flights over the period 19822000 , each featuring a somewhat different set of optical components and observing targets .
we have selected data from two of these for analysis here , as they provide the most complete sequence of well - observed fex emission lines .
the flight on 1989 may 5 ( henceforth referred to as serts89 ) carried a standard gold - coated toroidal diffraction grating , and was the first to observe a strong solar active region , noaa 5464 .
it detected hundreds of first - order emission lines in the 235450 wavelength range , plus dozens of features spanning 170225 , which appeared in second - order among the 340450 first - order lines ( thomas & neupert 1994 ) .
the spectrum was recorded on kodak 10107 emulsion , at a spectral resolution of 5080m ( fwhm ) in first - order , and a spatial resolution of approximately 7 arcsec ( fwhm ) .
this combination of high spectral resolution , strong signal , and broad wavelength coverage remains unsurpassed even today as the best available data set for investigating imaged solar emission features over the full wavelength range of 235450 , and hence are used in the present paper .
subsequent serts flights employed either a multilayer - coated diffraction grating or ccd - detector arrays which provided important technical advantages , but which also restricted the spectral bandpass that could be covered , making their data less suitable for the present study ( see keenan et al . 2007 and references therein ) .
the one exception was the flight on 1995 may 15 ( henceforth referred to as serts95 ) which incorporated a multilayer - coated toroidal diffraction grating that enhanced its sensitivity to second - order features in the 170225 wavelength range ( brosius et al
. 1998a ) .
it observed active region noaa 7870 , again using kodak 10107 emulsion , and had an improved spatial resolution of approximately 5 arcsec .
its unique shortwave multilayer coating led to the detection of many second - order emission lines not seen on previous serts flights ( thomas & neupert 1994 ; brosius et al .
1996 ) , and furthermore obtained the highest spectral resolution ( 0.03 fhwm ) ever achieved for spatially resolved active region spectra in this wavelength range .
the serts95 data therefore provide the best observations for investigating fex emission lines in the 170225 region , and hence are employed in this paper .
further details of the serts95 observations , and the wavelength and absolute flux calibration procedures employed in the data reduction , may be found in brosius et al .
( 1998a , b ) .
similar information for the serts89 spectrum is available from thomas & neupert ( 1994 ) .
we note that although the relative intensity calibration curves in both cases involved fitting to calculations of insensitive line - ratio pairs , fex lines made up less than 10% of the points utilized in those procedures .
furthermore , detailed comparisons of line ratio calculations for many ions with serts measurements have revealed generally very good agreement between theory and observation , providing support for the serts intensity calibrations ( see , for example , keenan et al .
2007 and references therein ) . thus ,
even without fully independent intensity calibrations , the serts results still provide a valid test for the new fex calculations presented here .
it should also be noted that the serts95 first - order calibration curve was extrapolated from its fitted range of 238336 to cover its complete bandpass . hence there is some extra uncertainty in the intensities of lines with wavelengths greater than 340 measured from this flight .
however , no such problem exists with the serts89 measurements of these lines .
we have searched for fex emission lines in the serts89 and serts95 spectra using the detections of thomas & neupert ( 1994 ) and brosius et al .
( 1998b ) , supplemented with those from other sources , including the nist database , the latest version ( v5.2 ) of the chianti database ( dere et al .
1997 ; landi et al .
2006 ) , the atomic line list of van hoof , peter / atomic/ ] and in particular the excellent summary of line identifications by del zanna et al . ( 2004 ) .
the latter provides not only a comprehensive list of wavelengths for well - observed fex transitions , but also indicates alternative wavelengths where previous identifications are not consistent with their conclusions . in table 1
we list the fex transitions found in the serts89 and serts95 spectra , along with their measured wavelengths .
we also indicate possible blending features or alternative identifications as suggested by thomas & neupert or brosius et al . in their original line lists . [ cols="^,<,<",options="header " , ] @xmath11from datla et al .
+ @xmath12present theoretical line ratios and those from chianti calculated for the measured plasma parameters .
our comparison of theoretical fex emission - line intensity ratios with solar active region spectra from the serts 1989 and 1995 flights reveals generally very good agreement between theory and experiment , with several features identified for the first time in the serts data sets , including fex 193.72 , 220.26 and 324.73 , plus fexi 201.56 and the feix / x blend at 175.48 .
in addition , the 195.32 transition of fex is detected for the first time ( to our knowledge ) in an astronomical source .
we find that the ratios 175.27/174.53 and 175.27/177.24 provide the best fex electron density diagnostics , as they involve lines which are strong and free from blends , are close in wavelength and the ratios are highly n@xmath5sensitive .
should these lines not be available , then the 257.25/345.74 ratio may be employed as a diagnostic , although this requires an accurate determination of the instrument intensity calibration over a relatively large wavelength range . however ,
if the weak 324.73 line is reliably detected , then the use of 324.73/345.74 or 257.25/324.73 is recommended in preference to 257.25/345.74 .
kma acknowledges financial support from epsrc , while dbj is grateful to the department of education and learning ( northern ireland ) and nasa s goddard space flight center for the award of a studentship .
the serts rocket programme is supported by rtop grants from the solar physics office of nasa s space physics division .
jwb acknowledges additional nasa support under grant nag513321 .
fpk is grateful to awe aldermaston for the award of a william penney fellowship .
the authors thank peter van hoof for the use of his atomic line list .
chianti is a collaborative project involving the naval research laboratory ( usa ) , rutherford appleton laboratory ( uk ) , and the universities of florence ( italy ) and cambridge ( uk ) . | fully relativistic calculations of radiative rates and electron impact excitation cross sections for fex are used to derive theoretical emission - line ratios involving transitions in the 174366 wavelength range .
a comparison of these with solar active region observations obtained during the 1989 and 1995 flights of the solar extreme - ultraviolet research telescope and spectrograph ( serts ) reveals generally very good agreement between theory and experiment .
several fex emission features are detected for the first time in serts spectra , while the 3s@xmath03p@xmath1 @xmath0p@xmath23s@xmath03p@xmath3(@xmath4s)3d @xmath0d@xmath2 transition at 195.32 is identified for the first time ( to our knowledge ) in an astronomical source . the most useful fex electron density ( n@xmath5 ) diagnostic line ratios
are assessed to be 175.27/174.53 and 175.27/177.24 , which both involve lines close in wavelength and free from blends , vary by factors of 13 between n@xmath5 = 10@xmath6 and 10@xmath7@xmath8 , and yet show little temperature sensitivity .
should these lines not be available , then the 257.25/345.74 ratio may be employed to determine n@xmath5 , although this requires an accurate evaluation of the instrument intensity calibration over a relatively large wavelength range .
however , if the weak 324.73 line of fex is reliably detected , the use of 324.73/345.74 or 257.25/324.73 is recommended over 257.25/345.74 .
electron densities deduced from 175.27/174.53 and 175.27/177.24 for the stars procyon and @xmath9 cen , using observations from the extreme - ultraviolet explorer ( euve ) satellite , are found to be consistent and in agreement with the values of n@xmath5 determined from other diagnostic ratios in the euve spectra .
a comparison of several theoretical extreme - ultraviolet fex line ratios with experimental values for a @xmath10-pinch , for which the plasma parameters have been independently determined , reveals reasonable agreement between theory and observation , providing some independent support for the accuracy of the adopted atomic data .
[ firstpage ] atomic data sun : activity sun : corona sun : ultraviolet . |
there have been intensive interests in the realization of fulde - ferrell - larkin - ovchinnikov ( fflo ) states in superconductors under a high magnetic field , since the first theoretical prediction .
@xcite studies on fflo states have been expanded into various fields , ranging from condensed matter @xcite and cold atoms @xcite to neutron stars .
@xcite although several candidates for fflo superconductors have already been proposed by numerous works , direct evidence of fflo states has not been observed in experiments . in a pauli - limited superconductor , the transition from the bardeen - cooper - schrieffer ( bcs ) state to the fflo state
is considered as the second - order transition accompanied by the jackiw - rebbi soliton @xcite which neatly accommodates pauli paramagnetic moment .
@xcite the soliton provides a generic key concept for commonly understanding the essential physics of fflo phases in a single - band superconductor , incommensurate structures , @xcite and fermionic excitations bound at topological defects of a superconductor through @xmath2-phase shift of the background fields .
@xcite multi - band effects originating from multiple sheets of fermi surfaces have recently been observed in a variety of superconductors , such as mgb@xmath3 , iron - based superconductors , and sr@xmath3ruo@xmath4 .
@xcite among them , strong pauli effects have been reported in iron - pnictides .
@xcite it is now recognized that multi - band effects are crucial to understand unconventional superconductivity .
nevertheless , in contrast to extensive studies in single - band superconductors , @xcite the structure and thermodynamic stability of fflo phases in multi - band systems have not been clarified so far other than a few exception .
@xcite the pauli - limiting field @xmath5 for a single - band system is given by @xmath6 , which depends only on the gap , @xmath7 , at @xmath8 and @xmath9 .
this formula is easily extended to a two - band case as @xmath10 ^ 2 + \frac{\mathcal{n}_\mathrm{f}^{(2)}}{\mathcal{n}_\mathrm{f } } \left [ \delta_{0}^{(2 ) } \right ] ^2 \right ) } , \label{eq : pauli}\ ] ] where @xmath11 ( @xmath12 ) is the density of states ( gap ) for the @xmath13 band ( @xmath14 or @xmath15 ) and we introduce @xmath16 .
when the major gap @xmath17 dominates the minor gap @xmath18 ( @xmath19 ) , the pauli - limiting field in the case of two - band systems , @xmath20 $ ] , is smaller than @xmath21 .
for @xmath22 , eq
. reduces to @xmath23 with @xmath24 being the pauli - limiting field for the @xmath13 band .
these two limiting cases show that the fflo region could become enlarged towards the lower field .
this simple consideration motivates us to look into the pauli paramagnetic effects in multi - band superconductivity .
the fflo states are cooper pairing states with a finite center - of - mass momentum @xmath25 . the nonzero center - of - mass momentum generates a spatial modulation of cooper pair amplitudes in real space , _ e.g. _ , @xmath26 .
the modulation period @xmath27 is determined by competing effect between spin paramagnetism and superconductivity . for a single - band case ,
the analytic form of the modulation length is derived in ref . .
at zero temperature , the modulation length depends on the amplitude of the pair potential at zero field , the fermi wavelength of background electrons , and an external field . in the case of a two - band superconductor ,
each band has its own favorable value of the modulation period , @xmath28 and @xmath29 .
the competing and synergetic effect arises from a coherent coupling of two bands having two modulation periods @xmath28 and @xmath29 through interband cooper pair tunneling , which makes the phase diagram of fflo states rich and complex .
we also note that multi - band effects might be accompanied by exotic superconductivity .
@xcite in the previous work , @xcite it is found that in a pauli - limiting two - band superconductor , the coherent coupling of two bands through the interband cooper pair tunneling gives rise to a devil s staircase structure in the field dependence of physical quantities and a rich fflo phase diagram . in this paper
, we more extensively investigate the thermodynamic stability and electronic structures of pauli - limiting two - band superconductors in a wide region of parameters .
based on self - consistent calculation of the bogoliubov - de gennes ( bdg ) equation , we first start to examine the phase diagram of a simplest two - band superconductor with a passive band in which the pairing interaction is absent and the superconductivity is induced by the cooper pairing transfer from the active band .
the resultant phase diagram is accompanied by the positive slope of the bcs - fflo phase boundary line , @xmath30 , @xmath31 this is contrast to the phase boundary in a single - band superconductor , @xmath32 ( see the solid line of fig .
[ fig : phase_single ] and refs . and ) .
we investigate the phase diagram in the more general case that two bands with comparable intraband interactions are coherently coupled through the cooper pair tunneling .
it is demonstrated that the sequence of fflo phases with rational modulation length is widely observed when the tunneling rate is present .
for all parameters used in the current work , the tricritical lifshitz point
l " is found to be a fixed point , @xmath33 .
this article is organized as follows . in the next section ,
we introduce the bdg equation for a multi - band system .
the details of the derivation are supplemented in appendix [ sec : bdgformalism ] . in appendix
[ sec : b ] , we also describe the way to numerically solve the bdg equation with a periodic boundary condition . the analytic form for the phase boundaries between superconducting states and normal states is derived in the latter half of the section . in sec .
[ sec : single ] , we consider a two - band superconductor with a passive band , where the interaction for intraband cooper pairing formation is absent in the second ( passive ) band and the superconductivity is induced by the cooper pair tunneling from the first ( active ) band . in sec .
[ sec:2bands ] , we show the numerical results on fflo states of pauli - limiting two - band superconductors , where both the bands have intraband pairing interaction .
finally , sec .
[ sec : summary ] is devoted to summary .
we start from the second - quantized microscopic hamiltonian with multiple bands under a uniform external magnetic field .
the hamiltonian is given as @xmath34 where the single - particle hamiltonian and interaction hamiltonian are @xmath35 the sums of @xmath36 are taken for up- and down - spins @xmath37 , and electron bands are denoted by @xmath38 .
the single - particle hamiltonian density @xmath39 is a @xmath40 component of a two - dimensional matrix @xmath41 , @xmath42 with @xmath43 .
the creation and annihilation field operators , @xmath44 and @xmath45 , are defined for each band and spin state .
the electron mass and effective bohr magneton in the @xmath13 band are denoted as @xmath46 and @xmath47 , respectively .
the @xmath48-dimensional unit matrix is denoted as @xmath49 .
we assume the effective coupling of electrons in the bands @xmath13 and @xmath50 as @xmath51 . here
@xmath52 is a vector potential , @xmath53 is a chemical potential of the @xmath13 band , and @xmath54 is an external magnetic field coupled with spins through @xmath55 , where @xmath56 are @xmath57 pauli matrices . employing the mean field approximation and following the usual procedure , @xcite we have the following bdg equation for the quasiparticle wavefunction @xmath58^\mathrm{t}$ ] , @xmath59 where the @xmath57 bdg matrix is @xmath60 equation is self - consistently coupled with the pair potential , @xmath61 where @xmath62 is the fermi distribution function at a temperature @xmath63 . here
, we assume that @xmath54 is fixed to the @xmath64 axis without loss of generality .
the detailed formalism is explained in appendix [ sec : bdgformalism ] . in this formalism , we take account of multi - band effects through the gap equation .
the coupling constants , @xmath65 , lead to the formation of intraband cooper pairs , and @xmath66 ( @xmath67 ) induces the interband cooper pair tunneling .
hence , in the case of nonzero @xmath66 ( @xmath67 ) , multiple electron bands are coherently coupled with each other through the interband cooper pair transfer .
this multi - band effect is crucial for the sequence of fflo phase transitions discussed in sec .
[ sec:2bands ] . for simplicity and to stay in a minimal model
, we consider a one - dimensional system along the @xmath64 direction . throughout this paper , the effective masses @xmath46 , and
the effective bohr magnetons @xmath47 of electrons in each band are set to be the same , @xmath68 and @xmath69 , respectively .
since we are interested in the strong pauli paramagnetic limit , the vector potential @xmath52 is supposed to be spatially uniform , neglecting the orbital repairing effect .
@xcite note that the orbital depairing effect can be suppressed in low - dimensional superconductors .
@xcite we numerically solve the self - consistent equations and .
we consider a one - dimensional modulation of the pair potential with the period @xmath70 , @xmath71 where @xmath72 corresponds to fflo ( bcs ) states .
the period @xmath70 determines the fflo modulation wavenumber @xmath25 , @xmath73 this imposes the periodic boundary condition on the quasiparticle wave functions , @xmath74 where @xmath75 is the bloch vector and @xmath76 denotes the bravais lattice vector @xmath77 in the one - dimensional system , which satisfies @xmath78 with @xmath79 .
we numerically diagonalize the bdg equation with the periodic boundary condition by using an finite element method implemented with the gauss - lobatto discrete variable representation .
the details are described in appendix [ sec : b ] . using the self - consistent solutions , we evaluate the thermodynamic potential @xmath80 , @xmath81 .
\label{eq : thermodynamicpot}\end{aligned}\ ] ] where @xmath82 , with @xmath83 . the thermodynamic potential in eq .
is calculated for both bcs and fflo states under a fixed @xmath84 and @xmath63 . and the modulation period @xmath70 in equilibrium
is determined so as to minimize @xmath85 with respect to the fflo modulation @xmath25 .
the other physical quantities are also calculated .
the magnetization density is defined as @xmath86 where we introduce the magnetization in the @xmath13 band as @xmath87.\end{aligned}\ ] ] the boundary between normal and superconducting states , @xmath89 , can be analytically derived in multi - band systems .
the ff and lo states are energetically degenerate on the boundary .
therefore , to determine the @xmath89 line , we assume the pair potential @xmath90 to be the ff state , which is given by @xmath91 where @xmath92 is a constant and @xmath25 denotes the modulation wavenumber of the ff state .
the bdg equation can be transformed to the particle - hole symmetrized form , @xmath93 where @xmath94 and @xmath95 the symmetrized bdg matrix satisfies @xmath96 with @xmath97 , where @xmath98 is the complex conjugate operator .
this implies that the positive energy solution @xmath99 with @xmath100^\mathrm{t}$ ] in eq .
is mapped to the negative energy solution @xmath101 with @xmath102 .
then the gap equation is reduced to @xmath103 , \label{eq : gapequationpositive}\end{aligned}\ ] ] where @xmath104 is considered for the upper bound for the summation .
we carry out the @xmath105 transformation of the quasiparticle wavefunction as @xmath106^\mathrm{t } = \underline{u } ( q ) \bm{\varphi}_\nu^{(\gamma ) } ( z ) $ ] , where @xmath107 $ ] .
the @xmath105 transformation reduces the bdg matrix to @xmath108 .
the transformed bdg matrix satisfies @xmath109 = \underline{0}_2 $ ] , where @xmath110 is a @xmath48-dimensional zero matrix .
the operators @xmath111 and @xmath112 have the simultaneous eigenstates .
the eigenstates can be expressed as @xmath113 , where @xmath114 is constant , and @xmath115^\mathrm{t}$ ] satisfies the normalization condition , @xmath116 .
then , eq . is recast into @xmath117 where @xmath118 with @xmath119 .
finally , the eigenenergy @xmath120 is calculated as @xmath121 with @xmath122 ^ 2 } .\ ] ] the corresponding eigenstates are given as @xmath123 } , \label{eq : u_k } \\
v_k^{(\gamma ) } = & \sqrt{\frac{1}{2 } \left [ 1 - \frac{\varepsilon^{(\gamma)}_q ( k ) - \frac{\hbar^2}{2m_\mathrm{e } } kq}{\tilde{e}^{(\gamma ) } ( k , q ) - \frac{\hbar^2}{2m_\mathrm{e } } kq } \right ] } .
\label{eq : v_k}\end{aligned}\ ] ] using these eigenenergies and eigenfunctions , the pair potential becomes @xmath124 where @xmath125 and @xmath126 means summation in @xmath127 . during the calculation ,
positive sign is taken for @xmath120 in eq . .
then the reduced gap equation for @xmath128^\mathrm{t}$ ] is expressed in the @xmath129-dimensional matrix form as @xmath130 \bm{\delta } = \bm{0},\ ] ] with @xmath131_{\gamma\gamma^{\prime } } = g_{\gamma\gamma^{\prime } } \mathcal{i}^{(\gamma^{\prime } ) } ( q ) $ ] and @xmath132 .
we here replace @xmath133 by @xmath134 in the thermodynamic limit .
to have a nontrivial set of @xmath135 , we have the condition , @xmath136 = 0 .
\label{eq : analyticboudarydet}\ ] ] the analytical phase boundary @xmath89 for @xmath129-band system is obtained from eq . for various @xmath25 s .
in the present paper , we consider the case of @xmath137 .
then , eq . can be written as @xmath138 \left [ 1 - g_{11 } \mathcal{i}^{(1 ) } ( q ) \right ] \left [ 1 - g_{22 } \mathcal{i}^{(2 ) } ( q ) \right ] \nonumber \\ & - g_{12 } g_{21 } \mathcal{i}^{(1 ) } ( q ) \mathcal{i}^{(2 ) } ( q ) = 0 .
\label{eq : analyticboundary}\end{aligned}\ ] ] to derive the @xmath139 line , we vary @xmath140 and @xmath141 so as to find the set of parameters @xmath142 which satisfy eq . under a fixed temperature @xmath63 .
the highest value of @xmath84 gives the phase boundary @xmath139 at a fixed temperature @xmath63 . in a similar manner
, we derive the bcs - normal boundary from eq . with @xmath143 .
the point at which two boundaries meet corresponds to the tricritical point , the lifshitz point
let us start from a simplest , yet nontrivial two - band case with a passive band .
this corresponds to the case of @xmath145 and the intraband interaction in the second band is absent . in such a case ,
the superconductivity in the second ( passive ) band is induced by cooper pair tunneling from the first ( active ) band through the interband interaction @xmath146 .
we study the phase diagram for various @xmath147 s .
this gives a first step to understand the thermodynamic properties of fflo phases in multi - band superconductors . .
( i ) the solid ( black ) line corresponds to the phase boundaries in a single - band superconductor ( @xmath148 ) , ( ii ) the dashed ( blue ) line is the two - band case with a passive band @xmath149 and ( iii ) the dash - dotted ( red ) line is the case of @xmath150 , where @xmath151 ( @xmath152 ) in the former ( latter ) case .
the temperature of the lifshitz point , l " , is a fixed point @xmath33 , which is independent of the interband coupling @xmath146 .
the inset shows the spatial profile of the pair potential in the top row and the magnetization in the bottom row at @xmath153 marked as a filled ( red ) circle in the case ( iii ) .
the solid ( red ) lines and the dashed ( blue ) lines correspond to the quantities of the first and second band , respectively . ]
figure [ fig : phase_single ] shows the phase diagram in an @xmath84-@xmath63 plane for three sets of parameters : ( i ) the single - band case ( @xmath148 ) and two - band cases with a passive band , ( ii ) @xmath154 and ( iii ) @xmath150 . using the self - consistent solutions under the parameter sets of ( ii ) and ( iii ) , we estimate @xmath155 as @xmath151 and @xmath156 , respectively .
the gap amplitude of the @xmath13 band at zero temperature and zero magnetic field is denoted by @xmath12 . the inset of fig .
[ fig : phase_single ] shows the spatial profile of the pair potential @xmath157 and the magnetization density @xmath158 at the filled ( red ) circle in the main frame of fig .
[ fig : phase_single ] with the parameter set of the case ( iii ) .
the pair potential shows that the energy gap of the second band has a finite amplitude , implying that the cooper pairs are transferred from the active band even if the intraband interaction in the second band is absent .
the transferred cooper pairs also affect back to the electrons in the first band through the cooper pair tunneling from the passive band .
we found the enhancement of the gap amplitude in the first band by increasing the interband interaction .
as shown in fig .
[ fig : phase_single ] , the phase boundary between bcs and fflo phases shifts to lower fields as the interband interaction increases .
it is important to mention that the lifshitz point , l " , at which three second - order phase boundaries meet shifts to the lower field as @xmath159 increases , while the temperature of the lifshitz point , @xmath160 , is unchanged by the parameter @xmath147 .
we numerically confirm that @xmath160 is invariant under any parameter changes including intraband interaction @xmath161 .
it is found that the phase boundary , @xmath162 , between bcs and fflo states , at the zero temperature sharply varies as a function of the interband coupling @xmath147 , while the magnetic field of the lifshitz point , @xmath163 , is relatively insensitive to @xmath147 .
this leads to the positive slope of the phase boundary , @xmath164 , in the large interband interaction regime , where @xmath165 is the bcs - fflo transition line .
this feature indicates that the interband interaction stabilizes the fflo states even in the lower field regime .
note that all the phase boundaries in fig .
[ fig : phase_single ] are of the second order .
in this section we set the intraband interaction in the second band to be finite , @xmath167 . in this case , the electrons in both the two bands essentially contribute to the superconducting state and the bands are coherently coupled to each other through the cooper pair tunneling with @xmath168 . before going further ,
let us review the analytic result of the magnetic field @xmath84 dependence of the fflo wavenumber @xmath25 in a single - band case .
@xcite by introducing a parameter @xmath169 , @xmath84 and @xmath25 are expressed with the analytic form , @xmath170 where @xmath7 is the pairing amplitude at @xmath171 , @xmath172 is the fermi velocity , and @xmath173 and @xmath174 are the complete elliptic functions of the first and second kinds , respectively .
although the equations and are derived from a single - band hamiltonian , it is applicable to a two - band system when the interband coupling @xmath159 is small .
we here define two coherence lengths , @xmath175 , defined in each electron band , @xmath176 where @xmath177 is the fermi wavenumber of the @xmath13 band .
let us here introduce the averaged variables of various quantities .
the averaged gap at @xmath171 is @xmath178 , the averaged chemical potential is @xmath179 , and the averaged coherence length is defined as @xmath180 where the last equation is analogous to eq . .
using these expressions , @xmath181 and @xmath182 are characterized as , @xmath183 in this analysis , we can divide the parameter space into the two cases such that the ratio of the fflo wavenumber in the high field limit denoted as @xmath184 is bigger or smaller than @xmath185 . from eqs . and
, @xmath184 is defined as @xmath186 where @xmath187 is the critical field at @xmath8 . in the present model , @xmath187 diverges . without loss of generality , in this paper , we consider @xmath188 so that @xmath189 , where @xmath190 in . in such a situation , if @xmath191 , the curves , @xmath192 and @xmath193 , are crossed at certain @xmath84 , as shown in the left panel of fig .
[ fig : analyticqs ] . for the case of @xmath194 , @xmath192 and @xmath193
are never crossed in all @xmath84 s as shown in the right panel of fig .
[ fig : analyticqs ] .
the detailed structures on the two cases , @xmath195 , are described in the following two subsections . , based on the single - band analysis .
the left ( right ) figure corresponds to the case @xmath196 . ]
let us consider the @xmath191 case where the fflo modulation wavenumbers @xmath181 s in the single - band analysis have a crossing point at a certain magnetic field ( left panel of fig .
[ fig : analyticqs ] ) . using the self - consistent solution of the bdg equation and gap equation with @xmath197
, we estimate the ratio of the order parameters at @xmath198 and @xmath9 as @xmath199
. then the ratio of the fflo wavenumber is found to be @xmath200 .
the ground state is determined by minimizing the thermodynamic potential in eq . with respect to the fflo modulation period @xmath201 . , where @xmath202 and @xmath199 . the fflo phase is separated to two regions in the low temperature regime , @xmath0- and @xmath1-fflo phases . in the high temperature regime
, it is ambiguous wether the state is @xmath0- or @xmath1-fflo . the inset shows the phase diagram for a single - band superconductor . ] the resulting phase diagram is displayed in fig .
[ fig : phase_t - case1 ] . in the low temperature regime ,
the fflo phase is divided into two regions in terms of @xmath25 . in fig .
[ fig : phase_t - case1 ] , the phase in the high field regime is referred to as the @xmath0-fflo phase and in the lower field regime as the @xmath1-fflo phase . since the second band has the smaller gap , @xmath203 , the second band becomes passive in the high field regime .
the @xmath0-fflo phase originates from the first band and its field - dependence is understandable with the single - band picture .
the detailed structure will be discussed below . in the lower field regime , however , the gain of the condensation energy in the second band becomes more important , compared with that in the higher field . since both bands form fflo states with comparable amplitudes of the pair potentials , each band has the favored fflo modulation wavenumbers as described in eqs . and .
the coherent interband coupling with @xmath204 induces frustration of two fflo modulation wavenumbers , @xmath205 and @xmath206 .
the competing effect gives rise to emergence of new fflo phase , that is , the @xmath1-fflo phase .
the frustration is never seen in the case of single - band superconductors and the @xmath1-fflo is peculiar to multi - band systems . in the high temperature region ,
the phase boundary between the @xmath0- and @xmath1-fflo phases is indistinguishable . in fig .
[ fig : phase_t - case1 ] , the thick ( thin ) line indicates the second - order ( first - order ) phase transition . the bcs , @xmath0-fflo , and @xmath1-fflo phases are separated via the first - order phase transition , which is contrast to the second - order bcs - fflo transition in single - band superconductors . the thermodynamic potential @xmath207 for various @xmath84 s is shown in fig .
[ fig : omega - q_t - case1 ] . as @xmath84 increases , the global minimum of @xmath207 shifts from @xmath1-fflo phases to the @xmath0-fflo phase .
the @xmath0- and @xmath1-fflo phases are stabilized with either the condition @xmath208 or @xmath209 , respectively . since @xmath210 , the transition becomes the first order . as a function of the fflo modulation wavenumber @xmath25 at @xmath8 for various magnetic fields , @xmath211 ( i ) , @xmath212 ( ii ) , @xmath213 ( iii ) , @xmath214 ( iv ) , @xmath215 ( v ) , @xmath216 ( vi ) , and @xmath217 ( vii )
the parameters are same as which in fig . [
fig : phase_t - case1 ] .
the inset shows the enlarged plot of @xmath85 in the smaller @xmath25 region .
the ( red ) filled circles show the global minima of @xmath85 . ]
it is found that there are many local minima in @xmath207 in fig .
[ fig : omega - q_t - case1 ] . to understand the local minima ,
let us consider the simplified pair potential , @xmath218 .
since two bands are coherently coupled through the interband tunneling , the pair potentials have to satisfy the periodic boundary condition given by eq . with @xmath219 .
then @xmath220 and @xmath221 satisfy the condition @xmath222 with @xmath223 . in fig .
[ fig : omega - q_t - case1 ] , the local minima in the low @xmath25 s in the low fields correspond to the cases of @xmath224 and @xmath225 . in the current case ( @xmath191 ) ,
all the states with nonzero @xmath226 are energetically unstable , relative to the @xmath227 @xmath1-fflo state .
figures [ fig : delta_mag_t - case1](a1)-(c1 ) show the spatial profiles of the pair potentials and figs .
[ fig : delta_mag_t - case1](a2)-(c2 ) are the magnetizations of the ground states under various @xmath84 s . here , figs .
[ fig : delta_mag_t - case1](c1 ) and [ fig : delta_mag_t - case1](c2 ) correspond to the local minimum with the largest @xmath25 at the critical field shown in the curve ( vi ) in fig .
[ fig : omega - q_t - case1 ] .
the pair potential and magnetization density of the local minimum with the second largest @xmath25 are displayed in ( b1 ) and ( b2 ) of fig .
[ fig : delta_mag_t - case1 ] . the data in figs .
[ fig : delta_mag_t - case1](a1 ) and [ fig : delta_mag_t - case1](a2 ) illustrate the global - minimum state with the largest @xmath25 in the curve ( ii ) .
to understand the spatial profiles of @xmath228 , we here carry out the fourier expansion of the pair potential , @xmath229 $ ] where @xmath230 .
the periodic boundary condition imposes the following condition on @xmath231 : @xmath232 , where @xmath233 and @xmath25 is an fflo wavenumber .
then the fourier expansion of the pair potential is given for the first and second bands as @xmath234.\ ] ] due to the symmetry of the pair potential in fflo states , the fourier components only take odd numbers , @xmath235 .
the fourier components of the pair potential is given by @xmath236 dz .
\label{eq : dm}\ ] ] similarly , the magnetization density is expanded in terms of the fourier series as @xmath237,\ ] ] with an average value of the magnetization , @xmath238 the fourier components are given by @xmath239 dz .
\label{eq : mm}\ ] ] the fourier components defined in eqs . and are displayed in the bottom of fig .
[ fig : delta_mag_t - case1 ] . from the fourier components in figs . [ fig : delta_mag_t - case1](a3)-(c3 ) and figs .
[ fig : delta_mag_t - case1](a4)-(c4 ) , these states are found to be purely sinusoidal states with a single modulation wavenumber @xmath25 .
the other local minima in fig .
[ fig : omega - q_t - case1 ] are classified as the @xmath1-fflo states with @xmath240 .
these states can not be the ground state in the case of @xmath241 , while they have the lowest energy in the case of @xmath242 .
hence , the detailed structures of the @xmath1-fflo states with @xmath240 are described in sec .
[ subsec : r<1 ] . the fourier components in fig . [
fig : delta_mag_t - case1 ] indicate that the pair potentials and the magnetizations are well described with a sinusoidal form with a single fflo modulation wavenumber . to understand this ,
let us first clarify the @xmath243 curve in the case of a single - band superconductor . in a single - band superconductor ,
the bcs - fflo phase transition occurs at the critical field @xmath244 .
the critical field is associated with the one - soliton formation energy and an isolated kink state of @xmath245 is characterized by @xmath246 $ ] with @xmath247 .
@xcite the higher fourier components with @xmath248 immediately disappear as @xmath84 increases and the spatial modulation of @xmath249 results in a sinusoidal form @xmath250 .
the field dependence of @xmath25 in the single - band case is described with eqs . and .
the slope of @xmath25 with respect to @xmath84 in the limit of @xmath251 is obtained as @xmath252}{2 e(k)/[\pi k ] } = 2.\ ] ] this indicates @xmath253 in the limit of the high magnetic field . and ( b ) spatially averaged magnetization , @xmath254 and @xmath255 , as a function of @xmath84 at @xmath8 .
the other parameters are same as those in fig .
[ fig : phase_t - case1 ] .
the ( black ) thin lines indicate the @xmath256 curves obtained from the single - band analysis .
the lines at lower @xmath25 correspond to the family of the @xmath1-fflo , @xmath257 , with @xmath258 .
the symbol `` @xmath259 '' ( `` @xmath260 '' ) indicates the local minimum state whose thermodynamic potential is lower ( higher ) than that in the bcs state .
the right panels show the detailed structures in the lower @xmath84 region . ] in the case of two - band superconductors with @xmath261 , electrons in the second band feel the relatively high magnetic field even in the vicinity of the phase boundary between bcs to @xmath1-fflo phases .
therefore , the higher fourier components never appear .
since pair potentials in two bands are coherently coupled through the interband cooper pair tunneling , the pair potential in the first band is also pulled into a sinusoidal form by the effect of the second band . figure [ fig : q_m - h_t - case1 ] shows the field dependence of the fflo wavenumber @xmath25 and the spatially averaged magnetization @xmath262 .
the resulting @xmath243 curve fits well with the @xmath205 and @xmath206 curves obtained from the single - band analysis in eqs . and .
the ground state in low magnetic field is identified as the @xmath1-fflo state with @xmath227 whose @xmath243 follows the @xmath206 curve , while it turns to the @xmath0-fflo state with @xmath263 in the higher field regime .
figure [ fig : q_m - h_t - case1 ] clearly indicates that the transition from @xmath1- to @xmath0-fflo phases is of the first - order .
the transition between bcs to @xmath1-fflo state is also of the first - order , as shown in fig .
[ fig : q_m - h_t - case1 ] . by lowering the magnetic field , the transition occurs and @xmath25 drops to zero around @xmath264 . as shown in fig .
[ fig : q_m - h_t - case1 ] , the net magnetization jumps at each critical field .
these jumps manifest the onset of the @xmath1- and @xmath0-fflo phases and might be observable in experiments . in fig .
[ fig : q_m - h_t - case1 ] , it is also found that the net magnetization in the second band decreases at the critical field @xmath265 between @xmath1- and @xmath0-fflo phases , while the total magnetization increases .
this is also observed in the case of @xmath194 . as a function of the fflo modulation wavenumber @xmath25 at @xmath266 for various temperatures , @xmath267 ( i ) , @xmath268 ( ii ) , @xmath269 ( iii ) , and @xmath270 ( iv ) .
each plot is scaled by @xmath271 for visibility .
the inset shows the calculated points in the phase diagram in fig .
[ fig : phase_t - case1 ] . ] as seen in fig .
[ fig : phase_t - case1 ] , the phase boundary between the @xmath0- and @xmath1-fflo states are indistinguishable in the high temperature region of the phase diagram .
this is because the two local minima originated from @xmath272 and @xmath273 merge into the same value with increasing the temperature , as shown in fig .
[ fig : omega - q_t - case1_ambiguos ] . the @xmath1-fflo state with @xmath274 can be continuously connected to the @xmath0-fflo state with @xmath225 if they share the same @xmath25 .
note here that , as discussed later , this is different from the case of @xmath194 in which the family of the @xmath1-fflo states with @xmath240 becomes the ground state and successive first - order transitions occur .
-fflo states at @xmath275 .
the ( blue ) dash - dotted line ( ii ) and the ( green ) dotted line ( iii ) correspond to @xmath1-fflo states with @xmath274 at @xmath275 and @xmath212 , respectively .
the density of states in ( i)-(iii ) are displayed in ( b1)-(b3 ) , respectively .
the other parameters are same as those in fig .
[ fig : phase_t - case1 ] . ]
let us confirm the realization of the @xmath1-fflo states from the perspective of the energy spectra .
figure [ fig : spectra_dos_t - case1 ] illustrates the energy spectra and the density of states ( dos ) of three fflo states .
the corresponding pair potentials are displayed in fig .
[ fig : delta_mag_t - case1 ] .
the energy spectra in the first and second bands are plotted in figs .
[ fig : spectra_dos_t - case1](a1 ) and [ fig : spectra_dos_t - case1](a2 ) , respectively .
the ( red ) solid line corresponds to the @xmath0-fflo state and the other lines depict the @xmath1-fflo states for various @xmath84 s .
the energy spectrum of the first - band is qualitatively unchanged by the magnetic field , which always opens a finite energy gap at the fermi level . the band structures around @xmath276 in figs .
[ fig : spectra_dos_t - case1](a1 ) and [ fig : spectra_dos_t - case1](a2 ) are interpreted as the lattice formation of the jackiw - rebbi solitons bound at the fflo nodes .
@xcite in the case of the @xmath1-fflo with @xmath227 , as shown in fig .
[ fig : spectra_dos_t - case1](a2 ) , the energy dispersion of the soliton band does not cross the fermi level .
as @xmath84 increases , however , the fflo modulation period @xmath201 shortens and thus the dispersion of the jackiw - rebbi soliton band becomes dispersive . as shown in fig .
[ fig : q_m - h_t - case1](a ) , @xmath25 jumps at the critical @xmath277 between the @xmath1- and @xmath0-fflo phases . in the @xmath0-fflo phase
, the jackiw - rebbi soliton band crosses the fermi level .
these features are also confirmed in the dos of the second band ( blue lines ) displayed in figs .
[ fig : spectra_dos_t - case1](b1)-(b3 ) , where the zero - energy dos is absent in the @xmath1-fflo states in figs .
[ fig : spectra_dos_t - case1](b2 ) and [ fig : spectra_dos_t - case1](b3 ) .
the change of low - energy structures implies that the @xmath1-fflo states gain the condensation energy at low temperatures and thus they can be stable relative to the @xmath0-fflo state .
since the soliton band is occupied ( unoccupied ) by only up - spin ( down - spin ) electrons in the low temperature regime , paramagnetic moments are accumulated by the fflo nodal planes .
hence , the spatial oscillation of magnetization density in figs .
[ fig : delta_mag_t - case1](a2 ) , [ fig : delta_mag_t - case1](b2 ) , and [ fig : delta_mag_t - case1](c2 ) , reflects the fflo modulation period @xmath70 and jackiw - rebbi soliton states in the low temperatures , which might be detectable in experiments .
let us now turn to the case of @xmath194 where @xmath206 is always bigger than @xmath205 .
the resulting phase diagram is shown in fig .
[ fig : phase_t - case2 ] . here , we take the two sets of parameters : @xmath278 and @xmath279 . using the former ( latter ) set , the ratio of the pair potential at @xmath198 and @xmath9 is estimated as @xmath280 ( @xmath281 ) .
then the ratio of the wavenumbers is @xmath282 and @xmath283 , respectively .
for both the sets of parameters , the region of high fields and low temperatures is occupied by two fflo phases , the @xmath0- and @xmath1-fflo phases .
in contrast to the case of @xmath241 , however , the @xmath1-fflo phase is further divided into a family of the @xmath1-fflo states through successive first - order phase transitions .
the detailed structures will be discussed below .
we here calculate the physical quantities in both the cases , which yield qualitatively same behaviors . in this paper , therefore , we focus on the results with the set of parameters , @xmath278 . the more information of the latter case can be found in ref . . in the main frame and for @xmath284 in the inset
the former ( latter ) case corresponds to @xmath285 ( @xmath283 ) and @xmath286 ( @xmath281 ) .
the fflo phase is separated by the first - order transition to @xmath0- and @xmath1-fflo states .
the @xmath1-fflo phase is divided into a family of the @xmath1-fflo states through the first - order transitions . ] at @xmath8 for @xmath287 ( i ) , @xmath288 ( ii ) , @xmath289 ( iii ) , @xmath290 ( iv ) , @xmath291 ( v ) , and @xmath292 ( vi ) .
the set of parameters are @xmath278 .
the left panel is an enlarged plot of the low field region .
the ( red ) filled circles denote the global minima of @xmath85 . ] in fig .
[ fig : omega - q_t - case2 ] we illustrate the thermodynamic potential @xmath293 for various @xmath84 s as a function of the fflo wavenumber @xmath25 . in fig .
[ fig : omega - q_t - case2 ] , the data ( i)-(iii ) show @xmath85 in the vicinity of the phase boundary between the bcs and @xmath1-fflo phases , and ( iv)-(vi ) are in the higher field regime around the boundary between @xmath1- and @xmath0-fflo phases . there
exist several local minima in @xmath85 . in the lower field regime
, the fflo phase with small @xmath25 s can be the ground state , while the value of @xmath25 in the ground state successively jumps to larger values with increasing @xmath84 .
figures [ fig : delta_mag_fourier_t - case2](a1)-(c1 ) show the spatial profile of the pair potentials and the corresponding magnetization densities are displayed in figs .
[ fig : delta_mag_fourier_t - case2](a2)-(c2 ) .
the corresponding fourier components are displayed in the bottom .
as seen in figs .
[ fig : delta_mag_fourier_t - case2](c1 ) and [ fig : delta_mag_fourier_t - case2](c2 ) , the pair potentials of two bands yield a simple sinusoidal modulation , and the spatial oscillations are characterized by a single fflo modulation wavenumber , corresponding to the fourier component with @xmath294 in eq . .
similarly , the spatial modulation of @xmath158 is composed of a single fourier component with @xmath295 in eq . .
it turns out that the modulation wavenumber @xmath25 coincides with the modulation wavenumber favored by the first band , @xmath205 . in this sense
, we refer to this state as the @xmath0-fflo state .
this state corresponds to the local - minimum state with the largest value of @xmath25 in fig .
[ fig : omega - q_t - case2 ] .
the @xmath0-fflo state always occupies the high field region of the phase diagram , regardless of @xmath241 and @xmath242 ( see figs . [
fig : phase_t - case1 ] and [ fig : phase_t - case2 ] ) and the parameters @xmath296 . figures [ fig : delta_mag_fourier_t - case2](b1 ) and [ fig : delta_mag_fourier_t - case2](a1 ) illustrate the spatial profiles of the pair potential which correspond to the local minima with the second and third largest @xmath25 in fig .
[ fig : omega - q_t - case2 ] , respectively .
the corresponding magnetization densities are displayed in fig .
[ fig : delta_mag_fourier_t - case2](b2 ) and ( a2 ) .
it is seen from figs .
[ fig : delta_mag_fourier_t - case2](b3 ) and [ fig : delta_mag_fourier_t - case2](a3 ) that @xmath157 is composed of multiple wavenumbers , @xmath297 , in addition to @xmath25 . in the case of fig .
[ fig : delta_mag_fourier_t - case2](b1 ) , the wavenumber @xmath298 of @xmath157 is favored by the pair potential of the second band @xmath299 , namely , @xmath300 . since the two bands are coherently coupled through the interband cooper pair tunneling and @xmath301 favors the relatively longer modulation period , the extra fourier component with @xmath25 then mixes the longer modulation period in @xmath157 , which is favored by the first band @xmath301 . as a result , the spatial modulation of fig .
[ fig : delta_mag_fourier_t - case2](b1 ) is described as @xmath302 .
we refer to this state as the @xmath1-fflo state with @xmath303 ( see fig . [
fig : omega - q_t - case2 ] ) .
similarly , @xmath157 in the case of fig .
[ fig : delta_mag_fourier_t - case2](a1 ) is composed of mainly two fourier components , @xmath25 and @xmath304 , where the shorter period @xmath305 is favored by the second band .
the fflo state in fig .
[ fig : delta_mag_fourier_t - case2](a1 ) is identified as the @xmath1-fflo state with @xmath306 . and ( b ) spatially averaged magnetizations @xmath254 and @xmath255 as a function of @xmath84 for @xmath278 and @xmath198 .
the ( black ) thin lines indicate the @xmath243 curve obtained form the single - band analysis .
we also depict the curves @xmath307 , corresponding to the family of the @xmath1-fflo .
the symbol``@xmath259 '' ( `` @xmath260 '' ) denotes the local minimum state whose thermodynamic potential is lower ( higher ) than that in the bcs state .
the right panels illustrate the detailed structures in the lower @xmath84 region . ]
note that in low temperatures , the modulation of the magnetization density reflects the spatial oscillation of @xmath157 in the fflo states . as seen in fig .
[ fig : delta_mag_fourier_t - case2](c4 ) , the spatial modulation of the magnetization density in the @xmath1-fflo state is expressed by a single fourier component with @xmath308 . in fig .
[ fig : delta_mag_fourier_t - case2](b4 ) and ( a4 ) , we show the fourier components of @xmath158 in @xmath1-fflo states with @xmath306 and @xmath309 , where the spatial modulation consists of multiple fourier components .
this feature in @xmath1-fflo states is qualitatively different from the single-@xmath25 behavior in the @xmath0-fflo state .
hence , the multiple-@xmath25 modulations in the magnetization density manifest the stability of the @xmath1-fflo states .
let us now consider the transition from @xmath1- to @xmath0-fflo states .
the field - dependence of the fflo modulation wavenumber , @xmath25 , is displayed in fig .
[ fig : q_m - h_t - case2](a ) . the corresponding phase diagram in fig .
[ fig : phase_t - case2 ] indicates that in the higher magnetic field regime , the @xmath0-fflo state is a ground state .
this is because the minority pair potential , @xmath299 , becomes passive in the high field regime and the fflo modulation is determined by the majority band .
hence , the field - dependence of @xmath25 follows the curve @xmath192 which is obtained from the single - band analysis in eqs . and .
in the lower field regime where the pair potential in the second band is competitive to that of the first band , the frustration between @xmath205 and @xmath206 gives rise to anomalous field - dependence of the fflo modulation wavenumber @xmath25 . in the case of @xmath194 which we consider ,
the fflo wavenumber favored by the second - band , @xmath206 , is always larger than @xmath205 , as shown in fig .
[ fig : analyticqs](b ) .
hence , as discussed above , the lower field region is occupied by a family of the @xmath1-fflo states , whose pair potentials are characterized by multiple wavenumbers , @xmath310 , with @xmath311 . as illustrated in fig .
[ fig : q_m - h_t - case2](a ) , the magnetic field drives successive first - order phase transition among @xmath1-fflo phases with different @xmath226 s , which are accompanied by the jump of the @xmath25 value .
the transition from @xmath1- to @xmath0-fflo phases is also of the first - order . as seen in fig .
[ fig : q_m - h_t - case2 ] , the field - dependence of the net magnetization clearly indicates the successive first - order transitions .
the so - called _ devil s staircase _
@xcite is peculiar to fflo phases in multi - band superconductors and might be detectable in experiments .
note that although there exist more steps in the @xmath312 of fig .
[ fig : q_m - h_t - case2](b ) , further calculations in the lower field region are technically difficult .
this is because @xmath1-fflo states with an infinitely large @xmath226 are competing to the spatially uniform bcs state .
we find that the thermodynamic potential of the bcs state is still higher than that of @xmath1-fflo states around @xmath313 .
therefore , the critical field below which the bcs state is stabilized will be located in the region of @xmath314 , where @xmath315 denotes the pauli - limiting field in the second band and @xmath316 with the parameters of fig .
[ fig : q_m - h_t - case2 ] . finally , to explain the thermodynamic stability of @xmath1-fflo states relative to the @xmath0-fflo state , we illustrate in fig . [
fig : spectra_dos_t - case2 ] the quasiparticle energy spectra and dos of the @xmath0-fflo state and @xmath1-fflo states with @xmath317 and @xmath306 . similarly with the case of @xmath241 in sec .
[ subsec : r>1 ] , the dispersion of the @xmath0-fflo state crosses the fermi level , while the @xmath1-fflo states open a finite energy gap around the fermi level . .
the ( red ) line ( i ) corresponds to @xmath0-fflo states at @xmath318 .
the ( blue ) dash - dotted line ( ii ) and ( green ) dotted line ( iii ) illustrate the @xmath1-fflo states with @xmath317 at @xmath319 and with @xmath320 at @xmath321 , respectively .
the corresponding density of states to the cases ( i)-(iii ) are displayed in ( b1)-(b3 ) . ] in the case of the @xmath306 @xmath1-fflo state , as shown in fig .
[ fig : delta_mag_fourier_t - case2](a3 ) , @xmath157 is composed of two modulation periods @xmath322 and @xmath323 , @xmath324 . the fflo state with the short period @xmath325 is accompanied by a finite zero energy density of states .
@xcite the mixing of the fourier component with the longer modulation period @xmath27 , however , folds back the original band into the small reduced brillouin zone with the @xmath326 size . as a result
, the band gap opens around the fermi level in the reduced brillouin zone of the minor band ( fig .
[ fig : spectra_dos_t - case2](a2 ) ) , resulting in the gain of condensation energy .
the multiple - band structure appears around @xmath327 , @xmath328 , and @xmath329 in the @xmath1-fflo states with @xmath317 and @xmath306 .
this helps the stability of the @xmath1-fflo states in the lower field regime , compared with the @xmath0-fflo state with a finite zero energy dos .
we have studied multi - band effects on superconductivity under an external magnetic field .
we have here focused on fulde - ferrell - larkin - ovchinnikov ( fflo ) states in two - band systems . extending the basic theory based on the bogoliubov - de gennes equation to a multi - band system
, we have developed the formalism to calculate analytic phase boundaries between superconducting states and the normal state .
we have first examined the phase diagram in a two - band system with a passive band , where intraband interaction in the second band is absent . as the interband interaction increases , the critical field , @xmath330 , above which the fflo state is stable , decreases and becomes lower than the tricritical lifshitz point .
this indicates that the bcs - fflo phase boundary line @xmath165 yields a positive slope , @xmath331 , contrast to that in a single - band superconductor , @xmath332 .
@xcite we have numerically checked that the temperature at the lifshitz point is invariant with respect to interband and intraband interactions . for a two - band superconductor where intraband interactions are present in both bands , there are two different fflo modulation wavenumbers , @xmath205 and @xmath206 , which are favored by the first and second bands , respectively .
the competing effect between two length scales @xmath28 and @xmath29 plays an important role in the fflo states of multi - band superconductors , giving rise to the existence of two fflo phases : @xmath0-fflo and @xmath1-fflo phases . in this paper , we divide the parameter region into two cases : @xmath333 and @xmath242 . in both cases
, it has been demonstrated that the phase diagram in the low temperature and high field region is occupied by the @xmath0-fflo phase , whereas the @xmath1-fflo becomes stable in the lower field regime around the pauli - limiting fields . in the case of @xmath191 , the spatial modulation of the pair potential in the @xmath0 and @xmath1-fflo phases
is well characterized by a single modulation wavenumber @xmath334 and @xmath335 , respectively .
the transition from the bcs to @xmath1-fflo phase is of the first order . with increasing an applied field
, the @xmath1-fflo phase undergoes the first - order transition to the @xmath0-fflo phase in the low temperature regime , while the transition becomes ambiguous in high temperatures . in the case of @xmath242 ,
the @xmath1-fflo phase is divided into a family of the @xmath1-fflo states , where the fflo modulation is composed of multiple wavenumbers .
the transitions between all the superconducting phases are of the first order .
the field - dependence of the fflo modulation wavenumber and net magnetization is accompanied by the devil s staircase , which clearly indicates the successive first - order transitions between the @xmath1-fflo substates .
recently , strong pauli - limited superconductivity has been reported in an iron - based superconductor , kfe@xmath3as@xmath3 .
@xcite it is known that kfe@xmath3as@xmath3 has @xmath336- , @xmath337- , @xmath338- , and @xmath339-bands , where the @xmath339-band has a relatively larger fermi velocity than that in other bands and @xmath340 .
@xcite the fflo modulation wavenumbers in each band are estimated as @xmath341 . @xcite this corresponds to the case of @xmath342 in the present context .
hence , it is inferred that the @xmath1-fflo phase appears in the low temperature and high field regime and with increasing the magnetic field , it undergoes the first - order transition to the @xmath0-fflo phase .
note that the generalized whh approach recently makes the pauli - limit effect of kfe@xmath3as@xmath3 questionable .
@xcite to fully understand the phase diagram of kfe@xmath3as@xmath3 , however , we must take account of orbital depairing effect which induces vortices . in a single - band superconductor
, the formation of vortices is competitive to the stability of the fflo state .
@xcite the study on the interplay of vortices with @xmath0- and @xmath1-fflo states remains in the future work .
this work was supported by jsps ( nos .
21340103 , 23840034 , 25287085 , and 25800199 ) and by grant - in - aid for scientific research from mext of japan , `` topological quantum phenomena '' nos . 22103005 and 25103716 .
in this appendix , we describe the details on the derivation of the bdg equation and gap equation .
let us start from the second - quantized microscopic hamiltonian with multiple bands under a uniform external magnetic field , which is presented in eqs . - . employing the mean field approximation
, the hamiltonian reduces to @xmath343 \label{eq : hmf}\ ] ] where @xmath344 . here
, @xmath228 denotes the pair potential defined by @xmath345 from now on , without loss of generality , @xmath54 is fixed to the @xmath64 axis : @xmath346 .
the second term of eq . becomes @xmath347 , and the single - particle hamiltonian density becomes diagonal .
it is convenient to introduce the following spinor notation : @xmath348^\mathrm{t } $ ] , where @xmath349 means transposition .
the mean - field hamiltonian in eq . is rewritten as follows : @xmath350 where the bdg hamiltonian density is given by @xmath351 we introduce the following bogoliubov transformation , @xmath352 with quasiparticle basis , @xmath353^\mathrm{t}$ ] . with the transformation ,
the mean - field hamiltonian with the bdg matrix is diagonalized to @xmath354 the creation and the annihilation operators of the quasiparticles , @xmath355 and @xmath356 , satisfy fermionic anti - commutation relations .
the quasiparticle wave function in the matrix form @xmath357 satisfies the orthonormal condition , @xmath358 and the completeness condition , @xmath359 by substituting eq . to eq . and using eqs . and
, we derive the following bogoliubov - de gennes ( bdg ) equation @xmath360 the bdg matrix @xmath361 described above yields double eigenstates for spins . to see this
, we set the eigenfunction of up spin states with an eigenvalue @xmath362 as @xmath363^\mathrm{t}$ ] .
then , the corresponding bdg equation is given by @xmath364 .
it is found that the bdg equation simultaneously has eigenstates for down spins of @xmath365^\mathrm{t}$ ] with the eigenvalue of @xmath366 : @xmath367 .
therefore , once one solves the bdg equation for the quasiparticle @xmath368^\mathrm{t}$ ] and @xmath369 , @xmath370 then the solution gives the eigenstate for both spin states with positive and negative energies .
one can realize that the bdg equation is the same form as that in a single band superconductor .
multi - band effects appear through the pair potential @xmath90 .
the gap equation is derived by substituting eq . to eq . and replacing variables in eq . ,
@xmath371\\ = & \sum _ { \gamma^{\prime},\nu}g_{\gamma \gamma ' } u_\nu^{(\gamma ' ) } ( \bm{r } ) v_\nu^{(\gamma ' ) \ast } ( \bm{r } ) f ( e_\nu^{(\gamma ' ) } ) , \label{eq : gapequation2}\end{aligned}\ ] ] where @xmath372 is the fermi distribution function with inverse temperature @xmath373 and @xmath374 is the boltzmann constant . here
, we utilize the following relations , @xmath375 , @xmath376 , and @xmath377 . the derived gap equation has a finite energy cutoff @xmath378 in numerics , @xmath379 in the case of a single - band superconductor with a spatially uniform gap @xmath380 , the gap equation reduces to @xmath381 where @xmath382 .
this gives the relation between the cutoff energy and physical quantities such as the intraband coupling constant @xmath383 and the pair potential @xmath380 at @xmath8 and @xmath384 .
hence , we estimate the amplitude of the intraband coupling constant @xmath383 from eq . with the energy cutoff @xmath378 , the chemical potential @xmath385 , and the pair potential @xmath386 at @xmath8 and @xmath384 .
we start by considering a periodic boundary condition , @xmath387 where @xmath388 . the bravais lattice vector is defined as @xmath389 with @xmath390 and the primitive vectors of the lattice , @xmath391 ( @xmath392 ) .
we assume that the phase factor @xmath393 yields additivity , @xmath394 . for simplicity
, we do not take account of the local gauge transformation , which implies that @xmath393 is independent of @xmath395 .
the bdg equation is @xmath396 where the matrix is given by @xmath397 let us now consider the periodic boundary condition for the bdg equation given by @xmath398 = e_{\mu}t_{\bm{r}}\bm{\varphi}_{\mu}(\bm{r})$ ] , where the translational operator , @xmath399 , obeys @xmath400 .
it is obvious from that the periodicity of the bdg matrix is expressed as the u(1 ) gauge transformation , @xmath401 = u(\bm{r})\mathcal{h}(\bm{r})u(\bm{r})$ ] , where @xmath402 yields additivity , @xmath403 and is supposed to be @xmath404 ^ 4=1 $ ] .
in addition , the eigenfunction of the bdg equation , @xmath405 , is the simultaneous eigenfunction of the translation operator , @xmath406 = 0.\ ] ] using the bloch vector @xmath407 with the reciprocal lattice vectors @xmath408 $ ] , we obtain the periodic boundary condition on the quasiparticle wavefunction as @xmath409 we here impose the born - von karman boundary condition , @xmath410 , where @xmath411 and @xmath412 . in the current problem
, we set @xmath413 in eq . .
the phase shift is taken to be @xmath219 for fflo states and @xmath414 for spatially uniform bcs states .
then , the periodic boundary condition on the quasiparticle wavefunction is recast into @xmath415 where the bloch vector is taken to be @xmath416 with @xmath417 and @xmath418 .
we numerically solve the resulting one - dimensional bdg equation with the periodic boundary condition in eq .
using the finite - element method with the gauss - lobatto discrete variable representation [ [ finite - element - method - with - the - gauss - lobatto - discrete - variable - representation ] ] finite element method with the gauss - lobatto discrete variable representation ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ here we describe the details about how to numerically solve the bdg equation with the periodic boundary condition in eq . .
in order to map the bdg equation into the eigenvalue equation , we apply the the finite ( fe ) element method implemented with the gauss - lobatto discrete variable representation ( dvr ) .
@xcite for simplicity , let us consider the following differential equation in the coordinate space , @xmath419 $ ] , @xmath420 without loss of generality , @xmath421 is composed of spatial derivatives and potential terms as @xmath422 .
the eigenfunction @xmath423 obeys the periodic boundary condition , @xmath424 where @xmath425 and @xmath426 .
we start by dividing the continuous variable @xmath427 into the @xmath129 intervals , where the @xmath428-th interval is defined as @xmath429 $ ] with the @xmath430 discrete grids @xmath431 @xmath432 ) .
each interval is further divided into sub grids labeled by @xmath433 as @xmath434 ( see fig . [
fig : grids ] ) . in order to represent wave functions in each interval ,
we define the set of finite - element basis functions , @xmath435 , which are identically zero outside a given interval : @xmath436 for @xmath437 $ ] .
the index @xmath438 denotes the local basis function in each interval and represents the dvr grids .
@xcite following ref .
, we introduce the boundary condition on @xmath439 and for @xmath440 @xmath441 it is convenient to construct the set of the orthonormal basis functions , @xmath442 , from @xmath443 , which obeys the condition , @xmath444 we expand the eigenfunction @xmath423 in terms of the basis functions , @xmath445 since the eigenfunction must be continuous across each interval , the coefficients for @xmath446 are required to satisfy the following conditions @xmath447 for @xmath446 and @xmath448 this indicates that the coordinate @xmath449 $ ] is divided into @xmath450 grid points .
each grid @xmath451 is accompanied by the local basis functions @xmath452 .
we now explicitly define the local basis functions , @xmath443 , that is , @xmath453 . following ref .
, the dvr is implemented to the finite - element method .
the discrete grids in each interval , @xmath454 , are chosen to be a set of gauss - lobatto quadrature points . in the interval of @xmath455
$ ] , the gauss - lobatto quadrature rule turns the integral of an arbitrary function @xmath456 over the continuous variable @xmath457 to the summation on the discrete grids @xmath458 , @xmath459 , which gives an exact result if the function @xmath456 is a polynomial of degree @xmath460 .
in contrast to the gauss - legendre quadrature , the endpoints are constrained to @xmath461 and @xmath462 .
the points and weights @xmath463 are chosen to be the @xmath336-th zeroth of the first derivative of the @xmath464-th degree legendre polynomials , @xmath465 , and @xmath466 ^ 2 } } , & \mbox{otherwise}. \end{array } \right.\ ] ] using the gauss - lobatto quadrature points and weights , we now define the grid points @xmath451 and the weights @xmath467 in each interval as @xmath468 now let us construct the dvr basis functions based on the gauss - lobatto quadrature , where arbitrary functions given at each fe - dvr grids @xmath454 are interpolated by lagrange polynomials , @xmath469 .
\end{array } \right.\ ] ] the set of the basis functions is localized at each fe - dvr grid , @xmath470 and obeys the orthogonal condition with the accuracy of the gauss - lobatto quadrature , @xmath471 .
we can now construct the set of orthonormal functions , @xmath472 , from the local basis functions , @xmath473 , which allows to expand the eigenfunction @xmath423 .
the basis functions are continuous in each interval , @xmath474 $ ] , while the continuity at the interval boundaries is ensured by introducing the so - called `` bridge '' function , @xmath475 .
the normalized bridge function at the grid @xmath476 and @xmath477 is defined as @xmath478 the bridge functions for interval boundaries with @xmath479 and @xmath480 are given by @xmath481 for each fe - dvr grid except for the boundaries , the orthonormal functions @xmath482 are obtained by normalizing the local basis function @xmath483 as @xmath484 the set of the functions in eqs .
- satisfies the normalization condition in eq . by expanding the eigenfunction @xmath485 in terms of the fe - dvr functions @xmath482 , as shown in eq .
, the original equation can be mapped onto the eigenvalue problem . since @xmath482 is localized at each fe - dvr grid , the potential term @xmath486 is diagonalized as @xmath487 the matrix elements of the second derivative term for @xmath488 are evaluated in terms of the fe - dvr basis as @xmath489 where the diagonal elements for @xmath490 are given by @xmath491,\end{aligned}\ ] ] otherwise @xmath492 the other elements are @xmath493 @xmath494 and @xmath495 and @xmath496 .
the elements @xmath497 and @xmath498 are obtained with the phase factor in eq . as @xmath499 we here set @xmath500 for @xmath501 and @xmath502 and @xmath503 for @xmath504 and @xmath505 , otherwise @xmath506 .
the first derivatives of the local basis functions @xmath507 at quadrature points , @xmath508 , are given by @xmath509 in the fe - dvr method , non - zero hamiltonian matrix elements are composed of only the gradient terms and the potential terms is fully diagonalized .
the number of non - zero elements , @xmath510 , are determined by the number of dvr grids in each interval , @xmath254 ; @xmath511 .
this drastically reduces the number of non - zero elements in the standard dvr method @xcite with @xmath512 .
hence , the fe - dvr method is accompanied by the sparse matrix and convenient for the implementation of the lanczos / arnoldi algorithm and massive parallel computation .
@xcite t. terashima , m. kimata , n. kurita , h. satsukawa , a. harada , k. hazama , m. imai , a. sato , k. kihou , c .- h .
lee , h. kito , h. eisaki , a. iyo , t. saito , h. fukazawa , y. kohori , h. harima , and s. uji , j. phys .
* 79 * , 053702 ( 2010 ) .
s. kittaka , y. aoki , n. kase , t. sakakibara , t. saito , h. fukazawa , y. kohori , k. kihou , c .- h .
lee , a. iyo , h. eisaki , k. deguchi , n. k. sato , y. tsutsumi , and k. machida , j. phys .
. jpn . * 83 * , 013704 ( 2014 ) . | multi - band effects on fulde - ferrell - larkin - ovchinnikov ( fflo ) states of a pauli - limiting two - band superconductor are studied theoretically , based on self - consistent calculations of the bogoliubov - de gennes equation .
first , we examine the phase diagrams of two - band systems with a passive band in which the intraband pairing interaction is absent and superconductivity is induced by a cooper pair tunneling from an active band .
it is demonstrated that the temperature of the lifshitz point at which three second - order transition lines meet is independent of the cooper pair tunneling strength .
the bcs - fflo critical field becomes lower than the lifshitz point with increasing the interband tunneling strength , and the resultant phase diagram is qualitatively different from that in a single - band superconductor .
we also study the thermodynamics of pauli - limiting two - band superconductors with comparable intraband pairing interactions . as a consequence of a competing effect between two bands ,
the fflo phase is divided into two phases : @xmath0- and @xmath1-fflo phases .
the @xmath0-fflo is favored in a high field regime and the @xmath1-fflo becomes stable in the lower field . in a particular case , the latter
is further subdivided into a family of fflo states with rational modulation lengths , leading to a devil s staircase structure in the field - dependence of physical quantities .
the critical field , above which the fflo is stabilized , is lower than that in a single band superconductor , while the temperature of tricritical lifshitz point is invariant under the change of two - band parameters . |
the mass distribution of white dwarfs ( wds ) in the solar neighborhood peaks at @xmath20.6 @xmath1 ( holberg et al . 2002 ; liebert et al . 2005 ; kepler et al . 2007 ) .
separate low mass ( @xmath30.45 @xmath1 ) and high mass ( @xmath40.8 @xmath1 ) components are also observed with each one contributing 10% and 15% to the total mass budget , respectively ( liebert et al . 2005 ) .
the magnitude - limited sample of wds from the palomar green ( pg ) survey includes wds with masses as low as 0.32 @xmath1 .
recent discoveries of several extremely low mass ( @xmath5 0.2 @xmath1 ) wds in the field ( liebert et al .
2004 ; eisenstein et al . 2006 ; kawka et al . 2006
; kilic et al .
2007a ) and around pulsars
( bassa et al . 2006 ; van kerkwijk et al . 1996 ) show that wds with mass as low as 0.17 @xmath1 are formed in the galaxy .
the formation of low mass wds require significant mass loss from a post main sequence star .
the only viable scenario for such mass loss has traditionally been the formation of a wd in a close binary system .
if a post main sequence star has a close binary companion , it can lose its outer envelope without reaching the asymptotic giant branch and without ever igniting helium , ending up as a helium core wd .
confirmation of the binary nature of several low mass wds by marsh et al .
( 1995 ) , marsh 2000 , maxted et al .
( 2006 ; 2007 ) , and kilic et al . ( 2007b ) supports this binary formation scenario .
an alternative scenario for the formation of low mass wds through extreme mass loss from single metal - rich stars was proposed by hansen ( 2005 ) to explain the wd sequence of ngc 6791 .
hubble space telescope observations of ngc 6791 showed that it has a population of wds brighter than expected for a 7 gyr old open cluster .
this population of wds implied an age of 2.7 gyr assuming they are carbon - oxygen core stars ( bedin et al .
hansen ( 2005 ) suggested that the age degeneracy is removed by assuming the stars are he - core wds and that most of the observed wds in this cluster are actually low mass wds created as a result of significant mass loss on the red giant branch in this metal - rich system ( [ fe / h ] = + 0.3 @xmath6 + 0.5 ; origlia et al .
2006 ; gratton et al . 2006 ) .
follow up spectroscopic work by kalirai et al .
( 2007a ) showed that the majority of the brighter wds in this cluster are actually 0.4 @xmath1 wds . since the binary star fraction of ngc 6791 seems to be low ( on the order of 14% ; janes & kassas 1997 ) , these low mass wds are most likely produced from single star progenitors .
there is therefore direct observational evidence that low mass wds may be copiously produced in a sufficiently old and metal rich star without requiring binary interactions . in this paper
, we provide evidence for the existence of a metal - rich population of field stars in the past 10 billion years of the galactic history and we show that such stars are likely to produce single low mass wds . the age - metallicity relation for the local galactic disk is discussed in 2 , while the observed populations of single and binary low mass wds among the field wd population are discussed in 3 and 4 , respectively . the implications of our findings are discussed in 5 .
the metallicity of the galactic disk is expected to increase with time as older generations of stars return nucleosynthesis products to the interstellar environment and new generations are born with more metal - rich gas .
the age - metallicity relation for fgk dwarfs in the solar neighborhood confirms this expectation , however the presence of old , metal - rich clusters like ngc 6791 and ngc 188 shows that there is an apparent scatter in age - metallicity relation for the galactic disk .
valenti & fischer ( 2005 ) presented a large sample of nearby fgk field dwarfs with high resolution spectroscopy which enabled them to determine stellar properties for these stars including metallicities .
using hipparcos parallaxes and the above dataset , reid et al .
( 2007 ) created a volume limited sample of nearby stars within 30 pc of the sun .
they have found that the mean metallicity of nearby stars has increased by @xmath7 dex in the last 10 gyr .
figure 1 shows the metallicity histograms for different lookback times for the reid et al . sample ( kindly made available to us by i. n. reid ) .
only stars from the valenti & fischer ( 2005 ) analysis are shown as they have reliable metallicity and age estimates .
this figure shows that the mean metallicity of the nearby sample of fgk dwarfs has increased with time , and that there is a large scatter in metallicity at all ages .
even though the metallicity distribution peaks below solar metallicity until about 5 gyrs ago , there have been metal rich stars with [ fe / h ] @xmath4 0 at all times in the local galactic disk .
the fraction of metal rich stars increased from 21% nine billion years ago ( middle left panel ) to 42% five billion years ago ( top right panel ) .
if we consider only stars with [ fe / h ] @xmath4 + 0.2 and [ fe / h ] @xmath4 + 0.3 ( similar to ngc 6791 ) , the fraction of these stars has been around 14% and 5% in the past 10 gyr , respectively .
this analysis shows that ngc 6791 is not unique in terms of its metallicity ; @xmath25% of the stars in the solar neighborhood had similar metallicities around the same time ngc 6791 was formed .
earlier studies by edvardsson et al .
( 1993 ) and nordstrm et al .
( 2004 ) also found metal - rich stars at all times in the history of the galactic disk , including stars with [ fe / h ] @xmath4 + 0.3 ( see the right two panels in figure 7 of reid et al .
hence , these studies also confirm that metal - rich stars has existed in the galaxy for a long time .
since the metal rich stars in ngc 6791 have formed low mass wds through extreme mass loss in the red giant branch , there is a priori no reason to assume that the same mechanism would not work for metal - rich field stars .
a confirmation of this mechanism would come from the discovery of single low mass wds in the field wd sample .
the binary formation scenario for low mass wds has been tested several times .
the first efforts to find binary companions to low mass wds came from marsh et al .
they have obtained radial velocity measurements of 7 low mass wds found by bergeron et al .
( 1992 ) and they have discovered 5 double degenerate close binary systems .
two of the stars in their sample , wd1353 + 409 and wd1614 + 136 did not show any radial velocity variations and their predicted companions still remain to be discovered ( maxted & marsh 1998 ) .
based on the high success rate of their observations , the idea that all low mass wds has to be produced through close binary evolution became widely accepted .
a recent analysis of 348 hydrogen atmosphere wds from the pg survey has revealed 30 low mass wds with @xmath8 @xmath1 ( liebert et al .
liebert et al .
predicted that 100% of these low mass wds require binary evolution most likely with other helium or carbon / oxygen core wds .
fifteen of these wds have already been searched for companions through radial velocity observations or near - infrared photometry .
eight ( 53% ) of these low mass wds have companions found by radial velocity surveys ( marsh 1995 ; marsh et al .
1995 ; holberg et al .
1995 ; maxted & marsh 1999 ; nelemans et al .
2005 ) two of which also confirmed by near - ir photometry ( farihi et al . 2005 ) .
maxted et al .
( 2000 ) did not find any radial velocity companions to 5 pg low mass wds .
in addition , farihi et al .
( 2005 ) did not see any near - ir excess for 3 wds in this sample .
one object with no evidence of a companion , wd1614 + 136 , is commmon to both maxted et al . and farihi et al .
overall , 47% of the pg low mass wds that have been searched for companions seem to be single .
maxted et al .
( 2000 ) presented their findings from a larger radial velocity survey of 71 wds including 14 low mass wds with no detectable radial velocity variations .
they had found 16 low mass wds with binary companions in their previous surveys , bringing their low mass wd sample size to 30 .
this implies that 47% of the low mass wds in their extended survey are single .
maxted et al .
( 2000 ) found that a mixed single / binary population of low mass wds is at least 20 times more likely to explain their data than a pure binary population .
perhaps due to the boldness of the idea at the time , they concluded that this result depends on several factors including assumed period distribution for possible binary companions , and they found the evidence for a population of single low mass wds not sufficient . the eso sn ia progenitor survey ( spy ) searched for radial velocity variations in more than a thousand wds using the very large telescope ( napiwotzki et al .
they have observed about 75% of the known wds accessible by the vlt with magnitudes brighter than b=16.5 ( nelemans et al . 2005 ) including 26 wds with @xmath9 @xmath1 .
they have found that 15 of these low mass wds do not show any radial velocity variations , corresponding to a single low mass wd fraction of 58% ( napiwotzki et al .
their detailed analysis of the detection probability of main sequence star , wd , and brown dwarf companions show that their observations of low mass wds are inconsistent with the idea that all of them reside in close binary systems .
the spy observations are less sensitive to brown dwarf companions , because of their smaller mass , but even in this case 100% binary fraction rate can be ruled out ( r. napiwotzki 2007 , private communication ) .
brown dwarf companions to wds are rare ( less than 0.5% ; farihi et al . 2005 ) and such companions are expected to end up in short period systems ( @xmath10 day ) after the common envelope phase
. such short period brown dwarf companions would reveal their presence through h@xmath11 emission as in the case of wd0137 - 349 ( maxted et al .
2006 ) .
all of the above searches to find companions to low mass wds show that the fraction of binary low mass wds is not 100% . the surveys with more than 10 low mass wds ( the pg , maxted et al . , and spy surveys ) reveal a single low mass wd fraction of 47 - 58% .
in this section we infer the space densities of both single and binary helium wds and compare them with expectations .
we contend that the observed single he wd population is comparable to what is expected for the fraction of old metal - rich stars .
furthermore , the binary he wd population is compatible with simple theoretical estimates , and in particular we argue that the latter population is probably underrepresented because of the presence of close and bright main sequence ( ms ) companions .
liebert et al .
( 2005 ) identified three peaks in the wd mass distribution - a low mass component ( he ) and two c / o components that correspond to low and high mass evolution . to convert their observations to formation rates they divided their sample into cooling curve age bins and divided the observed number in each mass bin by the lifetime in each bin .
this provides an observational test of incompleteness , in the sense that if cooler wds are missed in the survey , artificially low formation rates would be derived from older and cooler wds .
this is clearly seen in the most numerous middle category , where the formation rates derived from the three youngest ( and hottest ) bins are the same and then decline monotonically for the four oldest bins . with this procedure they derived formation rates of 0.4 , 4.5 , and 0.9 @xmath12 for the 0.4 @xmath1 , 0.6 @xmath1 , and @xmath40.8 @xmath1 components , respectively .
they also applied a binary correction to the he wd population . in the discussion that follows
we will omit this because we are interested in counting systems rather than individual wds .
in the traditional picture , making a he wd involves truncating the evolution of a low to intermediate mass star in a close binary system .
the maximum mass is set by the size of the core at the end of main sequence evolution .
stars above 2.3 @xmath1 have core masses large enough upon leaving the main sequence to produce c / o wds even for interacting binaries ( for a good review see iben & livio 1993 ) .
the maximum orbital separation depends on both the mass and mass ratio ; it can be as small as 0.5 au for a higher mass ( @xmath22 @xmath1 ) primary with a low mass companion and as large as 1.5 au for a lower mass ( @xmath21 @xmath1 ) primary with a higher mass companion .
we adopt 1 au as an average value . for the duquennoy & mayor ( 1991 ) period distribution ,
12% of binaries have orbital periods less than 1 au . for a 50% binary population , this yields a predicted frequency of 6% for low mass systems with at least one he wd .
if we exclude the high mass wds ( as close binaries in this mass range would not necessarily produce he wds ; see nelemans et al .
2001 ) the inferred formation rate of single he wds in the pg survey is 4% ( assuming 50% single wd fraction as implied by the surveys discussed in the previous section ) and that of binary he wds is also 4% .
given both the roughness of the calculation and the small he wd sample in liebert et al .
( 2005 ) , the binary fraction is consistent with expectations .
however , there is a potential - and checkable - systematic error that will lead to a differential undercounting of even hot he wds from the presence of ms companions . for a typical distance of 100
pc and a duquennoy & mayor ( 1991 ) period distribution , one would expect 50% of c / o wds to be single , 20% to be in resolved binaries , and 30% to be in unresolved binaries . by their nature ,
binary he wds come from a population that consists entirely of unresolved binaries .
furthermore , the secondaries of close binaries with periods less than one year are systematically higher in mass than those of distant binaries ; as a result the unresolved main - sequence ( ms ) companions of he wds will be substantially brighter than those of c / o wds .
therefore , the detectability of even hot he wds can be strongly influenced by companions , and we argue that the liebert et al . estimate is probably a lower bound to the number of he wds from binaries .
the median temperature for the 30 low mass wds in the pg survey is @xmath13 k corresponding to @xmath14 for a 0.4 @xmath1 wd ( bergeron et al .
this is comparable to the blue magnitude for a 0.65 @xmath1 ms star ; we adopt this as the mass threshold for secondaries that could influence the detectability of wds at this temperature .
if we adopt a typical age of 5 gyr , secondaries with masses below 1.3 @xmath1 will on average be unevolved .
the ratio of double degenerate systems to those with ms companions has been addressed in detailed population synthesis studies ( iben et al . 1997 ) .
iben et al . assumed that only wds with companions less than 0.3 @xmath1 were detectable , and such low mass systems would be 37% of the detectable total . for a flat relative mass function
this implies that in the underlying population 70% of all systems would have ms companions with masses less than 1.3 @xmath1 .
however , for hotter wds one would detect a larger fraction of systems .
out of all systems 32% would be undetected ( @xmath15 0.65 @xmath1 ) , 38% detected with an ms companion ( @xmath16 0.65@xmath1 ) , and 30% would be double degenerates . with this assumption an observed relative formation rate of 4% for binary wds becomes an intrinsic population of 5.9% .
this is entirely consistent with the predicted binary fraction of 6% from duquennoy & mayor ( 1991 ) period distribution .
we can use the 2mass data to check on the fraction of systems with low mass companions . with a field binary fraction of 50%
we would expect half of the `` single '' wds to actually have distant and unresolved non - interacting binary companions ; these companions will typically be low mass stars .
we expect 70% of the low mass wds produced from binary interactions to have close ms companions and the other 30% to be double degenerate systems . in the liebert et al .
sample there are 14 low mass wds with @xmath17 k that were used to estimate the formation rate without applying an incompleteness correction .
we would expect 2 double degenerate systems , 5 close binary companions , and 3.5 distant binary companions .
three systems are known to have degenerate companions , and we find significant j and k band excesses for 4 of the systems . based upon a comparison with isochrones , we infer masses of 0.3 - 0.35 @xmath1 for 3 companions and @xmath20.5 @xmath1 for one .
this indicates a deficit of companions relative to expectations , and in particular we note that there are no higher mass companions seen even though a significant number would be naturally produced by binary evolution .
if all systems were from interacting binaries , and the sample was complete , we would expect 10 ms companions .
we therefore conclude that even the hot sample of he wds is likely to be incomplete due to the presence of nearby companions . until a radial velocity survey is completed , it is also not clear whether the ms companions detected are close binaries ( attributed to the binary formation channel ) or distant ones ( attributed to the single formation channel ) .
follow - up work of this type will be essential to empirically set the intrinsic space densities .
more metal - rich post - main sequence stars are expected to lose more mass on the red - giant branch compared to less metal - rich stars , and this effect is directly observed in the red - giant branch luminosity functions of several old open clusters including ngc 188 and ngc 6791 ( see fig . 8 in kalirai
an important question is the significance of the mass loss process for different metallicities .
kalirai et al .
( 2005 ; 2007a ) claimed to see a small dependence of mass loss on metallicity in hyades ( [ fe / h ] = + 0.17 ) and ngc 2099 ( m37 ; [ fe / h ] = @xmath60.1 ) . for a given fraction of low mass wds produced from severe giant branch mass loss
, one can compare with the field metallicity distribution to infer a critical abundance above which the late stages of stellar evolution are truncated .
our implied assumption is that all old stars above some threshold truncate their evolution above a critical metallicity and none do below it ; this is certainly a simplification given the stochastic nature of mass loss .
we also note that this process is expected to occur only for stars that are old enough ( and have low enough turnoff mass ) to shed their envelopes prior to experiencing a helium flash .
an intermediate mass metal - rich star would still produce an ordinary c / o wd even with enhanced mass loss . as a result , the boundary that we derive is an upper limit , and the process may become important for stars with even lower metal abundances . according to the reid et al .
( 2007 ) analysis , the fraction of [ fe / h ] @xmath4 0 stars in the solar neighborhood has been more than 21% in the past 10 gyrs .
if the mass loss process was significant for [ fe / h ] @xmath4 0 , we would expect to find 21% of the field wd population as single low mass wds . since the single low mass wd formation rate among all wds is around 4% , the mass loss is not significant enough for [ fe / h]@xmath4 0 to produce wds with helium cores ( @xmath30.45 @xmath1 ) .
the fraction of metal rich stars with [ fe / h ] @xmath4 + 0.2 has been more than 10% in the past 10 gyr , therefore the metallicity effect seems to kick in for more metal - rich stars .
we find that the mass loss becomes significant enough to produce low mass wds for [ fe / h ] @xmath4 + 0.3 as the fraction of [ fe / h ] @xmath4 + 0.3 stars has been @xmath183% for the past 10 gyr .
a caveat in this analysis is that the valenti - fischer dataset is biased toward more metal rich stars than the complete volume - limited sample of reid et al .
( 2007 ; see their figure 6 ) .
therefore , the actual metallicity limit where the mass loss process is extreme enough to create low mass wds may be lower than [ fe / h ] = + 0.3 .
for example , allende prieto et al . (
2004 ) found that among the 1477 thin disk stars in their extended sample , the fraction of stars with [ fe / h ] @xmath4 + 0.2 and + 0.3 is 5.6% and 1.8% , respectively .
observations of several star clusters with different metallicities can help refine the constraint on metallicity at which the mass loss is significant enough to create low mass wds , however the field population of metal rich stars and single low mass wds imply that low mass stars with [ fe / h ] @xmath18 + 0.3 are likely to lose enough mass on the red giant branch to skip the asymptotic giant branch evolution and directly evolve into single helium - core wds .
the confirmation of the mass loss - metallicity dependence for both ngc 6791 and the field stars has important implications : even though several groups have been searching for planets around wds through infrared ( debes & sigurdsson 2002 ; burleigh et al .
2002 ) and mid - infrared excess around wds ( mullally et al .
2007 ) , there are no known planetary companions to wds yet .
this may be caused by a target selection effect . as all of these groups have targeted typical @xmath20.6 @xmath1 wds , they have selectively targeted descendants of lower metallicity stars .
fischer & valenti ( 2005 ) found that at [ fe / h ] @xmath4 + 0.3 , 25% of the stars have gas giant planets whereas less than 3% of the stars with -0.5 @xmath3 [ fe / h ] @xmath3 0.0 have doppler - detected planets . since the super - solar metallicity stars are expected to be the progenitors of single low mass wds , these wds are prime targets for planet detection .
if the progenitors of these single low mass wds have planets ( 25% likely ) , the planets are more likely to survive the post - main sequence evolution , as these stars do not go through the agb phase .
even though the fraction of binary low mass wds seems to be around 50% for @xmath20.4 @xmath1 wds , we expect this fraction to be 100% for the extremely low mass wds with @xmath19 @xmath1 .
the mass loss process is most likely not significant enough to produce 0.2 @xmath1 wds even at higher metallicities . the initial - final mass relation for wds (
weidemann 2000 ; kalirai et al .
2007b ) predict that a 1 @xmath1 star would produce a 0.55 @xmath1 wd . on the other hand , at [ fe / h ] @xmath2 + 0.4 , dcruz et al .
( 1996 ) predict that the same star would produce a [email protected] @xmath1 wd , well above the 0.2@xmath1 limit .
so far most of the known extremely low mass wds are found as companions to pulsars . only recently , several 0.2 @xmath1 wds are found in the field population ( liebert et al .
2004 ; kawka et al .
2006 ; eisenstein et al . 2006 ; kilic et al .
however , only one of them has been searched for companions and found to be in a binary system ( kilic et al .
2007b ) . a systematic search for companions to wds with masses [email protected] @xmath1 is required to evaluate the binary fraction at different mass ranges .
a prediction of this scenario is that we should see very few truly metal rich red clump giants .
this effect is seen in the red giant branch luminosity function of ngc 6791 ; the stars are thinning out of the red giant branch and some stars never reach the tip of the giant branch ( kalirai et al .
. a similar effect is also seen in the local population of giants .
figure 9 in luck & heiter ( 2007 ) compares the metallicity histograms for dwarfs and giants within 15 pc of the sun .
it is clear from their figure that even though the dwarf population has a high metallicity tail extending up to [ fe / h]@xmath2 + 0.6 ( for both of their dwarf samples including and excluding the planet - hosting stars ; top three panels ) , the giants show a significant drop in numbers after [ fe / h ] = + 0.2 and no giants with [ fe / h ] @xmath4 + 0.45 are observed in the field population . using hipparcos parallaxes and metallicities for 284 nearby red giant stars , udalski ( 2000 )
showed that there is a very well defined red clump for lower metallicity giants , i.e. [ fe / h ] @xmath200.05 , but above that the red clump becomes really broad ( see their figure 2 ) .
the higher metallicity bin is poorly defined due to the small number of metal - rich red clump stars in the solar neighborhood .
depending on the metallicity of the progenitors , the mass loss process is likely to create red giants with a range of envelope masses which would create a broad distribution of red clump stars .
these observations also suggest that the mass loss can thin out the envelopes of red giants and that most super - solar metallicity stars will not reach the tip of the red giant branch .
in our scenario , the mass loss rate in cool giants increases dramatically with increased metallicity .
this could impact the winds from more massive stars as well , which in turn could affect the type ia sne production rate .
there are two main formation scenarios for type ia sne ; accretion from a non - degenerate companion star onto a wd close to the chandrasekhar mass limit ( e.g. whelan & iben 1973 ) and mergers of two wds ( e.g. webbink 1984 ) . in the former case
, a finely tuned mass accretion rate is required to avoid either nova explosions or common envelope evolution .
changes in the stellar wind properties will certainly impact this balance .
a low snia production rate is actually observed in elliptical galaxies ( e.g. sullivan et al .
2006 ) and in low - redshift galaxy clusters ( sharon et al . 2007 ) relative to star forming galaxies .
this has been interpreted as evidence that a prompt channel for snia is more important than the slower phase which had previously been believed to be the main channel ( e.g. panagia et al .
however , a metallicity dependent production efficiency could explain the same observations and should be explored .
our understanding of the formation of planetary nebulae ( pne ) and whether they all require binary companions or not is questionable ( moe & de marco 2006 ) .
if both binary and single stars produce pne , another interesting consequence of our finding is that these low - mass , helium core wds would most likely not produce pne , as they completely avoid the agb phase .
indeed , gesicki & zijlstra ( 2007 ) find that the central stars of pne have a very sharply peaked mass distribution , with a mean mass of @xmath21 and a range of @xmath22 @xmath1 .
the relatively low number of pne ( per unit galaxy luminosity ) in more metal - rich elliptical galaxies ( buzzoni et al .
2006 ; see their figure 11 ) is consistent with this prediction .
we examined the age - metallicity relation for the local galactic disk and found that there have been super - solar metallicity stars in the solar neighborhood in the past 10 gyr . since high metallicity stars in ngc 6791 produced single low mass wds
, we expect that there should be a population of single low mass wds in the field wd population as well .
such a population of single low mass wds is actually observed in several surveys including the palomar green sample and the spy survey .
the observed fraction of single low mass wds is around 50% , corresponding to a formation rate of 4% among the entire wd population . by comparing the fraction of high metallicity stars in the solar neighborhood ,
we show that @xmath183% of the stars in the solar neighborhood have had [ fe / h ] @xmath4 + 0.3 in the past 10 gyr , therefore such stars are likely progenitors of the single low mass wds observed today .
our analysis indicates that this channel of stellar evolution actually happens not only in ngc 6791 , but in the field stars as well .
the major implication of our finding is that a substantial fraction of old metal - rich stars end their lives prior to igniting helium and skip interesting and visible phases of stellar evolution that such stars would ordinarily experience .
this also leads to an inversion of prior searches for planetary companions to wds , which avoided low mass wds as being the products of binary interactions . instead
, some of these objects may be the descendants of the sort of metal - rich stars that are known to have a high specific frequency of planet detection .
we would like to thank i. n. reid for a careful reading of this manuscript .
we would also like to thank the participants of the morning `` astronomy coffee '' at the department of astronomy , the ohio state university , for the daily and lively astro - ph discussion , one of which prompted us to investigate the problem described in this paper .
allende prieto , c. , barklem , p. s. , lambert , d. l. , & cunha , k. 2004 , , 420 , 183 bassa , c. g. , van kerkwijk , m. h. , koester , d. , & verbunt , f. 2006 , , 456 , 295 bedin , l. r. , salaris , m. , piotto , g. , king , i. r. , anderson , j. , cassisi , s. , & momany , y. 2005 , , 624 , l45 bergeron , p. , saffer , r. a. , & liebert , j. 1992 , , 394 , 228 burleigh , m. r. , clarke , f. j. , & hodgkin , s. t. 2002 , , 331 , l41 buzzoni , a. , arnaboldi , m. , & corradi , r. l. m. 2006 , , 368 , 877 dcruz , n. l. , dorman , b. , rood , r. t. , & oconnell , r. w. 1996 , , 466 , 359 debes , j. h. , & sigurdsson , s. 2002 , , 572 , 556 duquennoy , a. , & mayor , m. 1991 , , 248 , 485 edvardsson , b. , andersen , j. , gustafsson , b. , lambert , d. l. , nissen , p. e. , & tomkin , j. 1993 , , 275 , 101 eisenstein , d. j. et al .
2006 , , 167 , 40 farihi , j. , becklin , e. e. , & zuckerman , b. 2005 , , 161 , 394 fischer , d. a. , & valenti , j. 2005 , , 622 , 1102 gesicki & zijlstra 2007 , a&a , 467 , l29 gratton , r. , bragaglia , a. , carretta , e. , & tosi , m. 2006 , , 642 , 462 hansen , b. m. s. 2005 , , 635 , 522 holberg , j. b. , saffer , r. a. , tweedy , r. w. , & barstow , m. a. 1995 , , 452 , l133 holberg , j. b. , oswalt , t. d. , & sion , e. m. 2002 , , 571 , 512 iben , i. j. , & livio , m. 1993 , , 105 , 1373 iben , i. j. , tutukov , a. v. , & yungelson , l. r. 1997 , , 475 , 291 janes , k. , & kassas , m. 1997 , the third pacific rim conference on recent development on binary star research , 130 , 107 kalirai , j. s. , richer , h. b. , reitzel , d. , hansen , b. m. s. , rich , r. m. , fahlman , g. g. , gibson , b. k. , & von hippel , t. 2005 , , 618 , l123 kalirai , j. s. , bergeron , p. , hansen , b. m. s. , kelson , d. d. , reitzel , d. b. , rich , r. m. , & richer , h. b. 2007a , , in press ( astro - ph/0705.0977 ) kalirai , j. s. , hansen , b. m. s. , kelson , d. d. , reitzel , d. b. , rich , r. m. , & richer , h. b. 2007b , , submitted ( astro - ph/0706.3894 ) kawka , a. , vennes , s. , oswalt , t. d. , smith , j. a. , & silvestri , n. m. 2006 , , 643 , l123 kepler , s. o. , kleinman , s. j. , nitta , a. , koester , d. , castanheira , b. g. , giovannini , o. , costa , a. f. m. , & althaus , l. 2007 , , 375 , 1315 kilic , m. , allende prieto , c. , brown , w. r. , & koester , d. 2007 , , 660 , 1451 kilic , m. , brown , w. r. , allende prieto , c. , pinsonneault , m. h. , & kenyon , s. j. 2007 , , 664 , 1088 liebert , j. , bergeron , p. , eisenstein , d. , harris , h. c. , kleinman , s. j. , nitta , a. , & krzesinski , j. 2004 , , 606 , l147 liebert , j. , bergeron , p. , & holberg , j. b. 2005 , , 156 , 47 luck , r. e. , & heiter , u. 2007 , , 133 , 2464 marsh , t. r. 1995 , , 275 , l1 marsh , t. r. , dhillon , v. s. , & duck , s. r. 1995 , , 275 , 828 marsh , t. r. 2000 , new astronomy review , 44 , 119 maxted , p. f. l. , & marsh , t. r. 1998 , , 296 , l34 maxted , p. f. l. , & marsh , t. r. 1999 , , 307 , 122 maxted , p. f. l. , marsh , t. r. , & moran , c. k. j. 2000 , , 319 , 305 maxted , p. f. l. , napiwotzki , r. , dobbie , p. d. , & burleigh , m. r. 2006 , , 442 , 543 maxted , p. f. l. , odonoghue , d. , morales - rueda , l. , napiwotzki , r. , & smalley , b. 2007 , , 376 , 919 moe , m. , & de marco , o. 2006 , , 650 , 916 mullally , f. , kilic , m. , reach , w. t. , kuchner , m. j. , von hippel , t. , burrows , a. , & winget , d. e. 2006 , , in press ( astro - ph/0611588 ) napiwotzki , r. , et al .
2001 , astronomische nachrichten , 322 , 411 napiwotzki , r. , et al . 2007 , 15th european workshop on white dwarfs , 372 , 387 nelemans , g. , yungelson , l. r. , portegies zwart , s. f. , & verbunt , f. 2001 , , 365 , 491 nelemans , g. , et al . 2005 , , 440 , 1087 nordstrm , b. , et al .
2004 , , 418 , 989 origlia , l. , valenti , e. , rich , r. m. , & ferraro , f. r. 2006 , , 646 , 499 panagia , n. , della valle , m. , & mannucci , f. 2007 , aip conf .
, eds . l.burderi et al .
reid , i. n. , turner , e. l. , turnbull , m. c. , mountain , m. , & valenti , j. a. 2007 , , in press ( astro - ph/0704.2420 ) sharon , k. , gal - yam , a. , maoz , d. , filippenko , a. v. , & guhathakurta , p. 2007
, , 660 , 1165 sullivan , m. , et al .
2006 , , 648 , 868 udalski , a. 2000 , , 531 , l25 valenti , j. a. , & fischer , d. a. 2005 , , 159 , 141 van kerkwijk , m. h. , bergeron , p. , & kulkarni , s. r. 1996 , , 467 , l89 webbink , r. f. 1984 , , 277 , 355 weidemann , v. 2000 , , 363 , 647 whelan , j. , & iben , i. j. 1973 , , 186 , 1007 | low mass helium - core white dwarfs ( @xmath0 @xmath1 ) can be produced from interacting binary systems , and traditionally all of them have been attributed to this channel . however , a low mass white dwarf could also result from a single star that experiences severe mass loss on the first ascent giant branch .
a large population of low mass he - core white dwarfs has been discovered in the old metal - rich cluster ngc 6791 .
there is therefore a mechanism in clusters to produce low mass white dwarfs without requiring binary star interactions , and we search for evidence of a similar population in field white dwarfs .
we argue that there is a significant field population ( of order half of the detected systems ) that arises from old metal rich stars which truncate their evolution prior to the helium flash from severe mass loss .
there is a consistent absence of evidence for nearby companions in a large fraction of low mass white dwarfs .
the number of old metal - rich field dwarfs is also comparable with the apparently single low mass white dwarf population , and our revised estimate for the space density of low mass white dwarfs produced from binary interactions is also compatible with theoretical expectations .
this indicates that this channel of stellar evolution , hitherto thought hypothetical only , has been in operation in our own galaxy for many billions of years .
one strong implication of our model is that single low mass white dwarfs should be good targets for planet searches because they are likely to arise from metal - rich progenitors .
we also discuss other observational tests and implications , including the potential impact on sn ia rates and the frequency of planetary nebulae . |
the dynamic response of the ising systems has recently been studied extensively employing computer simulations [ 1 ] . in particular , the study of dynamical response of ising systems to oscillating magnetic field [ 2,3 ] has led to many intriguing dynamic phenomena , like dynamic hysteresis and the fluctuation induced dynamic symmetry breaking transitions in ( low e.g. , one , two or three dimensional ) ising system in presence of an oscillating field . acharyya and chakrabarti also noted [ 3,4 ] some anomalous behaviour in the growth of pulse susceptibility , in ising systems under pulsed magnetic fields of finite durations . usually , when a cooperatively interacting thermodynamic system in equilibrium is perturbed ( with the perturbation having a step function like variation with time ) , then the relaxation of the system ( to the equilibrium appropiate to the perturbed state ) is observed to follow the common debye type form with a single relaxation time .
the standard ( debye ) form for any response function ( say magnetisation of ising system ) @xmath11 is @xmath12 where @xmath13 is the relaxation time , @xmath14 denotes the new equilibrium value , @xmath15 is a constant .
as the critical temperature is approached @xmath13 shows a critical slowing down ; @xmath13 diverges at the critical temperature @xmath2 : @xmath16 where @xmath17 is the correlation length , @xmath18 is the dynamic exponent and @xmath19 is the correlation length exponent [ 1 ] . here
, we have investigated in details the response of pure ising systems to pulsed magnetic fields of finite duration , using monte carlo simulations for two dimensional ising systems , and solving numerically the meanfield equation of motion .
we have studied the response behaviour for positive pulses , where the pulsed field is in the direction of the sponteneous magnetisation in the ordered phase . in the disordered phase , of course , this notion is immaterial .
one can also study the effect of negative pulses on the sponteneous order , where the field direction is opposite to sponteneous magnetisation .
although many intriguing features of the domain growth etc are expected for such negative pulse problem , we confine here to the study of positive pulses only .
we have measured the ratio @xmath0 of the response magnetisation ( pulse ) half - width ( @xmath20 ) to that ( @xmath5 ) of the pulsed external field , and the ratio @xmath1 of the response magnetisation peak height ( @xmath21 ) to the field pulse height ( @xmath22 ) giving the pulse susceptibility .
the temperature variation of these two quantities for various pulse width durations and heights of the external field , have been investigated .
we find that for weak pulses , while the width - ratio @xmath0 diverges at the order - disorder transition point @xmath2 of the unperturbed ising system , the pulse susceptibility @xmath1 does not diverge and for low dimensions , e.g. , in two dimension studied here , a smeared peak in @xmath1 occurs at an effective @xmath4 , which approaches @xmath2 as the field pulse width @xmath5 increases ( @xmath1 diverges at @xmath2 as @xmath23 ) . a meanfield analysis for the absence of divergence of @xmath1 for finite @xmath5 has been developed .
also , a finite time scaling analysis , similar to fisher s finite size ( length ) scaling [ 5 ] , has been developed and compared with the observation of the effective transition temperature @xmath4 with the external field pulse width @xmath5 .
we have organised this paper as follows : in section ii , the model and the simulation techniques have been described . in section iii ,
the results are given .
the paper ends with concluding remarks in section iv .
a ferromagnetically interacting ( nearest neighbour ) ising system in presence of a time dependent magnetic field can be described by the hamiltonian @xmath24 where @xmath25 s are spin variable having their value @xmath26 and @xmath27 is the time varying longitudinal magnetic field . here
, we have considered the time variation of @xmath27 as follows : @xmath28 where @xmath22 is the amplitude of the field and @xmath5 is the duration or the active period of the external field . in our simulation , we have considered a 500@xmath29500 square lattice in two dimension . at each site of the lattice there is a spin variable @xmath25 = ( @xmath26 ) .
we update the lattice by stepping sequentially over it following the glauber single spin flip dynamics .
one such full scan over the lattice is unit time step ( monte carlo step or mcs ) here .
first , we allowed the system to reach the equilibrium ( at any temperature @xmath30 ) and only after that the magnetic field @xmath27 has been switched on ( @xmath31 is thus much larger than the relaxation time of the system ) .
one can also use random updating sequences .
however , it takes more time ( mcs ) to stabilise the system .
we expect all the features of the response studied here remains qualitatively unchanged for random updating sequences , and we give the results here for sequential updating only . we have measured the maximum height ( above the equilibrium value ) @xmath21 and half - width @xmath20 of the response magnetisation . here , as mentioned before , the following two quantities have been defined to characterise the response of the system : @xmath32 and @xmath33 it may be noted that @xmath1 reduces to normal static susceptibility as @xmath34 and @xmath23 . at each temperature for
fixed @xmath22 and @xmath5 , the numerical values of @xmath0 and @xmath1 are obtained averaging over 20 random monte carlo realisations ( initial seed ) .
we have studied the temperature variation of these two quantities .
the results are given and discussed in the next section .
the observation for finite peak in @xmath1 at an effective transition temperature @xmath35 is analysed in view of finite time scaling behaviour , as mentioned before .
to study the similar response in the case where the fluctuations are absent , we have considered the following meanfield dynamical equation of motion for the kinetic ising system : @xmath36 here @xmath37 is the microscopic relaxation time , @xmath38 is the average magnetisation ( in the meanfield approximation ) , @xmath27 is the time varying pulse field having the same time variation described in equation [ ham ] and @xmath30 denotes the temperature .
we have solved numerically the above equations using fourth order runge - kutta method .
we have evaluated the pulse width ratio @xmath0 and pulse susceptibility @xmath1 at various temperatures for fixed pulse width @xmath5 and height @xmath22 .
the numerical results are given in the next section , where we have compared the results for finite peak height in @xmath1 ( at @xmath2 = 1 ) , with an approximate analytic estimate of @xmath1 in such cases .
as mentioned before , our results here are for two dimensional ising system on square lattice of size 500@xmath29500 .
we applied an external field of amplitude @xmath22 for a duration @xmath5 after bringing the system into a steady state .
for this , the typical number of monte carlo steps required for this size of the lattice chosen here , is observed to be of the order of @xmath39 .
the response magnetisation has amplitude @xmath21 ( measured from the equilibrium value ) and a half - width @xmath20 .
1 shows a typical time variation of magnetic field @xmath27 and the corresponding response magnetisation @xmath11 .
the dynamical response is characterised by two quantities ; the width - ratio @xmath0 and the pulse - susceptibility @xmath1 .
the temperature variation of these two quantities has been studied .
2 shows the temperature variations of @xmath0 and @xmath1 for fixed values of @xmath40 and for three values of @xmath41 mcs ) .
since @xmath21 is bounded from above , a large @xmath22 would saturate @xmath21 and hence @xmath1 becomes small ( due to the saturation ) .
also , for extremely small @xmath22 , it becomes difficult to identify @xmath21 from the noise , and hence the estimate of @xmath1 becomes erroneous .
we found @xmath42 to be well within the above optimal range . from the figure it is clear that @xmath0 has a sharp divergence at @xmath43 , the onsager transition temperature , due to the small size of the system ) almost irrespective of the values of the pulse width @xmath5 .
but @xmath1 shows peak at different points ( significantly above @xmath44 ) depending upon the values of the field pulse width @xmath5 . as @xmath5 value increases
, it is observed that the peak shifts towards @xmath2 from above ( and also the peak height grows ) .
let us try to understand why the width - ratio @xmath0 diverges at @xmath2 , while height - ratio or pulse susceptibility @xmath1 shows peak at some higher value @xmath45 depending upon the value of @xmath5 : @xmath46 as @xmath23 .
the pulse - like perturbation probes the response of the system at finite frequencies .
consequently , the @xmath47 can not diverge as the response magnetisation height @xmath21 is not an equilibrium value corresponding to the pulse height @xmath22 ; rather the @xmath21 results are bounded by the time window of width @xmath5 . on the other hand
, the response will take its own relaxation time to come to its equilibrium value ( irrespective of the value of @xmath5 ) , when the field is switched off ( at @xmath48 ) .
this leads to the divergence of @xmath20 , due to critical slowing down , as @xmath30 approaches @xmath2 .
the sharp divergence of width - ratio @xmath0 is identified as the consequence of critical slowing down and the point of divergence is the critical temperature @xmath2 for the ferro - para transition .
in fact , since the relaxation after the withdrawal of the pulse is unrestricted by the pulse - width , we can assume that @xmath49 .
we therefore plot in fig .
3 , @xmath50 versus @xmath30 and find a straight line plot with @xmath3 in two dimensional case .
this compares well with the previous estimates of the value of @xmath51 [ 6 ] .
as the growth of the height of magnetisation response ( and its maximum value @xmath21 ) is very much bounded by the time window @xmath5 of the applied field pulse , the anomalous behaviour of @xmath1 ( having finite peak at a shifted temperature @xmath45 ) may be considered to be due to the finite size ( in time ) effect .
similar to the finite size ( in length ) scaling theory of fisher [ 5 ] , where a finite size system shows effective ( non - singular or non - divergent ) pseudo - critical behaviour at @xmath52 when the correlation length @xmath17 becomes of the order of the system size @xmath53 , we suggest a finite time @xmath54 scaling behaviour here for @xmath1 : if the relaxation time @xmath55 , where @xmath19 is the correlation length exponent and @xmath18 is the dynamic exponent , then @xmath1 would show peak at the temperature @xmath4 here when @xmath56 or @xmath57 , or @xmath58 where @xmath59 is some constant .
in fact , fig .
4 shows that the effective peak position @xmath4 indeed approaches @xmath2 as @xmath23 .
the inset shows the plot of @xmath4 with @xmath60 , which gives a straight line for @xmath7 .
this again suggests @xmath3 and also the extrapolated value of @xmath2 becomes about 2.29 , which compares well with the onsager value , comparable to the previous estimate .
we have solved the meanfield equation for response magnetisation [ mfe ] using fourth order runge - kutta method .
1 shows the typical variation of response magnetisation and field . here
also we have measured the width - ratio and the pulse susceptibility and studied the temperature variation of these two quantities .
fig.5 shows the temperature variation of @xmath0 and @xmath1 for different values of @xmath5 .
@xmath0 diverges and @xmath1 peaks at the same order - disorder transition point ( @xmath61 here ) .
we have also studied the variation of the maximum value of @xmath1 ( @xmath62 at @xmath63 ) with respect to the duration of the pulsed field .
it may also be mentioned that the peak height was found to increase with increasing pulse width @xmath5 , and @xmath1 diverges as @xmath23 : @xmath64 fig .
6 shows the variation of @xmath65 with @xmath5 . in order to comprehend these observations , we solve the equation [ mfe ] in a linearised limit ( large @xmath30 and small @xmath22 ; specifically @xmath66 , @xmath34 ) . in such limit ,
the equation of motion becomes @xmath67 one can solve the above equation using a form @xmath68 which gives @xmath69 giving @xmath70 , and @xmath71 integrating the last equation for @xmath72 for a finite time width @xmath5 and @xmath73 elsewhere , one gets @xmath74 at @xmath75 .
this gives @xmath76 we have checked the above linear relationship of @xmath62 with @xmath5 for extremely small values of pulse field amplitude @xmath22 ( see fig .
we have studied the glauber ( order parameter nonconserving ) dynamics of an ising system under a time varying external magnetic field , when the field is applied as a pulse of finite time width , after the system reaches the equilibrium .
the time variation of the response magnetisation is studied as a function of pulse width @xmath5 , height @xmath22 and temperature @xmath30 of the system .
we have measured specifically @xmath77 and @xmath78 , where @xmath20 is the time width of the response magnetisation and @xmath21 is the maximum height of the response magnetisation above its equilibrium value .
our computer simulation results for square lattice showed @xmath79 with @xmath80 , the onsagar value , and @xmath1 has a peak @xmath62 at @xmath81 , such that a finite size ( in time ) scaling behaviour is observed : @xmath82 , with @xmath3 .
the numerical solutions of the mean field equation [ mfe ] showed @xmath83 and the pulse susceptibility peak value @xmath84 , occuring at @xmath85 .
theoretical analysis for the finite size ( in time ) scaling behaviour ( of @xmath4 in the monte carlo case ) and for the peak @xmath86 ( in the mean field case ) are also given .
ma acknowledges jncasr for financial support and serc , iisc bangalore for computational facilities .
e-mail:[email protected] see e.g. , k. binder , ed .
, _ application of the monte carlo method in statistical physics _ , springer , heidelberg ( 1984 ) ; d. stauffer , ed . , _ annual reviews of computational physics _ , vol 1 , world - scientific , singapore ( 1994 ) m. rao , h. r. krishnamurthy and r. pandit , phys
b * 42 * 856 ( 1990 ) ; d. dhar and p. b. thomas , j. phys .
a * 25 * 4967 ( 1992 ) m. acharyya and b. k. chakrabarti , phys .
b * 52 * 6550 ( 1995 ) , and references therein m. acharyya and b. k. chakrabarti , physica a * 224 * 254 ( 1996 ) see e.g. , m. n. barber in _ phase transition and critical phenomena _ , vol 8 , eds . c. domb and m. green , academic press , n.y .
, pp 203 ( 1980 ) j. c. angles d auriac et al , j. stat . phys .
* 28 * 307 ( 1982 ) fig.2
. temperature variations of ( a ) @xmath0 and ( b ) @xmath1 for different values of @xmath5 in the monte carlo case .
the symbol circle is for @xmath5 = 5 ; the square is for @xmath5 = 10 and the cross is for @xmath5 = 25 .
fig.5 temperature variation of ( a ) @xmath0 and ( b ) @xmath1 for different values of @xmath5 in the mean field case .
the symbol triangle is for @xmath5 = 8 ; the plus is for @xmath5 = 16 and the cross is for @xmath5 = 32 . | the dynamical response to a pulsed magnetic field has been studied here both using monte carlo simulation and by solving numerically the meanfield dynamical equation of motion for the ising model .
the ratio @xmath0 of the response magnetisation half - width to the width of the external field pulse has been observed to diverge and pulse susceptibility @xmath1 ( ratio of the response magnetisation peak height and the pulse height ) gives a peak near the order - disorder transition temperature @xmath2 ( for the unperturbed system ) .
the monte carlo results for ising system on square lattice show that @xmath0 diverges at @xmath2 , with the exponent @xmath3 , while @xmath1 shows a peak at @xmath4 , which is a function of the field pulse width @xmath5 . a finite size ( in time ) scaling analysis shows that @xmath6 , with @xmath7 .
the meanfield results show that both the divergence of @xmath8 and the peak in @xmath1 occur at the meanfield transition temperature , while the peak height in @xmath9 , @xmath10 for small values of @xmath5 .
these results also compare well with an approximate analytical solution of the meanfield equation of motion . |
the rotational properties of spiral galaxies are the original experimental evidence@xcite that dark matter must exist .
subsequently , modern cosmology in the form of cold dark matter requires dark matter to be a large contributor to the energy density of the universe .
if neutrinos are dirac - type ( not majorana ) , they must posses a magnetic dipole moment , and it was shown@xcite that the instantaneous relativistic radiation loss for a magnetic moment @xmath10 is @xmath11\ ] ] where @xmath12 .
this is a higher power of @xmath13 than the familiar linard formula for charged particles q @xmath14 \ ; .\ ] ] only in turbulent , chaotic plasmas is this dipole radiation loss realizable .
however , in the early universe , large , chaotic , and turbulent fields are expected@xcite .
if the neutrinos lose energy in this manner , they are anticipated to condense and form degenerate neutrino matter .
if this happens , the cno are the largest and most massive objects in the universe@xcite and become the ` dark matter ' , which everybody is looking for . in this paper ,
we take the three spiral galaxies of the local group and use their rotational data to elucidate the properties of the cno which appears to embed them .
m33 , m31 and the milky way are the only spiral galaxies of the local group .
the rotational data of m33@xcite and m31@xcite possess a unique property : they have no noticable azimuth angle dependence .
the physics implication is that for a given fixed radial distance from the galaxy s center , the 360 degree rotation circle is a equipotential surface . for spherical dark matter , such as a cno
, this can only happen if the spin axis of the galaxy is aligned radially in the embedding dark matter sphere .
having two spiral galaxies with no noticable azimuth angle dependence in their rotation curves allows us to determine the center of the cno by extrapolating the two spin axis back to a point , figure 1 . in practice
, they will not actually cross , but join within an error volume that will be calculable .
the milky way has azimuth angle dependence in its rotation curve , because of the wide variation@xcite of the rotational speed with fixed radial distance .
this means that the milky way spin axis is ` canted ' in the cno spherical symmetry . in table 1
, we list the j2000 equatorial coordinates of m33 and m31 . to find the center of the cno embedding the lg
, we extrapolate the m31 and m33 spin axis using a monte carlo algorithm which is described in the appendix .
the spin axis of a spiral galaxy is described in reference@xcite .
the spin axis unit - vector @xmath15 of m31 and m33 is given by @xmath16 with @xmath17 where @xmath18 is the declination angle and @xmath19 is the right ascension angle of the galaxy in question .
the physical components are given by observables @xmath20 , @xmath21 and @xmath22 , where @xmath23 is the axial ratio and @xmath24 , where @xmath25 is the position angle , with @xmath26 .
we use a computer algorithm ( see appendix ) that finds the smallest extrapolated m33 and m31 spin - axis intersection for values of @xmath23 and @xmath27 within @xmath289 % of their quoted values within the literature .
this will reveal the center of the cno to within a computable error .
cccc + name & distance ( kpc ) & dec ( degree ) & ra ( degree ) + m31 & 788.333 & 41.2689 & 10.6846 + m33 & 862.417 & 30.6581 & 23.4662 + milky way & 7.61113 & -29.0078 & 266.417 cccc + quantity & derived value & quoted value & reference + m33 axial ratio & 0.653903 & 0.615 & reference@xcite + m33 position angle & 22.1965@xmath29 & @xmath30 & reference@xcite + m31 axial ratio & 0.575201 & .6 & reference@xcite + m31 positional angle & 40.1856@xmath29 & 38@xmath29 & reference@xcite in table 2 , we present the monte carlo s algorithm s estimate of the m33 , m31 axial ratio and position angles that gives a small error in their extrapolated intersection . . ]
this spin axis extrapolation will reveal the approximate center of the cno which is given in table 3 ; there , the cno is designated cno - lg .
the size and mass of the cno are determined by three parameters : the degenerate neutrino mass @xmath5 and two boundary conditions@xcite at the center , of which the only one that varies is @xmath31 where @xmath31 is the quantity @xmath32 at the center , with @xmath33 the fermi momentum and @xmath34 the speed of light . in general , @xmath35 from fitting galaxy clusters embedded in cno@xcite .
m33 has the best rotational data@xcite because that reference includes an estimate for the contribution of m33 s own mass . at about 15 kpc radial distance , the measured @xmath36 100 km
/ s , of which @xmath37 40 - 50 km / s is the dark matter contribution .
we now explain how this identifies the parameters of the cno that embeds the local group .
ccc + quantity & predicted value & unit + center cno - lg distance from earth & 675.933 & kpc + center cno - lg distance from m33 & 740.423 & kpc + center cno - lg distance from m31 & 656.015 & kpc + algorithm error center cno - lg & .706162e-3 & kpc + right ascension of cno - lg center & -26.6588 & @xmath29 + declination of cno - lg center & 0.91773 & @xmath29 + galactic longitude of cno - lg center & 62.83928785 & @xmath29 + galactic latitude of cno - lg center & -42.77834848 & @xmath29 + milky way ca nt angle & 47.221 & @xmath29 the cno dark matter contribution to the rotational speeds of a galaxy embedded in the cno is @xmath38 where @xmath39 is the distance of the galaxy s center from the cno center , and @xmath40 is the distance from the cno center to the rotational arm of the galaxy .
the notation @xmath41 means the centripetal speed is relative to the galaxy s center .
the m33 data requires this to be @xmath42 40 km / s at 15 kpc from m33 center , perpendicular to the radial direction . in this equation
, @xmath43 is the cno enclosed mass at radial coordinate r. it can only be computed numerically@xcite by solving the hydro - static equation of equilibrium using the neutrino equation of state . to work out @xmath44
, we use figure 2 . in this cno solution ,
m33 center is at a radial distance 740.423 kpc , while its 15 kpc spiral arm is at 740.575 kpc .
the difference in rotation speeds is given by eq(5 ) where @xmath40 = 740.575 kpc and @xmath45 = 740.423 kpc . by requiring this to be @xmath42 40 km / s at 15 kpc from m33 center , perpendicular to the radial direction , we determine the properties of the cno embedding the lg . for @xmath46 with @xmath47 0.8 ev / c@xmath9
, we get the values in table 4 .
the value for the milky way galaxy is consistent with the highest measured values@xcite , but they include an unknown mass contribution .
the mathematical reason why the milky way has a higher rotational speed at 8 kpc than m33 at 15 kpc is due to the non - zero ca nt angle of milky way .
assuming that the monte carlo solution @xmath46 with @xmath47 0.8 ev / c@xmath9 is reasonably close to the physical answer , we now can give the mass and radius of this cno embedding the lg , table 5 .
ccc + galaxy & spiral arm distance from its center & @xmath44 + m33 & 15 kpc & @xmath48 ( 40 km / s)@xmath9 + milky way & 8 kpc & @xmath48 ( 260 km / s)@xmath9 ccc + quantity & value & unit + cno - lg boundary condition & @xmath46 & @xmath32 + degenerate neutrino mass @xmath5 & @xmath49 & ev / c@xmath9 + cno - lg radius & 1.316 & mpc + cno - lg mass & 5.17@xmath6 & m@xmath7 a galaxy embedded in a cno has a gravitational force pulling it in the direction of the cno center
. eventually the galaxy will undergo simple harmonic motion about the center@xcite .
those spiral galaxies that have a non - zero ca nt angle are also expected to cartwheel about the arm closest to the cno center .
if a cluster of galaxies is embedded , the center of mass of the cluster is itself expected to undergo simple harmonic motion .
the back - reaction on the cno will cause it to vibrate in its spectrum of normal modes@xcite with the lowest frequency a quadrupole oscillation . in figure 3 , we show the 3-d geometry of the calculated cno solution in relationship to the local group .
it should be pointed out that cno can not overlap each other ( pauli exclusion principle ) .
hence , it is possible that cnos of various radii and masses could be adjacent to one another ( like kepler s stacked greengrocer oranges ) . in reference@xcite , we point out that condensed neutrinos might even form filaments between quasi - spherically shaped cnos ( bi - spherical symmetry ) .
the value of @xmath47 0.8 ev / c@xmath9 violates the upper - bound limit of neutrinos published by the planck satellite consortium@xcite , based on their assumption that neutrinos ( and anti - neutrinos ) are in thermodynamic equilibrium in the early big bang , before their decoupling of matter .
this thermodynamic equilibrium assumption is conventional wisdom@xcite , which does not take into account possible neutrino ( and anti - neutrino ) radiation losses .
the planck satellite analysis produces other issues : the discrepancy between the hubble expansion parameter as indirectly measured by the planck satellite consortium and the wilkinson microwave anisotropy probe , with its direct determination by type 1a supernova@xcite .
this tension / disagreement can no longer be ignored because the discrepancy is at the 5%-9% level@xcite .
what is going on ?
the key here is this assumption that the cosmological black - body radiation ( cbr ) scientists make in reducing their raw data .
they assume that the cosmological neutrinos are in thermodynamic equilibrium with baryonic matter right up to their decoupling and that they contibute to the perturbations on the acoustic peaks . in so doing , they constrain neutrino masses to be extremely small and they derive a small value of the hubble expansion parameter .
in fact , if neutrinos radiate their energy in the early universe turbulent plasmas , they will never be in equilibrium with baryonic matter .
if so , something else has to be present giving their degrees - of - freedom to the cbr .
actually there is : adams et al @xcite show that a cosmic magnetic field has the same contribution to the cosmic microwave background as kinetic degrees of freedom .
we anticipate that if the raw cbr data is re - processed using cosmic magnetic fields in place of neutrino ones , the discrepancy of the value of the hubble parameter should disappear .
whatever dark matter is , it must have an equation of state ( eos ) .
this is the key physics describing the aggregate of dark matter particles .
a critical requirement is that the eos allows for stable dark matter .
stability in this context means that it must be stable against its own self - gravity .
some internal pressure must exist to withstand its own gravitational collapse .
if dark matter is made up of bosons , it must have an internal energy source where the kinetic energy of the bosons prevents collapse .
no one has shown that candidate bosons have internal energy generation that can prevent the bosons from collapsing into a black hole configuration .
bosons are poor candidates for dark matter for this very reason and probably impossible for cold dark matter .
fermions , on the other hand , have degeneracy pressure because of the pauli exclusion principle . we know that electron degeneracy pressure allows white dwarfs to exist , and neutron degeneracy pressure allows neutron stars to exist .
if neutrinos radiate , they should condense and form stable objects .
if cold dark matter is in fact a condensed object composed of neutrinos , then there are constraints that can be immediately deduced .
one constraint is that these neutrinos must be cosmological in origin , since the number of neutrinos produced after the big bang is too small .
next , neutrino condensation requires them to loose their kinetic energy . in eq(1 ) above , we have a natural way to do this since the early universe had huge turbulent plasmas @xcite .
however , majorana neutrinos , have no permanent magnetic dipole moment , and so eq(1 ) is not available to them .
we conclude that if dark matter are objects made of neutrinos , they are of the dirac - variety .
most importantly , the normal set of neutrinos and their anti - neutrinos ( @xmath50 ) interact so weakly , that whatever numbers were produced in the big bang , those numbers still remain and the baryon matter - anti - matter disparity mystery does nt exist for neutrinos and their antineutrinos .
thus cno satisfy the large difference between the baryonic contribution ( 17% ) to the universe matter density and the dark matter contribution ( 83%)@xcite .
the katrin experiment @xcite will have the sensitivity to determine the mass of the electron anti - neutrino down to @xmath51 , well within the value of @xmath5 used here .
the mass range for the degenerate neutrino mass 0.75 ev / c@xmath52 ev / c@xmath9 found from fitting galaxy cluster data@xcite is in direct contradiction to the upper bound claimed by the planck satellite consortium .
if katrin discovers a neutrino mass in this range , we contend that the cbr raw data analysis must be revisited and that it would be a major finding endorsing condensed neutrinos as the so - called dark matter , which everyone has been looking for .
we describe the algorithm that finds the approximate cno center for two radially - aligned spiral galaxies .
let @xmath53 and @xmath54 be the vectors describing the position ( @xmath55 ) of spiral galaxy numbered 1 and a line through its spin axis ( @xmath56 ) , with spiral galaxy numbered 2 similiar .
the squared - distance between them @xmath57 is @xmath58^{2}\ ] ] we find the extremum @xmath59 , @xmath60 by solving for @xmath61 which gives the solutions @xmath62 } { [ ( \vec{a } \cdot \vec{b})^{2 } -1]}\ ] ] @xmath63 for unit spin vectors @xmath64 .
we vary the axial ratio and position angle of each spiral galaxy randomly within @xmath289 % of their quoted values within the literature , determine the separation distance between the lines that those probabilty draws create and either keep them as the running smallest value or throw them away before the next probability draw series .
a loop of 4 million series draws is sufficient for the approximate answer .
99 p. d. morley and d. j. buettner , _ ijmpd _
( 2014 ) doi : 10.1142/s0218271815500042 .
d. morley and d. j. buettner , _ ijmpd _ ( 2016 ) doi : 10.1142/s0218271816500899 .
https://www.katrin.kit.edu f. zwicky , reprinted in _ gen .
( 2009 ) doi : 10.1007/s10714 - 008 - 0707 - 4 .
joel / primack - lect1-lcdm_history.pdf p. d. morley and d.j .
buettner , _ astroparticle physics _ ( 2015 ) doi : 10.1016/j.astropartphys.2014.07.005 .
d. grasso and h. r. rubinstein , _ phys .
rep _ * 348 * ( 2001 ) 163 .
jacques m. wagstaff et al . , _
arxiv_:1304.4723v1 [ astro-ph.co ] edvige corbelli and paolo salucci , _ mnras _ ( 2000 ) doi : 10.1046/j.1365 - 8711.2000.03075.x .
claude carignan et al .
* 641 * ( 2006 ) l109 .
xiao - sheng xin and xing - wu zheng , _ res .
_ * 13 * ( 2013 ) 849 . arthur n. cox editor , 4th edition ( springer , 2013 ) .
pen , j. lee and u. seljak , _ apj _ * 543 * ( 2000 ) l107 .
g. de vaucouleurs , _ apj _ * 130 * ( 1959 ) 728 .
p. j warner , m. c. h. wright and j. e. baldwin , _ mnras _ * 163 * ( 1973 ) .
robin ciardullo et al . , _ ap.j . _
* 318 * ( 1987 ) 520 .
g. de vaucouleurs , _ ap.j . _
* 128 * ( 1958 ) 465 .
planck consortium in _
planck 2015 results .
cosmological parameters _ arxiv:1502.01589 .
edward w. kolb and michael s. turner in _ the early universe _
( addison - wesley , 1990 ) 94 .
news release : hubblesite.org / newscenter / archive / releases/2016/17/text/. j. adams et al . , _ phys
_ * b388 * ( 1996 ) 253 . hyperphysics.phy-astr.gsu.edu/hbase/astro/denpar.html . | we describe the neutrino flavor ( e = electron , @xmath0 = muon , @xmath1 = tau ) masses as @xmath2 with @xmath3 and probably @xmath4 .
the quantity @xmath5 is the degenerate neutrino mass .
because neutrino flavor is not a quantum number , this degenerate mass appears in the neutrino equation of state@xcite .
we apply a monte carlo computational physics technique to the local group ( lg ) of galaxies to determine an approximate location for a dark matter embedding condensed neutrino object@xcite ( cno ) .
the calculation is based on the rotational properties of the only spiral galaxies within the lg : m31 , m33 and the milky way .
cnos could be the dark matter everyone is looking for and we estimate the cno embedding the lg to have a mass 5.17@xmath6 m@xmath7 and a radius 1.316 mpc , with the estimated value of @xmath8 ev / c@xmath9 .
the up - coming katrin experiment @xcite will either be the definitive result or eliminate condensed neutrinos as a dark matter candidate .
dark matter in the local group of galaxies p.d .
morley + system of systems analytics , inc .
+ 11250 waples mill road + suite 300 + fairfax , va 22030 - 7400 + _ e - mail address : _ [email protected] + and + d.
j. buettner + aerospace corporation + p. o. box 92957 + los angeles , ca 90009 - 2957 + _ e - mail address : _
[email protected] _ keywords _ : dark matter ; neutrinos ; cosmology .
pacs number(s ) : 95.35.+d , 95.30.cq , 98.35.gi , 98.80.cq |
racks are algebraic structures with axioms derived from reidemeister moves type ii and type iii . they have been considered by knot theorists in order to construct knot and link invariants and their higher analogues
( see for example @xcite and references therein ) .
racks allow a refined and a complete algebraic framework in which ones investigates links and @xmath4-manifolds .
they have been studied by many authors and appeared in the literature with different names such as automorphic sets and in a special case quandles , distributive groupoids , crystals etc .
rack cohomology was introduced by fenn , rourke and sanderson @xcite . for each rack @xmath5 and abelian
group @xmath6 , they defined cohomology groups @xmath7 . since then there has been a number of results about this cohomology ( see @xcite , @xcite and @xcite ) with studies from different perspectives .
a modification of rack cohomology theory led to quandle cohomology theory which was developed in @xcite in order to define invariants of classical knots and knotted surfaces in state - sum form , called quandle cocycle ( knot ) invariants .
these invariants can be understood as enhancements of the quandle counting invariant , where quandle colorings of a link diagram are counted with a weight determined by a quandle cocycle . in @xcite ,
the quandle homology theory was generalized to the case when the coefficient groups admit the structure of alexander quandles , by including an action of the infinite cyclic group in the boundary operator . using generalizations of quandle homology theory provided by andruskiewitsch and graa @xcite ,
the quandle cocycle invariants were generalized in three different directions in @xcite .
the second author has studied several enhancements of quandle counting invariants of a link @xmath8 with respect to a finite target quandle @xmath9 ( see for example @xcite and @xcite ) .
+ the paper is organized as follows . after reviewing the basics of racks and recalling the notion of rank for racks in section 2 , rack homology
is considered in section 3 . as our main result
, we prove that the degenerate chains form a sub - complex of the classical complex defining rack homology and define a homology theory analogous to quandle homology for non - quandle racks which reduces to the usual quandle homology when the rack is a quandle . in section 4
we use cocycles in @xmath10 to enhance the rack counting invariant , obtaining a family of link invariants which generalize the cjkls invariants to allow non - quandle racks . note
that while a related invariant was defined in @xcite , the fact that @xmath0-degenerate chains form a subcomplex was not proved in @xcite , only invariance under blackboard framed isotopy and the @xmath0-phone cord move . in section 5
explicit reduced cocycles are given and used to perform computations of the invariants for prime knots and links .
we end the paper with section 6 in which we make some remarks and suggest some open questions for future investigations .
we begin with a definition from @xcite , give examples and then recall from @xcite the notion of rank for racks .
@xmath11 @xmath12
. [ ex1 ] @xmath13 @xmath14 @xmath15 @xmath16 @xmath17 @xmath18 the rack operations are generally non - associative , with right self - distributivity taking the place of associativity .
it is a standard exercise to show that in any rack @xmath5 we also have @xmath19 @xmath20 , @xmath21 , and @xmath22 . to minimize parentheses
, any rack word not containing parentheses will be associated left - to - right , so that @xmath23 in particular , the expression @xmath24 is an abbreviation for @xmath25 where we have @xmath26 total @xmath27s .
a similar notion is a _ rack power _
@xmath28 , defined recursively by the rules @xmath29 and @xmath30 . the map @xmath31 given by @xmath32 is a bijection known as a _ kink map _
; we have @xmath33 . given any @xmath34 , we can ask what is the minimal @xmath35 such that @xmath36 . @xmath37 note
that a quandle is a rack with rack rank @xmath38
. we will often write @xmath0 in place of @xmath39 when the rack in question is understood .
the rack rank is analogous to the characteristic of a field or ring ; indeed , we might consider the alternative term `` rack characteristic . '' for ease of computation
, a finite rack @xmath40 can be represented with a _ rack matrix _ @xmath41
whose @xmath42 entry is @xmath43 where @xmath44 .
that is , the rack matrix encodes the operation table of @xmath45 .
note that the operation table of @xmath46 can be recovered from @xmath41 , so we do not need to specify both operation tables to determine a rack structure on @xmath5 . as observed in @xcite ,
every finite rack has finite rack rank equal to the order in @xmath47 of the permutation @xmath48 given by the diagonal of the rack matrix .
[ example3 ] @xmath49\ ] ] every framed oriented link @xmath8 has a _ fundamental rack _ @xmath50 with generators corresponding to arcs in a blackboard - framed diagram of @xmath8 and relations determined at crossings as pictured : @xmath51 @xmath52 the rack axioms encode blackboard - framed isotopy , so that two link diagrams which are framed isotopic have isomorphic fundamental racks .
indeed , in @xcite it is shown that the fundamental rack is a complete invariant of unsplit oriented links in homology 3-spheres .
@xmath53 for any rack @xmath5 , a rack homomorphism @xmath54 can be visualized as a _ coloring _ or labeling of a diagram @xmath8 by elements of @xmath5 where each arc , say @xmath55 , in @xmath8 gets a label @xmath56 .
the set of homomorphisms @xmath57 is invariant under blackboard - framed isotopy . by summing these numbers of colorings over a complete set of framings modulo @xmath0
, we obtain an invariant of ambient isotopy ( @xcite ) .
let @xmath5 be a finite rack with rack rank @xmath0 , @xmath58 a link with @xmath59 components , @xmath60 , and @xmath61 the fundamental rack of @xmath8 with framing vector @xmath62 .
then the sum @xmath63 is an invariant of links , known as the _ integral rack counting invariant_. the theorem follows from the observation that while colorings of link diagrams by racks are not preserved under reidemeister i moves , they are preserved by the _ @xmath0-phone cord move _ where @xmath0 is the rack rank of @xmath5 : @xmath64 thus , if two links are equivalent by blackboard - framed moves and @xmath0-phone cord moves , their sets of colorings by a rack @xmath5 with rack rank @xmath0 are in one - to - one correspondence .
we can compute the set of rack colorings of a link @xmath8 by putting the link in braid form , assigning a generator to each strand at the top and pushing the colors down the braid ; closing the braid , we obtain a system of equations with one equation for each strand .
we must repeat this computation with @xmath0 stabilization moves on each component to get the total set of colorings .
@xmath65 @xmath66
let @xmath5 be a finite rack with rack rank @xmath0 and let @xmath67 be the free abelian group generated by @xmath26-tuples @xmath68 of elements of @xmath5 ; for @xmath69 set @xmath70 .
recall that for an @xmath26-chain , the differential @xmath71 ( see @xcite for example ) is defined on the generator @xmath72 by @xmath73\end{aligned}\ ] ] for @xmath74 and @xmath75 for @xmath76 and extended to @xmath77 by linearity .
[ ndegenerate ] @xmath78 we now come to our main result : let @xmath5 be a finite rack with rack rank @xmath0 . then the @xmath0-degenerate @xmath26-chains form a subcomplex of the complex @xmath79 .
let @xmath80 one checks that the @xmath81-chain @xmath82 is @xmath0-degenerate : @xmath83 % } \nonumber \\ & = & \sum_{k=1}^n \partial_{n } ( x_1,\dots , x_{i-1},x_i^{\tr k } , x_i^{\tr ( k+1 ) } , x_{i+2 } , \dots , x_n ) \nonumber \\ & = & \sum_{k=1}^n \left\ { \sum_{j=2}^{i-1 } ( -1)^{j } [ ( x_1,\dots , x_{j-1 } , x_{j+1},\dots , x_i^{\tr k } , x_i^{\tr ( k+1 ) } , x_{i+2 } , \dots , x_n ) \right . \nonumber \\ & & \quad \quad \left . - ( x_1 \tr x_j,\dots , x_{j-1 } \tr x_j , x_{j+1},\dots , x_i^{\tr k } , x_i^{\tr ( k+1 ) } , x_{i+2 } , \dots , x_n ) ] \right\ } \nonumber \\ & & + \sum_{k=1}^n \left\ { ( -1)^i\left [ ( x_1,\dots , x_{i-1 } , x_i^{\tr ( k+1 ) } , x_{i+2 } , \dots , x_n ) \right .
\nonumber \\ & & \quad \quad - \left .
( x_1 \tr x_i^{\tr k } , \dots , x_{i-1 } \tr x_i^{\tr k } , x_i^{\tr ( k+1 ) } , x_{i+2 } , \dots , x_n)\right ] \right .
\nonumber \\ & & \quad\quad + ( -1)^{i+1}\left [ ( x_1,\dots , x_{i-1 } , x_i^{\tr k } , x_{i+2 } , \dots , x_n ) \right .
\nonumber \\ & & \quad\quad - \left .
\left.(x_1 \tr x_i^{\tr ( k+1 ) } , \dots , x_{i-1 } \tr x_i^{\tr ( k+1 ) } , x_{i}^{\tr k } \tr x_i^{\tr ( k+1 ) } , x_{i+2 } , \dots , x_n)\right ] \right\ } \nonumber \\ & & + \sum_{k=1}^n \left\{\sum_{j = i+2}^{n } ( -1)^{j } \left [ ( x_1,\dots , x_{i-1},x_i^{\tr k } , x_i^{\tr ( k+1 ) } , \dots , x_{j-1 } , x_{j+1},\dots , x_n ) \right .
\nonumber \\ & & \quad \quad \left.\left .
- ( x_1 \tr x_j,\dots , x_{i-1 } \tr x_j , x_i^{\tr k } \tr x_j , x_i^{\tr ( k+1 ) } \tr x_j , \dots , x_{j-1 } \tr x_j , x_{j+1 } , \dots , x_n ) \right ] \right\}.\end{aligned}\ ] ] one observes that the following sum vanishes : @xmath84 -\right.\ ] ] @xmath85 \right\ } \\\ ] ] because @xmath86 and thus @xmath87=0\ ] ] and @xmath88 this last difference is zero since by [ 15 , lemma 1 ] we have @xmath89 and similarly @xmath90 . using self - distributivity , the last term in equation ( [ eq2 ] ) can be re - written as ( so it makes the last sum fits in the definition of degenerate chains ) @xmath91 recall from @xcite that @xmath92 is a chain complex .
let @xmath93 be the subset of @xmath94 generated by the @xmath0-degenerate @xmath26-chains as in definition [ ndegenerate ] if @xmath74 ; otherwise let @xmath95 .
if @xmath5 is a rack , then @xmath96 and @xmath97 is a sub - complex of @xmath98 . put @xmath99 and @xmath100 , where @xmath101 is the induced homomorphism .
henceforth , all boundary maps will be denoted by @xmath102 . for an abelian group @xmath103 , define the chain and cochain complexes @xmath104 in the usual way , where @xmath105 @xmath106 , @xmath107 , @xmath108 .
@xmath109 as an example , let us now compute the first and second homology groups of the @xmath110-rack @xmath111 with @xmath112 and @xmath113 as in example [ example3 ] .
we have the following let @xmath111 with @xmath114 , the rack in example [ example3 ] .
the group of @xmath115-degenerate 2-chains @xmath116 is generated by @xmath117 now since @xmath114 and @xmath118 , the image group , @xmath119 , is two - dimensional , generated by two generators @xmath120 and @xmath121 .
this implies the following the groups @xmath122 and @xmath123 are two - dimensional .
the group @xmath124 is 14-dimensional .
a straightforward computation gives the following fourteen generators @xmath125 , @xmath126 , @xmath127 for @xmath128 and @xmath129 for @xmath130 of the group @xmath124 .
now we calculate the dimension of the second homology group .
the group @xmath131 is 11-dimensional , giving dim @xmath132 .
again a straightforward computation gives the following eleven generators @xmath133
in this section we use cocycles in @xmath134 to define an enhancement of the rack counting invariant .
let @xmath8 be a link diagram with @xmath59 components , @xmath5 a finite rack with rack rank @xmath0 , @xmath60 and @xmath135 a nondegenerate 2-cocycle with coefficients in an abelian group @xmath103 .
for a given coloring @xmath136 , we define the _ boltzmann weight _ of @xmath137 in the following way . at each positive
crossing in @xmath8 , we have a contribution of @xmath138 where @xmath55 is the color on the inbound underarc and @xmath139 is the color on the overarc ; at each negative crossing , we have a contribution of @xmath140 where @xmath55 is the color on the outbound underarc and @xmath139 is the color on the overarc .
the boltzmann weight @xmath141 is then the sum over all crossings in @xmath8 of these contributions .
@xmath142 the rack cocycle condition is precisely the condition required so that the boltzmann weight of a rack - colored diagram is unchanged by reidemeister iii moves ; the coloring condition is chosen to guarantee invariance under reidemeister ii and blackboard - framed type i moves .
@xmath143 @xmath144 the @xmath0-degenerate subcomplex is generated by the characteristic chains of @xmath0-phone cord tangles . by setting these equal to zero ,
i.e. selecting @xmath145 as opposed to @xmath146 , we obtain invariance under the @xmath0-phone cord move .
thus , we have for any cocycle @xmath147 , the multiset @xmath148 is an invariant of ambient isotopy .
we can also use generating function style notation to rewrite the multiset as a one - variable polynomial with @xmath149 coefficients and @xmath103 exponents : @xmath150 note that there are four oriented @xmath0-phone cord moves ; we have used only one to define our degenerate subcomplex .
one of the other moves yields the same degenerate subcomplex , but the other two yield a slightly different subcomplex , generated by chains of the form @xmath151 note that these chains are called @xmath0-_reduced _ in @xcite .
the phone cord moves which yield this degeneracy condition are related to the one we chose through a combination of reidemeister ii and iii moves together with blackboard - framed i moves , so it follows that such alternative degenerate cocycles are homologous to cocycles in @xmath152 .
in particular , we have : [ phixx ] first note that since @xmath153 is a 2-cocycle we have @xmath154 for all @xmath155 . then in particular , setting @xmath156 , @xmath157 and @xmath158 we have @xmath159 let @xmath5 be a rack of rack rank @xmath160 and let @xmath161 be the set of linear combinations of chains of the form @xmath151 then @xmath162 is also a subcomplex of @xmath163 and the quotient complexes @xmath164 and @xmath165 are isomorphic .
in this section we collect a few examples of the rack cocycle invariants .
these examples were computed using ` python ` code available at ` www.esotericka.org ` with signed gauss codes transcribed by hand from diagrams at the knot atlas @xcite and checked with ` maple ` . @xmath167 @xmath168 @xmath169 the previous examples both use racks which are disjoint unions of one rack and one quandle , with each orbit acting trivially on the other . for our final example
, we use a more `` pure rack '' with no elements of rack rank @xmath170 .
@xmath171\ ] ] @xmath172 @xmath173
we conclude the paper by the following remarks and open questions : in @xcite a homology theory was developed for set - theoretic yang - baxter equations and used to define invariants of classical and virtual knots .
this approach was used in @xcite and extended to detect non - classicality of some virtual links . within these lines
generalize the cohomology theory in this paper to biracks and use it study invariants of classical and virtual knots .
our computer experiments suggest that small cardinality racks define 2-cocycle invariants which are stronger on links than on knots , due to the orbit subracks being fairly trivial constant action racks .
fast algorithms for computing the homology of large - cardinality racks should improve the practical utility of the rack cocycle invariants .
given any rack @xmath5 , there are a number of quandles associated to @xmath5 such as the _ maximal subquandle _ of @xmath5 consisting of all elements of @xmath5 of rack rank @xmath176 ( this may be the empty quandle ) and the _ operator quandle _ obtained by taking the quotient of @xmath5 by the congruence
@xmath177 what relationship , if any , can be found between the quandle homology of these quandles and the @xmath175 homology of @xmath5 ?
department of mathematics , + university of south florida , + 4202 e fowler ave . , + tampa , fl 33620 + + and + + department of mathematics , + claremont mckenna college , + 850 columbia ave . , + claremont , ca 91711 | the aim of this paper is to define a homology theory for racks with finite rank @xmath0 and use it to define invariants of knots generalizing the cjkls 2-cocycle invariants related to the invariants defined in @xcite . for this purpose ,
we prove that @xmath0-degenerate chains form a sub - complex of the classical complex defining rack homology .
if a rack has rack rank @xmath1 then it is a quandle and our homology theory coincides with the ckjls homology theory @xcite .
nontrivial cocycles are used to define invariants of knots and examples of calculations for classical knots with up to @xmath2 crossings and classical links with up to @xmath3 crossings are provided .
keywords : finite racks , rack homology , enhancements of counting invariants , cocycle invariants 2000 msc : 57m27 , 57m25 |
x - ray novae ( xn ) provide the strongest evidence for the existence of stellar mass black holes . xn are interacting binaries where a neutron star or a black hole accretes matter from a companion ( usually a dwarf - like k or m star ) . the accretion rate on to the black hole increases substantially during the outburst phase , and
hence the x - ray luminosity can vary by large amounts ( e.g. factors of @xmath9 or more ) .
the majority of xn spend most of their time in a `` quiescent '' state where the x - ray luminosity is on the order of the optical luminosity of the companion star .
it is in quiescence where the companion star can be best studied in the optical , where the observed radial velocity and light curves of the companion lead to dynamical mass measurements for the compact primary .
there are 18 x - ray binaries ( 15 of them xn ) where the mass of the primary has been shown to exceed the maximum mass of a stable neutron star ( @xmath10 , kalogera & baym 1996 ) , confirming the presence of black holes in these systems : gro j0422 + 32 ( orosz & bailyn 1995 , filippenko , matheson , & ho 1995 ) ; a0620 - 00 ( mcclintock & remillard 1986 ) ; grs 1009 - 45 ( filippenko et al .
1999 ) ; xte j1118 + 480 ( mcclintock et al .
2001 , wagner et al .
2001 ) ; gs 1124 - 683 ( remillard , mcclintock , & bailyn 1992 ) ; 4u 1543 - 47 ( orosz et al .
1998 ) ; xte j1550 - 564 ( orosz et al . 2002 ) ; gro j1655 - 40 ( bailyn et al . 1995 ) ; gx 339 - 4 ( hynes et al .
2003 ) ; h1705 - 250 ( remillard et al . 1996 ) ; sax j1819.3 - 2525 ( orosz et al . 2001 ) ; xte j1859 + 226 ( filippenko & chornock 2001 ) ; grs 1915 + 105 ( greiner , cuby , & mccaughrean 2001 ) ; gs 2000 + 25 ( casares , charles , & marsh 1995 ) ; gs 2023 + 338 ( casares , charles , & naylor 1992 ) ; cyg x-1 ( gies & bolton 1986 ) ; lmc x-3 ( cowley et al . 1983 ) ; lmc x-1 ( hutchings et al .
these sources open up the possibility of studying general relativity in the strong field regime .
for example , the study of high frequency quasiperiodic oscillations in the x - ray light curves of certain xn may lead to a measurement of black hole spin ( e.g. mcclintock & remillard 2004 and cited references ) .
we must press hard to obtain further observations of black hole masses in order to fully pursue these opportunities .
xte j1650 - 500 ( hereafter j1650 ) was discovered by rxte on 2001 september 5 ( remillard 2001 ) and subsequently reached a peak x - ray intensity of 0.5 crab .
based on subsequent observations , j1650 was established as a strong black hole candidate based on its x - ray spectrum and variability in the x - ray light curve ( markwardt , swank , & smith 2001 ; revnivtsev & sunyaev 2001 ; wijnands , miller , & lewin 2001 ) .
the radio counterpart was discovered with the australia telescope compact array ( atca ) by groot et al .
further radio observations sampled the behavior of xte j1650 - 500 along all its x - ray states ( corbel et al . in preparation ) .
we highlight two key results obtained during its outburst phase .
first , rxte observations during the third and fourth weeks of the outburst yielded a strong x - ray qpo with an rms amplitude of @xmath11 at a frequency of @xmath12 hz ( homan et al .
second , _ xmm - newton _ observed a broad , skewed emission line due to fe k@xmath13 ( miller et al . 2002 ) .
those authors argue that their results imply the primary is a nearly maximal kerr black hole that is delivering its spin - down energy to the accretion flow . the first significant observational program in the optical was that of sanchez - fernandez et al .
( 2002 , hereafter sf2002 ) .
they observed j1650 on the night of 2002 june 10 with the fourth 8.2 m telescope at the european southern observatory , paranal , and reported the following orbital elements : an orbital period of @xmath14 days and a velocity semiamplitude of @xmath15 km s@xmath3 , resulting in an optical mass function of @xmath16 .
the results in this paper contradict these findings .
in this paper we report the results of our photometric study of j1650 . a time series analysis of our photometry rules out the orbital period reported by sf2002 .
consequently we also report herein our reanalysis of the sf2002 data obtained from the eso archives .
we show that the spectroscopic period we derive from these data is consistent with our photometric period @xmath17 days , and we go on to determine the orbital elements of the system .
we outline below our observations and reductions , and our analysis techniques .
we end with a brief discussion of the implications of our results regarding the fast qpo observed for j1650 .
we observed j1650 with the 6.5 m clay telescope at las campanas observatory of the carnegie institution 2003 may 31 and june 1 using the magellan instant camera ( magic ) and an @xmath0-band filter .
a total of 47 @xmath0-band images with typical exposure times between 500 and 600 seconds were obtained over the two nights in good conditions : the average seeing was @xmath18 arcsec and it was photometric about 75% of the time .
m. holman and p. schechter kindly provided an additional 65 images obtained 2003 august 1 and 2 using the same instrumentation , and the observing conditions were comparable to those just described . in order to minimize readout time , the magic camera is read out simultaneously using four amplifiers .
thus , bias and flat - fielding is done separately for each quadrant of the detector .
these calibrations and the merging of the quadrant images into a single image were performed using the publicly - available magic reduction pipeline .
stetson s programs daophot iie , allstar , and daomaster @xcite were used to extract the stellar intensities and derive the light curve for j1650 ( see orosz & wade 1999 for a detailed discussion of the overall procedure used ) .
this suite of codes gives robust instrumental magnitudes .
daophot fits the point spread function ( psf ) for each image using several relatively isolated bright stars , and allstar uses the psf and finds instrumental magnitudes for all of the stars on an image simultaneously ( local background subtraction is included ) .
finally daomaster iteratively solves for zero - point shifts in the magnitude scales from image to image by essentially using all of the stable stars as `` comparison '' stars .
thus the final results are insensitive to changes in the seeing and to changes in the sky transparency . fig .
[ figfc ] shows a finding chart for j1650 made from the magic data .
sf2002 observed j1650 on the night of 2002 june 10 with the fourth 8.2 m telescope at the european southern observatory , paranal , using the fors2 imaging spectrograph with an mit ccd mosaic , the 1200r grism , the gg435 blocking filter , and a fixed slit width of 0.7 arcsecond .
this instrumental configuration yields a wavelength range of 5750 - 7310 with a resolution of about 0.76 fwhm .
sf2002 obtained a total of 15 spectra of j1650 in somewhat marginal seeing ( typically between 1.0 and 1.5 arcseconds ) .
they also obtained the spectra of six bright g- and k - type dwarfs using the same instrumental configuration .
we obtained these data from the eso archive , and used tasks in iraf to perform the ccd reductions and to extract the spectra . since the flat field and wavelength calibration images were obtained exclusively during the daytime hours at paranal , it was necessary to correct for flexure in the spectrograph by applying small corrections ( @xmath19 ) to the wavelength scales of the extracted spectra using bright night - sky emission lines
the magellan light curve of j1650 showed considerable variability , so we phased the data on the period reported by sf2002 ( @xmath20 days ) . to our surprise , the light curve phased on that period showed considerable scatter .
consequently , we searched the magellan light curve for periodicities using the elc code with its genetic fitting algorithm ( orosz & hauschildt 2000 ; orosz et al .
about 1,106,000 ellipsoidal models were generated and compared with the data using the absolute deviation as the merit function : @xmath21 where @xmath22 denotes the model value computed at @xmath23 , @xmath24 is the observed quantity at the same @xmath23 , and @xmath25 is the uncertainty in @xmath24 .
this merit function is more robust than the @xmath26 function because it is less sensitive to a few outlying points .
models were computed using a period range of 0.15 to 0.50 days .
after all of the fits were performed , the merit function hypersurface was projected onto the trial period axis ( see orosz et al .
2002 for a more in - depth discussion of this error estimation technique ) ; the resulting periodogram is shown in the top of fig .
the absolute deviation has a minimum value at a trial period of 0.3205 days , and a secondary minimum at a trial period of 0.3785 days .
[ fig2 ] ( top ) shows the magellan light curve phased on a period of @xmath17 days .
there is no dip in the absolute deviation near the period reported by sf2002 ( @xmath20 days ) , and the light curve phased on that period is essentially a scatter plot ( bottom of fig .
[ fig2 ] ) . as a check on our results
, we used the ` pdm ' task in iraf , which is an implementation of the phase dispersion technique of stellingwerf ( 1978 ) .
a similar periodogram was obtained . to estimate the error on our photometric measurement of the orbital period
, we did additional fits using elc s genetic algorithm and the normal @xmath26 as the merit function .
we find a @xmath27 uncertainty of about 0.0007 day , so we adopt @xmath28 day .
we then analyzed the vlt spectra in an attempt to understand the disagreement of our photometric period with the spectroscopic period reported by sf2002 .
we tried to extract radial velocities using the standard cross - correlation technique of tonry & davis ( 1979 , implemented in the iraf ` fxcor ' task ) .
however , we found that the spectra were very noisy and yielded only a few marginal measurements of velocity .
we therefore used the `` restframe '' analysis that we developed for h1705 - 250 ( remillard et al .
this technique is similar to the `` skew mapping '' technique sometimes used for cataclysmic variables ( smith , cameron , & tucknott 1993 ; vande putte et al .
2003 ) , and is quite simple to employ .
suppose one has a time series of spectroscopic observations .
if the spectra are doppler shifted to zero velocity and coadded using the correct orbital elements ( e.g.the period @xmath29 , the semiamplitude @xmath30 and the time of maximum velocity @xmath31 ) , then the absorption features of the companion star will appear at the same wavelengths in all of the individual spectra .
consequently , the lines in the summed spectrum will have a higher signal - to - noise ratio than in the individual spectra .
if , on the other hand , one uses the incorrect orbital elements , then the absorption lines from the companion will be at different wavelengths in different individual spectra , and hence will be `` averaged out '' in the resulting summed spectrum . in our implementation of the restframe analysis , we used a fortran program that writes iraf scripts to do the doppler shifting and the spectrum summation .
the fxcor task was used to provide a measure of how well a restframe spectrum matched the template spectrum .
in particular , we used the peak cross - correlation value as the measure of the goodness - of - fit of the template spectrum to the restframe spectrum .
the wavelength region used for the cross - correlation analysis was 5907 - 6261 , 6304 - 6513 , 6610 - 6800 , and 7000 - 7239 .
this region excludes telluric lines , and the h@xmath13 emission line .
we are interested in three orbital elements : the period @xmath29 , the semiamplitude @xmath30 and the time of maximum velocity @xmath31 .
this three dimensional parameter space was searched by making two dimensional grids of restframe spectra in the @xmath32 plane at several different values of the period @xmath29 . specifically , we used periods in the range 0.1500 to 0.5000 days in steps of 0.0020 days .
the restframe spectra for each trial period were stored in a separate subdirectory on disk . for each period
, we constructed a grid of restframe spectra in the @xmath33 plane as follows .
the range considered for @xmath34 was 300600 km s@xmath3 in steps of 2 km s@xmath3 .
for @xmath35 the range was centered on hjd 2,452,436.6 and the step size was 0.005 days . the extent of the range for @xmath35 was adapted to accommodate the trial value of the period .
an individual restframe spectrum was generated for each set of orbital elements . in total
we generated 1,166,146 restframe spectra using this same number of scripts stored as separate files .
next , we did a cross - correlation analysis on the restframe spectra .
it was therefore necessary to find a good template spectrum . to do this
, we adopted a period of 0.320 days found from the photometry and considered the restframe spectra in the @xmath32 plane at this period .
we then cross - correlated these restframe spectra against all six of the template spectra in turn and determined that the k4v star bs5568 provided the best match since it gave the strongest cross - correlation peaks .
next , all of the restframe spectra at all periods were cross - correlated in batch mode against the k4v template .
we made a periodogram by parsing the fxcor log files to get the peak cross - correlation value for each restframe spectrum .
the maximum peak cross - correlation value within each period subdirectory was saved , resulting in the periodogram shown at the bottom of fig .
the maximum cross - correlation value occurs for a trial period of @xmath36 days , in agreement with the photometric results , which thereby establishes the orbital period .
thus , we rule out the alias photometric period near 0.38 days and again rule out the 0.212 day spectroscopic period reported by sf2002 ( fig . [ fig1 ] ) .
we also made a periodogram using a template with spectral type g5v .
overall , the cross - correlation values were much lower and the periodogram was much noisier .
however , there still was a peak near @xmath36 days , and we conclude our adopted spectroscopic period is not sensitive to the choice of the template . adopting a period of @xmath36 days , we performed a finer search of the @xmath32 plane , with step sizes of 1 km s@xmath3 in @xmath30 and 0.001 days in @xmath31 .
the results are shown in fig .
the cross - correlation values have a fairly well - defined peak value along the @xmath31 axis , where the maximum cross - correlation is for @xmath37 ( hjd ) . on the other hand ,
the peak along the @xmath30 axis is broad with a maximum that occurs in the range @xmath38 km s@xmath3 .
judging from the width of the peak and on our experience with h1705 - 250 ( remillard et al.1996 ) , we adopt @xmath39 km s@xmath3 .
the optical mass function is then @xmath40 the best restframe spectrum found for the finer grid is shown in fig .
[ plotspect ] .
although few obvious k - star absorption features are apparent , the cross - correlation of this restframe spectrum using the k4v star bs 5568 as a template does yield a significant peak near zero velocity ( fig .
[ plotspect ] ) .
since the spectroscopic period agrees with the photometric period , our initial determination that the k4v template provides the best match needs no refinement . in order to be more quantitative ,
we cross - correlated the best restframe spectrum from the fine grid against all six template spectra .
table [ tabcor ] gives the cross - correlation value at the peak and the tonry & davis ` @xmath41 ' value ( which is a measure of the significance of the peak ) for each template . as noted earlier
, the k4v template has the best cross - correlation , so we adopt a spectral type of k4v for the companion star .
however , one should note that the gv5 star bs 7330 provides the second - best match , and as such our assignment of a spectral type of k4v should be treated with caution .
we attempted to decompose the restframe into its stellar and accretion disk components using the technique outlined in marsh , robinson , & wood ( 1994 ) . using the k4v template bs 5568
, we found that the accretion disk dominates in the @xmath0-band , where its contribution is @xmath42 .
however , owing to the limited number of templates and the poor signal - to - noise , this result should also be treated with caution .
our measurement of the optical mass function sets a lower limit on the mass of the compact primary . to find the actual mass we need to find the mass ratio of the two components and inclination of the orbital plane .
if the companion star fills its roche lobe and is in synchronous rotation , the mass ratio can be computed if one can measure @xmath30 and the projected rotational velocity of the companion star ( e.g. wade & horne 1988 ) .
one can also get an estimate of the mass ratio if the radial velocities of the h@xmath13 emission line can be reliably measured ( e.g. orosz et al .
however , in the case of j1650 , the poor quality of the spectroscopic data prevents us from making either measurement .
we can use our magellan light curve to place some constraints on the inclination .
however , the results are sensitive to the amount of light the accretion disk contributes in the @xmath0-band , and our estimate of the disk contamination outlined above is quite uncertain .
nevertheless , if we assume the limiting case of no light from the disk in the @xmath0-band , then we find an inclination of @xmath43 ( see fig .
[ fig2 ] for a representative model for this case ) .
if we add a substantial contribution from the disk in the @xmath0-band ( @xmath42 ) , then the inclination is @xmath44 or higher ( a representative ellipsoidal model for this case is also shown in fig . [ fig2 ] ) .
owing to the lack of x - ray eclipses , the inclination is less than about @xmath45 ( the exact limit depends on the mass ratio ) . if , for the sake of discussion , we assume the mass ratio of j1650 is similar to that of a0620 - 00 or grs 1124 - 683 ( @xmath46 ) , then an inclination of @xmath47 combined with our measured value of the optical mass function yields a black hole mass of @xmath48 .
likewise , an assumed inclination of @xmath44 yields a black hole mass of @xmath49 .
is general relativity ( gr ) the correct theory of gravity in the strong fields found near a black hole ?
one promising approach to answering this fundamental question is offered by the key discovery of the _ rossi x - ray timing explorer _ ( rxte ) that seven galactic black holes ( including j1650 ) display quasiperiodic ( qpo ) x - ray oscillations in the range of @xmath50 hz ( mcclintock & remillard 2004 and cited references ) .
these fast qpos must be produced near the event horizon since the x - rays originate there and since such high frequencies are comparable to the kepler frequency of the innermost stable circular orbit around a black hole ( @xmath51 depending on the dimensionless spin parameter @xmath52 , where the full range of @xmath52 is 0 to 1 e.g.kato , fukue , & mineshige 1998 ) .
rxte has made a further important discovery : four of these seven black holes produce pairs of stationary , fast qpos that have frequencies in a 3:2 ratio ( mcclintock & remillard 2004 ; homan et al .
such commensurate frequencies are a hallmark of non - linear resonance phenomena .
thus , this discovery has promoted a `` resonance model '' that invokes enhanced emissivity at the radius in the accretion disk where two of the three spatial coordinate frequencies ( e.g. keplerian and radial ) are commensurate ( abramowicz & kluzniak 2001 ; abramowicz et al .
for the three black holes with the qpo pairs _ and _ measured masses ( e.g. gro j1655 - 40 , xte j1550 - 564 , and grs 1915 + 105 ) , the qpo frequency is correlated with the black hole mass , where @xmath53 ( e.g.mcclintock & remillard 2004 ) .
this scaling is expected for gr oscillations , but different sources can lie on the same curve only if they have similar values of the dimensionless spin parameter @xmath52 .
schnittman & bertschinger ( 2004 ) performed ray tracing calculations in the kerr metric for emitting blobs orbiting a black hole at a radius where there is a 3:1 resonance between the azimuthal ( @xmath54 ) and radial coordinate frequencies .
it was shown that general relativistic effects may impart a qpo at @xmath54 and a beat oscillation at @xmath55 with a relative strength that depends on the angular width of the emitting blob of material .
this interpretation , when combined with accurate mass measurements , yields values for the dimensionless spin parameter : @xmath56 for gro j1655 - 40 and @xmath57 for xte j1550 - 564 ( remillard et al .
this illustrates the potential diagnostic power of fast qpos , if we can specify the correct oscillation mechanism .
furthermore , measurements of black hole mass and qpo frequency allow _ model - independent _ comparisons of spin differences between black hole subclasses as distinguished , for example , by relativistic jets or binary period . with our measurement of the orbital period and optical mass function of j1650 ,
we have taken the first steps need to fully exploit the potential of this system . if the qpo at @xmath12 hz observed by homan et al .
( 2003b ) represents the @xmath58 oscillation , then the predicted mass of the black hole would be about @xmath59 , if j1650 lies on the same @xmath60 curve on which gro j1655 - 40 , xte j1550 - 564 , and grs 1915 + 105 lie ( i.e. if the black holes all have similar values of the dimensionless spin parameter @xmath52 ) .
although there are substantial uncertainties in both our measured value for the optical mass function and in our inclination estimate , it appears that the black hole mass might be less than @xmath59 , which in the context of the orbital resonance model would imply that the value @xmath52 in j1650 is much smaller than it is for the other three sources ( e.g. abramowicz et al .
2004 ) . on the other hand , on the basis of x - ray spectroscopy , miller et al .
( 2002 ) have argued that j1650 contains a maximal kerr black hole .
thus j1650 may be a challenging case for the orbital resonance model of the high frequency qpos , although we again point out that our mass estimate has large uncertainties .
recent models for black hole formation ( e.g.fryer & kalogera 2001 ) predict a continuous and roughly exponential distribution of masses for black holes in binary systems , where the number of black holes formed falls off as the mass increases . depending on their model assumptions ,
these authors predict the number of black holes with masses in the range @xmath61 should be roughly two to three times the number of black holes with masses in the range @xmath62 .
bailyn et al.(1998 ) performed a statistical analysis of seven black holes with dynamical mass estimates and found strong evidence for a peak in the mass distribution centered near @xmath63 and a `` gap '' in the distribution between the neutron star masses and the peak near @xmath63 . since that time , the sample of objects has more than doubled , and it now appears that the evidence for a peak in the distribution near @xmath64 is weak ( see the references cited in 1 and orosz 2003 ) . however , black holes with masses in the @xmath61 range still seem to be rare , with gro j0422 + 32 being the only good candidate ( gelino & harrison 2003 ) . in this regard
, j1650 is an interesting system because of its probable low mass .
clearly additional data will be required to make more definitive statements about the mass of the compact object in j1650 .
it should not be unduly difficult to obtain higher quality spectra .
in hindsight , we now know that the companion star is a k - star , and as such it should have relatively strong absorption features near 5170 .
this region was not included in the vlt spectra owing to the relatively high resolving power of the 1200r grism .
it should be easier to measure radial velocities in spectra with more wavelength coverage ( at the expense of resolving power ) . also , spectral observations obtained in good to excellent seeing conditions ( @xmath65 arcseconds ) obviously will have better signal - to - noise than the current spectra do ( the seeing varied between 1 and about 1.4 arcseconds ) .
additional photometry should be obtained , and it would be helpful if two or more bandpasses could be used , since the amount of contaminating light from the disk should vary with color .
using @xmath0-band photometry of xte j1650 - 500 obtained with the 6.5 m clay telescope , we have measured a photometric period of @xmath28 days . this value is confirmed by our reanalysis of archival spectroscopic data .
our reanalysis of the spectroscopy also yields a velocity semiamplitude for the companion star of @xmath66 km s@xmath3 , which when combined with the orbital period , gives an optical mass function of @xmath67 .
the spectral type of the companion star is not well constrained because we only have six template spectra available to us .
a template spectrum with a spectral type of k4v provides the best match ; next best matches are provided by template spectra with spectral types of g5v and k2iii , respectively .
the summed spectrum suggests that the accretion disk may contribute a substantial amount of light in @xmath0 ( @xmath68 or more ) . using the @xmath0-band light curve
, we find a lower limit to the inclination of @xmath43 , which gives an upper limit to the mass of the black hole of @xmath69 , assuming a mass ratio of @xmath70 .
if we assume that the accretion disk contributes 80% of the light in @xmath0 , then the inclination would be about @xmath44 or more , and the black hole mass would be about @xmath8 ( again assuming a mass ratio of @xmath70 ) .
rcc @xmath71 ( deg ) & 79.0 & 51.7 @xmath72 & 20.7 & 1.1 @xmath73 ( k ) & 4500 & 4500 @xmath74 & 0.099 & 0.099 @xmath75 ( k ) & 15000 & @xmath76 & 0.01 & @xmath77 & 0.99 & @xmath78 & -0.3 & @xmath79 ( deg ) & 2.0 & | we present @xmath0-band photometry of the x - ray transient and candidate black hole binary xte j1650 - 500 obtained between 2003 may and august with the 6.5 m clay telescope .
a timing analysis of these data reveals a photometric period of @xmath1 days ( i.e. 7.63 hr ) with a possible alias at 0.3785 days ( 9.12 hr ) .
our photometry completely rules out the previously published spectroscopic period of 0.212 days ( 5.09 hr ) .
consequently , we reanalyzed the 15 archival eso / vlt spectra ( obtained 2002 june by sanchez - fernandez et al . )
that were the basis of the previously published spectroscopic period .
we used a `` restframe search '' technique that is well suited for cases when the signal - to - noise ratio of individual spectra is low . for each of roughly 1.1 million binary ephemerides , we summed all of the spectra in a trial restframe of the secondary star , and each restframe spectrum was cross - correlated against a template spectrum .
we then searched for the set of orbital parameters that produced the strongest cross - correlation value .
the results confirmed the photometric period of 0.3205 days , and rule out the alias period near 0.38 days .
the best value for the velocity semiamplitude of the companion star is @xmath2 km s@xmath3 , and the corresponding optical mass function is @xmath4 .
the spectral type of the companion star is not well constrained because we only have six template spectra available to us .
the k4v template provides the best match ; next best matches are provided by the g5v and k2iii templates .
we also find that the accretion disk dominates the light in the @xmath0-band where the disk fraction is 80% or higher , although this value should be treated with caution owing to the poor signal - noise - ratio and the limited number of templates .
the amplitude of the phased @xmath0-band light curve is 0.2 magnitudes , which gives a lower limit to the inclination of @xmath5 in the limiting case of no contribution to the @xmath0-band light curve from the accretion disk . if the mass ratio of xte j1650 - 500 is similar to the mass ratios of other black hole binaries like a0620 - 00 or grs 1124 - 683 ( e.g.@xmath6 ) , then our lower limit to the inclination gives an upper limit to the mass of the black hole in xte j1650 - 500 of @xmath7 .
however , the mass can be considerably lower if the @xmath0-band flux is dominated by the accretion disk .
for example , if the accretion disk does contribute 80% of the flux , as our preliminary results suggest , then the black hole mass would be only about @xmath8 . |
in this work , we studied the excitonic energy levels in an asymmetric vertical double quantum well @xcite-@xcite as a function of the magnetic and electric field strengths . within the effective - mass approach we expanded the excitonic wave function in an orthogonal basis formed by products of electron and hole wave functions along the crystal growth direction @xmath4 , and one - particle solutions of the magnetic hamiltonian in the @xmath5-@xmath6 plane .
the coulomb potential between electrons and holes produces off - diagonal terms thus mixing our basis states .
we obtained the energy spectra and wave functions by diagonalizing the excitonic hamiltonian in a truncated basis .
we applied our method to study the excitonic states in a gaas al@xmath0ga@xmath1as dqw , for the specific case of a dqw composed of a left well of @xmath7 nm , a barrier of @xmath8 nm , and a right well of @xmath9 nm . the effects of external electric and magnetic fields on the luminescence ( pl ) intensity for this heterostructure has been recently studied experimentally by krivolapchuk et al.@xcite , @xcite .
the effective - mass hamiltonian for excitons in a double quantum well in the diagonal approximation @xcite and in the presence of a magnetic field pointing towards @xmath4 , can be written as @xmath10 where @xmath11 is the one - dimensional hamiltonian for electrons , @xmath12 @xmath13 is the one - dimensional hamiltonian for holes , @xmath14 and @xmath15 ( @xmath16 ) is the potential that defines the double quantum well for electrons ( holes ) in the five regions of @xmath4 .
we included the electric field in @xmath15 and @xmath16 by a shift in the potential in stair steps similar to fig .
[ pot - z ] .
@xmath17 is the magnetic hamiltonian in the symmetric gauge , which depends on the relative coordinates of electrons and holes in the @xmath18 plane , @xmath19 where @xmath20 , @xmath21 and @xmath22 are the momentum operator , reduced mass , and hole mass respectively , defined on the @xmath18 plane .
@xmath23 is the coulomb potential between electrons and holes , including an effective dielectric constant for the system .
we expanded the solutions of the hamiltonian ( [ h - total ] ) , as a linear combination of products of eigenfunctions of the magnetic hamiltonian in the @xmath18 plane ( [ h - mag ] ) , and eigenfunctions of the electron and hole hamiltonians in the @xmath4 direction [ ( [ h - ele ] ) and ( [ h - hole ] ) ] , @xmath24 in which , in the symmetric gauge , @xmath25 where @xmath26 , and only @xmath27 functions are considered .
the electron wave functions defined in the five regions of @xmath4 , shown in fig
. [ pot - z ] , are given in terms of @xmath28 , @xmath29 , and @xmath30 .
the hole wave functions are given by similar expressions .
the coulomb interaction produces off - diagonal terms by mixing our basis states . in order to obtain the system of equations for the coefficients in expansion ( [ expansion ] ) , we need to evaluate the coulomb integrals @xmath31 the @xmath32 integral is trivial , because of @xmath33 conservation . using the explicit expansion of laguerre polynomials ( @xmath34 ) in @xmath35 and @xmath36 ( [ f - mag ] ) and after solving the @xmath37 integral , eqn .
( [ i - total ] ) can be written as a sum of terms of the form @xmath38 the @xmath39- and @xmath40- integrals can be solved analytically . the evaluation of these integrals is cumbersome due to the large number of terms , resulting from the five different regions of the potential .
the @xmath39- and @xmath40- integrals contain both decoupled terms in which the @xmath39- and @xmath40- integrals are independent of each other and coupled terms where the integration limits of the @xmath40-integral contains @xmath39 .
the remaining @xmath41 integral must be calculated numerically . by diagonalizing the system of equations resulting for the coefficients in expansion ( [ expansion ] ) , in a truncated basis
, we obtained the energies and wave functions for the first excitonic levels .
evaluating the oscillator strength we obtained the optical absorption spectra .
we calculated the excitonic energy levels for a slightly asymmetric 10.18/3.82/9.61 nm gaas al@xmath0ga@xmath1as double quantum well , in the presence of electric and magnetic fields .
the band gap used in our calculations is given by @xmath42 ( @xmath43 ) .
the band - gap offset considered was @xmath44 for the conduction band and @xmath45 for the valence band .
for all five regions in the double quantum well , we used the same electronic mass @xmath46 , the @xmath18 plane heavy - hole mass , @xmath47 , the @xmath48 heavy - hole mass @xmath49 , the light - hole masses @xmath50 and @xmath51 , and a dielectric constant @xmath52 . in our calculations
we used a truncated basis set composed of @xmath53 landau wave functions , four electronic wave functions , five heavy - hole wave functions , for which most of our results converged to less than a tenth of mev . for magnetic fields below @xmath54 t , we would need to increase the number of landau levels considered .
we included a fictitious width of @xmath55 mev for each optical absorption peak .
in this section we present our results for the energies and wave functions of electron and holes in a one - dimensional dqw , as a function of the external electric field .
we are interested in regions of electric field where two different electron ( or hole ) energy levels cross each other leaving a gap of a few mev between them .
when the wave functions involved in this kind of anticrossing are important for the optical absorption , we expect to find interesting behaviors in the peaks of the excitonic optical absorption .
figure [ e_f ] shows the electron ( upper plot ) and hole ( lower plot ) energy levels for this dqw as a function of the external electric field . in the range from @xmath56 to @xmath57
kv / cm ( not shown in the plot ) , there are crossings in the hole energy levels near @xmath54 , @xmath58 , @xmath59 and @xmath60 kv / cm .
these crossings do not met the two conditions explained before , and will not produce a noticeable behavior in the excitonic optical absorption . in the range from @xmath57 to @xmath61
kv / cm there is a region of electric fields near @xmath62 kv / cm , indicated with an arrow , where the second and the third electronic energy levels anticross . to understand the excitonic optical absorption it is necessary to know the behavior of the electron and hole dqw wave functions for electric field below , inside , and above the anticrossing region .
figure [ wf_0 ] shows the first four electron ( upper plot ) and hole ( lower plot ) wave functions for zero external electric field .
numbers indicate the order of the energy levels , and the most important states for the excitonic optical absorption are represented by a solid line .
the first state mostly localized in the left ( wider ) well is called * l1 * and the first state mostly localized in the right well is called * r1*. it is clear from this figure that holes are almost completely localized in the left or right wells , and that electrons have strong components over both wells .
this different behavior for electrons and holes is a consequence of the effective mass differences for electron and holes and the size of the wells and barrier in this sample .
figure [ wf_71 ] shows the electron ( upper plot ) and hole ( lower plot ) wave functions for an external electric field of @xmath63 kv / cm , where the most important states for the excitonic optical absorption are represented by a solid line . comparing with the previous case of electrons in zero electric field ,
we note that : the first level * l1 * get much more localized in the left well , the second level * r1 * changes dramatically into * l2 * , and the third level behaves as * r1*. for holes we can see that the electric field forces the wave functions to be localized in the right well , and the fourth level is localized in the left well ( corresponding to * l1 * ) .
figure [ wf_85 ] shows the electron ( upper plot ) and hole ( lower plot ) wave functions for an external electric field of @xmath61 kv / cm . again the most important states for the excitonic optical absorption are represented by solid lines . comparing with the previous case of a @xmath63
kv / cm electric field , we note that electron wave functions remains almost unchanged , and that the fifth hole wave function behaves as * l1*. increasing the electric field in a double quantum well produces a shift in the electron and hole energy levels and a change in the localization of the respective wave - functions .
the combination of both effects changes the behavior of the excitonic states and affects , in a non trivial way , the optical absorption spectrum of these systems .
figure [ abs_f ] shows the electric - field effects on the optical absorption spectra for an external magnetic field of @xmath64 t. each curve , corresponding to a different value of the electric field , has been displaced for clarity . for an electric field of @xmath57 kv / cm ( upper curve )
there are two peaks in the excitonic optical absorption .
the first peak corresponds to a direct exciton * l1l1 * , where the constituent electron and hole are localized in the left ( wider ) well .
the second peak corresponds to a direct exciton * r1r1 * , where the electron and hole are localized in the right well . for an electric field of @xmath61 kv / cm ( lower curve )
the same two peaks occur , corresponding again to the left direct exciton * l1l1 * and the right direct exciton * r1r1*. the main difference between the @xmath57 kv / cm and the @xmath61 kv / cm case is the exciton composition in terms of the base states .
this different composition of base states can not be observed experimentally in the previous limiting cases , but has a strong influence in the optical absorption behavior for electric fields between @xmath57 kv / cm and @xmath61 kv / cm .
the complex structure in the optical absorption spectra in this range of electric fields can be explained by the formation of excitonic states , which have a strong component of indirect exciton states . in this case
the indirect exciton state corresponds to * l2r1 * , which is formed by an electron in the second left state and a hole in first right state .
the composition in terms of our base states can be obtained from figures [ wf_0 ] , [ wf_71 ] , and [ wf_85 ] . when comparing this figure with fig . 2 of @xcite
, we see a similar behavior in the pl peaks .
this suggests that the peak attributed to an impurity - bound direct exciton in the experimental results , could be interpreted as a mixed state with direct and indirect exciton components .
quantitative agreement between our results and those in the experimental work is not possible because we included a strong magnetic field .
peaks originating from light holes states appear at the right end of this figure ( not shown ) . increasing the magnetic field in a double
quantum well produces a shift in the excitonic energy levels towards higher energies and an increase in the interaction energy .
the increase in the interaction energy between electrons and holes is a consequence of two effects : first , the wave function confinement in the @xmath5-@xmath6 plane produces a stronger interaction along the @xmath4 direction , as the wave functions penetrate more into the barrier .
second , the interaction between electrons and holes in the same plane increases , because their wave functions are confined to a smaller region .
figure [ abs_b : f=3d71 ] shows the magnetic - field effects on the optical absorption spectra for an external electric field of @xmath63 kv / cm .
the differences in the interaction energies in the layer plane and along the @xmath4 @xmath65 , when the magnetic field is varied , produces a different behavior in the optical absorption peaks with direct or indirect exciton components .
this difference produces the splitting of the first peak for a magnetic field of @xmath54 t into two peaks for stronger magnetic fields .
figure [ abs_b : f=3d78 ] shows the magnetic - field effects on the optical absorption spectra for an external electric field of @xmath66 kv / cm . in this case the magnetic field produces no mixing between excitonic states , and produces different shifts in energies for direct excitons ( * l1l1 * and * r1r1 * ) and the indirect exciton ( * l2r1 * ) .
this effect is clearly observed as a function of magnetic field .
in this work we studied the energies and wave functions of excitonic states in a 10.18/3.82/9.61 nm gaas al@xmath0ga@xmath1as double quantum well , in a magnetic field pointing along the growth direction @xmath4 .
we calculated the optical absorption spectra as a function of the external electric and magnetic fields and studied the behavior in the pl peaks for the indirect exciton @xmath2 and direct exciton @xmath3 .
we compared our results and those in the experimental work of @xcite and found similar behavior in the pl peaks .
this suggests that the peak attributed to an impurity - bound direct exciton in the experimental results , could be interpreted as a mixed state with direct and indirect exciton components . in our method of calculation
we used an orthogonal basis that involves functions of single - particle solutions of the double quantum well in @xmath4 direction , which makes our method appropriate for small and medium barrier widths .
the study of these systems for magnetic fields below @xmath54 tesla within this method requires to increase the number of landau levels in our base states , making the numeric computation excessively time consuming .
this work was supported by the fondecyt postdoctoral project no .
3990051 and by the `` catedra presidencial en ciencias 1998 , francisco claro '' .
kamizato t and matsuura m 1989 phys
. rev . * * * * b * 40 * 8378 dignam m and sipe j 1991 phys .
b * 43 * 4084 ferreira r , delalande c , liu h w , bastard g , etienne b and palmier j f 1990 phys . rev .
b * 42 * 9170 ferreira r and bastard g 1997 rep . prog
60 * 345 ten s , krol m f , guerreiro p t and peyghambarian n 1996 appl .
. lett . * 69 * 3387 haas h , magnea n and dang l 1997 phys .
b * 55 * 1563 kasapoglu e ,
sari h , balkan n , sokmens i and ergun y 2000 semicond .
* 15 * 219 butov l v , shashkin a a , dolgopolov v t , campman k l and gossard a c 1999 phys.rev .
b * 60 * 8753 bastard g 1990 wave mechanics applied to semiconductor heterostructures , les editions de physique , les ulis france p. 139
krivolapchuk v v , moskalenko e s , zhmodikov a l , cheng t s and foxon c t 1999 phys .
solid state * 41 * 291 krivolapchuk v v , mazurenko d a , moskalenko e s , poletaev n k , zhmodikov a l , cheng t s and foxon c t 1998 phys
. solid state * 40 * 737 | we studied , theoretically , the optical absorption spectra for a slightly asymmetric double quantum well ( dqw ) , in the presence of electric and magnetic fields .
recent experimental results for a 10.18/3.82/9.61 nm gaas al@xmath0ga@xmath1as dqw show clearly the different behavior in the luminescence peaks for the indirect exciton @xmath2 and left direct exciton @xmath3 as a function of the external electric field .
we show that the presence of a peak near the @xmath3 peak , attributed to an impurity bound left @xmath3 in the experimental results , could be a consequence of the non - trivial mixing between excitonic states . |
here we report bright solitons in a gaas semiconductor microcavity ( see methods section for sample and experimental details ) .
the point of inflection of the polariton dispersion is found at the in - plane polariton momentum @xmath0m@xmath1 ( see fig .
1(a ) ) , which corresponds to a group velocity of noninteracting polaritons @xmath2m / ps , where @xmath3 is the polariton energy .
we conduct our experiments with polaritons having momenta above the point of inflection , where the polariton mass is negative , @xmath4 kg ( @xmath5 ) at @xmath6m@xmath1 . as for the lossy and pumped gross - pitaevskii equations , see , e.g. ref.@xcite , reasonable estimates of the nonequilibrium soliton parameters can be obtained from the conservative limit of the system .
we can estimate the polariton soliton width for a given polariton density , @xmath7 using the well - known expression for the healing length of a quantum fluid @xmath8 .
this is obtained by equating the characteristic kinetic energy @xmath9 of the dispersive spreading of the wavepacket with width @xmath10 ( @xmath11 ) to the potential energy @xmath12 of the polariton - polariton repulsion ( @xmath13 , where @xmath14 @xmath15ev@xmath16m@xmath17 is the two - body interaction coefficient ) . for the typical soliton potential energies realised in our experiment @xmath18 mev ( corresponding to @xmath19m@xmath20 ) from which we deduce @xmath21 .
the short free polariton lifetime ( @xmath22 ps ) means that solitons emerging from this balance will traverse distances of @xmath23 m , before they dissipate . in order to sustain these solitons for
longer we need to provide a continuous supply of energy . to sustain these solitons ,
energy is pumped into the microcavity using a cw pump beam , focused to a @xmath24 ( fwhm ) gaussian spot ( fig . [ fig1](b ) ) .
we highlight the important properties of dissipative solitons which justify the choice of our experimental parameters to realise soliton production .
the polariton density varies strongly across the bright soliton profile reaching its maximum at the center , while the pump beam profile tends to hold the density at a quasi - constant level across the much larger pump spot @xcite .
furthermore , due to localisation in real space the soliton profile in momentum space is broad . as a result
, solitons can only be expected under conditions when the pump state is unstable with respect to spatially inhomogeneous perturbations at momenta different from that of the pump @xcite .
if the microcavity is driven by a pump beam slightly blueshifted with respect to the unperturbed lp branch , the pump polariton field exhibits bistability as a function of the pump beam power and angle ( fig .
[ fig1](c ) , fig .
[ fig4 ] ) .
the polariton bistability is usually accompanied by parametric ( modulational ) instability , which was extensively studied in microcavities @xcite .
this instability is a particular case of polaritonic four - wave mixing .
bright solitons are excited on top of the stable background of the lower branch of the bistability loop ( fig .
[ fig1](c ) ) and can be qualitatively interpreted as locally excited islands of the modulationally unstable upper branch solution @xcite .
broadband four - wave mixing of polaritons , expanding well beyond the momenta intervals with the parametric amplification , enables coherent scattering from the locally perturbed pump to the continuum of momenta forming the soliton ( fig .
[ fig1](a ) ) .
simultaneously , the soliton formation requires the transverse momentum of the pump @xmath25 , i.e. the incident angle of the laser beam , to be such that the effective polariton mass is negative , ensuring self - focusing of the repulsively interacting polaritons @xcite .
this mechanism can lead to self - localisation only along the direction of the pump momentum ( 1d solitons ) @xcite .
simultaneous localisation in two dimensions involves more subtle physics and the corresponding solitons exist only in a narrow range of pump intensities @xcite .
[ fig1](c ) shows the bistable dependence of the polariton emission intensity collected from our device at nearly zero transverse momenta ( direction normal to the cavity plane ) as a function of the pump momentum ( @xmath25 ) at the fixed energy of the pump @xmath26ev , which is @xmath27mev above the unperturbed lp branch at @xmath28m@xmath1 . with increasing @xmath25 beyond the bistable interval the intracavity polariton field increases abruptly over the excitation pump spot
this transition is also accompanied by the strong parametric generation of polaritons into the state with @xmath29 @xcite . according to the above discussion and the predictions of @xcite
, bright polariton solitons should exist within the bistability interval ( fig . [
fig1](c ) ) .
m wide writing beam with in - plane momentum the same as that of the pump .
( f ) is the same as ( d ) , but the writing beam has half the momentum of the pump .
( g ) is the same as ( f ) , but the writing beam width is @xmath30 m . ] we excite bright polariton solitons using a ps writing beam ( wb ) focused into a spot ( fig .
[ fig1](b ) ) of diameter in the range @xmath31 m to @xmath32 m , which is small compared to the diameter of the pump beam of @xmath33 m .
the writing beam is te and the pump beam is tm polarized . the experimental arrangement of the pump and writing beams is shown schematically in fig .
1(b ) , with both beams incident along the x - direction .
the unperturbed polariton density ( before the application of the writing beam ) was in the state corresponding to the lower branch of the bistability loop ( see arrow in fig .
[ fig1](c ) ) with no indication of the parametric generation of polaritons with momenta different from the pump .
the bistability domain was scanned by changing the pump momentum , where it was found that the optimal conditions for clear soliton observations exist close to the right boundary of the bistability interval shown in fig .
[ fig1](c ) .
the solitonic emission was collected along a @xmath34 m stripe of the streak camera image ( see methods ) parallel to the direction of incidence of the pump .
the light was collected inside the finite interval of angles corresponding to momenta @xmath35 , thereby avoiding collection of the reflected pump beam , which otherwise leads to detector saturation . figs . [ fig1](d ) and ( e ) show spatio - temporal traces of the intensities in te and tm polarisations , for typical non - diffracting and non - decaying propagating wavepackets excited with the writing beam . here
the @xmath31 m writing beam arrives at position @xmath36 m at time , @xmath37 where the transverse momentum is the same as that of the pump @xmath38 .
we now present the experimental results , supported by theory in the next section , which prove that the wavepackets arise from bright polariton solitons propagating across the excitation spot .
firstly , we demonstrate that the velocity of the soliton is independent of @xmath39 .
the role of the writing beam is to create a local perturbation of the pump state , which in turn results in soliton formation due to scattering from the pump state ( fig . [ fig1](b ) ) .
it has been shown numerically for linearly polarised polariton solitons that their velocity is close to the group velocity of the polaritons at the pump momentum @xcite . to test this prediction , we changed the momentum of the writing beam to half of the pump momentum , @xmath40 ( figs .
1(f , g ) .
we observe a soliton velocity in fig .
[ fig1](f ) of @xmath41m / ps , the same as that in figs .
[ fig1](d ) and ( e ) , where @xmath42 .
the independence of the polariton soliton velocity on @xmath39 is in sharp contrast to that expected for conservative solitons , where the soliton velocity is solely determined by the momentum of the excitation pulse @xcite .
secondly , we show that the size of the soliton is determined by the pump and cavity parameters and is independent of the size of the writing beam , as is true for other types of dissipative solitons @xcite .
the size of the excited wavepacket ( @xmath10 ) in the soliton regime is expected to be fixed by the potential energy @xmath12 of the solitons , where @xmath12 is of the order of the pump energy detuning with respect to the energy of the unperturbed lower branch polaritons ( @xmath43 as discussed earlier ) .
this is illustrated in figs .
[ fig2](a - f ) , which show the profiles of polariton wavepackets along their propagation direction ( x ) at different times and positions for writing beam sizes of @xmath44 m ( fig . [ fig2](a - c ) ) and @xmath45 m ( fig .
[ fig2](d - f ) ) . in the initial stage of the soliton excitation
, these writing beams produce polariton wavepackets of very different widths ( see figs . [
fig2](a ) and ( d ) ) , which then quickly evolve into solitons of the same size @xmath46 m as shown in figs . [ fig2](b ) and ( e ) , as expected for soliton formation .
m writing beam at different times , showing excitation and decay of the soliton ; ( d)-(f ) intensity profiles of a soliton created by the @xmath32 m writing beam ; ( g , h ) dynamics of the peak intensity and width of the wavepackets excited by the @xmath31 m writing beams with different powers as they propagate across the cavity .
the writing powers where the above parameters remain quasi - constant in the interval of @xmath47 to @xmath48 m correspond to the formation of solitons.,width=415 ] thirdly , we demonstrate the unambiguous non - spreading and non - decaying features of the soliton wave packets .
figures [ fig2](g ) and ( h ) show the dependences of the intensity and width of the excited wavepackets , respectively , versus their position as a function of the writing beam power , @xmath49 . within the range of @xmath49 from 1.6 mw to 2.3 mw ,
the soliton intensity and width are nearly constant .
furthermore , the width versus position dependence ( fig .
[ fig2](h ) ) exhibits a plateau in the interval from @xmath50 to @xmath51 m .
these signatures of polariton soliton formation are further discussed in the modeling section . at the edge of the pump spot ,
the pump intensity is insufficient to maintain the system within the bistability region and thus to sustain solitons .
this is the main factor determining the practical extent of the soliton trajectory , and leads to spreading and dissipation of the excited wavepackets at times greater than 40 - 45 ps when the edge of the pump spot is approached , as shown in figs .
[ fig1](d - g ) and fig .
[ fig2](c),(f ) .
we note that the intensity of the writing beam must be high enough to enable soliton switching .
[ fig2](h ) ) . as one reduces the pump intensity below the threshold of @xmath52mw ( fig .
[ fig2](h ) ) for soliton switching , we observe an abrupt decrease in the intensity and an increase in width of the wavepacket .
further reducing the pump intensity @xmath53 , only quickly decaying non - solitonic wavepackets are excited . for pump momenta
that are below the point of inflection , the writing beam triggers switching of the whole pump spot @xcite and no soliton formation is observed .
( a - c ) .
cross sections taken along the y direction indicates a soliton full width at half maximum of @xmath54 m . ]
figure [ fig3d ] shows two dimensional images of solitons .
the soliton size along the y coordinate perpendicular to the propagation direction ( see fig . [ fig1](b ) ) at @xmath55ps is @xmath56 m , narrower than the initial beam size .
this indicates suppression of diffraction and localization in the direction where polaritons have a positive effective mass .
whereas localisation in the x direction described above arises from the interplay between the negative effective mass and repulsive interactions , the origin of localisation along y is different .
it can be interpreted in terms of the phase dependent parametric nonlinearity and the interaction of propagating fronts , see ref .
@xcite .
we note that our observations have a very different physical origin to the triggered optical parametric oscillator observations of ref .
@xcite . in that case ,
the system is already in a high density phase before a writing beam is applied @xcite and furthermore the system was pumped at low @xmath57 in the region of positive effective mass , where soliton generation can not occur . as a result , propagating excitations of the polaritonic condensate
were studied , as opposed to the bright solitons reported by us ( see supplementary information for further discussion ) .
in order to provide further insight into our experimental observations , we have performed a series of numerical simulations .
this is important since the previous theoretical studies of polariton solitons @xcite are restricted to the linearly polarized case , while in the present setting , with orthogonally polarised pump and writing beams , coupling between the two polarisations is important and leads to the formation of elliptically polarised solitons .
the model we have used includes equations for the te ( polarisation parallel to the cavity plane ) and tm polarised optical modes and for the respective excitonic fields ( see methods for details ) . transforming from the laboratory frame to the frame moving with an unknown velocity , which is consistently determined together with the soliton profile @xcite
, we have found branches of elliptically polarised solitons , see figs .
[ fig4](a , b ) .
these solutions retain their spatial profile and all other characteristics for indefinitely long times . the soliton branches with positive slope in fig .
[ fig4](a ) represent propagating solutions , which are stable with respect to perturbations .
the negatively sloped branches correspond to unstable solitons .
these ideal solitons have been calculated on top of the spatially homogeneous pump and move with velocities @xmath58m / ps , in good agreement with experiment .
another series of numerical simulations has been conducted to demonstrate the excitation of solitons with @xmath59ps pulses on top of the gaussian pump beam , see figs .
[ fig4](c , d ) .
[ fig4](c ) ( cf . fig .
[ fig2](g ) ) and fig .
[ fig4](d ) ( cf .
fig [ fig2](h ) ) show changes of the peak intensity and of the width of wavepackets traveling across the microcavity for several intensities of the writing beam .
the saturation of both parameters with increasing intensity of the writing beam ( fig .
[ fig4](c ) ) as well as the quasi - constant soliton width in the region from 5 @xmath15 m to @xmath60 @xmath15 m of the propagation length ( fig .
[ fig4](d ) ) demonstrate the transition to the soliton regime , and is fully consistent with the experimental observations in figs .
[ fig2](g - h ) .
the observed saturation behaviour indicates that as for other dissipative solitons @xcite the soliton width in our case is fixed by the pump and cavity parameters@xcite and not by the writing beam powers .
this is in strong contrast to conservative solitons , which for the given system parameters can exist with an arbitrary width determined by the density @xmath7 initially induced by the writing beam
. has been filtered out .
( c ) and ( d ) dynamics of the soliton width and peak intensity as functions of position across the microcavity .
the writing beam is @xmath31 m wide and its duration is @xmath59ps .
different line colors correspond to the different intensities of the writing beam indicated in the inset in ( d ) .
one unit of all the plotted dimensionless field amplitudes corresponds to the pump beam amplitude @xmath61 at the right end of the lower branch of the bistability loop , see ( a ) . ]
the soliton formation can also be viewed as scattering of pump polaritons into a continuum spanning a broad range of transverse momenta from @xmath62 to @xmath63 ( see fig . [ fig1](a ) ) @xcite .
our measurements of the energy - momentum ( @xmath64 ) profile of the solitonic wavepackets over time , as shown in fig .
[ fig5 ] , confirm this point .
initially , for @xmath65ps , the polariton emission is concentrated mostly close to @xmath25 , whereas at later times ( @xmath55 ps ) the emission is distributed over a broad range of momenta a characteristic feature of soliton formation @xcite .
the soliton spectra are expected to form a straight line tangential to the dispersion of nonsolitonic radiation @xcite .
indeed , the measured dispersion at @xmath55ps can be approximated by a straight line ( dashed lines in fig .
[ fig5](b , c ) ) in agreement with our numerical modeling , see fig .
[ fig5 ] ( k , m ) .
we note however that soliton emission forms at @xmath66-vectors @xmath67 @xmath15m@xmath1 and spans over the point of inflection , where the dispersion of non - interacting polaritons is also close to linear .
nevertheless , the difference between the solitonic dispersion observed in experiment at around @xmath55ps ( in modeling at @xmath68 ps ) and the dispersion of the nonsolitonic polaritons observed at later times @xmath69ps ( in modeling at @xmath70ps ) is clear from the data shown in fig .
[ fig5 ] . the soliton velocity , @xmath71 , deduced from the dashed lines in figs . [ fig5](b , c ) is @xmath72m / ps , consistent with the measurements in fig .
[ fig1 ] and the numerical modeling .
the difference between the experimental te and tm emission in momentum space may arise from anisotropy of polariton scattering dynamics ( see for example @xcite ) , which is not taken into account in our modeling .
( k - n ) numerical simulations of polariton emission in time intervals of @xmath68 ps ( k , m ) and @xmath70 ps ( l , n ) after the application of the writing beam .
note , the span in the momentum space is wider than in the experimental figure .
dashed lines show the dispersion of the noninteracting lp branch . ]
the experimental observations of bright polariton solitons reported above open opportunities for the exploration of their potential applications in ultrafast information processing , since their picosecond response time is three orders of magnitude faster than that observed for the pure light cavity solitons in vcsels @xcite . furthermore polaritonic nonlinearities are 2 - 3 orders of magnitude larger than nonlinearities in vcsels @xcite .
the measured transverse dimensions of polaritonic solitons are @xmath56 m ( resolution limited ) , whereas the numerical model using the experimental parameters predicts @xmath73 m , several times less than the @xmath74 m width typical for vcsel solitons .
polariton solitons can potentially be used as more natural information bits than the propagating domain walls in the recently proposed integrated polariton optical circuits and gates ( polariton neurons ) @xcite .
the number of polaritons in the solitons observed is of the order of hundreds , which puts them into the category of mesoscopic structures .
realisation of spatially modulated 1d microcavity structures , with reduced numbers of particles may create conditions for observation of quantum solitons , and nonlinear functionality relying on few polariton quanta , potentially allowing designs of quantum polariton - soliton circuits .
we note finally that unlike the well - studied polariton condensates which correspond to macroscopic occupation of a single state in momentum space , the highly occupied polariton soliton is strongly localised in real space with a broad spread in energy and momentum .
in parallel with our work , amo _ et .
al . _ have very recently reported observations of dark 1d polariton solitons @xcite .
these solitons are formed by polaritons with positive mass in collision with obstacles .
a most important physical difference with the bright polariton solitons reported here arises from the fact that in the experiment of ref .
@xcite , the solitons are observed outside the pump spot , necessary to avoid locking of the dark soliton phase to the pump phase in their one - beam experiments . such an arrangement selects conservative solitons propagating in the presence of the unbalanced absorption .
we use a gaas - based device with six @xmath75 nm thick quantum wells ( qws ) grown by molecular beam epitaxy .
the experiments were performed at a temperature of @xmath76k .
propagation of the wave packets in real time was recorded using a hamamatsu streak camera with a time resolution of @xmath77ps .
an aspheric lens with a focal length of 4 cm and numerical aperture @xmath78 , provided optical resolution of about 4 - 5 @xmath15 m for both soliton excitation and imaging .
two - dimensional images of solitons at different times were reconstructed from an array of 1d images versus position x such as in fig . 1 obtained at different positions y on the pump spot .
solitons were driven by a tm polarised cw pump beam and triggered by a pulsed te polarised writing beam with duration of 5 ps .
the pump and writing beams were directed to the microcavity through the same optical path using a polarising beam splitter .
the orthogonal polarisations of the beams were chosen to ensure transmission of maximum available powers from each of the lasers to the sample .
numerical studies were performed using mean - field equations describing the evolution of slowly varying amplitudes of the tm and te cavity modes and of the corresponding excitonic fields @xmath79 : @xmath80\psi_{tm } -\frac{ig}{4}(1-r)\psi_{tm}^*\psi_{te}^2=i\omega_r e_{tm}\ ; , \label{eqpsix}\\ & & \partial_t\psi_{te } + \left(\gamma_e - i\delta_e\right)\psi_{te}+\frac{ig}{4}\left[(1+r)|\psi_{te}|^2 + 2|\psi_{tm}|^2\right]\psi_{te } -\frac{ig}{4}(1-r)\psi_{te}^*\psi_{tm}^2=i\omega_r e_{te}. \label{eqpsiy}\end{aligned}\ ] ] here @xmath81 kg is the effective cavity photon mass , @xmath82mev is the rabi splitting , @xmath83 mev are the cavity photon and the exciton coherence decay rates , @xmath84 mev , @xmath85 mev , @xmath86 is the nonlinear parameter , which can be easily scaled away , @xmath87 parameterizes the nonlinear interaction between the two modes @xcite .
@xmath88 is the pump amplitude with the momentum @xmath25 , the corresponding the angle of incidence @xmath89 $ ] and @xmath90 is the writing beam amplitude . for the case of the homogeneous pump @xmath91 and @xmath92 the soliton solutions are sought in the form @xmath93 , @xmath94 .
the soliton profiles and the unknown velocity @xmath95 are found self - consistently using newton - raphson iterations . in our simulations of the soliton excitation
the system of eqs .
( [ eqex])-([eqpsiy ] ) has been solved directly using the split - step method .
the sheffield group thanks epsrc ( ep / g001642 ) , the fp7 itn clermont 4 and the royal society for support of this work , and a. amo for a helpful discussion .
all authors prepared the manuscript and analysed experimental and numerical data ; m.s . and d.n.k .
conducted the experimental measurements ; a.v.g .
, r.h . , d.v.s . conducted the theoretical and numerical work ; d.v.s . proposed the concept ; k.b . and r.h .
fabricated the microcavity .
the authors declare no competing financial interests .
d. n. krizhanovskii and d. v. skryabin 10 url # 1`#1`urlprefix[2]#2 [ 2][]#2 , & .
_ _ * * , ( ) . & _ _ ( , ) . & .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ _ * * , ( ) . , , & _ _
. semiconductor science and technology ( , ) . , & .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) . , , , &
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _ . _
_ * * , ( ) . , , , & .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) . & ( eds . )
_ _ ( , ) . , & . in & ( eds . )
_ _ * * , ( , ) . & . in & ( eds . )
_ _ ( , ) . , , &
. _ _ * * , ( ) . , , &
_ _ * * , ( ) . .
_ _ * * , ( ) . & .
_ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _ . _ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _ . _
_ * * , ( ) . , &
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) . , &
_ _ * * , ( ) .
as was discussed in the main text , optical bistability of the pump field together with the negative polariton effective mass at @xmath57 are necessary conditions for formation of a stable solitonic wave . in this case the soliton results from switching in the bistable region ( @xmath96 ) from the lower to the upper state locally in a region of a few microns .
by contrast , if the whole pump spot is switched on , i.e the system is in the upper state in the bistability region or the k - vector of the pump is above the upper bistability threshold then a soliton can not be formed .
this is demonstrated in the following experiment .
firstly , we set the system to the upper state ( fig .
1(c ) of the main text ) , where a condensed signal state is formed at @xmath97 , with long range spatial coherence over @xmath98 , due to polariton - polariton parametric scattering from the pump state into signal and idler states at @xmath99 and @xmath100 .
this corresponds to the polariton optical parametric oscillator ( opo ) , which was studied in ref .
@xcite .
secondly , we introduce a pulsed writing beam ( wb ) at @xmath101 focused to a small spot of @xmath102 and observe propagation of the resulting wavepacket .
the wavepacket is amplified due to polariton - polariton scattering from the switched on pump state and hence propagates macroscopic distances , superimposed on the signal condensate at @xmath97 .
such a process corresponds to the so - called triggered opo ( topo ) , firstly reported by amo et .
al . in ref .
@xcite .
the topo propagating wavepacket exhibits very different physical properties to the soliton .
s1(a ) shows the topo intensity as a function of time and position @xmath103 ( see fig .
1(b ) of the main text ) for @xmath104 .
it is seen that the topo wavepacket propagates over long distances of up to @xmath105 .
however in contrast to the soliton propagation , it broadens significantly from @xmath106 up to @xmath107 within @xmath108 time and propagates at a speed 1.3 times slower than the observed soliton .
moreover the topo wavepacket intensity decays with time , again in marked contrast to soliton behaviour of fig .
the topo wavepacket created by the writing beam can be considered as a condensate excitation , which has the properties of as a diffusive goldstone mode @xcite .
there is competition between polariton - polariton scattering from the pump to the signal state at @xmath97 and to the localised topo wavepacket at @xmath39 .
since the pump state is switched on over the whole excitation spot there are no excitations in the pump state itself .
the topo wavepacket propagation speed is determined only by the writing beam k - vector @xmath101 and not by that of the pump as in the case of soliton .
the topo wavepacket expands since the polariton dispersion at @xmath101 is described by a positive effective mass .
the time dependence of the topo wavepacket fwhm is expected to be strongly dependent on the initial size of the writing beam .
it can be easily shown from solution of a time dependent schrodinger equation that the fwhm of a noninteracting wavepacket with initial fwhm @xmath109 with a dispersion described by positive effective mass @xmath110 is given by @xmath111 .
it is seen that the size of the wavepacket will increase with time more slowly for initially larger @xmath109 .
this is consistent with the above observation ( wavepacket with initial size @xmath112 is expected to spread up to @xmath113 within 50 ps ) and those in ref .
@xcite , where the topo wavepacket of a size 15 - 20 @xmath114 along the propagation direction is initially created and its size changes very little within 40 - 50 ps . indeed , using the above formula we can estimate that it should increase only by 2 microns for @xmath115 kg . | microcavity polaritons are composite half - light half - matter quasi - particles , which have recently been demonstrated to exhibit rich physical properties , such as non - equilibrium bose - einstein condensation , parametric scattering and superfluidity . at the same time
, polaritons have some important advantages over photons for information processing applications , since their excitonic component leads to weaker diffraction and stronger inter - particle interactions , implying , respectively , tighter localization and lower powers for nonlinear functionality . here
we present the first experimental observations of bright polariton solitons in a strongly coupled semiconductor microcavity .
the polariton solitons are shown to be non - diffracting high density wavepackets , that are strongly localised in real space with a corresponding broad spectrum in momentum space . unlike solitons known in other matter - wave systems such as bose condensed ultracold atomic gases , they are non - equilibrium and rely on a balance between losses and external pumping .
microcavity polariton solitons are excited on picosecond timescales , and thus have significant benefits for ultrafast switching and transfer of information over their light only counterparts , semiconductor cavity lasers ( vcsels ) , which have only nanosecond response time .
= 1 non - spreading localised wave - packets or solitons occur in a variety of non - linear classical and quantum photonic and matter - wave situations , where they reflect the distinct physical properties of the system under consideration .
light only solitons were first discovered @xcite and most actively researched in optical fibers @xcite .
fibers provide a very good experimental realisation of the one dimensional ( 1d ) nonlinear schrodinger ( nls ) equation - a paradigm model in soliton physics , which has simple soliton solutions shaped by opposing dispersive spreading and nonlinearity induced self - phase modulation of optical pulses .
subsequently , the concept of optical solitons has been extended into the spatial domain , where diffraction of propagating beams can be compensated by nonlinear changes in the material refractive index @xcite .
formally , anomalous group velocity dispersion is equivalent to diffraction : both require positive ( self - focusing ) kerr nonlinearity to form either temporal or spatial bright solitons @xcite .
recently , matter wave solitons have been demonstrated in bose condensed atomic gases , where the dispersive spreading of the localized wavepackets induced by the kinetic energy of positive mass atoms is balanced by the attractive interatomic interaction @xcite .
interactions between the atoms used in condensation experiments are often repulsive . in order to create bright solitons in these cases ,
the effective atomic mass must be made negative , as demonstrated using optical lattices @xcite .
the same effect can be achieved for light in photonic crystals @xcite .
strong exciton photon coupling in microcavities results in an unusual and advantageous lower polariton branch dispersion exhibiting regions of either positive or negative effective mass @xcite depending on the values of transverse momentum @xcite .
the transition from positive to negative mass is associated with the point of inflection of the energy - momentum diagram , shown in fig . [ fig1](a ) .
the negative mass of polaritons coupled with repulsive polariton - polariton interactions favour the formation of the bright solitons which we realise in the present work .
recent discoveries made with microcavity polaritons include condensation @xcite , vortices @xcite , superfluidity @xcite and follow a pathway similar to the one which led to the observation of coherent matter wave solitons @xcite .
the observation of low threshold bistability , polarization multistability @xcite and parametric scattering @xcite of polaritons have further prepared the necessary foundations for the realisation of half - light half - matter solitons .
in contrast to atomic systems , polariton microcavities operate under non - equilibrium conditions .
atomic systems are typically described by conservative hamiltonian models , such as nls or gross - pitaevskii equations , while the polariton system is intrinsically non - hamiltonian . in the photonic context , this system is often referred to as dissipative , which implies the importance of losses , but also implicitly assumes the presence of an external energy supply @xcite . theoretically , conservative half - light half - matter solitons have been predicted in ref .
@xcite , while dissipative bright polariton solitons in microcavities have been numerically studied in refs .
@xcite . |
consider the @xmath0 wave equation with non - homogeneous boundary conditions : @xmath4 system ( [ contwavecontrolled ] ) is said to be _ exactly controllable _ in time @xmath5 if , for all @xmath6 , there exists a control function @xmath7 such that the solution of ( [ contwavecontrolled ] ) can be driven at rest at time @xmath8 , i.e. @xmath9 we also introduce the adjoint @xmath0 wave equation with homogeneous boundary conditions : @xmath10 this system is well known to be well posed in the energy space @xmath11 and the energy below is conserved in time : @xmath12 the hilbert uniqueness method ( hum ) introduced in @xcite allows showing that the property of _ exact controllability _ for ( [ contwavecontrolled ] ) is equivalent to the _ boundary observability _ of ( [ contwaveadjoint ] ) .
the _ observability property _ of the wave equation ensures that the following _ observability inequality _ holds for all solutions of ( [ contwaveadjoint ] ) provided @xmath5 : @xmath13 the best constant @xmath14 in ( [ contobsineq ] ) is the so - called _
observability constant_. the observability time @xmath8 has to be larger than the characteristic time @xmath15 which is needed by any solution associated to initial data @xmath16 supported in a very narrow neighborhood of @xmath17 to travel along the characteristic rays @xmath18 , get to the boundary @xmath19 and come back to the boundary @xmath17 along the characteristics @xmath20 .
it is also well known that for all @xmath21 and all solutions @xmath22 of the adjoint problem ( [ contwaveadjoint ] ) with initial data @xmath23 , the following _ admissibility inequality _ holds : @xmath24 so that for all @xmath5 , @xmath25 and @xmath26 are equivalent norms . as a consequence of these results ,
it is easy to see that for all @xmath6 , there exists a control @xmath7 driving the solution of ( [ contwavecontrolled ] ) to the rest at @xmath27 , i.e. such that ( [ contnullcontrol ] ) holds .
this turns out to be equivalent to the fact that @xmath28 for all solutions @xmath22 of ( [ contwaveadjoint ] ) , where @xmath29 is the duality product between @xmath30 and @xmath31 . the hum control @xmath32 , the one of minimal @xmath33-norm , has the explicit form @xmath34 where @xmath35 is the solution corresponding to the minimum @xmath36 of the quadratic functional @xmath37 the effect of substituting the continuous controlled wave equation ( [ contwavecontrolled ] ) or the corresponding adjoint problem ( [ contwaveadjoint ] ) by discrete models has been intensively studied during the last years , starting from some simple numerical schemes on uniform meshes like finite differences or linear @xmath2 - finite element methods in @xcite and , more recently , more complex schemes like the discontinuous galerkin ones in @xcite .
in all these cases , the convergence of the approximation scheme in the classical sense of the numerical analysis does not suffice to guarantee that the sequence of discrete controls converges to the continuous ones , as one could expect .
this is due to the fact that there are classes of initial data for the discrete adjoint problem generating high frequency wave packets propagating at a very low group velocity and that , consequently , are impossible to be observed from the boundary of the domain during a finite time , uniformly as the mesh - size parameter tends to zero .
this leads to the divergence of the discrete observability constant as the mesh size tends to zero .
high frequency pathological phenomena have also been observed for numerical approximation schemes of other models , like the linear schrdinger equation ( cf .
@xcite ) , in which one is interested in the uniformity of the so - called _ dispersive estimates _ , which play an important role in the study of the well - posedness of the non - linear models .
several _ filtering techniques _ have been designed to face these high frequency pathologies : the _ fourier truncation method _
@xcite ) , which simply eliminates all the high frequency fourier components propagating non - uniformly ; the _ bi - grid algorithm _ introduced in @xcite and @xcite and rigorously studied in @xcite , @xcite or @xcite in the context of the finite differences semi - discretization of the @xmath0 and @xmath38 wave equation and of the schrdinger equation ( cf.@xcite ) , which consists in taking initial data with slow oscillations obtained by linear interpolation from data given on a coarser grid ; and the _ numerical viscosity _ method , which by adding a suitable dissipative mechanism damps out the spurious high frequencies ( @xcite , @xcite ) .
we should emphasize that the _ mixed finite element method _ analyzed in @xcite is , as far as we know , the unique method that preserves the propagation and controllability properties of the continuous model uniformly in the mesh size without requiring any extra filtering .
the interested reader is referred to the survey articles @xcite and @xcite for a presentation of the development and the state of the art in this topic .
the purpose of the present paper is to analyze the behavior of the quadratic @xmath1 - finite element semi - discretization of the problems ( [ contwavecontrolled ] ) and ( [ contwaveadjoint ] ) from the uniform controllability / observability point of view . in * section [ sectintrop2 ]
* we introduce in a rigorous way the discrete analogue of ( [ contwavecontrolled ] ) and ( [ contwaveadjoint ] ) and explain the minimization process generating the discrete controls . in * section [ sectfour ] * we analyze the spectral problem associated to this discrete model and reveal the co - existence of two main types of fourier modes : an _ acoustic _ one , of physical nature , related to the nodal components of the numerical solution , and an _ optic _ one , of spurious nature , related to the curvature with which the quadratic finite element method perturbs the linear approximation .
we also study finer properties of the spectrum , for example the _ spectral gap _ , identifying three regions of null gap : the highest frequencies on both acoustic and optic modes and the lowest frequencies on the optic one .
the content of this section is related to previously existing work .
for instance , the dispersive properties of higher - order finite element methods have been analyzed in @xcite in the setting of the helmholtz equation .
an explicit form of the acoustic dispersion relation was obtained for approximations of arbitrary order .
it was shown that the numerical dispersion displays three different types of behavior , depending on the approximation order relative to the mesh - size and the wave number . in * section [ sectboundobseig ]
* we obtain some spectral identities allowing us to analyze the discrete observability inequality for the adjoint system . in * section [ sectingham ] * we show that the _ fourier truncation _ of the three pathological regions of the spectrum leads to an uniform observability inequality . in * section [ sectbigrid ] * we prove that a filtering mechanism consisting in , firstly , considering piecewise linear initial data and , secondly , preconditionning the nodal components by a bi - grid algorithm guarantees uniform observability properties . within the proof
, we use a classical _ dyadic decomposition argument _
@xcite ) , which mainly relies on the fact that for this class of initial data the total energy can be bounded by the energy of the projection on the low frequency components of the acoustic dispersion relation .
we should emphasize that our results are finer than the ones in @xcite or @xcite , where one obtains uniform observability properties for finite element approximations of any order , but by filtering the fourier modes much under the critical scale @xmath39 .
here we only consider the particular case of quadratic finite element approximation on @xmath0 meshes , but we get to the critical filtering scale @xmath39 .
note however that the results in @xcite and @xcite apply in the context of non - uniform grids as well . in * section [ sectconvergence ]
* we present the main steps of the proof of the convergence of the discrete control problem under the assumption that the initial data in the corresponding adjoint problem are filtered through a fourier truncation or a bi - grid algorithm .
* section [ sectopenpbms ] * is devoted to present the conclusions of the paper and some related open problems .
let @xmath40 , @xmath41 and @xmath42 be the _ nodes _ of an _ uniform grid _ of the interval @xmath43 $ ] , with @xmath44 , @xmath45 , constituted by the subintervals @xmath46 , with @xmath47 .
we also define the _ midpoints _ @xmath48 of this grid , with @xmath47 .
let us introduce the space @xmath49 of polynomials of order @xmath50 on the interval @xmath51 and the _ space of piecewise quadratic and continuous functions _ @xmath52 .
the space @xmath53 can be written as @xmath54 , where the two classes of basis functions are explicitly given below ( see fig .
[ figbasis ] ) @xmath55 ( left ) and @xmath56 ( right).,title="fig:",width=207,height=170 ] ( left ) and @xmath56 ( right).,title="fig:",width=207,height=170 ] the quadratic approximation of the adjoint problem ( [ contwaveadjoint ] ) is as follows : @xmath57 since @xmath58 , it admits the decomposition @xmath59 .
the function @xmath60 can be identified with the vector of its coefficients , @xmath61 ( in the sequel , all vectors under consideration will be column vectors ) .
thus , using @xmath62 , @xmath63 , as test functions in ( [ p2waveadjointvar ] ) , system ( [ p2waveadjointvar ] ) can be written as the following system of @xmath64 second - order linear differential equations ( odes ) : @xmath65 where @xmath66 and @xmath67 are the following @xmath68 _ pentha - diagonal mass _ and _ stiffness _ matrices @xmath69 and @xmath70 we introduce the discrete analogue of @xmath71 , @xmath72 and @xmath73 spaces to be @xmath74 for the elements of the space @xmath75 we have to impose the additional requirement @xmath76 .
the inner products defining the discrete spaces @xmath77 , @xmath78 , are given by @xmath79 and the norms are given by @xmath80 , for all @xmath78 .
here , @xmath81 is the inner product in the euclidian space @xmath82 , defined by @xmath83 ( the overline symbol denotes complex conjugation ) . for @xmath84 with coefficients
@xmath85 , we introduce the following notations for the three possible discrete derivatives on each nodal point ( the _ forward _ , the _ backward _ and the _ midpoint _ one ) : @xmath86 and the values of @xmath87 at @xmath88 and @xmath89 , @xmath47 : @xmath90 with these notations , it is easy to check that the @xmath91 and @xmath92 - norms admit the following representations : @xmath93 set @xmath94 and its dual @xmath95 , the duality product @xmath96 between @xmath97 and @xmath98 being defined as @xmath99 . problem ( [ p2adjoint ] ) is well posed in @xmath100 .
the total energy of its solutions is conserved in time : @xmath101 one of the goals of this paper is to study discrete versions of the observability inequality ( [ contobsineq ] ) of the form @xmath102 where @xmath103 is a @xmath104 _ observability matrix operator_. the observability inequality ( [ p2observabilityinequalitygeneral ] ) makes sense for rather general matrices @xmath103 , corresponding , for instance , to the observability from any open subset contained in the spatial domain @xmath105 .
but , within this paper , we focus on the particular case of _ boundary observation operators _ @xmath103 , in the sense that they approximate the normal derivative @xmath106 of the solution of the continuous adjoint problem ( [ contwaveadjoint ] ) at @xmath17 .
the main example of such boundary matrix operators @xmath103 that will be used throughout this paper is as follows : @xmath107 the operator @xmath103 is also the one used for the finite difference semi - discretization in @xcite . let us remark that at the discrete level there are different ways to approximate the normal derivative of the continuous solution of ( [ contwaveadjoint ] ) . since @xmath108 is a vector and @xmath106 is a scalar , the way in which @xmath108 approximates @xmath109 needs to be further explained .
remark that @xmath103 in ( [ observationoperators ] ) is almost a null matrix , excepting the penultimate component on the last row .
the last component of @xmath108 , the only non - trivial one , equals to @xmath110 .
the consistency analysis shows that if @xmath111 is a sufficiently regular function and @xmath87 is its quadratic interpolation , then @xmath112 is a first - order approximation of @xmath113 .
we are interested in observability inequalities ( [ p2observabilityinequalitygeneral ] ) in a finite , but sufficiently large observability time , say @xmath114 . in this paper
we show that , when working on the whole discrete space @xmath115 , the observability constant @xmath116 blows - up as @xmath117 , whatever @xmath21 is .
one of the main contributions of this paper is to design appropriate subspaces @xmath118 on which the observability constant @xmath116 is uniformly bounded as @xmath117 .
we will also prove that the discrete version of ( [ contdirectinequality ] ) below holds uniformly as @xmath117 in the approximate space @xmath98 for all @xmath21 : @xmath119 once the observability problem is well understood , we are in conditions to address the discrete control problem . for a discrete control function @xmath98 , we consider the following non - homogeneous problem : @xmath120 here , the superscript @xmath121 denotes matrix transposition
. multiplying system ( [ p2controlledpbm ] ) by any solution @xmath122 of the adjoint problem ( [ p2adjoint ] ) , integrating in time and imposing that at @xmath27 the solution is at rest , i.e. @xmath123 we obtain that @xmath124 necessarily satisfies the following identity which fully characterizes all possible exact controls @xmath124 : @xmath125 in view of this , we introduce the following discrete version of the quadratic functional ( [ contfunctional ] ) : @xmath126 @xmath122 being the solution of the adjoint problem ( [ p2adjoint ] ) with initial data @xmath127 and @xmath128 the initial data to be controlled in ( [ p2controlledpbm ] ) .
the functional @xmath129 is continuous and strictly convex .
thus , provided it is coercive ( which is actually what the uniform observability inequality guarantees ) , it has an unique minimizer @xmath130 whose euler - lagrange equations are as follows : @xmath131 for all @xmath132 and @xmath122 the corresponding solution of ( [ p2adjoint ] ) .
the discrete hum control is then @xmath133 let us briefly comment the analogies between the identities ( [ conthumcontrol ] ) and ( [ p2optimalcontrol ] ) .
as we said , when @xmath103 is a boundary observability matrix operator , like for example the one in ( [ observationoperators ] ) , @xmath134 is a vector whose last component @xmath135 , the only non - trivial one , approximates the normal derivative of the solution to the adjoint continuous wave equation ( [ contwaveadjoint ] ) .
accordingly , the controls @xmath136 only act on @xmath137 when @xmath124 is the numerical control obtained by ( [ p2optimalcontrol ] ) .
consequently , the boundary observability operator @xmath103 does not act really at @xmath17 , but at @xmath138 , being in fact an internal control acting on a single point which is closer and closer to the boundary as the mesh size parameter becomes smaller , so that in the limit as @xmath117 it becomes a boundary control .
observe also that the control problem we deal with is a coupled system of non - homogeneous odes modeling the interaction between the nodal and the midpoint components .
thus , the node @xmath139 lies in fact on the boundary of the midpoint component .
consequently , the controls @xmath136 in ( [ p2controlledpbm ] ) , with @xmath124 as in ( [ p2optimalcontrol ] ) , are really natural approximations of the continuous boundary controls @xmath32 in ( [ contwavecontrolled ] ) .
for the sake of completeness , we recall the spectral analysis of this quadratic finite element method , following @xcite . the spectral problem associated to the adjoint system ( [ p2waveadjointvar ] )
is as follows : @xmath140 due to the symmetry and the coercivity of the bi - linear forms generated by the scalar products @xmath141 and @xmath142 , we have @xmath143 .
let @xmath144 be the components of the eigenfunction @xmath145 .
the pair @xmath146 is a generalized eigensolution corresponding to the pair of matrices @xmath147 , i.e. @xmath148 consider the normalized eigenvalues @xmath149 .
the system ( [ p2spectral ] ) is a pair of two equations : @xmath150 for @xmath47 , and @xmath151 depending on @xmath152 and with the boundary condition @xmath153 .
* the acoustic and optic modes . * from ( [ p2spectral1 ] )
, we obtain that for @xmath154 , the values of @xmath155 at the midpoints can be obtained according to the two neighboring nodal values as follows : @xmath156 replacing ( [ p2spectral3 ] ) into ( [ p2spectral2 ] ) , we obtain that @xmath157 it is easy to check that @xmath158 , with @xmath159 , solves ( [ p2spectral4 ] )
. then the normalized eigenvalues @xmath160 verify the identity @xmath161 for each @xmath162 $ ] , consider the following second - order algebraic equation in @xmath163 : @xmath164 whose solutions are @xmath165 with @xmath166 .
the superscripts @xmath167 and @xmath168 stand for _ acoustic _ and _ optic_. we will also need the square roots of the fourier symbols ( [ fouriersymbols ] ) , the so called _ dispersion relations _ : @xmath169 for each @xmath159 , set @xmath170 we refer to @xmath171 and @xmath172 as the _ acoustic _ and the _ optic branch _ of the spectrum .
the corresponding eigenvectors are @xmath173 * the resonant mode . * up to this moment , we have explicitly calculated @xmath174 solutions of the eigenvalue problem ( [ p2spectral ] ) . to do this
, we have supposed that @xmath154 .
but @xmath175 is also an eigenvalue corresponding to the _ resonant _ mode , the superscript @xmath176 standing for _ resonant_. the components of the corresponding eigenvector @xmath177 are @xmath178 for any normalized eigenvalue @xmath160 , define @xmath179 , @xmath180 , @xmath181 ( see figure [ figp2eigenvalues ] ) .
( left ) versus their square roots @xmath182 ( right ) : the continuous ones ( blue ) , the acoustic ( red ) , the optic ( black ) and the resonant ( green ) modes .
, title="fig:",width=283,height=188 ] ( left ) versus their square roots @xmath182 ( right ) : the continuous ones ( blue ) , the acoustic ( red ) , the optic ( black ) and the resonant ( green ) modes .
, title="fig:",width=283,height=188 ] * normalized eigenvectors . * for any eigenvector @xmath183 , we define the @xmath3-normalized eigenvector @xmath184 . . in blue / red ,
we represent their nodal / midpoint components.,title="fig:",width=207,height=170 ] . in blue / red ,
we represent their nodal / midpoint components.,title="fig:",width=207,height=170 ] . in blue / red ,
we represent their nodal / midpoint components.,title="fig:",width=207,height=170 ] + . in blue / red ,
we represent their nodal / midpoint components.,title="fig:",width=207,height=170 ] . in blue / red ,
we represent their nodal / midpoint components.,title="fig:",width=207,height=170 ] . in blue / red ,
we represent their nodal / midpoint components.,title="fig:",width=207,height=170 ] + . in blue / red ,
we represent their nodal / midpoint components.,title="fig:",width=207,height=170 ] . in blue / red ,
we represent their nodal / midpoint components.,title="fig:",width=207,height=170 ] . in blue / red , we represent their nodal / midpoint components.,title="fig:",width=207,height=170 ] using the expression ( [ discretenorms ] ) of the discrete norm @xmath92 and @xmath91 and the identity ( [ p2spectral3 ] ) , we obtain the following representations of the @xmath92 and @xmath91-norms of the acoustic and optic eigenvectors in terms of their nodal components , for all @xmath185 and @xmath159 : @xmath186h\sum\limits_{j=0}^n|\tilde{\varphi}_{j+1}^{\alpha , k}+\tilde{\varphi}_j^{\alpha , k}|^2 + \frac{h}{12}\sum\limits_{j=0}^n|\tilde{\varphi}_{j+1}^{\alpha , k}-\tilde{\varphi}_j^{\alpha , k}|^2\label{h0normeig}\ ] ] and @xmath187 then , using the representation formula ( [ h0normeig ] ) , the identities @xmath188 and ( [ p2spectral5 ] ) , we obtain @xmath189 let us remark that @xmath190 blows - up as @xmath191 . with the above notation , @xmath192 where @xmath193 @xmath194 and @xmath195 standing for the _ nodal _ and _ midpoint _ components . using the explicit form of the @xmath92 - norm and the characterization of the resonant mode ( [ p2resonanteigenvector ] ) , we obtain that @xmath196 and therefore , the normalized resonant mode @xmath197 satisfies @xmath198 let us introduce the sets of eigenvalues and of @xmath199-normalized eigenfunctions , i.e. @xmath200 * fourier representation of discrete solutions .
* since @xmath201 is an orthonormal basis in @xmath199 , the initial data in ( [ p2adjoint ] ) admit the following fourier representation : @xmath202 with @xmath203 , @xmath159 and @xmath204 .
correspondingly , the solution of ( [ p2adjoint ] ) can be represented as follows : @xmath205 , \label{p2adjointsol}\ ] ] where @xmath206 the total energy ( [ p2energy ] ) of the solutions of ( [ p2adjoint ] ) is then as follows : @xmath207 * limits and monotonicity of the eigenvalues .
* firstly , let us remark that as @xmath191 , @xmath208 , @xmath209 and as @xmath210 , @xmath211 . on the other hand , the so - called _ group velocities _ , which are first - order derivatives of the fourier symbols ( [ fouriersymbols ] ) or of the corresponding dispersion relations ( [ dispersionrelations ] ) , verify the following positivity condition @xmath212 which means that _ the acoustic branch is strictly increasing _ and _ the optic one is strictly decreasing _ in @xmath213 .
consequently , the high frequency wave packets involving only the acoustic or the optic modes and concentrated around a given frequency @xmath214 propagate in opposite directions .
moreover , at @xmath215 or @xmath216 , the group velocities satisfy @xmath217 which , according to the analysis in @xcite , shows , in particular , that there are waves concentrated on each mode which propagate at arbitrarily slow velocity .
the main result of this section is as follows : for all @xmath203 and all @xmath159 , the following identity holds for both acoustic and optic eigensolutions : @xmath218 moreover , for the resonant mode , the following identity holds : @xmath219 [ propobservabilityeigenvectors ] the identity ( [ observabilityeigenvectorsao ] ) is the discrete analogue of the continuous one @xmath220 , where @xmath221 is the @xmath3-normalized eigenfunction corresponding to the eigenvalue @xmath222.[remark0 ] due to the monotonicity of the fourier symbols , we have that @xmath223 and @xmath224 , for all @xmath159 . the quadratic equation @xmath225 has the roots @xmath226 and @xmath227 .
this allows us to guarantee that @xmath228 , for all @xmath229.[remark1 ] due to the form of the denominator in the right hand side of ( [ observabilityeigenvectorsao ] ) and from the above lower and upper bound of the fourier symbols and the behavior of the group velocities , we deduce that the coefficient of @xmath230 in the right hand side of ( [ observabilityeigenvectorsao ] ) is singular as @xmath231 both when @xmath232 or @xmath233 and when @xmath234 and @xmath233 .
[ remark2 ] fix @xmath203 and @xmath159 .
obviously , it is enough to prove ( [ observabilityeigenvectorsao ] ) for the un - normalized eigenvectors @xmath235 .
we will use two approaches to prove the identity ( [ observabilityeigenvectorsao ] ) .
the first one consists on using the classical multiplier @xmath236 ( which is a discrete version of the continuous one @xmath237 ) in the simplified spectral problem ( [ p2spectral6 ] ) and then to apply the _ abel summation by parts formula _
@xmath238 for all @xmath239 and @xmath240 . in
what follows , we will add the superscript @xmath241 to the solution @xmath155 of ( [ p2spectral4 ] ) . in this way , we deduce the following identity : @xmath242 replacing the crossed sum @xmath243 obtained from the identity with @xmath244 into the one with @xmath245 , we get the following equality : @xmath246 using the representation ( [ h1normeig ] ) of the @xmath91-norm of the optic and acoustic eigenvectors in terms of the nodal components , we obtain that ( [ obseig ] ) is equivalent to @xmath247 replacing the representation of the @xmath92-norm of the eigenvectors ( [ h0normeig ] ) into the one of the @xmath91-norm , ( [ h1normeig ] ) , we obtain : @xmath248 on the other hand , the @xmath92 and @xmath91-norms of the eigenvetors are related as follows : @xmath249 replacing ( [ h0h1normeig ] ) into ( [ obseig2 ] ) , we get @xmath250 } h^2||\mathbf{\tilde{\varphi}}^{\alpha , k}_h||_{h,1}^2.\label{obseig3}\ ] ] by combining ( [ obseig1 ] ) and ( [ obseig3 ] ) , we obtain @xmath251 from which the identity ( [ observabilityeigenvectorsao ] ) follows immediately . the second approach to prove ( [ observabilityeigenvectorsao ] )
is much more direct .
it consists in using the representation ( [ h1normeig ] ) of the @xmath91-norm of the eigenvectors , the trigonometric identities ( [ trigonometricidentities ] ) , the fact that @xmath252 and the relation ( [ p2spectral5 ] ) .
thus , for @xmath253 as in ( [ observabilityeigenvectorsao ] ) , we get : @xmath254 the identity ( [ observabilityeigenvectorsr ] ) follows by combining the explicit expressions of the components of the resonant eigenvector ( [ p2resonanteigenvector ] ) and ( [ discretenorms ] ) .
this concludes the proof of ( [ observabilityeigenvectorsr ] ) .
* the observability inequality . * in this section we prove that the discrete observability inequality ( [ p2observabilityinequalitygeneral ] ) holds uniformly as @xmath117 in a truncated class of initial data for the observation operator @xmath103 introduced in ( [ observationoperators ] ) .
more precisely , consider @xmath255 and @xmath256 and correspondingly the wave numbers @xmath257 and introduce the subspace of @xmath258 given by @xmath259 consider the truncated subspace @xmath118 defined by ( see figure [ figp2truncatedclass ] ) : @xmath260 .
in red / black / green , the acoustic / optic / resonant mode . , width=340,height=170 ] the main result of this subsection is as follows : for all @xmath261 and @xmath262 independent of @xmath263 , all initial data @xmath264 introduced in ( [ p2subsptruncation ] ) , with @xmath265 , @xmath266 given by ( [ wavenumbers ] ) , and all @xmath267}\partial_{\eta}\lambda^a,\min\limits_{\eta\in[\eta_+^o,\eta_-^o ] } ( -\partial_{\eta}\lambda^o(\eta))\}}\]][theoremfouriertruncation ] the observability inequality ( [ p2observabilityinequalitygeneral ] ) holds uniformly as @xmath117 for the operator @xmath103 in ( [ observationoperators ] ) . the fact that the eigenmodes involved in the class @xmath268 are such that the gap in each branch has a strictly positive lower bound uniformly as @xmath117 allows us to apply ingham theorem ( cf .
@xcite , theorem 4.3 , pp .
more precisely , the spectral gap on each branch is bounded as follows : @xmath269}\partial_{\eta}\lambda^a(\eta)>0,\ \forall k\pi h,(k+1)\pi h\in[0,\eta_+^a]\ ] ] and @xmath270}(-\partial_{\eta}\lambda^{o}(\eta))>0,\ \forall k\pi h,(k+1)\pi h\in[\eta_+^o,\eta_-^o].\ ] ] then @xmath271}\partial_{\eta}\lambda^a(\eta ) , \min\limits_{\eta\in[\eta_+^o,\eta_-^o]}(-\partial_{\eta}\lambda^{o}(\eta))\}>0 $ ] is the uniform gap needed to apply the ingham theory . from ( [ p2adjointsol ] ) and the definition of the class @xmath268
, we have @xmath272}\widehat{u}^{a , k}_{\pm}\exp(\pm it\lambda_h^{a , k})\varphi^{a , k}_n + \sum\limits_{k\pi h\in[\eta_+^o,\eta_-^o]}\widehat{u}^{o , k}_{\pm}\exp(\pm it\lambda_h^{o , k})\varphi^{o , k}_n\big].\ ] ] by applying the _ inverse inequality _ in ingham theorem ( cf .
@xcite , pp .
60 , ( 4.9 ) ) , we can guarantee that for all @xmath273 , there exists a constant @xmath274 independent of @xmath263 such that @xmath275}|\widehat{u}^{a , k}_{\pm}|^2\big|\frac{\varphi_n^{a , k}}{h}\big|^2 + \sum\limits_{k\pi h\in[\eta_+^o,\eta_-^o]}|\widehat{u}^{o , k}_{\pm}|^2\big|\frac{\varphi_n^{o , k}}{h}\big|^2 \big]\leq \int\limits_0^t\big|\frac{u_n(t)}{h}\big|^2\,dt.\ ] ] using the identities ( [ observabilityeigenvectorsao ] ) and @xmath276 , we obtain that ( the function @xmath253 below is as in ( [ observabilityeigenvectorsao ] ) ) : @xmath275}\lambda_h^{a , k}|\widehat{u}^{a , k}_{\pm}|^2w(\lambda^{a , k } ) + \sum\limits_{k\pi h\in[\eta_+^o,\eta_-^o]}\lambda^{o , k}_h|\widehat{u}^{o , k}_{\pm}|^2w(\lambda^{o , k } ) \big]\leq \int\limits_0^t\big|\frac{u_n(t)}{h}\big|^2\,dt.\ ] ] taking into account that for our filtering algorithm @xmath277\cup[\lambda_-^o,\lambda_+^o]\subset[0,10)\cup(12,60)$ ] ( with uniform inclusion as @xmath117 due to the fact that @xmath278 and @xmath279 do not depend on @xmath263 ) , and that for @xmath280 , the weight @xmath253 is strictly positive , we can guarantee that @xmath281 then the proof concludes by taking in ( [ p2observabilityinequalitygeneral ] ) with @xmath103 given by ( [ observationoperators ] ) the observability constant @xmath282 which is independent of @xmath263 . * the admissibility inequality .
* using the _ direct inequality _ in ingham theorem ( cf .
@xcite , pp .
60 , ( 4.8 ) ) , we can also prove that the inequality ( [ p2directinequalitygeneral ] ) , with @xmath103 as in ( [ observationoperators ] ) , holds uniformly as @xmath117 for all @xmath21 , with @xmath283 , where @xmath284 is the constant of the direct -ingham inequality and @xmath285
* 1 . the observability inequality . * in this section , @xmath286 will be an _ odd _ number .
we consider the space @xmath287 containing _ piecewise linear functions _ given below @xmath288 and the space @xmath289 of the discrete functions whose _ nodal components _ are given by a _ bi - grid algorithm _ , i.e. the even components are arbitrarily given and the odd ones are computed as average of the two even neighboring values : @xmath290 we also define the subspace @xmath291 of @xmath98 @xmath292 the aim of this section is to prove that the observability inequality ( [ p2observabilityinequalitygeneral ] ) still holds uniformly as @xmath117 for initial data in the bi - grid subspace @xmath291 introduced in ( [ p2subsbigrid ] ) : [ theorembigrid]for all @xmath293 and all initial data @xmath127 in the adjoint problem ( [ p2adjoint ] ) belonging to @xmath291 introduced in ( [ p2subsbigrid ] ) , the observability inequality ( [ p2observabilityinequalitygeneral ] ) with @xmath103 given by ( [ observationoperators ] ) holds uniformly as @xmath117 .
note that , in the bi - grid filtering mechanism we have designed , the data under consideration have been taken , before filtering through the classical bi - grid algorithm , to be piecewise linear in each interval @xmath294 , @xmath295 , which imposes a further restriction .
this allows to obtain the sharp observability time .
the bi - grid filtering algorithm proposed in theorem [ theorembigrid ] yields optimal observability time , i.e. the characteristic one @xmath296 .
this is due to the fact that for a numerical scheme the minimal time required for the observability to hold is @xmath297 , where @xmath32 is the minimal group velocity involved in the corresponding solution . from our analysis
, we will see that the bi - grid filtering algorithm above acts mainly as a fourier truncation of the whole optic diagram and of the second half ( the high frequency one ) of the acoustic one .
consequently , @xmath298}\partial_{\eta}\lambda^{a}(\eta)$ ] .
since the group velocity of the acoustic branch , @xmath299 , is increasing on @xmath300 $ ] , we conclude that @xmath301 and then the observability time of the numerical scheme is sharp : @xmath302 .
the following two auxiliary results hold : if the initial data @xmath127 in ( [ p2adjoint ] ) belong to @xmath303 , then the resonant fourier coefficients in ( [ p2adjointsol ] ) vanish , i.e. @xmath304 and the optic and acoustic ones are related by the following two identities : @xmath305 and @xmath306 [ proplineardata ] taking squares in ( [ opticalacoustical1 ] ) and ( [ opticalacoustical2 ] ) and in view of ( [ p2eignormalized ] ) , we deduce that @xmath307 we will prove only ( [ nullresonant ] ) and ( [ opticalacoustical1 ] ) , the proof of ( [ opticalacoustical2 ] ) being similar to the one for ( [ opticalacoustical1 ] ) .
observe that the fourier representation of the identity that characterizes @xmath308 is @xmath309\sin(k\pi x_{j+1/2})\\&+(\widehat{u}^r_++\widehat{u}^r_- ) \frac{\sqrt{15}}{2\sqrt{2}}\sin((n+1)\pi x_{j+1/2}),\nonumber\end{aligned}\ ] ] for all @xmath47 . multiplying ( [ bigrid1 ] ) by @xmath310 , @xmath311 , adding in @xmath47 and taking into account that @xmath312 , for all @xmath313 , we conclude the two identities ( [ nullresonant ] ) and ( [ opticalacoustical1 ] ) .
the total energy of data @xmath314 in ( [ p2adjoint ] ) can be written only in terms of the nodal components and coincides with the one of the @xmath2-finite element method @xmath315 taking into account the form of the fourier coefficients ( [ nullresonant ] ) and ( [ opticalacoustical ] ) corresponding to linear initial data , we obtain that the fourier representation of the total energy ( [ p2energylineardata ] ) is as follows : latexmath:[\[\mathcal{e}_h(\mathbf{u}_h^0,\mathbf{u}_h^1)=\frac{1}{2 } \sum\limits_{k=1}^n\lambda^{a , k}_h\big[\big(1+\frac{\lambda^{o , k}}{\lambda^{a , k}}\frac{w_1(\lambda^{a , k})}{w_1(\lambda^{o , k})}\big ) + \big(1+\frac{w_1(\lambda^{a , k})}{w_1(\lambda^{o ,
k})}\big ) the second auxiliary result establishes that for initial data in the bi - grid subspace @xmath291 in ( [ p2subsbigrid ] ) , the high frequency fourier coefficients on the acoustic branch can be evaluated in terms of the low frequency ones : [ proplowhighfreqacoustical]for each element @xmath132 introduced in ( [ p2subsbigrid ] ) , the following identities hold : @xmath317 @xmath318 and @xmath319 for all @xmath320 , with @xmath321 taking squares in ( [ lowhighfreqacou1]-[lowhighfreqacou2 ] ) , we obtain that @xmath322 where @xmath323 due to the orthogonality in @xmath199 of the eigenvectors in @xmath201 , to the fact that @xmath308 and using the representation ( [ p2eignormalized ] ) of the normalized eigenvectors , the following identity holds @xmath324 now , using the fact that @xmath325 , we have @xmath326 taking ( [ p2spectral5 ] ) into account ,
we have @xmath327 for @xmath328 , we obtain @xmath329 . since @xmath330 , we obtain ( [ lowhighfreqacou1 ] ) by equating the expressions of @xmath331 form the identity ( [ proof2 ] ) corresponding to @xmath213 and to @xmath332 .
the proof of ( [ lowhighfreqacou2 ] ) is similar , based on the fact that @xmath333 , from which for @xmath328 we obtain that @xmath334 , which concludes ( [ nullmidfrequency ] ) . replacing the fourier representations ( [ lowhighfreqacou3 ] ) and ( [ lowhighfreqacou4 ] ) into the total energy with linear initial data ( [ p2energylineardatafourier ] )
, we obtain that energy corresponding to initial data @xmath132 in ( [ p2subsbigrid ] ) is given by @xmath335,\label{p2energlinearbigriddatafourier}\ ] ] where the low and high frequency coefficients are given by @xmath336 and @xmath337 for any @xmath338 which does not depend on @xmath263 and any solution ( [ p2adjointsol ] ) of ( [ p2adjoint ] ) , let us introduce its projection on the first @xmath339 frequencies of the acoustic branch to be @xmath340 observe that the projection ( [ p2projacoudelta ] ) is still a solution of ( [ p2adjoint ] ) , therefore its total energy is conserved in time .
set @xmath341 the following result provides a bound of the total energy of the solutions of ( [ p2adjoint ] ) with initial data @xmath132 as in ( [ p2subsbigrid ] ) in terms of the total energy of their projections on the first half of the acoustic mode : for any solution @xmath342 of ( [ p2adjoint ] ) with initial data @xmath132 introduced in ( [ p2subsbigrid ] ) , there exists a constant @xmath343 which does not depend on @xmath263 such that @xmath344[propprojacousticallowfreq ] in order to obtain the estimate ( [ estimprojacousticallowfreq ] ) , we claim that it is sufficient to prove that there exist @xmath345 independent of @xmath263 such that @xmath346 and @xmath347 , for all @xmath320 . assuming this for a moment
, we can take @xmath348 for which ( [ estimprojacousticallowfreq ] ) holds .
let us analyze the boundedness of the four coefficients . observe that @xmath349 involves the product of @xmath350 and @xmath351 for @xmath352 .
but @xmath353 is singular only for @xmath354 , whereas for @xmath352 , due to the increasing nature of @xmath355 , we have @xmath356 . also @xmath357 is singular as @xmath358 , but since @xmath359 is decreasing in @xmath360 we have @xmath361 for all @xmath352 . in the same way , @xmath362 contains the product of @xmath363 with @xmath364 , for @xmath352 . the second factor
@xmath365 has a singularity as @xmath358 , but is evaluated for @xmath366 which is far from the singularity for all @xmath352 . since @xmath353 contains a factor @xmath367 , we deduce that @xmath368 has the same singularities as @xmath353 , i.e. @xmath369 , but since we work on the first half of the acoustic diagram , @xmath370 is far from that singularity .
we conclude the existence of @xmath371 independent of @xmath263 such that @xmath346 .
the coefficient @xmath372 contains two terms .
the first of them is constituted by the factors @xmath373 and @xmath374 .
since @xmath375 is not singular for any @xmath229 , then @xmath373 is bounded for all @xmath352 . on the other hand
, @xmath376 has three singularities : @xmath369 , @xmath377 and @xmath378 .
since all @xmath160-s situated on the first half of the acoustic mode are far for all the three singularities , we deduce the boundedness of @xmath374 .
the second term is a product of four factors : @xmath379 , @xmath373 , @xmath380 and @xmath374 . in view of our previous analysis
, we deduce the boundedness of the last three factors .
the first factor blows - up like @xmath381 for small @xmath213 .
nevertheless , in the same range of @xmath213-s , @xmath373 behaves like @xmath382 , compensating in this way the singularity of the first factor @xmath379 , so that @xmath383 is bounded for all @xmath352 .
the coefficient @xmath384 also contains two terms .
the first of them includes the factors : @xmath385 , @xmath386 , @xmath373 and @xmath374 .
we have already analyzed the first , the third and the fourth ones . the second one blows - up like @xmath387 for small @xmath213 .
but , as we said , @xmath373 behaves like @xmath388 for small @xmath213 , compensating the singularity of the second factor , so that @xmath389 is bounded for all @xmath352 .
the second term in @xmath384 contains the factors @xmath386 , @xmath390 , @xmath373 and @xmath374 and is bounded by the same arguments we used for the first term . consequently , there exist @xmath391 such that @xmath347 , which concludes the proof .
[ remark3]set @xmath392 and @xmath393 .
observe that @xmath394 and @xmath395 , where @xmath291 is the space in ( [ p2subsbigrid ] ) .
we want to point out that the result of theorem [ theorembigrid ] is not longer true when replace @xmath291 by @xmath396 , with @xmath397 , so that the condition on the initial data to be linear is sharp . indeed , when replacing @xmath291 by @xmath396 , in particular @xmath362 in ( [ p2energlinearbigriddatafourier ] ) has to be substituted by @xmath398 one can show that , for @xmath397 , it is not longer true that @xmath399 as @xmath210 , so that this factor can not compensate the singularity of @xmath386 as @xmath210 , like for @xmath400 .
consequently , for @xmath397 , at least @xmath401 is not bounded for @xmath320 .
proposition [ proplowhighfreqacoustical ] ensures that the total energy of initial data in @xmath291 introduced in ( [ p2subsbigrid ] ) is uniformly bounded by the energy of their projection on the first half of the acoustic mode .
on the other hand , theorem [ theoremfouriertruncation ] guarantees that the observability inequality ( [ p2observabilityinequalitygeneral ] ) with @xmath103 as in ( [ observationoperators ] ) holds uniformly as @xmath117 in the class of truncated data lying on the first half of the acoustic mode .
combining these two facts , one can apply a dyadic decomposition argument as in @xcite and conclude the proof of theorem [ theorembigrid ] .
* 2 . the admissibility inequality . * in the rest of this section , our aim is to prove the direct inequality ( [ p2directinequalitygeneral ] ) for the boundary operator @xmath103 in ( [ observationoperators ] ) using the spectral identities ( [ observabilityeigenvectorsao ] ) and ( [ observabilityeigenvectorsr ] ) .
firstly , let us observe that for all matrix operator @xmath103 and all @xmath21 , we have the following inequality : @xmath402 for any matrix @xmath403 , the matrix norm @xmath404 involved in the right hand side of ( [ directineq1 ] ) is defined as @xmath405 in the above definition of the norm of a matrix , we can reduce the set of test functions to @xmath406 introduced in ( [ p2seteigenvectors ] ) , which is an orthonormal basis in @xmath407 . let us remark that for any eigenfunction @xmath406 , the corresponding eigenvalue @xmath408 verifies the identity @xmath409 consequently , for any matrix @xmath103 and @xmath406 , we have @xmath410 for @xmath103 in ( [ observationoperators ] ) , using the identities ( [ directineq2 ] ) , ( [ observabilityeigenvectorsao ] ) and ( [ observabilityeigenvectorsr ] ) , we conclude that @xmath411 is a quantity independent of @xmath263 .
in this section , we describe the algorithm of constructing the discrete controls of minimal @xmath33-norm and we prove their convergence towards the continuous hum boundary control @xmath412 in ( [ conthumcontrol ] ) as @xmath117 , under the hypothesis that both inverse and direct inequalities ( [ p2observabilityinequalitygeneral ] ) and ( [ p2directinequalitygeneral ] ) hold uniformly as @xmath117 .
as we saw in the previous sections , the above hypothesis holds when the initial data in the adjoint system ( [ p2adjoint ] ) is filtered through a fourier truncation or a bi - grid algorithm . using the _ admissibility inequality _ ( [ p2directinequalitygeneral ] ) and the _ observability one _ ( [ p2observabilityinequalitygeneral ] )
, one can prove the _ continuity _ and the _ uniform coercivity _ of @xmath129 defined by ( [ p2functionalgeneral ] ) on @xmath118 , where @xmath291 can be both the truncated space ( [ p2subsptruncation ] ) or the bi - grid one ( [ p2subsbigrid ] ) .
moreover , it is _ strictly convex_. applying the _ direct method for the calculus of variations _ ( dmcv ) ( cf .
@xcite ) , one can guarantee the existence of an unique minimizer @xmath413 of @xmath129 , i.e. : @xmath414 moreover , the euler - lagrange equation ( [ p2eulerlagrange ] ) associated to @xmath129 characterizes the optimal control ( [ p2optimalcontrol ] ) .
remark that when the space of initial data in ( [ p2adjoint ] ) is restricted to a subspace @xmath118 , for example the ones given by ( [ p2subsptruncation ] ) or ( [ p2subsbigrid ] ) , the exact controllability condition ( [ p2exactcontrol ] ) holds for all @xmath132 .
this does not imply that the final state @xmath415 in the controlled problem ( [ p2controlledpbm ] ) is exactly controllable to the rest , but its orthogonal projection @xmath416 from @xmath98 on the subspace @xmath291 , i.e. @xmath417 set @xmath135 to be the last component of the control @xmath418 in ( [ p2optimalcontrol ] ) ( the other ones vanish ) .
since @xmath419 and taking into account the inverse inequality ( [ p2observabilityinequalitygeneral ] ) , we obtain : @xmath420 where @xmath421 and @xmath116 is the observability constant in ( [ p2observabilityinequalitygeneral ] ) under filtering .
set @xmath422 , @xmath423 , @xmath424 and @xmath425 be the space of square summable sequences . since @xmath426 is a hilbertian basis in each @xmath427 , @xmath428 , the initial data @xmath429 to be controlled in the continuous problem ( [ contwavecontrolled ] ) admits the following fourier decomposition : @xmath430 moreover , their @xmath431-norm has the following fourier representation : @xmath432 since the set @xmath201 introduced in ( [ p2seteigenvectors ] ) is a basis in @xmath407 , the initial data @xmath128 to be controlled in the discrete problem ( [ p2controlledpbm ] )
admit the following decomposition @xmath433 with @xmath434 moreover , their @xmath435-norm can be written in terms of the fourier coefficients as follows : @xmath436 ) , we consider initial data with the following two properties : @xmath437 and @xmath438 then @xmath439 where @xmath440 is the last component of the discrete optimal control @xmath418 given by ( [ p2optimalcontrol ] ) obtained by the minimization of the functional @xmath129 on the subspace @xmath291 defined in ( [ p2subsptruncation ] ) or in ( [ p2subsbigrid ] ) and @xmath412 is the continuous hum control ( [ conthumcontrol ] ) . moreover , if the convergences in ( [ p2convdatacontracoustical ] ) and ( [ p2convdatacontropt ] ) are strong , then the convergence of controls in ( [ convcontr ] ) is also strong .
[ theoremweakconvergence ] firstly , let us observe that from ( [ p2convdatacontracoustical ] ) and ( [ p2convdatacontropt ] ) , we obtain that there exists a constant @xmath343 independent of @xmath263 such that @xmath441 by combining ( [ boundp2control ] ) and ( [ p2boundeddatacontrol ] ) , we get the uniform boundedness as @xmath117 of the discrete control @xmath440 in @xmath33 , so that @xmath442 it is sufficient to prove that the weak limit @xmath443 coincides with the continuous hum control @xmath444 given by ( [ conthumcontrol ] ) . the control @xmath444 can be characterized as the unique control @xmath32 in ( [ contwavecontrolled ] ) which can be expressed as the space derivative of a solution of the adjoint problem ( [ contwaveadjoint ] ) evaluated at @xmath17 .
then , we have to prove that @xmath443 is an admissible control of the continuous wave equation , i.e. it verifies the identity ( [ identitycontcontrol ] ) , and that @xmath445 , where @xmath446 is the solution of the adjoint problem ( [ contwaveadjoint ] ) for some initial data @xmath447 .
* step i - the weak limit @xmath443 is a control in the continuous problem ( [ contwavecontrolled ] ) . * since @xmath448 is an orthonormal basis in @xmath11 , then the fact that @xmath32 is a control in ( [ contwavecontrolled ] ) , so it satisfies ( [ identitycontcontrol ] ) , it is equivalent to prove ( [ identitycontcontrol ] ) for all initial data of the form @xmath449 , @xmath450 .
the solution of the adjoint problem ( [ contwaveadjoint ] ) with this kind of initial data is @xmath451 .
the condition ( [ identitycontcontrol ] ) is equivalent to the following one : @xmath452 let us check that @xmath443 satisfies ( [ identitycontrolfourier ] ) . to do it ,
we distinguish between the two cases of subspaces @xmath291 of filtered data .
firstly , let us consider the case of truncated initial data , i.e. @xmath291 is given by ( [ p2subsptruncation ] ) .
a particular class of initial data in @xmath291 is @xmath453 for which the solution of the discrete adjoint problem ( [ p2adjoint ] ) is @xmath454 , for all @xmath455 . from ( [ p2eulerlagrange ] )
, we see that the discrete control @xmath440 verifies the identity @xmath456 let us fix @xmath450 ( independent of @xmath263 ) . in that case , @xmath457 as @xmath117 strongly in @xmath33 , so that , taking into account the weak convergence ( [ p2weakconvcontrol ] ) , we can pass to the limit as @xmath117 in the left hand side of ( [ p2eulerlagrangefourier ] ) and we obtain the left hand side of ( [ identitycontrolfourier ] ) with @xmath32 substituted by @xmath443 . on the other hand , taking into account the condition ( [ p2convdatacontracoustical ] ) , which is valid for all test sequences in @xmath425 and in particular for the basis functions of @xmath425 , @xmath458 ( meaning that the weak convergence in @xmath425 is a pointwise convergence ) , and additionally the fact that @xmath459 and @xmath460 as @xmath117 for each fixed @xmath213 , passing to the limit as @xmath461 in the right hand side of ( [ p2eulerlagrange ] ) , we obtain the right hand side of ( [ identitycontrolfourier ] ) .
then , the weak limit @xmath443 of the optimal control @xmath440 obtained by minimizing the functional @xmath129 on @xmath291 in ( [ p2subsptruncation ] ) is a control for the continuous problem .
let us consider now the case of linear initial data given by a bi - grid algorithm , i.e. @xmath291 is given by ( [ p2subsbigrid ] ) . taking into account propositions [ proplineardata ] and [ proplowhighfreqacoustical ]
, we see that for initial data @xmath132 the fourier representation ( [ fourierrepresentationdata ] ) has the more particular form @xmath462 the basis function @xmath463 for the space @xmath291 is given by @xmath464 where @xmath465 and @xmath466 is defined by ( [ normh0 ] ) .
let us fix @xmath320 and consider the homogeneous problem ( [ p2adjoint ] ) with initial data @xmath467 , @xmath468 , for which the solution takes the form @xmath469 where the low frequency coefficients are @xmath470 and the high frequency ones are @xmath471 and @xmath472 by considering the particular class of solutions given by ( [ typicalsolutionbigrid ] ) into ( [ p2eulerlagrange ] ) , we see that the control @xmath135 satisfies the identity @xmath473 where the error terms are @xmath474 and @xmath475 passing to the limit as @xmath117 in the left hand side and in the first term in the right hand side of ( [ p2eulerlagrangefourierbigrid ] ) can be done as for the truncated space @xmath291 in ( [ p2subsptruncation ] ) .
therefore , in order to prove that the weak limit @xmath443 satisfies ( [ identitycontrolfourier ] ) , it is enough to show that the error terms are small as @xmath117 , i.e. @xmath476 from the fact that the @xmath33-norm of the discrete control @xmath135 is uniformly bounded as @xmath117 , the cauchy - schwartz inequality and the explicit expressions of @xmath477 , @xmath478 and @xmath479 , we obtain @xmath480 with @xmath481 @xmath482 and @xmath483 where @xmath484 and @xmath485 are as in ( [ basisbigrid ] ) . on the other hand ,
since the @xmath435 - norm of the initial data @xmath486 to be controlled is uniformly bounded as @xmath117 , we see that @xmath487 with @xmath488 @xmath489 and @xmath490 for a fixed @xmath450 , let us study the convergence as @xmath117 of the terms @xmath491 , @xmath492 , @xmath493 : * from @xmath494 ( @xmath466 introduced in ( [ normh0 ] ) ) , @xmath495 , @xmath496 , @xmath497 and @xmath498 , we see that @xmath499 as @xmath117 . * from @xmath500 , @xmath501 , @xmath502 and @xmath503 , we see that @xmath504 as @xmath117 .
* remark that @xmath505 , @xmath506 , @xmath507 , but @xmath508 .
nevertheless , @xmath509 , so that at the end @xmath510 as @xmath117 .
* remark that @xmath211 and @xmath511 .
observe that @xmath512 like @xmath513 .
this compensates the singularity of @xmath386 so that @xmath514 as @xmath117 . *
@xmath515 since it involves the factor @xmath388 .
this cancels the singularity introduced by @xmath516 and ensures that @xmath517 as @xmath117 .
* @xmath509 since it contains the factor @xmath518 which compensates the singularity of @xmath516 so that @xmath519 as @xmath117 .
this concludes ( [ converrorbigrid ] ) and the fact that the weak limit @xmath443 of the sequence of discrete hum controls obtained by minimizing the functional @xmath129 over the bi - grid class @xmath291 in ( [ p2subsbigrid ] ) is a control in the continuous problem ( [ contwavecontrolled ] ) .
* step ii - the weak limit @xmath443 is the normal derivative of a solution of the continuous adjoint problem ( [ contwaveadjoint ] ) . *
consider @xmath130 ( which in what follows can be both the subspace in ( [ p2subsptruncation ] ) or the one in ( [ p2subsbigrid ] ) ) to be the minimum of the functional @xmath129 .
due to the uniform nature of the observability inequality ( [ p2observabilityinequalitygeneral ] ) , @xmath520 is uniformly bounded , i.e. there exists a constant @xmath343 independent of @xmath263 such that @xmath521 due to property ( [ nullresonant ] ) , the resonant mode in the solution of the adjoint problem ( [ p2adjoint ] ) for initial data in the filtered space @xmath291 in ( [ p2subsptruncation ] ) or ( [ p2subsbigrid ] ) vanishes , so that the fourier representation of the total energy in the left hand side of ( [ unifboundminimizer ] ) is valid for both filtered spaces @xmath291 in ( [ p2subsptruncation ] ) and ( [ p2subsbigrid ] ) .
remark however that the high frequency components vanish for data in the truncation subspace @xmath291 in ( [ p2subsptruncation ] ) . on the other hand , for data in the bi - grid space @xmath291 in ( [ p2subsbigrid ] ) , the relations between the optic and the acoustic modes and the high frequencies in the acoustic mode and the lowest ones described in propositions [ proplineardata ] and [ proplowhighfreqacoustical ] hold . from ( [ unifboundminimizer ] )
, we deduce that @xmath522 set @xmath523 , observe that @xmath447 and denote by @xmath446 the corresponding solution of ( [ contwaveadjoint ] ) .
firstly , we prove that @xmath524 in fact , for arbitrary functions @xmath525 and @xmath526 , we will prove the following estimate : @xmath527 where @xmath528 is the projection on the acoustic branch defined by ( [ p2projacoudelta ] ) . in a similar way , we define the projection on the optic branch , @xmath529 . in order to prove ( [ estimateweakconv ] )
, we decompose its right hand side as follows : @xmath530 taking into account that @xmath531 , by integration by parts , we have @xmath532 from the cauchy - schwartz and the admissibility inequality , the bound ( [ unifboundminimizer ] ) and the fact that @xmath533 for all @xmath534 , we deduce that @xmath535 ) , we conclude ( [ p2weakconvdiscnormalderiv ] ) by using the following three ingredients : * the weak convergence ( [ p2convdatacontracoustical ] ) combined with the strong convergence @xmath536 , @xmath537 and @xmath538 allowing us to pass to the limit as @xmath117 in the sense of @xmath425 in the sum in the right hand side of @xmath539 so that we can guarantee that the first term in the right hand side of ( [ estimateweakconv ] ) is small as @xmath117 . * the density of @xmath540 in @xmath33 , allowing to choose @xmath541 so that @xmath542 is arbitrarily small . * an appropriate choice of the mesh size @xmath263 according to @xmath541 , so that @xmath543 is arbitrarily small .
let us check that @xmath445 in @xmath33 .
indeed , using ( [ p2weakconvcontrol ] ) and then ( [ p2weakconvdiscnormalderiv ] ) , we have : @xmath544 for an arbitrary @xmath525 .
therefore , @xmath545 in @xmath33 and also @xmath546 , which , jointly with the continuous observability inequality ( [ contobsineq ] ) gives that @xmath547 in @xmath31 and then @xmath548 in @xmath425 , @xmath204 .
this means that once we have identified that the weak limit of the discrete controls is the continuous hum control , we have the @xmath549-convergence of the discrete minimizer to the continuous one .
* step iii - strong convergence of the discrete controls . * in order to prove that @xmath440 converges strongly in @xmath33 to @xmath444 as @xmath117 , it is enough to prove that @xmath550 using as test function @xmath122 in ( [ p2eulerlagrange ] ) the minimizer @xmath551 , we have that @xmath552 where @xmath553 let us remark that using the strong convergence of the acoustic part of the initial data to be controlled ( [ p2convdatacontracoustical ] ) and the boundedness of the energy of the minimizer of @xmath129 , @xmath554 on the other hand , taking into account the uniform boundedness of the energy of the minimizer of @xmath129 , we get @xmath555 which concludes ( [ p2convl2normcontr ] ) and the strong convergence of the optimal control . in @xcite , it was proved that , for initial data @xmath556 in the continuous control problem ( [ contwavecontrolled ] ) belonging to the more regular space @xmath557 , the numerical controls obtained for the finite difference or the linear finite element semi - discretization of the wave equation ( [ contwavecontrolled ] ) converge to the continuous hum controls with an error order @xmath558 .
this is due to the fact that @xmath559 , where @xmath560 can be each one of the dispersion relations for the finite difference or finite element approximation of the wave equation . in @xcite , we observed the fact that the acoustic dispersion relation @xmath561 of the @xmath1 - finite element method approximates the continuous one @xmath562 with error order @xmath563 for all @xmath564 $ ] , so that the convergence error for the numerical controls obtained by the bi - grid algorithm in the quadratic approximation of the wave equation increases to @xmath565 under the same regularity assumptions on the continuous initial data to be controlled .
[ remark5 ]
* all the results in this paper can be extended to finite element methods of arbitrary order @xmath213 , with the additional difficulty that when computing the eigenvalues , the quadratic equation ( [ p2spectral6 ] ) has to be replaced by a @xmath213-th order algebraic equation in @xmath160 which is technically complicated to be solved explicitly .
the same difficulty arises when passing to several space dimensions . * the results in @xcite providing a general method to obtain uniform observability results for time discretizations of conservative
system lead to the extension of our observability results for the @xmath1 - space semi - discretization to fully discrete @xmath1 conservative approximations of the wave equation . *
the extension of the results in this paper to non - uniform meshes is a completely open problem . *
the last open problem we propose is related to @xcite and concerns the dispersive properties of the schrdinger equation approximated in space using the @xmath1 - finite element method . designing appropriate bi - grid algorithms taking care of all the singularities of both group velocity and acceleration simultaneously is an open problem .
99 ainsworth m. , _ discrete dispersion relation for hp - version finite element approximation at high wave number _
, siam j. numer .
anal . , 42(2004 ) , 553575 .
castro c. , micu s. , _ boundary controllability of a linear semi - discrete @xmath0 wave equation derived from a mixed finite element method _ ,
math . , 102(3)(2006 ) , 413462 .
dacorogna b. , _ direct methods in the calculus of variations _ , springer - verlag , 1989 .
ervedoza , s. , _ spectral conditions for admissibility and observability of wave systems _ , numer .
math . , 113(3)(2009 ) , 377415 .
ervedoza s. , _ admissibility and observability for schrdinger systems : applications to finite element approximation schemes _ , asymptot .
, to appear .
ervedoza s. , zheng c. , zuazua e. , _ on the observability of time - discrete conservative linear systems _ , j. functional analysis , 254(12)(2008 ) , 30373078 .
ervedoza s. , zuazua e. , _ the wave equation : control and numerics _ , in _ control and stabilization of pdes _
, p. m. cannarsa and j. m. coron eds . , lecture notes in mathematics , cime subseries , springer verlag , to appear .
glowinski r. , _ ensuring well posedness by analogy : stokes problem and boundary control for the wave equation _ , j. comput .
physics , 103(2)(1992 ) , 189221 .
glowinski r. , li c.h .
, lions j.l . , _ a numerical approach to the exact boundary controllability of the wave equation .
i. dirichlet controls : description of the numerical methods _ , japan j. appl . math . , 7(1)(1990 ) , 176 .
hughes t.r.j .
, reali a. , sangalli g. , _ duality and unified analysis of discrete approximations in structural dynamics and wave propagation : comparison of p - method finite elements with k - method nurbs _ , comput . meth .
, 197(49- 50)(2008 ) , 41044124 .
ignat l. , zuazua e. , _ numerical dispersive schemes for the nonlinear schrdinger equation _
, siam j. numer .
anal . , 47(2)(2009 ) , 13661390 .
ignat l. , zuazua e. , _ convergence of a two - grid algorithm for the control of the wave equation _
soc . , 11(2)(2009 ) , 351391 .
infante j .- a .
, zuazua e. , _ boundary observability for the space semidiscretization of the @xmath0 wave equation _ , m2an ,
33(1999 ) , 407438 .
lions j.l . , _ contrlabilit exacte , perturbations et stabilisation des systmes distribus _ , vol .
1 , masson , paris , 1988 .
loretti p. , komornik v. , _ fourier series in control theory _ , springer , 2004 .
loreti p. , mehrenberger m. , _ an ingham type proof for a two - grid observability theorem _
, eaim : cocv , 14(3)(2008 ) , 604631 .
marica a. , zuazua e. , _ localized solutions and filtering mechanisms for the discontinuous galerkin semi - discretizations of the @xmath0 wave equation _ , c. r. acad .
paris ser .
i , 348(2010 ) , 10871092 .
marica a. , zuazua e. , _ on the quadratic finite element approximation of @xmath0 waves : propagation , observation , control and numerical implementation _ ,
proceedings of cfl-80 : a celebration of 80 years of the discovery of cfl condition " , c. kubrusly and c. a. moura , eds . , springer proceedings in mathematics , springer verlag , to appear .
micu s. , _ uniform boundary controllability of a semidiscrete 1-d wave equation with vanishing viscosity _ , siam j. control optim . , 47(6)(2008 )
, 28572885 .
negreanu m. , zuazua e. , _ convergence of a multigrid method for the controllability of a @xmath0 wave equation _ , c. r. math .
paris , 338(2004 ) , 413418 .
tcheugou tbou l.r . ,
zuazua e. , _ uniform boundary stabilization of the finite difference space discretization of the @xmath0 wave equation _ , advances comput .
math . , 26(2007 ) , 337365 .
zuazua e. , _ propagation , observation , control and numerical approximations of waves _ ,
siam review , 47(2)(2005 ) , 197243 . | we study the propagation , observation and control properties of the @xmath0 wave equation on a bounded interval discretized in space using the quadratic @xmath1 - classical finite element approximation .
a careful fourier analysis of the discrete wave dynamics reveals two different branches in the spectrum : the _ acoustic _ one , of physical nature , and the _ optic _ one , related to the perturbations that this second - order finite element approximation introduces with respect to the @xmath2 one . on both modes there
are high frequencies with vanishing group velocity as the mesh size tends to zero .
this shows that the classical property of continuous waves of being observable from the boundary fails to be uniform for this discretization scheme . as a consequence of this , the controls of the discrete waves may blow - up as the mesh size tends to zero . to remedy these high frequency pathologies ,
we design filtering mechanisms based on a bi - grid algorithm for which one can recover the uniformity of the observability constant in a finite time and , consequently , the possibility to control with uniformly bounded @xmath3 - controls appropriate projections of the solutions .
this also allows showing that , by relaxing the control requirement , the controls are uniformly bounded and converge to the continuous ones as the mesh size tends to zero . |
we examine the spectroscopic data from spontaneous emission of the organic molecule dibenzoterrylene ( dbt ) in an anthracene crystal .
the molecular system is effectively described by two energy eigenstates that are delocalized across the dbt .
the molecule is coupled to a low excitation heat bath , and the joint system number state basis evolves unitarily under the action of : @xmath2 where @xmath3 .
this coupling can be viewed as a partial swap between qubit and environment , parametrized by @xmath4 .
it has the effect of transferring @xmath5 ( @xmath6 is the energy of the @xmath7 level ) between the molecule and the environment , with a probability dependent on the coupling strength . within this experiment
all of the photons are emitted along the zero phonon line , and as such we do not observe the effects of transitions within the vibrational modes . if we perform periodic measurements on the output field we can perform a partial measurement on the state of the molecule , and measure the heat flux from the molecule to the environment .
this produces trajectories of the joint system which are shown in fig .
[ fig : trajplot ] .
we convert the time resolved trace of the photon statistics into the heat distribution which we find to be consistent with : @xmath8 where @xmath9 are the initial occupation probabilities of each level .
the heat distribution is plotted in fig .
[ fig : dataplot ] , for two different laser intensities .
* heat distribution and the jarzynski equality * for an organic molecule undergoing spontaneous emission at a range of different temperatures .
the solid curve is the result of overlaying the theoretical prediction of @xmath10 with the experimentally calculated value .
this time evolution is interpreted using eq .
[ eq : heatdist ] .
the two results have such close agreement it is not possible to distinguish between them , demonstrating that the jarzynski equality holds for all times during the emission .
the lower dashed curve represents part of the heat distribution , which has been calculated directly from the spectroscopic data.,title="fig : " ] * heat distribution and the jarzynski equality * for an organic molecule undergoing spontaneous emission at a range of different temperatures .
the solid curve is the result of overlaying the theoretical prediction of @xmath10 with the experimentally calculated value .
this time evolution is interpreted using eq .
[ eq : heatdist ] .
the two results have such close agreement it is not possible to distinguish between them , demonstrating that the jarzynski equality holds for all times during the emission .
the lower dashed curve represents part of the heat distribution , which has been calculated directly from the spectroscopic data.,title="fig : " ] by examining the heat distribution given above we can directly evaluate the jarzynski equality .
in particular we find that the system satisfies @xmath11 for all time points during the period of spontaneous emission .
this demonstrates that the decay of organic molecules from an excited to the ground state is a thermodynamic process , even when it is considered at the level of single photon emission events .
we can utilize the jarzynski equality to detect shifts in the energy spectra of the molecule . if we consider taking the energy levels from : @xmath12 where @xmath13 is a common shift of the energies and @xmath14 is the difference in shifts between energy levels 1 and 2 .
this operation produces a change in the equilibrium free energy ( @xmath15 ) of the molecule : @xmath16 here we have defined @xmath5 and @xmath17 . by observing the fluorescence emitted by the excited molecules we can determine the value of both of these quantities . combining this relation with the jarzynski equality for heat we can determine the value of @xmath13 .
this would allow for probing the strengths of local field effects with high sensitivity .
recently there has been substantial interest in developing systems which can be used to test quantum thermodynamics , @xcite .
here we have demonstrated that the photon emission from organic molecules satisfies one of the fundamental relationships of the field .
this is true for the single photon emission from the molecules , at all points during the relaxation time .
we believe that this provides an important platform for the study of quantum thermodynamics in real systems .
organic spectroscopy is a well developed and established field of physics , and here we demonstrate that it possible to tap into these resources to investigate fundamental questions in quantum thermodynamics . in the initial set up of this investigation we focused on emissions along the zero phonon line .
one major advantage of investigating quantum thermodynamics by utilizing fluorescence from organic molecules is that it is experimentally straightforward to relax this assumption and probe how the situation changes when we incorporate the phononic transitions .
this would have the effect of creating an extra heat bath for energy transferred into the molecule to dissipate into , and would also provide for a much larger number of potential trajectories to be taken . with single molecule
spectroscopy it is possible to have sufficient resolution of the photons emitted to determine how much energy is being `` lost '' to the phonon modes
. we would then be able to determine if the jarzynski equality is still satisfied in this open system . in this analysis of the molecule - heat bath system we have modelled it as initially being in a mixed state .
we believe this initial condition to be reasonable as the lifetime of any entanglement between the molecule and environment is very short lived .
if instead the molecule were to be left in a superposition of energy eigenstates ( e.g. @xmath18 ) after the work is done on it then in the current setup there is no difference in the measured heat statistics relative to the corresponding de - phased mixed state .
the reason this occurs is due to the repeated suppression of the coherence in the heat bath via measurement in the photon number basis and discarding the measurement outcomes .
however as the initial state before the spontaneous emission is very close to a pure product state between the molecule and the heat bath , and the hamiltonian time evolution is unitary , the total state before measurements is approximately pure , and there will exist a basis in which measurements would not destroy coherence . as is well known , see e.g. @xcite , coherence implies extra extractable work so it would be valuable to identify how to implement this alternative basis measurement experimentally .
consider the two level molecule .
we assume it is initially in a thermal state at the same temperature as the environment . in this set - up and for the modes of concern
this means both the molecule and the environment are essentially in the ground state with a very small probability of a thermal excitation . the system is then pulsed by a laser which causes a swap in the occupation probabilities of the joint state . ( the laser is non - resonant but there is then near - instantaneous vibrational de - excitation to the excited level of interest - see fig .
2 . ) now the joint system undergoes evolution under its own hamiltonian which has the effect of inducing a partial swap , parametrized by @xmath19 , see eq .
[ eq : unitary ] .
this unitary arises naturally from the jaynes - cummings hamiltonian which describes the coupling between the environment and the molecule @xcite .
this corresponds to the emission of a photon to the environment as the molecule is de - excited . framing this in the language of quantum jumps , such that we can evaluate the jarzynski equality la @xcite , produces the trajectories in table [ tab : trajtable ] .
these trajectories are defined by considering an infinitesimal time step during which one of two events happens - either the system evolves under the joint hamiltonian or we perform a projective measurement on the environment .
the sequence of pure states produces a trajectory . .
[ tab : trajtable ] * possible trajectories * of the local state of the molecule . [ cols="^ " , ] + 1 & @xmath20 & @xmath21 & @xmath22 & 0 & @xmath23 + 2 & @xmath24 & @xmath21 & 0 & @xmath21 & @xmath25 + 3 & @xmath26 & @xmath22 & 0 & @xmath22 & @xmath27 + 4 & @xmath28 & @xmath22 & @xmath21 & 0 & @xmath29
+ in this framework we consider the system to always be occupying a definite state and determine what possible evolutions each state can take .
each of these evolutions constitutes a possible trajectory the system can undertake , which have associated values of internal energy change , work and heat .
having identified what the possible trajectories are , we must now determine which correspond to the experimental data we obtain . as we are detecting photon emission from the molecules using an avalanche photodiode ( apd ) , in principle we detect trajectories 1 , 3 and 4 .
however due to the electronics of the apd there is a delay time in receiving the signal from the photons @xcite .
trajectories 3 and 4 occur on a very fast time scale due to the stimulating effect of the laser causing the transition and so are unobserved . as such , we only observe clicks in the detector due to traj .
1 . the probabilities of the remaining trajectories turn out , as described below , to be determined uniquely by the initial state being assumed to be thermal together with demanding that the probabilities of all the trajectories sum to 1 .
the time resolution is very high and the probability of receiving more than one photon in a given time - bin is negligible here , even taking into account the number of molecules scattered in the illuminated part of the crystal .
thus one click means one photon . *
the swapping parameter * that controls the interaction with environment is plotted as a function of temperature and the time .
it can be seen that the swapping parameter is very stable for the majority of time during the experiment , indicating that the model we employ to interpret the data is sound , see eq .
[ eq : unitary ] .
we utilize this in connecting the experimental data to the quantum jumps model that enables us to evaluate the jarzynski equality . ]
having established the concurrence between clicks and photons , we now treat the environment as a low excitation heat bath .
the emitted photons are then simply the heat flux into the bath , and by relating the cumulative probability of emission to the heat distribution , eq .
[ eq : heatdist ] , we can determine the values of @xmath30 and @xmath19 , see fig . [
fig : swapplot ] .
this information corresponds to trajectories 1 and 2 , and by employing the fact that @xmath31 we can explicitly calculate the jarzynski equality .
this equality is plotted in fig .
[ fig : dataplot ] .
we conclude that this fluorescence experiment amounts to a scenario where jarzynski s equality holds . as further evidence that this experimental set - up is promising for probing quantum thermodynamics we note / reiterate the following : ( i ) molecular spectroscopy is a powerful and well - established technique , ( ii ) the set - up maps naturally to the quantum jumps model for work and heat ( used e.g. in crook s proof of quantum jarzynski @xcite ) . a natural and exciting next step in this program is to prepare two entangled molecules to implement the quantum maxwell s demon of @xcite .
this would build on the important first steps we have taken with this work , and reinforce the applicability of quantum thermodynamics to spectroscopy .
a further area of research would be exploring the technological applications of utilizing these organic molecules to probe external fields by employing the relationship between spectroscopy and the jarzynski equality .
we are grateful for comments from john goold and mark mitchison .
the authors thank the oxford martin school , wolfson college and the university of oxford , the leverhulme trust ( uk ) , the john templeton foundation , the eu collaborative project thermiq ( grant agreement 618074 ) , the cost action mp1209 and the epsrc ( uk ) .
this research is also supported by the national research foundation , prime minister s office , singapore , under its competitive research programme ( crp award no .
nrf- crp14 - 2014 - 02 ) and administered by centre for quantum technologies , national university of singapore .
sanghyun park , fatemeh khalili - araghi , emad tajkhorshid , and klaus schulten .
free energy calculation from steered molecular dynamics simulations using jarzynski s equality .
, 119(6):35593566 , august 2003 .
tiago batalho , alexandre m. souza , laura mazzola , ruben auccaise , ivan s. oliveira , john goold , gabriele de chiara , mauro paternostro , and roberto m. serra .
experimental reconstruction of work distribution and verification of fluctuation relations at the full quantum level . , 113(14 ) , october 2014 .
arxiv : 1308.3241 .
shuoming an , jing - ning zhang , mark um , dingshun lv , yao lu , junhua zhang , zhang - qi yin , h. t. quan , and kihwan kim . experimental test of the quantum jarzynski equality with a trapped - ion system .
, 11(2):193199 , february 2015 . | we demonstrate with an experiment how molecules are a natural test - bed for probing fundamental quantum thermodynamics .
single - molecule spectroscopy has undergone transformative change in the past decade with the advent of techniques permitting individual molecules to be distinguished and probed . by considering the time - resolved emission spectrum of organic molecules as arising from quantum jumps between states ,
we demonstrate that the quantum jarzynski equality is satisfied in this set - up .
this relates the heat dissipated into the environment to the free energy difference between the initial and final state .
we demonstrate also how utilizing the quantum jarzynski equality allows for the detection of energy shifts within a molecule , beyond the relative shift .
marrying the language of thermodynamics with quantum phenomena is giving rise to a quantum thermodynamics whose development defines a new frontier where the transfer of energy at the level of individual quantum objects can now be studied .
the success of thermodynamics owes much to being open to experimental testing in a variety of systems @xcite .
these have included systems operating in the quantum regime and exhibiting features like quantum coherence , such as quantum dots @xcite .
here we seek to demonstrate that the fluorescence of organic molecules forms a natural testbed for these theories .
our result provides an important proof of concept for the validity of quantum thermodynamics in this regime , and opens the door for more advanced tests of the theory . + non - equilibrium thermodynamics is evolving rapidly , and a key result is the jarzynski equality @xcite .
this concerns the probability distribution of work into a driven system in contact with a heat bath .
it essentially equates the average of the exponential work with something which is constant regardless of the driving rate . that constant is the exponential of the equilibrium free energy , i.e. the free energy between a thermal state at the initial boundary conditions and at the final respectively .
a key use of this equation is to determine the equilibrium free energy difference from non - equilibrium experiments @xcite .
the equation also holds for quantum systems @xcite .
one important area within this field , which has received a large amount of recent attention , is the development of experimental techniques and protocols which can test the theoretical predictions @xcite .
+ parallel developments in physical chemistry with the advent of single molecule spectroscopy @xcite over the past decade have opened up unprecedented opportunities for the study of single quantum systems . unlike studies in bulk where spectroscopic signatures are washed out by the averaging effect of ensembles , these novel techniques now allow individual molecules to be identified , tracked and probed . thus it is becoming possible to study energy transfers at the level of single molecules , some of which offer ideal test - beds owing to their well - defined spectroscopic signatures and quantum state dynamics .
their properties are extremely reproducible , more so than rival quantum systems such as quantum dots and nanocrystals @xcite .
schematic of a two - level molecule driven by laser pulses generating single photons from decaying electronic states with our thermodynamical representation of the process below it . ] in this work we connect these two approaches , namely quantum thermodynamics and molecular spectroscopy .
we consider the spontaneous emission of photons from a single excited organic molecule as a thermodynamic process .
we determine the probability distribution of work and heat in the experiment , see [ fig : engine ] , and find that it satisfies the quantum jarzynski equality .
a key point is to treat the free space around the molecule as an extremely low excitation heat bath , and the fluorescent light as heat transferred into that environment .
the laser applied initially corresponds to the driving force , as depicted in fig .
[ fig : engine ] . in the theoretical analysis
we use quantum jumps .
the core concept in the quantum jump formalism is that the system always occupies a distinct state that undergoes a stochastic evolution which enables it to transition between different energy levels @xcite as illustrated in figs .
[ fig : dbt ] and [ fig : trajplot ] .
furthermore , we demonstrate how use of the quantum jarzynski equality can be used to infer absolute energy - level shifts in molecules from the emitted photons , as opposed to just the relative shift .
we thus establish that these organic molecules are a natural and powerful test - bed for non - equilibrium thermodynamics and suggest how they have use in the enhanced detection of force fields . .
decay from @xmath0 into the ground state @xmath1 gave rise to a sharp lifetime limited line ( 30mhz ) , which , in a single molecule , gives rise to the emission of a single photon for each decay .
the narrow width of the zero phonon line ( zpl is lifetime - limited with a width of approx . 30mhz . * ( c ) * waveform of the decay of excited electronic states in dbt at a temperature of 4k obtained by time - resolved fluorescence spectroscopy .
the waveform represents a probability distribution of the arrival times of single photons from the decaying state rather than the shape of an optical signal .
the time - resolved technique relies on time - correlated single photon counting of pulses obtained by exciting the sample with a 730 nm pulsed laser with a repetition rate of 80mhz .
the probability of detecting more than one photon per excitation period is negligible .
the excited state which gives to the zpl emission line is found to have a characteristic lifetime of 5ns aftering fitting the exponential decay constant . ] |
true to `` stigler s law of eponymy '' ( stigler , 1980 ) , simpson s paradox has a long history in statistics going back to yule s ( 1903 ) ` spurious ' association , but it is currently credited to simpson ( 1951 ) for reframing it as a ` paradox ' ; see blyth ( 1972 ) . the paradox seems to have a number of alternative conceptions , and thus , it is often described interchangeably as a counter - intuitive statistical result pertaining to : \(a ) statistical associations that reverse themselves , such as `` a marginal association can have a different direction from each conditional association '' ( agresti , 2013 ) .
\(b ) either the magnitude or the direction of an association between two variables is influenced by a third variable , such as " the association between a pair of variables @xmath0 reverses sign upon conditioning on a third variable , @xmath1 .
( pearl , 2014 ) .
\(c ) apparent statistical associations that after closer scrutiny of the data are rendered ` spurious ' ( yule , 1903 ) .
the recent discussions in statistics have focused on adopting one of the perspectives ( a)-(c ) , and using actual or hypothetical data to either explain away the paradox or criticize other proposed ` solutions ' .
the current dominating view revolves around perspective ( b ) that differs from ( a ) in so far as it emphasizes the causal dimension of conditioning on a confounder ; see pearl ( 2009 ) , spirtes et al .
armistead ( 2014 ) put forward a dissenting view by arguing that perspective ( b ) is rather narrow to explain the different facets of the paradox : a strong case can be made that simpson s paradox has different dimensions that are often conflated or ignored in the literature .
as argued by wasserman ( 2004 ) : the primary aim of this paper is to shed light on the different conceptions of the paradox by bringing out the similarities and differences between perspectives ( a)-(c ) .
the key is provided by yule s idea of ` spuriousness ' in ( c ) .
beginning with yule ( 1903 ) , the problem of ` fictitious ' associations and ` spurious ' correlations was a recurring theme in yules papers that culminated in yule ( 1926 ) on `` nonsense - correlations '' .
although he shed some light on the issues involved , he did not succeed in establishing a direct link between spurious associations and invalid probabilistic assumptions for reasons to be discussed in the sequel .
the notion of statistical misspecification can be used to formalize the term ` spurious ' as ` statistically untrustworthy ' results , stemming from unreliable inference procedures .
this enables one to delineate between two distinct cases of association reversal : case 1 .
the reversal is statistically trustworthy due to statistical adequacy .
the reversal is statistically untrustworthy due to statistical misspecification .
it turns out that the statistical misspecification perspective suggests that in both cases there is nothing counterintuitive to explain . in section 2
, we discuss the case where the reversal is statistically trustworthy due to the fact that the statistical models involved are _ statistically adequate _ : the invoked probabilistic assumptions are valid for the particular data . when this is not the case , the inference results are likely to be statistically untrustworthy ( spurious ) .
this is discussed in section 3 using two empirical examples that bring out the distinction between statistical and substantive misspecification .
the statistical misspecification argument is illustrated further in section 4 using several widely discussed examples of the paradox . in section 5
, we revisit the causal ` resolution ' of the paradox in an attempt to delineate the modeling and inference issues raised by the statistical misspecification perspective .
consider the case of a linear regression ( lr ) model:@xmath2{c}y_{t}\mathit{=}\beta_{0}+\beta_{1}x_{1t}+\beta_{2}x_{2t}\mathit{+}u_{t},\medskip\\ ( u_{t}\mathit{\mid}x_{1t}\mathit{=}x_{1t},x_{2t}\mathit{=}x_{2t})\mathit{\backsim}\text{\textsf{niid}}(0,\sigma_{u}^{2}),\ t\mathit{\in } \mathbb{n } , \end{array } \vspace*{-0.07in}\label{rm1}\ ] ] where ` niid ' stands for ` normal , independent and identically distributed ' .
it is often insufficiently appreciated that the error assumptions imply a particular statistical parameterization for the unknown parameters @xmath3 in terms of the moments of the observable process @xmath4 underlying data @xmath5 ( see appendix ) .
alternatively , one can derive the parameterization directly using the joint distribution of the observable random variables involved:@xmath2{c}\left ( \begin{array } [ c]{c}y_{t}\\ x_{1t}\\ x_{2t}\end{array } \right ) \backsim\text { \textsf{niid}}\left ( \left ( \begin{array } [ c]{c}\mu_{1}\\ \mu_{2}\\ \mu_{3}\end{array } \right ) , \left ( \begin{array } [ c]{ccc}\sigma_{11 } & \sigma_{12 } & \sigma_{13}\\ \sigma_{12 } & \sigma_{22 } & \sigma_{23}\\ \sigma_{13 } & \sigma_{23 } & \sigma_{33}\end{array } \right ) \right ) \end{array } \vspace*{-0.1in}\label{j}\ ] ] in this case , the regression and skedastic functions take the form:@xmath2{c}e(y_{t}\mathit{\mid}x_{1t}\mathit{=}x_{1t},x_{2t}\mathit{=}x_{2t})\mathit{=}\beta_{0}\mathit{+}\beta_{1}x_{1t}\mathit{+}\beta_{2}x_{2t},\ var(y_{t}\mathit{\mid}x_{1t}\mathit{=}x_{1t},x_{2t}\mathit{=}x_{2t})\mathit{=}\sigma^{2 } , \end{array } \vspace*{-0.1in}\ ] ] where the parameterizations of @xmath6 are ( table 1):@xmath2{c}\beta_{0}\mathit{=}\mu_{1}\mathit{-}\beta_{1}\mu_{2}\mathit{-}\beta_{2}\mu _ { 3},\ \beta_{1}\mathit{=}\frac{(\sigma_{12}\sigma_{33}-\sigma_{13}\sigma _ { 23})}{(\sigma_{22}\sigma_{33}-\sigma_{23}^{2})},\ \beta_{2}\mathit{=}\frac{(\sigma_{13}\sigma_{22}-\sigma_{12}\sigma_{23})}{(\sigma_{22}\sigma _ { 33}-\sigma_{23}^{2})}\end{array } \vspace*{-0.1in}\label{b}\]]@xmath2{cl}\sigma_{u}^{2 } & \mathit{=}\sigma_{11}\mathit{-}\sigma_{12}\left ( \frac{\sigma_{12}\sigma_{33}-\sigma_{13}\sigma_{23}}{\sigma_{22}\sigma _ { 33}-\sigma_{23}^{2}}\right ) \mathit{-}\sigma_{13}\left ( \frac{\sigma _ { 13}\sigma_{22}-\sigma_{12}\sigma_{23}}{\sigma_{22}\sigma_{33}-\sigma _ { 23}^{2}}\right ) \mathit{=}\sigma_{11}\mathit{-}\sigma_{12}\beta _ { 1}\mathit{-}\sigma_{13}\beta_{2}\end{array } \label{s}\ ] ] these results offer the key to elucidating perspectives ( a)-(b ) on simpson s paradox . *
perspective ( a ) on simpson s paradox*. the correlation between @xmath7 and @xmath8@xmath9__@xmath10__@xmath11 is positive ( @xmath12 ) , but the coefficient @xmath13 in ( [ rm1 ] ) is negative ( @xmath14 ) . * is this reversal of association possible , and under what circumstances ? * in light of the parameterization of @xmath13 in ( [ b ] ) , its numerator is negative when:@xmath15@xmath2{c}\left [ ( \sigma_{12}\sigma_{33}\mathit{-}\sigma_{13}\sigma_{23})<0\right ] \rightarrow\left [ \frac{\sigma_{13}\sigma_{23}}{\sigma_{33}}>\sigma _ { 12}\right ] \end{array } \vspace*{-0.07in}\ ] ] multiplying both terms in the last expression by @xmath16 yields:@xmath15 @xmath2{c}\frac{\sigma_{13}\sigma_{23}}{\sigma_{33}\sqrt{\sigma_{11}\sigma_{22}}}\mathit{=}\rho_{13}\rho_{23}>\rho_{12}\mathit{=}\frac{\sigma_{12}}{\sqrt{\sigma_{11}\sigma_{22 } } } , \end{array } \vspace*{-0.05in}\ ] ] where @xmath17__@xmath10__@xmath18 and @xmath19__@xmath10__@xmath20 .
hence , @xmath12 and @xmath14 occur when the following conditions hold : \(i ) the correlation coefficients @xmath21 and @xmath22 have the _ same sign _ , \(ii ) the product of @xmath21 and @xmath22 is greater than @xmath23 , i.e. @xmath24 and \(iii ) the determinant of the correlation matrix of @xmath25 is positive:@xmath15@xmath2{c}corr(\mathbf{z}_{t})\mathit{=}1\mathit{-}\rho_{12}^{2}\mathit{-}\rho_{13}^{2}\mathit{-}\rho_{23}^{2}\mathit{+}2\rho_{12}\rho_{13}\rho_{23}\mathit{>}0 . \end{array } \vspace*{-0.07in}\ ] ] condition ( iii ) ensures that @xmath26 in ( [ j ] ) is proper , giving rise to a well - defined conditional distribution @xmath27 ; see spanos and mcguirk ( 2002 ) .
assuming @xmath28 without any loss of generality , let the relevant correlations be : @xmath29 which satisfy ( i)-(iii ) above .
\(a ) for values @xmath30 : @xmath31 \(b ) for values @xmath32 : @xmath33 note that the sign of @xmath34 reflects the common sign of @xmath35 in light of these results , it is clear that there is nothing paradoxical , or surprising , about the reversal of sign between the simple correlation @xmath12 [ stemming from the joint distribution @xmath36 , and the regression coefficient @xmath14 [ stemming from the conditional distribution @xmath37 .
this reversal is due to the conditions ( i)-(iii ) above , which are easily testable in practice ; see spanos ( 2006b ) .
it is well - known that there is a direct connection between @xmath38 and the regression coefficient of @xmath39 in the context of the simple linear regression:@xmath2{c}y_{t}\mathit{=}\alpha_{0}\mathit{+}\alpha_{1}x_{1t}+\varepsilon_{t},\medskip\\ ( \varepsilon_{t}\mathit{\mid}x_{1t}\mathit{=}x_{1t})\mathit{\backsim } \text{\textsf{niid}}(0,\sigma_{\varepsilon}^{2}),\ t\mathit{\in}\mathbb{n } , \end{array } \vspace*{-0.08in}\label{rm2}\ ] ] whose implicit statistical parameterization of @xmath40 is:@xmath2{c}\alpha_{0}\mathit{=}\mu_{1}\mathit{-}\alpha_{1}\mu_{2},\ \alpha_{1}\mathit{=}\frac{\sigma_{12}}{\sigma_{22}},\ \sigma_{\varepsilon}^{2}\mathit{=}\sigma_{11}\mathit{-}\frac{\sigma_{12}^{2}}{\sigma_{22}}. \end{array } \vspace*{-0.08in}\ ] ] this is because @xmath38 is a scaled @xmath41 reparameterization of @xmath42:@xmath2{c}\rho_{12}\mathit{=}\frac{\sqrt{\sigma_{22}}}{\sqrt{\sigma_{12}}}\alpha_{1}. \end{array } \vspace*{-0.12in}\label{cr}\ ] ] in the above numerical example , @xmath43 and @xmath44 confirming the sign reversal .
this implies that one can consider the question of association reversal by comparing the inference results in ( [ rm1 ] ) and ( [ rm2 ] ) . in conclusion
, it is very important to emphasize that in the above example , both lr models , ( [ rm1 ] ) and ( [ rm2 ] ) , are assumed to be statistically adequate : their probabilistic assumptions are valid . in the case of real data on @xmath45 , one needs to establish the statistical adequacy of both models using comprehensive misspecification testing .
what are the probabilistic assumptions that need to hold for data @xmath5 ?
in this section we bring out more explicitly the probabilistic assumptions comprising the linear regression ( lr ) model with a view to illustrate the role of statistical misspecification in shedding light on the various aspects of simpson s paradox .
traditionally , the probabilistic assumptions underlying the linear regression ( lr ) model are specified in terms of the error term ; see appendix .
it turns out , however , that such specifications are often incomplete and sometimes include non - testable assumptions .
table 1 specifies the lr , generically defined by : in terms of the statistical generating mechanism ( gm ) and assumptions [ 1]-[5 ] that constitute a complete , internally consistent and testable set of assumptions in terms of the observable process @xmath46 underlying the data @xmath47 .
this provides a purely probabilistic construal for the notion of a statistical model , viewed as a particular parameterization of the process @xmath46 .
intuitively , the statistical model comprises the totality of probabilistic assumptions one imposes on the process @xmath48 with a view to render data @xmath5 a ` typical ' realization thereof .
the ` typicality ' is testable using thorough misspecification testing ; see spanos ( 2006a).@xmath49@xmath50{l}\hline\begin{tabular } [ c]{l}\textbf{table 1 : linear regression model}\end{tabular } \\\hline\hline $ \overset{\qquad}{\begin{tabular } [ c]{ll}statistical gm : & $ y_{t}\mathit{=}\beta_{0}+\mathbf{\beta}_{1}^{\top } \mathbf{x}_{t}+u_{t},\ t\mathbf{\mathit{\in}}\mathbb{n}$. \end{tabular } } $ \\ $ \underset{\qquad}{\left .
\begin{tabular } [ c]{lll}\lbrack1 ] & normality : & $ \left ( y_{t}\mathbf{\mid
x}_{t}\mathit{=}\mathbf{x}_{t}\right ) \backsim\mathsf{n}(.,.),$\\ \lbrack2 ] & linearity : & $ e\left ( y_{t}\mathbf{\mid x}_{t}\mathit{=}\mathbf{x}_{t}\right ) \mathit{=}\beta_{0}+\mathbf{\beta}_{1}^{\top}\mathbf{x}_{t } , $ \\
\lbrack3 ] & homoskedasticity : & $ var\left ( y_{t}\mathbf{\mid x}_{t}\mathit{=}\mathbf{x}_{t}\right ) \mathit{=}\sigma^{2},$\\ \lbrack4 ] & independence : & $ \{\left ( y_{t}\mathbf{\mid x}_{t}\mathit{=}\mathbf{x}_{t}\right ) , \
t\mathbf{\mathit{\in}}\mathbb{n}\}$ indep .
process$,$\\ \lbrack5 ] & t - invariance : &
$ \left ( \beta_{0},\mathbf{\beta}_{1},\sigma ^{2}\right ) $ are \textit{not } changing with $ t,$\\ \multicolumn{3}{l}{$\ \ \ \ \ \
\beta_{0}\mathit{=}e(y_{t})\mathit{-}\mathbf{\beta}_{1}^{\top}e(\mathbf{x}_{t}),\ \mathbf{\beta}_{1}\mathit{=}[cov(\mathbf{x}_{t})]^{-1}cov(\mathbf{x}_{t},y_{t}),\ $ } \\ & \multicolumn{2}{l}{$\sigma^{2}\mathit{=}var(y_{t})\mathit{-}cov(\mathbf{x}_{t},y_{t})^{\top}[cov(\mathbf{x}_{t})]^{-1}cov(\mathbf{x}_{t},y_{t})$}\end{tabular } \right\ } \ t\mathbf{\mathit{\in}}\mathbb{n}.}$\\\hline \end{tabular}\ ] ] * statistical adequacy*. an estimated lr model is said to be _ statistically adequate _ when all assumptions [ 1]-[5 ] are valid for data @xmath51 in practice , statistical adequacy can be appraised using comprehensive misspecification testing ; see spanos ( 1999 , 2015 ) .
the importance of establishing statistical adequacy stems from the fact that it secures the statistical reliability of inference based on such a model .
that is , the inference propositions associated with the lr model , including the optimal properties of the mle estimators and the relevant error probabilities of the t and f tests , are reliable in the sense that their actual sampling distributions approximate closely the theoretical ones derived by invoking the validity of assumptions [ 1]-[5 ] .
* unreliability of inference*. when any subset of the assumptions [ 1]-[5 ] are invalid , the reliability of inference of such procedures is called into question .
statistical misspecifications are likely to give rise to inconsistent estimators as well as induce sizeable discrepancies between the nominal ( assumed ) error probabilities and the actual ones in testing . for instance , when any of the assumptions [ 2 ] , [ 4]-[5 ] are invalid , the ols estimators of @xmath52 are likely to be inconsistent , and the nominal error probabilities associated with the significance t - tests for the coefficients @xmath53 are likely to have significant discrepancies from the actual error probabilities ; see spanos and mcguirk ( 2001 ) , spanos ( 2010 ) .
applying a @xmath54 significance level t - test when the actual type i error is closer to @xmath55 is likely to give rise to unreliable inferences .
it is important to emphasize that for assumptions [ 4 ] and [ 5 ] to be testable , one needs to select an ordering of interest for data @xmath5 . in the case of time - series data ,
the ordering of interest is invariably ` time ' , which is an interval scale variable . for cross - section data , however ,
there are often several orderings of interest , depending on the individual unit being observed , and the modeler needs to think about such potential orderings as they relate to [ 4]-[5 ] .
potential orderings for cross - section can vary from gender ( nominal scale ) , to age ( ratio scale ) , etc .
let us return to example 1 , where the problem of association reversal can be viewed in the context of comparing the regression coefficients of @xmath56 @xmath42 and @xmath57 in the context of two linear regression models:@xmath58@xmath59{ll}model 1 : & $ y_{t}\mathit{=}\beta_{0}+\beta_{1}x_{1t}+\beta_{2}x_{2t}\mathit{+}u_{t},\ ( u_{t}\mathit{\mid}\mathbf{x}_{t}\mathit{=}\mathbf{x}_{t})\mathit{\backsim}\text{\textsf{niid}}(0,\sigma_{u}^{2}),\ t\mathit{\in } \mathbb{n},\medskip$\\ model 2 : & $ y_{t}\mathit{=}\alpha_{0}\mathit{+}\alpha_{1}x_{1t}+\varepsilon _ { t},\ ( \varepsilon_{t}\mathit{\mid}x_{1t}\mathit{=}x_{1t})\mathit{\backsim } \text{\textsf{niid}}(0,\sigma_{\varepsilon}^{2}),\ t\mathit{\in}\mathbb{n},$\end{tabular } \vspace*{-0.08in}\ ] ] where @xmath60 in the previous section , it was argued that when both models are statistically adequate , it could happen that the estimated coefficients @xmath42 and @xmath13 differ in both sign and magnitude . there is , however , a sizeable literature on ` omitted variables ' which would call model 2 misspecified when @xmath61 turns out to be statistically significant ; see greene ( 2011 ) . in what sense is model 2 misspecified if its assumptions [ 1]-[5 ] ( table 1 ) are valid ?
similarly , the literature on causal modeling would test the significance of the covariances @xmath62 and @xmath63 as they relate to the regression coefficients , to decide whether @xmath64 is a confounder ; see pearl ( 2011 ) .
how does this relate to the statistical misspecification perspective ? a closer look at the literature suggests that statistical misspecification is often conflated with substantive misspecification , using confusing and confused claims , such as the ols estimator of @xmath42 in model 2 is an inconsistent estimator of @xmath13 in model 1 ( greene , 2011 ) , ignoring the fact that the two coefficients represent very different parameterizations:@xmath65 to make any sense of such comparisons , one needs to distinguish between _ statistical _ and _ substantive adequacy _ because the former requires only that assumptions [ 1]-[5 ] are valid for @xmath5 .
assumptions [ 1]-[5 ] have nothing to do with : the lr model includes all ` substantively ' relevant variables .
the latter is a substantive assumption that pertains to the explanatory potential of the estimated model as it relates to the phenomenon of interest .
substantive inadequacy can arise from missing but relevant variables , false causal claims , etc .
the crucial importance of this distinction stems from the fact that when models 1 - 2 are statistically misspecified , both the test for an omitted variable , as well as the tests for deciding whether @xmath64 is a confounder , or a mediator , are likely to give rise to untrustworthy results ; see spanos ( 2006b ) .
this distinction is also important when the term ` spurious ' is employed without being qualified to differentiate between _ statistically _ and _ substantively spurious _ inference results .
indeed , the term ` spurious correlation ' is often used to describe the case where the statistical significance of a correlation coefficient is taken at face value , and an attempt is made to explain it away using substantive arguments ; see sober ( 2001 ) . more often than not , however
, one can show that the statistical significance is more apparent than real , because it is just an untrustworthy result stemming from a statistically misspecified model .
the problem of ` spurious ' associations , first noted by pearson ( 1896 ) , was high up in yule s agenda during the first quarter of the 20th century , returning to it on several occasions ; see yule ( 1909 , 1910 , 1921 ) .
yule ( 1926 ) is the culmination of his efforts to unravel the puzzle of ` spurious ' results using the high correlations between time series data as an example .
he used data measuring the ratio of church of england marriages to all marriages ( @xmath66 ) and the mortality rate ( @xmath67 ) over the period 1866 - 1911 , to demonstrate that their estimated correlation @xmath68 was both very high and statistically significant .
he described this result as ` nonsense - correlation ' because .
he went on to reject any attempt , however ingenious , to rationalize such a statistical result on substantive grounds : yule ( 1926 ) attempted to articulate the premise that ` nonsense - correlations ' have something to do with the fact that his time series data are _ not _ ` random series ' .
he could not establish a clear and direct link between ` spurious ' associations and statistical misspecification , however , because he was missing two key components that were yet to be integrated into statistics .
the first is the notion of a ` parametric statistical model ' , innovated by fisher ( 1922 ) , and the second is the theory of ` stochastic processes ' founded by kolmogorov ( 1933 ) .
the former comprises all the probabilistic assumptions imposed on the data , and the latter formalizes the notions of a ` random series ' into a realizationn of an iid stochastic proceses , as well as departures from it in the form of probabilistic concepts for dependence and heterogeneity .
* yule s reverse engineering*. given that there was no notion of a prespecified parametric statistical model , comprising the probabilistic assumptions imposed on the data , yule resorted to ` reverse engineering ' : he went on to consider the formula for estimating the sample standard error and elicit the implicit probabilistic assumptions that render it a ` good ' estimator of the distribution standard error .
let us emulate yule s reverse engineering using the sample correlation coefficient , which is the focus of his paper:@xmath58@xmath2{c}\widehat{corr(x_{t},y_{t})}\mathit{=}\frac{\begin{array } [ c]{c}\frac{1}{n}\sum\nolimits_{t=1}^{n}(y_{t}-\overline{y})(x_{t}-\overline{x } ) \end{array } \smallskip}{\sqrt{\left [ \frac{1}{n}\sum\nolimits_{t=1}^{n}(x_{t}-\overline{x})^{2}\right ] \left [ \frac{1}{n}\sum\nolimits_{t=1}^{n}(y_{t}-\overline{y})^{2}\right ] } } , \medskip\\\begin{array } [ c]{c}\overline{x}\mathit{=}\frac{1}{n}\sum\nolimits_{t=1}^{n}x_{t},\ \overline { y}\mathit{=}\frac{1}{n}\sum\nolimits_{t=1}^{n}y_{t},\ \widehat{var(x_{t})}\mathit{=}\frac{1}{n}\sum\nolimits_{t=1}^{n}(x_{t}-\overline{x})^{2},\medskip\\ \widehat{var(y_{t})}\mathit{=}\frac{1}{n}\sum\nolimits_{t=1}^{n}(y_{t}\mathit{-}\overline{y})^{2},\ \widehat{cov(x_{t},y_{t})}\mathit{=}\frac{1}{n}\sum\nolimits_{t=1}^{n}(y_{t}\mathit{-}\overline{y})(x_{t}\mathit{-}\overline{x } ) , \end{array } \end{array } \label{cor}\ ] ] as a ` good ' estimator of the distribution correlation coefficient:@xmath69@xmath2{c}corr(x_{t},y_{t})\mathit{=}\frac{cov(x_{t},y_{t})}{\sqrt{var(x_{t})var(y_{t})}}\end{array } \vspace*{-0.08in}\ ] ] the first assumption implicit in these formulae is the _ constancy _ of the moments:@xmath2{c}e(y_{t})\mathit{=}\mu_{1},\ e(x_{t})\mathit{=}\mu_{2},\ var(y_{t})\mathit{=}\sigma_{11},\ var(x_{t})\mathit{=}\sigma_{22},\ cov(x_{t},y_{t})\mathit{=}\sigma_{12},\ t\mathit{\in}\mathbb{n } , \end{array } \vspace*{-0.08in}\ ] ] which corresponds to a form of the _ i d assumption_. the formulae for @xmath70 and @xmath71 implicitly assume _ non - correlation _ over @xmath72 , otherwise they should have included covariances over @xmath72 terms .
yule also sought to unveil the implicit distributional assumption the sample moments are not always ` optimal ' estimators of the distribution moments .
for instance , the estimators in ( [ cor ] ) will be ` optimal ' under normality , but they will be non - optimal if the distribution is uniform ; see carlton ( 1946 ) . in light of the fact that under normality the assumption of i d reduces to the constancy of the first two moments , and non - correlation
coincides with _ independence _ , one could make a case that the implicit parametric statistical model underlying the above formulae is the * * simple bivariate normal * * in table 2.@xmath59{l}\hline\begin{tabular } [ c]{l}\textbf{table 2 - the simple } ( bivariate)\textbf{\ normal model}\end{tabular } \medskip\\\hline\hline $ \left .
\begin{tabular } [ c]{lll}\multicolumn{3}{l}{statistical gm:\qquad\qquad$\mathbf{z}_{t}=\mathbf{\mu } + \mathbf{u}_{t},$}\\ \lbrack1 ] & normal : & $ \mathbf{z}_{t}\backsim\mathsf{n}(.,.),$\\ \lbrack2 ] & constant mean : & $ e(\mathbf{z}_{t})\mathit{=}\mathbf{\mu},$\\ \lbrack3 ] & constant covariance : & $ var(\mathbf{z}_{t})\mathit{=}\mathbf{\sigma},$\\ \lbrack4 ] & independence : & $ \{\mathbf{z}_{t},\ t$\textbf{$\in$}$\mathbb{n}\}$ is independent.\\\hline \end{tabular } \right\ } \;t$\textbf{$\in$}$\mathbb{n},$\end{tabular}\]]@xmath2{c}\mathbf{z}_{t}\mathit{:=}\left ( \begin{array } [ c]{c}y_{t}\\ x_{t}\end{array } \right ) , \
\mathbf{\mu}\mathit{:=}\left ( \begin{array } [ c]{c}\mu_{1}\\ \mu_{2}\end{array } \right ) , \ \mathbf{\sigma}\mathit{:=}\left ( \begin{array } [ c]{cc}\sigma_{11 } & \sigma_{12}\\ \sigma_{12 } & \sigma_{22}\end{array } \right ) \end{array } \label{bn}\ ] ] when any of the assumptions [ 1]-[4 ] are invalid for the particular data @xmath5 , the estimated correlation coefficient is likely to be ` spurious ' ( statistically untrustworthy ) .
granted , certain departures from particular assumptions , such as [ 2]-[4 ] , are more serious than other departures , say from [ 1 ] .
a glance at the t - plots of yule s ( 1926 ) data suggests , to borrow his phrase on p. 5 , that : ( aka iid).@xmath73{2.7959in}{\begin{center } \includegraphics [ natheight=3.464400 in , natwidth=5.190600 in , height=1.7331 in , width=2.7959 in ] { .. / .. / .. /swp55/docs - toshiba / o6m4v000.wmf}\\\protect\begin{tabular } [ c]{l}fig .
1 : t - plot of $ x_{t}$-ratio of church of\protect\\ england marriages to all marriages \protect\end{tabular } \end{center } } } { \parbox[b]{2.6922in}{\begin{center } \includegraphics [ natheight=3.464400 in , natwidth=5.190600 in , height=1.7331 in , width=2.6922 in ] { .. / .. / .. /swp55/docs - toshiba / o6m4v001.wmf}\\\protect\begin{tabular } [ c]{l}fig .
2 : t - plot of $ y_{t}$-the mortality rate\protect\\ for the period 1866 - 1911 \protect\end{tabular } \end{center } } } \ ] ] both data series exhibit clear departures from iid ( fig .
6 ) in the form of mean @xmath74-heterogeneity ( trending mean ) and dependence ( irregular cycles ) . to bring out the cycles in the original data more clearly one needs to subtract the trending means using , say , a generic 3rd degree trend polynomial.@xmath73{2.6714in}{\begin{center } \includegraphics [ natheight=3.464400 in , natwidth=5.190600 in , height=1.7331 in , width=2.6714 in ] { .. / .. / .. /swp55/docs - toshiba / o6of6x00.wmf}\\ fig .
3 : t - plot of detrended $ x_t$ \end{center } } } { \parbox[b]{2.629in}{\begin{center } \includegraphics [ natheight=3.464400 in , natwidth=5.190600 in , height=1.7331 in , width=2.629 in ] { .. / .. / .. /swp55/docs - toshiba / o6of6x01.wmf}\\ fig .
4 : t - plot of detrended $ y_t$ \end{center } } } \ ] ] in light of the direct relationship between the correlation ( @xmath38 ) and the regression coefficient ( @xmath13 ) in ( [ cr ] ) , one can pose the question of statistical adequacy in the context of the linear regression model , which will yield:@xmath2{c}y_{t}\mathbf{\mathit{=}}\underset{(1.416)}{-10.847}\mathit{+}\underset{(.020)}{.419}x_{t}+\widehat{u}_{t},\;r^{2}\mathbf{\mathit{=}}.905,\;s\mathbf{\mathit{=}}.664,\;n\mathbf{\mathit{=}}46 , \end{array } \vspace*{-0.1in}\label{eq5}\ ] ] where the standard errors are reported in brackets below the coefficient estimates . both coefficients @xmath52
seem statistically significant since the t - ratios are:@xmath2{cc}\tau_{0}(\mathbf{z}_{0})\mathit{=}\frac{10.847}{1.416}\mathit{=}7.660[.000 ] , & \tau_{1}(\mathbf{z}_{0})\mathit{=}\frac{.419}{.020}\mathit{=}20.95[.000 ] , \end{array } \vspace*{-0.1in}\ ] ] and the p - values are given in square brackets .
note that the implied correlation ( see ( [ cor ] ) ) yields the value in yule ( 1926 ) : @xmath75 $ ] . a glance at the t - plot of the residuals ( fig .
5 ) , however , indicates that ( [ eq5 ] ) is statistically misspecified ; assumptions [ 4]-[5 ] are likely to be invalid .
the residual t - plot differs from that of a niid realization ( fig .
6 ) in so far as it exhibits distinct trends and cycles .
these misspecifications are confirmed formally by the statistical significance of the trends and lags in the auxiliary regression based on the residuals ( @xmath76 ) from ( [ eq5]):@xmath2{c}\widehat{u}_{t}\mathbf{\mathit{=}}\underset{(1.267)}{11.987}\mathit{-}\underset{(.016)}{.413}x_{t}\mathit{-}\underset{(.998)}{1.670}t\mathit{-}\underset{(.216)}{.406}t^{2}\mathit{+}\underset{(.076)}{.885}y_{t-1}\mathit{+}\underset{(.015)}{.006}x_{t-1}\end{array } \label{eq6}\]]@xmath73{2.7821in}{\begin{center } \includegraphics [ natheight=3.464400 in , natwidth=5.190600 in , height=1.7331 in , width=2.7821 in ] { .. / .. / .. /swp55/docs - toshiba / o6kgis02.wmf}\\ fig .
5 : t - plot of the residuals from ( \ref{eq5 } ) \end{center } } } { \parbox[b]{2.6221in}{\begin{center } \includegraphics [ natheight=3.464400 in , natwidth=5.190600 in , height=1.7331 in , width=2.6221 in ] { .. / .. / .. /swp55/docs - toshiba / o6kgis03.wmf}\\ fig . 6 : t - plot of niid data \end{center } } } \ ] ] these results suggest that the estimator of @xmath13 is inconsistent _ _ , _ _ and the t - test for its significance is statistically untrustworthy . taking mean deviations from @xmath77 when the actual means are trending ,
will render all the above estimators in ( [ cor ] ) inconsistent . in light of these departures from the iid assumptions , ( [ cor ] ) is an _
inconsistent _ estimator of the correlation coefficient , and thus statistically spurious .
indeed , one can easily show that when the data are de - trended and de - memorized ( subtract the temporal dependence using 2 lags ) to render ( [ cor ] ) an appropriate estimator , the estimated correlation is : @xmath78,$ ] which is totally statistically insignificant . in summary
, the notion of statistical adequacy provides a direct and testable link between statistical misspecification and statistically untrustworthy ( spurious ) associations , or inference results more generally .
a likely criticism of this link is that the probability assumptions of the assumed model in ( [ bn ] ) are too strong , in contrast to the current statistical practice favoring as weak a set of assumptions as possible .
the short reply to such a charge is that weaker but non - testable assumptions ( i ) do not render the assumed model less vulnerable to statistical misspecifications , and ( ii ) they underestimate the importance of securing statistical adequacy .
in addition , weak assumptions often rely on asymptotic sampling distributions without testing the validity of the assumptions invoked by limit theorems ; see spanos ( 2015 ) .
the truth of the matter is that the trustworthiness of all inference results will rely exclusively on the approximate validity of the probabilistic assumptions imposed on @xmath79 and nothing else . as argued by le cam ( 1986 , p. xiv ) : @xmath80@xmath80 in this sub - section we consider an empirical example based on cross - section data because statistical adequacy is less well appreciated in such a context .
consider the case where a practitioner wants to evaluate the effect of education on a person s income .
the data refer to education , @xmath66-years of schooling , and income , @xmath67-thousands of dollars , for @xmath81 working people within the age group of 30 - 40 years old selected from a city s population .
the estimated lr model yields:@xmath2{c}y_{t}\mathbf{\mathit{=}}\underset{(1.957)}{53.694}\mathit{-}\underset{(.147)}{.474}x_{t}+\widehat{u}_{t},\;r^{2}\mathbf{\mathit{=}}.096,\;s\mathbf{\mathit{=}}3.307,\;n\mathbf{\mathit{=}}100 . \end{array } \vspace*{-0.08in}\label{eq1}\ ] ] both coefficients @xmath52 appear to be statistically significant since the t - ratios are:@xmath2{cc}\tau_{0}(\mathbf{z}_{0})\mathit{=}\frac{53.694}{1.957}\mathit{=}27.437[.0000 ] , & \tau_{1}(\mathbf{z}_{0})\mathit{=}\frac{.474}{.147}\mathit{=}3.224[.001 ] .
\end{array } \vspace*{-0.08in}\ ] ] the practitioner is surprised by the negative sign of the coefficient of @xmath82 since that implies that additional years of education contribute negatively to one s income .
he takes a closer look at the data and decides to run separate linear regressions for men ( @xmath83 ) and women ( @xmath84 ) .
the estimated lr model for men yields : @xmath2{c}y_{1t}\mathbf{\mathit{=}}\underset{(2.236)}{45.229}\mathit{+}\underset{(.172)}{.409}x_{1t}+\widehat{u}_{1t},\;r^{2}\mathbf{\mathit{=}}.973,\;s\mathbf{\mathit{=}}2.371,\;n_{1}\mathbf{\mathit{=}}50 . \end{array } \vspace*{-0.1in}\label{eq2}\ ] ] the estimated lr model for women yields:@xmath2{c}y_{2t}\mathbf{\mathit{=}}\underset{(2.937)}{35.106}\mathit{+}\underset{(.199)}{.675}x_{2t}+\widehat{u}_{2t},\;r^{2}\mathbf{\mathit{=}}.193,\;s\mathbf{\mathit{=}}2.124,\;n_{2}\mathbf{\mathit{=}}50 . \end{array } \vspace*{-0.1in}\label{eq3}\ ] ] the estimation results in ( [ eq2])-([eq3 ] ) indicate that for both estimated regressions : the coefficients @xmath52 are statistically significant , and the sign of the coefficient @xmath57 of education variable @xmath85 , is positive . the positive sign of the estimated @xmath13 clearly contradicts the negative sign in ( [ eq1 ] ) , which is usually interpreted as a case where a statistical association is reversed .
this is considered as an example of simpson s paradox when viewed from perspective ( b ) , where gender ( @xmath86 ) is viewed as a confounding variable that correlates with both @xmath67-income and @xmath66-education . in econometrics ,
this is usually viewed as a case of ` omitted - variable bias ' ; see greene ( 2011 ) . according to pearl ( 2014 ) , p. 10
, the only way to decide whether to rely on the aggregated data regression in ( [ eq1 ] ) or the disaggregated data regressions ( [ eq2])-([eq3 ] ) is to use causal calculus .
@xmath73{3.039in}{\begin{center } \includegraphics [ natheight=3.464400 in , natwidth=5.190600 in , height=1.7331 in , width=3.039 in ] { .. / .. / .. /swp55/docs - toshiba / o6hx9702.wmf}\\ fig . 7 : residuals from equation ( \ref{eq1 } ) \end{center } } } \ ] ] upon reflection , however , the statistical misspecification perspective provides an alternative way to resolve the paradox on statistical adequacy grounds .
the above estimation and testing results in ( [ eq1])-([eq3 ] ) are trustworthy only when the model assumptions [ 1]-[5 ] are valid for the particular data for each of the three estimated equations . estimating the aggregated data equation ( [ eq1 ] ) using ` gender ' as the ordering of interest , and plotting the residuals ( fig .
7 ) suggests that ( [ eq1 ] ) is _ statistically misspecified _ because the t - plot is far from being normal white - noise .
assumption [ 5 ] is clearly invalid since its sample mean is not constant around zero , but shifts from positive for the first half to negative for the second , and the variance appears smaller for the second half ; see spanos ( 1999 ) , ch .
this form of t - heterogeneity differs from that in yule s data discussed above .
this is confirmed by the auxiliary regression using the residuals ( @xmath76 ) : @xmath2{c}\widehat{u}_{t}\mathbf{\mathit{=}}\underset{(2.00)}{-15.986}\mathit{+}\underset{(.134)}{.967}x_{t}+\underset{(.616)}{6.556}d_{t},\;r^{2}\mathbf{\mathit{=}}.54,\;s\mathbf{\mathit{=}}2.263,\;n\mathbf{\mathit{=}}100 , \end{array } \vspace*{-0.12in}\ ] ] where @xmath87 , 1-male , 0-female , since its coefficient is statistically significant : @xmath88.$ ] this suggests that the statistical misspecification perspective provides a very different interpretation of the reversion results and offers an alternative way to resolve the apparent paradox .
first , the key to resolving any seemingly conflicting inference results is not the notion of ` confounding ' ( pearl , 2014 ) , but that of statistical adequacy
. before one can talk about any form of reversal of a statistical association , one needs to establish that all the associations involved are statistically trustworthy .
any claim that there is a ` reversal of association ' between equation ( [ eq1 ] ) and ( [ eq2])-([eq3 ] ) is misleading since the aggregated data equation ( [ eq1 ] ) is statistically misspecified .
therefore , the inference that the coefficient of @xmath66 is negative and statistically significant is untrustworthy ; an artifact of imposing invalid probabilistic assumptions on data @xmath89 .
hence , the aggregate data misrepresent the relationship between @xmath67 and @xmath90 @xmath73{2.6221in}{\begin{center } \includegraphics [ natheight=3.464400 in , natwidth=5.190600 in , height=1.7331 in , width=2.6221 in ] { .. / .. / .. /swp55/docs - toshiba / o6plix02.wmf}\\ fig . 8 : t - plot of income ( $ y_t)$ \end{center } } } { \parbox[b]{2.6221in}{\begin{center } \includegraphics [ natheight=3.464400 in , natwidth=5.190600 in , height=1.7331 in , width=2.6221 in ] { .. / .. / .. /swp55/docs - toshiba / o6plix03.wmf}\\ fig .
9 : t - plot of education ( $ x_t)$ \end{center } } } \ ] ] second , the diagnosis that the variable ` gender ' , represented by @xmath86 is a missing ` confounder ' seems rather misleading for two reasons .
the information pertaining to the ordering(s ) of potential interest is already in the original data @xmath89 ( see figures 8 - 9 ) .
in addition , defining the confounder as an omitted variable @xmath1 which is related to the included variables @xmath91 and @xmath92 in the right way , requires that @xmath1 is stochastic variable , not a deterministic ordering . for statistical inference purposes , the inclusion of generic terms such as shifts in the mean , trends and lags in the estimated equation could , in certain cases , secure statistical adequacy , without having to resort to finding additional explanatory variables . in the case of ( [ eq1 ] ) , a more pertinent explanation is that the modeler neglected , or chose to ignore , the heterogeneity in the data by assuming constant mean and variance ( i d ) for both data series with respect to the ordering , _
gender_. such forms of misspecification pertain to statistical information contained in the data , which could be generically modeled using shift functions or / and trend polynomials in @xmath74 or / and lags , respectively ; see spanos ( 1999 ) . in the case of example 3 , one could attempt to respecify the original equation in ( [ eq1 ] ) by including the dummy variable ( @xmath86):@xmath2{c}y_{t}\mathbf{\mathit{=}}\underset{(2.00)}{37.639+}\underset{(.616)}{6.556}d_{t}+\underset{(.134)}{.501}x_{t}+\widehat{u}_{t},\;r^{2}\mathbf{\mathit{=}}.58,\;s\mathbf{\mathit{=}}2.272,\;n\mathbf{\mathit{=}}100 .
\end{array } \vspace*{-0.08in}\label{eq4}\ ] ] as it stands , the coefficient of @xmath66 represents an misleading weighted average of the two coefficients from the disaggregated data in ( [ eq2])-([eq3 ] ) . the residuals from this equation ( fig .
10 ) do not indicate any major departures from assumptions [ 1]-[5 ] , but in practice one needs to apply thorough misspecification testing to confirm or deny such a claim .
for instance , one needs to test that the variances of the residuals in the two sub - samples are equal ; see spanos ( 1986 ) , p. 481 - 3 . finally and most importantly , using the statistical misspecification perspective
, one can distinguish clearly between example 1 ( section 2 ) , and examples 2 and 3 above .
the key difference is that in example 1 both lr models ( [ rm1 ] ) and ( [ rm2 ] ) are statistically adequate .
in contrast , in example 3 the estimated lr model ( [ eq1 ] ) based on aggregated data , is statistically misspecified which renders the estimated coefficients and t - tests statistically untrustworthy .
hence , there was never a statistically trustworthy result at the aggregate level that gave rise to a reversal of associations .
in example 3 , only the disaggregated data give rise to statistically reliable inferences .
this calls seriously into question the conventional wisdom that these two cases as identical , as stated by samuels ( 1993 ) , p. 87 : @xmath73{3.039in}{\begin{center } \includegraphics [ natheight=3.464400 in , natwidth=5.190600 in , height=1.7331 in , width=3.039 in ] { .. / .. / .. /swp55/docs - toshiba / o6lwna00.wmf}\\ fig . 10 : residuals from equation ( \ref{eq4 } ) \end{center } } } \ ] ] in concluding this section , it is important to emphasize that the statistical misspecification perspective requires one to know the complete set of probabilistic assumptions imposed on the data , i.e. the statistical model .
more often than not , practitioners have an incomplete picture of the statistical model , they rarely test its assumptions , and thus the ensuing inference results are often untrustworthy . hence , in evaluating published empirical papers , it is sometimes useful to employ yule s reverse engineering to uncover the statistical model .
in this section we will revisit two cross - section data sets that have been widely discussed in the statistics literature , using the statistical misspecification perspective .
bickel et al . ( 1975 ) published an influential paper in _ science _
, where they illustrated simpson s paradox using cross - section data based on uc berkeley admissions , for the fall of 1973 .
their perspective relates to perspective ( a ) and pertains to the reversal of a statistical relationship between the aggregated data , at the university level , and the disaggregated data , at the department level .
the aggregate data are shown in table 3 and the data for the largest 5 departments , denoted by a - f , are given in table 4 ; see .
the estimated parameter @xmath93 based on the aggregate data ( table 3 ) indicates that the rate of admissions for female candidates ( @xmath94 ) is smaller that for male candidates ( @xmath95 ) , and a test for the difference indicated a statistically significant difference ; see bickel et al .
( 1975 ) . at the department level
, however , the admissions rate for females is greater than that of males in five out of the six departments shown in table 4 .
this is interpreted as an apparent reversal of the inference based on the aggregate data .
the statistical misspecification perspective , however , suggests that the estimated admissions rate using the aggregated data is statistically untrustworthy .
let us unpack this claim.@xmath59{c}\hline\begin{tabular } [ c]{l}\textbf{table 3 : } admissions aggregate data \end{tabular } \medskip\\\hline\hline\begin{tabular } [ c]{|l||ll||r|}\hline & \textbf{m}$\text{\textbf{ales}}$ & \textbf{f}$\text{\textbf{emales}}$ & $ \text{\textbf{total}}$\\\hline\hline admit & $ 3738 $ & $ 1494 $ & $ 5232$\\\hline deny & $ 4704 $ & $ 2827 $ & $ 7531$\\\hline\hline total & $ 8442 $ & $ 4321 $ & $ 12763$\\\hline\hline \multicolumn{4}{|l|}{$\begin{array } [ c]{c}\widehat{\theta}_{m}\mathit{=}\frac{3738}{8442}\mathit{=}.44,\ \widehat{\theta } _ { f}\mathit{=}\frac{1494}{4321}\mathit{=}.35 \end{array } $ } \\\hline \end{tabular } \end{tabular}\ ] ] @xmath59{ll}\hline \multicolumn{2}{c}{\begin{tabular } [ c]{l}\textbf{table 4 : admissions disaggregated data for departments a - f}\end{tabular } \medskip}\\\hline\hline $ \begin{tabular } [ c]{|l||ll||r|}\hline \fbox{\begin{tabular } [ c]{l}a \end{tabular } } & \textbf{m}$\text{\textbf{ales}}$ & \textbf{f}$\text{\textbf{emales}}$ & $ \text{\textbf{total}}$\\\hline\hline admit & $ 512 $ & $ 89 $ & $ 601$\\\hline deny & $ 313 $ & $ 19 $ & $ 332$\\\hline\hline total & $ 825 $ & $ 108 $ & 933\\\hline\hline \multicolumn{4}{|l|}{$\begin{array } [ c]{c}\widehat{\theta}_{am}\mathit{=}\frac{512}{825}\mathit{=}.62,\ \widehat{\theta } _ { af}\mathit{=}\frac{89}{108}\mathit{=}.82 \end{array } $ } \\\hline \end{tabular } $ & $ \begin{tabular } [ c]{|l||ll||r|}\hline \fbox{\begin{tabular } [ c]{l}b \end{tabular } } & \textbf{m}$\text{\textbf{ales}}$ & \textbf{f}$\text{\textbf{emales}}$ & $ \text{\textbf{total}}$\\\hline\hline admit & $ 353 $ & $ 17 $ & $ 370$\\\hline deny & $ 207 $ & $ 8 $ & $ 215$\\\hline\hline total & $ 560 $ & $ 25 $ & $ 585$\\\hline\hline \multicolumn{4}{|l|}{$\begin{array } [ c]{c}\widehat{\theta}_{bm}\mathit{=}\frac{353}{560}\mathit{=}.63,\ \widehat{\theta } _ { bf}\mathit{=}\frac{17}{25}\mathit{=}.68 \end{array } $ } \\\hline \end{tabular } $ \\ \multicolumn{2}{l}{}\\
$ \begin{tabular } [ c]{|l||ll||r|}\hline \fbox{\begin{tabular } [ c]{l}c \end{tabular } } & \textbf{m}$\text{\textbf{ales}}$ & \textbf{f}$\text{\textbf{emales}}$ & $ \text{\textbf{total}}$\\\hline\hline admit & $ 120 $ & $ 202 $ & $ 322$\\\hline deny & $ 205 $ & $ 391 $ & $ 596$\\\hline\hline total & $ 325 $ & $ 593 $ & $ 918$\\\hline\hline \multicolumn{4}{|l|}{$\begin{array } [ c]{c}\widehat{\theta}_{cm}\mathit{=}\frac{120}{325}\mathit{=}.37,\ \widehat{\theta } _ { cf}\mathit{=}\frac{202}{593}\mathit{=}.34 \end{array } $ } \\\hline \end{tabular } $ & $ \begin{tabular } [ c]{|l||ll||r|}\hline
\fbox{\begin{tabular } [ c]{l}d \end{tabular } } & \textbf{m}$\text{\textbf{ales}}$ & \textbf{f}$\text{\textbf{emales}}$ & $ \text{\textbf{total}}$\\\hline\hline admit & $ 139 $ & $ 131 $ & $ 270$\\\hline deny & $ 278 $ & $ 244 $ & $ 522$\\\hline\hline total & $ 417 $ & $ 375 $ & $ 792$\\\hline\hline \multicolumn{4}{|l|}{$\begin{array } [ c]{c}\widehat{\theta}_{dm}\mathit{=}\frac{139}{417}\mathit{=}.33,\
\widehat{\theta } _ { df}\mathit{=}\frac{131}{375}\mathit{=}.35 \end{array } $ } \\\hline \end{tabular } $ \\ \multicolumn{2}{l}{}\\\cline{2 - 2}$\begin{tabular } [ c]{|l||ll||r|}\hline \fbox{\begin{tabular } [ c]{l}e \end{tabular } } & \textbf{m}$\text{\textbf{ales}}$ & \textbf{f}$\text{\textbf{emales}}$ & $ \text{\textbf{total}}$\\\hline\hline admit & $ 53 $ & $ 94 $ & $ 147$\\\hline deny & $ 138 $ & $ 199 $ & $ 337$\\\hline\hline total & $ 191 $ & $ 293 $ & $ 484$\\\hline\hline \multicolumn{4}{|l|}{$\begin{array } [ c]{c}\widehat{\theta}_{em}\mathit{=}\frac{53}{191}\mathit{=}.28,\ \widehat{\theta } _ { ef}\mathit{=}\frac{94}{293}\mathit{=}.32 \end{array } $ } \\\hline \end{tabular } $ & \multicolumn{1}{l|}{$\begin{tabular } [ c]{|l||ll||r|}\hline \fbox{\begin{tabular } [ c]{l}f \end{tabular } } & \textbf{m}$\text{\textbf{ales}}$ & \textbf{f}$\text{\textbf{emales}}$ & $ \text{\textbf{total}}$\\\hline\hline admit & $ 22 $ & $ 23 $ & $ 45$\\\hline deny & $ 351 $ & $ 318 $ & $ 669$\\\hline\hline total & $ 373 $ & $ 341 $ & $ 714$\\\hline\hline \multicolumn{4}{|l|}{$\begin{array } [ c]{c}\widehat{\theta}_{fm}\mathit{=}\frac{22}{373}\mathit{=}.06,\
\widehat{\theta } _ { ff}\mathit{=}\frac{23}{341}\mathit{=}.07 \end{array } $ } \\\hline \end{tabular } $ } \end{tabular}\ ] ] what is missing from the discussion of the traditional association reversal interpretation is any evidence that the above inferences based on the estimated @xmath96 is trustworthy
. such evidence can be secured by testing the validity of the assumptions invoked by the above inferences , which comprise the underlying statistical model : a _ bivariate _ version of the simple bernoulli model ( table 5 ) , with @xmath96 replaced with a vector of unknown parameters @xmath97 ; see bishop et al .
( 1975 ) . in relation to the bernoulli model , it is important to point out that @xmath96 is also the mean of process underlying the data , as well as determining the variance , i.e.@xmath15@xmath2{c}e(x_{t})\mathit{=}\theta,\ var(x_{t})\mathit{=}\theta(1\mathit{-}\theta),\ 0\leq\theta\leq1,\ \forall t\mathbf{\mathit{\in}}\mathbb{n}. \end{array } \vspace*{-0.07in}\ ] ] @xmath59{l}\hline\begin{tabular } [ c]{l}\textbf{table 5 - the simple bernoulli model}\end{tabular } \medskip\\\hline\hline $ \left .
\begin{tabular } [ c]{lll}\multicolumn{3}{l}{statistical gm:\qquad\qquad$x_{t}=\theta+u_{t},\;$}\\ \lbrack1 ] & bernoulli : & $ x_{t}\backsim\mathsf{ber}(.,.),$\\ \lbrack2 ] & constant mean : & $ e(x_{t})\mathit{=}\theta,$\\ \lbrack3 ] & constant variance : & $ var(x_{t})\mathit{=}\theta(1-\theta),$\\ \lbrack4 ] & independence : & $ \{x_{t},\
t$\textbf{$\mathit{\in}$}$\mathbb{n}\}$ is independent .
\end{tabular } \right\ } t$\textbf{$\mathit{\in}$}$\mathbb{n}$\\\hline \end{tabular}\ ] ] a glance at the estimated @xmath96 for males and females at the department level in table 4 , indicate clearly that the estimated means and variances differ , not only from those based on aggregate data , but also between departments .
this renders assumptions [ 2 ] and [ 3 ] invalid at the aggregate level .
that is , when the data are aggregated the process @xmath98 is no longer identically distributed ( i d ) with respect to the ordering ` gender ' . in light of this , the association reversal is spurious because the estimated values:@xmath15@xmath2{c}\widehat{\theta}_{m}\mathit{=}\frac{3738}{8442}\mathit{=}.44,\ \widehat{\theta } _
{ f}\mathit{=}\frac{1494}{4321}\mathit{=}.35 \end{array } \vspace*{-0.07in}\ ] ] from the aggregated data .
this is because the estimators of @xmath96 using the aggregated data will be _ inconsistent _ estimators of the true @xmath96.@xmath99{2.6645in}{\begin{center } \includegraphics [ natheight=3.464400 in , natwidth=5.190600 in , height=1.7331 in , width=2.6645 in ] { .. / .. / ..
/swp55/docs - toshiba / o6lwrg03.wmf}\\\protect\begin{tabular } [ c]{l}fig . 11 : t - plot of $ x_{t}\backsim$\textsf{beriid}$(.2,.16)$\protect\end{tabular } \end{center } } } { \parbox[b]{2.7475in}{\begin{center } \includegraphics [ natheight=3.464400 in , natwidth=5.190600 in , height=1.7331 in , width=2.7475 in ] { .. / .. / .. /swp55/docs - toshiba / o6lwrg04.wmf}\\\protect\begin{tabular } [ c]{l}fig . 12 : t - plot of $ z_{t}\backsim$\textsf{beriid}$(.6,.24)$\protect\end{tabular } \end{center } } } \ ] ] to see how this arises in practice , consider figures 11 - 12 that represent the t - plots of two bernoulli iid [ ( @xmath100 $ ] processes with @xmath101 and @xmath102 respectively .
as can be seen from these figures , the concentration of longer ` runs ' [ group of successive values of 0 s or 1 s ] switches from the value 0 to the value 1 as @xmath96 increases above @xmath103 .
hence , any attempt to ignore the differences in the two moments of such processes will give rise to a misspecified bernoulli model .
that invalidates any inferences based on the aggregate data , and the only potentially reliable inference can be drawn from the disaggregated data .
the above statistical misspecification perspective can be used to explain the seemingly contradictory results in lindley and novick s ( 1981 ) hypothetical data shown below .
this is a particularly interesting example because , as argued by armistead ( 2014 ) , the ordering of interest might become apparent after the data are collected .
for instance , in a clinical trial the ` gender ' or / and ` age ' ordering(s ) might turn out to be relevant after the data are collected .
the estimated @xmath96 s for the aggregated data in table 6:@xmath104@xmath2{c}\widehat{\theta}_{w}\mathit{=}\frac{20}{40}\mathit{=}.5,\ \widehat{\theta}_{b}\mathit{=}\frac{16}{40}\mathit{=}.4 , \end{array } \vspace*{-0.07in}\ ] ] are very different from those based on the disaggregated data ( table 7 ) , rendering the former statistically untrustworthy because it imposes an invalid assumption : the means of the bernoulli process underlying the disaggregated data are constant.@xmath59{l}\hline\begin{tabular } [ c]{l}\textbf{table 6 : lindley - novick}\\ \textbf{\ \ \ \ aggregated data}\end{tabular } \medskip\\\hline\hline\begin{tabular } [ c]{|l||ll||r|}\hline & \textbf{white } & \textbf{black } & $ \text{\textbf{total}}$\\\hline\hline high & $ 20 $ & $ 16 $ & $ 36$\\\hline low & $ 20 $ & $ 24 $ & $ 44$\\\hline\hline total & $ 40 $ & $ 40 $ & $ 80$\\\hline \multicolumn{4}{|l|}{$\begin{array } [ c]{c}\widehat{\theta}_{w}\mathit{=}\frac{20}{40}\mathit{=}.5,\ \widehat{\theta}_{b}\mathit{=}\frac{16}{40}\mathit{=}.4 \end{array } $ } \\\hline \end{tabular } \end{tabular}\]]@xmath59{ll}\hline \multicolumn{2}{c}{\begin{tabular } [ c]{l}\textbf{table 7 : lindley - novick disaggregated data}\end{tabular } } \\\hline\hline\begin{tabular } [ c]{|l||ll||r|}\hline short & \textbf{white } & \textbf{black } & $ \text{\textbf{total}}$\\\hline\hline high & \multicolumn{1}{||c}{$2 $ } & \multicolumn{1}{c||}{$9 $ } & \multicolumn{1}{||c|}{$11$}\\\hline low & \multicolumn{1}{||c}{$8 $ } & \multicolumn{1}{c||}{$21 $ } & \multicolumn{1}{||c|}{$29$}\\\hline\hline total & \multicolumn{1}{||c}{$10 $ } & \multicolumn{1}{c||}{$30 $ } & \multicolumn{1}{||c|}{$40$}\\\hline\hline \multicolumn{4}{|l|}{$\begin{array } [ c]{c}\widehat{\theta}_{sw}\mathit{=}\frac{2}{10}\mathit{=}.2,\ \widehat{\theta } _ { sb}\mathit{=}\frac{9}{30}\mathit{=}.3 \end{array } $ } \\\hline \end{tabular } & \begin{tabular } [ c]{|l||ll||r|}\hline tall & \textbf{white } & \textbf{black } & $ \text{\textbf{total}}$\\\hline\hline high & \multicolumn{1}{||c}{$18 $ } & \multicolumn{1}{c||}{$7 $ } & \multicolumn{1}{||c|}{$25$}\\\hline low & \multicolumn{1}{||c}{$12 $ } & \multicolumn{1}{c||}{$3 $ } & \multicolumn{1}{||c|}{$15$}\\\hline\hline total & \multicolumn{1}{||c}{$30 $ } & \multicolumn{1}{c||}{$10 $ } & \multicolumn{1}{||c|}{$40$}\\\hline\hline \multicolumn{4}{|l|}{$\begin{array } [ c]{c}\widehat{\theta}_{tw}\mathit{=}\frac{18}{30}\mathit{=}.6,\ \widehat{\theta } _ { tb}\mathit{=}\frac{7}{10}\mathit{=}.7 \end{array } $ } \\\hline \end{tabular } \end{tabular}\ ] ]
pearl s ( 2014 ) claims that the only way to resolve the paradox is to use causal calculus : @xmath1@xmath1 viewing examples 2 - 5 from the misspecification perspective , however , lends support to the armistead s ( 2014 ) key argument : indeed , in cases where the ` third variable ' represents an _ ordering _ of potential interest for the particular data , the only relevant criterion to decide which orderings are relevant for the statistical analysis of the particular data is the _ statistical adequacy _ of the estimated equations .
that is , when two or more alternative orderings are potentially relevant for a particular data set , one needs to test the statistical adequacy of all three equations relative to each of these orderings before one could draw reliable conclusions concerning how to resolve any apparently paradoxical results .
where does this leave the pearl ( 2014 ) claim quoted above ? cartwright ( 1979 ) rightly points out that reliance on regularities and frequencies for statistical inference purposes is not sufficient for representing substantively meaningful causal relations . on the other hand ,
imposing causal relations that belie the chance regularities in the data would only give rise to untrustworthy inference results . while the causal dimension remains an important component in delineating the issues raised by simpson s paradox , it is not the only relevant , or even the most important , dimension in unraveling the puzzle .
indeed , the suggestion that in cases where the third variable ( ordering of interest ) is noncausal one should accept the results based on the aggregated data ( pearl , 2009 ) , is called into question by examples 2 - 5 .
this is because when the model estimated using the aggregated data is statistically misspecified , the causal inference results pertaining to conditional independence are likely to be untrustworthy .
one way or another , the modeler needs to account for the statistical information not accounted for by the original statistical model , with a view to ensure the trustworthiness of the ensuring statistical results .
it is interesting to note that yule ( 1926 ) considered the third variable causal explanation , but questioned its value as a general ` solution ' to the problem : a crucial issue that needs to be addressed by the causal explanation is that conditioning on a third variable is not as straightforward as adherents to this ` explanation ' of simpson s paradox would have us believe . in practice
, the question whether a particular variable @xmath105 constitutes a confounder is not just a matter of testing whether @xmath105 relates to @xmath106 and @xmath107 the right way ; see pearl ( 2009 ) , spirtes et al .
before such testing can even begin , one needs to test for the statistical adequacy of the estimated model with respect to a relevant ordering .
although ` time ' is the obvious ordering for time series , it is no different than other deterministic orderings for cross - section data such as ` gender ' , marital status , age , geographical position , etc . ; only the scale of measurement differs .
when the original model is statistically misspecified , it needs to be respecified with a view to secure statistical adequacy .
often one can restore statistical adequacy using generic terms relating to that ordering . to secure substantive adequacy
, however , one needs to replace such generic terms with proper explanatory random variables without foregoing the statistical adequacy .
the latter ensures the reliability of testing whether @xmath105 is a confounder or not ; see spanos ( 2006b ) .
yule ( 1926 ) considered ` time ' as a third variable and expressed his misgivings : viewing his comment from the vantage point of today s probabilistic perspective , the proposal to ` condition ' on a third variable raises technical issues , since the conditional distribution , defined by:@xmath15@xmath2{c}f(y_{t}\mathit{\mid}x_{t},d_{t};\mathbf{\varphi})\mathit{=}\frac { f(y_{t}\mathit{,}x_{t},d_{t};\mathbf{\psi})}{f(d_{t};\mathbf{\phi})},\ \forall y_{t}\mathit{\in}\mathbb{r}_{y } , \end{array } \vspace*{-0.07in}\ ] ] makes no probabilistic sense when @xmath108 is a _ determinist ordering _ ( variable ) such as time ; see williams ( 1991 ) .
this issue arises more clearly in cases where the ordering was deemed potentially important after the data have been collected , such as having plants grow short or tall , blood pressure being high or low , black or white plants , etc . ; see armistead ( 2014 ) .
how does one bridge the gap between a deterministic ordering of interest and conditioning on a third random variable related to that ordering ? * separating modeling from inference*. the statistical misspecification perspective suggests that to ensure the reliability of inference one needs to separate the initial stages of _ specification _ ( initial model selection ) _ misspecification testing _ and _ respecification _ , from _ inference _ proper .
the latter includes testing for _ substantive adequacy _ , such as attributing causality to statistical associations . in practice
, this requires focusing first on the ordering(s ) of interest that could potentially reveal statistical misspecifications that pertain to dependence and heterogeneity uncovered by misspecification testing .
the next step is to respecify the initial model with a view to account for the statistical information revealed by the misspecification testing .
this is usually achieved by employing _
generic _ terms , such as shifts , trends and lags , to ` capture ' such forms of systematic statistical information .
once statistical adequacy is secured one can then proceed to ` model ' such information by replacing the generic terms with appropriate explanatory variables with a view to improve the _ substantive adequacy _ without forgoing the statistical adequacy .
this is because a third degree trend polynomial might capture the mean heterogeneity in the data to ensure the statistical reliability of inference , but from the substantive perspective it represents ignorance .
replacing the trend polynomial with explanatory variables without forsaking statistical adequacy will add to our understanding of the phenomenon of interest ; see spanos ( 2010 ) .
viewing the problem from a broader perspective , the primary reason for the untrustworthiness is that the question of probing for the nature of any causal connections pertains to _ substantive _ , and not _ statistical adequacy _
, even though the distinction between the two might not always be clear cut or obvious ; see spanos ( 2010 ) .
this distinction is crucial because any attempt to probe for substantive adequacy , including causal connections , before securing statistical adequacy of the assumed statistical model is likely to give rise to unreliable results . to avoid this problem of unreliable inferences
, one needs to establish the statistical adequacy of the original model first before probing for any form of substantive adequacy , such as attributing a causal interpretation to statistical associations .
these include probing for the appropriateness of a particular confounder or choosing between different potential confounders ; see spanos ( 2006b ) for an extensive discussion .
this distinction is crucial in differentiating between _ statistically _ and _ substantively _ ` spurious ' inferential results .
unfortunately , in the statistics and philosophy of science literatures the term ` spurious ' is often used to describe the latter ; see blyth ( 1972 ) .
what is often insufficiently appreciated is that one needs to establish first that there is a statistically trustworthy statistical association , before attempting to explain it away as substantively spurious .
the statistical misspecification perspective also calls into question certain philosophical discussions of simpson s paradox that focus primarily on the ` numbers ' associated with the relevant probabilities / associations as in the case of example 1 . a typical representation of simpson s paradox in terms of events @xmath109 is:@xmath15@xmath2{c}\begin{tabular } [ c]{l}$p(a\mathit{\mid}b)\mathit{<}p(a\mathit{\mid}\lnot b),\text { but}\medskip$\\ $ p(a\mathit{\mid}b , c)\mathit{>}p(a\mathit{\mid}\lnot b , c)$ and$\medskip$\\ $ p(a\mathit{\mid}b,\lnot c)\mathit{>}p(a\mathit{\mid}\lnot b,\lnot c),$\end{tabular } \end{array } \vspace*{-0.07in}\label{ph}\ ] ] where ` @xmath110 ' denotes the ` negation ' operator .
malinas and bigelow ( 2016 ) illustrate ( [ ph ] ) using made up numbers that satisfy the above inequalities , and describe the source of the paradox as follows : they proceed to claim that their artificial illustration provides a way to explain an empirical example from cohen and nagel ( 1934 ) concerning death rates in 1910 from tuberculosis in richmond , virginia and new york city .
as argued above , however , in the case of observed data some of the ` numbers ' used in such arguments might be statistically untrustworthy , undermining the soundness of the logical argument in ( [ ph ] ) .
indeed , oversimplifications of the form ( [ ph ] ) , contribute to the perpetuation of the misconceptions beleaguering the paradox .
what is often insufficiently appreciated in statistical modeling and inference is that the inference propositions ( optimal estimators , tests , and predictors and their sampling distributions ) depend crucially on the validity of the probabilistic assumptions one imposes on the data .
the totality of these assumptions comprise the underlying statistical model , which is used to define the distribution of the sample and the likelihood function .
if the statistical model is misspecified , in the sense that any of its assumptions are invalid for the particular data , the reliability of inference based on such a model is usually undermined , giving rise to untrustworthy evidence .
the paper revisited simpson s paradox using the statistical misspecification perspective with a view to shed light on several silent features of the paradox . using this perspective
, it was argued that the key to unraveling the various counterintuitive results associated with this paradox is to formalize the vague notion of ` spurious ' inference results into ` statistically untrustworthy ' results which can be evidenced using misspecification testing .
this enables one to distinguish between two different cases of the paradox as it relates to the reversal of statistical associations .
case 1 , where the reversal is statistically trustworthy because the underlying statistical models are statistically adequate ( example 1 ) .
case 2 , where the apparent reversal is statistically untrustworthy due to statistical misspecification ( examples 2 - 5 ) .
the real issue is whether the inference results pertaining to statistical associations are statistically trustworthy or not , and the key criterion to appraise that is statistical adequacy .
hence , the statistical misspecification perspective puzzles out simpson s paradox because in both cases there is nothing counterintuitive to explain .
the statistical misspecification perspective is also used to revisit the causal dimension of the paradox by distinguishing between statistical and substantive inadequacy ( spuriousness ) . to ensure the reliability of any inferences relating to testing whether a third variable constitutes a confounder , requires that the underlying statistical model is statistically adequate .
this is particularly problematic for the causal resolution of the paradox when the third variable is related to a relevant ordering of interest which is revealed after the data are collected . in such cases one needs to account for any departures from the model assumptions as they relate to the ordering in question , and replace the generic terms used to capture the neglected statistical information with substantively meaningful explanatory variables.@xmath111 99 armistead , t.w .
( 2014 ) , resurrecting the third variable : a critique of pearl s causal analysis of simpson s paradox , _ the american statistician _ , 68 : 1 - 7 .
bickel , p.j .
hammel , and j.w .
oconnell ( 1975 ) , sex bias in graduate admissions : data from berkeley , _ science _ , 187 , 398404 .
bishop , y.m .
fienberg , p.w .
holland ( 1975 ) , _ discrete multivariate analysis : theory and practice _ , mit press , cambridge , ma .
blyth , c.r . , ( 1972 ) , on simpson s paradox and the sure thing principle , _ journal of the american statistical association _ , 67 : 364366 .
carlton , a. g. ( 1946 ) , estimating the parameters of a rectangular distribution , _ the annals of mathematical statistics _
, * 17 * , 355 - 358 .
cartwright , n. , ( 1979 ) , causal laws and effective strategies , _ nos _ , 13 ( 4 ) : 419437 .
cohen , m. , and nagel , e. ( 1934 ) , _ an introduction to logic and the scientific method _
, new york : harcourt , brace and company .
fisher , r. a. ( 1922 ) , on the mathematical foundations of theoretical statistics , _ philosophical transactions of the royal society a _ , * 222 * : 309 - 368 .
greene , w.h .
( 2011 ) , _ econometric analysis _ , 7th ed .
, new jersey : prentice hall .
kolmogorov , a. n. ( 1933 ) , _ foundations of the theory of probability _ , 2nd english edition , chelsea publishing co. ny . le cam , l. ( 1986 ) , _ asymptotic methods in statistical decision theory _ , springerverlag ,
lindley , d.v . , and m.r .
novick ( 1981 ) , the role of exchangeability in inference , _ journal of the american statistical association _
, 9 : 4558 .
malinas , g. and j. bigelow ( 2016 ) , simpson s paradox , the stanford encyclopedia of philosophy ( summer 2016 edition ) , edward n. zalta ( ed . ) , forthcoming url = @xmath112http://plato.stanford.edu / archives / sum2016/entries / paradox - simpson/@xmath113 .
pearl , j. ( 2009 ) , _ causality : models , reasoning , and inference _
( 2nd ed . ) , cambridge university press , ny . pearl , j. ( 2011 ) , why there is no statistical test for confounding , why many think there is , and why they are almost right , department of statistics , ucla .
pearl , j. ( 2014 ) , comment : understanding simpson s paradox , _ the american statistician _
, 68(1 ) , 8 - 13 .
pearson , k. ( 1896 ) , mathematical contributions to the theory of evolution on a form of spurious correlation which may arise when indices are used in the measurement of organs , _ proceedings of the royal society of london _
, 60(359 - 367 ) , 489 - 498 .
samuels , m.l .
( 1993 ) , simpson s paradox and related phenomena , _ journal of the american statistical association _
, 88:421 , 81 - 88 .
simpson , e.h .
( 1951 ) , the interpretation of interaction in contingency tables , _ journal of the royal statistical society_. series b ( methodological ) , 238 - 241 . sober , e. ( 2001 ) , venetian sea levels , british bread prices , and the principle of the common cause , _ the british journal for the philosophy of science _ , 52 : 331 - 346 .
spanos , a. , ( 1986 ) , _ statistical foundations of econometric modelling _ , cambridge university press , cambridge .
spanos , a. ( 1999 ) , _ introduction to probability theory and statistical inference _ , cambridge university press .
spanos , a. ( 2006a ) , where do statistical models come from ? revisiting the problem of specification , pp .
98 - 119 in _ optimality : the second erich l. lehmann symposium _ , edited by j. rojo , lecture notes - monograph series , vol .
49 , institute of mathematical statistics .
spanos , a. ( 2006b ) , revisiting the omitted variables argument : substantive vs. statistical adequacy , _ journal of economic methodology _
, 13 : 179 - 218 .
spanos , a. ( 2010 ) , statistical adequacy and the trustworthiness of empirical evidence : statistical vs. substantive information , _ economic modelling _
, 27 : 14361452 .
spanos , a. ( 2015 ) , statistical mis - specification testing in retrospect and prospect , working paper , virginia tech .
spanos , a. and a. mcguirk ( 2001 ) , the model specification problem from a probabilistic reduction perspective , _ journal of the american agricultural association _ , * 83 * : 1168 - 1176 .
spanos , a. and a. mcguirk ( 2002 ) , the problem of near - multicollinearity revisited : erratic vs. systematic volatility , _ journal of econometrics _ , * 108 * : 365 - 393 . spirtes , p. , c.n .
glymour , r. scheines ( 2000 ) , _ causation , prediction , and search_. mit press , ma .
stigler , s. m. ( 1980 ) , stigler s law of eponymy , _ transactions of the new york academy of sciences _ , 39 : 147 - 157 .
wasserman , l. ( 2004 ) , _ all of statistics : a concise course in statistical inference _ , springer , ny . williams , d. ( 1991 ) , _ probability with martingales _
, cambridge university press , cambridge .
yule , g.u .
( 1903 ) , notes on the theory of association of attributes in statistics , _ biometrika _ , 2 : 121 - 134 .
yule , g.u .
( 1909 ) , the applications of the method of correlation to social and economic statistics , journal of the royal statistical society , 72 : 721 - 730 .
yule , g. u. ( 1910 ) , on the interpretation of correlations between indices or ratios , _ journal of the royal statistical society _ , 73 : 644 - 647 .
yule , g. u. ( 1921 ) , on the time - correlation problem , with especial reference to the variate - difference correlation method , _ journal of the royal statistical society _ , 84 : 497 - 537 .
yule , g.u .
( 1926 ) , why do we sometimes get nonsense - correlations between time - series ? a study in sampling and the nature of time - series , _ journal of the royal statistical society _ , 89 : 1 - 64 .
the traditional specification of the lr model takes the form:@xmath2{c}y_{t}\mathit{=}\beta_{0}+\mathbf{\beta}_{1}^{\top}\mathbf{x}_{t}\mathit{+}u_{t},\medskip\\ \text{\lbrack i]\ } e(u_{t}\mathit{\mid}\mathbf{x}_{t}\mathit{=}\mathbf{x}_{t})\mathit{=}0,\ \text{[ii]\ } e(u_{t}^{2}\mathit{\mid}\mathbf{x}_{t}\mathit{=}\mathbf{x}_{t})\mathit{=}\sigma^{2},\medskip\\ \text{\lbrack iii ] } e(u_{t}u_{s}\mathit{\mid}\mathbf{x}_{t}\mathit{=}\mathbf{x}_{t})\mathit{=}0,\ \text{[iv ] } u_{t}\mathit{\backsim}\text{\textsf{n}}(.,.),\ t\mathit{\in}\mathbb{n}. \end{array}\ ] ] assumptions [ i]-[iii ] relating to the first two moments of the conditional distribution @xmath115 imply that the model parameters @xmath116:**=**@xmath117 have the following statistical parameterizations in terms of the primary parameters of the joint distribution @xmath118 , @xmath119 : @xmath59{ll}$\beta_{0}\mathbf{=}e(y_{t})\mathit{-}\mathbf{\beta}_{1}^{\top}e(\mathbf{x}_{t}),$ & $ \mathbf{\beta}_{1}\mathit{=}\mathbf{[}cov(\mathbf{x}_{t})]^{-1}cov(\mathbf{x}_{t},y_{t}),\medskip$\\ \multicolumn{2}{l}{$\sigma^{2}\mathbf{=}var(y_{t})\mathit{-}cov(\mathbf{x}_{t},y_{t})^{\top}[cov(\mathbf{x}_{t})]^{-1}cov(\mathbf{x}_{t},y_{t})$}\end{tabular } \label{p}\ ] ] assumption [ i ] implies that:@xmath15@xmath2{c}e(u_{t}\mathit{\mid}\mathbf{x}_{t}\mathit{=}\mathbf{x}_{t})\mathbf{=}0\leftrightarrow e(y_{t}\mathit{\mid}\mathbf{x}_{t}\mathit{=}\mathbf{x}_{t})\mathbf{=}\beta_{0}+\mathbf{\beta}_{1}^{\top}\mathbf{x}_{t}. \end{array } \label{ce}\ ] ] the law of iterated expectations ( williams , 1991 ) : @xmath120{c}\left ( e\left [ e(y\mathit{\mid}\sigma(\mathbf{x}))\right ] \right ) \mathbf{=}e(y ) , \end{array } $ ] where @xmath121 denotes the sigma - field generated by @xmath122 implies that:@xmath123 \mathbf{=}e(y_{t})\mathbf{=}\beta_{0}\mathit{+}\mathbf{\beta}_{1}^{\top}e(\mathbf{x}_{t})\longrightarrow\beta_{0}\mathit{=}e(y_{t})-\mathbf{\beta}_{1}^{\top } e(\mathbf{x}_{t})\ ] ] substituting @xmath124 back into @xmath125 yields:@xmath126 \mathit{=}\mathbf{\beta}_{1}^{\top}\left [ \mathbf{x}_{t}-e(\mathbf{x}_{t})\right ] + u_{t}.\ ] ] post - multiplying both sides by @xmath127 ^{\top}$ ] and taking expectations yields:@xmath2{cl}cov(y_{t},\mathbf{x}_{t } ) & \mathit{:=}e\left ( \left [ y_{t}\mathit{-}e(y_{t})\right ] \left [ \mathbf{x}_{t}\mathit{-}e(\mathbf{x}_{t})\right ] ^{\top}\right ) \mathit{=}\\ & \mathit{=}\mathbf{\beta}_{1}^{\top}\left [ \mathbf{x}_{t}\mathit{-}e(\mathbf{x}_{t})\right ] \left [ \mathbf{x}_{t}\mathit{-}e(\mathbf{x}_{t})\right ] ^{\top}\mathit{+}e(u_{t}\left [ \mathbf{x}_{t}\mathit{-}e(\mathbf{x}_{t})\right ] ^{\top } ) .
\end{array}\ ] ] since , the last term is zero : @xmath120{c}e(\mathbf{x}_{t}^{\top}u_{t})\mathit{=}e\left [ e(u_{t}\mathit{\mid}\sigma(\mathbf{x}_{t}))\right ] \mathit{=}0 , \end{array } \smallskip\newline$]it follows that : @xmath128^{-1}cov(\mathbf{x}_{t},y_{t})$ ] . in the case of
@xmath129 we use a theorem analogous to the _ lie _ for the variance ( williams , 1991):@xmath2{c}var(y_{t})=e\left [ var(y_{t}\mathit{\mid}\sigma(\mathbf{x}_{t}))\right ] + var\left [ e(y_{t}\mathit{\mid}\sigma(\mathbf{x}_{t}))\right ] , \end{array}\ ] ] where , by definition @xmath130 \mathbf{=}\sigma^{2}.$ ] the mean deviation of ( [ ce ] ) is:@xmath2{c}\left [ \beta_{0}\mathit{+}\mathbf{\beta}_{1}^{\top}\mathbf{x}_{t}\right ] -e\left ( \left [ \beta_{0}\mathit{+}\mathbf{\beta}_{1}^{\top}\mathbf{x}_{t}\right ] \right ) \mathit{=}\mathbf{\beta}_{1}^{\top}\left [ \mathbf{x}_{t}-e(\mathbf{x}_{t})\right ] , \end{array}\ ] ] and thus , by definition:@xmath2{c}var\left [ e(y_{t}\mathit{\mid}\mathbf{x}_{t})\right ] \mathbf{=}e\left [ \mathbf{\beta}_{1}^{\top}\left [ \mathbf{x}_{t}\mathit{-}e(\mathbf{x}_{t})\right ] \left [ \mathbf{x}_{t}\mathit{-}e(\mathbf{x}_{t})\right ] ^{\top}\mathbf{\beta}_{1}\right ] \mathit{=}\mathbf{\beta}_{1}^{\top}\left [ cov(\mathbf{x}_{t})\right ] \mathbf{\beta}_{1}. \end{array } \vspace*{-0.1in}\ ] ] from this , it follows that : @xmath2{c}var(y_{t})\mathbf{=}\sigma^{2}\mathit{+}\mathbf{\beta}_{1}^{\top}\left [ cov(\mathbf{x}_{t})\right ] \mathbf{\beta}_{1}\longrightarrow\sigma ^{2}\mathit{=}var(y_{t})\mathit{-}\mathbf{\beta}_{1}^{\top}\left [ cov(\mathbf{x}_{t})\right ] \mathbf{\beta}_{1 } , \end{array } \vspace*{-0.1in}\ ] ] which yields the parameterization in ( [ p ] ) . | the primary objective of this paper is to revisit simpson s paradox using a statistical misspecification perspective .
it is argued that the reversal of statistical associations is sometimes spurious , stemming from invalid probabilistic assumptions imposed on the data .
the concept of statistical misspecification is used to formalize the vague term ` spurious results ' as ` statistically untrustworthy ' inference results .
this perspective sheds new light on the paradox by distingusing between statistically trustworthy vs. untrustworthy association reversals .
it turns out that in both cases there is nothing counterintuitive to explain or account for .
this perspective is also used to revisit the causal ` resolution ' of the paradox in an attempt to delineate the modeling and inference issues raised by the statistical misspecification perspective .
the main arguments are illustrated using both actual and hypothetical data from the literature , including yule s `` nonsense - correlations '' and the berkeley admissions study .
keywords : association reversal ; spurious correlation ; statistical misspecification ; statistical vs. substantive adequacy ; misspecification testing ; untrustworthy evidence ; causality ; confounding . |
quantum rabi model ( qrm ) describes a two - level atom ( qubit ) coupled to a cavity electromagnetic mode ( an oscillator ) @xcite , a minimalist paradigm of matter - light interactions with applications in numerous fields ranging from quantum optics , quantum information science to condensed matter physics .
the solutions to the qrm are however highly nontrivial .
whether an analytical exact solution even exists is uncertain for a long time .
recently , braak presented an analytical exact solution @xcite to the qrm using the representation of bosonic creation and annihilation operators in the bargmann space of analytical functions @xcite .
a so - called @xmath0-function with a single energy variable was derived yielding exact eigensolutions , which is not in the closed form but well defined mathematically .
alternatively , using the method of extended coherent states ( ecs ) , this @xmath0-function was recovered in a simpler , yet physically more transparent manner by chen _
et al . _
braak s solution has stimulated extensive research interests in the single - qubit qrm @xcite .
as quantum information resources , such as the quantum entanglement @xcite and the quantum discord @xcite , can be easily stored in two qubits , two qubits in a common cavity have potential applications in quantum information technology .
such a model system with two qubits now can be constructed in several solid devices @xcite .
recently , some analytical studies to the qrm with two qubits have been attempted within various approaches @xcite . by using the ecs technique ,
a @xmath0-function for the qrm with two equivalent qubits , resembling the simplest one without a determinant in the single - qubit qrm @xcite , was obtained @xcite .
while in the bargmann representation , the @xmath0-function was built as a high order determinant , such as @xmath2 determinants for qrm with two different qubits @xcite . practically , the qrm with two arbitrary qubits is the most general one , and can be realized in experimental device with the greatest possibility .
we believe that a simpler @xmath0-function is more convenient to obtain the eigensolutions , and can also shed light on many physical processes more clearly . in this work , employing the ecs , we demonstrate a successful derivation of a very concise @xmath0-function , which is just a @xmath3 determinant for the qrm with two arbitrary qubits .
furthermore , for the same coupling , the @xmath0-function even for two different qubits can be reduced to a simplest one without the use of the determinant , like that in the single - qubit qrm @xcite .
the hamiltonian of the qrm with two qubits can be generally written as @xcite @xmath4 where @xmath5 is the energy splitting of the @xmath6-th qubit , @xmath7 creates one photon in the common single - mode cavity with frequency @xmath8 , @xmath9 describes the coupling strength between the @xmath6-th qubit and the cavity , @xmath10 @xmath11 and @xmath12 are the usual pauli matrices of the @xmath6-th qubit . after a rotation with respect to the @xmath13-axis by an angle @xmath14 the hamiltonian in the two - qubit basis @xmath15 , @xmath16 , @xmath17 , and @xmath18 , which are eigenstates@xmath19of @xmath20 , can be written as the following symmetric matrix ( in unit of @xmath21 ) @xmath22 where @xmath23 and @xmath24 . for later use , we express the wavefunction in terms of the fock space as @xmath25 + b_n\left [ \left| 1,-1\right\rangle \pm \left ( -1\right ) ^n\left| -1,1\right\rangle \right ] \right\ } |n\rangle , \label{wave_d}\ ] ] where @xmath26 corresponds to even ( odd ) parity , @xmath27is the photonic number state .
the schrdinger equation leads to the recurrence relation @xmath28 b_m-\left ( m - e\right ) a_m - ga_{m-1}}{g\left ( m+1\right ) } , \nonumber \\
b_{m+1 } & = & \frac{\left [ \delta _ 2\pm \delta _ 1\left ( -1\right ) ^m\right ] a_m-\left ( m - e\right ) b_m - g^{\prime } b_{m-1}}{g^{\prime } \left ( m+1\right ) } .
\label{coeff_d}\end{aligned}\ ] ] note that they can not be reduced to a linear three - term recurrence form .
the coefficients @xmath29 can be obtained in terms of two initial values of @xmath30 and @xmath31 recursively . in this paper , we will first study the general case of different coupling strengths with the same cavity , then we turn to the special equal coupling case . to employ the ecs approach , we first perform the following pair of bogoliubov transformations with finite displacements @xmath33 by which some diagonal matrix element @xmath10can be reduced to the free particle number operators plus a constant , which is very helpful for the further study . the wavefunction can be expanded in the fock space of the operator @xmath34 @xmath35 |n\rangle _ { a_{+ } } , \label{wave_a+}\ ] ] where@xmath36is the number state in the @xmath37-space , and termed as the ecs previously @xcite .
the schrdinger equation straightforwardly gives @xmath38 which can not be reduced to the linear three - term recurrence relation either .
note that if three initial coefficients @xmath39 are given , all other coefficients will be uniquely determined recursively .
considering the conserved parity , the wavefunction can be also expressed in series expansion in the fock space of @xmath40 @xmath41 |n\rangle _ { a_{-}}.\ ] ] the wavefunction for non - degenerate state should be the same , so we have @xmath42 projecting onto @xmath43 and with the use of @xmath44 , we have one linear equation @xmath45 g^n=0 , \label{ga}\ ] ] where @xmath26 in left - hand - side is for even ( odd ) parity
. we can also expand the wavefunction in the fock space of the second displaced operators @xmath46 @xmath47 |n\rangle _ { b_{+}}. \label{wave_b}\ ] ] the schrdinger equation yields @xmath48 all coefficients for @xmath49 can be determined from three initial parameters @xmath50 recursively . through the similar procedure , we can obtain the second linear equation as @xmath51 ( g^{\prime})^n=0 .
\label{gb}\ ] ] there seems to be @xmath52 initial coefficients in eqs .
( [ ga ] ) and ( [ gb ] ) .
fortunately , they can be determined by coefficients in the series expansion eq .
( [ wave_d ] ) in the original fock space .
for example , the same wavefunction for non - degenerate states implies @xmath53 .
projecting onto @xmath54 yields @xmath55 where we have used @xmath56 , and removed irrelevant constants .
then we can obtain @xmath57and@xmath58 in eq .
( [ ga ] ) through eq .
( [ coeff_a ] ) recursively .
similarly , @xmath59 can also be obtained by coefficients @xmath60 and @xmath61 . through eq .
( [ coeff_b ] ) , we can get all @xmath62 and @xmath63 in eq .
( [ gb ] ) . inserting these coefficients into eqs .
( [ ga ] ) and ( [ gb ] ) , we can arrive at two linear equations for @xmath64 and @xmath31@xmath65 where @xmath66and @xmath67 are obtained from eqs .
( [ ga ] ) and ( [ gb ] ) @xmath68by setting @xmath69 and @xmath70 ; @xmath71and @xmath72are obtained bysetting @xmath73 and @xmath74 . the @xmath0-function is then defined with the following @xmath3 determinant @xmath75 nonzero coefficients require the vanishing @xmath0-function in the real physics problems . so the zeros of this @xmath0-function will give the energy spectrum for the present model .
in addition , from the first one in eq .
( [ coeff_a ] ) and the second one in eq .
( [ coeff_b ] ) , we immediately know that @xmath76 and @xmath77 are two sets of the exceptional solutions , which are not the zeros of the @xmath0-function .
it is interesting to note from the coefficients in eqs .
( [ coeff_a ] ) , ( [ coeff_b ] ) , and ( [ initial ] ) that the present @xmath0-function is a well defined transcendental function .
thus analytical exact solutions have been formally found .
-functions for the two - qubit qrm with even ( blue ) and odd ( red ) parity using different scales in four energy intervals connected successively .
crosses denote the unstable zeros .
( bottom ) @xmath78 for different zeros with the serial number @xmath79 .
@xmath80 , @xmath81 and @xmath82 . ,
width=302 ] we plot the @xmath0-function for @xmath83 @xmath84and @xmath85 in fig .
[ g_function_n2_diff ] .
the stable zeros reproduce all regular spectra , which can be confirmed by the numerical exact solutions .
typically , the convergence is assumed to be achieved if zeros ( i.e. @xmath86 ) are determined within relative errors @xmath87 , where @xmath88 is the truncated number of the series expansions in the space of the displaced operators .
we also calculate @xmath78 for all zeros , which are exhibited in fig .
[ g_function_n2_diff ] ( b ) .
for the stable zeros , @xmath89 is almost the same as the relative errors and should be around the @xmath90 line within error @xmath91 , much smaller than the symbol size . while , very few unstable zeros , which are not the true eigenvalues , are also present in practical calculations because of unavoidable finite truncations .
fortunately , they can be excluded very easily .
they are very sensitive to the truncated number @xmath88 and can not converge with increasing @xmath88 , because the corresponding coefficients oscillate with increasing magnitudes as @xmath92 increases . in sharp contrast , for the stable zeros , the coefficients converges to zero rapidly with @xmath92 .
the positions of unstable zeros must change even increasing @xmath88 by @xmath93 .
so they can be easily figured out in fig .
[ g_function_n2_diff ] ( b ) for apparent deviation from @xmath90 line .
it should be pointed out here that these unstable zeros are absolutely not the true zeros of @xmath0-function , and will disappear if the summations are really performed infinitely .
the baselines shown in fig . [ g_function_n2_diff ] ( a ) are close to @xmath94 and @xmath95 exceptional eigenvalues due to the divergence in the @xmath0-functions .
generally , the exceptional eigenvalues hardly occur for given rational model parameters .
it is interesting to note that the present @xmath0-function within ecs is only a @xmath3 determinant , much simpler than those with @xmath2 determinants in the same model using the bargmann representation @xcite .
for the same coupling , @xmath97 the hamiltonian matrix ( [ h_matrix ] ) is of higher symmetry due to @xmath98 , the solution will become simpler .
the wavefunction in the series expansion in the original fock space is the same as eq .
( [ wave_d ] ) , but the recurrence relation obtained from the schrdinger equation should take a simpler form @xmath99 a_m & = & \left ( m - e\right ) b_m , \label{coeff_ab } \\ \left
[ \delta _ 2\pm \delta _ 1\left ( -1\right ) ^m\right ] b_m & = & \left ( m - e\right ) a_m+g\left [ a_{m-1}+\left ( m+1\right ) a_{m+1}\right ] .
\nonumber\end{aligned}\ ] ] note that one - to - one relation of @xmath100and @xmath101is found , and thus a three - term recurrence relation can be obtained , in contrast with the case of @xmath32 .
the coefficients can be obtained in terms of only one initial value @xmath74 or @xmath69 recursively , so the continued - fraction technique is directly applicable , similar to that in the single - qubit qrm @xcite .
we can choose initial parameter @xmath74 for @xmath102,@xmath10 and @xmath103for @xmath104 to avoid the artificial divergence .
alternatively , we propose a mathematically well - defined technique . the wavefunction can be expressed as eq .
( [ wave_a+ ] ) in terms of the sole pair of the displaced operators @xmath105 in eq .
( [ operator_a ] ) . set @xmath106 , eq .
( [ coeff_a ] ) can be reduced to the following recurrence relation @xmath107 although they can not be reduced to a linear three - term relation either , the coefficients can be uniquely given recursively by only two initial coefficients @xmath108 and @xmath109 in this case .
to this end , we can define the @xmath0-function as @xmath110 g^n=0 . \label{g_fun_same}\ ] ] next we need build the relationship between two sets of coefficients for series expansions in @xmath111 and @xmath105 .
for the same wavefunction , we have @xmath112 b_n|n\rangle , \;\sum_{n=0}^\infty \sqrt{n!}z_n^a|n\rangle _ { a_{+}}=\pm r\sum_{n=0}^\infty \left ( -1\right ) ^n% \sqrt{n!}a_n|n\rangle .\ ] ] projecting onto @xmath54 yields @xmath113 both @xmath108 and @xmath109 are determined by coefficients @xmath60 and @xmath61 , which are only dependent on the initial parameter
. hence @xmath114 and @xmath115 in eq . ( [ g_fun_same ] ) can be obtained recursively from eq .
( [ r_coeff_same ] ) .
thus for the qrm of two different qubits but with the same coupling , we have derived the simplest @xmath0-function without the use of the determinant , similar to that in the one - qubit qrm @xcite .
note that the @xmath0-function in this case in ref .
@xcite is a @xmath116 determinant .
similarly , we also plot the @xmath0-function for @xmath83 @xmath81 and @xmath117 in fig .
[ g_function_n2_same ] .
the stable zeros give all eigenvalues and the unstable zeros can be distinguished by the same trick outlined above .
-functions for the two - qubit qrm with the same coupling constants with even ( blue ) and odd ( red ) parity .
crosses denote the unstable zeros .
( bottom ) @xmath118 for different zeros with the serial number @xmath119 .
@xmath117 and @xmath120 .
, width=302 ] from eq .
( [ r_coeff_same ] ) , we know that @xmath121 is an exceptional solution . however , @xmath122 is neither an exceptional solution nor a singularity for @xmath10 @xmath123 , in contrast with the previous observation @xcite .
we attribute the difference to the possible enlarged dimension where their g function is defined .
the final results should be the same in both kinds of treatments , but the present scheme is much more concise and allows an in - depth discussion . by eq .
( [ coeff_ab ] ) we may write the recurrence relation as @xmath124 ^2}{\left ( m - e\right ) } a_m-\left ( m - e\right ) a_m - ga_{m-1}.\ ] ] for @xmath125 , the analyticity of the wave function requires @xmath122 for the even @xmath126 for odd parity and odd @xmath126 for even parity , independent of the coupling strength @xmath127 .
it is just the trivial eigenvalue for the spin - singlet state . for @xmath123 ,
the first two coefficients for the states with even parity are @xmath128 a_0;\;a_2=\lim_{e\rightarrow 1}\frac 1{2g}\left [ \frac{a_1}{e-1}\left ( \delta _ 2-\delta _ 1\right ) ^2-ga_0\right ] .
\nonumber\ ] ] the non - analyticity of the eigenfunction only occurs for the possible divergence of @xmath129 where the denominator is zero .
to lift the pole of @xmath130 , it is required that @xmath131 , which immediately yields @xmath132 while the first two coefficients for states with odd parity are @xmath133 a_0;\;a_2=\lim_{e\rightarrow 1}\frac 1{2g}\left [ \frac{a_1}{e-1}\left ( \delta _ 2+\delta _ 1\right ) ^2-ga_0\right ] .
\nonumber\ ] ] the analyticity of @xmath134requires @xmath135 which gives @xmath136 equations ( [ cond_1 ] ) and ( [ cond_2 ] ) are just the conditions for special dark states with @xmath137 found by peng et al .
, @xcite .
so they can be easily figured out in the continued - fraction technique . .
( left ) @xmath83 @xmath138 ; ( right ) @xmath139 @xmath140 .
@xmath122 denoted by dashed lines .
, title="fig:",width=226 ] .
( left ) @xmath83 @xmath138 ; ( right ) @xmath139 @xmath140 .
@xmath122 denoted by dashed lines . ,
title="fig:",width=226 ] figure [ energy_e1 ] presents the energy spectra for two sets of @xmath141 , @xmath142 , which satisfy eqs .
( [ cond_1 ] ) and ( [ cond_2 ] ) respectively .
it is shown clearly that @xmath137 for @xmath143 and @xmath144 are @xmath127 independent eigenvalues for even and odd parity respectively .
the states with @xmath145 are not the dark states , which are exhibited obviously in fig .
[ energy_e1 ] .
in this work , we have derived the concise @xmath0-function for the qrm with two arbitrary qubits by using ecs , which leads to simple , analytic solutions .
although the coefficients in the recurrence relations look complicated , actually they can be uniquely and straightforwardly given by the schrdinger equations .
our @xmath0-function is only a @xmath146 determinant for the general case and a rather simple one without a determinant for the same coupling case , much more concise than those derived recently in the bargmann space .
this work is to extend the methodology of a compact @xmath0-function in the qrm with one qubit to the qrm with two arbitrary qubits in the simplest way , thereby allowing a conceptually clear , practically feasible treatment to energy spectra .
it is our expectation that the present concise approach will find more applications in the future .
we stress that the present analytic solution is well defined mathematically , because of no built - in truncations , which is essentially different from the previous finite truncation approaches @xcite and the continued - fraction technique , therefore of both fundamental and practical interest .
qhc acknowledge useful discussions with daniel braak and jie peng .
this work was supported by national natural science foundation of china under grant no . 11174254 , national basic research program of china under grant no .
2011cba00103 .
a. moroz , europhys . lett .
* 100 * ( 2012)60010 ; i. travenec , phys . rev . a * 85 * ( 2012)043805 ; b. gardas , j. dajka , j. phys .
a * 46 * ( 2013)265302 ; y. z. zhang , j. phys .
a * 54 * ( 2013)102104 ; a. j. maciejewski , m. przybylska , t. stachowiak , phys
. lett . a * 378 * ( 2014)3445 ;
h. h. zhong et al . , j. phys .
a * 46 * ( 2013)415302 ; inbid * 47 * ( 2014)045301 .
m. a. sillanpaa , j. i. park , and r. w. simmons r w , nature * 449 * ( 2007)438 ; g. haack , f. helmer , m. mariantoni , f. marquardt and e. solano , phys . rev .
b * 82 * ( 2010)024514 ; f. altintas and r. eryigit , phys .
lett . a * 376 * ( 2012)1791 .
j. benhelm , g. kirchmair , c. f. roos , and r. blatt , nature phys .
* 4 * ( 2008)463 ; c. a. ryan , m. laforest , and r. laflamme , new j. phys .
* 11 * ( 2009)013034 ; a. d. corcoles , phys .
a * 87 * ( 2013)030301 . | using extended coherent states , an analytical exact study has been carried out for the quantum rabi model ( qrm ) with two arbitrary qubits in a very concise way .
the @xmath0-functions with @xmath1 determinants are generally derived .
for the same coupling constants , the simplest @xmath0-function , resembling that in the one - qubit qrm , can be obtained .
zeros of the @xmath0-function yield the whole regular spectrum .
the exceptional eigenvalues , which do not belong to the zeros of the @xmath0 function , are obtained in the closed form . the dark states in the case of the same coupling
can be detected clearly in a continued - fraction technique .
the present concise solution is conceptually clear and practically feasible to the general two - qubit qrm and therefore has many applications .
two - qubit quantum rabi model , analytic solution , extended coherent state 03.65.ge , 42.50.ct , 42.50.pq |
after the discovery of hawking radiation @xcite in 1975 , it was widely anticipated that the semi - classical approach to gravity based on quantum field theory ( qft ) in curved spacetime will open the opportunity to explore various interesting physical problems , in which a quantum field interacts with the curved spacetime where it resides . among these problems , of special interest is the self - consistent dynamical evaporation of a black hole ( bh ) .
another outstanding problem in this class is the evolution of quantum fluctuations in the very early universe , and the resulting cosmological structure formation .
indeed , in the latter problem of cosmological quantum - field perturbations a remarkable progress has been achieved over the last few decades .
@xcite however , the problem of analyzing the semiclassical evolution of an evaporating bh still remains a serious challenge .
to understand the difficulties in analyzing this interesting problem of self - consistent bh evaporation , let us briefly review the basic structure of semiclassical gravity .
the metric @xmath2 is treated as a classical field , and it is assumed to satisfy the _
semiclassical einstein equation _
@xmath3 where @xmath4 is the einstein tensor associated with @xmath2 . and the @xmath5 signature . ] the source term @xmath6 is the regularized expectation value of the stress - energy tensor associated with a quantum field @xmath7 . for the sake of simplicity
we shall take @xmath8 here to be a scalar field .
it is supposed to satisfy the field equation @xmath9 where @xmath10 and @xmath11 respectively denote the mass and coupling constant of the scalar field .
note that the theory is semi - classical , as the field @xmath8 is quantized but the metric @xmath12 is classical , nevertheless the two are coupled . the major difficulty in solving ( or even analyzing ) the field equation ( [ eq : intro - qft - cst basic eqs ] ) has to do with the regularization of the divergent quantity @xmath13 .
as opposed to qft in flat spacetime in which one can use the normal - ordering procedure , in curved spacetime the outcome of this procedure depends on the choice of time slicing which is completely arbitrary .
there exists a regularization method named _ point - splitting _ ( ps ) , also known as covariant point separation , which gives a general prescription how to regularize quantities which are quadratic in the field and its derivatives such as @xmath13 .
alas , implementing this prescription in situations where the solution of the field equation ( [ eq : field_equation ] ) is known only numerically turns out as a surprisingly difficult problem . in this paper we shall focus on the regularization of @xmath1 instead of @xmath13 .
the quantity @xmath1 is also divergent , but not as strong as @xmath13 .
in addition , the scalar character of @xmath1 ( as opposed to the tensorial @xmath13 ) makes it easier to regularize .
these properties make @xmath1 a convenient tool to examine and explain new ideas concerning regularization . in order to calculate @xmath14 (
i.e. the regularized @xmath15 ) using the ps method we split the point @xmath16 and write it as a product of @xmath8 at two different points , namely @xmath17 .
this is known as the _ two - point function _ ( tpf ) .
we then subtract from the tpf a known _ counter - term _ and take the limit @xmath18 .
this limit , however , is what makes the numerical implementation so hard . in 1984
candelas and howard @xcite developed a method to numerically implement ps if a high - order wkb approximation for the field modes is known . using this method , they calculated @xmath0 in schwarzschild spacetime , and subsequently this method was used by howard @xcite to calculate @xmath6 in schwarzschild . later
, this method was extended to a general static spherically symmetric background , first for @xmath0 by anderson @xcite , and subsequently to @xmath6 by anderson , hiscock and samuel @xcite .
the limitation of the method proposed by candelas and howard is that it requires a high - order wkb approximation for the field modes : at least second - order for @xmath0 and forth order for @xmath6 . in the ordinary ( lorentzian - signature )
schwarzschild metric this task of high - order wkb expansion is very difficult , especially because of the presence of a turning point : for typical modes of large @xmath19 and @xmath20 , there is a turning point on the @xmath21 axis , at a value @xmath22 where the effective potential @xmath23 equals @xmath24 . and
@xmath20 there are usually _ two _ such roots . here
we shall explicitly refer to the larger one , but the same complications arise also at the smaller root . ]
the mode s radial function is essentially oscillatory at @xmath25 and exponential at @xmath26 . both these basic wkb approximations the oscillatory approximation at @xmath25 and the exponential approximation at @xmath26 break down and actually diverge at @xmath27 . to correctly match the two approximations
, one has to use another , intermediate approximation valid in the neighborhood of @xmath28 .
this _ turning - point approximation _ is based on the airy function . whereas the leading - order matching is manageable
, it becomes exceedingly hard to go to higher - order wkb , because each succeeding order will now require its own turning - point matching .
furthermore , to the best of our understanding , the series of powers involved in the airy - based turning - point expansion proceeds in powers of @xmath29 ( rather than @xmath30 ) . correspondingly , we may expect that in the presence of a turning point , to implement the wkb - based expansion to order @xmath31 ( required for calculating @xmath6 ) , one would have to carry matched asymptotic expansion up to _ twelfth _ order in @xmath29 ( or six such orders for @xmath0 ) a formidably difficult task . to overcome these difficulties , candelas and howard @xcite and several others @xcite used an elegant trick : they used wick rotation to analytically extend the background metric to the euclidean sector .
any static spacetime is guaranteed to have such a real - metric euclidean sector .
in the latter , the radial equation does not admit a turning point .
this way , it was possible to carry the wkb analysis to the desired order and to implement the above regularization scheme for @xmath0 as well as for @xmath6 . our ultimate goal , however , is to develop a regularization scheme for @xmath1 and @xmath13 , applicable to _ time - dependent _ backgrounds as well .
such a time - dependent metric ( even if spherically symmetric ) does @xmath32 generically admit a euclidean sector .
we are thus led to carry the analysis directly in the lorentzian sector , which in turn implies the presence of a turning point in the radial equation , hampering any attempt to carry high - order wkb expansion .
we shall therefore refrain from establishing our regularization scheme on the wkb analysis .
there is another obvious reason for avoiding wkb analysis : consider a time - dependent spherically - symmetric background .
the field equation for a given mode may still be expressed as a one - dimensional ( namely @xmath33 ) wave equation with an effective potential , but now the potential will be time - dependent .
in such a situation , even the leading - order wkb ( and even if we forget for the moment about the turning point ) becomes a non - trivial task , let alone higher - order wkb analysis . for these reasons , we shall not base our regularization scheme on high - order wkb expansion .
instead , in our method we extract the required information concerning the high - frequency field s modes directly from the well - known counter - term ( [ eq : basic ps - the counter - terms gds ] ) for @xmath0 ( and , for computing @xmath6 , from the counter - term developed by christensen @xcite ) .
we point out that this approach , namely extraction of the high - frequency asymptotic behavior of the modes from a known local counter - term , was actually initiated by candelas @xcite already before he and howard resorted to the euclidean sector @xcite .
however , candelas analysis was restricted to the schwarzschild case ( which in particular means restriction to staticity , spherical symmetry , and to vacuum ) . we should also comment that even in that case the analysis in ref .
@xcite was not completed , because the required integral of the regularized mode contribution over @xmath19 was not carried out .
when we attempted to implement this integral over @xmath19 , we found that it actually fails to converge in the usual sense , due to growing oscillations ( see below ) , which led us to introduce the notion of generalized integral .
our approach is in this sense a completion of candelas method , as well as its generalization beyond the vacuum case ( and with the scope of further extending it to dynamical backgrounds ) .
our method requires the field modes to admit a trivial decomposition in at least one of the coordinates ( e.g. through @xmath34 or spherical harmonics ) .
this usually corresponds to having a killing field in spacetime .
the splitting is then done in that trivial coordinate which enables us to treat the coincidence limit analytically .
so far we have developed two different variants of our method : ( i ) the _ @xmath35-splitting _ variant , which requires a time - translation symmetry ; ( ii ) the _ angular - splitting _ variant , which requires spherical symmetry .
we are also exploring a third variant , _ azimuthal splitting _ ( which would only require axial symmetry ) , but this one is still in progress . in all these variants we assume for simplicity that the background is asymptotically - flat , although this requirement can probably be relaxed . since our ultimate goal is to analyze @xmath6 on the time - dependent background of an evaporating bh , the @xmath35-splitting method is insufficient , and we shall actually need the angular - splitting variant .
it turns out , however , that the @xmath35-splitting variant is in some sense conceptually simpler and easier to present at first stage , because certain additional complications arise in the angular - splitting method .
although we know how to address these complications , they make the method s logical structure a bit more obscure and harder to explain . for this reason , for the sake of introducing our basic regularization strategy
we choose to present here the @xmath35-splitting variant , which is logically simpler . and for exactly the same reason , in this paper we shall display the regularization of @xmath1 rather than @xmath13 . then in the next paper we plan to present the angular - splitting variant , which is going to be our main tool for analyzing dynamical bh evaporation .
the tpf diverges for any pair of points connected by a null geodesic , even if they are far from each other @xcite .
as it turns out , this long - distance divergence of the tpf leads to undamped oscillations in the mode contributions at large @xmath19 . to address this issue we use the concept of _ generalized integral _ , in which these oscillations are properly damped upon integration over @xmath19 , which fully cures the oscillations problem .
the origin of this complication ( the presence of connecting null geodesics ) is discussed in sec . [ sub : remarks - about - the - tpf ] and also in appendix [ sec : appendix - b ] ; and the resolution of the oscillations problem by means of generalized integral ( and particularly the so called `` self - cancellation integral '' ) is described in sec . [ sub : remarks - about - the - tpf ] and further in appendix [ sec : appendix - a ] . in the description of the @xmath35-splitting method in sec .
[ sec : the - new - scheme ] we assume a spherically - symmetric static background for the sake of simplicity .
however , as was discussed above , the @xmath35-splitting method does not require spherical symmetry , and in principle it may be applied to a generic ( asymptotically - flat ) stationary spacetime .
we outline this generalization of the method to stationary backgrounds in sec . [
sub : our method - non spherical symmetric ] .
we point out , however , that some completion is still required in the case of a stationary eternal bh ( see therein ) .
next we apply our method explicitly to the schwarzschild case , computing @xmath0 in boulware state .
we compare our results to those obtained previously by anderson ( using the euclidean sector ) , and find full agreement .
this article is divided as follows : in sec .
[ sec : basic point splitting ] we present the basic ps method , and then briefly outline the procedure developed by candelas and howard @xcite .
note that sec .
[ sub : remarks - about - the - tpf ] discusses certain subtleties of the tpf in some detail , and may be skipped in first reading . in sec .
[ sec : the - new - scheme ] we present our @xmath35-splitting method .
we first describe it for spherically - symmetric static background , and then outline its generalization to a generic stationary background . in sec .
[ sec : calc in schwarzschild ] we harness this method for the calculation of @xmath0 in the schwarzschild metric .
finally , in sec .
[ sec : discussion ] we discuss the implications of our new method and try to pave the path towards our ultimate goal of investigating self - consistent bh evaporation .
we start by sketching the basic ps regularization method .
this method is aimed to regularize the expectation value of various quantities which are quadratic in the field operator ( and its derivatives ) . among these quantities ,
the most important one is probably the energy - momentum tensor @xmath13 .
however , in this first paper we shall consider @xmath1 as a simpler example ( although we shall occasionally remark on the analogous calculation of @xmath13 ) . for simplicity
we consider here a quantum scalar field @xmath7 living in a static , spherically symmetric , asymptotically flat spacetime with metric @xmath36 where @xmath37 . the field operator may then conveniently be expressed as @xmath38 here , @xmath39 and @xmath40 are the creation and annihilation operators of the field s @xmath41 mode , @xmath42 is a complete , orthonormal , family of modes and @xmath43 is @xmath44 . ]
taking the form @xmath45 and @xmath46 are the usual spherical harmonics .
the radial functions @xmath47 are obtained by solving the field equation for @xmath7 with the decomposition ( [ eq : basic ps - expression for f ] ) and with appropriate boundary conditions .
the operation of summation over @xmath48 and integration over @xmath19 repeats many times in the analysis below
. we shall generally refer to this operation as the `` mode sum '' ( despite the slight abuse of terminology ) .
we point out , however , that the decomposition ( [ eq : basic ps - field decomposition to a , ad ] ) applies as - is in the case of an asymptotically - flat background spacetime with simple asymptotic structure ( like e.g. minkowski or a star ) .
but if the background spacetime is an eternal bh with a past horizon , then for each @xmath41 combination there are actually two orthonormal modes , namely the `` in '' and `` up '' modes .
the `` in '' modes are those described above ( namely monochromatic waves propagating from past null infinity ) , and the `` up '' modes describe monochromatic waves that emerge from the past horizon . in this case of eternal bh , the mode sum should also include a summation over the contributions of these two independent modes for each @xmath41 ( as explicitly described in sec . [ sec : calc in schwarzschild ] for the schwarzschild case ) .
the quantity @xmath1 obviously depends on the quantum state . naturally , one would like to evaluate it in the vacuum state .
the latter is defined to be the quantum state annihilated by each of the above @xmath40 operators .
we shall consider the boulware vacuum state . ] trying to naively calculate @xmath15 in the vacuum state yields the divergent expression @xmath49 although the sum over @xmath48 does converge for a given @xmath19 , the integral over @xmath19 diverges .
in fact , one can easily check that already in minkowski spacetime the integrand is @xmath50 ( and @xmath51 for @xmath13 ) , and the same divergence occurs in the schwarzschild case as well .
if one tries to integrate over @xmath19 before the summation , one finds that the integral over @xmath19 again diverges ( for a given @xmath52 ) .
one therefore needs to somehow regularize the expression for @xmath1 . in flat spacetime
this regularization may be easily achieved by normal ordering , but in curved spacetime this yields a slicing - dependent , non - unique result .
this is where point splitting comes into play .
dewitt @xcite proposed that @xmath15 ( and , more generally , quantities quadratic in the field operator and its derivatives ) can be treated by taking the product of the field operators in two separate points @xmath53 and then considering the coincidence limit @xmath54 .
more specifically , he showed that the regularized expectation value of @xmath55 can be defined as @xmath56.\label{eq : basic ps - phi ren limit}\ ] ] here @xmath57 is the dewitt - schwinger _ counter - term _ , namely a local term which fully captures the singular piece of the tpf . for a scalar field with mass @xmath10 and coupling constant @xmath11
it takes the form @xcite
@xmath58\\ -\frac{m^{2}}{16\pi^{2}}+\frac{1}{96\pi^{2}}r_{\alpha\beta}\frac{\sigma^{;\alpha}\sigma^{;\beta}}{\sigma}.\label{eq : basic ps - the counter - terms gds}\end{gathered}\ ] ] here @xmath59 and @xmath60 are respectively the ricci scalar and tensor , @xmath61 denotes the euler constant , and @xmath62 is the biscalar associated with the short geodesic connecting @xmath16 and @xmath63 .
the value of @xmath64 is half the geodesic distance squared ( see ref .
@xcite ) . more specifically , for a timelike separation @xmath65 , where @xmath66 denotes the proper time between @xmath16 and @xmath63 .
the parameter @xmath67 is unknown and it corresponds to the well - known ambiguity in the regularization procedure @xcite .
to be the field s mass @xmath10 .
@xcite ] it is not easy to directly implement the point - splitting method , especially because in the cases of interest ( e.g. black hole backgrounds ) the radial functions @xmath68 are only known from numerics .
furthermore , in eq .
( [ eq : basic ps - phi ren limit ] ) the tpf @xmath17 , which is to be computed numerically , diverges as @xmath69 ( and as @xmath70 for @xmath13 ) as @xmath63 approaches @xmath16 , where @xmath66 denotes the geodesic distance between @xmath16 and @xmath63 .
this problem was addressed by candelas , howard and later anderson and collaborators ( see refs .
@xcite ) a long time ago .
they developed a calculation scheme that allows regularization of @xmath1 or @xmath13 numerically , provided that an analytic approximation for the field is known , up to a sufficiently high order . , and a fourth order approximation to compute @xmath71 . ] in that scheme , one first analytically constructs the approximate singular piece of the field , which we denote @xmath72 .
it is composed of the contribution of the modes of large @xmath19 and @xmath20 , up to second order in @xmath73 and @xmath74 ( fourth order for @xmath13 ) , which is usually computed using wkb analysis . from this quantity
one then constructs the approximate tpf @xmath75 , and recasts eq .
( [ eq : basic ps - phi ren limit ] ) in the form @xmath76\\ + \lim_{x'\to x}\left[\left\langle \phi_{sing}(x)\phi_{sing}(x')\right\rangle -g_{ds}\left(x , x'\right)\right].\label{eq : basic ps - phi ren two limits}\end{gathered}\ ] ] since @xmath75 contains the entire singular piece of the tpf , both limits at the r.h.s .
are well defined .
we denote the second limit by @xmath77 .
this quantity is well - defined and regular , and it is computed analytically by summing / integrating over the wkb expressions for the large @xmath78 modes which comprise @xmath79 . the first term in the r.h.s .
is now expressed as a mode sum , and owing to its regularity the limit may trivially be taken by replacing @xmath63 by @xmath16 .
putting it all together one obtains @xmath80 the mode sum should now converge ( even though it is taken in coincidence ) , owing to the subtraction of the singular piece . using this method with splitting in the @xmath35 direction , @xmath0 ( as well as @xmath6 )
was calculated for the schwarzschild case @xcite , and later also for a generic static spherically - symmetric spacetime @xcite . besides the need to numerically compute the various mode functions
@xmath47 , this method also includes a challenging analytical component : it requires a high - order wkb analysis . as was mentioned in the introduction , the presence of a turning point makes this an extremely difficult task . to overcome these difficulties , candelas and howard @xcite and others @xcite actually carried the analysis in the euclidean sector .
the point - splitting method is based on the presumption that the two - point function is well - behaved as long as the two points are separated .
a few caveats are associated to this issue : first , even in flat spacetime , the two - point function diverges when the separation is in a null direction .
let us therefore assume , for the sake of simplicity , that the points are separated in a timelike ( or possibly spacelike ) direction .
the second caveat is that in curved spacetime , assuming that the two points are indeed separated by a timelike geodesic , if the ( proper - time ) distance between @xmath16 and @xmath63 is sufficiently large , there may also be a null geodesic connecting these two points .
for example , consider the schwarzschild spacetime and an approximately - static timelike geodesic @xmath81 located in the asymptotic region very far from the bh .
let @xmath16 be a point on @xmath81 .
there is a null geodesic which emanates from @xmath16 and moves towards the bh ( but with an appropriate miss ) , makes a turn around the bh , then returns to @xmath81 and hits it at a point @xmath82 .
it turns out that @xmath17 develops a singularity at @xmath83 @xcite .
in fact , along the timelike geodesic @xmath81 there is an infinite discrete set of points
@xmath84 which are connected to @xmath16 by null geodesics that make @xmath85 turns around the bh before returning to @xmath81 , and @xmath17 diverges at all points @xmath86 .
note , however , that the problematic points @xmath84 are all located far away from @xmath16 , outside the normal neighborhood ; and for the point - splitting procedure only points @xmath63 in the immediate neighborhood of @xmath16 are relevant .
we shall thus restrict our attention now to points @xmath63 in the close neighborhood of @xmath16 , with a timelike ( or alternatively spacelike ) separation between @xmath16 and @xmath63
. then the tpf should be well - behaved .
naively one might expect that since the tpf is well behaved for a short timelike or spacelike separation , the mode sum associated to it should converge .
it turns out , however , that the situation is more subtle : the involved sum / integral usually fails to converge ( in the literal sense ) .
typically this failure to converge is associated with undamped oscillations .
this phenomenon , the failure of strict convergence , occurs already in flat spacetime .
the ( non- ) convergence situation may depend on the direction of splitting , on the order of operations ( summation over @xmath52 and integration over @xmath19 ) , and on the specific quantity being calculated ( whether it is the tpf , or a component of @xmath13 ) . as an example , consider the calculation of @xmath17 in minkowski spacetime , using standard mode decomposition in spherical harmonics and temporal modes @xmath34 .
the mode functions are given by @xmath87 where @xmath88 is the bessel function of the first kind .
consider now a point @xmath16 located at some @xmath89 , and a point @xmath63 displaced in the @xmath35 direction by an amount @xmath90 the sum over @xmath52 ( for a given @xmath19 ) then converges , and one finds that @xmath91 the mode - sum expression for the tpf then becomes @xmath92 the problem is that this integral does not converge in the usual sense ( even conditionally ) , due to the growing oscillations at large @xmath19 . in this example of pure @xmath35-splitting ,
in which we first sum over @xmath48 , it was the integral over @xmath19 that failed to converge .
in another application of point splitting , in which the splitting is in both @xmath35 and @xmath93 , and in which one first integrates over @xmath19 and sums over @xmath48 afterward , one finds that this time the sum over @xmath20 fails to converge , again due to growing oscillations .
converges for the tpf but only conditionally , and fails to converge even conditionally for @xmath94 , due to growing oscillations . ]
this situation , of non - converging oscillatory integrals ( or sums over @xmath20 ) , is fairly common in various qft calculations .
the common practice ( which may be justified by several arguments ) is to contend that the large-@xmath19 oscillations should be damped in some appropriate manner .
this situation motivates us to introduce the notion of generalized integral , which properly incorporates oscillation damping .
we should emphasize that all the @xmath19 integrals in this paper are in fact generalized integrals .
we shall consider here two specific procedures of oscillation damping , which yield two ( mutually consistent ) definitions of generalized integral .
[ [ abel - summation - integral ] ] abel - summation integral : + + + + + + + + + + + + + + + + + + + + + + + + this is a commonly - used method for giving a meaning for such oscillatory integrals .
the abel - summed integral is defined as @xmath95 provided of course that the integral at the r.h.s .
is well defined for @xmath96 , and the limit @xmath97 exists .
for example , with abel summation , the integral in eq .
( [ eq : a3 ] ) reads @xmath98 yielding the standard expression @xmath99 for the flat - space tpf [ cf .
( [ eq : basic ps - the counter - terms gds ] ) ] .
a crucial property of the abel - summation integral is that it is consistent with the standard integral .
namely , whenever the function @xmath100 is integrable in the strict sense , its abel - summation integral coincides with the standard integral of this function .
[ [ self - cancellation - integral ] ] self - cancellation integral : + + + + + + + + + + + + + + + + + + + + + + + + + + + for the calculation of @xmath0 in a black - hole spacetime we shall have to carry ( generalized ) integral of oscillatory functions @xmath100 that we determine numerically .
in such a case , this procedure of abel summation which must now be implemented numerically is fairly inconvenient .
furthermore , in practice we only determine @xmath100 in a restricted range @xmath101 , which makes the abel - summation integral even harder to implement . and @xmath102 in an appropriate manner .
and it turns out that the convergence of this generalized abel integral with increasing @xmath103 is rather slow .
] we therefore find it much more convenient to use another concept of generalized integral , which we name _ self - cancellation _
( of the oscillations ) .
this type of generalized integral is applicable whenever the oscillations have well - defined frequencies which is indeed the situation in our problem ( see appendix [ sec : appendix - b ] ) . to formulate this concept ,
we first define the integral function @xmath104 the standard integral may then be expressed as @xmath105 ( whenever this limit exists ) . instead
, our self - cancellation generalized integral is defined as @xmath106.\label{eq : a22}\ ] ] here , @xmath107 denotes the `` wave - length '' of the oscillation in @xmath100 ( which is also inherited by @xmath108 ) .
for example , in eq .
( [ eq : a3 ] ) the oscillatory factor is @xmath109 , hence the period of oscillation is @xmath110 .
the idea is simple : the non - oscillating piece of @xmath111 is unaffected by this averaging , but the oscillatory piece will be very effectively annihilated by such averaging with half - wavelength shift . as a simplest example , consider the case @xmath112 .
then @xmath113/l$ ] , and obviously @xmath114 .
clearly @xmath111 fails to have a limit @xmath115 . yet
the self - cancellation integral is perfectly well - defined : the term in squared brackets in eq .
( [ eq : a22 ] ) is simply @xmath116 , entirely independent of @xmath19 .
this example demonstrates the potential of the self - cancellation method to yield extremely fast convergence in @xmath19 .
this last property is important , especially because in an actual calculation we have to determine @xmath100 numerically , and we do so in a restricted range of @xmath19 . for later convenience
we re - formulate this notion of self - cancellation integral as follows : @xmath117,\label{eq : a27}\ ] ] where @xmath118 is the `` self - cancellation operation '' defined by @xmath119\equiv\frac{f(\omega)+f(\omega+\lambda/2)}{2}.\ ] ] in the actual calculation of @xmath1 we shall have to repeat the self - cancellation operation several times ( for several different oscillation frequencies ) . in appendix
[ sec : appendix - a ] we shall introduce the `` multiple self - cancellation operation '' @xmath120 , formed by combining several @xmath118 operations .
it is easy to show that the notion of self - cancellation integral is fully consistent with the standard integral whenever the latter is well defined .
furthermore , the self - cancellation integral is also fully consistent with the abel - summation integral , in the following sense : if the self - cancellation integral converges , then the abel integral converges too , and the two generalized integrals yield the same result .
note , however , that the abel summation method is more general than self - cancellation .
namely , there are functions @xmath100 for which the abel - summed integral is well defined but the self - cancellation integral is not ( that is , the limit in eq .
( [ eq : a22 ] ) is non - existent ) .
nevertheless , for the functions @xmath100 involved in the analysis below , the self - cancellation integral is well - defined and extremely powerful .
as was already noted in the introduction , since our ultimate goal is to address dynamical background metrics as well , we shall carry the analysis directly in the lorentzian sector .
hence , due to the inevitable presence of a turning point ( and also due to the pde nature of the time - dependent mode equation ) , we shall avoid using wkb expansion in our method .
instead we extract the required information about the high - frequency modes of the field directly from the counter - term ( [ eq : basic ps - the counter - terms gds ] ) .
we shall present here the @xmath35-splitting variant , which requires a time - translation killing field . although this variant should be applicable to a rather generic stationary asymptotically - flat background , we shall restrict our attention first to the more specific case of static spherically - symmetric background , for the sake of simplicity . then in sec .
[ sub : our method - non spherical symmetric ] we outline the generalization of the @xmath35-splitting formulation beyond spherical symmetry ( and beyond staticity ) , to a more generic stationary asymptotically - flat background .
we split the points in the @xmath35 direction , namely @xmath121 the tpf then takes the form @xmath122 as was already mentioned above , the sum over @xmath52 converges , and we denote it by @xmath123 .
in fact , the sum of @xmath124 over @xmath10 yields @xmath125 , and therefore @xmath126 this function is to be computed numerically .
the tpf now reduces to @xmath127 ( note that in the coincidence limit @xmath128 this integral would diverge .
however , the oscillatory factor @xmath129 regularizes it . ) equation ( [ eq : basic ps - phi ren limit ] ) now reads @xmath130,\label{eq : a2}\ ] ] where by @xmath131 we refer to @xmath57 with @xmath63 given by eq .
( [ eq : our scheme - the points x , x in the t split ] ) . note that for a given @xmath16 , @xmath64 is uniquely determined by @xmath132 . by a fairly straightforward taylor expansion of the geodesic equation ( and @xmath64 ) in @xmath132
, one finds that @xmath133 in eq .
( [ eq : basic ps - the counter - terms gds ] ) takes the general form @xmath134+d\left(r\right)+o(\varepsilon),\label{eq : our scheme - the counter - terms}\ ] ] where @xmath135 are certain ( real ) functions that depend on the background metric and the parameters of the field
. appears in the brackets because @xmath64 is negative . ]
the explicit form of these functions is not important in the present discussion ( although it is certainly needed for the actual calculation of @xmath0 ) . to proceed
, we now decompose the @xmath132-dependent terms in @xmath133 using the laplace transform .
we have the following identities : @xmath136 @xmath137 inserting these identities into eqs .
( [ eq : a2],[eq : our scheme - the counter - terms ] ) we obtain @xmath138 where @xmath139 and @xmath140 consider now the r.h.s . of eq .
( [ eq : our scheme - phi analytic ] ) .
if the integral of @xmath141 converges ( in either the strict sense or the generalized sense ) , we can interchange the limit and integration , and get rid of the @xmath142 limit altogether .
we should expect the convergence of the integral of @xmath141 , because the singular piece @xmath143 has already been removed from @xmath123 .
in fact , we find that in the schwarzschild case the integral of @xmath141 indeed converges , although only in the generalized sense due to oscillations ( see next section ) .
we were unable to _ prove _ the ( even generalized ) convergence of the integral of @xmath141 , but nevertheless since this convergence is naturally expected , and since the schwarzschild example confirms this expectation , we shall hereafter _ assume _ that this integral indeed converges in the generalized sense . will demonstrate this non - convergence right away . ]
we therefore write our final result as @xmath144 where , recall , the integral over @xmath19 is a generalized one ( as defined in sec .
the implementation of this generalized integral is demonstrated in sec .
[ sec : calc in schwarzschild ] for the schwarzschild case .
as was already pointed out above , our @xmath35-splitting method does not require the background metric to be spherically symmetric or even static : it should be applicable to a generic stationary asymptotically - flat background . here
we outline this extension . on account of asymptotic flatness
, we now choose our coordinates ( @xmath145 such that the large@xmath146 asymptotic metric still takes its standard minkowski form @xmath147 .
takes the standard weak - field form @xmath148 , where here @xmath149 denotes the system s asymptotic mass . ]
due to lack of spherical symmetry , the expression ( [ eq : basic ps - expression for f ] ) for the field modes is now replace by @xmath150 where @xmath151 is a set of solutions to the ( @xmath19-dependent ) spatial part of the field equation , which is now a pde ( in @xmath152 ) rather than ode .
these solutions are required to be regular everywhere , and to satisfy the large-@xmath21 boundary condition @xmath153 where @xmath154 denotes the reflected field which is @xmath155 times some function of @xmath93 and @xmath156 .
is to be regarded as the standard function of @xmath21 given in eq .
( [ eq : r - star ] ) . ]
once the modes @xmath42 were defined , the expression ( [ eq : basic ps - field decomposition to a , ad ] ) for the field operator is unchanged .
the tpf now takes the form @xmath157 it is important to recall that even though the metric is not spherically symmetric , the sums over @xmath48 should still converge ( for a given @xmath19 ) : due to asymptotic flatness , at large @xmath21 there will be a centrifugal barrier @xmath158 in the effective potential , just like in minkowski , preventing the penetration of modes with too large @xmath20 .
thus we can again define @xmath159 and the tpf still takes the form ( [ eq : our scheme - phi2 splitted ] ) [ although now with @xmath160 .
the calculation now proceeds just as in the spherically symmetric case except that all the quantities in sec .
( [ sub : our method - spherical symmetric ] ) that were dependent on @xmath21 only , now depend on @xmath93 and @xmath156 as well .
the final result is @xmath161 with @xmath162 and @xmath163 it should be pointed out that although the choice of the coordinates @xmath152 is certainly non - unique ( due to lack of spherical symmetry ) , the resultant mode decomposition is still unique .
this is because of asymptotic flatness ( and the standard weak - field metric that we require our @xmath164 coordinates to satisfy at large @xmath21 ) , and because the mode functions @xmath151 are defined through their asymptotic form at @xmath165 .
the unique mode decomposition in turn leads to a unique fock space associated to it , and to a well - defined vacuum state .
note , however , that in the above formulation we implicitly assumed a stationary background metric with no past horizon ( e.g. a spinning star ) , hence it was sufficient to construct the `` in '' modes . in the case of eternal non - spherical bh the situation becomes more subtle , because now we also need to specify the `` up '' modes .
here asymptotic flatness will not be of much help , because these modes are to be defined by boundary conditions at the past horizon . with the lack of spherical symmetry and staticity
, one still needs to figure out how to make a unique mode decomposition and to obtain a `` natural '' vacuum state . coordinates , and with a corresponding arbitrary construction of the set of `` up '' modes
, we have no guarantee that the resultant `` vacuum state '' would at all be a well - defined hadamard state . in this regard
, we should mention the observation that there is no hadamard state that respects the symmetries of kerr spacetime and regular everywhere .
@xcite ] the implementation of this method in the non - spherical case is of course technically more challenging ( even for a non - bh background ) , because now the mode functions @xmath166 which comprise @xmath167 are to be obtained by numerically solving pdes rather than just odes .
using the method presented in the last section , we compute @xmath0 in the exterior region of schwarzschild spacetime , in the boulware vacuum state , for a minimally coupled massless scalar field .
the schwarzschild metric is @xmath168 in this metric the quantum field @xmath169 can be expanded in the form presented in eqs .
( [ eq : basic ps - field decomposition to a , ad])-([eq : basic ps - expression for f ] ) , where @xmath47 is conveniently recast as @xmath170 and @xmath171 satisfies the radial equation @xmath172\psi_{\omega l}\left(r\right).\label{eq : schwarzschild calc - radial equation}\ ] ] henceforth , @xmath173 will denote the tortoise coordinate given by @xmath174 and @xmath175\ ] ] is the effective potential .
the general solution of eq .
( [ eq : schwarzschild calc - radial equation ] ) ( for given @xmath78 ) is spanned by two basic solutions , that we denote @xmath176 and @xmath177 .
the boundary conditions for these two basic solutions are taken to be @xmath178{cc } \tau_{\omega l}\ , e^{-i\omega r _ { * } } , & r_{*}\to-\infty\\ e^{-i\omega r_{*}}+\rho_{\omega l}\ , e^{i\omega r_{*}},\,\,\,\,\,\ , & r_{*}\to\infty \end{array}\right.\nonumber \\
\psi_{\omega l}^{up}\left(r\right)=\left\ { \begin{array}[t]{cc } e^{i\omega r_{*}}+\tilde{\rho}_{\omega l}\ , e^{-i\omega r_{*}},\,\,\,\,\,\ , & r_{*}\to-\infty\\ \tau_{\omega l}\ , e^{i\omega r _ { * } } , & r_{*}\to\infty \end{array}\right.\label{eq : schwarzschild calc - basic solutions bounadry conditions}\end{gathered}\ ] ] where @xmath179 , @xmath180 , and @xmath181 represent the transmission and reflection amplitudes .
these two basic solutions are properly normalized and mutually orthogonal .
the presence of two independent modes for each @xmath41 ( as opposed to a single such mode in e.g. minkowski ) requires a slight modification of the formalism above : we now have _ two _ sets of annihilation operators , @xmath182 and @xmath183as well as their conjugate operators @xmath184 , @xmath185 .
correspondingly , in eq .
( [ eq : basic ps - field decomposition to a , ad ] ) , in addition to the integral over @xmath19 and the sum over @xmath20 and @xmath10 , we also have to sum over the separate contributions of the `` in '' and `` up '' modes .
we shall consider here the b__oulware vacuum _ _ , namely the quantum state annihilated by all the operators @xmath183 as well as @xmath182 .
revisiting the analysis of the previous sections , one finds that everything remains intact , except that all the equations that involve summation over @xmath52 should now also include a summation over the `` in '' and `` up '' contributions . correspondingly , eq . ( [ eq : our scheme - f defintion ] )
is now replaced by @xmath186 but otherwise the results of the previous section , and in particular eqs .
( [ eq : a5],[eq : a6],[eq : our scheme - phi analytic - final ] ) , are unaffected .
since the schwarzschild metric is a vacuum solution , and since we are dealing with a massless field , the counter - term ( [ eq : basic ps - the counter - terms gds ] ) now reduces to @xmath187 @xmath64 is related to @xmath188 through the proper - time @xmath66 of the short geodesic connecting @xmath16 to @xmath63 , via @xmath65 . by conducting a second - order expansion of the geodesic equation we obtain @xmath189 and correspondingly @xmath190 comparing this to eq .
( [ eq : our scheme - the counter - terms ] ) we find that @xmath191 therefore in the schwarzschild case eqs .
( [ eq : a5],[eq : a6],[eq : our scheme - phi analytic - final ] ) reduce to @xmath192 with @xmath193 summarizing the analytical part of the calculation , @xmath14 is given by eq .
( [ eq : a13 ] ) along with eqs .
( [ eq : a12 ] ) , ( [ eq : a10 ] ) and ( [ eq : schwarzschild calc - f omega ] ) .
note that basically this expression for schwarzschild was already obtained by candelas , @xcite , but here we also complete the calculation by implementing the numerical part as well ( and by doing so , we encounter the oscillations problem and address it ) .
we have numerically solved the radial equation for @xmath176 and @xmath177 , using the standard mathematica numerical ode solver , in the domain @xmath194 , at a set of @xmath19 values with a uniform separation @xmath195 .
hereafter , we use units in which the bh mass is @xmath196 ( in addition to @xmath197 ) , hence @xmath19 as well as @xmath21 are dimensionless . for a given @xmath19 ,
the contribution of the different @xmath20 modes to @xmath123 starts to decay exponentially fast beyond a certain @xmath20 value , typically of order @xmath198 .
( this decay may be interpreted as tunneling into the potential barrier . ) correspondingly , for each @xmath19 value , we truncate the sum ( [ eq : schwarzschild calc - f omega ] ) at an @xmath20 value where the contribution becomes negligible ( @xmath199 . then from @xmath123 we construct the regularized function @xmath141 according to eq .
( [ eq : a12 ] ) .
figure [ fig:1a ] displays @xmath123 for @xmath200 , which we choose here as our representative @xmath21 value .
then fig . [ fig:1b ] displays the regularized function @xmath141 , obtained from @xmath123 by removing the linear piece @xmath201 .
clearly , the linear divergence has been removed , but there remain oscillations that grow as @xmath202 , which we address below .
the origin of these oscillations ( the aforementioned connecting null geodesics ) , and the determination of their frequencies , are discussed in appendix [ sec : appendix - b ] .
to calculate @xmath14 we need the generalized integral of @xmath141 from @xmath203 to infinity , see eq .
( [ eq : a13 ] ) . to this end
we define the integral function @xmath204 the strict integral of @xmath205 would of course correspond to the limit @xmath115 of @xmath111 .
figure [ fig:2 ] displays @xmath111 , and makes it clear that this limit does not exist , due to the growing oscillations ( which were inherited directly from @xmath205 ) .
we therefore have to resort to the generalized integral instead , as discussed in sec .
[ sub : remarks - about - the - tpf ] .
the abel - summation integral ( [ eq : a72 ] ) is well defined in this case .
however , since we know the precise frequency of oscillations , it is much more convenient and more efficient to employ the self - cancellation integral ( which is fully consistent with the abel integral ) . as it turns out ,
there are multiple oscillation frequencies in @xmath206 ( see appendix [ sec : appendix - b ] ) , and we cancel each of the four dominant ones by a forth - order self - cancellation operation . adapting the notation of appendix [ sec : appendix - a ] to the present specific context ( integration of @xmath205 ) , the desired generalized integral is @xmath207 where @xmath208 $ ] , @xmath111 is given in eq .
( [ eq : a39 ] ) , and @xmath120 denotes the multiple self - cancellation operation as generally defined in eq .
( [ eq : a32 ] ) and detailed in eq . ( [ eq : a49 ] ) .
( that is , the integral of @xmath205 from zero to @xmath19 ) .
[ fig:2 ] ] ( obtained from @xmath108 after self - cancellation of the oscillations ) .
notice the quick convergence .
[ fig:3 ] ] figure [ fig:3 ] displays the function @xmath209 .
it is remarkable that after the oscillations have been removed the function converges very quickly .
this allows a fairly precise determination of the limit @xmath115 of this function , which constitutes the generalized integral in eq .
( [ eq : a34 ] ) .
we then substitute this integral , as well as @xmath210 of eq .
( [ eq : a10 ] ) , in eq .
( [ eq : a13 ] ) . in figs .
[ fig:4a ] and [ fig:4b ] we present our results for @xmath0 as a function of @xmath21 , and compare them to results obtained previously by anderson @xcite using a very different method ( analytic extension to the euclidean sector ) .
the differences are typically of order a few parts in @xmath211 , consistent with the estimated numerical errors .
we presented here a new approach for implementing point - splitting regularization numerically , for the computation of @xmath0 and @xmath6 in various asymptotically - flat spacetimes .
our main motivation in developing this approach is to allow systematic investigation of self - consistent semiclassical evaporation of bhs .
this would require the calculation of @xmath6 in a time - dependent bh background .
so far we developed two variants of our basic approach ( both for a quantum scalar field in asymptotically - flat background ) : ( i ) @xmath35-splitting , applicable to stationary spacetimes , and ( ii ) angular splitting , applicable to spherically - symmetric spacetimes .
in this first paper we focused on the simplest of the two , the @xmath35-splitting variant . in presenting this method , we restricted our attention to static spherically - symmetric backgrounds , and to @xmath0 rather than @xmath6 , for the sake of simplicity .
but we also described the extension of the method to more generic , non - spherical , stationary asymptotically - flat backgrounds .
this extension suffices for the non - bh case , e.g. for a rotating star . in the case of a stationary bh which is not a spherically - symmetric static one
, our method still needs a completion : we still have to formulate the construction of an appropriate `` up '' state at the past horizon , which would constitute a physically - meaningful ( and properly hadamard ) `` vacuum '' state .
we then implemented the @xmath35-splitting variant to the specific case of boulware state in schwarzschild spacetime ( for a minimally - coupled massless scalar field ) .
the analytical part of the regularization procedure coincides in this case with the one developed by candelas @xcite .
however , here we also implemented the numerical part , which involves the numerical solution of the radial equation for the various modes , and the summation / integration over the mode contributions .
doing so , we found that the regularized function @xmath141 which was naively expected to be well - behaved at large @xmath19 actually suffers from growing oscillations , which make the @xmath19-integral non - convergent .
as it turns out , this phenomenon has little to do with the short - distance behavior of the tpf in the coincidence limit @xmath212 . instead ,
the oscillations originate from divergences of the tpf at _ remote _ points @xmath63 which are connected to @xmath16 by null geodesics .
we used the notion of generalized integral and particularly the pragmatic method of self - cancellation integral in order to handle these oscillations and to carry the desired integration over @xmath19 .
doing so , we found excellent agreement with previous results obtained by a different method , the euclidean extension @xcite .
putting aside for a moment the ultimate goal of analyzing the time - dependent evaporation process , we wish to emphasize that even the simplest version of the method , the @xmath35-splitting variant presented here , makes it possible to do ps regularization in non - spherical stationary spacetimes , e.g. that of a strong - field spinning star ( although in the generic stationary bh case our construction still needs a completion ) .
this was not possible so far , due to the difficulties in conducting high - order wkb expansion in such spacetimes .
however , to achieve our primary goal of self - consistent semiclassical evaporation , we must deal with time - dependent backgrounds . to this end
we shall need the angular - splitting variant , which is slightly more complicated than @xmath35-splitting .
we shall describe this variant in a separate paper .
we already applied the angular - splitting method to @xmath0 and also to @xmath6 in boulware state in schwarzschild , and again we found very good agreement with previous calculations @xcite .
these results will be presented elsewhere @xcite .
we are grateful to paul anderson and bob wald for many interesting and fruitful discussions .
in addition ao thanks both paul anderson and bob wald for kind hospitality in their institutions , where many of these discussions took place . in particular
the understanding of the nature of the oscillations , and their relation to connecting null geodesics , emerged in a discussion with bob wald during a visit at the university of chicago .
we also wish to thank paul anderson for sharing his unpublished data with us .
for a given integrand function @xmath100 , the self - cancellation integral was defined in sec .
2.2 as @xmath117,\label{eq : a27 - 1}\ ] ] where @xmath213 @xmath118 is defined by @xmath119\equiv\frac{f(\omega)+f(\omega+\lambda/2)}{2},\ ] ] and @xmath107 is the oscillation s wavelength . consider next the situation in which @xmath100 contains oscillations in two different frequencies ( say , with @xmath19-independent amplitudes ) .
then @xmath111 will take the asymptotic form @xmath214 with two different wave - lengths @xmath215 .
if we apply a self - cancellation with respect to ( say ) @xmath216 , then we will be left with a function @xmath217 $ ] in which the @xmath218 is of course unaffected , and the term @xmath219 has entirely been annihilated .
the other oscillatory term @xmath220 is still present in @xmath217 $ ] , although its amplitude will decrease by the factor @xmath221 $ ] , as one can easily verify .
the self - cancellation integral ( associated to @xmath222 ) is thus still non - convergent . a second self - cancellation , this time with respect to @xmath223 ,
will yield a convergent self - cancellation integral , which may be formulated as @xmath224.\ ] ] the order of the two operations @xmath222 and @xmath225 is unimportant , as one can easily verify . with both orderings , the double self - cancellation integral yields the `` @xmath218 '' in eq .
( [ eq : a51 ] ) . , but with a `` mistaken '' wavelength parameter @xmath226 that differs from the true oscillation wavelength @xmath107 , does not `` spoil '' the generalized integral in any way : the mismatch in @xmath107 does not lead to any new oscillatory terms , nor does it modify the value of the generalized integral . the effect of the mismatch in @xmath107 is merely to limit the efficiency of the self - cancellation operation : it decreases the oscillation s amplitude by the factor @xmath227 $ ] instead of fully annihilating it . ]
this process is straightforwardly generalized to functions @xmath111 with any number @xmath228 of oscillation frequencies that need to be annihilated .
to handle this situation , we define the multiple self - cancellation operator @xmath229 then the multiple self - cancellation integral may be expressed as @xmath230 where @xmath231.\label{eq : a62}\ ] ] as was demonstrated in sec .
[ sub : remarks - about - the - tpf ] , if @xmath111 contains an oscillation with fixed amplitude , then the @xmath118 operation fully nullifies this oscillation .
( this was demonstrated there for the case @xmath112 , but it equally applies to the more general cases @xmath232 . )
however , if the oscillation s amplitude varies with @xmath19 , the cancellation is not complete . consider the case @xmath233 where @xmath234 is some slowly - varying function namely , a function whose typical length of variation @xmath235 becomes @xmath236 at large @xmath19 . to be more specific ,
let us further assume that @xmath235 diverges as @xmath115 . , because in this class @xmath237 .
( but it also includes much more general classes of functions , e.g. @xmath238@xmath239 for any @xmath240 . ) ] then one can easily show that @xmath241\simeq-\frac{\lambda}{2}\,\frac{dg}{d\omega}\ , e^{i\omega l}\ ] ] at large @xmath19 .
in particular , if @xmath242 then @xmath241\propto p\,\omega^{p-1}e^{i\omega l}.\ ] ] consider now the case @xmath243 for some @xmath244 . then @xmath245 $ ]
still does not converge as @xmath115 .
nevertheless , @xmath246\right)$ ] is @xmath247 and hence it converges .
this illustrates that in certain circumstances one may want to apply the same self - cancellation operation several times , say @xmath85 times , an operation which we shall denote as @xmath248 .
quite generally , in the case @xmath243 one finds that with @xmath249 , in which @xmath250 $ ] strictly vanishes . ]
@xmath251\propto\omega^{p - n}e^{i\omega l}.\ ] ] therefore , for this class of @xmath111 functions , the convergence criterion for the ( single - frequency ) multiple self - cancellation integral is simple : it converges if and only if @xmath249 . in the case
@xmath252 a single self - cancellation operation would be sufficient for achieving convergence .
however , one would then be left with a slowly - decaying amplitude @xmath253 . to speed the large-@xmath19 convergence ,
one may repeat @xmath118 several times .
this is in fact the situation in our specific problem , where @xmath254 and hence @xmath255 : the single self - cancellation integral converges , but rather slowly , with amplitude @xmath256 , and we therefor repeat @xmath118 several times to achieve faster convergence ( which is crucial for numerical implementation ) .
if there are several different oscillation frequencies that need to be annihilated , one can freely choose how many times to repeat the @xmath118 operation for each frequency .
we therefore generalize the above expression ( [ eq : a30 ] ) for the multiple self - cancellation operator : @xmath257 the multiple self - cancellation integral ( @xmath258 ) is thus obtained by using this @xmath120 operator in eqs .
( [ eq : a31],[eq : a62 ] ) .
finally we point out that this multiple self - cancellation integral is fully consistent with the abel - summation integral in the same sense discussed in sec.[sub : remarks - about - the - tpf ] ( concerning the single self - cancellation integral ) .
as can be clearly seen in fig .
[ fig:1b ] , the function @xmath141 admits growing oscillations in @xmath19 .
these oscillations are of course inherited directly from @xmath123 , as can be seen in fig .
[ fig:1a ] ( although in the latter the oscillations are overshadowed by the linearly - growing term ) . in this appendix
we shall discuss the origin of these oscillations , their frequencies , and their amplitudes .
then at the end we shall specify the self - cancellation operator that we apply in order to practically remove these oscillations .
owing to spherical symmetry and staticity , in @xmath35-splitting the tpf may depend only on @xmath21 and on @xmath259 .
let us then introduce the abbreviated notation for the tpf @xmath260 equation ( [ eq : our scheme - phi2 splitted ] ) actually tells us that @xmath123 is the fourier transform of @xmath261 . as a function that vanishes for all @xmath262 .
note also that in the present context one should not think of @xmath132 as a small parameter : instead , it is allowed to take all real values . ]
if @xmath261 were regular and smooth for all @xmath132 , then its fourier transform @xmath123 would decay quickly at large @xmath19 , faster than any power of the latter .
the undamped oscillations in @xmath123 must therefore indicate some irregularity in @xmath261 . we still need to understand the nature of this singularity , and its location on the @xmath132 axis .
the singularity of the tpf at @xmath128 indeed leads to a divergence of @xmath123 at large @xmath19 ( this is the linearly - growing term shown in fig .
[ fig:1a ] ) but not to oscillations ; and this linear singularity is no longer present in @xmath141 .
the oscillations in fig .
[ fig:1b ] must then indicate another singularity , located at some point @xmath263 . to illustrate this ,
consider for example the delta - function case : the fourier transform of @xmath264 is @xmath265 .
it is oscillatory if and only if @xmath266 . the frequency of the @xmath19-oscillations is just @xmath267 namely , it directly tells us the distance of the singularity from the point @xmath268 .
one more example is the singularity @xmath269 , whose transform is @xmath270 . but
this phenomenon is of course more general : if a certain function @xmath271 admits a fourier transform @xmath272 , then the transform of @xmath273 is @xmath274 .
the oscillations seen in e.g. fig .
[ fig:1b ] have a certain `` @xmath19-wavelength '' @xmath275 ( in units in which the bh mass is @xmath196 ) .
they must therefore correspond to a singularity in the tpf , located at a distance @xmath276 from point @xmath16 .
what is the nature of this non - local singularity of the tpf ?
as already pointed out in sec .
[ sub : remarks - about - the - tpf ] , the function @xmath17 admits a singularity whenever a null geodesic exists which connects @xmath16 and @xmath63 . in the present context of schwarzschild background and @xmath35-splitting
, we are dealing here with a null geodesic which emanates from a certain spatial point @xmath277 , makes a round trip around the bh , and then returns to that same spatial point , but obviously with a certain delay in @xmath35 which should correspond to the shift parameter @xmath267 .
in fact there is not only one but an _ infinite , _ discrete set of such connecting null geodesics ( for each @xmath21 ) .
this is because a null geodesic emanating from a spatial point @xmath277 can make any integer number of revolutions around the bh before returning to that spatial point .
therefore , the tpf will actually admit an infinite number of singular points at a discrete set of values @xmath278 , one for each integer @xmath85 . correspondingly , there will be a discrete spectrum of oscillation modes in @xmath141 , with ( @xmath21-dependent ) @xmath19-frequencies @xmath279 and corresponding wavelengths @xmath280 .
we point out , however , that the dominant mode is always @xmath281 , and the oscillation s amplitude quickly decays with @xmath85 ( see next subsection ) . to perform the self - cancellation of oscillations
, we shall need to know the spectrum of frequencies @xmath279 , at any desired @xmath21 value .
this requires integration of the null geodesic equation , to find the connecting null geodesics .
for @xmath282 the situation is especially simple , because in that case the connecting null geodesic is circular .
one then finds that @xmath283 . at
other @xmath21 values the spectrum becomes more complicated , and needs to be calculated numerically .
the null orbits in schwarzschild are characterized by a single constant of motion , namely the angular momentum per unit energy ( or the `` impact parameter '' ) .
the calculation of @xmath279 involves ( i ) numerical integration of the null geodesic equation ( for prescribed values of that constant of motion ) , and ( ii ) using the newton - raphson method for adjusting this constant , in order to find the connecting geodesics , which return to the original spatial point after @xmath85 rounds .
the required parameters @xmath279 are the @xmath35-duration of these connecting null geodesics .
overall , this is an easy numerical procedure . in our representative case @xmath284
, the first few oscillation frequencies are found to be @xmath285 in fig .
[ fig:1b ] we predominantly see the basic oscillation @xmath281 , and the frequency agrees very well with this value of @xmath286 .
we can also notice the residual effect of the @xmath287 oscillation , which causes the small distortion ( i.e. small deviation from the smooth @xmath288 envelope ) in the pattern of peaks of @xmath141 , seen in fig .
[ fig:1b ] .
after the basic oscillation @xmath281 is removed by self - cancellation , the next one ( @xmath287 ) dominates and becomes very clear , and again , its frequency is found to agree very well with the above value of @xmath289 . by this procedure
it is possible to expose a few more modes , and to confirm their agreement with the above @xmath279 values , that were obtained from the connecting null geodesics .
as was already mentioned above , for @xmath282 the frequencies @xmath279 form an arithmetic sequence with a common difference @xmath290 . for other @xmath21 values
this is no longer the case .
still , the difference @xmath291 quickly approach the standard spacing @xmath292 .
for example , from the four @xmath279 values specified above for the @xmath284 case , one sees that @xmath293 , and then @xmath294 . this is simply because for any @xmath21 , a connecting null geodesic with large @xmath85 makes most of the revolutions around the bh along an orbit very close to the circle @xmath282 .
the numerical data indicate that the oscillation s amplitude grows as @xmath202 .
this in turn implies that the divergence of the tpf at @xmath295 should be @xmath296 .
we shall not address this issue here in detail , but we point out that this @xmath297 power is just what one would expect from simple arguments . to this end
one has to recall that since the background is spherically symmetric , and since the splitting is only in the @xmath35 direction ( implying that @xmath298 and @xmath299 ) , whenever @xmath63 is located on a null geodesic emanating from @xmath16 , it is placed _
exactly _ at a _ caustic point _ of that null geodesic .
qualitative arguments suggest that on crossing such a caustic point , the tpf should indeed diverge as @xmath300 .
but the discussion of this issue is far beyond our present scope .
the @xmath301 oscillations , too , grow as @xmath202 at large @xmath19 ( the numerics confirms this , at least for the first few @xmath85 values ) .
this is for the same reason as that described above for @xmath281 .
we thus express the large-@xmath19 amplitudes of the various @xmath279 modes as @xmath302 , where @xmath303 is a set of ( @xmath21-dependent ) amplitude parameters .
it is important to explore how @xmath303 behaves with increasing @xmath85 , in order to control the possible effect of the infinite number of oscillating modes .
the simplest case to analyze is again @xmath282 , because the connecting null geodesic is then the circular geodesic at @xmath282 .
simple analytical considerations suggest that in this case @xmath303 should form an almost - exact geometric sequence with @xmath304 .
geodesic , suggests that in this case @xmath303 should be exactly proportional to @xmath305^{-1/2}$ ] .
this expression deviates from the geometric sequence @xmath306 by a tiny relative amount @xmath307 . this deviation is smaller than one part in @xmath308 even for @xmath281 , too small to be detected by our numerics , but nevertheless our numerical results are fully consistent with that expression .
] our numerical results for @xmath309 allows reliable evaluation of the first three amplitudes , and the calculated ratios @xmath310 and @xmath311 agree very well with @xmath312 , to about one part in @xmath211 .
( for the @xmath313 modes the oscillations are too weak to reliably measure their @xmath303 . ) for other @xmath21 values the situation is more complicated , and we do not expect to find such a well - approximated geometric sequence ; yet , we still expect that as @xmath85 increases , @xmath314 should quickly approach the above `` canonical '' value @xmath312 .
the reason is that , for large @xmath85 , the connecting null geodesic makes most of the @xmath85 revolutions around the bh along an orbit very close to the circle @xmath282 .
hence , the decrease in the van - vleck determinant at each revolution is approximately the same as in the analogous @xmath282 case ( an approximation that ever improves with increasing @xmath85 ) . at
very small @xmath85 , however , the ratio between two successive amplitudes may slightly differ from @xmath312 .
we numerically find that the ratio @xmath310 ranges from @xmath315 at @xmath282 to @xmath316 at @xmath317 .
overall , at least in the range @xmath318 that we have numerically explored , the numerical data as well as the theoretical considerations are all consistent with an almost - geometric sequence ( even for small @xmath85 ) , with @xmath314 ranging between @xmath319 and @xmath320 . in turn
this implies that for practical computation of the generalized integral of @xmath141 , we shall have to cancel the first few @xmath85 modes , but the contribution of large-@xmath85 modes may be neglected .
we self - cancel the first four frequencies @xmath321 .
the higher modes @xmath322 are too weak to notice . for _ any _ @xmath85 , because after all we only integrate along a finite interval of @xmath19 [ which is in turn allowed due to the fast convergence of @xmath209 ] , and because of the fast decay of @xmath303 with @xmath85 , the large-@xmath85 oscillation terms do not have any noticeable effect on the integral . ]
the values of the frequencies @xmath279 are numerically obtained from the connecting null geodesics , for any desired @xmath21 , as explained above .
the corresponding wavelengths are then given by @xmath280 .
for each frequency , we apply a fourth - order self - cancellation .
thus , in the terminology of eq .
( [ eq : a32 ] ) , our actual multiple self - cancellation operator is @xmath324 | computation of the renormalized stress - energy tensor is the most serious obstacle in studying the dynamical , self - consistent , semiclassical evaporation of a black hole in 4d .
the difficulty arises from the delicate regularization procedure for the stress - energy tensor , combined with the fact that in practice the modes of the field need be computed numerically .
we have developed a new method for numerical implementation of the point - splitting regularization in 4d , applicable to the renormalized stress - energy tensor as well as to @xmath0 , namely the renormalized @xmath1 .
so far we have formulated two variants of this method : _ t - splitting _ ( aimed for stationary backgrounds ) and _ angular splitting _ ( for spherically - symmetric backgrounds ) . in this paper
we introduce our basic approach , and then focus on the t - splitting variant , which is the simplest of the two ( deferring the angular - splitting variant to a forthcoming paper ) .
we then use this variant , as a first stage , to calculate @xmath0 in schwarzschild spacetime , for a massless scalar field in the boulware state .
we compare our results to previous ones , obtained by a different method , and find full agreement .
we discuss how this approach can be applied ( using the angular - splitting variant ) to analyze the dynamical self - consistent evaporation of black holes . |
while passing through a narrow temperature range around @xmath7 , da ( hydrogen ) white dwarfs exhibit pulsations with periods between 2 and 20 minutes .
these oscillations are identified with g - modes having radial orders @xmath8 , and spherical harmonic degrees @xmath9 .
mode overstability occurs when driving in the thin and fast reacting surface convection zone exceeds damping in the underlying radiative interior ( brickhill @xcite , gautschy et al .
@xcite , goldreich & wu @xcite , hereafter paper i ) .
the flux perturbation entering the convection zone from the underlying radiative interior is in phase with the pressure perturbation .
because the convection is fast and efficient , the specific entropy perturbation is nearly depth - independent throughout the bulk of the convection zone .
consequently , the magnitude of the flux perturbation decreases outward .
the phase relation between the pressure and flux perturbations then results in convective driving .
gravity - modes have simple structures in their upper evanescent regions .
this permits an analytic derivation of a stability criterion based on the quasiadiabatic approximation for modes whose evanescent regions extend well below the base of the convection zone .
overstability is predicted provided @xmath10 , where @xmath4 is a few times the thermal time scale at the bottom of the convection zone .
applied blindly , this criterion predicts overstability in cool zz cetis for modes whose periods far exceed those observed .
our numerical calculations enable us to drop both the quasiadiabatic approximation and the restriction that the evanescent region extend well below the convection zone .
they also allow us to incorporate the effect of turbulent viscosity as discussed in brickhill ( @xcite ) and goldreich & wu ( @xcite , hereafter paper ii ) .
this paper is the third in a series dedicated to the overstability of gravity - modes in zz cetis .
it is largely numerical , and presents results of nonadiabatic calculations of eigenvalues and eigenfunctions .
nonadiabaticity is important for modes which have @xmath11 , since they have @xmath12 in their driving and damping regions .
this category includes both overstable modes with @xmath13 , which are easily detectable by photometric measurements , and damped modes with @xmath14 , whose excitation by parametric instability acts to limit the amplitudes of overstable modes .
it is imperative that our calculations accurately represent them .
the organization of this paper is as follows . in [ sec : nonad - eqsetup ] , we collect the equations which govern linear , nonadiabatic oscillations , establish appropriate boundary conditions , and describe the method we use to construct envelope models for zz cetis . in [ sec : nonad - important ] we assess the magnitude of nonadiabatic effects in the evanescent and propagating regions of g - modes .
numerical techniques employed to solve the eigenvalue problem are elaborated in [ sec : nonad - numer ] .
the impact of nonadiabaticity and turbulent viscosity on g - mode eigenvalues and eigenfunctions is described in [ sec : nonad - result ] .
we then discuss in [ sec : nonad - maximum ] several possibilities to reconcile the nonadiabatically calculated spectra of overstable modes with those obesrved .
a short summary follows in [ sec : nonad - sum ] .
the appendix is devoted to analysis of a toy model which elucidates the asymptotic damping rates of high order modes .
symbols used in this paper are defined in table 1 .
their usage is consistent with that in previous papers of this series .
cll[t ] @xmath15 & stellar radius @xmath16 & surface gravity @xmath17 & depth below photosphere @xmath18 & depth at bottom of convection zone @xmath19 & depth at top of mode s cavity , @xmath20 @xmath3 & radian mode frequency , @xmath21@xmath22 & radial order of mode @xmath23 & angular degree of mode @xmath24 & horizontal wave vector , @xmath25 @xmath26 & vertical wave vector @xmath27 & mass density , @xmath28 mass density at @xmath29 @xmath30 & pressure @xmath31 & temperature @xmath32 & specific entropy in units of @xmath33 @xmath34 & energy flux @xmath35 & adiabatic sound speed , @xmath36 @xmath37 & @xmath38 @xmath39 & denotes lagrangian perturbation @xmath40 & horizontal component of displacement vector @xmath41 & vertical component of displacement vector @xmath42 & horizontal component of velocity vector @xmath43 & coefficient measuring convective inefficiency @xmath44 & convective velocity , @xmath45 @xmath46 & response time for convection , @xmath47 @xmath48 & dimensionless constants approximately @xmath49 & @xmath50 for zz cetis @xmath51 & thermal constant at depth z , @xmath52 in the convection zone @xmath53 & thermal time at @xmath54 @xmath55 & unconventional thermal time constant at @xmath18 , @xmath56 at @xmath18 .
@xmath4 & time constant of low pass filter for convection zone , @xmath57 @xmath58 & frequency@xmath59 , @xmath60 & opacity derivatives , @xmath61 , @xmath62 @xmath63 & dimensionless heat capacity , @xmath64 @xmath65 & temperature gradient , @xmath66 @xmath67 & adiabatic temperature gradient , @xmath68
the linearized equations describing nonadiabatic pulsations read @xmath69 for plane - parallel geometry with the cowling approximation , and an assumed time - dependence of @xmath70 .
equations - express the conservation of mass , horizontal and vertical momentum , as well as energy .
equation describes the equation of radiative diffusion .
closure of this system of equations requires constitutive relations for the equation of state and the opacity .
setting @xmath71 reduces the above equations to the adiabatic ones studied in paper i. we choose @xmath72 and @xmath73 as our independent thermodynamic variables and set @xmath74 then the linear perturbation equations may be written as four first - order differential equations for the four dependent variables @xmath75 , @xmath76 , @xmath77 , and @xmath73 ; @xmath78\left({{\delta p}\over p}\right ) -x + \left({p\over { g\rho}}\right ) { { ( gk_h)^2-\omega^4}\over{g \omega^2 } } \rho_s \delta s , \label{eq : nonad - dppds } \\ { d\over d\ln p } \left({{\delta f}\over
f}\right ) & = & { -i\omega k_b t p\over g m_p f } \delta s , \label{eq : nonad - dsdff } \\
{ 1\over c_p \nabla}{d\delta s\over d\ln p } & = & - \left [ ( 4-\kappa_t ) \nabla_{ad } - { { \kappa_\rho}\over \gamma_1 } -\left ( 1- { { \omega^2 p } \over{g^2\rho } } \right ) { { ( gk_h)^2}\over{(gk_h)^2-\omega^4 } } + { 1\over \nabla } { { d \nabla_{ad}}\over{d\ln p}}\right ] \left({{\delta p}\over p}\right ) \nonumber \\ & & \hskip-0.1 in -\left [ { { \nabla_{ad}}\over{\nabla } } - { { ( gk_h)^2}\over{(gk_h)^2-\omega^4 } } \right ] x + \left({{\delta f}\over f}\right ) - \left [ { { ( 4-\kappa_t)}\over{c_p } } -\kappa_\rho \rho_s + { 1\over\nabla } { { d}\over { d\ln p}}\left({1\over { c_p}}\right ) \right ] \delta s. \label{eq : nonad - dffall } \end{aligned}\ ] ] the perturbation equations are to be solved as an eigenvalue problem . since equation includes a factor @xmath79 , all four dependent variables and the eigenfrequency are _
complex_. we write each complex variable in the form @xmath80 the physical perturbation is given by @xmath81 . inside the convection zone ,
g - mode perturbations are constrained by the rapid response of convection to the instantaneous pulsational state ( brickhill @xcite & @xcite , papers i & ii ) .
rapid momentum diffusion enforces |x|=|ddp(pp)||pp| .
[ eq : nonad - xconst ] and fast entropy mixing ensures that |dsdp||s| , [ eq : nonad - dscvz ] except in the thin superadiabatic layer .
the entropy gradient in the superadiabatic layer is non - negligible and increases with increasing convective flux .
but as this region is thin and has low mass , it is sufficient to incorporate its effect into the boundary conditions .
the total , convective plus radiative , flux perturbation follows from equation . whereas equation determines the radiative flux perturbation in terms of @xmath72 and @xmath73 .
solving four linear , homogeneous , first - order , ordinary differential equations to obtain eigenvalues and eigenfunctions , requires a total of five boundary conditions .
four of these express physical constraints imposed by the environment outside the domain in which the differential equations are to be integrated .
the fifth merely sets the magnitude scale and phase for the eigenfunctions .
approximations described in [ subsec : nonad - eqnconv ] enable us to lower the outer boundary from the photosphere to the top of the radiative interior at @xmath18 .
the three boundary conditions applied there read : @xmath82 , \label{eq : nonad - ddppzb } \\
\delta s & = & { ( b+c)\over 1-i\omega\tau_c}\left({{\delta f}\over f}\right ) .
\label{eq : nonad - dszb}\end{aligned}\ ] ] the constant @xmath83 in equation sets the scale and phase of the eigenfunction .
equation follows from the near vanishing of @xmath84 in the convection zone , as expressed by equation , together with equations ( 20 ) and ( 21 ) of paper ii , which account for the jump in @xmath84 across the convective - radiative boundary .
equation relates the entropy perturbation in the main part of the convection zone to the flux perturbation that enters from below ( see paper i ) .
it is advantageous to raise the bottom boundary from the center of the star to a depth , @xmath85 , where the quasiadiabatic approximation is valid and the plane - parallel approximation still applies .
this step alters the spectrum of eigenvalues @xmath3 .
however , a simple procedure to be described later allows us to recover values for @xmath86 appropriate to a complete stellar model .
we generally take @xmath85 to be the level at which @xmath87 .
since @xmath88 at @xmath89 , the quasiadiabatic approximation is valid there for all g - modes of interest to our investigation . and in da
white dwarf models provided to us by bradley ( @xcite ) , the region above @xmath89 extends over the outer @xmath90 percent of the stellar radius and includes about @xmath91 of the stellar mass .
the two boundary conditions imposed at @xmath85 are @xmath92 the mechanical boundary condition given by equation states that @xmath93 at @xmath85 .
it follows from equations and .
thus our model is bounded from below by a rigid wall .
equation is our thermal boundary condition .
it is the quasiadiabatic limit of the radiative diffusion equation which defines the coefficients @xmath94 and @xmath95 .
our lower boundary conditions are somewhat arbitrary . for example
, we could have adopted @xmath96 and @xmath97 as mechanical and thermal boundary conditions . however , because @xmath73 decreases rapidly with depth , taking @xmath97 as the thermal boundary condition makes it more difficult for our numerical schemes to converge on eigenvalues . instead of using complete white dwarf models , we work with plane - parallel , hydrogen envelopes computed on fine grids .
these are produced by integrating downward from the photosphere . ] using the lawrence livermore equation of state and opacity tables ( rogers @xcite , iglesias & rogers @xcite ) .
our envelopes have @xmath98 and cover the range @xmath99 .
we model convection by invoking the mixing length ansatz .
this involves introducing a dimensionless parameter @xmath43 such that = f(|_s| p)^13 ( m_p f_cvt)^23 .
[ eq : nonad - cvzparameter ] the parameter @xmath43 is of order unity and is related to the conventional mixing length ratio @xmath100 by @xmath101 . a smaller @xmath43 signifies more efficient convection and yields a thicker convection zone .
the radiative flux is related to the entropy gradient by = 3 p f_rad - _ ad .
[ eq : nonad - superparameter ] we determine @xmath102 , @xmath103 , and @xmath104 from equations and together with the relation @xmath105 .
the grids of our model envelopes are chosen fine enough to resolve the steep entropy profiles in the superadiabatic layer .
figure [ fig : na - env - model ] displays some characteristics of the surface convection zones in envelope models produced with @xmath106 .
this value of @xmath43 enables us to match our model to that of bradley s at @xmath107 .
notice that the eddy turn - over time , @xmath46 , is of order a few seconds in even the coolest models of interest .
nonadiabaticity is quantified by how much the presence of the entropy perturbation , @xmath73 , affects @xmath75 and @xmath77 .
our analysis for the radiative interior leads to separate criteria for the evanescent region and the propagating cavity .
the effect of @xmath73 on @xmath75 is contained in equation . noting that @xmath75 varies on the scale @xmath108
, we obtain @xmath109 \left({{\delta p}\over p}\right ) & \sim & \left({k_h z n\over\omega}\right)^2\left({{\delta p}\over p}\right ) \sim { z\over{z_\omega } } \left({{\delta p}\over p}\right ) , \label{eq : nonad - scalesp}\end{aligned}\ ] ] where @xmath110 and @xmath111 .
the nonadiabatic term is of order , _
s s ~z s. [ eq : nonad - scalesds ] so the nonadiabatic correction to @xmath75 is of order @xmath73 . to relate @xmath73 to @xmath75 , we turn to equations and . with appropriate scalings they yield to @xmath77 in equation
varies on scale @xmath17 , although weakly . ]
@xmath112 together , equations - imply that the ratio of the nonadiabatic to adiabatic contributions to both @xmath113 and @xmath77 is of order @xmath114 , as is commonly cited in the literature . here , all perturbation quantities vary on a vertical scale @xmath115 , where @xmath116
. moreover , it follows from the local dispersion relation derived in paper i that @xmath117 . scaling the adiabatic terms in equation yields @xmath118 \left({{\delta p}\over p}\right ) & \sim & { { z}\over{z_\omega } } \left({{\delta p}\over p}\right ) \sim ( k_z z)^2 \left({{\delta p}\over p}\right ) .
\label{eq : nonad - scalesp2}\end{aligned}\ ] ] to order of magnitude , the nonadiabatic term is given by _
s s ~z s ~(k_z
z)^2 s. [ eq : nonad - scalesds2 ] these equations imply that the nonadiabatic correction to @xmath75 is of order @xmath73 .
we scale equations and to relate @xmath73 to @xmath75 ; @xmath119 combining equations - , we determine that the ratio of nonadiabatic to adiabatic contributions to @xmath113 and @xmath77 is of order @xmath120 .
this is not surprising ; @xmath121 is the timescale of thermal diffusion across distance @xmath115 .
nonadiabaticity is measured by the ratio of the mode period to this timescale .
we follow a two step procedure in solving the linear pulsation equations - for the four dependent variables subject to the five boundary conditions given by equations - .
the initial step consists of guessing a value for the complex eigenfrequency and then applying a relaxation method ( cf .
press @xcite ) to solve the differential equations subject to four out of the five boundary conditions .
the step is complete when the dependent variables satisfy both the pulsation equations and the boundary conditions to @xmath122 of a scaling factor provided by the corresponding adiabatic eigenfunction at individual points .
working in double precision , this is routinely achieved .
the second step determines the eigenvalue by requiring the remaining boundary condition to be satisfied .
both minimization and root finding techniques work well .
normally we reserve equation for our fifth boundary condition , but identical results are obtained when others are used instead .
nonadiabatic eigenvalues of our plane - parallel hydrogen envelopes are denoted by primes , @xmath123 . to relate @xmath124 to @xmath86 of a complete da white dwarf model
, we proceed as follows .
we compute adiabatic modes for a corresponding complete model and indentify those whose frequencies bracket @xmath125 . as the mode of the plane - parallel envelope . ]
we determine by interpolation the non - integer number @xmath22 of nodes in @xmath75 to assign to the adiabatic mode of the complete model whose frequency @xmath126 .
suppose that @xmath127 of these nodes lie above @xmath89 . then _
i_i^. [ eq : nonad - real ] validation of equation is provided by equation in the appendix .
its physical justification is that whereas energy is dissipated near the surface , it is stored throughout the mode cavity .
our procedure for converting @xmath124 to @xmath86 requires nonadiabatic effects to be small well above @xmath89 , as is the case for all modes of interest to our investigation .
however , it only applies to modes which have at least one node above @xmath128 .
thus it can not handle the highest frequency g - modes .
values of @xmath86 for these are computed from the work integral .
very low order modes ( @xmath129 ) are evanescent above @xmath128 .
we therefore can not enforce the boundary condition equation on them to solve for the eigenvalues .
we rely on work integral for stability analysis of these modes . calculating the work integral
provides an approximate method for evaluating the driving or damping rate of a mode .
the work integral computes @xmath130
. this method is well defined and accurate when the quasiadiabatic approximation applies .
moreover , it reveals regions of driving and damping .
since overstable g - modes of zz cetis have high quality factors , their linear pulsations are nearly periodic .
the work integral may be calculated as follows ( unno @xcite ) : @xmath131 .
\label{eq : nonad - workps}\end{aligned}\ ] ]
the effects of nonadiabaticity are illustrated by comparing nonadiabatic and adiabatic eigenfunctions for three @xmath0 g - modes of the same stellar model .
the model is characterized by @xmath132 , @xmath98 , and @xmath106 , which together imply @xmath133 at @xmath134 .
the modes have periods of @xmath135 , @xmath136 , and @xmath137 .
quasiadiabatic calculations predict overstability for each of these modes .
radiative diffusion acts to soften sharp temperature gradient perturbations .
its importance increases with mode period .
these characteristics are illustrated in figures [ fig : na - nad430]-[fig : na - nad2400 ] .
at @xmath138 by setting it equal to its adiabatic counterpart .
thus the imaginary component of @xmath77 is due to nonadiabaticity . ]
the @xmath135 mode , with @xmath139 at @xmath140 , is quite adiabatic in the radiative interior .
radiative diffusion is more pronounced in the evanescent region of the @xmath136 mode , which has @xmath141 at @xmath140 .
the nonadiabatic eigenfunction of the @xmath137 mode , for which @xmath142 at @xmath140 , deviates significantly from its adiabatic counterpart in both the evanescent and propagating regions .
the phase shift between @xmath77 and @xmath75 at @xmath18 is so large that the convection zone contributes to mode damping .
this mode is stable .
the recent detection of velocity signals associated with g - modes by van kerkwijk ( @xcite ) is an important advance in the study of zz cetis .
what is being detected are horizontal velocities near the stellar limb .
phase differences and amplitude ratios of flux and velocity variations due to individual modes offer unique clues about both the star and the modes .
nonadiabatic calculations , such as that depicted in figure [ fig : na - nafun4085 - 800 ] , provide a theoretical basis for interpreting these observable quantities .
deep inside the radiative interior , the maximum flux perturbation preceeds the maximum horizontal velocity by a quarter of a period , and the flux to velocity amplitude ratio is determined by the quasiadiabatic approximation .
these relations are modified in the upper envelope by the convective transport of heat and momentum and by radiative diffusion .
convection delays and diminishes the flux variation ( see paper i )
. it also flattens the horizontal velocity profile above @xmath18 and forces it to jump at @xmath18 ( see paper ii ) .
radiative diffusion smears flux perturbations below the convection zone . in the example
illustrated in figure [ fig : na - nafun4085 - 800 ] , a photospheric flux variation of @xmath143mma mma variation means @xmath144 .
] corresponds to a horizontal velocity of @xmath145 .
nonadiabaticity , mostly in the convection zone , reduces the phase lag between velocity maximum and light maximum from the quasiadiabatic value of @xmath146 to @xmath147 at the photosphere . employing the conventions in van kerkwijk ( @xcite )
, this mode would exhibit @xmath148 .
this is consistent with the @xmath149 mode which they observed to have @xmath150 and @xmath151 .
where applicable , direct calculations of @xmath86 yield results consistent with those based on the work integral .
this provides a measure of confidence in both .
modest discrepancies are found for some marginally overstable modes .
the eigenvalue calculations are more reliable , as work integrals suffer from inaccuracies due to cancellation between comparable magnitudes of driving and damping
. values of @xmath86 for individual g - modes evolve with decreasing @xmath152 .
effects of nonadiabaticity on driving and damping rates of g - modes are presented in figures [ fig : na - nonadratecombi]-[fig : na - vis - jumpglobal ] .
general trends are described below .
upper lids of cavities of short period modes ( e.g. , the @xmath135 mode ) lie far beneath the bottom of the convection zone in even in the coolest variables .
their driving / damping rates are largely immune to both nonadiabatic effects and the depth of the convection zone , and hardly vary across the instability strip .
the quasiadiabatic estimate , @xmath153 , pertains to these modes ( see paper i ) .
longer period modes ( e.g. , the @xmath154 mode ) become overstable at lower @xmath152 .
values of their @xmath155 exhibit a steady decline with decreasing @xmath152 .
this is a consequence of the increase in mode inertia with decreasing @xmath152 .
as the convection zone deepens , it depresses the upper lid of a mode s cavity .
this decreases the relative size of the perturbation amplitude near the surface , where nonadiabaticity is greatest , with respect to that in the interior , where most of the mode energy is stored
. modes with even longer periods ( e.g. , the @xmath156 mode ) are weakly overstable in cooler and narrower temperature ranges .
nonadiabatic effects tend to stabilize these modes . in the limit of strong dissipation , @xmath157 , where @xmath158 is the reflection coefficient at the cavity lid .
we derive this relation using a toy model in the appendix .
magnitudes of nonadiabatic damping rates are an important input to calculations of parametric instability , an amplitude limiting mechanism for overstable modes that is explored in the next paper of this series .
the derivation of the overstability criterion @xmath160 depends on the quasiadiabatic approximation and the condition that @xmath161 ( brickhill @xcite , paper i ) . since @xmath162 at @xmath18 ,
the quasiadiabatic approximation is suspect for modes with @xmath163 .
the upper and lower panels of figure [ fig : na - vis - jumpglobal ] display the boundary between overstable and damped modes as determined by quasiadiabatic and nonadiabatic calculations . both show that the boundary value of @xmath164 is close to unity for hot zz cetis but rises as stars cool .
inaccuracy of the quasiadiabatic approximation does not dramatically modify the overstability criterion , but the violation of the condition @xmath165 does . the latter occurs because @xmath18 increases with decreasing @xmath152 and @xmath19 decreases with decreasing @xmath3 . modes with @xmath166 propagate just below @xmath18 .
their short wkb wavelengths enhance radiative damping above the estimate given by equation ( 54 ) in paper i on which the derivation of the overstability criterion @xmath160 rests .
both our quasiadiabatic and nonadiabatic calculations consistently predict maximum periods for overstable modes that are about a factor two longer than those observed ( cf .
[ fig : na - vis - l1mode - new ] ) .
we have not been able to resolve this discrepancy .
several possible explanations which we have considered are described in what follows .
both brickhill ( @xcite ) and paper ii stress that linear damping due to turbulent viscosity , although negligible inside the convection zone for all modes , may be significant in the region of convective overshoot for long period modes .
we incorporate this effect into our analysis , while recognizing that it can not be quantified precisely .
the rate of turbulent damping in the overshoot region is expressed as ( eq . [ 36 ] in paper ii ) , _ vis - os , [ eq : na - vis - os ] where @xmath167 is the normalized jump in the horizontal displacement across the convective overshoot region within which the turbulent viscosity decays on length scale @xmath168 . to cover our ignorance , we present results for @xmath169 of @xmath170 and @xmath171 .
figure [ fig : na - vis - l1mode - new ] demonstrates that with the larger value , the longest period overstable mode at a given @xmath152 is compatible with the observed upper limit of @xmath172 .
however , it seems unlikely that @xmath173 .
the overshoot region probably extends less than a pressure scale height below @xmath18 , while @xmath174 , and @xmath168 must be several times smaller than @xmath175 .
we modify our envelope models to simulate convective overshoot by adding an isentropic layer below the base of the convection zone . within this layer
the radiative flux exceeds the total flux and turbulence transports energy downward .
the top of the radiative interior immediately below the overshoot layer differs from that below a convective layer of similar depth .
the temperature gradient is shallower and the frequency has a finite positive value .
these differences enhance the effects of radiative diffusion and modify the relations between horizontal velocity and light perturbations at the photosphere .
figure [ fig : na - nad1500-os ] provides such an example .
our nonadiabatic calculations indicate that one pressure scale height of overshoot suffices to stabilize all modes with @xmath176 ( see fig .
[ fig : na - vis - l1mode - new ] )
. moreover , by effectively deepening the convection zone , it shifts the instability strip to higher @xmath177 .
our nonadiabatic calculations indicate that , at fixed @xmath152 , the longest period for an overstable mode scales approximately as @xmath178 .
thus maximum periods of overstable modes around @xmath179 would require @xmath180 , which is well outside observational constraints .. ] hansen ( @xcite ) show that damping due to upward propagation of a gravity wave above the photosphere is unimportant for modes having frequencies exhibited zz cetis .
we concur with this conclusion .
detection of lower frequency modes requires longer observational data streams .
noise due to variations of atmospheric transparency increases at lower frequencies ( see fig . 4 of winget @xcite ) . these may result in an observational bias against the detection of low frequency modes .
evidence that photometric amplitudes of modes decline for @xmath181 , suggests that the observational cutoff at @xmath182 ) is genuine ( see fig . 5 in clemens @xcite ) .
taken literally , it also hints that the cutoff is the result of a nonlinear mechanism which saturates pulsation amplitudes .
there is little doubt that convective driving , as originally proposed by brickhill ( @xcite ) , is responsible for the linear overstability of g - modes in da white dwarfs .
it is physically self - consistent , and convincingly rationalizes observational facts .
it accounts for the general location of the instability strip , although a precise specification depends upon the modeling of convection ( e.g. , the mixing - length parameter ) .
convective driving also explains why longer period modes become overstable as a star cools ( see fig .
[ fig : na - vis - l1mode - new ] ) .
although we agree with brickhill that @xmath10 is a necessary condition for mode overstability , we find that it is not a sufficient condition for modes whose periods exceed @xmath183 .
this stems from enhanced radiative damping of modes whose upper cavity lids approach @xmath18 , as is apparent from both quasiadiabatic and nonadiabatic calculations ( see fig .
[ fig : na - vis - jumpglobal ] ) .
our nonadiabatic calculations yield a maximum period of about @xmath184 for overstable modes .
this clashes with the maximum period of @xmath6 for observationally detected modes .
we also agree with brickhill s deduction ( @xcite ) that turbulent convection forces the horizontal velocity to be nearly independent of depth within the convective envelope .
consequently , mode damping due to turbulent dissipation within the convection zone is reduced to a negligible level
. however , suppression of the horizontal shear in the convective envelope results in a shear layer at the top of the radiative interior .
when convective overshoot is accounted for , this provides linear turbulent damping .
figure [ fig : na - vis - l1mode - new ] suggests that turbulent dissipation might depress the maximum period of observable modes to some degree .
convective overshoot also alters the thermal structure of the upper radiative layer .
a rough treatment of this effect with one scale height of overshoot predicts a maximum period of @xmath185 for overstable modes .
nonlinear interactions which limit the amplitudes of overstable modes may also play a part in determining the maximum period . in conclusion
, the blue edge of the theoretical instability strip seems to be set by the condition that @xmath186 for the lowest order @xmath0 mode .
but the location of the red edge is more nebulous , and may result from a combination of decreased photometric visibility , convective overshoot , and nonlinear effects .
additional detections of velocity variations associated with g - modes could provide important clues .
convective driving makes the testable prediction that velocity variations become relatively more observable than photometric variations toward the red edge of the instability strip .
bergeron , p. , wesemael , f. , lamontagne , r. , fontaine , g. , saffer , r. a. , & allard , n. f. 1995 , , 449 , 258 bradley , p. a. 1995 , baltic astronomy , 4,536 bradley , p. a. 1996 , , 468 , 350 brickhill , a. j. 1990 , , 246 , 510 brickhill , a. j. 1991 , , 251 , 673 clemens , j. c. 1995 , baltic astronomy , 4 , 142 gautschy , a. , ludwig , h. g. , & freytag , b. 1996 , , 311 , 493 goldreich , p. & wu , y. 1998 , , accepted for publication , ( paper i ) goldreich , p. & wu , y. 1998 , , submitted , ( paper ii ) hamada , t. & salpeter , e. e. 1961 , , 134 , 683 hansen , c. j. , winget , d. e. & kawaler , s. d. 1985 , , 297 , 544 iglesias , c.a . , & rogers , f. j. 1996 , , 464 , 943 van kerkwijk , m. h. , clemens , j. c. , wu , y. 1998 , submitted to press , w. h. , teukolsky , s. a. , vetterling , w. t. , & flannery , b. p. 1992
, numerical recipes in fortran , ( cambridge university press ) , 403 , 753 , rogers , f. j. , swenson , f. j. & iglesias , c. a. 1996 , , 456 , 902 winget , d. e. , nather , r. e. , clemens , j. c. , 1991 , , 378 , 326 unno , w. , osaki , y. , & ando , h. 1989 , nonradial oscillations of stars , ( university of tokyo press ) , 238
we describe a simple toy model for nonadiabatic modes .
it is particularly useful for interpreting damping rates in the limit of strong dissipation . consider waves which satisfy the one - dimensional , homogeneous , acoustic wave equation = u^2 ^2 , [ eq : na - homowave ] in the interval @xmath187 .
here @xmath188 is the lagrangian displacement and @xmath189 is the constant propagation speed .
the dispersion relation connecting frequency , @xmath3 , and wave vector , @xmath190 , reads @xmath191 .
we take the lower boundary to be a rigid , perfectly reflecting wall , so = 0 z = l .
[ eq : na - bdcl ] dissipation is introduced by means of a partially reflective upper boundary , where @xmath158 denotes the reflection coefficient of the incident wave .
this is expressed through the boundary condition ( t - uz)= -r(t+u z ) z=0 .
[ eq : na - bdcu ] note that in the limit @xmath192 , the upper boundary becomes a perfectly reflecting wall .
eigen - solutions of equation are composed of oppositely directed waves ; =
a e^-it - i k z + b e^-i t + i k z. [ eq : na - homowavesol ] application of the boundary conditions given by equations and yields b =- ra , [ eq : na - reflect ] and kl = n - i2 ^ -1 , [ eq : na - k ] where @xmath22 is the number of half wavelengths between the walls
. then the dispersion relation implies _
i= k_i _ r = -_r2n^-1 , [ eq : na - omi ] where @xmath22 is an integer .
equation is the key result of our toy model .
it demonstrates that @xmath86 grows logarithmically with @xmath193 , and provides an order - of - magnitude estimate for the damping rate of strongly nonadiabatic stellar modes .
radiative diffusion substantially suppresses the effective reflection coefficients of modes having @xmath194 . in hot da stars
, this applies to modes with periods in excess of @xmath183 ( cf .
[ fig : na - nonadratecombi ] ) .
finally , equation justifies the procedure we apply in [ subsec : nonad - stellar ] to translate @xmath124 , the damping rate for a mode of an envelope model , to @xmath86 , the damping rate for a mode of a complete stellar model . | we report on numerical calculations of nonadiabatic eigenvalues and eigenfunctions for g - modes in zz ceti variables .
the spectrum of overstable @xmath0 modes delineates the instability strip .
its blue edge occurs where @xmath1 for the @xmath2 mode . here
@xmath3 is radian frequency and @xmath4 is about four times the thermal timescale at the bottom of the surface convection zone .
as a zz ceti cools , its convection zone deepens , longer period modes become overstable , but the critical value of @xmath5 separating overstable and damped modes rises .
the latter is a consequence of enhanced radiative damping for modes which propagate immediately below the convection zone .
the critical value of @xmath5 is of observational significance because modes with the smallest value of @xmath5 are most observable photometrically .
maximum periods for overstable modes predicted for our cooler model envelopes are about a factor two longer than the observational upper limit of @xmath6 .
we assess a number of plausible resolutions for this discrepancy among which convective overshoot and nonlinear saturation look promising .
the nonadiabatic eigenfunctions enable us to predict relative amplitudes and phases of photospheric variations of flux and velocity , quantities made accessible by recent observations .
we also present asymptotic formula for damping rates of high order modes , a result of consequence for future investigations of nonlinear saturation of the amplidues of overstable modes .
# 1([#1 ] ) |
the theoretical @xcite@xcite and experimental @xcite discoveries of quantum spin hall effect in 2d followed by its extension to the idea of topological insulators in 3d @xcite have become one of the most exciting developments in condensed matter physics in the @xmath3 century @xcite.topological insulators ( tis ) are characterized by novel response properties the most striking of which is an electromagnetic response called magnetoelectric ( tme ) effect in 3d topological insulators , where an applied magnetic field @xmath4 induces electric ( charge ) polarization @xmath5 ( or applied electric field @xmath6 induces magnetization @xmath7).in topological field theory @xcite , this effect is derived from the effective action @xmath8 known as axion electrodynamics @xcite , giving rise to magnetoelectric response @xmath9.axion term also leads to witten effect where a unit magnetic monopole binds a generally fractional electric charge in a medium with @xmath10 , as also studied in topological insulator context @xcite.the @xmath11 is the pseudoscalar axion field ; odd under parity and also odd under time reversal.for static time - reversal invariant systems , @xmath12 or @xmath13 , corresponding to ordinary and topological insulators respectively.when time reversal symmetry is broken , @xmath14 .
when the system is dynamical @xcite , the @xmath11 becomes a time - dependent function , which we will later denote by @xmath15 .
we will study this dynamical case and focus on @xmath16 response . to study this magnetoelectric response of 3d topological insulators
, we will consider fu - kane - mele ( fkm ) model of 3d ti @xcite given by @xmath17 where @xmath18
are the sites of 3d diamond lattice , @xmath19 is the electron creation ( annihilation ) operator at site @xmath20 , @xmath21 the hopping integral , @xmath22 the spin - orbit coupling strength , @xmath23 the cubic unit cell size , @xmath24 the pauli spin matrix vector , @xmath25 and @xmath26 are two consecutive nearest - neighbor vectors connecting two second - neighbor sites.the 3d topological insulators are characterized by a set of four @xmath27 topological invariants @xmath28 @[email protected] phase diagram of fkm model of 3d tis contains weak topological insulator ( wti ) phase ( @xmath29 , essentially equivalent to ordinary insulator ) which is not robust to weak perturbation and disorder and strong topological insulator ( sti ) phase ( @xmath30 ) , which is robust to weak perturbation and disorder.a number of materials have been predicted to be 3d stis , such as @xmath31 , @xmath32 , and @xmath33 under axial strain @xcite.the topological phase transition between the two phases is normally marked by a gap closing point @xcite.however , the gap itself can not be used as an ` order parameter ' of the transition as the gap is nonzero in both ordinary and topological insulating phases of 3d tis.this is what motivates our work here where we consider dynamical axion response of 3d tis and obtain as our main results ; 1)tme effect can be described in linear response theory with a vector potential involving axion field.2)correlation function of topological current is used to define a dynamical current susceptibility which in its static limit surprisingly gives an ` order parameter ' of the topological phase transition.3)the ( real part of ) dynamical current susceptibility shows a sign - changing singularity whose strength is suppressed in the topological insulating state.the latter two results provide the signatures of the topological phase transition in 3d tis .
most of the existing studies on the electromagnetic response in solids employed the abstract berry phase formalism @xcite@[email protected] this paper , we employ current density formalism @xcite@xcite which is more physically motivated as it works directly with such measurable quantities as current.we study the response of a charged particle system to electromagnetic fields described by 4-vector potential @xmath34.we focus on the case with an applied magnetic field and choose a gauge where @[email protected] hamiltonian is given by @xmath36 where we use a non - relativistic fermion system of mass @xmath37 and charge @xmath38 as an example , giving rise to charge current operator @xmath39 @xmath40\ ] ] @xmath41 the first and the second terms are known as the paramagnetic current @xmath42 and the diamagnetic @xmath42 current , respectively ; @xmath43.such decomposition also works for relativistic fermion systems @xcite@xcite . in 3d tis , the applied magnetic field
is expected to induce polarization @xmath5.we can express the polarization in terms of the value of the current @xmath44 @xmath45 within linear response theory , the induced current in response to vector potential ( in a gauge where @xmath35 ) can be written in normalized units as @xcite@xcite@xcite @xmath46 where @xmath47 is the paramagnetic - paramagnetic current - current correlation function , @xmath48 is the ground - state density , and @xmath49 is the vector potential.the first and second terms in eq .
( [ current ] ) represent the paramagnetic and diamagnetic contributions respectively to the induced current.using this current - based formalism and linear response theory to study fkm model of inversion - symmetric 3d tis , we found that tme effect can not arise from purely maxwellian electrodynamics @xcite.a novel form for vector potential @xmath50 in accordance with axion electrodynamics is required in order to consistently describe the tme effect.we will derive this new form of vector potential in the following section .
we find the expression for @xmath50 by solving the field equations derived from the lagrangian involving the axion term.in the presence of axion term mentioned in the beginning , rewritten as @xmath51 where @xmath52 is the coupling constant , and @xmath15 is the generally time - dependent axion field added to the standard maxwellian electrodynamics described by @xmath53 , the field equations are modified and are given as @xcite@xcite @xmath54 @xmath55 we note the presence of extra terms in the form of additional charge density in eq.([mae1 ] ) and current density in eq.([mae2 ] ) due to the axion field @xmath15.considering a situation with no free charge density ( @xmath56 ) and no free current density ( @xmath44 ) _ within _ a periodic solid subjected to an _ external _ spatially uniform magnetic field @xmath4 , we found that the axion field @xmath15 @xcite generates an electric field @xmath6 given by @xmath57 this electric field then induces polarization @xmath5 in the solid , manifesting the tme effect .
since the axion field @xmath15 is time - dependent , the resulting electric field @xmath6 and polarization @xmath5 are also time-dependent.as a result , one obtains a charge current following from @xmath58.dynamical axion ensures the presence of a charge current even when the magnetic field is static.this current further modifies the vector potential @xmath50.we found that the vector potential @xmath50 for a general time - dependent uniform magnetic field becomes @xcite @xmath59 @xmath60 we note that @xmath50 now acquires a new term in addition to the standard @xmath61 ; the axion term @xmath62 , which is proportional to the applied uniform magnetic field @xmath4 , implying that the polarization @xmath5 acquires a component along the direction of @xmath4 , according to eqs.([p - j ] ) and ( [ current]).this is the topological polarization of the tme effect .
we now consider inversion - symmetric fu - kane - mele model of 3d tis @xcite for concreteness , where the maxwellian @xmath63 term generates no polarization @xcite , so that we can focus on the topological part of @xmath50 of interest.considering general space - time dependent axion field @xmath64 and treating the paramagnetic and diamagnetic parts on equal footing @xcite , we obtain the following equation for the topological induced current , @xmath65 @xmath66\ ] ] where the paramagnetic and diamagnetic contributions take parallel form.we seek to characterize the response of topological insulators in terms of this quantity .
consider the ( retarded ) total current - total current correlation function , directly connected to linear response theory , given by @xmath67|g\rangle\ ] ] where @xmath68 represents the ground state expectation value , @xmath69 the current operator in interaction picture @xcite.now we define the dynamical current susceptibility @xmath70 where @xmath71 are integrals over the unit cell of the lattice ( with unit volume in our unit system).using eqs.([current]),([totalcurrentomega ] ) , and ( [ retardedcorrfunction ] ) , exact derivation @xcite gives @xmath72 where @xmath73 with @xmath47 is given by eq.([retardedcorrfunction ] ) but with @xmath74 in place of @xmath75 and @xmath76 is the ground state average number of electrons per unit cell for each spin species , which is a given number.the second term in eq.([totaldyncurrsuscept ] ) thus merely acts as a constant shift to the first term .
the result eq.([totaldyncurrsuscept ] ) suggests that we can adopt @xmath2 to study the response of 3d tis and its dependence on the frequency and microscopic parameters of @xmath77 , especially the hopping integral , as we tune the system across its topological phase transition.this @xmath2 can be used as a dynamical characterization of the tme effect of 3d ti in response to a uniform magnetic field applied along an appropriate direction.the response will be dominated by the diagonal elements of @xmath2 because in tme effect @xmath5 is collinear with @xmath4.we will explicitly compute @xmath78 as example .
this response function eq.([etafunction ] ) is determined entirely by the intrinsic band structure properties of the insulators.physically , @xmath79(@xmath2 ) represents the correlation between the total currents ( induced topological currents ) at two different points , averaged over space , as it is related to two - particle green s function @xcite.this quantity therefore encodes and should be able to detect the nonlocal global topological properties of the band structure that manifests in the electromagnetic response of the insulator .
we calculated the function @xmath78 using the general expression for @xmath47 @xcite derived from two - particle green s function formalism @xcite , @xmath80|g\rangle\ ] ] @xmath81\ ] ] where @xmath82 is the number of unit cells ( the number of @xmath83 points in the first brillouin zone ) , @xmath84 , @xmath85 being the fermi - dirac distribution function at @xmath86 of the @xmath87 bloch state @xmath88 at wavevector @xmath83 corresponding to the @xmath87 energy band @xmath89.the paramagnetic current operator is @xmath90 where @xmath91 acts on everything to its right while @xmath92 acts on every thing to its left.the @xmath20 and @xmath93 denote the initial and final states , taken to be the ground state and excited states respectively.in this work , the @xmath94 simply equals 0 for @xmath95 and 1 for @xmath96 .
we computed the bloch states from the fkm model eq.([3dtifkmmodel]).our calculation applies to 3d tis described by @xmath77 in eq.([3dtifkmmodel ] ) with @xmath97 .
we show the profile for @xmath78 in the static limit @xmath98 corresponding to static susceptibility in fig.1 , demonstrating its variation as function of the deformation @xmath99 of hopping integral @xmath100 along two different directions ; @xmath101 $ ] and @xmath102 $ ] directions , each separately and simultaneously.the deformation of hopping integral can be accomplished by appropriate chemical doping which creates point defects of the type vacancies or substitutional defects in the semiconductors @xcite.the topological phase diagram obtained in fu - kane - mele s seminal work @xcite suggests that one has wti ( reducible to ordinary insulator ) for @xmath103 , sti ( true topological insulator ) for @xmath104 , and quantum critical point semimetal at @xmath105 for deformation along [ 111 ] while keeping @xmath105 along @xmath102 $ ] and vice versa , as well as when the hopping integral is deformed along the two directions simultaneously .
fig.1 interestingly suggests that the two topologically distinct insulating phases are distinguished by different behaviors of our static susceptibility in the following aspects:1.the curves corresponding to separate hopping deformations are linear functions with distinct slopes.the deflection occurs right at the quantum critical point semimetal.2.the simultaneous deformation gives rise to a close to the average of the static current susceptibilities of the separate deformations in the wti phase whereas in the sti phase , the former shows clear deviation away from the other two.these changes in behavior can be taken as an indication of the topological phase transition .
@xcite as function of hopping integral deformation @xmath99 ( proportional to band gap @xmath106 ) along directions @xmath101$],@xmath102 $ ] , and both , with equal @xmath99 .
b ) the resulting ` order parameter ' @xmath107 of the topological phase transition @xmath108 given in eq.([orderparameter]).the units are such that @xmath109 and @xmath110 and we have set @xmath111.,title="fig : " ] @xcite as function of hopping integral deformation @xmath99 ( proportional to band gap @xmath106 ) along directions @xmath101$],@xmath102 $ ] , and both , with equal @xmath99 .
b ) the resulting ` order parameter ' @xmath107 of the topological phase transition @xmath108 given in eq.([orderparameter]).the units are such that @xmath109 and @xmath110 and we have set @xmath111.,title="fig : " ] [ fig : pvsgap ] the above observations suggest that we can define a quantity as function of deformation @xmath99 which can serve as an ` order parameter ' of the topological phase transition ; @xmath112+[1,-1,-1]}(dt)-\overline{\chi^{\mathbf{j}_p\mathbf{j}_p}_{zz}(0)(dt)}\ ] ] where @xmath113}(dt)+\chi^{\mathbf{j}_p\mathbf{j}_p}_{zz}(0)_{[1,-1,-1]}(dt)}{2}\ ] ] fig.1 shows that @xmath107 goes from zero to nonzero across the topological phase transition and can therefore be used as an unambiguous signature of the occurrence of such transition.combining eqs.([totaldyncurrsuscept ] ) and ( [ orderparameter ] ) , it is clear that the total current susceptibility @xmath114 measurable in experiment will also give the same critical behavior .
this critical behavior of @xmath107 is nontrivial because naively , the susceptibility should only depend on the projection of the direction of electron hopping to the direction of applied magnetic field and thus should not depend on how the deformation of the hopping integral is implemented , given the two fixed directions.this turns out to hold only if the solid is always in the same phase relevant to its current response , e.g. weak topological ( equivalent to ordinary ) insulating phase.now the current susceptibility shows such unexpected behavior as one crosses a topological phase transition , even if the two directions of hopping deformation are fixed .
@xmath107 thus can serve as an ` order parameter ' for this topological phase transition which fig.1 suggests to be continuous , in agreement with experiments @xcite.this is also consistent with the expectation for a phase transition obtained by continuously tuning a parameter in a non - interacting fermion model eq.([3dtifkmmodel]).the computed static orbital polarizability also agrees with the result of berry phase theories @xcite .
theoretically , this nontrivial behavior originates from the @xmath115 symmetry of the band structure of the fkm model eq.([3dtifkmmodel ] ) under @xmath13 rotation about the @xmath116 axis preserved under simultaneous deformation of the hopping integrals with equal strength along @xmath101 $ ] and @xmath102 $ ] directions , but broken when the deformation is applied separately in one of the two directions.this @xmath115 symmetry manifests its @xmath27 effect @xmath117 in the current susceptibility for ( strong ) topological insulator since the surface fermi arc encloses odd number of dirac points ( @xmath30 ) while the effect vanishes in the weak topological ( ordinary ) insulator since the arc encloses even number of them ( @xmath29)@xcite.we have checked and confirmed numerically that such a critical behavior in @xmath118 would never occur if we set the spin - orbit coupling to zero for example , where only one type of insulator ; ordinary insulator ( at any @xmath119 ) could exist @xcite .
we now present the outcome for the dynamical response function @xmath78 in the ordinary and topological insulator phases as well as at the quantum critical point between them in fig.2.apart from the numerical fluctuations due to the finite - size effect around some critical frequencies , the profile of @xmath78 is consistent with the familiar @xmath120 behavior of the two - body ( e.g.density-density , current - current ) response functions that have a sign - changing singularity at a critical frequency @xmath121 at which the functions flip sign.using eq.([totaldyncurrsuscept ] ) and noting the magnitude of @xmath98 from figs.1 and 2 , it is clear that the singularity and the sign change also occur to the @xmath122 measurable in experiment.the ordinary and topological insulator phases can be distinguished by the overall strength of the singularity around @xmath121.in general , we find that topological insulating phase leads to suppression of the strength of singularity in the dynamical response function at around @xmath121.the switch from negative to positive correlation across @xmath121 is nontrivial but can be nicely explained in terms of the continuity equation obeyed by the charge density and the current induced by the polarization generated by the magnetic field via the tme effect @xcite .
@xcite for sti ( topological insulator ) , wti ( or ordinary insulator ) , and the critical point ( semimetal ) as function of frequency @xmath123.we have used @xmath124 along direction @xmath125 $ ] for the topological ( ordinary ) insulator phase as example.the units are such that @xmath109 and @xmath110 and we have set @xmath111.,title="fig : " ] [ fig : dynrespsti ]
topological phase transition occurs between different phases which can not be distinguished by a local order parameter with true symmetry breaking in the sense of ginzburg - landau theory @xcite.rather , they are described by a global topological property.the quantum hall transition @xcite is one of the first examples of such non - ginzburg - landau type of phase transition.the quantum spin hall effect and topological insulators also display topological phase transition between ordinary insulator and quantum spin hall ( topological insulator ) phases , marked by a gap closing in the bulk excitation spectrum.our work discovered an ` order parameter ' derived from the nonlocal two - particle green s function to probe this latter phase transition between the states distinguished by the global topological properties .
the response function that we define is determined by the current - current correlation function which amounts to a current noise power measurement . here
we propose the experimental setup for such measurement on 3d ti candidate materials described by fkm model @xcite to detect such topological phase transition.consider a thick slab sample of such 3d ti materials having front and rear surfaces normal to the @xmath101 $ ] or @xmath102 $ ] direction but parallel to a uniform magnetic field and are coated with insulating ferromagnet @xcite.the magnon of the ferromagnet provides the dynamical axion field that couples to the magnetic field and the induced electric field @xcite .
time - dependent uniform magnetic field induces current on the two surfaces , assumed uniformly distributed , and we propose to measure their correlation .
the current noise cross - corrrelation is given by @xcite @xmath126 where the current noise is @xmath127 at the @xmath128-th surface with @xmath129 the corresponding steady - state equilibrium current ( without the applied magnetic field ) .
the surface current will induce a time - dependent change in the charge density at the top and bottom ends of the sample , satisfying continuity equation so as to ensure charge conservation.our theory predicts that the behavior of @xmath130 is given by the profiles of the response function @xmath78 shown in figs.1 and 2.we have thus demonstrated a characterization of topological phase transition in 3d topological insulators from their dynamical axion response properties.it is important to note that our results based on current - based formalism are consistent with those obtained from berry phase formalism , as both approaches manifest the direct correspondence between the band topology and electromagnetic response properties that characterize topological insulators.our results are also universal as they hold for any 3d tight - binding fermion systems that host topological and ordinary insulating phases , including ultracold atom system realizing the axion electrodynamics @xcite .
the author was supported by the grant no.anr-10-labx-0037 of the programme des investissements davenir of france.the author thanks dr.p.romaniello and dr.a.berger for the guidance all along and prof.i.affleck for the insightful discussion and comments .
in these appendices , we first prove the absence of topological magnetoelectric effect in purely maxwellian electrodynamics .
we next discuss the linear response theory and the associated current - current correlation function and derive the eq.([totaldyncurrsuscept ] ) in the main text .
we then show a consistency check of our results .
finally , we give the details of the experimental setup of the current noise correlation measurement and explain how our results agree with simple physical intuition .
we will show that topological magnetoelectric effect can not arise from purely maxwellian electrodynamics .
we illustrate this from linear response theory applied to maxwellian electrodynamics to study the electromagnetic response of inversion - symmetric lattice periodic solids , where magnetic field is found to not generate any polarization . in inversion - asymmetric system ,
such polarization will be nonzero .
topological polarization coming from tme effect should be nonzero , regardless of symmetry property of the system under inversion and should therefore come from new term in the vector potential @xmath50 in addition to the familiar @xmath131 term , as we derived in the main text . here follows the proof .
standard maxwellian electrodynamics suggests that we can write the solution of the equation @xmath132 as usual , which in symmetric gauge is given by @xmath133 valid for spatially uniform magnetic field @xmath4 . for single molecule systems
, it has been found that the diamagnetic and paramagnetic contributions can be treated on equal footing and yield the reference - independent expression for current and polarization in the @xmath134 limit @xcite .
@xmath135_j\ ] ] the above expression with its relative position vector term @xmath136 is also the one required for lattice periodic solid systems based on the requirement of lattice translational invariance , which is realized by the @xmath136 .
the current - current correlation function itself is also translationally invariant by definition .
@xmath137 for translations by arbitrary lattice vectors @xcite .
note that the two independent position vectors @xmath138 and @xmath139 are translated by the same lattice vector @xmath140 in this definition of translational invariance . for lattice - periodic solid systems , the eq.([p - jsinglemolecule ] ) holds naturally due to the lattice periodicity and so we have @xmath141 in this work , we are studying the linear response to a spatially uniform macroscopic magnetic field , taken to be in @xmath142 for concreteness . writing the explicit form for each component of the polarization from maxwellian vector potential given in eq.([p - j0 ] )
, it can be easily checked that in current - based formalism , one can only obtain polarization parallel to the applied magnetic field from the contribution of the off - diagonal elements of the current susceptibility tensor .
diagonal elements of the susceptibility tensor only contribute to the polarization perpendicular to the magnetic field ; polarization on @xmath143 plane for magnetic field on @xmath142 direction .
the tme response described by the equation @xmath144 stated in the introduction in the main text on the other hand implies that the topological polarization is parallel to the magnetic field .
we will first show that for inversion symmetric systems , both the diagonal and off - diagonal elements of the tensor @xmath145 give no contribution to the polarization @xmath5 in eq.([p - j0 ] ) from the standard form of vector potential @xmath146 .
this conclusion is applicable to inversion - symmetric 3d topological insulators , such as the fkm model of 3d topological insulators on diamond lattice @xcite@xcite with diamond - shaped brillouin zone , which has inversion symmetry and results in the two - fold degeneracy of the upper and lower energy bands .
the key to the cancellation of the off - diagonal tensor element contribution to polarization is this inversion symmetry of the model .
the absence of this contribution to polarization can be shown as follows .
we want to compute polarization from eq.([p - j0 ] ) .
let us make the following substitutions @xmath147 .
as a result , we obtain @xmath148 where indices appearing twice are to be summed over . for a system with inversion symmetry
, we have @xmath149 . therefore , eq.([p - j1 ] ) can be written as @xmath150 comparing eqs.([p - j0 ] ) and ( [ p - j2 ] ) we arrive at the conclusion that @xmath151 . in summary , the contribution of the off - diagonal elements of the susceptibility tensor @xmath152 to the polarization from the maxwellian vector potential @xmath153 in our problem is identically zero at all frequencies @xmath123 .
this result does not imply that tme effect , which only requires time reversal symmetry on system with spin - orbit coupling , occurs only in inversion asymmetric 3d topological insulators .
tme effect does exist in inversion symmetric ti s but their net tme vanishes .
for inversion asymmetric system , the familiar @xmath146 contributes nonzero @xmath5 but which does not have topological character .
tme effect should exist in 3d ti regardless of whether the system is inversion symmetric or not and so this effect can not arise from the maxwellian solution for @xmath50 in order for the linear response theory formulation of the effect to be consistent .
the complete form of the vector potential @xmath50 is as given in eq.([totala ] ) in the main text , consisting of the maxwellian and axionic parts @xcite .
in linear response theory , the linear response of the system to an external perturbation is represented by the change in the expectation value of an observable coupled to the external potential .
this is described by the kubo formula @xmath154|g\rangle\ ] ] where @xmath155 represents the ground state expectation value , @xmath156 is the observable operator , and @xmath157 is the external perturbation hamiltonian , which are given by @xmath158 @xmath159 in the interaction picture , where @xmath77 is the unperturbed hamiltonian and @xmath160 with @xmath161 an appropriate perturbation field . in our work ,
the observable is the total charge current @xmath43 where the first and second terms represent the paramagnetic and diamagnetic contributions respectively given by @xmath162\ ] ] @xmath163 the total induced current in response to vector potential is therefore given by @xmath164|g\rangle\ ] ] where the current operator in the interaction picture is given by @xmath165 . using @xmath166
we obtain @xmath167 where we have used planck units where @xmath168 and defined a retarded total current - total current correlation function @xmath169|g\rangle\ ] ] taking the time fourier transform of this retarded correlation function , we obtain @xmath170 @xmath171|g\rangle\ ] ] where we have defined @xmath172 , which takes the range @xmath173 due to the @xmath174 .
the time fourier transform of eq.([totalinducedcurrent ] ) is @xmath175 it is known from standard many - body physics literature @xcite that the total induced current in linear response theory is given by @xmath176 @xmath177a_j(\mathbf{x}',t')\ ] ] where @xmath178 , the ground state fermion density @xmath179 , @xmath180 , and a unit mass @xmath181 is taken .
the @xmath182 is defined in the same way as @xmath183 in eq.([retardedcorrfunctiontotalj ] ) but with @xmath184 in place of @xmath39 .
taking the time fourier transform of eq.([totalinducedcurrentstrinati ] ) , we obtain @xmath185 @xmath186a_j(\mathbf{x}',\omega)\ ] ] comparing with eq .
( [ totalinducedcurrentft ] ) , we find that @xmath187 now consider the space - averaged response function that we propose ; the dynamical current susceptibility .
@xmath188 using eq.([magicrelation ] ) we obtain @xmath189 where the last term is nothing but the ground state average number of electrons per unit cell @xmath190 for each spin species .
this completes the derivation of eq.([totaldyncurrsuscept ] ) in the main text .
there are several consistency checks that can be done in order to verify the correctness of our current - based and linear response theory formalism .
it is first to be noted that we are working in linear response regime , which means the applied magnetic field @xmath4 is weak and the response physical quantity ; the polarization @xmath5 is linear in @xmath4 . in this case , in the linear response eq.(5 ) in the main text , the susceptibility tensor reflects a ground state property and the hamiltonian from which it is computed does not include the zeeman magnetic field applied to the system .
the axionic tme effect however enters via the vector potential @xmath50 , which has to be modified accordingly based on axion electrodynamics in order to describe such effect , as we noted earlier . as function of hopping integral deformation @xmath99 ( proportional to band gap @xmath106 ) along directions
@xmath101$],@xmath102 $ ] , and both directions simultaneously , in tight - binding model on diamond lattice without spin - orbit coupling .
the units are such that @xmath109 and @xmath110 and we have used @xmath111.,title="fig : " ] [ fig : pvsgapnoso ] as a consistency check , we present the result of our static susceptibility for model of tight - binding electron on diamond lattice without spin - orbit coupling .
one expects to have normal metal for isotropic case @xmath105 and ordinary insulator for any @xmath191 .
since the system is then ordinary insulator for both @xmath104 and @xmath103 with a change in the gap @xmath192 relative to that of the metal state at @xmath105 , the current susceptibility should behave the same way or become symmetric with respect to the critical gapless metal at @xmath105 for hopping integral deformation along any direction .
since ordinary insulator is topologically trivial , clearly we expect the topological order parameter as we have defined to vanish all the way .
that would confirm the consistency of the theory .
this is indeed what we obtain and confirm numerically as shown in fig.3 , where we use the same parameters as those used in obtaining fig.1 in the main text , except that we set the spin - orbit coupling to infinitesimally small number , e.g. @xmath193 .
we note that the magnitude of the static response without spin - orbit coupling is several orders of magnitude smaller than that of fig.1 in the main text with spin - orbit coupling .
in addition , the former shows no regular pattern and this simply means it corresponds to numerical fluctuations .
this simply suggests that this quantity vanishes in the zero spin - orbit coupling limit @xmath194 ( and thermodynamic limit ) . using eq.([orderparameter ] ) in the main text
, the order parameter @xmath107 is thus always zero in the absence of spin - orbit coupling , as should indeed be the case .
the residual nonzero total current susceptibility @xmath79 , which from eq.([totaldyncurrsuscept ] ) is simply a constant equals to the ground state average number of particles per unit cell per spin species for @xmath195 and zero otherwise , represents the ground state current correlation which does not come from the tme effect as the latter necessarily comes from the current induced by the applied magnetic field .
as we know 3d tis have surface state current even in the absence of external bias field and this contributes to the ground state current that gives rise to nonzero ground state correlation .
this consistency further corroborates our identification of the dynamical current susceptibility @xmath79 as a characterization of the tme effect and the order parameter @xmath107 as an order parameter of the topological phase transition in 3d topological insulators .
the experiment that we propose in the main text can be illustrated as in fig.4 . a thick slab sample of 3d ti candidate material is coated on each of its surfaces with a layer of ferromagnet with magnetization @xmath7 perpendicular to the interface @xcite , whose dynamical spin wave ( magnon ) provides the dynamical axion field @xmath15 that couples to the electric field @xmath6 induced by a generally time - dependent spatially uniform magnetic field @xmath4 applied parallel to the front and rear surfaces , which are normal to @xmath101 $ ] or @xmath102 $ ] direction .
this magnetization is also needed to obtain the time reversal symmetry breaking in analogy to the effect of magnetic field in quantum hall effect , in order to observe the tme effect @xcite .
consider a static magnetic field first . then there will be a polarization parallel to the magnetic field and which induces a charge density of opposite signs at the bottom and top ends of the sample . since the axion field is dynamical ,
this polarization is readily time - dependent and gives rise to a current , following from @xmath58 and a time - dependent charge density according to continuity equation @xmath196 .
applying time - dependent magnetic field gives rise to additional current which together with the charge density always satisfies the continuity equation . in this setup ,
the current will be uniformly distributed over the front and rear surfaces ( assuming edge effect is negligible ) so that point contact measurement is equivalent anywhere on the surface , corresponding therefore directly to the space - averaged value of the quantity of interest .
now suppose that the current is slowly increasing on the front surface , then the charge density on the bottom front as a result is decreasing .
but charge conservation demands that the current on the rear surface should be decreasing and the charge density at the rear bottom end be slowly increasing , for slow enough time variation of the current .
this implies that the currents on the front and rear surfaces are anticorrelated at small frequencies .
this is in complete agreement with our explicit calculations represented in fig . 1 and fig . 2 where the current cross - correlation is negative at dc and low frequencies @xmath197 .
at larger than critical frequencies , the currents on the two surfaces are positively correlated and there will be flow of charge from inside the bulk to supply for charge accumulation or depletion at the top and bottom ends of the sample .
this is easy to understand if we note that in the energy band dispersion in @xmath83 space , low frequency photons only excite surface state electrons or states near fermi energy whereas high frequency photons excite electrons deep within bulk bands .
the explanation for why the strength of sign - changing singularity is weaker in the ( strong ) topological insulating than in the weak topological ( or ordinary ) insulating phase is more curious but intuitively it relates to the surface state .
in general , the presence of surface state facilitates the flow of charge and helps the electromagnetic field to excite low energy electrons near fermi energy .
since the surface state of sti encloses an odd number of dirac points and so has at least one such dirac point , the transition from negative to positive current correlation is smoother as it is facilitated by the surface state , compared to a wti or ordinary insulator , which has no surface state or has surface state that encloses an even number of dirac points ( the effect of which may cancel out each other ) , the photons work by themselves in exciting the electrons without extra assistance from the surface state and this results in more dramatic sudden change from the negative to positive current correlation across the critical frequency @xmath121 .
xu , y. xia , l. a. wray , s. jia , f. meier , j. h. dil , j. osterwalder , b. slomski , a. bansil , h. lin , r. j. cava , and m. z. hasan , science 332 , 560 ( 2011 ) ; t. sato , k. segawa , k. kosaka , s. souma , k. nakayama , k. eto , t. minami , y. ando , and t. takahashi , nat .
phys . 7 , 840 ( 2011 ) ; l. wu , m. brahlek , r. v. aguilar , a. v. stier , c. m. morris , y. lubashevsky , l. s. bilbro , n. bansal , s. oh , n. p. armitage , nat .
9 , 410 ( 2013 ) .
the axion field @xmath15 can in general be function of space and time but for the response of topological insulators , topological field theory @xcite shows that the tme effect comes from the time dependence of the axion field @xmath15 .
the side surfaces of the ti should also be coated with ferromagnetic layer ( not shown in the figure ) but in such a way that still allows the measurement of the currents on the front and rear surfaces . | axion electrodynamics , first proposed in the context of particle physics , manifests itself in condensed matter physics in the topological field theory description of 3d topological insulators and gives rise to magnetoelectric effect , where applying magnetic ( electric ) field @xmath0 induces polarization ( magnetization ) @xmath1 .
we use linear response theory to study the associated topological current using the fu - kane - mele model of 3d topological insulators in the presence of time - dependent uniform weak magnetic field . by computing the dynamical current susceptibility @xmath2
, we discover from its static limit an ` order parameter ' of the topological phase transition between weak topological ( or ordinary ) insulator and strong topological insulator , found to be continuous.the @xmath2 shows a sign - changing singularity at a critical frequency with suppressed strength in the topological insulating state .
our results can be verified in current noise experiment on 3d ti candidate materials for the detection of such topological phase transition . |
80 gustavsson m _ et al .
_ 2008 interference of interacting matter waves _ preprint _
arxiv:0812.4836 will s _ et al . _ 2009 multi - orbital quantum phase diffusion _
gustavsson m _
_ 2008 , 080404 bloch i , dalibard j and zwerger w 2008 885 bloch i 2008 nature * 453 * 1016 greiner m , mandel o , hnsch t w and bloch i 2002 nature * 6902 * 51 giovannetti v _ et al .
_ 2004 science * 306 * 1330 ; esteve j _ et al .
_ 2008 nature * 455 * 1216 - 1219 .
morsch o and oberthaler m 2006 179 ; zenesini a _ et al .
_ 2008 053038 morsch o _
_ 2001 140402 tomadin a , mannella r and wimberger s 2007 130402 salger t _
_ 2007 190405 ; salger t _
_ 2009 science * 326 * 1241 breid b m , witthaut d and korsch h j 2006 110 tomadin a , mannella r and wimberger s 2008 a * 77 * 013606 tomadin a 2006 _ quantum chaos with ultracold atoms in optical lattices _ , ( master thesis , universit di pisa ) fukuyama h , bari r a and fogedby h c 1973 b * 8 * 55795586 kolovsky a r and buchleitner a 2003 e * 68 * 056213 press w _
_ 1992 _ numerical recipes in c _
( cambridge university press ) nakamura y , pashkin y a and tsai j s 2001 246601 meystre p and sargent m 1991 _ elements of quantum optics _ ( springer ) abramowitz m and stegun i a 1968 _ handbook of mathematical functions _ ( dover publications ) buchleitner a , delande d and zakrzewski j 2002 _ physics reports _ * 368 * 409 547 eckardt a , weiss c and holthaus m 2005 260404 zenesini a _ et al .
_ 2009 100403 anderlini m _ et al .
_ 2006 s199 ; sebby - strabley j _ et al .
_ 2007 200405 | we study the effect of a many - body interaction on inter - band oscillations in a two - band bose - hubbard model with external stark force .
weak and strong inter - band oscillations are observed , where the latter arise from a resonant coupling of the bands .
these oscillations collapse and revive due to a weak two - body interaction between the atoms .
effective models for oscillations in and out of resonance are introduced that provide predictions for the system s behaviour , particularly for the time - scales for the collapse and revival of the resonant inter - band oscillations .
_ introduction .
_ recent experiments proved the possibility to study the coherent dynamics of interacting many - particle systems @xcite .
such realisations of many - body systems with ultra - cold gases in optical lattices have a short but impressive history and open immense possibilities for various fields of physics @xcite .
the demonstration of the well - known phenomenon of collapse and revival , the latter being a pure quantum effect , with ultra - cold atoms bears witness of this coherent evolution of a many - body wave function @xcite .
additionally , the high degree of control in such experiments allows a manipulation of many system parameters and makes them particularly interesting for various fields of physics as well as future applications @xcite .
different ways of addressing additional degrees of freedom in such ultra - cold bosonic gases have been suggested @xcite . in the present fast track communication
, we discuss a two - band model with an additional external force to control the coupling between the two bands . applying a force to atoms in optical lattices leads to bloch oscillations and
is a realization of a many - body wannier stark system @xcite .
the coupling of the low - lying energy bands in such systems has been demonstrated in different experiments , e.g. on landau - zener tunnelling @xcite and the influence of the many - body interaction on a weak coupling of the bands has also been studied theoretically @xcite .
we go beyond a weak coupling of energy bands and consider an isolated two - band system with a strong external force .
a closed two - band system is an idealisation but can be realised approximatively with ultra - cold atoms @xcite using different techniques as , e.g. , super lattices @xcite .
besides the possibility of experimental realisation , a closed two - band model is also interesting as a simple model system . for the latter , we focus on the two lowest energy bands of interacting bosons in a optical lattice @xmath0 , with the wave vector of the optical lattice @xmath1 .
then , all parameters of the model hamiltonian just depend on this external potential .
the parameters can be computed numerically ( see , e.g. , appendix a in @xcite ) and analytical approximations exist for them for not too small amplitudes @xmath2 @xcite . using this setup , we are able to identify regions of strong and weak inter - band coupling .
a weak two - body interaction introduces new energy scales in the coherent evolution of the many - body wave function , leading to a collapse of such oscillations , but the form of the interaction gives rise to subsequent revivals .
we give analytical expressions for all time - scales in this many - body realisation of a two - band collapse and revival phenomenon , and experimental ramifications for a realisation with ultra - cold bosonic gases are discussed . _ the many - particle model .
_ we study a two - band bose - hubbard model with an additional external stark force for a strong coupling of the two bands , introduced in @xcite @xmath3 . } \ ] ] here , @xmath4 ( @xmath5 ) annihilates ( creates ) a particle at site @xmath6 of totally @xmath7 sites in the lower band and @xmath8 ( @xmath9 ) in the upper band .
the corresponding number operators are @xmath10 , @xmath11 .
the bands are separated by a bandgap @xmath12 and have onsite - energies @xmath13 , respectively .
we include hopping between neighbouring sites in band @xmath14 with a hopping strength @xmath15 , and a repulsive interaction between particles occupying the same site in band @xmath16 with a strength @xmath17 .
the two bands are coupled via @xmath18 , with the external stark force @xmath19 and a coupling constant @xmath20 depending on the depth of the lattice @xmath2
@xcite , and also via the inter - particle interaction with a strength @xmath21 .
all parameters are measured in recoil energies @xmath22 and we set @xmath23 troughout .
focussing on a realisation with a single optical lattice rather than a superlattice , the relation between the parameters is generally : @xmath24 , as well as @xmath25 and @xmath26 .
we take the external force @xmath19 as a free parameter .
we assume that the interaction strength can be tuned ( experimentally , e.g , by the use of feshbach resonances @xcite ) and include a corresponding scaling factor @xmath27 to all interaction terms . to study the occupation of the upper band , we prepare the system in an initial state @xmath28 , with a uniform distribution of particles in the lower band only and evolve it in time by the many - body schrdinger equation .
the quantity we study is the ( normalised ) number of particles in the upper band @xmath29 where @xmath30 is the total number of particles .
we will refer to @xmath31 as _ occupation of the upper band_. let us discuss the non - interacting single - particle case @xmath32 first .
we apply the following transformation involving bessel functions of the first kind @xmath33 to our operators , which is known to remove the hopping terms in the single - band case @xcite @xmath34 with @xmath35 , @xmath36 . using relations for bessel functions , we arrive at @xmath37,\ ] ]
where @xmath38 and @xmath13 as above .
we obtain coupling between any two sites of the two different bands , weighted by bessel functions . in the coupling between different and possibly remote sites
originates from an on - site coupling and subsequent hoppings .
it can thus be considered a higher order process in the original basis .
but in the transformed hamiltonian , this coupling is now direct with a strength modified by the factor @xmath39 .
the hilbert space in the many - particle problem is spanned by fock states with fixed particle number @xmath40 with single - particle basis @xmath41 .
the total dimension of the fock space for a given number of atoms @xmath42 and lattice sites @xmath7 per band is given by @xmath43 $ ] . for numerical simulations ,
we change to the interaction - picture with respect to the external force @xcite which removes the tilt @xmath44 and replaces @xmath45 ( and likewise for @xmath46 ) .
the hamiltonian is then time - dependent with a periodicity of @xmath47 and decomposes into a direct sum of operators for specific quasi - momenta @xmath48 @xcite . as a consequence
, the size of the hilbert space is reduced by a factor of the order of @xmath7 @xcite .
since the different subspaces are physically equivalent , we restrain our discussion to the @xmath49-subspace of the hilbert space @xcite . + for the time - evolution of a given initial state , we use either a direct numerical integration ( with an adaptive step - size runge - kutta algorithm @xcite ) or an eigenbasis expansion after diagonalising the problem : @xmath50 here we use the eigenstates @xmath51 of the floquet - bloch operator ( @xmath52 denotes time - ordering ) @xmath53,\ ] ] since the hamiltonian is explicitly time - dependent , with a periodicity @xmath54 . in addition , this gives us the full spectrum and enables access to relevant energy scales of the problem , as well as an identification of the most important states participating in the time - evolution .
_ results .
_ for the specific system under consideration , i.e. , bosons in optical lattices , both hopping coefficients @xmath55 are smaller than unity and ( since we are interested in strong inter - band coupling ) we take the external stark force @xmath19 to be much larger than the hopping coefficients : @xmath56 .
the non - interacting hamiltonian now allows for simple solutions for two regimes : values of the external force @xmath19 not leading to a degeneracy between energy levels of different bands ( _ off - resonant _ regime ) and values of the force @xmath19 leading to such a degeneracy ( _ resonant _ regime ) .
for the off - resonant regime we make use of @xmath57 and neglect all bessel functions in except for @xmath58 .
the hamiltonian then decomposes @xcite into a sum of independent two - level systems and the occupation of the upper band when initially zero follows a simple rabi formula @xmath59^{-1}\sin^2 ( \tilde\delta t/2)$ ] , where @xmath60 .
this corresponds to rabi oscillations between the bands with an amplitude much smaller than unity and a period @xmath61 of the order of the _ bloch period _ @xmath62 .
an example is shown in the lower panel of where the off - resonant contribution to the oscillations is shown .
although the coupling from a site @xmath6 to sites in the other band with different index @xmath63 is usually small ( cf . discussion after ) , it is important when the two levels become degenerate in energy , i.e. , for resonant values of the force @xmath19 .
this happens when the energy gap between both bands is close to an integer multiple of the external force @xmath64 and we refer to this regime as _ resonant of order @xmath65_. in resonance , the coupling of the degenerate levels is most important and the hamiltonian of can similarly be reduced to a sum of independent two - level systems @xmath66,\ ] ] and diagonalised by @xmath67 and @xmath68 .
the resonant oscillations ( of order @xmath65 ) between the two bands have an amplitude of almost unity and a period given by @xmath69 an example of these resonant oscillations for @xmath70 is shown as the thin dashed line in . the period predicted by the reduction to independent two - level systems
is @xmath71 for the parameters given there , in excellent agreement with where the actual period is @xmath72 .
we found equally good agreement in numerical simulations for different lattice depths and other orders @xmath65 of resonance not explicitly reported here .
let us now study the effect of the many - body interaction in the original hamiltonian
. also shows the occupation of the upper band as a function of time for the initial state @xmath73 in a weakly interacting system with @xmath74 .
we observe a decay of the resonant oscillations followed by a major revival . at later times
several minor revivals occur ( not shown in ) .
this effect is stable against variations of the system parameters ( as number of particles and lattice sites , even for fillings @xmath75 of order but not always close to one ) , with the time scales of decay and revival depending on the interaction strength @xmath27 ( see the following section and ) .
we found the same phenomenon with different initial states in numerical simulations as long as the particles are mainly delocalised along the lattice and occupy only the lower band , i.e. , excluding fock states with all particles on one lattice site for instance .
the oscillatory behaviour depicted in strongly reminds of the collapse and revival effect known from quantum optics @xcite .
a specific feature in these systems is a linear dependence of the collapse and revival time on the inverse coupling strength , i.e. @xmath76 and @xmath77 where @xmath27 usually denotes the coupling strength between the light field and the two - level atom in the quantum optical systems @xcite .
we verified numerically for our system that the observed decay and revival times obey a similar dependence with the interaction parameter @xmath27 as coupling strength in our model .
we define @xmath78 as the time when the difference between the maximal and average amplitude of the inter - band oscillations has fallen to @xmath79 , i.e. , @xmath80 . the revival time is taken as the maximum of the revived oscillations .
clearly demonstrates the linear dependence of these times on the inverse interaction strength . to find an effective description , we try to understand the most relevant interaction processes
clearly , the repulsive two - body interaction disfavours double occupancy of sites since this will always cost an interaction energy @xmath81 for two particles occupying the same site in the lower ( upper ) band .
starting from an initial state with population of the ground band only , the strong stark force leads to an occupation of the upper band .
doubly occupied sites are also suppressed there , but two particles may sit at the same site in either band , i.e. on top of each other " . indeed , the most important interaction term in the time - evolution is @xmath82 , since it already gives a non - zero contribution when there is only one particle per site in each of the two bands .
in fact , comparing the time evolution of the initial state with all interaction terms and only the one mentioned shows almost no difference ( cf . ) .
we focus on fillings close to unity @xmath83 and study the time - evolution of states of the form @xmath84 , which is not an eigenstate of the system in resonance . note
that this is the most important contribution to the superfluid ground state in an expansion in our configuration state fock basis . for large enough systems the superfluid ground state of the untilted system ( and the single - occupancy state @xmath85 likewise ) becomes indistinguishable from a coherent state ( eq . ( 66 ) in @xcite ) that factorises into a product of local coherent states at each site @xmath6 : @xmath86 |{\rm vac}\rangle = \prod_l \big ( \rme^{\sqrt{\bar n } a_l^{{\dagger } } } |{\rm vac}\rangle_l\big ) = \prod_l |\varphi ; 0\rangle_l \equiv |\boldsymbol \varphi ; 0\rangle.\ ] ] we denote this coherent state with phase @xmath87 by @xmath88 .
we are now going to re - write this state in the resonant basis and determine the effect of the interaction term @xmath89 when acting on this state .
we start by inserting the transformation @xmath90 into to obtain @xmath91 |{\rm vac}\rangle \nonumber\\ % & = { \rm exp}\bigg[\sqrt{\frac{n}{2l}}\sum_{l , n}j_{l - n}(x_a)(\mu_n^{\dagger}- \nu_n\dag ) \bigg ] |{\rm vac}\rangle \nonumber\\ & = \prod_n \rme^{\sqrt{\bar n/2}\;\mu_n^{\dagger } } \rme^{-\sqrt{\bar n/2}\;\nu_n^{\dagger } } |{\rm vac}\rangle .
% \equiv \prod_n\big|\frac{\varphi}{\sqrt{2 } } ; -\frac{\varphi}{\sqrt{2}}\big)_n.\end{aligned}\ ] ] where we used @xmath92 $ ] for @xmath93 @xcite and @xmath94=0 $ ] which follows from the properties of the operators @xmath95 .
let us denote fock states with single - particle basis @xmath96 by round parentheses @xmath97 .
from we see , that the coherent state of the lower band in resonance is a product of local coherent states for both bands in the resonance basis @xmath98 with @xmath99 .
the time - evolution of this state ( for the non - interacting system in resonance ) is simple since it is diagonal in the eigenbasis of the hamiltonian in resonance .
we continue to study the phase evolution created by the most important term @xmath89 perturbatively by expressing the operators @xmath100 in the @xmath101-basis : @xmath102 |\boldsymbol{\tilde\varphi } ; -\boldsymbol{\tilde\varphi } ) = \exp\bigg[\frac{\rmi gw_x t } { 2}\sum_{l}\sum_{l_1,\ldots , l_4 } j_{l - l_1}(x_a)j_{l - l_2}(x_a)j_{l - l_3}(x_b ) \times \nonumber\\ j_{l - l_4}(x_b ) ( \mu_{l_1}^{\dagger } - \nu_{l_1}^{\dagger})(\mu_{l_2 } - \nu_{l_2})(\mu_{l_3}^{\dagger}+ \nu_{l_3}^{\dagger})(\mu_{l_4 } + \nu_{l_4})\bigg ] |\boldsymbol{\tilde\varphi } ; -\boldsymbol{\tilde\varphi}).\end{aligned}\ ] ] since @xmath103 is a product of local coherent states we can ignore the sum over @xmath6 and discuss the expected behaviour .
firstly , both @xmath104 are much smaller than unity and the main contribution in the sums over @xmath105 will come from the zeroth - order bessel functions @xmath106 . secondly , the product of operators gives @xmath107 different combinations of the field operators .
but due to the prefactors , the combinations with equal indices are the most important .
they simply give an integer number when applied to the product of local coherent states they are acting on . taking these two points together , the time evolution of this state should show an approximate revival at @xmath108 this result is valid for large systems and can not account for the effect of non - universal properties like a limited number of particles and lattice sites , but we expect it to yield the right order of magnitude for finite systems , and in particular the correct scaling with the parameters of external potentials ( c.f .
, inset , below ) .
additional finite size corrections to can be understood by using the decomposition into the eigenbasis .
the evolution of the occupation of the upper band certainly depends on the initial state , which we can take into account by studying the weights @xmath109 of the initial state expanded in the eigenbasis .
the result of a numerical diagonalisation for a system in resonance is depicted in , where the absolute values of the expansion coefficients with their corresponding quasi - energies are shown .
for vanishing two - body interaction strength @xmath110 , the quasi - energies from the states with different occupation numbers are degenerate as expected .
the energy difference between neighbouring lines of constant quasi - energies corresponds to the time - scale of the resonant inter - band oscillations and follows from the diagonalisation of the resonant non - interacting system as @xmath111 . in the non - interacting system ( @xmath110 )
, two coefficients are dominating and the difference of the quasi - energies yields a single time - scale @xmath112 .
when the interaction is turned on , the weight of states with many contributions from double- or higher occupancies of sites decrease significantly ( since they are energetically disfavoured ) and their quasi - energies are slightly shifted .
but , surprisingly , only a limited number of additional coefficients @xmath109 contributes significantly in the eigenbasis expansion even for @xmath113 .
the observed collapse and revival signal is now determined by a few expansion coefficients that are much larger than the others .
if we focus on the three largest coefficients , denoted as @xmath114 and sorted by their quasi - energies @xmath115 , we find that the latter are shifted by the interaction by different amounts . the differences between neighbouring quasi - energies , @xmath116 and @xmath117 ( shown for the example @xmath118 in ) , lead to a beating between two oscillations with periods @xmath119 and the revival time will thus be given by @xmath120 this estimate requires a numerical diagonalisation but gives a clear physical interpretation to the revival time observed in a specific realisation with @xmath42 atoms on @xmath7 lattice sites .
thus , taking and together , we have an understanding of the general physical mechanism triggering the collapse and revival of the resonant two - band oscillations .
we compare our prediction for the revival time , and , to actual numerical simulations in .
we find that gives the right order of magnitude for the revival in a specific realisation , and , in particular , it shows the correct scaling behaviour with the depth of the optical lattice ( shown in the inset of ) .
additionally , shows that the corrections for specific sizes of the system are well accounted for by , which only slightly underestimate the revival times by a few percent .
this deviation could be corrected by including more than three participating states , extending in this manner the arguments which lead to .
+ in an experimental realisation , the shorter the collapse time @xmath78 the easier it could be observed , and for an estimate we make use of the fact @xcite that the collapse time is proportional to the revival time @xmath121 where @xmath122 denotes the width of the distribution of coefficients @xmath109 necessary to expand the initial state in the eigenbasis . for the specific example of @xmath123 and @xmath124
we find @xmath125 for the width of the distribution ( i.e. the @xmath126 couplings include just one or very few additional states as in the derivation of above ) , such that we estimate @xmath127 compared to @xmath128 found numerically . in this work ,
we focused on realisations with ultra - cold atoms in a tilted periodic potential , and the observed effect can be manipulated by engineering the potential @xcite or the two - body interaction @xcite .
specifically , the revival time depends sensitively on parameters as the hopping strength and the external force close to the zeros of the bessel function in .
this is analogous to already realised manipulations by time - dependent forces as predicted by @xcite and realized in @xcite .
the observation of bloch oscillations over thousands of periods and a fine control on the two - body interaction have already been demonstrated experimentally @xcite .
therefore , the collapse and revival of resonant inter - band oscillations predicted here should be accessible in such state - of - the - art experiments .
we finally remark that the collapse and revival phenomena discussed in this section have their origin ( cf . ) in the degradation ( due to interactions ) of single particle _
inter - band _ oscillations .
so , even if there are analogies to the collapses and revivals observed in other experiments @xcite , the collapse - revival oscillations reported there arise from the interaction within a single - band .
therefore , in contrast to our results , those oscillations would not at all occur when the lower band interaction is suppressed , equivalent to @xmath129 in the model here discussed .
_ summary .
_ we studied the coupling between two energy bands in a two - band bose hubbard model with an additional tilting force .
the force can lead to a strong coupling of the bands and we found strong resonances in the inter - band oscillations in this lattice model .
furthermore , the two - particle interaction leads to a collapse and revival of the resonant inter - band oscillations . here , we made predictions for the relevant time scales which were verified numerically .
+ a closed two - band system is an idealisation , but it can approximately be realised in various parameter regimes with ultra - cold atoms @xcite using different techniques as e.g. super lattices @xcite .
in addition , the use of feshbach resonances allows a complete control and fine tuning of the interaction strength @xcite needed to test our predictions .
our work is a first important step in going beyond ground - band physics and accessing more degrees of freedom in bosonic ultra - cold gases .
this work was supported within the framework of the excellence initiative by the german research foundation ( dfg ) through the heidelberg graduate school of fundamental physics ( grant number gsc 129/1 ) , the frontier innovation fonds , and the global networks mobility fund .
p.p . is grateful to andrea tomadin and acknowledges support from the klaus tschira foundation . |
-8pt the stated systematic error in the batse 3b catalog@xcite burst locations is @xmath2 .
this value was estimated by taking the rms deviation of batse positions from known positions of 36 bursts , determined by the ipn , watch , and comptel@xcite .
unfortunately , many of these same known positions were used to calibrate the batse burst location software , i.e. , as a guide in determining what effects to include in the burst location algorithm . in order to calibrate properly the batse burst position errors , an independent set of burst locations
is needed .
such a set exists , in the form of 196 bursts for which single ipn arcs exist@xcite . in this paper
we analyze the systematic errors in the positions of bursts in the batse 1b , 2b and 3b catalogs using these 196 bursts .
our analysis is based upon the likelihood approach . -8pt
-8pt we assume that the systematic error @xmath0 and statistical error @xmath1 in position are circular gaussians , to be added in quadrature , as prescribed by the batse catalog .
the circular approximation should be good , since roughly @xmath3 of all batse bursts have @xmath4 positional contours that are nearly circular@xcite , and the bursts with ipn arcs tend to be those with larger fluences .
we further assume that the sky is flat ( i.e. , that the size of the total error in the batse burst position is not too large ) , and that the ipn arcs have zero width .
the former is a good approximation , since the largest batse @xmath5 among the bursts with ipn arcs .
the latter is a good approximation since the characteristic width of the 3-@xmath6 contours of the ipn arcs is a few arcminutes , which is much less than the smallest batse @xmath7 .
figure 1 .
configurations for a burst with a known position ( left diagram ) and for a burst with a single ipn arc ( right diagram ) .
the likelihood function for bursts with known positions ( e.g. , two intersecting ipn arcs ) is given by @xmath8 whereas the likelihood function for bursts with single ipn arcs is given by @xmath9 here , @xmath10 is the deviation of the batse position from the known position , @xmath11 is the perpendicular deviation of the batse position from the ipn arc ( see figure 1 ) , and @xmath12 .
the likelihood function @xmath13 allows exploration of any model that we may have for @xmath0 .
here we focus on three models : ( 1 ) @xmath0 = a constant , ( 2 ) @xmath14 erg @xmath15 , and ( 3 ) @xmath16 .
the first model has one free parameter ( @xmath0 ) ; the second and third models have two free parameters ( scale factor @xmath17 and power law index @xmath18 ) . -8pt
-6pt fitting the constant @xmath0 model to the 18 batse 3b bursts with ipn positions , we find a maximum likelihood ( ml ) value @xmath19 .
this is consistent with the value of @xmath20 quoted in the batse 3b catalog@xcite , which was found using these bursts and some others . with the value of @xmath20 quoted in the batse 3b catalog , as shown by the probability distribution in @xmath0 ( see figure 2 ) and comparison of the expected and observed distribution of normalized deviations @xmath22 assuming @xmath2 ( see figure 3 ) . figure 4 .
upper left - hand and lower and upper right - hand panels : 1-@xmath6 , 2-@xmath6 , and 3-@xmath6 contours in the ( @xmath23)-plane for the burst samples and data as labeled , assuming the @xmath1-dependent model .
lower left - hand panel : same as figure 2 , except for the @xmath24-dependent model . fitting the third model , in which @xmath29
, we find ml values @xmath30 and @xmath31 ( see figure 4 ) .
the probability @xmath32 , showing that the data strongly prefer the power - law model over the ml model with constant @xmath33 .
these results imply that @xmath0 is strongly correlated with @xmath34 and @xmath1 .
for example , @xmath35 while @xmath36 .
this is illustrated in figure 5 , which shows the ml @xmath34-dependent and @xmath24-dependent models .
figure 5 .
left - hand panel : the ml @xmath34-dependent model .
right - hand panel : the ml @xmath1-dependent model .
the points are _ not _ data points , but rather are the ml _ constant _ @xmath0 model for the given interval in @xmath34 or @xmath1 .
it is also interesting to examine the character of @xmath0 for subsets of the 3b catalog , in particular the 1b sample of bursts in which evidence for repeating was found .
fitting the three models to the `` old '' 1b positions , we find @xmath37 for the ml constant @xmath0 model , and that the data do not request a more complicated model(see figures 3 and 4 ) .
in contrast , fitting the 3b positions for the 1b sample of bursts , we find that the data request the @xmath1-dependent model at the 95% confidence level [ e.g. , @xmath38 ( again see figures 3 and 4 ) .
these results imply that for the 1b sample of bursts , the new ( 3b ) @xmath0 s are smaller than the old ( 1b ) ones for bursts with @xmath39 erg @xmath40 or with @xmath41 .
however , for the @xmath42 80% of bursts with lower fluences or larger @xmath1 s , they imply that the new @xmath0 s are larger than the old ones .
we are grateful to kevin hurley for providing us with the 3rd catalog of ipn arcs , without which this analysis would not have been possible .
we also wish to thank geoff pendleton and michael briggs for discussing batse burst location issues with us . | we analyze the systematic errors in the positions of bursts in the batse 1b , 2b and 3b catalogs , using a likelihood approach .
we use the batse data in conjunction with 196 single ipn arcs .
we assume circular gaussian errors , and that the total error is the sum in quadrature of the systematic error @xmath0 and statistical error @xmath1 , as prescribed by the batse catalog .
we find that the 3b burst positions are inconsistent with the value @xmath2 stated in the batse 3b catalog . -8pt |
it is well - known that extending string theory off - shell is a hard problem . in the first quantized string theory based on an action of a single string ,
the guiding principle is a local weyl invariance of string worldsheets .
this immediately translates into on - shellness of string scattering amplitudes .
off - shellness alone does not suggest any particular symmetry and one has to consider insertions of operators carrying dimensions of arbitrarily high degrees on the bulk of the string worldsheets : this is a hard task to carry out .
situation can be more manageable to handle if the local weyl invariance is broken only on the boundaries of the worldsheets . to date
there are two proposals for off - shell string theory through deformation on the boundary of the string worldsheets .
the one is the boundary string field theory ( bsft ) of witten @xcite on a unit disc and the other is the sigma model approach of open strings to the spacetime action @xmath0 .
( see for example ref . @xcite and the references therein . ) to put simply , @xmath1 where @xmath2 is a partition function for a single string .
the former proposal was successfully applied to the problem of open string tachyon condensation @xcite .
it is well known that open strings alone do not serve as perturbatively stable system in flat spacetime @xcite .
a possible next step common to these two approaches will be to extend the idea to the other worldsheet geometry which shares the same euler number as that of a disc , namely @xmath3 .
this amounts to considering the string field theory of unoriented open and closed strings at least at the linearized level @xcite .
this was considered and developed in refs .
@xcite .
the above thoughts have also suggested that the off - shell extension of boundary states @xcite and that of crosscap states @xcite are an expedient tool and can be used in more general contexts . in particular , the computation of @xmath4 function is equivalent to the determination of the normalization of these states .
( see refs . @xcite and the references therein . )
bsft @xcite is defined on a disc worldsheet .
let us consider an unit disc @xmath5 with worldsheet action i^disc = _ d^2 z x^ |x _ + _ ( ) , [ disc - action ] where @xmath6 is a generic scalar operator with ghost number @xmath7 let us consider the operator @xmath8 located on the boundary , where @xmath9 is the tangent component of the ghost field along the boundary .
the defining differential equation for @xmath10 is written as & & tr ^ ( ) \ { q_brs , o } ( ^ ) _ v , disc^ghost , + o ( ) & = & c^ ( ) v ( ) = _ _ c^ ( ) v^ ( ) = _ _ o^ ( ) , where @xmath11 denotes the trace over the chan - paton space .
the brs charge @xmath12 corresponds to a fermionic vector field @xmath13 which is a basic ingredient of the formalism of bsft . here
@xmath14 is the unnormalized path integral with respect to the worldsheet action ( [ disc - action ] ) , and can be represented as a matrix element @xmath15 .
the bra vector @xmath16 obeys [ bcondition ] b .
( ( z + |z ) x^ + ) |_z = e^i , |z= e^-i = 0 .
we refer to @xmath17 as an off - shell boundary state ( obs ) of the disc .
the condition ( [ bcondition ] ) is a consequence from the correspondence between the matrix element and the end point of the path integral .
the latter one obeys the boundary condition derived from the worldsheet action @xmath18 on the disc .
( [ bcondition ] ) tells us that , at the initial point of the coupling constant flow , the system obeys the neumann boundary condition , and the end point of the flow is described by zero of @xmath19 , namely the dirichlet boundary condition .
we now proceed to construct the @xmath3 extension @xcite of bsft .
@xmath3 is a non - orientable riemann surface of euler number one with no hole , no boundary and one crosscap .
we construct the @xmath3 worldsheet on a complex @xmath20-plane by using an involution where we identify @xmath20 and @xmath21 .
we choose the fundamental region @xmath22 to be @xmath23 .
the crosscap @xmath24 , the non - trivial closed loop of the @xmath3 worldsheet , is represented as half of unit circle @xmath25 in this case .
let us consider the @xmath3 worldsheet with the following action : i^rp^2 = _
^ -c d^2 z x^ | x _ + _ c ^ ( ) . [ rp2-action ] our first objective is to obtain the operator which corresponds to the fermionic vector field @xmath13 for the @xmath26 case .
let us consider the following operator : @xmath27 where the subscript @xmath28 implies that the modes are restricted to the even ones .
we can show that @xmath29 and @xmath30 , where @xmath31 is the normal component of the ghost field along the loop .
( @xmath32 is understood . )
we also obtain @xmath33 .
thus , the operator @xmath34 generates the brs transformations @xmath35 associated with the diffeomorphisms in the @xmath36 direction on @xmath32 .
it is now immediate to carry out the action of @xmath34 on a generic operator with ghost number one .
this also brings us an operator which is `` on - shell '' with respect to @xmath34 , that is , invariant under @xmath35 .
let @xmath37 be a scalar operator with ghost number one , @xmath38 , where @xmath39 , @xmath40 are generic scalar relevant operators .
we can show @xmath41 which is the condition we need in our formalism . following the disc case ,
we introduce a defining differential equation for @xmath42 & & _ c _ c ^ ( ) \{q_brs^ ( ) , o } ( ^ ) _ v , rp^2^ghost + & = & 2 _ c _ c ( - ) c_1 c_0 c_-1 _ rp^2 f(, ) , where @xmath43 , @xmath44 .
note that the allowed form of the nonvanishing ghost three point function has selected @xmath39 alone and @xmath40 has disappeared . here
the unnormalized path integral @xmath45 is evaluated with respect to the action ( [ rp2-action ] ) , and can be represented as a matrix element between the ket vector of the closed string vacuum and the bra vector of the off - shell crosscap state ( ocs ) @xmath46 , namely @xmath47 .
details of the ocs will be given in the next section .
in this section , we consider the @xmath3 worldsheet with background dilaton field @xmath48 from the viewpoint of the sigma model approach @xcite .
the basic idea of the sigma model approach is that the spacetime action for string fields is essentially the renormalized partition function of the worldsheet with corresponding background string fields . in the case of the disc for bosonic strings , for example , we need corrections of the disc partition function in order to subtract the divergence from the mbius infinity @xcite . however , the mbius group of @xmath3 is @xmath49 whose volume is finite , and we have no mbius infinity from the @xmath3 worldsheet .
therefore it is natural to assume that the partition function of the @xmath3 worldsheet with background fields itself is the exact loop correction term from the @xmath3 graph in the presence of the quadratic background fields . to begin with
, we set the metric inside the unit circle ( @xmath50 ) on the complex plane as @xmath51 , @xmath52 .
the metric outside the unit circle ( @xmath53 ) is obtained by the involution @xmath54 , @xmath55 as @xmath56 , where @xmath57 for @xmath58 .
the worldsheet curvature @xmath59 is then given by @xmath60 , and we obtain @xmath61 here @xmath4 is the determinant of the worldsheet metric written in the polar coordinate @xmath62 .
note that eq .
( [ dilaton - rp2 ] ) gives the correct euler number of the @xmath26 ( which is one ) if we set @xmath63 .
therefore the contribution of the background dilaton concentrates on the crosscap with the above gauge choice .
let us consider the @xmath3 worldsheet @xmath22 with the following action : @xmath64 where @xmath65 .
note that the worldsheet action is free in the bulk " region @xmath66 in this gauge choice .
the aim of this subsection is to extend the on - shell crosscap conditions into the case @xmath67 .
we should find , in other words , the constraints on the closed - string modes in the neighborhood of @xmath24 in the presence of interaction @xmath48 .
we call these constraints off - shell crosscap conditions .
we assert that the off - shell crosscap conditions can be written as @xmath68 where @xmath69 the right - hand side of eq .
( [ k ] ) indicates the meaning of the off - shell crosscap conditions ; eq .
( [ gen - cross - cond ] ) are the conditions so that the @xmath70 in the neighborhood of @xmath24 , as well as its image by the involution , connects smoothly with the @xmath70 on @xmath24 which obeys the equation of motion @xmath71 .
has no boundary , @xmath72 which comes from the total derivative survives only on the crosscap due to the involution . ]
the off - shell crosscap conditions ( [ gen - cross - cond ] ) are rewritten in terms of closed - string modes as @xmath73 where we do not sum over @xmath74 .
we can easily check that the off - shell crosscap conditions interpolate between the usual on - shell crosscap conditions and their t - duals .
we define ocs @xmath75 as @xmath76 , @xmath77 .
the explicit form of @xmath75 is given as @xmath78 where @xmath79 , and @xmath80 .
we can easily check that this ocs becomes ( the t - dual of ) the usual on - shell crosscap state if we take the limit @xmath81 ( @xmath82 ) .
therefore the ocs naturally interpolates between the crosscap state for a higher - dimensional o - plane and that for a lower - dimensional o - plane . the coupling constant @xmath83 , which is a parameter of the configuration of the background dilaton field , controls the dimension of the corresponding o - plane .
next , we show that the ocs is a useful tool to evaluate the quantities on the @xmath3 worldsheet . for example
, we can calculate the green s function and the partition function on the @xmath3 worldsheet in the presence of interaction @xmath84 on the crosscap .
let us consider one - dimensional target space and omit the superscript @xmath74 of @xmath85 and @xmath86 for simplicity .
the green s function for this case is given by @xmath87 the expectation value of composite operator @xmath88 which is defined as @xmath89 , \label{def - x2}\end{aligned}\ ] ] is obtained by substituting @xmath90 and @xmath91 into eq .
( [ green - rp2 ] ) : @xmath92 , \label{x2}\end{aligned}\ ] ] where @xmath93 and we have written the constant in eq .
( [ def - x2 ] ) as @xmath94 by using a positive constant @xmath95 .
the value for @xmath95 is ambiguous at this stage and depends on the renormalization scheme .
we will determine the value for @xmath95 later .
we can calculate the partition function of the @xmath3 worldsheet by using the relationship @xmath96 , and then we obtain @xmath97 up to the overall normalization factor . in general , the partition function for 26-dimensional target space with @xmath65 on the crosscap can be written as @xmath98 up to the overall normalization factor .
obs s are useful tools for analyses of two - dimensional systems with boundaries . in the case
the boundary interaction is quadratic , the proper form of the green s function and the @xmath4-function has been given by using obs @xcite . in the determination of the @xmath4-function which is defined as @xmath99 , the overall normalization of the obs is crucial .
the normalization can be fixed by computing the partition function of the cylinder worldsheet , @xmath100 , and demanding @xmath101 , without taking the long cylinder limit @xcite . here
@xmath102 is the hamiltonian on the cylinder of length @xmath103 , and @xmath104 and @xmath105 indicate the boundaries .
the left - hand side is calculated via path integral with @xmath106-function regularization .
for example , the @xmath4-function for bosonic theory ( [ disc - action ] ) with the quadratic boundary interaction @xmath107 is given by @xmath108 for supersymmetric extension with boundary mass , the fermionic sector gives @xmath109 which agree with the results of refs .
@xcite .
the descent relation among o - plane tensions can be also derived by using the result of sec .
[ rp2 ] @xcite .
let us define quantity @xmath110 as @xmath111 , where the limit is taken as @xmath112 and @xmath113 .
we do not touch the parameter @xmath114 here . according to the argument in sec .
[ rp2 ] , @xmath110 is equal to @xmath115 where @xmath116 and @xmath117 are the volume and the tension of an o@xmath118-plane . here , the dimension of the o - plane is @xmath119 and is defined as the number of parameters @xmath83 which are taken to zero .
then we can write @xmath120 we find @xmath121 where @xmath122 is the zero mode of @xmath123 and we have correctly extracted the integral of the zero - mode part of @xmath124 @xcite .
here we have assigned the value 2 to @xmath95 because we can obtain a finite and non - zero value of @xmath125 in the limit @xmath126 if and only if @xmath127 .
in other words , we have chosen the renormalization scheme in eq .
( [ def - x2 ] ) so that we obtain a finite and non - zero value of @xmath125 in the limit @xmath126 .
we therefore obtain @xmath128 this is precisely the ratio of the tension of an o@xmath129-plane and that of an o@xmath130-plane . in general
, we can show in a similar manner that @xmath131 .
off - shell crosscap states as well as off - shell boundary states are useful tools both for formulating off - shell theory ( worldsheet string field theory or off - shell non - linear sigma model ) of unoriented open and closed strings , and for performing computation . in configuration space of strings ,
interpolations of o - planes of various dimensions are certainly possible as are in d - branes , and the dilaton background is responsible for this .
one important physical issue is how these interpolations are materialized as physical processes in systems with moderate instabilities .
a.a . gerasimov and s.l .
shatashvili , jhep * 0010*(2000)034 , hep - th/0009103 ; d. kutasov , m. mario and g. moore , jhep * 0010*(2000)045 , hep - th/0009148 .
see also s. moriyama and s. nakamura , phys .
lett . * b506*(2001)161 , hep - th/0011002 , for example . b. grinstein and m. b.
wise , phys .
* d35*(1987)655 ; h. itoyama and p. moxhay , nucl . phys .
* b293 * ( 1987 ) 685 ; y. arakane , h. itoyama , h. kunitomo and a. tokura , nucl . phys . * b486*(1997)149 , hep - th/9609151 .
t. lee , phys . rev . *
d64*(2001)106004 , hep - th/0105115 ; a. fujii and h. itoyama , hep - th/0105247 ; m. laidlaw and g. w. semenoff , hep - th/0112203 ; e. t. akhmedov , m. laidlaw and g. w. semenoff , hep - th/0106033 .
h. itoyama and t. oota , hep - th/0206123 .
o.d . andreev and a.a .
tseytlin , nucl . phys .
* b311*(1988)205 ; a.a .
tseytlin , phys . lett .
* b208*(1988)221 ; a.a .
tseytlin , phys . lett . * b194*(1987)63 .
r. chatterjee , mod .
phys . lett .
* a10*(1995)973 , hep - th/9412169 ; a. leclair , g. mussardo , h. saleur and s. skorik , nucl . phys . * b453*(1995)581 , hep - th/9503227 . | the authors recent works on off - shell boundary / crosscap states are reviewed . |
the periodic anderson model ( pam ) gives a schematic description of very important systems with strongly correlated electrons , and there are several recent reviews on this subject @xcite .
the model consists of a lattice with two localized electronic states at each site , strongly correlated by a coulomb repulsion @xmath2 , plus a band of uncorrelated conduction electrons ( c - electrons ) that hybridize with the localized electrons ( f - electrons ) .
the cumulant expansion of the pam discussed in this work is a perturbative expansion around the atomic limit , and is an extension@xcite of the method originally employed by hubbard @xcite to study his well known model of correlated itinerant electrons@xcite .
he introduced operators @xmath3 that transform the local state @xmath4 at site @xmath5 into the local state @xmath6 at the same site , and developed a diagrammatic method that circumvents the fact that the @xmath7 operators are neither fermions nor bosons .
his work was the first application of the cumulant expansion to a quantum system of fermions@xcite and has several desirable properties : it seems to be a natural extension of the usual diagrammatic techniques , it does not have excluded site problems in the lattice summations and it is possible to derive a linked cluster expansion for the grand canonical potential ( see ref .
@xcite for more details ) . at the same time
, the use of @xmath7 operators allows to work in a subspace of the whole space of local states , that contains only the states that are relevant to the problem of interest .
it seems therefore useful to understand better this method , and the main purpose of the present paper is to study the properties of the cumulant expansion for the pam in the limit of infinite spatial dimensions ( @xmath8 ) .
metzner and vollhardt@xcite have recently applied the limit @xmath8 to strongly correlated fermion systems , providing non - trivial models of the hubbard type that are substantially simpler to analyze .
this development has stimulated many new works , and we could mentioned the very successful dynamical mean field theory@xcite , that uses the local properties of the self energy for @xmath9 as a starting point .
several methods have been employed to study the pam in this limit@xcite , and the cumulant expansion of the hubbard model was reconsidered by metzner@xcite , showing that there is an important reduction in the number and type of cumulant diagrams that appear in the expansion when @xmath9 in an hypercubic lattice . in the hubbard model there is no hybridization and the hopping hamiltonian is used as perturbation , while the pam is usually employed to model systems with a bandwidth much larger than the hybridization energies , and it is then preferable to diagonalize the conduction band and use the hybridization hamiltonian as perturbation .
it was then necessary to make substantial modifications to metzner s derivation in order to apply it to the pam when @xmath9 . in the following sections two types of hybridization models are considered : the purely `` local '' hybridization and the `` nearest neighbor '' ( n.n . )
one . in the n.n .
hybridization there are electronic transitions between localized ( `` f '' ) states at a site and conduction wannier states at the nearest neighbor sites , while in the local hybridization those transitions only occur at the same site .
the same type of diagram cancellation obtained by metzner for the hubbard model is obtained for these two hybridization models in the pam . as the local hybridization between @xmath10 and @xmath11 or @xmath12 electrons should vanish when the site has inversion symmetry , the n.n .
hybridization discussed in this paper is more realistic than the local one , because it does not necessarily vanish for that type of site . the local approximation is usually employed because it makes the calculation simpler ; and the changes introduced by an interatomic hybridization have been already discussed in the case of a diatomic - molecule model@xcite .
let us consider free electrons in an hypercubic lattice of dimension @xmath13 , described by a hamiltonian@xcite @xmath14 with translational invariance .
the position of the site @xmath15 is given by @xmath16 , where @xmath17 is the lattice parameter and the @xmath18 are integers .
hopping shall be considered , so that @xmath19 is non - zero and equal to @xmath20 when the components @xmath21 of @xmath22 satisfy @xmath23 for @xmath24 . employing the wannier transformation @xmath25\quad , \label{e2.2}\ ] ] with @xmath26 satisfying cyclic boundary conditions , one obtains @xmath27 where @xmath28 the electronic green s functions ( gf ) for imaginary time are defined by@xcite @xmath29 where @xmath30\ c_{{\bf % k,\sigma } } \ \exp [ -h_{c}\tau ] $ ] and the subindex @xmath31 in eq.([e2.5 ] ) is the usual chronological ordering of fermi operators , with @xmath32 increasing to the left . to emphasize that the hybridization was not considered in these gf
, they will be denoted with @xmath33 , and after fourier transformation in the imaginary time @xmath32 they are given by@xcite @xmath34 . to obtain a finite average energy per site
when @xmath35 , it is necessary to renormalize @xmath36 , taking a finite @xmath37 in the non - zero @xmath38 ( i.e. when @xmath39 and @xmath40 are n.n . ) . in that limit
the energy density of states is given by@xcite @xmath41 \quad , \label{e2.6}\ ] ] and @xmath42 the exact solution of @xmath36 is used in the present treatment of the pam , employing the hybridization as perturbation .
to study the behavior of the cumulant expansion in the limit @xmath9 , it is convenient to consider the direct space gf , given by @xmath43\quad , \label{e2.8}\ ] ] where @xmath44 , as well as its dependence with the minimum number @xmath45 of n.n
. jumps necessary to go from @xmath22 to @xmath39 .
when the gf are obtained for large @xmath11 using the hopping as a perturbation , one immediately shows that @xmath46,where @xmath47 .
a similar result is obtained for the `` exact '' @xmath48 of eq .
( [ e2.8 ] ) , but the discussion will be postponed to a latter section , in which a more general property will be shown .
the hamiltonian of the whole system is @xmath49 where @xmath50 is the hamiltonian of the conduction electrons ( c - electrons ) discussed in section [ sec2 ] .
the second term @xmath51 describes independent localized electrons ( f - electrons ) , where a simple index @xmath5 has been used to indicate the sites @xmath39 .
the last term is the hybridization hamiltonian@xcite , giving the interaction between the c - electrons and the f - electrons : @xmath52 the state space of the f - electrons , at a given site j , is spanned by four states : the vacuum state @xmath53 the two states@xmath54 of one f - electron with spin component @xmath55 and the state @xmath56 with two electrons of opposite spin . in the limit of infinite electronic repulsion ( @xmath0 ) considered here , the state @xmath57 is always empty , and can be projected out of the space .
the operator @xmath58 destroys the electron in state @xmath54 leaving the site in the vacuum state @xmath59 with energy @xmath60 , and its hermitian conjugate @xmath61 reverses that process .
the hubbard operators @xmath62 are not usually fermi or bose operators , and a product rule : @xmath63 should be employed when the two operators are at the same site . for different sites one should first classify the operators in two families : @xmath62 is of the `` fermi type '' when @xmath4 and @xmath64 differ by an odd number of fermions , and it is of the `` bose type '' when they differ by an even number of fermions . at different sites , two x - operators of the fermi type anticommute , and they commute when at least one of them is of the bose type . to complete the definition ,
the @xmath62 should anticommute ( commute ) with the @xmath65 and @xmath66 when they are of the fermi type ( bose type ) .
sometimes it will be convenient to employ @xmath67 to describe the hubbard operators , where @xmath68 , and @xmath69 gives the transition @xmath70 that destroys one electron , i.e. with @xmath71 having one electron less than @xmath72 as in ref . , and it is often convenient to abbreviate @xmath73 . the index @xmath74 is then introduced to describe the inverse transition , so that @xmath75 when @xmath76 and @xmath77 when @xmath78 , and for the pam with @xmath0 there are only two possible transitions @xmath79 .
the grand canonical ensemble ( gce ) of electrons is employed in this problem , and it is useful to introduce @xmath80 where @xmath81 is the number of electrons in state @xmath72 and @xmath82 is the chemical potential . as usual @xmath83
is split into @xmath84 where @xmath85 will be the perturbation hamiltonian , and the exact and unperturbed averages of any operator @xmath86 will be denoted respectively by @xmath87 and @xmath88 .
it is also convenient to denote the exact or heisenberg `` @xmath32 evolution '' of any operator @xmath86 with @xmath89\ a\ \exp [ -{\cal h}\tau ] \label{e3.5a}\ ] ] and employ @xmath90\ a\ \exp [ -{\cal h% }
_ { 0}\tau ] $ ] for the unperturbed case .
it is convenient to use the expressions @xmath91 because the energies appear almost always in those combinations .
the c - electron gf employed in the cumulant expansion are : @xmath92 where the subindex c was added to differentiate them from the unperturbed gf of the f - electron . @xmath93 where @xmath94 .
the general hybridization coefficients are given by ( cf .
( 2.3 ) in ref . )
@xmath95 the hybridization is purely local when @xmath96 is independent of @xmath26 , because using the wannier transformation ( cf .
eq.([e2.2 ] ) ) in eq .
( [ e3.2 ] ) it follows that @xmath97 the @xmath98 that corresponds to n.n .
hybridization is @xmath99 where the vectors @xmath100 give the position of the @xmath11 n.n .
sites of the origin , and the corresponding @xmath101 is immediately obtained : @xmath102 in figure [ f1 ] are shown some of the infinite diagrams that contribute to the exact gf @xmath103 .
the full circles ( f - vertices ) correspond to the cumulants of the f - electrons and the empty ones ( c - vertices ) to those of the c - electrons . each line reaching a vertex
is associated to one of the @xmath7 operators of the cumulant , and the free lines ( i.e. those that do not join an empty circle ) correspond to the external @xmath7 operators appearing in the exact gf . an explicit definition of the cumulants can be found in the references @xcite , and they can be calculated by employing a generalized wick s theorem@xcite . the first diagram in figure [ f1]a corresponds to the simplest free propagator @xmath104 , and the second diagram in that figure has an empty circle that corresponds to the conduction electron cumulant , equal to the free propagator @xmath105 .
the interaction is represented by the lines ( edges ) joining two vertices and , because of the structure of the hybridization , they always join a c - vertex to an f - vertex ; the number of edges in a diagram gives its order in the perturbation expansion .
cumulants containing statistically independent operators are zero , and those appearing in the present formalism ( with the hybridization as perturbation ) vanish unless they contain only @xmath7 operators at the same site or only @xmath106 or @xmath107 operators with the same @xmath108 and @xmath55 .
the only non - zero c - cumulants are of second order , because the uncorrelated c - operators satisfy wick s theorem . on the other hand , the f - vertices can have many legs , all corresponding to @xmath7 operators at the same site , like the fourth and sixth order cumulants appearing in the rather more complicated diagram shown in figure [ f1]c .
[ sec3a]all the infinite diagrams that contribute to the gf of the f - electron with cumulants of at most second order are shown in figure [ f1]a , and they define the `` chain approximation '' ( cha ) when all the other diagrams are neglected .
the corresponding approximation for the gf of the c - electrons corresponds to the diagrams of figure [ f1]b .
the diagrams of the cha usually appear as part of more complicated diagrams , and it is then useful to analyze their behaviour when @xmath9 . in the cha ,
the gf is given in frequency and @xmath109 space by ( cf .
( 3.10 ) in ref .
@xcite ) @xmath110 and it would be useful to transform it back to real space to show the dependence with the distance @xmath111 between the two sites @xmath112 and @xmath5 , as it was done in eq .
( [ e2.8 ] ) for the conduction electrons .
because of the lattice translational invariance , it is enough to use the distance @xmath113 of the site @xmath5 to the origin , and characterize this site with an @xmath114 ( cf .
[ sec2 ] ) .
the set of indices @xmath115 with @xmath116 will be denoted with @xmath117 , while @xmath118=@xmath119 is the minimum number of n.n .
jumps necessary to go from the origin to the site @xmath5 . in both the local and the n.n .
hybridization models , the @xmath120 depends on @xmath109 only through the @xmath121 , and for any given site @xmath5 one can write @xmath122 where @xmath123 and @xmath124 . substituting this relation in eq .
( [ e3.14 ] ) one can expand it in a power series of all the @xmath125 with @xmath126 and then transform back to real space to obtain @xmath127 employing the relation @xmath128 which is a consequence of the cyclic boundary conditions , it follows that @xmath129 the gf for the conduction electrons in the cha , corresponding to the diagrams in fig .
[ f1]b , is given by : @xmath130 and employing the same derivation used above for the f - electrons it follows that @xmath131 .
the same relation is obtained in the absence of hybridization , as stated at the end of section [ sec2 ] for @xmath132 . to close this section
let us emphasize that the present expansion employs @xmath133 as perturbation , and that the exact solution of the conduction problem in the absence of hybridization is included in the zeroth order hamiltonian .
the contribution to the gf joining two sites separated by @xmath118 n.n .
jumps , that is of order @xmath134 , includes then contributions of any order in the @xmath50 .
it is because of this difference that the derivation of metzner@xcite for the hubbard model had to be modified for the present problem .
all the contributions of @xmath50 would disappear in the case of a band with zero width , but the electronic wave functions would still be extended for the model with n.n . hybridization .
in the present section it will be shown that the only diagrams that remain in the limit @xmath9 are those that are topologically `` fully two particle reducible '' ( f.t.p.r . ) .
these diagrams are defined as those in which any pair of vertices can be separated by cutting one or two edges@xcite , and are `` topologies constructed by linking polygons''@xcite .
two points should be made here : the first is that two different polygons ( also called loops or rings in what follows ) can have at most one vertex in common in the f.t.p.r .
topology ( see fig [ f2]a , b ) .
the second point , already stressed by metzner for the hubbard model,@xcite is that the topology of a diagram can be different from those of its possible embeddings in the lattice , because in the cumulant expansions there is no excluded site restriction in the lattice sums@xcite , and two different vertices of a diagram can occupy the same site ( see fig .
[ f2]c , d ) .
the property stated above refers to the topology of the embeddings , and in the sum of contributions of diagrams that are not f.t.p.r .
there may be terms that give non zero contributions because they correspond to a f.t.p.r .
topology of the embedding : the diagram in fig . 2c could contribute when @xmath135 , because the topology of its embedding , given by fig .
2d , is f.t.p.r . to prove this property
it is necessary to modify the derivation given by metzner for the hubbard model , for the same reasons given in section [ sec3a ] for the gf dependence on @xmath136 .
the proof of this property is given in the following two subsections , and it can be summarized as follows . consider first the diagram in fig .
[ f3]a , that corresponds to the sum of loops with all possible lengths .
all the embeddings of these diagrams are f.t.p.r . , and they give a non - zero contribution to the calculation of the free energy . as a second step , one can consider diagrams like those in figs .
[ f3]c , d , obtained by linking loops of any length at different f - vertices
. one could obtain all these diagrams by a procedure similar to the vertex renormalization@xcite , with the difference that only insertions of a rather special type are considered .
all these diagrams , as well as their embeddings , are f.t.p.r . , and a finite contribution of all of them is expected .
the next step is to consider diagrams that are not f.t.p.r .
, like the one shown in fig [ f3]e , that can be obtained by joining two loops ( like that in fig .
[ f3]a ) at two different f - vertices . if the family in fig .
[ f3]a _ _ gives a finite contribution , one can show that all the contributions to the diagram in fig .
[ f3]e will vanish when the topology of the embedding coincides with that of the diagram , i.e. when the embedding is not f.t.p.r .
, as it would happen when @xmath137 . in the special case of @xmath138 , ( cf . figs .
[ f2 ] c , d for a special case of this situation ) the embedding is f.t.p.r . , and the diagram contribution does not necessarily vanish . of the two sums over the lattice sites @xmath139 and @xmath5 , only one over @xmath138 remains in the limit @xmath35 , and this simplifies the calculation of this diagram . in particular , the restriction that in reciprocal space there must be conservation of @xmath26 at each vertex is removed@xcite , leaving only a single conservation of @xmath26 for all the edges joining the collapsed vertices : this makes the calculation of the diagram much simpler in the limit @xmath9 . to prove that the contribution of a f.t.p.r
diagram can be finite , let us consider the family of diagrams represented in fig .
their contribution does not vanish and can be expressed as @xmath140 where @xmath141 is a local cumulant@xcite and @xmath142 .
the abbreviations @xmath143 @xmath144 and @xmath145 @xmath146 will be used when there is no possibility of confusion .
the symbol @xmath147 corresponds closely to the `` self - field''@xcite @xmath148 , that was employed in ref . to renormalize vertices , but it gives only `` insertions '' obtained from simple loops of any length , as represented by the diagrams in fig [ f3]b ; by it definition , all the @xmath144 correspond to the same site , indicated by @xmath149 .
the notation @xmath150 has the same meaning , and reflects the fact that this cumulant is zero unless @xmath151 because @xmath152 and @xmath153 are statistically independent in the unperturbed system when @xmath154 .
note that @xmath155 depends explicitly on the parameters @xmath156 ( where @xmath157 represents @xmath158 for @xmath159 ) through the two hybridization constants @xmath160 associated to the edges joining the insertion vertex . to consider the addition of a simple loop to any f - vertex of a diagram , it is convenient to consider first the simplest possible case , shown in fig .
[ f3]c : its contribution is obtained by substituting the cumulant @xmath161 in eq . [ e4.1 ] by @xmath162 when the loop is added to a site @xmath149 that is already joined by @xmath163 edges , the corresponding cumulant @xmath164 suffers a similar substitution . repeated application of this procedure at all f vertices
gives all the possible f.t.p.r diagrams , and as the cumulants are independent of the lattice dimension @xmath11 and of the site @xmath5 , this procedure should not affect the order of the contribution with respect to @xmath11 .
the simple loops of fig .
[ f3]a give a finite contribution , so that the contribution of any f.t.p.r diagrams is then of @xmath165 .
a similar type of procedure can be applied to the cha diagrams of the one - particle gf , by successive decoration of the f - vertices with any number of insertions corresponding to the @xmath155 discussed above .
it is then clear that these diagrams will still be f.t.p.r . , and that the corresponding gf joining two sites separated by @xmath12 n.n .
jumps will be of order @xmath166 as it was shown in section [ sec3 ] for the cha . to analyze the diagrams that are not f.t.p.r .
, consider any one of them as a `` mother '' diagram , and split it into two or more f.t.p.r linked `` daughter''diagrams without any edges in common but such that any of them has at least two vertices in common with another daughter diagram , as well as two edges arriving at each of the common vertices .
it is clear that any pair of vertices that are common to two daughter diagrams can not be separated in the mother diagram by cutting one or two edges , and one says that they are not `` two particle reducible '' ( t.p.r . ) . from all the daughter diagrams
choose one as a vacuum diagram and transform the remaining ones into gf by adding external lines of the bose type ( as indicated in appendix [ apa ] ) to all the vertices that each of them had in common with any other daughter diagram in the mother diagram .
three examples of this procedure are given in fig [ f4 ] .
consider now the real space calculation of the contribution of the daughter diagrams . in appendix [ apa ]
it is shown that by joining one daughter gf to another diagram and fixing the position of the common vertices , one obtains the order of the contribution of the resulting diagram as the product of those of the building diagrams . in appendix [ apb ] it is shown that unless all the external f - vertices of one daughter gf coincide at the same lattice site , the corresponding contribution is of @xmath167 with @xmath168 , where @xmath124 .
taking now the only daughter vacuum diagram and adding successively all the other daughter gf diagrams , the total contribution for given positions of all the gf external vertices will then be @xmath169 with @xmath168 , unless all the external vertices of each daughter gf coincide at the same lattice site , that can be different for different daughter gf .
a typical example of this type of diagram is given in fig [ f5]b , while in fig [ f5]c it is shown the corresponding f.t.p.r .
topology of the embedding when the vertices that are not t.p.r . in the diagram
are at the same lattice site_. _ it is then clear that in the sum over the vertices that are common to all the daughter diagrams , each of the terms will be at best of @xmath170 , and would vanish in the limit @xmath171 unless all the external vertices of each daughter diagram are at the same lattice site , that can be different for different daughter diagrams .
one can conclude that the contribution of the mother diagram would vanish when any pair of vertices @xmath139 and @xmath5 that are not t.p.r . occupy different lattice sites , because in that case , it is always possible to separate from the mother diagram a daughter gf that has @xmath139 and @xmath5 in common with the rest of the diagram .
in that case , only a sum over @xmath138 should remain from the unrestricted sum over @xmath139 and @xmath172 and those two vertices then become t.p.r . in the embedding of the diagram (
this is illustrated in figs .
[ f5]b , c ) .
this reduction of the terms that contribute to the lattice summations has been already described , and called `` collapse of vertices '' , in the study of the u - perturbation theory in high dimensions@xcite : this property will also be given the same name in the present work .
one important point to notice , is that the collapse of vertices occurs only in the embedding of the diagram , and that the rules for calculating the contribution should be applied to the original diagram ( i.e. it would be the diagram of fig .
[ f5]b in the examples given ) and not to the collapsed diagram of the embedding ( i.e. the diagram in fig .
[ f5]c ) . as in the collapse of @xmath139 and
@xmath5 the two independent lattice summations over the vertices @xmath139 and @xmath5 are replaced by a single lattice summation over @xmath138 , it is easy to see from the derivation of the contribution rules in reciprocal space ( cf . section iii d in ref . )
that the two independent momentum conservation at each of the two collapsed vertices becomes a single conservation of the wave vectors corresponding to all the internal edges joining them .
when all the pair of vertices that are not t.p.r .
have been collapsed , one can see that at every vertex of the embedding of a vacuum or one particle gf diagram , all the edges can be arranged in pairs with momentum @xmath173 and -@xmath173 respectively ( cf .
[ f5]c as a typical example ) it then follows that the momentum conservation at the collapsed vertices are automatically satisfied , and most of the usual restrains on the momentum integration disappear in the limit @xmath174
the cumulant expansion of the pam@xcite was considered in the limit of infinite spatial dimensions ( @xmath175 for two types of hybridization models : the purely `` local '' hybridization and the `` nearest neighbor '' ( n.n . )
one . as the systems usually described by the pam have a bandwidth much larger than the hybridization energies , the unperturbed hamiltonian is chosen to include the exact solution of the conduction band electrons in the absence of hybridization .
it was then necessary to modify the derivation employed by metzner@xcite for the cumulant expansion of the hubbard model , in which the hopping hamiltonian is used as perturbation .
the basic result presented here is that in spite of this change the pam shows , for the two hybridization models considered , the same type of simplifications that occur in the diagrammatic cumulant expansion of the hubbard model when @xmath9 .
only the linked diagrams contribute to the general cumulant expansion , and there is a vertex collapse when @xmath9 for those diagrams that are not f.t.p.r .
a diagram is not f.t.p.r .
when at least two vertices @xmath112 and @xmath5 can not be separated by cutting only one or two edges of the diagram , and their collapse means that from the two independent summations over @xmath139 and @xmath5 that are necessary to calculate their contribution , only a summation over @xmath138 remains .
the topology of the embedding is different from that of the diagram itself when @xmath138 , and the collapse is repeated until the topology of the embedding becomes f.t.p.r .
one important point to notice is that , after the collapse of vertices , the rules for calculating the contribution should be applied to the original diagram and not to the collapsed diagram of the embedding .
when the two independent summations over @xmath139 and @xmath5 of the diagram become a single summation over @xmath138 because the two vertices @xmath139 and @xmath5 collapse , the two separate conservations of @xmath26 at these vertices become a single conservation of the @xmath26 corresponding to all the edges joining both @xmath139 and @xmath5 . as a consequence ,
most of the usual restrains on the momentum integration disappear in the limit @xmath171 and the calculation of many diagrams is very much simplified by this change .
one should note that the vertex collapse does not alter the calculation over frequencies , which keep their conservation at all the diagram vertices , even when @xmath9 . employing the cumulant expansion , metzner has shown@xcite that the single - particle properties of the hubbard model in the limit @xmath35 , can be described as that of independent electrons `` hopping between dressed atoms characterized by an effective green s function '' .
a similar derivation can be employed for the pam , and the exact one electron gf is given by the family of diagrams in fig .
[ f1]a , but using an effective cumulant @xmath176 for the f - electron vertices instead of the bare one @xmath177 the effective cumulant @xmath176 is given by the contribution of all the diagrams of @xmath178 that can not be separated by cutting a single edge ( usually called `` irreducible diagrams '' ) , where @xmath178 is the exact gf of the f - electrons in the real space representation for @xmath179 this property is only valid in the limit @xmath9 , so that all the diagrams that are not f.t.p.r have their vertices collapsed until the associated embeddings are f.t.p.r .
the exact gf can then be written @xmath180 a practical difficulty in the study of correlated electron systems with cumulant expansions , is that the higher order cumulants rapidly become very laborious to calculate .
to consider in some way the higher order cumulants , we are studying the substitution of @xmath176 by an approximate quantity @xmath181 , derived from an exactly soluble model . to this purpose
we use the same anderson periodic model but in the atomic limit@xcite , i.e. when the hopping of the conduction electrons is eliminated by taking a conduction band of zeroth width .
this attempt will be discussed in another publication , and the present work provides an essential guidance to that study , by showing that all the cumulant diagrams present in @xmath176 are also present in the approximate @xmath181 .
although the hopping is missing from the @xmath181 , the exact solution of the conduction band shall appear in the `` hopping between dressed atoms''@xcite through the @xmath182 in eq .
( [ e5.1 ] ) . the authors would like to thank prof .
roberto luzzi for critical comments , and to acknowledge financial support from the following agencies : capes - picd ( msf ) , fapesp and cnpq ( mef ) .
this work was done ( in part ) in the frame of the associate membership programme of the international centre for theoretical physics , trieste italy ( mef ) .
in section [ sec4 ] it is shown that the contribution of a `` mother '' vacuum diagram that is not f.t.p.r . can be estimated from those of several f.t.p.r .
`` daughter '' diagram that are derived from the original one .
the daughter diagrams are obtained by separating the mother diagram into several f.t.p.r .
subdiagrams : a vacuum diagram plus several gf diagrams obtained by adding external lines of the bose type to the vertices that these separated diagrams have in common in the mother diagram . to show the procedure , it is convenient to analyze a simple case , like that of fig .
[ f5]a , and consider only a single vertex in real space , as shown in fig [ f6]a , assuming first that it does not belong to the vacuum daughter diagram . writing only that part @xmath183 of the total contribution to the mother diagram that have parameters connected to the particular vertex under consideration , one obtains @xmath184 \nonumber \\ & & \times v(1)v^{*}(2)v(3)v^{*}(4)g_c^0(1)g_c^0(2)g_c^0(3)g_c^0(4)m_4 ^ 0(j,1,2,3 , 4 ) \label{apa.1}\end{aligned}\ ] ] where the @xmath185 are the c - electron gf , and the other abbreviations are those introduced in section [ sec4 ] .
the time or frequency dependence has been left out because it only plays a trivial role in this proof .
the corresponding contribution @xmath186 to one daughter diagram for fixed values @xmath187 of the internal lines joining the vertex , and with an added external line of momentum @xmath188 and parameters @xmath189 is @xmath190 \nonumber \\ & & \times m_3 ^ 0(j , a,1,2)\ v(1)v^{*}(2)g_c^0(1)g_c^0(2)\quad , \label{apa.2}\end{aligned}\ ] ] and the contribution @xmath191 of the other daughter diagram is trivially obtained by replacing @xmath192 by @xmath193_. _ transforming @xmath194 to real space , the contribution associated to a site @xmath195 is given by @xmath196 v(1)v^{*}(2)g_{c}^{0}(1)g_{c}^{0}(2)m_{3}^{0}(j , a,1,2)\quad \label{apa.3}\end{aligned}\ ] ] and a similar expression @xmath197 is obtained for the other daughter diagram .
it is clear that the only difference between the product @xmath198 and the term corresponding to @xmath199 and fixed @xmath200 in eq .
[ apa.1 ] is given by the cumulants @xmath201 , which are all independent of both @xmath202 and of the space dimension @xmath11 .
it then follows that @xmath203 gives the dependence of @xmath183 with @xmath11 . when one of the two daughter diagrams is the chosen as the vacuum one , it is necessary to consider the corresponding contribution @xmath204 v(1)v^{*}(2)g_c^0(1)g_c^0(2)m_2 ^ 0(j,1,2)\quad , \label{apa.5}\ ] ] that was obtained replacing @xmath205 by @xmath206 in eq .
( [ apa.3 ] ) : this is the quantity that appears in the corresponding vertex of the vacuum daughter diagram . in this case , instead of eq .
( [ apa.4 ] ) one has the relation @xmath207 this are the relations employed in section [ sec4 ] to prove that only embeddings that are f.t.p.r
. give a non zero contribution , thus leading to the collapse of vertices when they can not be separated by cutting at most two edges .
in the procedure discussed in section [ sec4 ] a `` mother '' vacuum diagram that is not f.t.p.r . is separated into several subdiagrams that have that property . as discussed before ,
one of the daughter diagrams is chosen as a vacuum diagram , and the remaining ones are gf diagrams obtained by adding external lines of the bose type to the vertices that these gf diagrams share with other daughter diagrams in the mother diagram .
let us consider one of the gf daughter diagrams in the reciprocal space , with momenta @xmath208 indices @xmath209 and transitions @xmath210 ( @xmath211 assigned to the @xmath212 external vertices , where the @xmath210 correspond to pair of states with equal number of electrons , i.e. to operators @xmath213 of the bose type .
all these external parameters will be indicated by @xmath214 , while @xmath215 will be used for the set of @xmath216 associated to the internal edges joining all the external vertices .
the contribution corresponding to fixed @xmath215 of a split diagram is denoted by @xmath217 , where the factor @xmath218 , associated to each external line joining an f - vertex ( cf .
rule 3.7 f in ref . ) , has been explicitly written . to transform from @xmath219 to position variables @xmath220 one
should calculate @xmath221 \label{apb.1}\ ] ] where @xmath222 denotes all the @xmath216 , both internal and external . from the translational invariance of the system ( or from wave vector conservation at all the vertices ) it follows that @xmath223 , so that @xmath224 \right ) \right\ } \ f\left ( \left\ { { \bf { k}}% _ \upsilon , \ell _ \upsilon \right\ } , \left\ { \ell \right\ } _ { int}\right ) \label{apb.2}\ ] ] in the two hybridization models considered in this work ,
the external wave vectors @xmath225 appear in @xmath226 only through the delta that gives the wave vector conservation at each external vertex .
the internal wave vectors @xmath26 appear through the @xmath227 in @xmath228 ( i.e. the c - electron gf of eq.([e3.9a ] ) ) and also in the @xmath229 ( cf .
( [ e3.10 ] ) ) for the n.n . hybridization model .
when the wave vector conservation at all vertices is considered explicitly , the number of summations over internal @xmath26 is reduced , but the arguments of the @xmath230 become linear combinations of both the remaining external and internal @xmath231 to illustrate this result , consider the diagram in fig . [ f7 ] : applying momentum conservation one can write @xmath232 where the frequencies are not explicitly written because they do nt play any role in this analysis .
there is only a single summation over @xmath233 because from the sum over @xmath234 only @xmath235 remains . to study the general case , it is convenient to write ( cf .
( [ e2.4 ] ) ) @xmath236 so that @xmath237 can be expanded in series of the @xmath238 , putting @xmath239 at the end of the derivation . to analyze the dependence of @xmath240 with @xmath241 , it is convenient to concentrate first on a given external vertex @xmath242 , and write @xmath243 and @xmath244 the eq .
( [ apb.2 ] ) shows that for the given external vertex @xmath242 , there is a @xmath245 applied to @xmath246 for each dimension @xmath247 of the space , and when for a given @xmath12 it is @xmath248 , all the terms independent of @xmath249 in the expansion of @xmath250 will cancel out ( cf .
( [ e3.17 ] ) ) .
it then follows that @xmath251 , where @xmath252 and @xmath253 is the number of dimensions for which there are non zero components @xmath12 for at least one of the @xmath254 vectors @xmath255 .
one can then conclude that unless all the @xmath256 coincide , the @xmath257is at least of @xmath258 and it vanishes when @xmath259 this property was used to prove the collapse of vertices discussed in section [ sec4 ] for diagrams that are not f.t.p.r .
hubbard j 1964 _ proc .
london , ser .
_ a * 276 * 238 ; hubbard j 1964 _ proc .
london , ser . _ a * 277 * 237 hubbard j 1964 _ proc .
london , ser . _ a * 281 * 401 ( these are the first three papers of a series of six ) | the diagrammatic cumulant expansion for the periodic anderson model with infinite coulomb repulsion ( @xmath0 ) is considered here for an hypercubic lattice of infinite dimension ( @xmath1 ) .
the nearest neighbor hopping of the uncorrelated electrons is described exactly by a conduction band , while two different models of hybridization are treated as a perturbation .
the same type of simplifications obtained by metzner for the cumulant expansion of the hubbard model in the limit of @xmath1 , are shown to be also valid for the periodic anderson model .
the derivation of these properties had to be modified because of the exact treatment of the conduction band . |
recently many authors have pointed out that the speeds of jets associated with different types of compact object ( pre - main - sequence star , white dwarf , neutron star , black hole ) are approximately equal to the keplerian or escape speed from the surfaces of those compact objects .
however , some caution must be exercised in taking this equivalence too literally .
a good example of this is the extragalactic radio source class
. all are believed to be powered by accretion onto black holes , from whose surface ( the horizon ) the escape velocity is exactly @xmath1 , the speed of light .
however , while all extragalactic jets appear to be at least mildly relativistic , the speeds obtained from very long baseline interferometry ( vlbi ) observations range from @xmath2 in centaurus a ( @xcite ) to @xmath3 in 3c 345 ( @xcite ) . in terms of the proper velocity @xmath4 , this represents a range of two orders of magnitude in jet speed .
clearly , even before traversing the main part of the interstellar medium , jet speeds are strongly affected by the details of the acceleration and collimation process near or in the accreting black hole system .
radio jets on the kiloparsec scale also differ dramatically in speed , and these differences apparently translate into a morphological difference .
subsonic or transonic jets with velocities less than @xmath5 are believed to be responsible for producing the fanaroff and riley class i sources ( @xcite ; @xcite ; @xcite ) .
the fr i sources are distinguished by generally low radio power , often complex morphology ( s - shaped , c - shaped , _ etc . _ ) , and diffuse lobes that have their bright region close to the associated galaxy all consistent with transonic flow that decelerates into subsonic flow .
class ii sources generally are much more powerful , have only a straight morphology , and have their brightest point at the far ends of the dumbbell - shaped source .
they are thought to be produced by highly supersonic jets that terminate in a strong shock where the high - speed flow strikes the interstellar medium .
a mechanism for determining the speed of jets at their site of acceleration was suggested by @xcite , hereinafter megpl . termed the `` magnetic switch '' because of the two distinct slow and fast jet states , this mechanism operates in the magnetized corona of the black hole accretion disk itself .
fast ( highly relativistic ) jets are produced when the coronal plasma is unbound by the magnetic field , and slow ( trans - relativistic ) jets at roughly the disk escape velocity are produced when the corona is initially bound .
this simple physical process is a good candidate for explaining the difference between the two fanaroff & riley classes , because it explains how jets with radically different speeds could occur in otherwise similar sources .
however , as discussed in section [ old_magsw_section ] below , the magnetic switch , triggered solely by the magnetic field strength ( or even by coronal density ) , fails to explain why these morphological differences have rather long lifetimes ( @xmath6 yr ) .
one way out of this dilemma is to assume that all radio jets for a given galaxy mass indeed do start out in the nucleus with about the same average speed , but , depending on local galaxy properties ( which do have long lifetimes ) , are affected differently by deceleration processes in the interstellar medium ( @xcite ) .
indeed , jet deceleration _ has _ been inferred from variations in the jet - to - counterjet ratio in observations of fr i sources ( @xcite ) . in this paper , however , we propose an alternate method for maintaining long - lived fr i or fr ii morphology one that preserves the magnetic switch model , along with all of its attractive properties , and suggests a reason _ why _ some jets may be more prone to deceleration than others . since the power in these mhd outflows also depends on the angular velocity of the magnetic field , _ it is suggested that the difference in radio power among galaxies of the same mass is due to different magnetic field rotation rates in their central engine , caused by different spin rates of their central black holes_. this has been suggested before ( @xcite ) based on observations alone , but this paper derives such a behavior directly from a generalized theory of the magnetic switch . as we show below , black holes of different spin should produce jets of different power and velocity , and a sharp fr i / ii - like transition is expected to occur .
unlike the ism model , this model predicts a different mean jet velocity in the central parsec for the fr i and fr ii classes , yielding a definite observational test for distinguishing between the two .
note that the magnetically - switched , rotating black hole model depends critically on the jet luminosity being a direct function of the black hole angular momentum in direct conflict with the conclusions of @xcite .
therefore , section [ rot_bh_subsection ] spends some time on this issue and shows that , while the basic tenets of these authors are correct , nevertheless , if the accretion disk within the ergosphere is taken into account , a blandford - znajek - like ( bz ) expression for the jet power is recovered even though the jet is ejected from the accretion disk itself by a blandford - payne - like ( bp ) mechanism .
the jet production model is similar to the punsly - coroniti ( pc ) ergospheric wind mechanism ( @xcite ) .
furthermore , due to an expected _ metric - shear_-driven dynamo ( see the appendix ) , the poloidal magnetic field in the ergosphere is expected to have a larger value than that suggested by @xcite , leading to radio jet powers comparable to those observed .
the outline of the paper is as follows .
section [ old_magsw_section ] briefly reviews the classical magnetic switch model and the reasons for its failure to explain the fr i / ii break .
section [ gen_magsw_section ] presents the generalized magnetic switch ( including rotation as a parameter ) and its role in the rotating black hole model for jet production and for the fr i / ii transition . section [ appl_section ] discusses applications of the model to recent observations of radio sources , with observational and theoretical tests of the model proposed in section [ tests_section ] .
in the original magnetic switch model presented in megpl , the flow behavior depends on the ratio of the alfvn velocity in the inner disk corona @xmath7 to the escape velocity there @xmath8 , where @xmath9 is the strength of the poloidal magnetic field protruding from the disk into the corona at radius @xmath10 , @xmath11 is the density of the corona there , @xmath12 is the mass of the black hole , and @xmath13 is the gravitational constant . when @xmath14 , gravity dominates magnetic forces ( the plasma is bound initially ) and the character of the flow is similar to a parker wind ( with collimation ) : azimuthal magnetic field coiled in the corona uncoils upward to slowly accelerate the flow through a series of critical points . depending on the parameters ,
the resulting jet speed is approximately equal to the escape speed at the base of the outflow ( @xcite ) , which is somewhat interior to the last stable orbit in the relativistic schwarzschild black hole case ( @xcite ) . when @xmath15 , magnetic forces dominate gravity and , for sufficiently high @xmath16 , render the latter negligible .
( the coronal plasma is unbound . ) as in pulsar wind theory , the outflow velocity then is determined not by gravitational effects , but by the magnetization parameter @xmath17 which is a measure of the amount of work done on the injected plasma ; here , @xmath18 is the velocity with which matter is injected , @xmath19 is the radius of the light cylinder , and @xmath20 is the angular velocity of the poloidal magnetic field at @xmath10 . for relativistic flows
, @xmath21 appears to be a good approximation for the final mhd wind / jet velocity ( @xcite ; @xcite ) .
a transition between the two types of outflow then is induced by increasing the magnetic field strength @xmath9 until @xmath15 , whereupon significantly higher jet speeds occur .
this magnetic switch has several properties that make it an attractive model for the difference in morphology between class i and class ii extragalactic radio sources ( @xcite ) : 1 ) the transition is sharp , occurring with increasing jet power ; 2 ) jets produced when @xmath22 are transonic , while those for @xmath15 are highly supersonic , and 3 ) even statistical behavior observed in large samples of radio galaxies of different power and optical magnitude is reproduced ( see @xcite and section [ olwl_subsection ] below ) .
however , as presented in megpl , the original magnetic switch , triggered by either increasing the coronal magnetic field or decreasing the coronal density , fails to explain the fr i / ii break , chiefly because it fails to explain why radio sources appear to maintain their fr morphology for long periods of time at least as long as the flow time from galactic center to lobe ( @xmath6 yr ) .
there is no reason to expect the strength of the magnetic field threading disk and hole ( or the coronal density ) to remain fixed at high or at low values over such long time scales . indeed
, one expects that these quantities will be distributed fairly equally about a mean , and on average will be the same for galaxies of the same mass .
the argument is as follows .
the structure of an accretion disk is a function of the black hole mass @xmath23 , the accretion rate @xmath24 , the viscosity parameter @xmath25 , and position in the disk scaled to the schwarzschild radius ( @xcite ; @xcite ) .
now , there is growing evidence that the black hole mass , even in presently inactive objects , is approximately proportional to the mass of the central spheroidal bulge of stars ( see , _ e.g. _ , @xcite ) such that @xmath26 where @xmath27 is the absolute blue optical magnitude of the bulge .
in addition , while the instantaneous accretion rate will vary dramatically on short time scales ( years ) , because the hole is fed by galactic stars and gas , over the very long term the average accretion rate also should be roughly proportional to the bulge mass ( a perhaps necessary condition for equation [ [ bh_mass_vs_gal_mag_eq ] ] to be valid ) .
therefore , to first order , we expect the average accretion rate @xmath28 to be a constant , proportional to the average amount of matter shed per unit time into the nucleus and onto the hole by a galactic bulge of a given optical magnitude , divided by the mass of the black hole in that nucleus .
( note that the definition of @xmath29 here does not include an efficiency of conversion of accreting matter into radiation , allowing for maximum values somewhat greater than unity [ @xmath30 , depending on the spin of the black hole .
it also assumes that the accretion is continuous ; to take intermittent accretion into account , we will add a duty cycle parameter below . ) furthermore , detailed simulations of the magneto - rotational instability ( @xcite ; @xcite ) have shown that @xmath31 is a fairly good approximation for turbulent accretion flows , presumably independent of black hole mass to first order . and ,
finally , jets are believed to be ejected from near the black hole horizon ( _ i.e. _ , from just inside the last stable orbit , @xcite ) , rendering the position parameter @xmath32 a near - constant value for all objects .
therefore , ignoring external influences , _ over periods of time comparable to the flow time from nucleus to lobe , all parameters that affect the disk structure equations are expected to be , at most , functions of the black hole / galactic bulge mass only_. on average , one expects bulges of the same mass to produce the same type of accretion disk in the nucleus , with similar magnetic field and coronal density , and , therefore , the same type of jet ( either all slow or all fast ) .
it is not expected that there will be a spectrum of jet structures that maintain themselves for millions of years . yet
the observations show that just such a range of radio powers and morphologies exists , and that a radio source remains an fr i or ii for longer than a flow time .
in order to lift the degeneracy in expected jet speeds and morphologies , we shall take as an additional parameter the rotation rate of the magnetic field , as determined by the black hole rotation rate itself .
this section develops the model of the accreting rotating black hole system , the production of radio jets , and the triggering of the magnetic switch . in order to compute observable quantities
, we must assume an accretion scenario in the black hole system .
for definiteness , we choose a rather benign ( standard ) accretion disk in which electron scattering dominates free - free absorption and gas pressure dominates radiation pressure , and we scale the results to a rather low accretion rate ( @xmath33 ) . while the details will change depending on the exact accretion model and rate , the basic conclusions herein will not , even in the extreme cases of near - eddington accretion ( radiation pressure dominated ) or very low , advection - dominated accretion flow ( adaf ; @xcite ) .
the first task is to generalize the magnetic switch theory to include rotation as a triggering parameter , in addition to magnetic field strength and plasma density . consider a general magnetic rotator of size @xmath10 , poloidal magnetic field @xmath9 , and angular velocity @xmath20 .
following @xcite , the work done on the plasma trapped in the field lines , resulting in an outflowing magnetohydrodynamic ( mhd ) wind or jet , is @xmath34 for non - relativistic flow . now , for a normal parker - type wind , the plasma in the injection region is in hydrostatic equilibrium , with only a slight force imbalance outward that slowly accelerates the outflow through a critical point until it reaches escape velocity and leaves the system ( see , _
e.g. _ , @xcite ) .
this is also true of mhd winds , although the critical point structure and geometry are more complex .
however , based on the magnetic switch mechanism described by megpl , there should exist a critical power , given by the liberation of an escape energy in a free - fall time @xmath35 such that if @xmath36 exceeds @xmath37 , then the plasma becomes unbound even in the injection region itself .
that is , the outward magnetic and centrifugal forces there strongly exceed gravity , accelerating the plasma to greater than the escape velocity within @xmath10 and eventually to a much greater final velocity . _ the magnetic switch luminosity
, then , plays the same role in mhd acceleration as the eddington luminosity plays in radiative acceleration . _ for keplerian disks ( with @xmath38 ) , the condition @xmath39 reduces to the original magnetic switch condition @xmath40 .
however , for a non - keplerian rotator of _ fixed _ magnetic field strength and plasma density , there should exist a critical angular velocity @xmath41 such that if @xmath20 exceeds @xmath42 , then the magnetic switch also should be triggered and a fast jet of the type found by megpl should ensue .
the author tested this hypothesis by performing numerical mhd simulations similar to megpl , but holding @xmath9 fixed and varying the disk rotation speed instead .
the left panel of figure [ sim_fig ] shows the resulting jet speed as a function of @xmath20 . for angular velocities below @xmath42
the terminal jet / wind speed is approximately the escape speed at the injection point .
when @xmath43 , the material is unbound and quickly accelerated ( in a well - collimated jet , similar to megpl ) to a speed such that a large fraction of the power @xmath36 is in the form of kinetic energy ( see the right panel of figure [ sim_fig ] ) .
it is this strong difference in the character of the outflows below and above @xmath42 that is so similar to the fr i / ii division : while the total power varies smoothly as @xmath36 exceeds @xmath37 , the character of the outflow varies dramatically from a transonic , magnetic - energy - dominated jet to a highly supersonic , kinetic - energy - dominated one .
the details of the behavior of the curves in figure [ sim_fig ] can be understood qualitatively as follows .
when @xmath44 , the injected plasma is no longer supported fully by rotation .
therefore , as soon as it emerges from the disk boundary , it dynamically contracts in cylindrical radius , seeking a new keplerian equilibrium .
the jet then is ejected from this new , smaller , radius @xmath45 with a speed similar to the keplerian velocity there and a power slightly _ larger _ than that given by equation ( [ nonrel_mhd_power_eq ] ) , because of compressional increase in the axial magnetic field component ( @xmath46 ) . if @xmath20 is decreased further , the plasma contracts more , increasing the eventual jet velocity slightly . on the other hand , when @xmath47 , the reverse occurs , with the plasma expanding to a new injection radius . as @xmath20 increases , the value of @xmath42 calculated from equation ( [ omega_crit_eq ] ) changes as well , chiefly because only the radial component of the magnetic field is important in the expanded state ( @xmath48 ) .
this increases the critical value to @xmath49 ( where @xmath50 is the _ radial _ component of the alfvn velocity at the new injection radius ) , which is greater than the value in equation ( [ omega_crit_eq ] ) .
finally , well above the magnetic switch ( when @xmath51 ) the well - collimated jet flow gives way to a less - collimated radial wind , and the increase in outflow speed with rotation speed begins to break down .
it is not clear whether such a decrease at high @xmath20 will occur in the relativistic case , where light cylinder effects will be important .
relativistic mhd simulations are needed to investigate this behavior .
some brief comments should be made about boundary conditions . in these simulations , the kinematic and magnetic properties of the inflowing matter at the equator ( the `` disk corona '' )
are fixed throughout the calculation . however , some authors allow these parameters to be determined by the solution , particularly the azimuthal and radial components of the field at the equator and the inflow velocity there ( see , _ e.g. _ , @xcite )
. both of these sets of conditions ( fixed or self - adjusting ) are easy to apply numerically , and each is appropriate under different circumstances .
the latter is useful for generating steady state solutions to compare with @xcite s self - similar , semi - analytic models . in this case ,
conditions at the base of the outflow re - adjust themselves , subject only to the strength of the vertical magnetic field , the density of inflow , and the speed and shear of the disk rotation .
the system finally reaches a steady state with parameters lying on ( or near ) the blandford - payne solution surface in @xmath52-space ( see bp ) , with the flow varying smoothly from the outer regions all the way back to the equatorial boundary . on the other hand ,
fixing the inflow conditions on the equator , as we have done here , is perhaps more appropriate for studying the astrophysical situation where the magnetic field is anchored in an accretion disk ( or other type of rotator ) . in that case
the local magnetic field is determined by processes internal to the rotator , not by the outflow that originates above it . when the boundary conditions are fixed , a steady flow still develops above the disk , and the azimuthal and radial fields and flow velocity in the corona also re - adjust .
however , because that resulting outflow does not necessarily match onto the fixed disk boundary conditions , there is a transition region between the disk and the base of the outflow region , in which the magnetic field and velocity dramatically change in magnitude and direction .
it is not yet fully clear whether this transition region merely provides a way to connect an outflow solution that would have occurred anyway to the disk conditions , or whether it plays an active role in determining what outflow solution is chosen .
what is clear is that the values of the parameters chosen on the disk _ below _ the transition region _ do _ matter : the flow changes dramatically with the strength of the poloidal _ and _ azimuthal field specified and , to a lesser extent , with the angle of the poloidal field ( @xcite ; megpl ) .
furthermore , jet simulations in which a substantial _ equatorial _ field ( azimuthal plus radial ) is specified seem to display more collimation , presumably because of the strong azimuthal field that is initially present or that develops after a few turns of the disk .
the proposed model of jet production from the accreting , rotating black hole system is essentially a marriage between the disk acceleration model of @xcite and the rotating black hole acceleration model of @xcite and @xcite .
the former authors considered the case when the magnetic field threads only the keplerian portion of the disk , while the latter examined the situation when it threads both hole and disk . in this paper
we consider the intermediate case when the magnetic field threads both a region of the system outside the static limit that does not experience appreciable frame dragging and a region closer to the black hole that does the ergosphere .
the main features of this model are the following : ( 1 ) unlike the bz model , the magnetic field does not have to thread the horizon itself in order to extract the black hole rotational energy , which avoids the problems discussed by @xcite , and the primary source of material in the jet can be a corona created by the accretion disk itself , not necessarily particles created in spark gaps near the horizon .
( 2 ) it is unlike the bp model because the disk and field there are frame - dragged by the hole relative to an observer at infinity , and thus the jet production couples well to the black hole spin , extracting rotational energy even though it is disk rotation that accelerates the jet .
this model is very similar to that of pc , who studied steady state winds from the ergospheric region of a rotating hole .
we note here an additional feature of such a system that occurs when time dependence is taken into account : as shown in the appendix , when a plasma with an appreciable poloidal magnetic field is present in the ergospheric region , the differential dragging of frames will increase the azimuthal magnetic field at the expense of the hole s rotational energy .
if , as a result of dynamo action , this also is accompanied by a corresponding increase in the poloidal component , this will have the effect of increasing the jet power produced near a rotating black hole to a value substantially larger than that for a disk around a schwarzschild hole .
when a black hole is spinning rapidly , the accretion disk can extend well into the ergosphere the region where local observers must rotate with the hole with equatorial radius @xmath53 .
for a nearly - maximal kerr hole ( @xmath54 ) , the last stable orbit reaches into this region very close to the horizon at @xmath55 , where @xmath56 is the normalized angular momentum of the hole ( equal to `` @xmath57 '' in geometric units ) .
now , consider a loop of coronal magnetic field @xmath9 with one foot anchored in the ergosphere and one outside . recently
@xcite and @xcite have shown that even in a keplerian flow , differential rotation in the inner disk can modify closed coronal loops , stretching and twisting them into long , virtually - open vertical helical structures of the type needed to accelerate mhd winds and jets .
if the differential rotation is further augmented by frame dragging , then the work done on the field and outflow derives its power not only from the keplerian rotation of the disk , but also from the rotation of the ergosphere itself relative to the point where the outer foot is anchored ( see the appendix ) .
the system then appears to an observer at infinity as a simple magnetic rotator with size @xmath58 , magnetic field @xmath9 , and angular velocity @xmath59 , where @xmath60 and @xmath61 are the angular velocity of the ergosphere equator and hole , respectively .
the form of equation ( [ nonrel_mhd_power_eq ] ) for a relativistic rotator ) also can be derived from equation ( [ sigma_eq ] ) and the relations @xmath62 and @xmath63 , where @xmath64 is the mass lost in the jet . ]
yields an mhd power for the accreting system of @xmath65 which is a comparable power to that of the bz mechanism for the same poloidal field strength .
@xcite have argued that the mhd power from the accretion disk ( the bp process , given by equation [ [ nonrel_mhd_power_eq ] ] ) and that from the black hole ( the bz process , given by equation[[rel_mhd_power_eq ] ] with @xmath66 ) are , at best , comparable , and , at worst , entirely dominated by the former if the hole spin is small .
they , therefore , conclude that the bz process will never be important and that the spin of the black hole is probably irrelevant for the expected electromagnetic power of the system .
this author agrees with the basic tenets of @xcite that the disk and hole power will be comparable , but not with their ultimate conclusion on the unimportance of the spin .
the key difference is that , in the ergosphere , the mhd power _ of the disk itself _
( equation [ bh_mhd_power_eq ] ) should depend on the spin of the black hole .
in addition , when the hole is rotating , a much larger poloidal magnetic field strength is allowed than estimated by these authors , yielding black hole rotation - driven powers comparable to those observed in extragalactic radio sources .
first , consider the non - rotating schwarzschild black hole case .
the disk will be a standard keplerian one , with internal disk field given by ( @xcite ) @xmath67 where @xmath68 is the local internal disk pressure .
now the source of a coronal loop of field @xmath9 will be this ( largely azimuthal ) disk field , first protruding into the corona and then growing as the loop is sheared and expanded upward .
however , in a thin keplerian disk , if the poloidal field is allowed to grow to the same order as @xmath69 , then the expected mhd power from equation ( [ nonrel_mhd_power_eq ] ) would be orders of magnitude greater than the accretion luminosity @xmath70 where @xmath71 is the sound speed in the disk , @xmath72 is the inward accretion drift speed , and the notation @xmath73 represents @xmath74 .
( as discussed earlier , we have used a gas - pressure , electron - scattering - dominated disk , but the results are similar for other types of disks . )
there is not enough accretion luminosity to support this mhd power by factors of several thousand .
as @xmath9 grows to a even a small fraction of @xmath69 , the mhd disk power becomes a substantial fraction of @xmath75 . at that point
, a significant amount of angular momentum is removed from the disk by the outflow , altering the disk structure and enhancing the local accretion rate . in order that the accretion draining the inner disk remain in a steady state with that feeding it from the outer disk ( and not , for example , launch into a series of bursts that each destroy the inner disk entirely ) , @xmath9 must be self - limited to about the value postulated by @xcite , or @xmath76 , where @xmath77 is the half - thickness of the disk at @xmath10 .
however , when the central object is a rotating black hole , the situation will be quite different .
as before , when @xmath9 grows to a fraction of @xmath69 , a significant mhd outflow will develop , exerting a braking torque on the inner portion of the disk .
because of the differential frame dragging , however , even if _ all _ the keplerian angular momentum were removed by the outflowing mhd power , the disk material still will be sheared _ by the metric itself _ , continuing to increase the local magnetic field strength and continuing to produce mhd power .
the appendix shows quantitatively how this occurs in the kerr metric , producing field enhancement everywhere in the ergospheric region .
tension in this growing , working field then produces a back - reaction on the disk plasma falling toward the hole , accelerating it relative to the rotating frame in a direction _ retrograde _ to the black hole spin . by the time the magnetized accretion flow enters the black hole
, it could have a significant _ negative _ angular momentum which , when added to the hole , would decrease the spin . therefore , because of frame dragging , power flows from hole to disk to field and finally to the outflow , exerting a braking torque on the hole itself .
no longer self - regulated by the accretion flow , @xmath9 in the ergospheric region above the disk now can continue to grow at the expense of the hole rotational energy to order @xmath69 , given by the disk solutions with a kerr metric ( @xcite ) . for
the gas pressure / electron - scattering accretion model assumed here , this value is @xmath78 equation ( [ bh_mhd_power_eq ] ) then becomes @xmath79 ( this differs from equation ( 16 ) of @xcite by a factor of @xmath25 , from equation ( [ ss_eq ] ) , and a factor of @xmath80 from the slightly smaller ergospheric rotation rate . )
while @xmath36 appears to scale with @xmath81 and @xmath29 in a manner similar to an accretion luminosity , it is important to note that it does so only because of the scaling on @xmath69 .
the energy liberated is derived mainly from the rotation of the hole itself and potentially much larger than the accretion luminosity ( @xmath82 ) . to turn @xmath36 into an observed radio power , we note that , over long periods of time , the central engine may generate a jet only a fraction @xmath83 of the time .
furthermore , only a fraction @xmath84 of this average power is radiated at ghz frequencies in a 1 ghz bandwidth and observed with standard radio telescopes . from millimeter and optical observations of repeated outbursts of radio galaxies and quasars ( see , _
e.g. _ , @xcite ) , we estimate typical duty cycle values to be @xmath85 .
the radio efficiency is also very uncertain , but we note that for the microquasars gro j1655 - 40 and grs 1915 + 105 , estimates of @xmath84 are of order @xmath86 or so , just for the instantaneous moving jet ( @xcite ) .
a canonical value of @xmath87 was chosen for these calculations by comparing equation ( [ lmhd_eq ] ) with the maximum observed radio source power ( @xmath88 ) for a radio galaxy that should have a @xmath89 hole in its nucleus ( @xmath90 ) .
this value for @xmath84 is consistent with total conversion rates of @xmath91 ( see , _ e.g. _ , @xcite ) , with only a fraction of that radiated in a 1 ghz bandwidth . scaling to these values , the calculated radio power for different black hole spins is @xmath92 where @xmath29 is still the average _ outburst _ accretion rate only , not that averaged over the entire duty cycle .
the final step in the model is to identify the magnetic switch as the cause of the fr i / ii break .
that is , one should observe a distinct change in the character of the jet produced when @xmath93 , or when the radio power attains the critical value @xmath94 to compute an observed break power we need an expression for the corona density @xmath11 at @xmath10 .
of course , coronal physics is poorly understood , even for the sun .
nevertheless , we can obtain a rough estimate by assuming that the corona is produced locally by the accretion disk , so the disk and corona densities should scale together as and the internal disk temperature @xmath95 ( _ i.e. _ , with @xmath68 ) .
this yields a slightly weaker @xmath96 slope ( -0.55 ) in equation ( [ mod_olwl_eq ] ) below , but still within the observational errors . ]
@xmath97 where @xmath98 is the internal disk density at @xmath10 and @xmath99 is a parameter which , while not necessarily a true constant , is assumed to vary sufficiently slowly with the other parameters that it does not affect significantly the scaling of our results with @xmath81 , @xmath100 , and @xmath29 .
the parameter @xmath101 is expected to be quite small ; in the solar corona , for example , @xmath102 in the region between the photosphere and @xmath103 . if the corona is a pair plasma , @xmath101 ( which is a mass ratio ) could be very small indeed .
combining equations ( [ lcrit_eq ] ) , ( [ pfr_eq ] ) , and ( [ rhoc_eq ] ) , and again using the gas pressure / electron - scattering disk , one obtains @xmath104 this is the magnetic switch model prediction for the locus of the fr i / ii break as a function of black hole mass .
this section applies the model to the fr i / ii break in the radio - optical plane , shows that the rotationally - triggered magnetic switch _ is _ consistent with long source morphology lifetimes , and addresses the more general questions of evolution and radio loud _ vs. _ radio quiet sources .
the important features of the model and its applications are shown in figure [ roplane_fig ] .
[ appl_section ] @xcite , @xcite , and @xcite have shown that fr i and fr ii type radio galaxies occupy different regions of the @xmath105 plane , where @xmath106 is the observed radio power at 1400 mhz ( in @xmath107 ) and @xmath108 is the galaxy optical red magnitude measured to a surface brightness of @xmath109 magnitudes per square arcsecond . in particular , the tendency for a galaxy of a given magnitude to transition from an fr i to an fr ii appears to be a strong function of optical magnitude .
the dividing line is given approximately by @xmath110 or @xmath111 .
the error in the coefficient of @xmath108 is at least @xmath112 and in the intercept at least @xmath113 .
the owen - laing - white - ledlow ( olwl ) relation is a more precise statement of the fanaroff & riley class division , so care must be taken when classifying radio galaxies based on their radio power only .
there exist fr ii radio galaxies with quite small powers ( occurring in less bright elliptical galaxies ) , as well as fr i objects with very high radio power ( occurring in very bright galaxies ) .
from the magnetic switch model for the fr i / ii break , it is possible to calculate a theoretical relation for the observed break power to compare with equation ( [ obs_olwl_eq ] ) . using equation ( [ th_frbr_eq ] ) to define the break , equation ( [ bh_mass_vs_gal_mag_eq ] ) to convert black hole mass to galaxy bulge blue optical magnitude , and a standard @xmath114 to convert to elliptical galaxy blue magnitudes to red , one obtains @xmath115 given the uncertainties in the observed slope , in the scaling of the coronal density , and in the parameters @xmath84 , @xmath83 , and , especially , @xmath101 , the good numerical agreement between equations ( [ obs_olwl_eq ] ) and ( [ mod_olwl_eq ] ) is rather fortuitous , of course .
the important points to note are as follows : * the basic trend that brighter galaxies , with their more massive black holes , should undergo the magnetic switch at a higher jet power is a general feature of the model ( see figure [ roplane_fig ] ) .
the ultimate source of this trend is that the critical jet power is proportional to the size of the ergosphere squared times the disk coronal density there .
if this density were uniform for all black holes , it would produce a trend slightly steeper than @xmath116 , because the galaxy @xmath117 ratio increases with galaxy @xmath118 ( @xcite ) ; however , as densities tend to decrease as the size of the accreting system increases , the trend should be _ less _ steep than this . *
although the uncertainties are great , the intercept is consistent with the values of the parameters chosen here , in particular with the very low value of @xmath101 .
that is , _ the observations are entirely consistent with the jet material originating in the coronae of the black hole accretion disks _ , and grossly _
in_consistent by more than ten orders of magnitude with the jet ejecting a large fraction of the accretion disk itself .
incidentally , the present model is _ not _ inconsistent with observations that fr i jets may decelerate on kiloparsec scales ( @xcite ) .
indeed , it suggests a reason for such deceleration to occur more readily in fr i sources .
jets with @xmath119 are transonic in the simulations here and in megpl .
such jets are more prone to deceleration than highly supersonic flows ( @xcite ) .
note that deceleration of relativistic parsec - scale jets to nonrelativistic flow on kiloparsec scales is needed in _ any _ model , magnetically switched or not , to explain head - tail and wide - angle - tail sources in clusters , whose jets are bent by motion of only a few thousand @xmath120 through the intracluster medium . because the black hole spin is the assumed source of power , there should be a critical spin @xmath121 corresponding to a critical hole angular velocity @xmath122 such that the magnetic switch is triggered when the hole is spun up to this value . equating @xmath123 and @xmath124 ( or @xmath36 and @xmath37 ) one obtains @xmath125 now , while this relation is a very weak function of @xmath81 and @xmath29 , and @xmath25
is considered to be known to within a factor of ten or better , the exact value of @xmath121 is still very uncertain because of the parameter @xmath101 . indeed ,
a large value ( dense corona ) would make @xmath126 , implying that even a maximal kerr hole would not trigger the switch .
below observational tests are proposed to confirm the low choice of @xmath101 in this paper .
note that , independent of the magnetic switch model , if the source of the radio jet power is , indeed , the black hole rotational energy , then simple arguments show that at the fr i / ii transition , @xmath100 must be of order @xmath127 in any case . from the olwl results and
the bivariate radio luminosity function itself , it appears that , for a galaxy of a given mass , the maximum possible radio power is about 100 times the power at the fr break .
therefore , for any model where @xmath128 , we find that @xmath129 note also that , like the accretion time @xmath136 @xmath135 is virtually independent of black hole mass ( although much shorter ) .
therefore , in the absence of other effects , even _
stellar _ mass black holes like gro j1655 - 40 or grs 1915 + 105 are expected to retain their spin over @xmath137 as they use their rotational energy to drive radio jets .
such a long spin lifetime should have implications for models for galactic microquasars . during the entire lifetime of the radio source
, an energy comparable to @xmath138 should be liberated into the lobes .
that is , extended radio sources are the remnants of formerly rapidly - rotating black holes .
this amount of energy is consistent with , and in fact somewhat larger than , the total energy content of extended radio sources in relativistic particles and magnetic fields ( @xcite ) .
this allows for a significant fraction of this energy to have been converted to other forms ( _ e.g. _ , heating of the intergalactic medium ) or radiated away .
the fact that @xmath138 is proportional to @xmath131 , and not to just @xmath100 , for example , means that the spindown time is the same no matter how rapidly or slowly the hole is rotating .
the model predicts therefore , that after turning into an fr i source , the object will continue to display pure luminosity evolution on a @xmath135 time scale .
however , recent analyses of the radio source counts and redshift data ( @xcite ; @xcite ) conclude that the fr i population evolves little or not at all . modeling
this sort of behavior is very difficult , especially in a universe where everything involved with agn seems to evolve on at least a cosmological time scale ( galaxy formation and mergers ; the rate of stellar death [ which may feed gas into the nucleus ] ; even radio quiet quasars ; _ etc . _ ) .
below are some ideas for halting , or at least slowing , the evolution of fr i sources in this model ; however , much more work needs to be done on this problem : * _ constant low - powered jet from outside the ergosphere . _ after spinning down , the black hole loses most of its ergosphere and its ability to create a large poloidal field strength in the disk .
the field strength is expected to drop to the value suggested by @xcite , yielding an expected jet power of @xmath139 which , with efficiency factors similar to those used above , could generate radio emission of up to @xmath140 .
a radio source of this power is not expected to evolve much , unless @xmath29 evolves .
however , this is a very weak power much smaller than many fr i radio galaxies , which also presumably do not evolve much .
furthermore , if the accretion events have rather random angular momenta , then this minimum jet has no preferred direction .
it is likely to point in random directions with each event , never developing into an extended source , and perhaps appearing radio quiet . *
_ periodic spinup of the black hole .
_ on the other hand , if the short - term accretion events have systematically the same angular momentum direction , as might be the case for a galaxy collision or merger , then while they would keep the above weak jet oriented in the same direction , they also would tend to re - orient and spin up the hole ( @xcite ; @xcite ) . this , or the direct merger of two black holes , would re -
kindle the ergospheric jet . then , if the galaxy merger rate evolved on a very long time scale ( say , @xmath141 ) , the shorter time scale spindown process would become slave to the spinup process , with the average radio power being a function of the rate at which angular momentum is accreted onto the hole . at low redshifts ,
then , both fr i and the occasional re - kindled fr ii objects would evolve on the longer merger evolution time , not @xmath135 .
the model proposed here deals mainly with the difference between two radio loud populations the fr i and ii sources .
however , it also has implications for the difference between radio loud and quiet quasars as well . in particular
, it suggests that radio quiet objects should be identified with very slowly - rotating , nearly - schwarzschild black holes . at first , the minimum jet power ( equation [ min_mhd_power_eq ] ) appears to be a problem with this suggestion , as it predicts that all active galaxies should produce some sort of extended radio emission .
however , even if an agn does produce an extended radio source power of @xmath142 the apparent flux would be quite small : less than a few @xmath143 for seyfert galaxies at @xmath144 and quasi - stellar objects at @xmath145 , and a few millijanskies total for nearby spiral galaxies .
this amount of extended emission in spirals could be hidden easily by emission from supernova remnants , _
etc._. therefore , there should be a minimum spin rate @xmath146 for the hole such that when @xmath147 , it will appear radio quiet .
while detailed simulations of jet production from rotating black holes , plus observational selection factors , need to be taken into account to determine a precise value of @xmath146 , we can estimate it here by equating the mhd power generated by the ergospheric and keplerian disk jets ( equations [ lmhd_eq ] and [ min_mhd_power_eq ] ) and solving : @xmath148 for spins slower than this , the wind / jet from the keplerian portion of the disk outside the ergosphere will dominate the output mhd power .
even if the minimum jet power turns out not to be a problem , one still needs to explain quantitatively as well as qualitatively why the radio power distribution is bimodal : radio quiet objects appear to be a separate class , not merely the fainter version of the powerful radio sources ( @xcite ) .
one possible solution to this problem is the re - kindling the radio source through galaxy mergers or collisions .
this is not a new idea ( @xcite ) , but the model presented here , especially equation ( [ spindown_time_eq ] ) , gives a physical and quantitative basis for it . in the early universe , holes would have been born with a spectrum of rotational speeds .
those rotating rapidly would have produced extended radio sources , including holes in some ( pre-)spiral galactic bulges , while those rotating slowly would have been radio quiet .
all of these objects should have spun down in a few @xmath135 times and become radio quiet , according to the model . however , a subgroup those in galaxies that experience repeated collisions and mergers will be spun up again , at least partially to fr i class and very occasionally to fr ii strength , re - kindling the radio source .
this scenario is consistent with powerful radio sources appearing predominately in elliptical galaxies , which now are believed to form from , or be strongly influenced by , mergers .
an important corollary is that , early in their history , even the bulges of a significant fraction of spiral galaxies should have undergone a one - time extended radio source phase , lasting a few @xmath135 times and producing both fr i and ii morphologies , but with considerably weaker luminosities than those associated with present - day giant elliptical radio sources .
the radio loud / quiet problem thus may be one of environment rather than being intrinsic to the system .
however , before adopting such a simple model , it will be important to understand the strange case of the microquasar gro j1655 - 40 , which can be either radio loud or radio quiet during its x - ray outbursts ( @xcite ) .
it is not clear yet whether this is simply a short - term phenomenon or is fundamentally related to the _ macro_quasar radio loud / quiet problem .
so far , no model , including the one in this paper , appears able to explain completely this object s behavior .
it has been assumed in this paper that on a spindown time the average @xmath29 is fairly independent of black hole mass and remains at roughly @xmath149 .
however , on a cosmological time scale this is not likely to be the case .
indeed , there is evidence ( @xcite ; @xcite ) that some sources ( class ` a ' ) have a high accretion rate , indicated by strong , broad line emission , while others ( class ` b ' ) have a low ( perhaps adaf - like ) rate . for the case where the latter objects are core - dominated and appearing as bl lac objects , wall & jackson note that at least some must have fr ii sources as their parent population , not just fr i sources .
one possible explanation of this is that the average @xmath29 may have been much higher in the early universe when the merger rate was high , and lower , perhaps even adaf - like , now .
indeed , it appears that it is only the class ` a ' objects that display rapid evolution ( @xcite ) .
then , present - day re - kindled fr i and ii sources would both have a low average accretion rate , as well as a low evolution rate , and both have the potential for appearing as a bl lac object if viewed end - on .
not all fr ii sources seen this way would have to appear as quasars . ascribing a high accretion rate to class ` a ' objects solves another puzzle concerning quasars the fact that few , if any , radio loud quasars display fr i morphology . for @xmath150 , equation ( [ jcrit_eq ] )
gives @xmath151 , implying that even slowly - rotating black holes would produce an fr ii morphology when the accretion rate is high .
the right panel of figure [ roplane_fig ] illustrates that the fr i portion of the quasar diagram is significantly smaller than that for radio galaxies ; the exact size of the region depends critically on the definition of radio quietness .
this paper has shown that the magnetic switch mechanism , triggered by a rapidly spinning black hole , is consistent with a variety of observations concerning the fr i / ii break .
however , bicknell ( 1995 ) has shown that an ism model is consistent as well . in order to distinguish between these two ,
the following observational tests are proposed .
the current model predicts the following : 1 .
the ensemble average of vlbi jet speeds for sources with fr i _ morphology _ should be distinctly slower than that for fr ii sources . until detailed general relativistic mhd simulations of very tenuous coronae in kerr
black hole ergospheres can be done ( _ cf .
_ @xcite ) , the exact values of the two mean velocities will be uncertain .
the fr ii mean should be highly relativistic ( _ i.e. _ , @xmath152 or greater ; see megpl ) .
however , the fr i value may be more than mildly relativistic itself , as the escape velocity from the ergosphere is very close to @xmath1 . note the emphasis on fr morphology , rather than radio power .
radio power should _ not _ be used to distinguish between fr class in this test , unless it can be shown ( using the olwl relation ) that the galaxy in question lies well within either the fr i or fr ii region of the radio - optical diagram .
faint galaxies can have low - power radio sources that still have fr ii morphology , and vice - versa .
2 . direct measurement of the accretion disk corona density in the ergosphere , using x - ray observations for example , should yield rather low mass densities ( @xmath153 ) , perhaps either a @xmath154 @xmath155 pair plasma or a very tenuous @xmath156 @xmath157 corona .
these low values are needed for @xmath37 to be such that a radio power of @xmath158 is generated at the fr break .
significantly higher measured densities would render the corona too heavy for the magnetic switch to be triggered by any reasonable jet power , and , in this model , always would predict slow jet velocities .
recently , methods have been developed to estimate the spin of black holes by measuring the rotation rate of the innermost part of their accretion disks ( @xcite ; @xcite ) .
if such methods can be applied to extragalactic radio sources , one should find that the ensemble average of black hole spins for sources with fr i morphology should be distinctly lower than for those with fr ii morphology .
there should exist a population of high - redshift , extended radio galaxies associated with the bulges of spiral or pre - spiral galaxies .
they should include both fr i and ii morphologies , but be weaker copies of the giant sources associated with ellipticals .
the radio sources themselves should be readily detectable ( for @xmath159 and a redshift of 3 , @xmath160 or @xmath161 ) , but their optical counterparts would be very faint if undergoing an initial burst of star formation ( @xmath162 for a @xmath163 bulge at a similar redshift ; @xcite ) and even fainter ( @xmath164 ) if not .
preliminary estimates of their numbers are on the order of a few hundred to a few thousand per square degree considerably smaller than the larger number of starburst galaxies at faint fluxes ( @xcite ) .
the first and third of these tests would provide support for the magnetic switch mechanism for the fr i / ii transition . the second would support the choice for a small @xmath101 . the third and fourth provide evidence for the proposal that the black hole spin is the switch triggering mechanism and the ultimate source of the radio power .
a number of numerical tests will be almost as important as the observational ones : 1 .
firstly , as suggested by megpl , the magnetic switch needs to be confirmed by other groups performing numerical simulations , especially with a general relativistic mhd code , even if just for schwarzschild black holes . 2 .
secondly , general relativistic mhd simulations in kerr geometry are needed to confirm that a substantial jet , whose power is directly related to the black hole spin , can be produced simply by threading the ergosphere with a magnetic field that is anchored further out
. a simple vertical field should be sufficient initially , but simulations similar to those of @xcite , with closed coronal loops , would provide a more realistic test .
finally , it will be important to reproduce in kerr geometry the simple simulations performed in this paper , showing that increasing the black hole spin eventually triggers the magnetic switch .
it also will be important to determine the speed , mach number , and collimation properties of the outflow when the black hole spin is well above @xmath42 to see if the flow de - collimates ( as it did in these non - relativistic simulations ) or remains well - collimated .
the magnetically - switched , rotating black hole model for the fr i / ii break is consistent with current black hole accretion models for active galactic nuclei , with mhd simulations of jet production , and with relativistic wind theory .
it predicts that a sharp transition in jet morphology should occur with increasing radio power , as observed , and that the high luminosity jets should be fast and kinetic energy dominated .
it yields the correct trend of this transition with galaxy luminosity ( larger galaxies should transition at larger radio power ) , and is consistent with the actual value of the fr i / ii break radio power , _ if _ the average density of accretion disk coronae in black hole ergospheres is rather low ( @xmath0 ) .
the model suggests a long - term memory mechanism the black hole spin for assuring that radio sources maintain the same morphology over periods of time that are at least as long as the flow time from galactic center to lobe ( @xmath165 ) . in that regard
, it builds on an earlier suggestion ( @xcite ) that black hole spin is the mechanism determining the _ direction _ of the radio jet ( although the black hole spin itself may ultimately be tied to the angular momentum of mergers or collisions ; see @xcite ) .
because black hole rotational energy drives the radio jet , the model predicts spindown and , consequently , cosmic evolution of the fr ii space density on a time scale of order @xmath166 or less .
it also suggests a possible model for radio quiet quasars ( schwarzschild black holes ) and for why radio loud quasars display predominantly fr ii morphology . finally , the total amount of rotational energy liberated by the jets from a rapidly - rotating hole is more than adequate to account for the entire energy content of a typical extended radio source .
several observational and numerical experiments have been proposed to test this model .
the most important of these is the vlbi jet velocity test , in which objects with fr i morphology should have a distinctly slower ensemble - averaged jet speed than that for fr ii sources .
other tests involve a similar effect for the black hole spin , a prediction for the typical density of a black hole accretion disk corona , and a search at high redshift for a population of faint extended radio sources associated with spiral or pre - spiral bulges . proposed
numerical experiments include general relativistic simulations confirming that the magnetic switch process works and that output mhd power depends critically on black hole spin .
* note added in proof : * in discussing the literature on the expected power output of a rotating black hole , the author omitted an important reference .
p. ghosh & m. a. abramowicz ( 1997 , _ mon . not .
royal astron .
_ * 292 * , 887 ) also discussed this subject at length , assuming , as is done here , that @xmath167 , and obtained expressions for the mhd power output of a rotating hole similar to those in the present paper .
( note that these authors took all of this power to be due to the blandford - znajek process occurring very near the horizon . )
however , the absolute magnitude of their power , is about two orders of magnitude smaller than that in equation ( 12 ) . part of this difference is due to a factor of @xmath168 in the definition of the mhd power ( which was discussed by these authors ) , but most of it is due to the present paper recognizing that the accretion disk can extend well interior to @xmath169 and even inside the ergosphere for rapidly rotating black holes .
this yields much larger magnetic field strengths near the static limit and black hole horizon .
in fact , one way in which ghosh & abramowicz ( 1997 ) , livio _ et al . _
( 1999 ) , and this paper can be brought into rough agreement is to 1 ) extend the disk into the ergosphere ( which would increase the power computed by ghosh & abramowicz ) and 2 ) treat that disk as an advection - dominated one with @xmath170 , yielding @xmath167 ( which would increase the power calculated by livio , ogilvie , & pringle ) .
the author is pleased to acknowledge the following : discussions with d. jones , s. koide , and d. murphy ; comments made by r. d. blandford and e. s. phinney during a seminar on this subject by the author ; and helpful e - mail exchanges with e. fomalont , c. jackson , and r. khanna .
this research was carried out at the jet propulsion laboratory , california institute of technology , under contract to the national aeronautics and space administration .
[ msd_appendix ]
this appendix shows that shear in a general relativistic metric , like that displayed in kerr geometry , can serve to enhance the local magnetic field strength .
magnetized plasma accreting onto a rotating black hole , therefore , will extract energy from the hole , converting it into enhanced local magnetic field strength .
the effect occurs even in absence of shear in the plasma fluid flow itself and even if the magnetic field lines do not thread the black hole . in a classical
mhd flow the magnetic field evolution equation , derived from faraday s and ohm s laws with infinite conductivity , is @xmath171 where @xmath172 and @xmath173 are the magnetic field and velocity vectors .
for an initially poloidal field and rotational flow only , the @xmath174 component of this equation becomes , in spherical - polar coordinates , @xmath175\ ] ] where @xmath176 is the angular velocity field of the flow .
thus , if there is shear in a classical rotating magnetized fluid flow , a poloidal magnetic field will serve as a seed for generation of a potentially large azimuthal magnetic field that grows with time .
this `` omega - effect '' is a necessary ( though not sufficient ) condition for an `` alpha - omega '' dynamo to exist ( @xcite ) . according to cowling s theorem ( @xcite ) ,
for a full dynamo , there also must be non - axisymmetric poloidal fluid motions ( due to turbulence , for example ) that enhance the poloidal magnetic field as well .
such turbulent motions are thought to be a common occurrence when the rotating shear flow occurs in a gravitational field ( @xcite ; @xcite ) .
so , the presence of an omega - effect in a gravitational field is a strong indication that an alpha - omega dynamo may be operating and generating a substantial magnetosphere .
the theory of gravito - magnetic dynamos in the kerr metric has been developed by @xcite , raising the possibility that even an _ axisymmetric _ flow can have growing , self - excited modes .
although such modes so far have not been found in kinematic numerical simulations ( @xcite ; @xcite ) , the non - validity of cowling s theorem in a general relativistic metric has been analytically confirmed by @xcite , who showed that there are growing axisymmetric modes when steep gradients exist in the plasma angular velocity @xmath20 relative to the rotating space . here
, however , we consider the simpler situation where the plasma flow itself does not shear , but there still is shear due to the general relativistic effect of frame - dragging by a rotating black hole .
while this does not demonstrate the existence of a dynamo in the system , it does satisfy several of the necessary conditions for such in particular , an omega - effect occurring in a gravitational field . to compute the evolution of the azimuthal magnetic field , we use two different approaches
one that uses the covariant four - dimensional formalism of @xcite ( mtw ) and one that uses the `` 3 + 1 '' decomposition of @xcite ( tpm ) . both give insight into the magnetic field enhancement process .
the general relativistic analogy of equation ( [ nonrel_b_evol_eq ] ) is ( @xcite ) @xmath177 which uses the standard mtw conventions ( einstein summation , semicolon to signify covariant differentiation , and comma for ordinary differentiation ) .
expanding the covariant derivatives and solving for evolution of @xmath178 @xmath179 where the @xmath180 are the connection coefficients and the antisymmetric commutation coefficients are given by the components of basis vector commutators @xmath181^{\alpha}\ ] ] the metric shear @xmath182 vanishes for @xmath178 expressed in a coordinate basis .
but it does not for a physical , locally - lorentz , orthonormal frame ( one in which local observers measure physical quantities , and in which magnetic field has the units of gauss for _ all _ components ) . in such an orthonormal frame , where we denote the coordinates with a `` hat '' accent , equation ( [ gen_metric_shear_eq ] ) becomes @xmath183 where @xmath184 is a coordinate transformation that diagonalizes the metric , converting to a locally - lorentz system , with inverse @xmath185 @xmath186 to compute @xmath187 , and therefore the evolution of the magnetic field for rotating black hole magnetospheres , we use the zero - charge boyer - lindquist version of the kerr metric ( mtw ) written in the following form @xmath188 ^ 2\ ] ] where the lapse function and angular velocity of frame dragging are given by @xmath189 the various geometric functions are given by @xmath190 @xmath191 is the specific angular momentum , and physical quantities are now in geometric units ( @xmath192 ) .
this metric can be diagonalized with many transformations , of course .
we choose one that rotates with the same angular velocity @xmath193 as the frame - dragged space .
this is the so - called zero angular momentum observer ( zamo ) system of @xcite .
the non - zero elements of this transformation and its inverse are @xmath194 with the @xmath195 and @xmath196 axes parallel to @xmath197 and @xmath198 , respectively . to illustrate that the metric shear acts like a fluid shear flow , consider the situation where there is no plasma motion with respect to the orthonormal system , so that the fluid four - velocity is given by @xmath199 . expressing equation ( [ rel_b_evol_eq ] ) in the orthonormal zamo frame , the evolution of the @xmath200 component of the magnetic field , using equation ( [ ortho_metric_shear_eq ] ) ,
is @xmath201\end{aligned}\ ] ] with the time and azimuthal parts vanishing . comparing with equation ( [ nonrel_bp_evol_eq ] ) ,
then , we see that there is an exact analogy with the classical rotating fluid shear case .
any poloidal magnetic field line penetrating the space near the rotating black hole is sheared into azimuthal field at the expense of the rotational energy of the black hole , whether or not the magnetic field line actually threads the hole .
this is independent of any shear in the fluid flow itself that may or may not be present . equation ( [ 4d_bp_evol_eq ] ) also can be derived using the `` 3 + 1 '' language of @xcite ( tpm ) . in that case the metric equation ( [ 4d_kerr_metric_eq ] ) is expressed in terms of the lapse and shift functions and a 3-metric @xmath202 where @xmath203 raises or lowers indices on the shift 3-vector @xmath204 and 3-velocity @xmath205 .
comparing ( [ 4d_kerr_metric_eq ] ) and ( [ 3p1_kerr_metric_eq ] ) gives @xmath206 , and components of ( the diagonal ) @xmath203 can be determined similarly .
physical quantities like magnetic field and velocity vectors are measured in a reference frame that rotates with the zamos , but they and the gradient operator @xmath207 are expressed in terms of absolute ( boyer - lindquist ) spatial and time coordinates . under these conditions ,
the general relativistic analog to equation ( [ nonrel_b_evol_eq ] ) is ( tpm ; @xcite ) @xmath208 where @xmath209 represents the lie derivative along @xmath210 @xmath211 and the frozen - field condition @xmath212 and solenoidal condition @xmath213 have been applied .
we again consider the situation where there is no fluid motion with respect to the rotating space ( @xmath214 ) .
furthermore , because the gradients now are expressed in a coordinate frame , the shear terms in equation ( [ lie_der_eq ] ) vanish , and the lie derivative becomes simply @xmath215 . as a result ,
the azimuthal component of equation ( [ 3p1_b_evol_eq ] ) becomes , in boyer - lindquist ( un - hatted ) coordinates , @xmath216 where @xmath217 only , since @xmath218 .
the left - hand side is proportional to the time derivative in the zamo frame ( @xmath219 ) , so transformation of equation ( [ 3p1_bp_evol_eq ] ) to the locally - lorentz orthonormal coordinates yields an equation identical to ( [ 4d_bp_evol_eq ] ) .
that is , even though the 3 + 1 formalism is expressed in the boyer - lindquist coordinate frame , the metric shear still is present ; but that shear is now embodied in the shift vector @xmath210 , rather than in the commutators @xmath182 .
since the sheared space near a kerr black hole is so similar in nature to that of a sheared accretion disk ( _ i.e. _ , rotating shear flow in a gravitational field ) , it is likely that an instability similar to the balbus - hawley one also is operating .
thus , there should be not only an omega - effect enhancing the azimuthal magnetic field , but also an alpha - effect , due to the magneto - rotational - generated turbulence , enhancing the poloidal field as well .
since this process draws its energy from the hole s rotation , the strength of the magnetosphere is expected to continue to grow long after all the accretion disk s angular momentum has been removed , and to levels far greater than could be supported by the accreting matter alone .
bicknell , g. v. 1985 , proc .
, 6 , 130 .
bicknell , g. v. 1995 , , 101 , 29 .
blandford , r. d. & payne , d. g. 1982 , , 199 , 883 ( bp ) .
blandford , r. d. & znajek , r. 1977 , , 179 , 433 ( bz ) .
brandenburg , a. 1996 , , 465 , l115 .
brandenburg , a. , nordlund , a. , stein , r.f .
& torkelsson , u. 1996 , , 458 , l45 .
baum , s. a. , zirbel , e. l. , & odea , c. p. 1995 , , 451 , 88 .
camenzind , m. 1989 , in accretion disks and magnetic fields in astrophysics , ed .
g. belevedere , ( dordrecht : kluwer ) , p. 129 .
cowling , t. g. 1934 , , 94 , 39 .
deyoung , d. 1976 , ann .
, 14 , 447 .
faber , s. m. et al .
1987 , in nearly normal galaxies , proc .
8th santa cruz summer wkshp on astron . & astrophys . , ( new york : springer - verlag ) , p. 175 .
fanaroff , b. l. & riley , j. m. 1974 , , 164 , 31p .
hayashi , m. r. , shibata , k. , & matsumoto , r. 1996 , , 468 , l37 .
henriksen , r. n. 1989 , in accretion disks and magnetic fields in astrophysics , ed .
g. belevedere , ( dordrecht : kluwer ) , p. 117 .
jackson , c. a. & wall , j. v. 1999 , , 304 , 160 .
kellermann , k. i. et al .
1989 , , 98 , 1195 .
khanna , r. & camenzind , m. 1996a , , 307 , 665 .
khanna , r. & camenzind , m. 1996b , , 313 , 1028 .
koide , s. , shibata , k. , & kudoh , t. 1998a , , 495 , l63 .
koide , s. , shibata , k. , & kudoh , t. 1998b , , submitted .
kormendy , j. & richstone , d. 1995 , ann .
, 33 , 581 .
kudoh , t. & shibata , k. 1996 , , 452 , l41 .
laing , r. a. 1996 , in energy transport in radio galaxies and quasars , eds .
e. hardee , a. h. bridle , and j. a. zensus , asp conf .
series vol .
100 , p. 241 .
leahy , j. p. 1991 , in beams and jets in astrophysics , ed .
p. hughes , ( cambridge : cambridge univ . press ) , p. 100 .
meier , d. l. , payne , d. g. , & lind , k. r. 1996 , in iau symposium 175 : extragalactic radio sources , eds .
r. ekers , c. fanti , and l. padrielli , ( dordrecht : kluwer ) , p. 433 .
meier , d. l. 1996 , , 459 , 185 .
meier , d. l. , edgington , s. , godon , p. , payne , d. g. , & lind , k. r. , 1997a , nature , 388 , 350 .
( megpl ) meier , d. l. et al .
1997b , in iau colloquium 164 : radio emission from galactic and extragalactic compact sources , eds .
j. a. zensus , g. b. taylor , & j. m. wrobel , asp conf .
series vol .
144 , p. 51 .
misner , c. w. , thorne , k. s. , & wheeler , j. a. 1973 , _ gravitation _
( san francisco : freeman ) ( mtw ) .
owen , f. n. & laing , r. a. 1989 , , 238 , 357 .
owen , f. n. & white , r. a. 1991 , , 249 , 164 .
punsly , b. & coroniti , f. v. 1990a , , 350 , 518 .
punsly , b. & coroniti , f. v. 1990b , , 354 , 583 ( pc ) .
rees , m. j. 1978 , nature , 275 , 516 .
robson , i. 1992 , in variability of blazars , eds .
e. valtaoja and m. valtonen , ( cambridge : cambridge univ .
press ) , p. 111 .
romanova , m. m. et al .
1998 , , 500 , 703 .
roberts , p. h. 1993 , in astrophysical fluid dynamics , eds . j .-
zahn & j. zinn - justin , ( amsterdam : elsevier ) , p. 229 .
schmidt , m. 1972 , 176 , 289 .
shakura , n. i. & sunyaev , r. a. 1973 , , 24 , 337 .
sobczak , g. j. , mcclintock , j. e. , remillard , r. a. , bailyn , c. d. , & orosz , j. a. 199 , , in press .
stone , j.m . ,
hawley , j.f .
, gammie , c.f .
, & balbus , s.a .
1996 , , 463 , 656 .
tavani , m. et al 1996 , , 473 , l103 .
thorne , k. s. , price , r. h. , & macdonald , d. a. 1986 , black holes : the membrane paradigm , ( new haven : yale univ .
press ) .
vermeulen , r. c. 1996 , in energy transport in radio galaxies and quasars , eds .
e. hardee , a. h. bridle , and j. a. zensus , asp conf .
series vol .
100 , p. 117 .
wall , j. v. & jackson , c. a. 1997 , , 290 , l17 .
wilson , a. s. & colbert , e. j. m. 1995 , , 438 , 62 .
windhorst , r. a. et al 1995 , nature , 375 , 471 .
zhang , s. n. , cui , w. , & chen , w. 1997 , , 482 , l155 . | a model is presented in which both fanaroff and riley class i and ii extragalactic jets are produced by magnetized accretion disk coronae in the ergospheres of rotating black holes .
it employs a hybrid version of the blandford - payne and blandford - znajek magnetohydrodynamic mechanisms ( similar to the punsly - coroniti model , with the addition of a metric - shear - driven dynamo ) and a generalized form of the magnetic switch , which is shown to be the mhd analog of the eddington luminosity .
while the jets are produced in the ergospheric accretion disk itself , the output power still is an increasing function of the black hole angular momentum . for high enough spin
, the black hole triggers the magnetic switch , producing highly - relativistic , kinetic - energy - dominated jets instead of magnetic - energy - dominated ones for lower spin .
the coronal mass densities needed to trigger the switch at the observed fr break power are quite small ( @xmath0 ) , implying that the source of the jet material may be either a pair plasma or very tenuous electron - proton corona , not the main accretion disk itself . the model explains the differences in morphology and mach number between fr i and ii sources and the observed trend for massive galaxies ( which contain more massive black holes ) to undergo the fr i / ii transition at higher radio power .
it also is consistent with the energy content of extended radio lobes and explains why , because of black hole spindown , the space density of fr ii sources should evolve more rapidly than that of fr i sources .
a specific observational test is proposed to distinguish between models like this one , in which the fr i / ii division arises from processes near the black hole , and models like bicknell s , in which the difference is produced by processes in the host galaxy s interstellar medium .
if the present model is correct , then the ensemble average speed of parsec - scale jets in sources distinguished by their fr i _ morphology _ ( not luminosity ) should be distinctly slower than that for sources with fr ii morphology .
the model also suggests the existence of a population of high - redshift , sub - mjy fr i and ii radio sources associated with spiral or pre - spiral galaxies that flared once when their black holes were formed but were never again re - kindled by mergers .
# 1 # 1 |
the generation of neutrino masses is an important issue in any realistic extension of the standard model . in general , the values of these masses ( of the order of , or less than , 1 ev ) that are needed to explain all neutrino oscillation data @xcite are not enough to put strong constraints on model building .
it means that several models can induce neutrino masses and mixing compatible with experimental data .
so , instead of proposing models built just to explain the neutrino properties , it is more useful to consider what are the neutrino masses that are predicted in any particular model which has motivation other than the explanation of neutrino physics .
for instance , the 3 - 3 - 1 model was proposed as a possible symmetry on the lightest lepton sector @xmath1 @xcite . once assumed that symmetry it has to be implemented in the rest of the leptons and also in the quark sector .
like in the standard model if we do not introduce right - handed neutrinos and/or violation of the total lepton number neutrinos remain massless at any order in perturbation theory . in this vain
it has been done some effort to produce neutrino masses in the context of that 3 - 3 - 1 model and some of its extensions @xcite . in this work
we consider the generation of neutrino masses in a supersymmetric 3 - 3 - 1 model with broken @xmath0 parity .
we show that , as an effect of the mixing among all leptons of the same charge , at the tree level only one charged lepton and one neutrino remain massless but they gain mass through radiative corrections . in order to compare this model
we do the same calculations in the context of the minimal supersymmetric standard model with @xmath0 broken parity also . in both cases
we are not assuming that sneutrinos gain non - vanishing vacuum expectation values ( vevs ) i.e. , the only non - zero vevs are those of the scalars of the non - supersymmetric models .
the outline of this work is as follows . in sec .
[ sec : mssm ] we review the origin of the lepton masses in the minimal supersymmetric standard model context under the same assumptions that we will use in the case of the 3 - 3 - 1 supersymmetric model . in sec .
[ sec:331s ] we consider the supersymmetric version of a 3 - 3 - 1 model which has only three triplets of higgs scalars .
we explicitly show that leptons gain mass only as a consequence of their mixing with gauginos and higgsinos .
our conclusions are found in the last section .
let us consider in this section the lepton masses in the minimal supersymmetric standard model ( mssm ) @xcite . in this model
the interactions are written in terms of the left - handed ( right - handed ) @xmath2 ( @xmath3 ) leptons , left - handed ( right - handed ) quarks @xmath4 @xmath5 ; and the higgs doublets @xmath6 . with those multiplets the superpotential that conserves @xmath0-parity is given by @xmath7 , where @xmath8 while the @xmath0-parity violating terms are given by @xmath9 , where @xmath10 and we have suppressed @xmath11 indices , @xmath12 is the antisymmetric @xmath11 tensor .
above , and below in the following , the subindices @xmath13 run over the lepton generations @xmath14 but a superscript @xmath15 indicates charge conjugation ; @xmath16 denote quark generations .
we have also to add the soft terms that break the supersymmetry : ' '' '' ' '' '' @xmath17 , \label{softmssm}\end{aligned}\ ] ] ' '' '' ' '' '' 2 where @xmath18 is an @xmath11 index and @xmath19 are the supersymmetric partners of the respective gauge vector bosons but we have omitted generation indices and the gluino - mass terms . with the interactions in eq .
( [ mssmrpc ] ) it is possible to give mass to all charged fermions in the model ( see below ) but neutrinos remain massless . hence , we must introduce @xmath0-parity violating term like those in eq .
( [ mssmrpv ] ) . some of the coupling constants in that expression should be set to zero in order to avoid a too fast proton decay . defining the basis @xmath20 , the mass term is @xmath21 $ ] , where @xmath22 is the mass matrix ' '' '' ' '' '' @xmath23 ' '' '' ' '' '' 2 with @xmath24 , @xmath25 , etc
are defined as @xmath26 and @xmath27 is the weak mixing angle .
the matrix in eq .
( [ mssm ] ) is generated only by the two usual vacuum expectation values of the two scalars and by the @xmath0-parity breaking terms @xmath28 .
the mass matrix is similar to that in ref .
@xcite but we have included the three neutrinos and we are neither assuming that sneutrinos gain nonzero vacuum expectation values nor have introduced sterile neutrinos like in ref .
@xcite . the mass matrix in eq .
( [ mssm ] ) has two zero eigenvalues : it has determinant equal to zero and its secular equation which give the eigenvalues , @xmath29 , has the form @xmath30 times a polynomial of five degree ; thus there are two neutrinos @xmath31 , which are massless at the tree level . using @xmath32 and @xmath33 gev , @xmath34 , @xmath35 ,
@xmath36 gev ( this value is consistent with that of ref .
@xcite ) , @xmath37 gev , @xmath38 gev , @xmath39 gev , we obtain besides the two massless neutrinos a massive one with @xmath40 ev , and four heavy neutralinos with masses @xmath41 gev .
these zero eigenvalues are a product of the matrix structure in eq .
( [ mssm ] ) and there is not a symmetry to protect the neutrinos to gain mass by radiative corrections . on the other hand , if @xmath42 , @xmath43 , all neutrinos remain massless at the tree level . in this case
it is the @xmath0-parity and total lepton number conservation that protect neutrinos of gain masses .
the neutralino masses above are consistent with those of ref .
@xcite : two states are massless and the other ones have masses of the order @xmath44 .
more realistic neutrino masses require radiative corrections @xcite . here
we will only consider the neutrino masses generated by radiative corrections arisen from the interactions given in eqs .
( [ mssmrpc ] ) and ( [ mssmrpv ] ) and only two vevs .
we have in this case the interactions @xmath45 and the 1-loop diagrams like those in ref .
@xcite arise .
notice however that if we introduce a discrete symmetry ( called @xmath46 later on ) , @xmath47 , and all other fields are even under this transformation , we have that @xmath48 and the @xmath49 neutrinos will remain massless at all order in perturbation theory .
it is also possible to choose the symmetry such as @xmath50 while all other fields remain invariant . in this case
we have that @xmath51 and @xmath52 remain massless . however , if no discrete symmetry is imposed neutrinos gain mass through 1-loop effect like in ref .
@xcite .
next , let us consider the charged sector . with the interactions in eq .
( [ mssmrpc ] ) it is possible to give mass to all charged fermions . denoting @xmath53 where all the fermionic fields are still weyl spinors
, we can define @xmath54 , and the mass term @xmath55 $ ] where @xmath56 is the mass matrix given by : @xmath57 with ' '' '' ' '' '' @xmath58 ' '' '' ' '' '' 2 with @xmath59 , @xmath60 , @xmath61 , @xmath62 we obtain from eq .
( [ clmmmssm ] ) the masses @xmath63 ( in gev ) for the usual leptons , and 4.3 and 81 tev for the charginos .
we see by comparing eq .
( [ mssm ] ) with eq .
( [ clmmmssm ] ) that there is no relation between the charged lepton masses and the neutralino masses .
notice also that all charged leptons gain masses at the tree level . we will not consider this model ( or some of its extensions ) further since it has been well studied in literature @xcite .
in the nonsupersymmetric 3 - 3 - 1 model @xcite the fermionic representation content is as follows : left - handed leptons @xmath64 , @xmath43 ; left - handed quarks @xmath65 , @xmath66 , @xmath67 , @xmath68 ; and in the right - handed components we have @xmath69 , that transform as in the sm , and the exotic quarks @xmath70 .
the minimal scalar representation content is formed by three scalar triplets : @xmath71 ; @xmath72 and @xmath73 , and one scalar sextet @xmath74 .
we can avoid the introduction of the sextet by adding a charged lepton transforming as a singlet @xcite .
notwithstanding , here we will omit both the sextet and the exotic lepton
. a see - saw - type mechanism will be implemented by the mixing with supersymmetric partners , higgsinos or gauginos .
the complete set of fields in the 3 - 3 - 1 supersymmetric model has been given in refs .
we will denote , like in the previous section , the respective superfields as @xmath75 and so on .
we recall that in the nonsupersymmetric 3 - 3 - 1 model with only the three triplets the charged lepton masses are not yet the physical ones : @xmath76 .
we will show how , in the present model supersymmetry and the @xmath0-violating interactions give the correct masses to @xmath77 and @xmath78 , even without a sextet or the charged lepton singlet .
we have the higgsinos @xmath79 and their respective primed fields which have the same charge assignment of the triplets @xmath80 and @xmath81 , for details see ref . @xcite . due to
the fact that in the supersymmetric model we have the gauginos and higgsinos , ( for details on the lagrangian of the model see @xcite ) , when the @xmath0-parity is broken we have in analogy with the mssm , but with important differences , a mixture between the usual leptons and the gauginos and higgsinos .
one part of the superpotential is given by @xmath82 where @xmath83 @xmath43 ; and @xmath84 where @xmath85 with @xmath12 the completely antisymmetric tensor of @xmath86 but we have omitted the respective indices ; the generation indices are as follows : @xmath87 and @xmath16 .
the gaugino masses come from the soft - terms shown in the appendix [ sec : a1 ] , eq .
( [ ess ] ) .
the @xmath88 , @xmath89 and @xmath90 terms break the @xmath0-parity defined in this model as @xmath91 where @xmath92 , b(l ) is the baryon ( total lepton ) number ; s is the spin .
the @xmath93 term of the superpotential @xmath94 implies interactions like ( see eq .
( [ nusint ] ) below ) @xmath95 and we have also the interactions @xmath96 where @xmath97 is the superfield related to the @xmath98 gauge boson of @xmath99 .
this interaction mixes higgsinos with gauginos as showed in ref .
@xcite . the parameters @xmath100 and @xmath101 are the equivalent of the @xmath102 parameter in the mssm @xcite .
the terms proportional to @xmath93 and @xmath103 have no equivalent in the mssm .
the @xmath104 and @xmath105 coupling constants are constrained by the proton decay such that @xcite @xmath106 assuming the superpartner masses in the range of 1 tev . in this model
there are interactions like @xmath107 \nonumber \\
& -&\frac{1}{2}\,\mu_{0a}[l\tilde{\eta}_1^{\prime+}+ \bar{l}\overline{\tilde{\eta}_1^{\prime+}}+ l^c\tilde{\eta}_2^{\prime- } + \bar{l^c}\overline{\tilde{\eta}_2^{\prime- } } ] , \label{newint}\end{aligned}\ ] ] which imply a general mixture in both neutral and charged sectors .
let us first considered the charged lepton masses . denoting @xmath108 where all the fermionic fields are still weyl spinors , we can also , as before , define @xmath109 , and the mass term @xmath110 $ ] where @xmath56 is given by : @xmath111 with ' '' '' ' '' '' @xmath112 ' '' '' ' '' '' 2 where we have defined @xmath113 the chargino mass matrix @xmath56 is diagonalized using two unitary matrices , @xmath114 and @xmath115 , defined by @xmath116 ( @xmath114 and @xmath115 sometimes are denoted , in non - supersymmetric theories , by @xmath117 and @xmath118 , respectively ) .
then we can write the diagonal mass matrix as @xmath119 to determine @xmath115 and @xmath114 , we note that @xmath120 and define the following dirac spinors : @xmath121 where @xmath122 is the particle and @xmath123 is the anti - particle @xcite .
we have obtained the following masses ( in gev ) for the charged sector : @xmath124 and the masses for the usual leptons ( in gev ) @xmath125 , @xmath126 and @xmath127 .
these values have been obtained by using the following values for the dimensionless parameters [ valpara ] @xmath128 @xmath129 and for the mass dimension parameters ( in gev ) we have used : @xmath130 @xmath131 we also use the constraint @xmath132 coming from @xmath133 , where , we have defined @xmath134 and @xmath135 . assuming that @xmath136 gev , @xmath137 gev , and @xmath138 tev , the value of @xmath139 is fixed by the constraint above .
notice , from eq .
( [ cm ] ) , that the electron is massless at the tree level .
this is again a result of the structure of the mass matrix in eq .
( [ clmmmssm ] ) and there is not a symmetry that protects the electron to get a mass by loop corrections .
hence , it can gain mass trough radiative corrections like that shown in fig .
the interactions of the leptons with the sleptons written in term of dirac fermions ( although we are using the same notation ) are given by ( and the respective hermitian conjugate ) ' '' '' ' '' '' @xmath140 + \overline{\tilde{\eta}^+_{1r}}\ , ( l^c_{al}\tilde{\nu}_{bl}-l^c_{bl}\tilde{\nu}_{al})\right ] \nonumber \\ & + & \frac{\lambda_{3a}}{3}\left [ -\overline{\tilde{\rho}^{--}_r}l_{al}\chi^-+ \bar{\tilde\rho}^-_r\ , ( l_{al}\chi^0- l^c_{al}\chi^{--})- \overline{\tilde{\chi}^+_r}\,(l_{al}\rho^{++}-l^c_{al}\rho^0)- \overline{\tilde{\chi}^{++}_r}l^c_{al}\rho^++ \overline{\tilde{\rho}^0_r}l^c_{al}\chi^{- } + \overline{\tilde{\chi}^0_r}l_{al}\rho^+\right]\nonumber \\ & + & \frac{\lambda^\prime_{\alpha ai}}{2 } \left[(\bar{u}_{\alpha r}l_{al}+ \bar{j}_{\alpha r}l^c_{al } ) \tilde{d}^*_i + \bar{d}^c_{ir}\ , ( l_{al}\tilde{u}_\alpha + l^c_{al}\tilde{j}_\alpha)\right ] .
\label{clint}\end{aligned}\ ] ] ' '' '' ' '' '' 2 the @xmath104 interactions generate the low vertices in fig . [ fig1 ] . on the other hand ,
the interactions between the squarks , sleptons and scalars , see the appendix [ sec : a1 ] , are given by the scalar potential .
the soft part contributes only through the trilinear interactions @xmath141 while the @xmath114-terms have only quartic interactions @xmath142 where @xmath143 ; and @xmath43 . from the @xmath144-terms we have contributions to both , the trilinear interactions @xmath145 and the quartic ones ' '' '' ' '' '' @xmath146 ' '' '' ' '' '' 2 where @xmath147 which will contribute to the upper quartic vertex in fig .
[ fig1 ] . due to the interactions given in eq .
( [ clint])([ft4 ] ) , we can generate the appropriate mass to the electron .
the dominant contributions , assuming the mass hierarchy @xmath148 where fermion means a fermion different from @xmath149 and scalar means @xmath150 ( @xmath151 denote the heaviest higgs scalar ) and using the values of the masses and the parameters given in eqs .
( [ cm ] ) , ( [ lambdas ] ) and ( [ para ] ) we obtain that the dominant contribution to the electron mass is , up to logarithmic corrections , @xmath152 and with all the indices fixed , @xmath153 denotes mixing matrix elements in the two dimension @xmath149 space , @xmath154 means the same but in the d - like squark sector .
we obtain @xmath155 gev if @xmath156 and @xmath157 have the values already giving above , and with @xmath158 , which imposes @xmath159 or @xmath160 gev . using @xmath161 gev @xcite
, we have @xmath162 gev . with @xmath163
we obtain squarks masses of the order of @xmath164 gev @xcite . like in the case of the charged sector ,
the neutral lepton masses are given by the mixing among neutrinos , induced by the @xmath28 term in eq .
( [ e2 ] ) , and the neutral higgsinos and gauginos @xcite , and also by @xmath93 and @xmath165 in eq .
( [ e22 ] ) .
the first two terms in eq .
( [ e2 ] ) give the interactions between neutrinos and higgsinos : @xmath166- \frac{\mu_\eta}{2}\,[\tilde{\eta}^{\prime0}\tilde{\eta}^{0}+ \overline{\tilde{\eta}^{\prime0 } } \,\ \overline{\tilde{\eta}^{0 } } ] , \label{e3}\ ] ] and from eq .
( [ e22 ] ) we have also the interactions @xmath167.\ ] ] these interactions imply a mass term for the neutrinos and neutralinos . the mass term in the basis @xmath168 is given by @xmath169 $ ] where ' '' '' ' '' '' @xmath170 ' '' '' ' '' '' 2 all parameters in eq .
( [ mmn ] ) , but @xmath171 , are defined in eqs .
( [ vevs ] ) , ( [ lambdas ] ) and ( [ para ] ) ; @xmath172 and @xmath173 denote the gauge coupling constant of @xmath99 and @xmath174 , respectively .
the neutralino mass matrix is diagonalized by a @xmath175 rotation unitary matrix @xmath176 , satisfying @xmath177 and the mass eigenstates are @xmath178 we can define the following majorana spinor to represent the mass eigenstates @xmath179 as above the subindices @xmath13 run over the lepton generations @xmath14 . with the mass matrix in eq .
( [ mmn ] ) , at the tree level we obtain the eigenvalues ( in gev ) , @xmath180 and for the three neutrinos we obtain ( in ev ) @xmath181 we have got the values in eqs .
( [ hnm ] ) and ( [ mnus1 ] ) by choosing , besides the parameters in eqs .
( [ lambdas ] ) and ( [ para ] ) , @xmath182 gev .
notice that the coupling constant @xmath173 and the parameter @xmath171 appear only in the mass matrix of the neutralinos , all the other parameters in eq .
( [ mmn ] ) have already been fixed by the charged sector , see eq .
( [ clmm ] ) , ( [ lambdas ] ) and ( [ para ] ) .
the neutrino masses in eq .
( [ mnus1 ] ) are of the order of magnitude for lsnd and solar neutrino data . on the other hand
if we choose @xmath182 gev and @xmath183 gev , we obtain ( in ev ) @xmath184 which are of the order of magnitude required by the solar and atmospheric neutrino data .
notice also the sensibility of the neutrino masses in @xmath185 and @xmath171 , and that if @xmath42 , @xmath43 , all neutrinos remain massless .
the masses of the charged sector are insensible to the values of @xmath28 for all @xmath43 which are suitable for generating the appropriated different neutrino mass spectra ( see the appendix [ sec : a2 ] ) .
we have obtained numerically the unitary matrices @xmath186 and @xmath176 which diagonalize the mass matrices in eq .
( [ clmm ] ) and ( [ mmn ] ) but we will not write them explicitly . the charged current is written in the mass - eigenstate basis as @xmath187 with the maki - nakagawa - sakata matrix @xcite defined as @xmath188 , where @xmath189 and @xmath190 are the @xmath191 submatrices of @xmath115 and @xmath176 , respectively . hence , we have : @xmath192 notice that this leptonic mixing matrix is not orthogonal as it must be since we are omitting the mixture with the heavy charginos and neutralinos and that one neutrino remains massless at the tree level .
we can always rotate the neutral fields in such a way that the electron neutrino is the one which remains massless ; or we can also assume @xmath193 , so that the electron neutrino decouples from the other neutrinos and neutralinos . in this case , diagonal and non - diagonal mass terms in eq .
( [ mmn ] ) will be induced by loop corrections like that in fig .
thus , a @xmath191 non - orthogonal mixing matrix will appear in eq .
( [ mns ] ) .
here we will only consider the order of magnitude of a mass generated by this process .
the massless neutrino can get a mass from the loop correction like that in fig .
[ fig2 ] as a consequence of the majorana mass term of the neutral lepton in the triplet .
this is equivalent to the mechanism of ref .
@xcite but now with a triplet of leptons instead of a neutral singlet .
for instance , the @xmath93 interactions will contribute in the left- and right vertices in fig .
[ fig2 ] : ' '' '' ' '' '' @xmath194+\frac{\lambda_{3a}}{3 } \left[\overline{\tilde{\chi}^{++}_r}\ , \nu_{al}\rho^{++}+ \overline{\tilde{\rho}^{--}_r}\nu_{al}\chi^{--}- \overline{\tilde{\chi}^0_r}\nu_{al}\rho^0- \overline{\tilde{\rho}^0_r}\nu_{al}\chi^0\right]\nonumber \\ & + & \frac{\lambda^\prime_{\alpha ai}}{3}\left[\bar{d}_{\alpha r}\nu_{al}\tilde{d}^c_i+ \bar{d}^c_{ir}\nu_{al}\tilde{d}_\alpha\right]+h.c .
, \label{nusint}\end{aligned}\ ] ] ' '' '' ' '' '' 2 these interactions generate the lower vertices of fig . [ fig2 ] .
the upper vertex are given in eqs .
( [ softnew])([ft4 ] ) . with these interactions
we can generate the following small mass to the electron neutrino .
in fact , assuming a hierarchy of the masses as in the last subsection , we obtain the dominant contribution to the @xmath51 mass , up to logarithmic corrections : @xmath195 where all the indices are fixed , @xmath196 is the mixing matrix element defined in eqs .
( [ m1 ] ) and ( [ m2 ] ) , @xmath197 denotes the mixing matrix element in the slepton sector . the charged lepton which gives the main contribution is the @xmath78 lepton
: it has a large mass and the mixing angle is not too small , in fact @xmath198 . since , @xmath199 we obtain an electron neutrino mass of the order of @xmath200 ev if @xmath201 gev . if @xmath202 we have @xmath203 gev @xcite .
in the nonsupersymmetric 3 - 3 - 1 model @xcite with only three scalar triplets @xmath204 it is not possible to generate the observed charged lepton masses .
then , it is necessary to introduce a scalar sextet in order to get the appropriate masses . when we supersymmetrized the model and allow @xmath0-parity breaking interactions we can give to all known charged leptons and neutrinos the appropriate masses even without the introduction of a scalar sextet . of course
, in order to cancel anomalies we have to introduce another set of three triplets @xmath205 . in this case although the correct values for the lepton masses can still be obtained , if the new vevs @xmath206 and @xmath207 are zero , it was shown in ref .
@xcite that in order to give mass to all the quarks in the model all these vevs have to be different from zero .
hence , we have considered that the six neutral scalar components got a nonzero vev . as can be seen from eq .
( [ valpara ] ) , the charged lepton masses arise from a sort of seesaw mechanism since there are small mass parameters , as in eq .
( [ mus0 ] ) , related with @xmath0-parity breaking interactions , and large ones as in eq .
( [ para ] ) , related with the mass scale of the supersymmetry breaking , this can be better appreciated in eqs .
( [ clmmn ] ) .
the same happens in the neutrino sector , see eq .
( [ mmnn ] ) . in a supersymmetric version of the model in which we add the sextet @xmath208
, there is a fermionic non - hermitian triplet under @xmath209 that is part of a sextet under @xmath210 .
this can also implement a seesaw mechanism for neutrino masses as it was pointed out in ref .
( @xcite ) .
the case of a hermitian fermion triplet was considered in the context of the standard model in refs .
@xcite .
it is interesting to note that in the context of mssm a @xmath211 symmetry @xcite , @xmath212 where @xmath213 is a matter , vector and scalar superfields , respectively , forbids the @xmath0-parity breaking terms in eq .
( [ mssmrpv ] ) . in the present model it happens the same :
the @xmath0-parity breaking terms in eqs .
( [ e2 ] ) and ( [ e22 ] ) are forbidden .
notwithstanding the @xmath214 symmetry @xcite , @xmath215 where @xmath216 , forbids the @xmath217 violating terms but allow the @xmath218 violating ones .
this also happens in the present model .
however , if we introduce an extra discrete @xmath46-symmetry , such that @xmath219 , and all other fields being even under this transformation , we have that @xmath220 , at all orders in perturbation theory .
this does not modify the mass matrix in the charged sector in eqs .
( [ ypm ] ) and ( [ clmm ] ) , but forbids the electron neutrino to get a mass , at all orders in perturbation theory . the present model will induce processes contributing to @xmath221 , @xmath222 , @xmath223 , and other exotic decays .
however these processes can be suppressed mainly by the scalar masses since these scalars do not enter explicitly in the mass matrix at the tree level .
some contributions to those processes are suppressed by the coupling constants themselves , like @xmath224 s in eq .
( [ lambdas ] ) , other ones which involve @xmath225 which are of the order unity can be suppressed by combining the mixing angles and masses of scalars or charginos sectors . a more detailed study of this issue will be done elsewhere .
@xcite in summary , we have analyzed the charged lepton and neutrino masses in a @xmath0-parity breaking supersymmetric 3 - 3 - 1 model .
unlike the mssm model the electron and its neutrino remain massless at the tree level but gain masses at the one loop level .
the resulting leptonic mixing matrix @xmath226 is non - orthogonal .
this work was supported by fundao de amparo pesquisa do estado de so paulo ( fapesp ) , conselho nacional de cincia e tecnologia ( cnpq ) and by programa de apoio a ncleos de excelncia ( pronex ) . ' '' '' ' '' ''
the interactions between the scalars of the theory is given by the scalar potential that is written as @xmath227 where the @xmath228 term is given by @xmath229 the @xmath144 term is @xmath230 , \label{esf}\end{aligned}\ ] ] finally , the soft term is ( the following soft - terms do not included the exotic quarks ) @xmath231 . \label{ess}\end{aligned}\ ] ] ' '' '' ' '' '' 2
here we show explicitly the numerical values of each entry of the mass matrices in eqs .
( [ clmm ] ) and ( [ mmn ] ) using the parameters given in eq .
( [ valpara ] ) and @xmath182 gev . for the charged sector
we have here we show that the relevant parameters for the leptons masses are @xmath233 .
we note that there are four type of parameters in the mass matrices in eqs .
( [ valpara ] ) .
firstly we have the dimensionless yukawa couplings in the usual leptons , @xmath233 and in the supersymmetric partners , @xmath234 .
we also have the mass dimension parameters @xmath28 and @xmath235 and @xmath236 and @xmath171 which are soft terms in eq .
( [ ess ] ) . of all these parameters
we expect that the relevant ones in the charged lepton and neutrinos are @xmath233 . to show this we consider several choices of the parameters as follows : ( below all masses are in gev ) + * case 1 : * @xmath237 charged sector masses : + 3186.03 , 3001.10 , 557.17 , 196.55,149.30 , 16.85 , 0 , 0 , 0 . + neutral sector masses : + -4162.22 , 3260.47 , 3001.10 , 557.79 , -557.17 , 450.14 , -330.68 , 17.18 , -17.01 , 0 , 0 , 0 .
* case 2 : * @xmath238 charged sector masses : + 3186.03,3001.10,557.17,196.55,149.30,16.85,1.92@xmath239 , 0 , 0 . + neutral sector masses : + -4162.22 , 3260.47 , 3001.10 , 557.79 , -557.17 , 450.14 , -330.68 , 17.18 , -17.01 , 2.80 @xmath240 , 0 , 0 .
* case 3 : * @xmath241 charged sector masses : + 3186.05 , 3001.11 , 584.85 , 282.30 , 149.41 , 204.55 , 2.10@xmath242 , 0 , 0 . + neutral sector masses : + -4162.22 , 3260.47 , 3001.10 , 585.18 , -585.18 , 453.22 , -344.14 , 283.14 , -271.99 , 1.23 @xmath243 , 0 , 0 .
* case 4 : * @xmath244 charged sector masses : + 3186.03 , 3001.10 , 557.17 , 196.55,149.30 , 16.85 , 1.85 , 1.85 , 0 . + neutral sector masses : + -4162.22 , 3260.47 , 3001.10 , 557.79 , -557.17 , 450.14 , -330.68 , 17.18 , -17.01 , 0 , 0 , 0 .
* case 5 : * @xmath245 charged sector masses : + 3186.03 , 3001.11 , 584.85 , 282.30 , 204.55 , 149.41 , 1.78 , 0.105 , 0 . + neutral sector masses : + -4162.22 , 3260.47 , 3001.10 , 585.19 , -585.19 , 453.22 , -344.14 , 283.14 , -271.99 , 1.23 @xmath243 , 0 , 0 . 99 y. fukuda _
( superkamiokande collaboration ) , phys .
lett . * 81 * , 1562 ( 1998 ) ; _ ibid _ , * 82 * , 2644 ( 1999 ) .
b. t. cleveland _
( homestake collaboration ) , astrophys .
j. * 496 * , 505 ( 1998 ) ; k. s. hirata _
( kamiokande collaboration ) , phys .
77 * , 1683 ( 1996 ) ; w. hampel _
( gallex collaboration ) , phys .
b477 * , 127 ( 1999 ) ; j. n. abdurashitov _ at al .
_ ( sage collaboration ) , phys . rev
. lett . * 77 * , 4708 ( 1996 ) ; phys .
c * 60 * , 055801 ( 1999 ) ; q. r. ahmad _ et al . _ ( sno collaboration ) , nucl - ex/0106015 . c. athanassopoulos _ et al . _
( lsnd collaboration ) , phys .
lett . * 77 * , 3082 ( 1996 ) ; _ ibid _ * 81 * , 1774 ( 1998 ) .
f. pisano and v. pleitez , phys .
d * 46 * , 410 ( 1992 ) ; p. h. frampton , phys .
* 69 * , 2889(1992 ) ; r. foot , o. f. hernandez , f. pisano and v. pleitez , phys . rev .
d * 47 * , 4158 ( 1993 ) .
u. okamoto and m. yasue , phys . lett . * b466 * , 267 ( 1999 ) ; t. kitabayashi and m. yasue , phys . rev . * d63 * , 095002 ( 2001 ) ; t. kitabayashi and m. yasue , phys . lett . *
b508 * , 85 ( 2001 ) ; t. kitabayashi and m. yasue , nucl .
* b609 * , 61 ( 2001 ) ; t. kitabayashi and m. yasue , phys . rev . * d63 * , 095006 ( 2001 ) ; j. c. montero , c. a. de s. pires and v. pleitez , phys . lett . *
502b * , 167 ( 2001 ) , hep - ph/0011296 ; hep - ph/0103096 ; hep - ph/0112203 . and
hep - ph/0112246 . h. e. haber and g. l. kane , phys .
rep . * 117 * , 75 ( 1985 ) .
l. j. hall and m. suzuki , nucl .
b231 * , 419 ( 1984 ) .
t. banks , y. grossman , e. nardi and y. nir , phys .
d * 52 * , 5319 ( 1995 ) .
m. a. diaz , j. c. romo and j. w. f. valle , nucl .
b524 * , 23 ( 1998 ) . f. borzumati and y. nomura , phys .
d * 64 * , 053005 ( 2001 ) ; f. borzumati , k. hamaguchi and t. yanagida , phys .
* b497 * , 259 ( 2001 ) ; f. borzumati , k. hamaguchi , y. nomura and t. yanagida , hep - ph/0012118 .
r. n. mohapatra , phys .
d * 34 * , 3457 ( 1986 ) .
j. c. romo and j. w. f. valle , nucl .
* b381 * , 87 ( 1992 ) . s. davison and m. losada , hep - ph/0010325 .
t. v. duong and e. ma , phys .
* b316 * , 307 ( 1993 ) .
j. c. montero , c. a. de s. pires and v. pleitez hep - ph/0103096 .
j. c. montero , v. pleitez and m. c. rodriguez , phys .
d * 65 * , 035006 ( 2002 ) , hep - ph/0012178 .
m. capdequi - peyranere and m.c .
rodriguez , phys .
d * 65 * , 035006 ( 2002 ) , hep - ph/0103013 . h. n. long and p. b. pal , mod
. phys . lett . * a13 * , 2355 ( 1998 ) .
p. das , p. jain and d. w. mckay , phys .
d * 59 * , 055011 ( 1999 ) .
d. e. groom _ et al .
_ ( particle data group ) , eur .
j. c * 15 * , 1 ( 2000 ) .
z. maki , m. nakagawa and s. sakata , prog .
theo . phys . * 28 * , 246 ( 1962 ) .
k. s. babu and e. ma , phys .
lett . * 61 * , 674 ( 1988 ) .
j. c. montero , c. a. de s. pires and v. pleitez , to be published in phys .
, hep - ph/0112246 .
r. foot , h. lew , x .-
he and g. c. joshi , z. phys .
c * 44 * , 1989 ) ; e. ma , phys .
lett . * 81 * , 1171 ( 1998 ) .
h. e. haber , hep - ph/0103095 . | we consider the mass generation for both charginos and neutralinos in a 3 - 3 - 1 supersymmetric model . we show that @xmath0-parity breaking interactions leave the electron and one of the neutrinos massless at the tree level .
however the same interactions induce masses for these particles at the 1-loop level . unlike the similar situation in the minimal supersymmetric standard model the masses of the neutralinos
are related to the masses of the charginos . 2 |
the evolution of star formation in the universe is described by determining the total luminosity density arising from young stars as a function of redshift ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
various studies have attempted to measure cosmic evolution of this star formation rate ( sfr ) by locating galaxies using ultraviolet , optical , and near - infrared techniques ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . because star - forming galaxies are discovered in various ways , it is vital to compare various methods for measuring the sfr in a galaxy
the galaxies which we are studying in the present paper have their observed multiwavelength spectral characteristics determined primarily by on - going , luminous star formation .
for this reason , we use the classification `` starburst galaxies '' , or `` starbursts '' . the starbursts which we include have sfrs ranging from @xmath9 0.1 m@xmath6 yr@xmath7to @xmath9 1000 m@xmath6 yr@xmath7 .
initial determinations of the evolution of sfr in the universe were based on optical observations which reveal rest - frame ultraviolet luminosities of starbursts at high redshifts @xcite .
uncertain corrections for extinction are the greatest limitation of such studies , and the significance of dust obscuration has become increasingly emphasized by the large populations of starbursts being discovered among the dusty sources found by the spitzer space telescope ( @xmath2 ) .
these `` dust obscured galaxies '' @xcite are revealed in various surveys at 24 @xmath0 m with the multiband imaging photometer for @xmath2 ( mips ; rieke et al .
2004 ) combined with 3 @xmath0 m to 8 @xmath0 m photometry with the @xmath2 infrared array camera ( irac ; fazio et al .
2004 ) .
luminous , dusty galaxies discovered with mips and irac have been characterized by spectra with the infrared spectrograph on @xmath2 ( irs ; houck et al .
2004 ) to redshifts z @xmath9 3 .
the starburst sources have a mid - infrared spectrum dominated by emission from polycyclic aromatic hydrocarbons ( pah ) ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
it is especially important to have methods for measuring sfr at high redshifts using various observations enabled by @xmath2 because the mid - infrared luminosity is often the only measure available for starbursts in these luminous , distant , and heavily extincted sources .
a technique to measure sfr in many mid - infrared sources using a consistent parameter is to use the luminosity of the pah features , because these arise in the photodissociation region at the inner edge of the dusty molecular clouds surrounding the starburst @xcite . in the present paper ,
we use a measurement of pah luminosity to determine the sfr as measured in the infrared [ sfr(pah ) ] in order to compare with independent measures of sfr that arise from ultraviolet continuum data [ sfr(uv ) ] and radio continuum data [ sfr(radio ) ] .
this comparison is enabled by the large number of starburst galaxies which have now been observed by the @xmath2 irs and also by the galaxy evolution explorer ( galex , martin et al .
2005 ) and at 1.4 ghz by the very large array , primarily in the first survey @xcite .
we also utilize and classify the available optical spectra from the sloan digital sky survey ( sdss , gunn et al .
1998 ) .
our comparison of sfr(pah ) , sfr(radio ) , and sfr(uv ) has three primary objectives : 1 . to test the validity of the infrared pah calibration of sfr ; 2 . to derive an estimate of extinction for the ultraviolet luminosity by an empirical measurement comparing sfr as measured in the infrared and the ultraviolet ; 3 . to determine if infrared - discovered starbursts are similar in dust content to starbursts discovered with other techniques .
we make these comparisons using a total sample of 287 starburst galaxies with @xmath2 irs spectra , described below . as a control sample for comparison to these starbursts chosen
because they have infrared observations , we use the original list of starburst galaxies derived from the markarian galaxy sample and classified as starbursts with optical spectroscopy @xcite .
this sample was selected without knowledge of infrared characteristics and was discovered as galaxies having unusually bright continua in the visible ultraviolet @xcite .
infrared flux densities measured at 25 @xmath0 m by the infrared astronomical satellite ( iras ) are compared to galex fluxes for a sample of 76 markarian galaxies as a test of whether there are different results for sfr(pah)/sfr(uv ) when using the optically - discovered markarian sample compared to the infrared - discovered @xmath2 sample .
a summary of 243 starbursts with irs spectra and determinations of sfr(pah ) is given by @xcite .
for comparison with galex and radio data , we consider only the 49 galaxies in this sample having z @xmath1 0.5 and not included by @xcite .
the redshift limit is chosen both to enable measurements at an accessible ultraviolet rest - frame wavelength using the galex data , and to have a sufficiently bright sample that most of the @xmath2 sources are detected with galax and vla observations .
the brandl et al .
starbursts are omitted , because part of our objective is to test the calibration for sfr determined from these galaxies . also , they are all resolved sources with the irs , so uncertain aperture corrections would have to be applied for comparison to galex and vla data .
the 49 sources which we use from @xcite are listed in table 1 , identified by footnote . to increase substantially the sample of @xmath2 starbursts ,
we have extracted new irs spectra from the archival data of two @xmath2 legacy programs .
the first is program 30742 ( d. schiminovich , p.i . ) .
this program was designed to obtain irs observations for sources within the lockman hole field of the spitzer wide area infrared extragalactic survey ( swire , lonsdale et al .
2003 ) that also have measured fluxes from galex .
this program contains 101 sources with irs spectra .
the second legacy program is program 40539 ( g. helou , p.i . )
designed to obtain a flux limited sample from various @xmath2 survey fields of sources with f@xmath4(24 @xmath0 m ) @xmath10 5 mjy .
this program contains 330 sources .
we extracted irs spectra of all 431 sources in the two programs described above to find those sources which satisfy our criterion for a `` pure starburst '' .
as in previous irs studies of starbursts @xcite , this criterion for selecting sources whose mid - infrared spectrum is dominated by luminosity from a starburst ( without being contaminated by agn luminosity ) is having equivalent width ( ew ) of the 6.2 @xmath0 m pah feature be ew(6.2 @xmath0 m ) @xmath10 0.4 @xmath0 m . the origin of this criterion is explained further in section 3.1 , below .
all of our new spectral extractions were done with the smart analysis package @xcite , beginning with the basic calibrated data products of the @xmath2 flux calibration pipeline , using the most recent version of calibration applied to the sources ( either v.15 or v.17 ) .
because our objective was only to measure the pah luminosity , we extracted only the irs short - low spectra , which extend to @xmath9 14 @xmath0 m . from the spectral extractions , we measured the redshift as determined from the pah features at 6.2 @xmath0 m , 7.7 @xmath0 m , and 11.3 @xmath0 m ; measured the f@xmath4(7.7 @xmath0 m ) ; and measured the rest - frame equivalent width of the 6.2 @xmath0 m feature , ew(6.2 @xmath0 m ) .
results are in table 1 . for the 101 sources in program 30742 ,
75 sources are found to satisfy the criterion of ew(6.2 @xmath0 m ) @xmath10 0.4 @xmath0 m .
these 75 sources are in table 1 , identified by footnote .
all of these sources have sdss spectra available .
all show optical spectra classified as `` hii '' , because the emission lines are narrow and the h@xmath12 emission line is strong compared to the higher ionization [ oiii ] lines .
this confirms the clear and strong correlation between the optical classification as a starburst and the presence of strong pah emission .
for the 330 sources in program 40539 , 152 sources are found to satisfy the starburst criterion
ew(6.2 @xmath0 m ) @xmath10 0.4 @xmath0 m .
these 152 sources are in table 1 , identified by footnote .
sdss spectra are available for 57 of these sources , and all show optical spectra classified as `` hii '' .
finally , we include 11 blue compact dwarfs ( bcds ) from @xcite for which we extracted spectra to measure f@xmath4(7.7 @xmath0 m ) .
most of these sources do not satisfy the ew(6.2 @xmath0 m ) @xmath10 0.4 criterion because pah features are systematically weak in bcds .
we include these to have a sample of low luminosity sources .
adding these bcds yields the total of 287 starbursts given in table 1 .
uncertainties in the measurement f@xmath4(7.7 @xmath0 m ) for these sources arise from uncertainties in the @xmath2 flux calibration , which is typically @xmath5 5% , and uncertainties in fitting the value of the pah peak at 7.7 @xmath0 m . for sources as bright as those we include
, the combination of these uncertainties leads to a typical observational uncertainty in the infrared flux and luminosity measures of @xmath5 10% .
the galex mission is an all - sky survey satellite which obtains images in the ultraviolet at wavelengths 134 - 179 nm ( fuv ) and 177 - 283 nm ( nuv ) @xcite ] .
point source images have full width at half maximum ( fwhm ) of 4.2 for fuv and 5.3 for nuv @xcite . for the @xmath2 irs ,
spatial resolution is wavelength dependent , so the fwhm of an unresolved source at the observed wavelength of the 7.7 @xmath0 m feature depends on source redshift and is 3 @xmath13 fwhm @xmath13 5 .
galex surveys report a one sigma positional uncertainty of 0.5 @xcite and @xmath2 surveys have a one - sigma uncertainty of 0.4 @xcite .
these image sizes mean that the spatial resolutions of galex and @xmath2 observations are well matched .
any mismatch does not affect measured fluxes for point sources , although there may be aperture - dependent effects between reported fluxes within the observing apertures if sources are very extended compared to the spatial resolution . to avoid such uncertainties
, we do not include in table 1 any sources which are well resolved by the irs ( which is one reason why the nearby starburst galaxies in @xcite are not included ) , so the comparison of fluxes for @xmath2 and galex should not be significantly affected by the slightly differing spatial resolutions , even for any sources that are partially resolved .
a more significant source of uncertainty is determining whether @xmath2 and galex have detected the same source .
confusion would arise , for example , if there were a highly obscured starburst found by @xmath2 but a less obscured , nearby starburst seen by galex . to minimize this uncertainty
, we use a positional - coincidence criterion .
of the 287 @xmath2 infrared sources in table 1 , 283 were included in galex observations . of these 283 sources , 272 infrared sources in table 1 have galex identifications to within 3 , and the galex observations for these sources are tabulated .
sources with positional differences @xmath10 3 are not tabulated because the galex source is outside the half - power location of the @xmath2 observing aperture .
the differences in position between @xmath2 and galex are shown in figure 1 for all sources compared to source flux densities .
for the 272 sources with distance @xmath1 3 , the median difference between galex and @xmath2 positions is only 0.64 .
this value is comparable to the quoted positional uncertainties , which gives confidence that the same starbursts are observed in both infrared and ultraviolet .
the observed flux density at rest frame fuv wavelength of 153 nm [ used for determining sfr(uv ) ] is determined by using the galex observed flux densities at fuv ( effective wavelength 153 nm ) and nuv ( effective wavelength 227 nm ) to fit a power law continuum between these wavelengths , and then interpolate to the flux density at observed wavelength ( 1+z)153 nm using the resulting power law index .
this is the value of f@xmath4(fuv ) listed in table 1 .
for comparing infrared flux densities to radio flux densities , similar considerations for spatial resolution and positional coincidence apply .
most of the radio detections are from the first survey @xcite , and the remainder are from targeted observations with the vla @xcite .
these surveys give spatial resolution with fwhm @xmath9 5 , so the same positional criterion ( agreement to within 3 ) is used for radio and infrared sources as for uv and infrared sources . of the sources in table 1 ,
38% have radio detections .
the radio limit which is used for sources that are not detected at 1.4 ghz is the 1 mjy limit of the first survey .
radio observations are corrected to rest frame 1.4 ghz by assuming a power law spectrum of index @xmath14 = -0.7 .
results for the radio luminosities are in table 1 .
many previous studies have shown a well defined correlation between the presence of pah features in the mid - infrared spectrum and the presence of a starburst as classified in other parts of the spectrum ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
this correlation occurs because the pah emission arises in the photodissociation region that lies between the hii region of a starburst and the surrounding molecular cloud within which the stars formed .
the exceptional uniformity among the pah spectrum of starburst galaxies is illustrated in figure 2 .
this figure shows the overplotted , normalized irs spectra of all 25 sources having an optical sdss classification as starburst which are included within a @xmath2 flux limited sample defined only by f@xmath4(24 @xmath0 m ) @xmath10 10 mjy @xcite .
the pah features arise from vibrational modes in the complex pah molecules ( e.g. * ? ? ?
variations in the intensity and hardness of the radiation field within starbursts change the relative strengths of the pah features , but these changes are not great in the integrated spectra of starbursts , as seen in figure 2 .
the 6.2 @xmath0 m and 7.7 @xmath0 m features have very similar relative strengths , for example , because both arise from the cc stretching mode vibration .
by contrast , the relative strength of the 11.3 @xmath0 m feature has greater variation because it arises from the ch bending mode .
@xmath2 irs spectra have been used to measure the sfr based on pah measures and empirical calibration between pah luminosities and bolometric luminosities ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the latter two papers describe and utilize the measure of pah luminosity @xmath3l@xmath4(7.7 @xmath0 m ) , where l@xmath4(7.7 @xmath0 m ) is determined only from the flux density at the peak of the 7.7 @xmath0 m feature .
the motivation for this measure of pah luminosity is illustrated by the spectra in figure 2 .
the pah features are complex and blended .
determining the total luminosity of any single feature requires deblending and adopting a level for the underlying continuum of dust emission .
while such sophisticated deblending can be undertaken with sufficient s / n @xcite , it is not possible to do accurately in sources of limited wavelength baseline or poor s / n . because of the objective to measure pah strength in high redshift sources with poor s / n and limited rest frame wavelength coverage , @xcite and @xcite define the pah luminosity simply as @xmath3l@xmath4(7.7 @xmath0 m ) .
the measure @xmath3l@xmath4(7.7 @xmath0 m ) is not a measure only of the pah luminosity , as would be an integrated measurement of a deblended pah feature .
the measure based on peak @xmath3l@xmath4(7.7 @xmath0 m ) includes some contribution from the dust continuum underlying the pah feature .
as illustrated in figure 2 , this contribution is negligible .
for the representative continuum fit illustrated , this continuum contribution is @xmath1 10% of the peak f@xmath4(7.7 @xmath0 m ) . for our use of @xmath3l@xmath4(7.7 @xmath0
m ) as a measure of sfr , separating the pah contribution from the dust continuum contribution is irrelevant so long as the dust continuum is also produced by the starburst .
this is the case because the calibration of sfr(pah ) in @xcite is done empirically using the total @xmath3l@xmath4(7.7 @xmath0 m ) , which includes both pah and dust contributions .
systematic error in this use of @xmath3l@xmath4(7.7 @xmath0 m ) to measure sfr(pah ) will arise only if the dust continuum has a significant contribution from dust heated by an agn . in this case
, the underlying dust continuum will be enhanced , and @xmath3l@xmath4(7.7 @xmath0 m ) will be artificially increased compared to that for a pure starburst .
this would lead to an overestimate of sfr(pah ) . to avoid this uncertainty
, we deal only with sources which have sufficiently strong pah features that they are pure starbursts without an agn contribution to the mid - infrared continuum .
as discussed in @xcite and @xcite , the criterion for such starburst - dominated sources is that the ew of the 6.2 @xmath0 m pah feature exceed 0.4 @xmath0 m in the source rest frame .
this criterion arises empirically from ew(6.2 @xmath0 m ) of the starbursts in @xcite classified optically and from the starbursts in @xcite which have sdss optical classification as starbursts .
this quantitative criterion is also verified by the synthetic spectra in @xcite which combine the prototype starburst ngc 7714 with the agn markarian 231 and show that the starburst fraction of @xmath3l@xmath4(7.7 @xmath0 m ) exceeds 90% when ew(6.2 @xmath0 m ) @xmath10 0.4 @xmath0 m .
the sources we utilize in the present paper are sufficiently bright that the ew(6.2 @xmath0 m ) can be reliably measured , and these are given in table 1 . while we could , therefore , use the total integrated luminosity in this feature as the measure of sfr for the sources in table 1 , a major objective of our study is to allow comparison of these sources to the many fainter sources discovered by @xmath2 at higher redshifts . for these faint sources , f@xmath4(7.7 @xmath0 m )
is a much more reliable observational measurement than is the total integrated flux in the 6.2 @xmath0 m feature . to restrict the present analysis to sources
for which we can be confident are dominated by a starburst , we also use the criterion of ew(6.2 @xmath0 m ) @xmath15 0.4 @xmath0 m .
the only exception are 11 bcds which are included from @xcite ; these systematically have weaker pah features than more luminous starbursts , although the bcds show no spectral evidence of an agn contribution . the ew criterion used for the remaining starbursts
is confirmed by the fact that all of the sources in table 1 which have sdss spectra show spectra classified as `` hii '' , with no indication of an agn of any type . for sources with a mid - infrared spectrum arising only from a starburst
, the luminosity @xmath3l@xmath4(7.7 @xmath0 m ) relates empirically to the total infrared luminosity of the starburst , @xmath16 , according to log @xmath16 = log [ @xmath3l@xmath4(7.7 @xmath0 m ) ] + 0.78 @xcite .
this transformation has no dependence on luminosity , as is seen by combining the iras - derived @xmath16 for the local , low luminosity starburst sample of @xcite with the high luminosity , high redshift submillimeter galaxies in @xcite .
using the relation from @xcite , this result for @xmath16 yields that log [ sfr(pah ) ] = log [ @xmath3l@xmath4(7.7 @xmath0 m ) ] - [email protected] , for @xmath3l@xmath4(7.7 @xmath0 m ) in ergs s@xmath7 and sfr in m@xmath6 yr@xmath7 .
the measurement of sfr(pah ) using @xmath3l@xmath4(7.7 @xmath0 m ) has been used to measure the local sfr density of the universe @xcite and to measure the evolution of starbursts @xcite .
it was found , for example , that star formation in the most luminous starbursts evolves as log[sfr(pah ] = 2.1(@xmath50.3 ) + 2.5(@xmath50.3 ) log(1+z ) , up to z = 2.5 for sfr(pah in m@xmath6 yr@xmath7 . in the present paper ,
our primary objective is to compare sfr(pah ) with sfr(uv ) . both to verify the previous conclusions based on sfr(pah ) and to undertake comparison with sfr(uv ) , it is useful to test the validity of the relation described above for sfr(pah ) .
@xcite showed that this relation was consistent to within @xmath9 10% for measure of sfr as determined from the luminosities of the infrared [ ne ii ] and [ ne iii ] lines @xcite , which arise in the hii region surrounding the starburst .
also , @xcite showed that the empirical calibration of @xmath3l@xmath4(7.7@xmath0 m ) to sfr leads to a measure of the integrated sfr in the local universe that is consistent with that measured for starburst galaxies using different samples and techniques for measuring sfr from infrared continuum luminosity .
comparison with sfr determined with radio observations can provide an additional test of sfr(pah ) .
the demonstration of a well defined correlation between far - infrared flux measured by iras and 1.4 ghz flux density from various radio observations @xcite led to a calibration of sfr(radio ) as measured by 1.4 ghz luminosity @xcite .
( radio luminosity from a starburst arises primarily from the non - thermal luminosity of supernova remnants , and the number of such remnants relates to the sfr of the starburst . )
this relation is log [ sfr(radio ) ] = log [ @xmath3l@xmath4(1.4 ghz ) ] - 37.07 , for @xmath3l@xmath4(1.4 ghz ) in ergs s@xmath7 and sfr(radio ) in m@xmath6 yr@xmath7 .
fundamentally , the calibration of sfr(radio ) and the calibration of sfr(pah ) based on @xmath3l@xmath4(7.7 @xmath0 m ) both arise from the premise of @xcite who equates the total bolometric luminosity radiated by dust to the bolometric luminosity of the young stars in the starburst .
nevertheless , the calibrations of sfr(radio ) and sfr(pah ) arise through different routes and using different sources so provide an independent consistency check on one another .
star formation rates
sfr(pah ) measured with @xmath3l@xmath4(7.7 @xmath0 m ) and sfr(radio ) measured with @xmath3l@xmath4(1.4 ghz ) are given in table 1 .
of the 287 infrared sources in table 1 , only 108 have actual radio detections . for the remainder ,
sfr(radio ) is determined as an upper limit assuming f@xmath4(1.4 ghz ) @xmath1 1 mjy .
for these sources , sfr(pah)/sfr(radio ) is a lower limit .
the comparison of sfr(pah)/sfr(radio ) for the detected radio sources is shown in figure 3a .
sources range over a factor of 10@xmath17 in luminosity but only a factor of 10@xmath18 without the bcds . the median value of the ratio log [ sfr(pah)/sfr(radio ) ] = 0.0 for these detected sources .
this indicates precise agreement between sfr(pah ) and sfr(radio ) for starbursts , and there is no statistically meaningful dependence on luminosity for the starbursts which are not bcds .
only the bcds depart from the agreement between sfr(pah ) and sfr(radio ) .
this result arises primarily because of the known weakness of pah features in some bcds @xcite , so that sfr(pah ) based on @xmath3l@xmath4(7.7 @xmath0 m ) underestimates the true sfr .
the sources in table 1 without radio detections are plotted as limits in figure 3b , using 1 mjy as the limit for f@xmath4(1.4 ghz ) .
the limits on sfr(pah)/sfr(radio ) shown in figure 3b are in the sense that the actual sfr(pah)/sfr(radio ) would be greater than the limit plotted .
the distribution of limits has a median slightly smaller than the median of 0.0 for detected sources , which is as expected for the limits .
the most direct measurement of sfr in a starburst is to observe the ultraviolet continuum luminosity from the young , hot stars in the starburst ( e.g. * ? ? ?
* ; * ? ? ?
the primary difficulty in obtaining sfr(uv ) in this way is in determining the fraction of the observed luminosity which has been absorbed by intervening dust .
corrections for this dust extinction can be determined observationally by measuring the reddening of the spectrum , using the balmer decrement for example , and then using an extinction law to relate reddening to total extinction @xcite .
such corrections can not account , however , for luminosity sources so extinguished that they make no contribution to the observed spectrum from which reddening corrections are determined .
the same dust that absorbs ultraviolet luminosity from these `` buried '' sources produces infrared luminosity . by comparing the infrared luminosity from the buried starburst with the observed ultraviolet luminosity
, a direct measurement is made of the fraction of intrinsic ultraviolet luminosity which is absorbed .
it is this comparison that we undertake by comparing sfr(uv ) and sfr(pah ) , using only the observed ultraviolet luminosity without applying any corrections for dust attenuation .
sfr(uv ) is determined using the relation sfr(uv)= 1.08x10@xmath19l@xmath4(fuv ) , for l@xmath4(fuv ) in ergs s@xmath7 hz@xmath7@xcite .
this is the intrinsic value , without applying any extinction corrections to the ultraviolet . for convenient comparison to sfr(pah ) and
sfr(radio ) in table 1 , we convert this to log [ sfr(uv ) ] = log @xmath3l@xmath4(fuv ) - 43.26 , for @xmath3l@xmath4(fuv ) in ergs s@xmath7 and sfr in m@xmath6 yr@xmath7 , using f@xmath4(fuv ) determined as described above in section 2.2 for the fuv rest frame wavelength of 153 nm .
the relations for sfr(pah ) and sfr(uv ) also describe the correspondence between intrinsic flux @xmath3f@xmath4(7.7 @xmath0 m ) for the pah feature and fuv magnitude which should arise from the same starburst . if sfr(pah ) = sfr(uv ) , and there is no extinction of the ultraviolet , log f@xmath4(fuv ) = log f@xmath4(7.7 @xmath0 m ) - 1.01 . when compared to an infrared source
for which f@xmath4(7.7 @xmath0 m ) is 1 mjy , the intrinsic fuv magnitude of the same starburst in the ab system would be 18.5 ( using the transformation ab = -2.5log f@xmath4 - 48.60 , for f@xmath4 in ergs @xmath20 s@xmath7 hz@xmath7 . ) the comparison sfr(pah)/sfr(uv ) for the sources in table 1 is shown in figure 4 .
in contrast to the close agreement between sfr(pah ) and sfr(radio ) shown in figure 3 , there is a very large offset between sfr(pah ) and sfr(uv ) in figure 4 . the median log [ sfr(pah)/sfr(uv ] = 1.67 .
if this entire offset is caused by extinction of the ultraviolet luminosity , the median result indicates that only 2% of the ultraviolet luminosity escapes the starburst , or that sfr(uv ) would be underestimated by a factor of 50 if extinction were not considered .
the median value is determined primarily by the large number of sources with log @xmath3l@xmath4(7.7 @xmath0 m ) @xmath10 43 .
there is , however , a significant luminosity dependence in the ratio sfr(pah)/sfr(uv ) .
fitting all of the points in figure 4 yields the result : @xmath21= ( 0.53$\pm$0.05)log [ $ \nu$l$_{\nu}$(7.7 $ \mu$m ) ] - 21.49$\pm$0.18 , for $ \nu$l$_{\nu}$(7.7 $ \mu$m ) in ergs s$^{-1}$.$$\ ] ] the quoted errors arise only from the fitting of the line in order to demonstrate the intrinsic dispersion of the ratio sfr(pah)/sfr(uv ) among different starbursts ; uncertainties for individual points are not included in the quoted errors .
the most luminous starbursts are those which have been found by @xmath2 at z @xmath9 2.5 .
the luminosities scale with redshift @xcite , with log [ @xmath3l@xmath4(7.7 @xmath0 m ) ] = 44.63 + 2.5log ( 1+z ) . at z = 2.5 ,
the most luminous starburst from this relation has log @xmath3l@xmath4(7.7 @xmath0 m ) = 46 .
equation ( 1 ) and figure 4 indicate that such starbursts should have values of sfr(pah)/sfr(uv ) @xmath10 700 !
this very large ultraviolet extinction is consistent with the observation that the population of luminous @xmath2 sources at high redshift are very faint when observed optically @xcite . at z = 2.5 ,
the rest frame fuv wavelengths correspond approximately to observed @xmath11 magnitude .
the brightest obscured infrared sources at z = 2.5 have f@xmath4(7.7 @xmath0 m ) @xmath9 2 mjy in the observed frame . according to the relation given above between 7.7 @xmath0 m flux and rest frame ultraviolet magnitude , such a source with a factor of 700 extinction would have optical @xmath11 mag .
@xmath9 24.9 . as was also
the case for sfr(pah)/sfr(radio ) in figure 3 , the low luminosity bcds in figure 4 show systematically smaller values of sfr(pah)/sfr(uv ) .
this result indicates low ultraviolet extinction in bcds combined with weak pah features .
various explanations related to low metallicity have been suggested to explain the weakness of pah features in bcds @xcite .
an alternative explanation for weak pah features based on the observation of little extinction is that the hii regions are not ionization bounded , implying that the starbursts in bcds are not enveloped by the dusty molecular clouds which produce the photodissociation region where pah features arise in more luminous starbursts
. the same calibration of sfr(pah ) which we use was used by @xcite to determine the sfr density in the local universe from a complete , flux limited sample of @xmath2 sources .
the result was 0.008 m@xmath6 yr@xmath7mpc@xmath22 .
the local sfr determined with galex results was measured as 0.003 m@xmath6 yr@xmath7mpc@xmath22 without ultraviolet extinction corrections @xcite , and increases to 0.016 m@xmath6 yr@xmath7mpc@xmath22 with an assumed extinction correction of a factor of @xmath9 5 .
based on comparisons with other infrared - derived measures of the local sfr density , houck et al .
consider their result based on sfr(pah ) to be underestimated by perhaps a factor of two because of the small number of low luminosity sources in their sample .
this implies that sfr(pah ) and sfr(uv ) are consistent when extinction by a factor of 5 is applied to the ultraviolet data .
why is this extinction correction by a factor of 5 for the local sfr so different from the median factor of 50 in sfr(pah)/sfr(uv ) in figure 4 ?
the answer is in the luminosity dependence for sfr(pah)/sfr(uv ) .
the sfr for the local sample @xcite is dominated by galaxies with 10@xmath23 @xmath1 @xmath16 @xmath1 10@xmath24 l@xmath6 . at these luminosities in figure 4 , 4 @xmath1 sfr(pah)/sfr(uv )
@xmath1 16 , which is consistent with the local correction of @xmath9 5 applied for ultraviolet extinction .
this result indicates that an infrared - discovered sample of dusty starbursts may contain extinction that is somewhat greater than that found in ultraviolet - discovered galaxies , but this excess extinction is less than a factor of two compared to the extinction within ultraviolet - discovered galaxies .
this consistency is evidence that results we find for starburst galaxies discovered in the infrared with @xmath2 can be used to consider the nature of starburst galaxies in general . as a further test of this conclusion
, we now consider a more detailed comparison with dust extinction in ultraviolet - selected galaxies .
the relations used above to determine sfr(pah ) and sfr(uv ) derive from independent assumptions and modeling regarding the stellar composition of starbursts
. it would not be surprising if these independent approaches led to different values for sfr in a starburst .
we have assumed in our interpretation , however , that both approaches should give the same result for sfr and that the only differences arise from dust attenuation of the ultraviolet luminosity . in order to test this assumption ,
we now consider independent measures of dust attenuation to determine if our conclusions regarding the amount of extinction by dust are consistent with other estimates .
the result in equation ( 1 ) and the distribution of points in figure 4 indicates that the measured values for sfr(pah)/sfr(uv ) without corrections to sfr(uv ) for dust extinction scale with starburst luminosity , such that sfr(pah)/sfr(uv ) is dependent on ( starburst luminosity)@xmath25 .
if the ratio sfr(pah)/sfr(uv ) differs only because of differing dust extinction for the observed sfr(uv ) , this indicates that more luminous starbursts are also more dusty , which is one of our most important results .
for example , the points in figure 4 indicate that dust extinction is negligible in starbursts of log[@xmath3l@xmath4(7.7 @xmath0 m ) ] @xmath9 41 ( typical luminosity of bcds ) but that dust extinction can reach a factor of 1000 in the most luminous starbursts .
our conclusion depends on the interpretation that the difference between observed values of sfr(pah ) and sfr(uv ) arises only because of dust extinction of some fraction of intrinsic ultraviolet luminosity from the starburst .
this interpretation can be checked independently of our result .
estimates of the dust extinction corrections that should be applied for ultraviolet luminosity have previously been made using extensive collections of galaxy sed templates to compare @xmath16 to the total observed ultraviolet luminosity , @xmath26 , from galaxies @xcite .
the ratio @xmath16/@xmath26 derived in this way is a measure of dust extinction which is equivalent to our measure sfr(pah)/sfr(uv ) .
these previous results also indicate a luminosity dependence for the dust extinction @xcite .
the fundamental principle in using @xmath16/@xmath26 to measure dust attentuation of ultraviolet luminosity is that all infrared luminosity from a starburst arises from re - radiation by dust of ultraviolet luminosity which has been absorbed .
the total luminosity radiated by the starburst is then measured as @xmath16 + @xmath26 .
this sum is the same as the intrinsic ultraviolet luminosity which is produced , which we define as @xmath26(starburst ) .
the fraction of @xmath26(starburst ) which emerges without dust attenuation is @xmath26/@xmath26(starburst ) , or @xmath26/(@xmath16 + @xmath26 ) .
equivalently for our measurements , @xmath26/@xmath26(starburst ) = sfr(uv)/[sfr(uv ) + sfr(pah ) ] . in order to use the results in @xcite with our relation in equation ( 1 )
, we convert log @xmath3l@xmath4(7.7 @xmath0 m ) to @xmath16 using the empirical relation for starbursts in @xcite , log l@xmath8 = log [ @xmath3l@xmath4(7.7 @xmath0 m ) ] - [email protected] for l@xmath8 in l@xmath6 and @xmath3l@xmath4(7.7 @xmath0 m ) in ergs s@xmath7 . with this transformation , equation ( 1 )
becomes @xmath21= 0.53 log l$_{ir}$ - 4.11 , for l$_{ir}$ in { l$_{\odot}$}.$$\ ] ] we use equation ( 2 ) to determine the fraction of emerging ultraviolet luminosity ,
@xmath26/@xmath26(starburst ) = sfr(uv)/[sfr(uv ) + sfr(pah ) ] , and show this in figure 5 as a function of luminosity .
our result is compared to the same parameter for ultraviolet - selected galaxies using @xmath26/(@xmath16 + @xmath26 ) in @xcite .
the slope of the luminosity dependence is very similar between the two results .
our relation from equation ( 2 ) is offset , however , by about a factor of two lower values of @xmath26/@xmath26(starburst ) .
this offset implies a systematically higher extinction by a factor of two at all luminosities for our infrared - discovered sample compared to the extinction estimates from ultraviolet - selected sources
. there could be various reasons for the offset .
part could be produced if our equation for sfr(pah ) is incorrect compared to sfr(uv ) and gives larger values for sfr .
we concluded in section 3.2 , however , that there is no offset between sfr(pah ) and sfr(radio ) .
another explanation is that the templates in @xcite are not purely starbursts , so ultraviolet luminosity from portions of a galaxy with less extinction may be included than is the case for the pure starbursts in the infrared sample .
finally , part of the excess extinction we find could also result from our selection based on infrared sources , which preferentially selects the most dusty starbursts .
we make another test for this final selection effect in section 3.5 , below .
all of the starbursts considered so far were selected because they have @xmath2 infrared measurements , and most were discovered in @xmath2 surveys at 24 @xmath0 m .
a concern is that this selection might lead to a selection effect which favors sfr(pah ) compared to sfr(uv ) . to test for such an effect , we utilize a sample of starbursts selected without regard to their infrared fluxes but selected instead because of unusually strong fluxes in the visible ultraviolet .
this sample derives from the markarian survey for `` galaxies with strong ultraviolet continuum '' @xcite .
it is the sample of 76 markarian starbursts which have galex identifications among the markarian galaxies defined as starbursts by @xcite using optical spectra which show only an hii spectrum with no indication of an agn . the optical colors and emission line properties of these markarian sources
are explained by the presence of a hot , young star population , and this list was the first compilation of a starburst galaxy sample following the original definition of this galaxy category @xcite .
this sample is a useful control sample because most of the markarian starbursts also have mid - infrared measurements by iras and ultraviolet measurements with galex , although neither mission existed when the original sample was assembled . in table 2 ,
the galex and iras data for these markarian galaxies are given .
for sources detected by galex but not by iras , the 3@xmath27 upper limit for f@xmath4(25 @xmath0 m ) is taken as 65 mjy . to adopt an infrared spectrum
, we use starburst spectra such as those in figure 2 which characterize the unbiased samples of @xmath2 sources described in @xcite . for sources in that sample having 42 @xmath1 log [ @xmath3l@xmath4 ( 15 @xmath0 m ) ] @xmath1 44 , covering the luminosity range of most markarian galaxies in table 2 ,
the average spectrum has f@xmath4(25
@xmath0m)/f@xmath4(7.7 @xmath0 m ) = [email protected] , which allows a transformation from f@xmath4(25 @xmath0 m ) observed by iras to a value of f@xmath4(7.7 @xmath0 m ) .
this allows the calculation of sfr(pah ) for the markarian sample using the same relations used above for sources in table 1 having observed values of f@xmath4(7.7 @xmath0 m ) .
results for sfrs are in table 2 and the ratios sfr(pah)/sfr(uv ) are shown in figure 6 .
fitting all of the points and limits in figure 6 yields the result : @xmath28= ( 0.38$\pm$0.08)log $ \nu$l$_{\nu}$(7.7 $ \mu$m ) - 15.27$\pm$0.25.$$\ ] ] as for the infrared - selected starbursts in equation ( 1 ) and figure 4 , the markarian starbursts in figure 6 also show a dependence of sfr(pah)/sfr(uv ) on luminosity , although the range of luminosities is less than for the infrared - selected starbursts in figure 4 . within the uncertainties of equations ( 1 ) and ( 3 ) , the ratio sfr(pah)/sfr(uv )
has the same dependence on luminosity . at the median luminosity for markarian galaxies in figure 6 of log @xmath3l@xmath4(7.7 @xmath0
m ) = 43 , the log [ sfr(pah)/sfr(uv ) ] = 1.07 . at this same luminosity ,
the infrared - selected starbursts in figure 4 have from equation ( 1 ) , log [ sfr(pah)/sfr(uv ) ] = 1.30 .
this result indicates moderately less extinction in the markarian galaxies ; 8.5% of the ultraviolet luminosity escapes compared to only 5% which escapes the infrared - selected galaxies .
these differences are small compared to the large extinctions which are observed in both samples .
this means that extreme extinction in the ultraviolet is characteristic of all starbursts regardless of selection technique , so that choosing starbursts from infrared observations does not lead to a sample which is more heavily extincted by dust than choosing starbursts based on optical spectroscopic classifications .
this comparison with markarian galaxies is additional confirmation that an infrared - selected sample of starbursts does not yield starbursts which are significantly dustier than other starbursts .
star formation rates determined from infrared ( @xmath2 ) , ultraviolet ( galex ) , and radio ( vla ) indicators have been determined for 287 starburst galaxies with spectra from the @xmath2 infrared spectrograph .
we find that sfr(pah ) and sfr(radio ) give the same values , but sfr(uv ) is typically a factor of 50 less than sfr(pah ) .
this is an empirical measure of extinction in the ultraviolet .
the measure of starburst luminosity used for sfr(pah ) is @xmath3l@xmath4(7.7 @xmath0 m ) for the peak luminosity of the 7.7 @xmath0 m pah feature .
the measure of starburst luminosity for sfr(uv ) arises from the observed continuum flux density at rest frame 153 nm .
the ratio log sfr(pah)/sfr(uv ) has a median value of 1.67 , with dependence on luminosity of form log [ sfr(pah)/sfr(uv ) ] = ( [email protected])log [ @xmath3l@xmath4(7.7 @xmath0 m ) ] - [email protected] , for @xmath3l@xmath4(7.7 @xmath0 m ) in ergs s@xmath7 .
using the conversion we adopt between @xmath3l@xmath4(7.7 @xmath0 m ) and @xmath16 for starbursts , this becomes log [ sfr(pah)/sfr(uv)]= 0.53 log l@xmath8 - 4.11 , for l@xmath8 in l@xmath6 .
these results indicate that sfrs determined with ultraviolet luminosities require extinction corrections by a factor of @xmath9 10 for typical local starbursts , but this correction factor exceeds 700 for the most luminous starbursts at z @xmath9 2.5 .
the trend of increasing extinction with increasing luminosity is similar to that reported previously for ultraviolet - selected galaxies , although our infrared - selected sample shows a systematically larger dust extinction by about a factor of two .
the extreme ultraviolet extinction explains why the most luminous starbursts discovered by @xmath2 at z @xmath9 2.5 are optically faint .
these sources have log @xmath3l@xmath4(7.7 @xmath0 m ) = 46 . with the derived amount of extinction ,
the optical magnitude of a starburst at z @xmath9 2.5 with a 1 mjy 7.7 @xmath0 m feature should be @xmath11 @xmath9 25.6 .
sfr(pah ) and sfr(uv ) are compared for a control sample of 76 markarian starburst galaxies having ultraviolet fluxes in order to test for selection effects arising from our use of infrared - selected starbursts .
the result indicates only moderately less extinction in the markarian galaxies ; 8.5% of the ultraviolet luminosity escapes galaxies compared to only 5% which escapes infrared - selected galaxies of the same luminosity log @xmath3l@xmath4(7.7 @xmath0
m ) @xmath9 43 .
this comparison with markarian galaxies is additional confirmation that an infrared - selected sample of starbursts does not yield starbursts which are significantly dustier than other starbursts .
overall results for sfr(pah ) indicate that infrared - discovered starbursts are not significantly dustier that starbursts of similar luminosity discovered in other ways .
this means that conclusions derived from @xmath2 samples of starbursts are generally applicable to understanding the evolution of star formation in the universe .
we thank d. barry for help in data acquisition , and p. hall , v. leboutiellier , and j. bernard - salas for aid in improving our irs spectral analysis with smart .
we thank the anonymous referee for suggestions which improved our paper .
this work is based primarily 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 at cornell university was provided by nasa through contract number 1257184 issued by jpl / caltech .
support was also provided by the us civilian research and development foundation under grant arp1 - 2849-ye-06 .
this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by jpl / caltech under contract with nasa .
ageros , m.a .
et al . , 2005 , , 130 , 1022 balzano , v.a 1983 , , 268 , 602 brandl , b. et al . 2006 , , 653 , 1129 .
calzetti , d. et al .
2007 , , 666 , 870 buat , v. et al .
2007 , , 173 , 404 condon , j. j. 1992 , ann.rev.astron.ap , 30 , 575 condon , j.j .
2003 , , 125 , 2411 dey , a. et al .
2008 , , 677 , 943 fadda , d. et al .
2006 , , 131 , 2859 .
farrah , d. , et al .
2008 , , 677 , 957 genzel , r. et al .
1998 , , 498 , 579 gunn , j. e. , et al .
1998 , , 116 , 3040 haarsma , d. b. , partridge , r. b. , windhorst , r. a.,and richards , e. a. 2000 , , 544 , 641 helou , g. , soifer , b. t. , and rowan - robinson , m. 1985 , , 298 , l7 higdon , s.j.u .
, et al . 2004 , , 116 , 975 ho , l , and keto , e. 2007 , , 658 , 314 houck , j. r. , et al .
2004 , , 154 , 18 houck , j. r. , weedman , d.w .
, lefloch , e. , and hao , l. 2007 , , 671 , 323 iglesias - paramo , j. et al .
2006 , , 164 , 381 izotov , y. i. , and thuan , t. x. 1998 , , 500 , 188 kennicutt , r.c .
1998 , ann.rev.astron.ap , 36 , 189 le floch , e. et al .
2005 , , 632 , 169 lonsdale c. j. , et al , 2003 , pasp , 115 , 897 madau , p. , pozzetti , l. , and dickinson , m. 1998 , , 498 , 106 mannucci , f. , buttery , h. , maiolino , r. , marconi , a. , and pozzetti , l. 2007 , , 461 , 423 markarian , b.e .
1967 , astrofizika , 3 , 55 .
martin , d.c .
2005 , , 619 , l1 .
martin , d.c .
2008 , , 173 , 415 morrissey , p. et al .
2005 , , 619 , l7 peeters , e. , spoon , h.w.w . , and tielens , a.g.g.m .
2004 , , 613 , 986 peeters , e. , hony , s. , van kerckhoven , c. , tielens , a.g.g.m . ,
allamandola , l.j . ,
hudgins , d.m . , and bauschlicher , c.w .
, 2002 , , 390 , 1089 .
pope , a. et al .
2008 , , 675 , 1171 rigopoulou , d. , spoon , h. w. w. , genzel , r. , lutz , d. , moorwood , a. f. m. , and tran , q. d. 2000 , , 118 , 2625 salim , s. et al . , 2007 , , 173 , 267 smith , j.d.t .
et al . 2007 , , 656 , 770 treyer , m. et al . 2007 , , 173 , 256 weedman , d. w. , feldman , f. r. , balzano , v. a. , ramsey , l. w. , sramek , r. a. and wu , c .- c . , 1981 , , 208 , 105 weedman , d.w .
et al . , 2006 , , 653 , 101 .
weedman , d.w . , and houck , j.r .
2008 , , 686 , 127 weedman , d.w . , and houck , j.r .
2009 , , 693 , 370 white , r. l. , becker , r. h. , helfand , d. j. , and gregg , m. d. 1997 , , 475 , 479 wu , y. , charmandaris , v. , hao , l. , brandl , b. r. , bernard - salas , j. , spoon , h. w. w. , and houck , j. r. 2006 , , 639 , 157 yan , l. et al .
2007 , , 658 , 778 1 & haro 11 & 003652.5 - 333319 & 1.1 & 0.021 & 0.13 & 20.4 & 43.84 & 2969 & 43.72 & 27.2 & 38.55 & 1.27 & 0.46 & 1.48 + 2 & sst24 & 021503.53 - 042422.0 & 1.5 & 0.141 & 0.60 & 3.2 & 43.76 & 1.8 & 42.2 & @xmath11 & @xmath138.81 & 1.19 & -1.06 & @xmath11.74 + 3 & sst24 & 021557.13 - 033728.3 & 0.9 & 0.034 & 0.49 & 7.3 & 42.85 & 53.2 & 42.42 & @xmath11 & @xmath137.55 & 0.28 & -0.84 & @xmath10.48 + 4 & sst24 & 021649.72 - 042554.3 & 0.3 & 0.141 & 0.57 & 6.1 & 44.03 & 52.5 & 43.67 & @xmath11 & @xmath138.81 & 1.46 & 0.41 & @xmath11.74 + 5 & sst24 & 021743.85 - 051751.7 & 1.6 & 0.032 & 0.65 & 21 & 43.26 & 201.5 & 42.94 & 1.45 & 37.66 & 0.69 & -0.32 & 0.59 + 6 & sst24 & 021754.89 - 035825.8 & 0.4 & 0.226 & 0.55 & 10.6 & 44.70 & 38.3 & 43.96 & 1.41 & 39.38 & 2.13 & 0.70 & 2.31 + 7 & sst24 & 021849.78 - 052158.4 & 0.4 & 0.292 & 0.71 & 6.2 & 44.70 & 3.5 & 43.16 & 0.44 & 39.11 & 2.13 & -0.10 & 2.04 + 8 & sst24 & 021912.72 - 050541.4 & 0.3 & 0.195 & 0.88 & 2.6 & 43.95 & 38.5 & 43.83 & 0.4 & 38.71 & 1.38 & 0.57 & 1.64 + 9 & sst24 & 021928.37 - 042240.6 & 1.0 & 0.047 & 0.63 & 7.4 & 43.14 & 21.6 & 42.31 & @xmath11 & @xmath137.84 & 0.57 & -0.95 & @xmath10.77 + 10 & sst24 & 021953.09 - 051824.3 & 0.5 & 0.072 & 0.62 & 24.1 & 44.03 & 25.9 & 42.77 & @xmath11 & @xmath138.21 & 1.46 & -0.49 & @xmath11.14 + 11 & sst24 & 021957.01 - 052439.9 & 0.3 & 0.082 & 0.69 & 8.8 & 43.71 & 58.6 & 43.24 & @xmath11 & @xmath138.33 & 1.14 & -0.02 & @xmath11.26 + 12 & sst24 & 022147.87 - 025730.3 & 0.5 & 0.069 & 0.73 & 18.2 & 43.87 & 31.1 & 42.81 & @xmath11 & @xmath138.18 & 1.30 & -0.45 & @xmath11.11 + 13 & sst24 & 022151.55 - 032912.1 & 0.7 & 0.165 & 0.54 & 5.5 & 44.13 & 30.7 & 43.58 & @xmath11 & @xmath138.95 & 1.56 & 0.32 & @xmath11.88 + 14 & sst24 & 022205.05 - 050536.7 & 1.5 & 0.259 & 0.52 & 7.7 & 44.68 & 1.9 & 42.78 & @xmath11 & @xmath139.36 & 2.11 & -0.48 & @xmath12.29 + 15 & sst24 & 022223.29 - 044319.8 & 0.6 & 0.072 & 0.46 & 4.5 & 43.30 & 7.4 & 42.22 & @xmath11 & @xmath138.21 & 0.73 & -1.04 & @xmath11.14 + 16 & sst24 & 022224.07 - 050550.2 & 0.3 & 0.150 & 0.62 & 5.8 & 44.07 & 6.6 & 42.82 & @xmath11 & @xmath138.87 & 1.50 & -0.44 & @xmath11.8 + 17 & sst24 & 022329.16 - 043209.4 & @xmath103 & 0.141 & 0.49 & 3.1 & 43.73 & & & @xmath11 & @xmath138.81 & 1.16 & & @xmath11.74 + 18 & sst24 & 022334.63 - 035230.2 & 2.8 & 0.175 & 0.56 & 3.4 & 43.98 & 6.8 & 42.97 & @xmath11 & @xmath139.01 & 1.41 & -0.29 & @xmath11.94 + 19 & sst24 & 022345.09 - 054234.4 & 1.3 & 0.142 & 0.55 & 10.7 & 44.29 & 8.2 & 42.87 & @xmath11 & @xmath138.82 & 1.72 & -0.39 & @xmath11.75 + 20 & sst24 & 022413.66 - 042227.3 & 0.4 & 0.119 & 0.61 & 7.6 & 43.98 & 13.7 & 42.93 & 0.41 & 38.27 & 1.41 & -0.33 & 1.2 + 21 & sst24 & 022422.50 - 040230.7 & 0.6 & 0.174 & 0.41 & 3.5 & 43.98 & 13.9 & 43.28 & 0.25 & 38.39 & 1.41 & 0.02 & 1.32 + 22 & sst24 & 022434.33 - 041530.9 & 0.7 & 0.261 & 0.60 & 7.9 & 44.70 & 11.8 & 43.58 & 0.78 & 39.26 & 2.13 & 0.32 & 2.19 + 23 & sst24 & 022447.00 - 040851.4 & 0.9 & 0.096 & 0.36 & 12.5 & 44.00 & 3.4 & 42.13 & 1.45 & 38.63 & 1.43 & -1.13 & 1.56 + 24 & sst24 & 022507.47 - 041835.7 & 0.7 & 0.105 & 0.49 & 4.4 & 43.62 & 42.0 & 43.31 & 0.35 & 38.09 & 1.05 & 0.05 & 1.02 + 25 & sst24 & 022522.60 - 045451.9 & 0.3 & 0.146 & 0.81 & 7.2 & 44.14 & 10.3 & 42.99 & 0.59 & 38.62 & 1.57 & -0.27 & 1.55 + 26 & sst24 & 022536.45 - 050011.1 & 1.9 & 0.052 & 0.61 & 49.3 & 44.06 & 226.8 & 43.42 & 12.95 & 39.04 & 1.49 & 0.16 & 1.97 + 27 & sst24 & 022548.25 - 050051.8 & 0.7 & 0.145 & 0.42 & 4.2 & 43.90 & 0.5 & 41.7 & @xmath11 & @xmath138.84 & 1.33 & -1.56 & @xmath11.77 + 28 & sst24 & 022549.79 - 040025.1 & 1.5 & 0.043 & 0.36 & 23.8 & 43.57 & 15.1 & 42.08 & 2.19 & 38.1 & 1.00 & -1.18 & 1.03 + 29 & sst24 & 022600.02 - 050145.1 & 0.6 & 0.203 & 0.55 & 6.4 & 44.38 & 3.0 & 42.76 & @xmath11 & @xmath139.14 & 1.81 & -0.50 & @xmath12.07 + 30 & sst24 & 022602.96 - 045307.1 & 0.3 & 0.052 & 0.57 & 7.1 & 43.21 & 29.7 & 42.54 & 0.39 & 37.51 & 0.64 & -0.72 & 0.44 + 31 & sst24 & 022637.83 - 035841.8 & 0.5 & 0.073 & 0.38 & 5.1 & 43.37 & 3.2 & 41.87 & @xmath11 & @xmath138.23 & 0.80 & -1.39 & @xmath11.16 + 32 & sst24 & 022655.90 - 040301.7 & 0.1 & 0.135 & 0.94 & 3.3 & 43.72 & 17.1 & 43.15 & 0.14 & 37.93 & 1.15 & -0.11 & 0.86 + 33 & sst24 & 022720.70 - 044537.4 & 0.3 & 0.055 & 0.56 & 34.9 & 43.95 & 94.9 & 43.09 & 2.96 & 38.45 & 1.38 & -0.17 & 1.38 + 34 & sst24 & 022738.55 - 044702.9 & 0.3 & 0.171 & 0.51 & 5.8 & 44.19 & 3.8 & 42.7 & 0.4 & 38.59 & 1.62 & -0.56 & 1.52 + 35 & sst24 & 022741.66 - 045650.0 & 1.7 & 0.056 & 0.59 & 16.7 & 43.65 & 35.9 & 42.68 & 1.09 & 38.03 & 1.08 & -0.58 & 0.96 + 36 & ngc 1140 & 025433.6 - 100140 & 1.2 & 0.005 & 0.53 & 4 & 41.98 & 6496 & 42.83 & 23.7 & 37.26 & -0.59 & -0.43 & 0.19 + 37 & sbs0335 - 052 & 033744.0 - 050240 & 0.4 & 0.014 & @xmath10.02 & @xmath115 & @xmath142.22 & 539.5 & 42.62 & 0.5 & 36.42 & @xmath1 - 0.35 & -0.64 & -0.65 + 38 & ngc 1569 & 043047.0 + 645059 & @xmath103 & 0 & 0.22 & 0.3 & 40.02 & & & 336.3 & 36.36 & -2.55 & & -0.71 + 39 & iizw 40 &
055542.6 + 032332 & & 0.003 & 0.03 & 3 & 41.72 & & & 32.6 & 36.93 & -0.85 & & -0.14 + 40 & ugc 4274 & 081313.0 + 455939 & 1.8 & 0.002 & 0.73 & 3 & 40.76 & 996.4 & 41.04 & 10.5 & 35.93 & -1.81 & -2.22 & -1.14 + 41 & sst24 & 103254.00 + 580633.2 & 0.8 & 0.070 & 0.58 & 5 & 43.32 & 46.6 & 43.00 & @xmath11 & @xmath138.19 & 0.75 & -0.26 & @xmath11.12 + 42 & sst24 & 103450.53 + 584418.2 & 1.0 & 0.090 & 0.52 & 12.9 & 43.96 & 39.9 & 43.16 & @xmath11 & @xmath138.42 & 1.39 & -0.10 & @xmath11.35 + 43 & sst24 & 103525.37 + 581709.7 & 0.5 & 0.113 & 0.43 & 3.3 & 43.57 & 19.6 & 43.04 & @xmath11 & @xmath138.61 & 1.00 & -0.22 & @xmath11.54 + 44 & sst24 & 103531.42 + 581234.1 & 0.3 & 0.178 & 0.43 & 7.7 & 44.34 & 11.2 & 43.21 & @xmath11 & @xmath139.02 & 1.77 & -0.05 & @xmath11.95 + 45 & sst24 & 103538.69 + 573543.2 & 1.5 & 0.100 & 0.45 & 6.3 & 43.74 & 4.9 & 42.34 & @xmath11 & @xmath138.5 & 1.17 & -0.92 & @xmath11.44 + 46 & sst24 & 103539.20 + 574244.4 & 0.2 & 0.116 & 0.53 & 3.8 & 43.65 & 21.7 & 43.11 & @xmath11 & @xmath138.64 & 1.08 & -0.15 & @xmath11.57 + 47 & sst24 & 103542.81 + 583312.9 & 0.7 & 0.085 & 0.65 & 6.5 & 43.61 & 22.7 & 42.86 & @xmath11 & @xmath138.36 & 1.04 & -0.40 & @xmath11.29 + 48 & sst24 & 103554.19 + 581836.2 & 0.2 & 0.155 & 0.39 & 5.5 & 44.07 & 28.7 & 43.49 & @xmath11 & @xmath138.9 & 1.50 & 0.23 & @xmath11.83 + 49 & sst24 & 103557.87 + 572500.4 & 0.5 & 0.104 & 0.50 & 3.6 & 43.53 & 9.0 & 42.63 & @xmath11 & @xmath138.54 & 0.96 & -0.63 & @xmath11.47 + 50 & sst24 & 103606.48 + 574702.7 & 0.6 & 0.045 & 0.49 & 6.8 & 43.07 & 64.0 & 42.74 & 0.69 & 37.64 & 0.50 & -0.52 & 0.57 + 51 & sst24 & 103615.53 + 573915.4 & 0.4 & 0.113 & 0.55 & 5.6 & 43.80 & 6.7 & 42.58 & 0.36 & 38.18 & 1.23 & -0.68 & 1.11 + 52 & sst24 & 103646.49 + 584330.5 & 1.7 & 0.139 & 0.60 & 5.9 & 44.01 & 6.0 & 42.72 & @xmath11 & @xmath138.8 & 1.44 & -0.54 & @xmath11.73 + 53 & sst24 & 103653.51 + 575443.1 & 0.7 & 0.101 & 0.60 & 5.9 & 43.72 & 25.3 & 43.06 & @xmath11 & @xmath138.51 & 1.15 & -0.20 & @xmath11.44 + 54 & sst24 & 103723.23 + 573115.0 & 0.3 & 0.071 & 0.65 & 33.3 & 44.16 & 42.8 & 42.97 & 1.96 & 38.49 & 1.59 & -0.29 & 1.42 + 55 & sst24 & 103732.05 + 572306.0 & 0.1 & 0.073 & 0.65 & 4.4 & 43.31 & 27.7 & 42.81 & @xmath11 & @xmath138.23 & 0.74 & -0.45 & @xmath11.16 + 56 & sst24 & 103804.72 + 580255.9 & 0.8 & 0.089 & 0.70 & 5.3 & 43.57 & 57.6 & 43.30 & @xmath11 & @xmath138.4 & 1.00 & 0.04 & @xmath11.33 + 57 & sst24 & 103818.22 + 583556.5 & 0.8 & 0.129 & 0.51 & 2.1 & 43.48 & 5.6 & 42.62 & @xmath11 & @xmath138.73 & 0.91 & -0.64 & @xmath11.66 + 58 & sst24 & 103856.59 + 571549.2 & 0.7 & 0.046 & 0.58 & 9.4 & 43.23 & 21.2 & 42.28 & @xmath11 & @xmath137.82 & 0.66 & -0.98 & @xmath10.75 + 59 & sst24 & 103858.63 + 582233.6 & 0.7 & 0.069 & 0.48 & 10.8 & 43.65 & 15.1 & 42.49 & @xmath11 & @xmath138.18 & 1.08 & -0.77 & @xmath11.11 + 60 & sst24 & 103911.23 + 571617.1 & 1.2 & 0.105 & 0.77 & 4.2 & 43.60 & 12.9 & 42.80 & @xmath11 & @xmath138.55 & 1.03 & -0.46 & @xmath11.48 + 61 & sst24 & 103926.00 + 575805.0 & 0.6 & 0.138 & 0.42 & 5.7 & 43.99 & 19.7 & 43.23 & @xmath11 & @xmath138.79 & 1.42 & -0.03 & @xmath11.72 + 62 & sst24 & 103928.22 + 584730.1 & 0.5 & 0.044 & 0.62 & 5.4 & 42.95 & 33.1 & 42.44 & @xmath11 & @xmath137.78 & 0.38 & -0.82 & @xmath10.71 + 63 & sst24 & 103935.70 + 575908.1 & 0.8 & 0.076 & 0.60 & 6 & 43.48 & 18.1 & 42.66 & @xmath11 & @xmath138.26 & 0.91 & -0.60 & @xmath11.19 + 64 & sst24 & 103956.43 + 572418.7 & 0.4 & 0.071 & 0.55 & 6.2 & 43.43 & 33.9 & 42.87 & @xmath11 & @xmath138.2 & 0.86 & -0.39 & @xmath11.13 + 65 & sst24 & 103957.23 + 583934.0 & 0.4 & 0.196 & 0.66 & 8 & 44.44 & 21.9 & 43.59 & @xmath11 & @xmath139.11 & 1.87 & 0.33 & @xmath12.04 + 66 & sst24 & 104016.34 + 570845.9 & 1.0 & 0.115 & 0.64 & 7.3 & 43.93 & 2.3 & 42.13 & @xmath11 & @xmath138.63 & 1.36 & -1.13 & @xmath11.56 + 67 & sst24 & 104024.26 + 580918.5 & 2.1 & 0.070 & 0.45 & 6.5 & 43.43 & 44.5 & 42.98 & @xmath11 & @xmath138.19 & 0.86 & -0.28 & @xmath11.12 + 68 & sst24 & 104026.26 + 572623.5 & 0.6 & 0.047 & 0.49 & 6.9 & 43.11 & 19.9 & 42.27 & @xmath11 & @xmath137.84 & 0.54 & -0.99 & @xmath10.77 + 69 & sst24 & 104050.35 + 572341.3 & 0.2 & 0.118 & 0.77 & 8.9 & 44.04 & 18.1 & 43.05 & @xmath11 & @xmath138.65 & 1.47 & -0.21 & @xmath11.58 + 70 & sst24 & 104110.48 + 581523.3 & 0.5 & 0.121 & 0.65 & 3.5 & 43.65 & 13.9 & 42.96 & @xmath11 & @xmath138.67 & 1.08 & -0.30 & @xmath11.61 + 71 & sst24 & 104113.73 + 570456.7 & 1.0 & 0.045 & 0.60 & 6.8 & 43.07 & 13.0 & 42.05 & @xmath11 & @xmath137.8 & 0.50 & -1.21 & @xmath10.73 + 72 & sst24 & 104116.03 + 565345.4 & 0.7 & 0.184 & 0.46 & 10.6 & 44.51 & 15.4 & 43.38 & @xmath11 & @xmath139.05 & 1.94 & 0.12 & @xmath11.98 + 73 & sst24 & 104122.60 + 585331.1 & 0.6 & 0.066 & 0.61 & 10.7 & 43.60 & 133.3 & 43.40 & @xmath11 & @xmath138.14 & 1.03 & 0.14 & @xmath11.07 + 74 & sst24 & 104132.52 + 565952.3 & 1.3 & 0.351 & 0.39 & 4 & 44.67 & 73.6 & 44.64 & @xmath11 & @xmath139.63 & 2.10 & 1.38 & @xmath12.57 + 75 & sst24 & 104139.09 + 583514.6 & 0.8 & 0.114 & 0.56 & 17.7 & 44.31 & 10.3 & 42.77 & 1.69 & 38.85 & 1.74 & -0.49 & 1.78 + 76 & sst24 & 104202.45 + 575349.5 & 1.2 & 0.109 & 0.45 & 2.9 & 43.48 & 14.6 & 42.88 & @xmath11 & @xmath138.58 & 0.91 & -0.38 & @xmath11.51 + 77 & sst24 & 104210.07 + 583939.6 & 0.3 & 0.031 & 0.46 & 5.5 & 42.65 & 37.6 & 42.19 & @xmath11 & @xmath137.47 & 0.08 & -1.07 & @xmath10.4 + 78 & sst24 & 104218.68 + 584059.2 & 0.4 & 0.120 & 0.40 & 3.5 & 43.65 & 23.7 & 43.18 & @xmath11 & @xmath138.67 & 1.08 & -0.08 & @xmath11.6 + 79 & sst24 & 104305.82 + 584152.2 & 0.3 & 0.120 & 0.58 & 8.5 & 44.03 & 27.2 & 43.24 & @xmath11 & @xmath138.67 & 1.46 & -0.02 & @xmath11.6 + 80 & sst24 & 104339.05 + 580250.5 & 0.7 & 0.133 & 0.59 & 9.7 & 44.18 & 28.4 & 43.35 & @xmath11 & @xmath138.76 & 1.61 & 0.09 & @xmath11.69 + 81 & sst24 & 104400.79 + 584537.5 & 1.4 & 0.072 & 0.56 & 19.8 & 43.95 & 6.0 & 42.13 & @xmath11 & @xmath138.21 & 1.38 & -1.13 & @xmath11.14 + 82 & sst24 & 104427.42 + 594429.6 & 0.9 & 0.117 & 0.53 & 5.8 & 43.85 & 7.0 & 42.63 & @xmath11 & @xmath138.64 & 1.28 & -0.63 & @xmath11.58 + 83 & sst24 & 104432.97 + 564042.1 & 1.1 & 0.066 & 0.39 & 12.9 & 43.68 & 25.3 & 42.68 & @xmath11 & @xmath138.14 & 1.11 & -0.58 & @xmath11.07 + 84 & sst24 & 104438.34 + 562211.2 & 0.8 & 0.026 & 0.50 & 50.8 & 43.46 & 105.5 & 42.48 & 4.59 & 37.98 & 0.89 & -0.78 & 0.91 + 85 & sst24 & 104445.22 + 582718.0 & @xmath103 & 0.034 & 0.45 & 10.7 & 43.02 & & & @xmath11 & @xmath137.55 & 0.45 & & @xmath10.48 + 86 & sst24 & 104454.15 + 574426.2 & 0.7 & 0.117 & 0.54 & 10.7 & 44.11 & 102.0 & 43.79 & @xmath11 & @xmath138.64 & 1.54 & 0.53 & @xmath11.58 + 87 & sst24 & 104501.26 + 574632.6 & 0.3 & 0.115 & 0.38 & 6.1 & 43.85 & 32.1 & 43.27 & @xmath11 & @xmath138.63 & 1.28 & 0.01 & @xmath11.56 + 88 & sst24 & 104530.42 + 581231.9 & 0.3 & 0.118 & 0.55 & 16.2 & 44.30 & 19.8 & 43.09 & 1.61 & 38.86 & 1.73 & -0.17 & 1.79 + 89 & sst24 & 104532.47 + 583005.7 & 0.6 & 0.115 & 0.45 & 4.2 & 43.69 & 7.5 & 42.64 & @xmath11 & @xmath138.63 & 1.12 & -0.62 & @xmath11.56 + 90 & sst24 & 104555.55 + 590915.9 & 0.5 & 0.046 & 0.40 & 5.6 & 43.01 & 96.2 & 42.94 & @xmath11 & @xmath137.82 & 0.44 & -0.32 & @xmath10.75 + 91 & sst24 & 104556.97 + 572718.7 & 1.0 & 0.117 & 0.85 & 6.9 & 43.92 & 23.9 & 43.16 & @xmath11 & @xmath138.64 & 1.35 & -0.10 & @xmath11.58 + 92 & sst24 & 104643.29 + 584715.5 & 0.1 & 0.142 & 0.46 & 5.6 & 44.00 & 14.6 & 43.12 & @xmath11 & @xmath138.82 & 1.43 & -0.14 & @xmath11.75 + 93 & sst24 & 104704.63 + 562024.6 & 0.7 & 0.047 & 0.58 & 42.7 & 43.90 & 8.0 & 41.88 & 2.3 & 38.2 & 1.33 & -1.38 & 1.13 + 94 & sst24 & 104709.03 + 593813.4 & 0.6 & 0.092 & 0.69 & 3.9 & 43.46 & 15.3 & 42.75 & @xmath11 & @xmath138.43 & 0.89 & -0.51 & @xmath11.36 + 95 & sst24 & 104713.43 + 581039.3 & 0.6 & 0.062 & 0.57 & 6.7 & 43.34 & 4.3 & 41.85 & @xmath11 & @xmath138.08 & 0.77 & -1.41 & @xmath11.01 + 96 & sst24 & 104729.91 + 572843.3 & 0.2 & 0.230 & 0.64 & 11.2 & 44.74 & 6.0 & 43.17 & 1.24 & 39.34 & 2.17 & -0.09 & 2.27 + 97 & sst24 & 104742.55 + 561850.5 & 0.7 & 0.154 & 0.56 & 8.6 & 44.26 & 25.9 & 43.44 & @xmath11 & @xmath138.89 & 1.69 & 0.18 & @xmath11.82 + 98 & sst24 & 104748.13 + 573438.2 & 0.4 & 0.116 & 0.61 & 7.3 & 43.94 & 22.7 & 43.13 & @xmath11 & @xmath138.64 & 1.37 & -0.13 & @xmath11.57 + 99 & sst24 & 104804.08 + 585136.4 & 0.3 & 0.086 & 0.48 & 5.2 & 43.52 & 37.4 & 43.08 & @xmath11 & @xmath138.37 & 0.95 & -0.18 & @xmath11.3 + 100 & sst24 & 104805.16 + 574854.3 & 0.4 & 0.073 & 0.64 & 10.9 & 43.70 & 19.2 & 42.65 & @xmath11 & @xmath138.23 & 1.13 & -0.61 & @xmath11.16 + 101 & sst24 & 104810.18 + 562249.9 & 0.7 & 0.070 & 0.42 & 7.4 & 43.50 & 15.4 & 42.52 & @xmath11 & @xmath138.19 & 0.93 & -0.74 & @xmath11.12 + 102 & sst24 & 104811.23 + 572430.9 & 0.5 & 0.072 & 0.54 & 4.1 & 43.26 & 33.7 & 42.88 & @xmath11 & @xmath138.21 & 0.69 & -0.38 & @xmath11.14 + 103 & sst24 & 104817.77 + 572410.1 & 0.3 & 0.071 & 0.57 & 5.2 & 43.36 & 72.1 & 43.20 & @xmath11 & @xmath138.2 & 0.79 & -0.06 & @xmath11.13 + 104 & sst24 & 104829.32 + 575355.7 & 1.0 & 0.074 & 0.77 & 9.3 & 43.64 & 68.4 & 43.21 & @xmath11 & @xmath138.24 & 1.07 & -0.05 & @xmath11.17 + 105 & sst24 & 104835.13 + 580124.0 & 0.1 & 0.175 & 0.38 & 3.9 & 44.03 & 17.1 & 43.38 & @xmath11 & @xmath139.01 & 1.46 & 0.12 & @xmath11.94 + 106 & sst24 & 104837.88 + 582642.1 & 0.4 & 0.232 & 0.65 & 12 & 44.78 & 17.6 & 43.64 & @xmath11 & @xmath139.26 & 2.21 & 0.38 & @xmath12.19 + 107 & sst24 & 104844.02 + 580340.9 & 0.8 & 0.164 & 0.46 & 4.9 & 44.07 & 10.9 & 43.12 & @xmath11 & @xmath138.95 & 1.50 & -0.14 & @xmath11.88 + 108 & sst24 & 104904.23 + 561326.4 & 1.4 & 0.071 & 0.40 & 14.3 & 43.79 & 14.1 & 42.49 & 1.4 & 38.35 & 1.22 & -0.77 & 1.28 + 109 & sst24 & 104905.29 + 573733.5 & 0.6 & 0.079 & 0.66 & 6.7 & 43.56 & 17.8 & 42.68 & @xmath11 & @xmath138.3 & 0.99 & -0.58 & @xmath11.23 + 110 & sst24 & 104907.19 + 565715.1 & 0.8 & 0.072 & 0.75 & 16.7 & 43.87 & 14.1 & 42.5 & 1.27 & 38.32 & 1.30 & -0.76 & 1.25 + 111 & sst24 & 104927.49 + 593252.8 & 1.0 & 0.219 & 0.62 & 8.8 & 44.59 & 10.2 & 43.35 & @xmath11 & @xmath139.21 & 2.02 & 0.09 & @xmath12.14 + 112 & sst24 & 104938.42 + 593503.4 & 0.6 & 0.115 & 0.58 & 4.1 & 43.68 & 1.9 & 42.05 & @xmath11 & @xmath138.63 & 1.11 & -1.21 & @xmath11.56 + 113 & sst24 & 104956.37 + 593030.1 & 1.1 & 0.116 & 0.37 & 5.7 & 43.83 & 2.1 & 42.11 & @xmath11 & @xmath138.64 & 1.26 & -1.15 & @xmath11.57 + 114 & sst24 & 105006.00 + 561500.4 & @xmath103 & 0.119 & 0.46 & 9.8 & 44.09 & & & @xmath11 & @xmath138.66 & 1.52 & & @xmath11.59 + 115 & sst24 & 105007.25 + 593254.3 & 0.3 & 0.113 & 0.69 & 10.2 & 44.06 & 23.0 & 43.11 & @xmath11 & @xmath138.61 & 1.49 & -0.15 & @xmath11.54 + 116 & sst24 & 105008.92 + 585543.3 & 0.6 & 0.118 & 0.56 & 5.9 & 43.86 & 57.9 & 43.55 & @xmath11 & @xmath138.65 & 1.29 & 0.29 & @xmath11.58 + 117 & sst24 & 105009.90 + 593029.0 & 0.4 & 0.066 & 1.05 & 4.4 & 43.21 & 26.0 & 42.69 & @xmath11 & @xmath138.14 & 0.64 & -0.57 & @xmath11.07 + 118 & sst24 & 105034.07 + 592221.7 & 0.7 & 0.153 & 0.63 & 7.8 & 44.21 & 17.5 & 43.27 & @xmath11 & @xmath138.88 & 1.64 & 0.01 & @xmath11.82 + 119 & sst24 & 105037.24 + 590544.8 & 0.1 & 0.029 & 0.55 & 16.1 & 43.06 & 111.3 & 42.60 & @xmath11 & @xmath137.41 & 0.49 & -0.66 & @xmath10.34 + 120 & sst24 & 105047.90 + 590348.3 & 1.2 & 0.130 & 0.71 & 5.8 & 43.94 & 36.9 & 43.44 & @xmath11 & @xmath138.74 & 1.37 & 0.18 & @xmath11.67 + 121 & sst24 & 105058.83 + 560550.2 & & 0.122 & 0.55 & 2.3 & 43.47 & & & @xmath11 & @xmath138.68 & 0.90 & & @xmath11.61 + 122 & sst24 & 105110.74 + 590436.9 & 0.3 & 0.117 & 0.40 & 3.9 & 43.67 & 1.3 & 41.90 & @xmath11 & @xmath138.64 & 1.10 & -1.36 & @xmath11.58 + 123 & sst24 & 105128.09 + 573501.7 & 0.6 & 0.070 & 0.57 & 6.6 & 43.44 & 59.3 & 43.1 & @xmath11 & @xmath138.19 & 0.87 & -0.16 & @xmath11.12 + 124 & sst24 & 105200.36 + 591933.4 & 0.9 & 0.118 & 0.37 & 3.2 & 43.59 & 9.0 & 42.75 & 1.23 & 38.74 & 1.02 & -0.51 & 1.67 + 125 & sst24 & 105206.63 + 580947.3 & 0.3 & 0.118 & 0.61 & 22.8 & 44.45 & 31.7 & 43.29 & 1.65 & 38.87 & 1.88 & 0.03 & 1.8 + 126 & sst24 & 105242.40 + 571914.9 & 0.2 & 0.091 & 0.38 & 5.2 & 43.57 & 13.0 & 42.67 & @xmath11 & @xmath138.42 & 1.00 & -0.59 & @xmath11.35 + 127 & sst24 & 105252.83 + 570754.2 & 0.2 & 0.165 & 0.49 & 3.9 & 43.98 & 4.1 & 42.71 & 0.29 & 38.41 & 1.41 & -0.55 & 1.34 + 128 & sst24 & 105301.42 + 570543.2 & 0.5 & 0.080 & 0.64 & 13.7 & 43.88 & 19.9 & 42.74 & 0.74 & 38.18 & 1.31 & -0.52 & 1.11 + 129 & sst24 & 105304.76 + 563328.8 & 0.3 & 0.070 & 0.45 & 8.9 & 43.57 & 7.0 & 42.17 & @xmath11 & @xmath138.19 & 1.00 & -1.09 & @xmath11.12 + 130 & sst24 & 105335.21 + 564431.5 & 0.8 & 0.130 & 0.37 & 10.6 & 44.20 & 15.2 & 43.06 & 2.32 & 39.1 & 1.63 & -0.20 & 2.03 + 131 & sst24 & 105336.93 + 580350.7 & 1.6 & 0.455 & 0.39 & 4 & 44.91 & 3.0 & 43.48 & @xmath11 & @xmath139.87 & 2.34 & 0.22 & @xmath12.8 + 132 & sst24 & 105432.35 + 570932.8 & 0.3 & 0.067 & 0.43 & 10.2 & 43.60 & 18.4 & 42.55 & @xmath11 & @xmath138.15 & 1.03 & -0.71 & @xmath11.08 + 133 & sst24 & 105450.97 + 582327.4 & 0.6 & 0.115 & 0.49 & 6.8 & 43.90 & 10.4 & 42.79 & @xmath11 & @xmath138.63 & 1.33 & -0.47 & @xmath11.56 + 134 & sst24 & 105634.99 + 584748.3 & 0.3 & 0.048 & 0.68 & 24.6 & 43.68 & 114.3 & 43.05 & 1.15 & 37.92 & 1.11 & -0.21 & 0.85 + 135 & sst24 & 105641.86 + 580045.3 & 0.4 & 0.129 & 0.62 & 10.4 & 44.19 & 43.3 & 43.51 & @xmath11 & @xmath138.73 & 1.62 & 0.25 & @xmath11.66 + 136 & sst24 & 105719.83 + 575706.3 & 0.5 & 0.075 & 0.59 & 12.8 & 43.79 & 84.8 & 43.32 & @xmath11 & @xmath138.25 & 1.22 & 0.06 & @xmath11.18 + 137 & sst24 & 105733.51 + 565737.1 & @xmath103 & 0.084 & 0.46 & 4.5 & 43.44 & & & @xmath11 & @xmath138.35 & 0.87 & & @xmath11.28 + 138 & sst24 & 105740.62 + 570616.8 & 0.4 & 0.071 & 0.51 & 4.8 & 43.32 & 13.5 & 42.47 & @xmath11 & @xmath138.2 & 0.75 & -0.79 & @xmath11.13 + 139 & sst24 & 105854.12 + 574130.0 & 1.1 & 0.234 & 0.39 & 2.5 & 44.11 & 10.4 & 43.42 & @xmath11 & @xmath139.27 & 1.54 & 0.16 & @xmath12.2 + 140 & sst24 & 105900.08 + 581526.0 & 0.3 & 0.133 & 0.54 & 4.8 & 43.88 & 14.1 & 43.05 & @xmath11 & @xmath138.76 & 1.31 & -0.21 & @xmath11.69 + 141 & sst24 & 105904.90 + 581920.0 & 0.5 & 0.134 & 0.56 & 3.8 & 43.78 & 24.1 & 43.29 & @xmath11 & @xmath138.77 & 1.21 & 0.03 & @xmath11.7 + 142 & sst24 & 110002.09 + 573141.9 & 0.6 & 0.392 & 0.42 & 7.3 & 45.03 & 9.3 & 43.84 & @xmath11 & @xmath139.73 & 2.46 & 0.58 & @xmath12.67 + 143 & sst24 & 110133.82 + 575205.8 & 0.6 & 0.282 & 0.54 & 8.2 & 44.78 & 1.3 & 42.69 & @xmath11 & @xmath139.44 & 2.21 & -0.57 & @xmath12.37 + 144 & sst24 & 110235.07 + 574655.9 & 0.3 & 0.226 & 0.55 & 4.7 & 44.35 & 8.2 & 43.29 & @xmath11 & @xmath139.24 & 1.78 & 0.03 & @xmath12.17 + 145 & mrk 1450 & 113835.6 + 575227 & & 0.003 & & @xmath13 & @xmath140.72 & & & 2 & 36.13 & @xmath1 - 1.85 & & -0.94 + 146 & um 461 & 115133.3 - 022222 & 2.1 & 0.004 & & @xmath13 & @xmath140.06 & 436.4 & 41.26 & 2.6 & 35.9 & @xmath1 - 2.51 & -2.00 & -1.17 + 147 & sst24 & 142500.18 + 325950.6 & 0.5 & 0.320 & & 2.5 & 44.37 & & & @xmath11 & @xmath139.55 & 1.8 & & @xmath12.48 + 148 & sst24 & 142504.04 + 345013.7 & 0.6 & 0.078 & 0.59 & 16.6 & 43.96 & 10.8 & 42.46 & 1.2 & 38.37 & 1.39 & -0.80 & 1.3 + 149 & sst24 & 142543.59 + 334527.6 & 0.5 & 0.072 & 0.55 & 12.3 & 43.75 & 94.7 & 43.32 & @xmath11 & @xmath138.21 & 1.18 & 0.06 & @xmath11.14 + 150 & sst24 & 142554.57 + 344603.2 & 0.4 & 0.035 & 0.56 & 38.6 & 43.62 & 27.5 & 42.16 & 2.7 & 38.01 & 1.05 & -1.10 & 0.94 + 151 & sst24 & 142607.93 + 345045.4 & 0.7 & 0.408 & 0.46 & 5 & 44.9 & 0.9 & 42.88 & @xmath11 & @xmath139.77 & 2.33 & -0.38 & @xmath12.70 + 152 & sst24 & 142618.65 + 345703.9 & 1.0 & 0.420 & 0.44 & 0.6 & 44.04 & & & @xmath11 & @xmath139.80 & 1.47 & & @xmath12.73 + 153 & sst24 & 142629.15 + 322906.7 & 0.3 & 0.103 & 0.37 & 6.9 & 43.82 & 11.9 & 42.75 & @xmath11 & @xmath138.53 & 1.25 & -0.51 & @xmath11.46 + 154 & sst24 & 142646.63 + 322125.6 & 2.1 & 0.268 & 0.49 & 9.2 & 44.81 & 11.6 & 43.60 & @xmath11 & @xmath139.39 & 2.24 & 0.34 & @xmath12.32 + 155 & sst24 & 142658.80 + 333314.0 & @xmath103 & 0.155 & 0.51 & 16.4 & 44.56 & & & 2.6 & 39.3 & 1.99 & & 2.23 + 156 & sst24 & 142659.15 + 333305.1 & 2.1 & 0.151 & 0.52 & 13 & 44.44 & 23.8 & 43.39 & 1.5 & 39.04 & 1.87 & 0.13 & 1.97 + 157 & sst24 & 142732.90 + 324542.0 & 2.9 & 0.370 & & 1.2 & 44.18 & 1.5 & 43.01 & @xmath11 & @xmath139.68 & 1.61 & -0.25 & @xmath12.61 + 158 & sst24 & 142819.51 + 335150.1 & 0.4 & 0.493 & 0.63 & 4 & 44.98 & 8.4 & 44.01 & @xmath11 & @xmath139.94 & 2.41 & 0.75 & @xmath12.87 + 159 & sst24 & 142849.79 + 343240.3 & 0.5 & 0.219 & 0.53 & 16.3 & 44.84 & 10.0 & 43.34 & 1.8 & 39.46 & 2.27 & 0.08 & 2.39 + 160 & sst24 & 142919.77 + 341506.2 & 1.3 & 0.424 & 0.76 & 2.6 & 44.65 & 9.5 & 43.86 & @xmath11 & @xmath139.81 & 2.08 & 0.60 & @xmath12.74 + 161 & sst24 & 142935.97 + 333713.4 & 0.8 & 0.424 & 0.55 & 3.3 & 44.76 & & & @xmath11 & @xmath139.81 & 2.19 & & @xmath12.74 + 162 & sst24 & 143000.39 + 353814.6 & 0.6 & 0.243 & & 7.1 & 44.58 & 5.2 & 43.15 & @xmath11 & @xmath139.30 & 2.01 & -0.11 & @xmath12.23 + 163 & sst24 & 143024.46 + 325616.4 & 1.0 & 0.040 & 0.58 & 22.2 & 43.48 & 335.0 & 43.36 & 1.1 & 37.75 & 0.91 & 0.10 & 0.68 + 164 & sst24 & 143039.27 + 352351.0 & 0.4 & 0.089 & 0.47 & 14.4 & 44.01 & 6.1 & 42.32 & 2 & 38.7 & 1.44 & -0.94 & 1.63 + 165 & sst24 & 143114.77 + 334623.0 & 2.6 & 0.230 & 0.45 & 7.6 & 44.56 & 5.4 & 43.12 & @xmath11 & @xmath139.25 & 1.99 & -0.14 & @xmath12.18 + 166 & sst24 & 143115.23 + 324606.2 & 0.3 & 0.024 & 0.59 & 9.8 & 42.71 & 179.3 & 42.65 & @xmath11 & @xmath137.25 & 0.14 & -0.61 & @xmath10.18 + 167 & sst24 & 143119.79 + 353418.1 & 0.8 & 0.033 & 0.68 & 39 & 43.58 & 42.3 & 42.30 & 1.7 & 37.77 & 1.01 & -0.96 & 0.7 + 168 & sst24 & 143120.00 + 343804.2 & 0.2 & 0.016 & & 2.7 & 41.76 & 209.5 & 42.33 & @xmath11 & @xmath136.89 & -0.81 & -0.93 & @xmath1 - 0.18 + 169 & sst24 & 143121.15 + 353722.0 & @xmath103 & 0.035 & & 51.9 & 43.74 & & & 4 & 38.17 & 1.17 & & 1.1 + 170 & sst24 & 143125.46 + 331349.8 & 2.3 & 0.023 & 0.77 & 15.1 & 42.85 & 947.8 & 43.34 & 1.1 & 37.26 & 0.28 & 0.08 & 0.19 + 171 & sst24 & 143126.81 + 344517.9 & 2.6 & 0.084 & 0.54 & 13.7 & 43.94 & 3.3 & 42.00 & @xmath11 & @xmath138.35 & 1.37 & -1.26 & @xmath11.28 + 172 & sst24 & 143156.25 + 333833.3 & 0.8 & 0.033 & 0.48 & 27.8 & 43.43 & 161.2 & 42.87 & 1.6 & 37.74 & 0.86 & -0.39 & 0.67 + 173 & sst24 & 143228.36 + 345838.8 & 2.3 & 0.131 & 0.37 & 5.2 & 43.9 & 6.6 & 42.70 & @xmath11 & @xmath138.75 & 1.33 & -0.56 & @xmath11.68 + 174 & sst24 & 143234.90 + 332832.3 & 1.7 & 0.251 & & 4.2 & 44.39 & 7.2 & 43.33 & 2.6 & 39.74 & 1.82 & 0.07 & 2.67 + 175 & sst24 & 143326.18 + 330558.6 & 0.8 & 0.243 & & 4.1 & 44.35 & 2.9 & 42.90 & @xmath11 & @xmath139.30 & 1.78 & -0.36 & @xmath12.23 + 176 & sst24 & 143445.32 + 331346.1 & 0.7 & 0.076 & 0.41 & 7.7 & 43.46 & 8.2 & 42.31 & 1.1 & 38.29 & 0.89 & -0.95 & 1.19 + 177 & sst24 & 143453.85 + 342744.4 & 0.8 & 0.340 & & 1 & 44.02 & 10.9 & 43.78 & @xmath11 & @xmath139.61 & 1.45 & 0.52 & @xmath12.54 + 178 & sst24 & 143619.14 + 332917.9 & 2.7 & 0.189 & 0.75 & 4.9 & 43.37 & 5.0 & 42.91 & @xmath11 & @xmath139.07 & 0.8 & -0.35 & @xmath12.00 + 179 & sst24 & 143628.12 + 333358.0 & 0.2 & 0.264 & 0.40 & 7.8 & 44.66 & 10.5 & 43.54 & @xmath11 & @xmath139.38 & 2.09 & 0.28 & @xmath12.31 + 180 & sst24 & 143632.01 + 335230.7 & 0.7 & 0.088 & 0.84 & 9.4 & 43.82 & 53.5 & 43.26 & @xmath11 & @xmath138.39 & 1.25 & 0.00 & @xmath11.32 + 181 & sst24 & 143636.65 + 345034.0 & 0.3 & 0.278 & 0.66 & 2.4 & 44.19 & 16.0 & 43.76 & @xmath11 & @xmath139.42 & 1.62 & 0.50 & @xmath12.35 + 182 & sst24 & 143641.26 + 345824.4 & @xmath103 & 0.029 & 0.61 & 21.8 & 43.21 & & & @xmath11 & @xmath137.41 & 0.64 & & @xmath10.34 + 183 & cg 0598 & 145920.6 + 421610 & 1.0 & 0.058 & 0.75 & 14.6 & 41.45 & 190.1 & 43.43 & @xmath11 & @xmath138.01 & -1.12 & 0.17 & @xmath10.95 + 184 & cg 0752 & 153121.3 + 470124 & 0.4 & 0.021 & 0.52 & 48.7 & 41.11 & 65.8 & 42.09 & 4.9 & 37.82 & -1.46 & -1.17 & 0.75 + 185 & sst24 & 160038.84 + 551018.4 & 0.3 & 0.143 & 0.42 & 9.7 & 44.24 & 2.2 & 42.31 & 1.32 & 38.94 & 1.67 & -0.95 & 1.87 + 186 & sst24 & 160322.81 + 544237.6 & 0.2 & 0.217 & 0.82 & 5.3 & 44.36 & 8.8 & 43.28 & @xmath11 & @xmath139.2 & 1.79 & 0.02 & @xmath12.13 + 187 & sst24 & 160341.36 + 552612.5 & 0.1 & 0.145 & 0.59 & 8.9 & 44.22 & 4.3 & 42.61 & @xmath11 & @xmath138.84 & 1.65 & -0.65 & @xmath11.77 + 188 & sst24 & 160358.21 + 555504.9 & 1.2 & 0.333 & 0.39 & 3.9 & 44.61 & 0.3 & 42.27 & @xmath11 & @xmath139.59 & 2.04 & -0.99 & @xmath12.52 + 189 & sst24 & 160408.22 + 542531.2 & 1.9 & 0.264 & 0.54 & 7 & 44.66 & 0.6 & 42.29 & @xmath11 & @xmath139.38 & 2.09 & -0.97 & @xmath12.31 + 190 & sst24 & 160408.36 + 545812.7 & 0.3 & 0.064 & 0.52 & 20.1 & 43.85 & 5.4 & 41.98 & 1.68 & 38.34 & 1.28 & -1.28 & 1.27 + 191 & sst24 & 160440.68 + 553409.4 & 0.3 & 0.078 & 0.43 & 15.8 & 43.92 & 29.6 & 42.89 & 1.41 & 38.43 & 1.35 & -0.37 & 1.36 + 192 & sst24 & 160730.53 + 554905.6 & 0.8 & 0.116 & 0.65 & 7 & 43.92 & 30.1 & 43.25 & @xmath11 & @xmath138.64 & 1.35 & -0.01 & @xmath11.57 + 193 & sst24 & 160743.12 + 554416.3 & 0.5 & 0.118 & 0.63 & 15.6 & 44.28 & 13.7 & 42.93 & @xmath11 & @xmath138.65 & 1.71 & -0.33 & @xmath11.58 + 194 & sst24 & 160801.86 + 555400.1 & 1.4 & 0.062 & 0.90 & 6.1 & 43.30 & 199.1 & 43.52 & @xmath11 & @xmath138.08 & 0.73 & 0.26 & @xmath11.01 + 195 & sst24 & 160832.64 + 552927.0 & 0.2 & 0.065 & 0.59 & 8.7 & 43.50 & 1.9 & 41.54 & @xmath11 & @xmath138.12 & 0.93 & -1.72 & @xmath11.05 + 196 & sst24 & 160839.78 + 552331.0 & 0.4 & 0.065 & 0.76 & 9.7 & 43.55 & 17.0 & 42.49 & @xmath11 & @xmath138.12 & 0.98 & -0.77 & @xmath11.05 + 197 & sst24 & 160858.44 + 553010.5 & 0.4 & 0.068 & 0.61 & 4 & 43.21 & 13.6 & 42.44 & @xmath11 & @xmath138.16 & 0.64 & -0.82 & @xmath11.09 + 198 & sst24 & 160858.70 + 563636.2 & 0.7 & 0.115 & 0.55 & 5.2 & 43.78 & 16.4 & 42.98 & @xmath11 & @xmath138.63 & 1.21 & -0.28 & @xmath11.56 + 199 & sst24 & 160907.60 + 552428.3 & 0.0 & 0.640 & 0.53 & 9.6 & 45.60 & 8.6 & 44.25 & @xmath11 & @xmath140.18 & 3.03 & 0.99 & @xmath13.11 + 200 & sst24 & 160908.34 + 552241.2 & 0.2 & 0.086 & 0.57 & 11.6 & 43.87 & 6.4 & 42.31 & @xmath11 & @xmath138.37 & 1.30 & -0.95 & @xmath11.3 + 201 & sst24 & 160931.57 + 541827.8 & 0.2 & 0.083 & 0.41 & 7 & 43.62 & 7.2 & 42.34 & 0.93 & 38.31 & 1.05 & -0.92 & 1.24 + 202 & sst24 & 160937.59 + 541259.6 & 0.3 & 0.085 & 0.65 & 12 & 43.88 & 7.1 & 42.35 & 0.79 & 38.26 & 1.31 & -0.91 & 1.19 + 203 & sst24 & 161103.78 + 544322.4 & 1.4 & 0.063 & 0.45 & 4.8 & 43.21 & 34.5 & 42.77 & 0.45 & 37.74 & 0.64 & -0.49 & 0.67 + 204 & sst24 & 161123.53 + 545158.5 & 0.2 & 0.076 & 0.47 & 6 & 43.47 & 6.4 & 42.21 & 0.43 & 37.89 & 0.90 & -1.05 & 0.82 + 205 & sst24 & 161223.43 + 540339.8 & 2.0 & 0.139 & 0.57 & 7 & 44.08 & 32.2 & 43.45 & 0.61 & 38.58 & 1.51 & 0.19 & 1.51 + 206 & sst24 & 161233.50 + 545630.3 & 0.9 & 0.085 & 0.40 & 7.6 & 43.68 & 55.6 & 43.24 & @xmath11 & @xmath138.36 & 1.11 & -0.02 & @xmath11.29 + 207 & sst24 & 161241.14 + 543957.3 & 0.9 & 0.034 & 0.78 & 3.6 & 42.54 & 18.9 & 41.97 & @xmath11 & @xmath137.55 & -0.03 & -1.29 & @xmath10.48 + 208 & sst24 & 161254.15 + 545525.4 & 0.5 & 0.065 & 0.45 & 14.6 & 43.73 & 15.0 & 42.44 & @xmath11 & @xmath138.12 & 1.16 & -0.82 & @xmath11.05 + 209 & sst24 & 161301.86 + 552123.6 & 1.5 & 0.010 & 0.52 & 22.7 & 42.28 & 24.2 & 41.01 & @xmath11 & @xmath136.48 & -0.29 & -2.25 & @xmath1 - 0.58 + 210 & sst24 & 161403.03 + 560756.6 & 1.3 & 0.062 & 0.61 & 18.2 & 43.78 & 33.3 & 42.74 & @xmath11 & @xmath138.08 & 1.21 & -0.52 & @xmath11.01 + 211 & sst24 & 161411.61 + 540554.4 & 0.6 & 0.305 & 0.45 & 3.1 & 44.43 & 4.2 & 43.27 & @xmath11 & @xmath139.51 & 1.86 & 0.01 & @xmath12.44 + 212 & sst24 & 161528.12 + 534402.1 & 0.5 & 0.125 & 0.60 & 7.3 & 44.00 & 35.6 & 43.39 & @xmath11 & @xmath138.7 & 1.43 & 0.13 & @xmath11.63 + 213 & sst24 & 161548.38 + 534551.2 & 0.2 & 0.149 & 0.56 & 7.8 & 44.19 & 27.5 & 43.44 & @xmath11 & @xmath138.86 & 1.62 & 0.18 & @xmath11.79 + 214 & sst24 & 161551.52 + 541536.1 & 0.9 & 0.215 & 0.56 & 5.2 & 44.35 & 6.2 & 43.12 & @xmath11 & @xmath139.19 & 1.78 & -0.14 & @xmath12.12 + 215 & sst24 & 161644.47 + 533733.9 & 0.4 & 0.149 & 0.52 & 11.3 & 44.35 & 29.2 & 43.47 & @xmath11 & @xmath138.86 & 1.78 & 0.21 & @xmath11.79 + 216 & sst24 & 161659.99 + 560027.2 & 0.7 & 0.063 & 0.41 & 16.8 & 43.76 & 32.3 & 42.74 & @xmath11 & @xmath138.1 & 1.19 & -0.52 & @xmath11.03 + 217 & sst24 & 161712.33 + 551852.7 & 0.1 & 0.036 & 0.39 & 2.9 & 42.50 & 39.9 & 42.34 & @xmath11 & @xmath137.6 & -0.07 & -0.92 & @xmath10.53 + 218 & sst24 & 161716.59 + 550920.4 & 0.5 & 0.085 & 0.39 & 4.2 & 43.42 & 4.9 & 42.19 & @xmath11 & @xmath138.36 & 0.85 & -1.07 & @xmath11.29 + 219 & sst24 & 161819.33 + 541859.6 & 1.3 & 0.081 & 0.39 & 16 & 43.96 & 53.0 & 43.18 & 1.86 & 38.59 & 1.39 & -0.08 & 1.52 + 220 & sst24 & 161823.14 + 552721.3 & 0.3 & 0.083 & 0.64 & 36.3 & 44.34 & 31.4 & 42.98 & 2.07 & 38.66 & 1.77 & -0.28 & 1.59 + 221 & sst24 & 161827.74 + 552209.2 & 0.7 & 0.082 & 0.53 & 8.5 & 43.70 & 0.8 & 41.37 & @xmath11 & @xmath138.33 & 1.13 & -1.89 & @xmath11.26 + 222 & sst24 & 161843.39 + 554433.1 & 0.5 & 0.156 & 0.57 & 9.6 & 44.32 & 29.6 & 43.51 & @xmath11 & @xmath138.9 & 1.75 & 0.25 & @xmath11.83 + 223 & sst24 & 161950.56 + 543715.4 & 0.1 & 0.144 & 0.54 & 10.3 & 44.28 & 5.2 & 42.69 & @xmath11 & @xmath138.83 & 1.71 & -0.57 & @xmath11.76 + 224 & sst24 & 162034.04 + 542324.0 & 1.3 & 0.133 & 0.48 & 6.7 & 44.02 & 36.4 & 43.46 & @xmath11 & @xmath138.76 & 1.45 & 0.20 & @xmath11.69 + 225 & sst24 & 162038.14 + 553521.6 & 0.2 & 0.190 & 0.57 & 9.4 & 44.49 & 6.8 & 43.05 & 1.24 & 39.17 & 1.92 & -0.21 & 2.1 + 226 & sst24 & 162058.88 + 542512.8 & 2.6 & 0.082 & 0.60 & 28.6 & 44.22 & 75.4 & 43.34 & 1.11 & 38.37 & 1.65 & 0.08 & 1.3 + 227 & sst24 & 162059.07 + 542601.8 & 0.7 & 0.042 & 0.54 & 2.7 & 42.61 & 134.4 & 43.01 & @xmath11 & @xmath137.74 & 0.04 & -0.25 & @xmath10.67 + 228 & sst24 & 162128.00 + 551452.9 & 0.8 & 0.099 & 0.52 & 8.7 & 43.87 & 14.6 & 42.8 & @xmath11 & @xmath138.5 & 1.30 & -0.46 & @xmath11.43 + 229 & sst24 & 162133.03 + 551830.1 & 0.5 & 0.244 & 0.37 & 2.9 & 44.21 & 5.2 & 43.16 & @xmath11 & @xmath139.3 & 1.64 & -0.10 & @xmath12.24 + 230 & sst24 & 162150.86 + 553009.4 & 0.8 & 0.098 & 0.69 & 11.3 & 43.97 & 65.2 & 43.44 & @xmath11 & @xmath138.49 & 1.40 & 0.18 & @xmath11.42 + 231 & sst24 & 162210.98 + 550253.5 & 1.1 & 0.030 & 0.42 & 31.4 & 43.38 & 3.7 & 41.15 & 3.3 & 37.96 & 0.81 & -2.11 & 0.89 + 232 & sst24 & 162214.76 + 550614.6 & 0.9 & 0.238 & 0.45 & 7.4 & 44.59 & 16.8 & 43.65 & @xmath11 & @xmath139.28 & 2.02 & 0.39 & @xmath12.21 + 233 & sst24 & 162313.17 + 551111.9 & @xmath103 & 0.235 & 0.59 & 11.5 & 44.77 & & & @xmath11 & @xmath139.27 & 2.20 & & @xmath12.2 + 234 & sst24 & 163001.55 + 410953.3 & 0.8 & 0.123 & 0.58 & 7.5 & 44.00 & 1.1 & 41.87 & @xmath11 & @xmath138.69 & 1.43 & -1.39 & @xmath11.62 + 235 & sst24 & 163220.42 + 402334.5 & 1.0 & 0.079 & 0.42 & 9.4 & 43.70 & 15.3 & 42.62 & @xmath11 & @xmath138.3 & 1.13 & -0.64 & @xmath11.23 + 236 & sst24 & 163310.94 + 405642.2 & 1.6 & 0.136 & 0.43 & 4.2 & 43.84 & 7.5 & 42.8 & @xmath11 & @xmath138.78 & 1.27 & -0.46 & @xmath11.71 + 237 & sst24 & 163335.92 + 401530.1 & 1.8 & 0.028 & 0.63 & 15.6 & 43.01 & 54.5 & 42.26 & @xmath11 & @xmath137.38 & 0.44 & -1.00 & @xmath10.31 + 238 & sst24 & 163359.13 + 405305.0 & 1.0 & 0.031 & 0.58 & 15 & 43.08 & 13.7 & 41.75 & 0.63 & 37.27 & 0.51 & -1.51 & 0.2 + 239 & sst24 & 163401.84 + 412052.3 & 0.5 & 0.028 & 0.57 & 24 & 43.20 & 154.9 & 42.71 & 1.83 & 37.65 & 0.63 & -0.55 & 0.58 + 240 & sst24 & 163506.08 + 411038.2 & 0.8 & 0.078 & 0.52 & 14.8 & 43.89 & 1.1 & 41.46 & 0.89 & 38.23 & 1.32 & -1.80 & 1.16 + 241 & mrk 1499 & 163521.1 + 521253 & & 0.009 & 0.83 & 10.2 & 39.63 & & & 1.5 & 36.59 & -2.94 & & -0.48 + 242 & sst24 & 163546.91 + 403903.3 & 0.2 & 0.122 & 0.77 & 10.6 & 44.14 & 25.6 & 43.23 & @xmath11 & @xmath138.68 & 1.57 & -0.03 & @xmath11.61 + 243 & sst24 & 163608.15 + 410507.8 & 0.7 & 0.167 & 0.41 & 14.6 & 44.56 & 10.5 & 43.12 & 2.94 & 39.43 & 1.99 & -0.14 & 2.36 + 244 & sst24 & 163705.34 + 413156.0 & 0.3 & 0.120 & 0.65 & 15.8 & 44.30 & 19.5 & 43.1 & 0.99 & 38.66 & 1.73 & -0.16 & 1.59 + 245 & sst24 & 163715.60 + 414934.2 & 0.4 & 0.122 & 0.52 & 5.8 & 43.88 & 4.6 & 42.49 & @xmath11 & @xmath138.68 & 1.31 & -0.77 & @xmath11.61 + 246 & sst24 & 163729.31 + 405248.3 & 0.3 & 0.027 & 0.41 & 19.2 & 43.07 & 11.5 & 41.55 & 1.52 & 37.53 & 0.50 & -1.71 & 0.46 + 247 & sst24 & 163751.28 + 401439.6 & 0.3 & 0.071 & 0.66 & 13.9 & 43.78 & 34.5 & 42.88 & @xmath11 & @xmath138.2 & 1.21 & -0.38 & @xmath11.13 + 248 & sst24 & 163751.95 + 401503.6 & 0.8 & 0.072 & 0.69 & 12.4 & 43.74 & 5.9 & 42.12 & @xmath11 & @xmath138.21 & 1.17 & -1.14 & @xmath11.14 + 249 & sst24 & 163805.89 + 413508.3 & 0.4 & 0.118 & 0.44 & 4.5 & 43.74 & 30.4 & 43.27 & @xmath11 & @xmath138.65 & 1.17 & 0.01 & @xmath11.58 + 250 & sst24 & 163808.55 + 403214.2 & 1.3 & 0.225 & 0.46 & 9.2 & 44.63 & 3.4 & 42.91 & @xmath11 & @xmath139.23 & 2.06 & -0.35 & @xmath12.16 + 251 & sst24 & 163809.68 + 402844.4 & 0.7 & 0.070 & 0.57 & 6.3 & 43.43 & 16.9 & 42.56 & @xmath11 & @xmath138.19 & 0.86 & -0.70 & @xmath11.12 + 252 & sst24 & 163906.20 + 404003.7 & 0.2 & 0.035 & 0.63 & 7.4 & 42.89 & 40.7 & 42.33 & @xmath11 & @xmath137.58 & 0.32 & -0.93 & @xmath10.51 + 253 & sst24 & 164043.74 + 413309.8 & 0.3 & 0.153 & 0.59 & 7 & 44.17 & 12.0 & 43.1 & @xmath11 & @xmath138.88 & 1.60 & -0.16 & @xmath11.82 + 254 & sst24 & 164115.40 + 410320.8 & 1.8 & 0.136 & 0.58 & 10.7 & 44.25 & 17.9 & 43.17 & 1.53 & 38.96 & 1.68 & -0.09 & 1.89 + 255 & sst24 & 164211.93 + 410816.4 & 1.0 & 0.143 & 0.51 & 9 & 44.22 & 4.8 & 42.65 & @xmath11 & @xmath138.82 & 1.65 & -0.61 & @xmath11.75 + 256 & sst24 & 171233.85 + 594026.2 & 0.2 & 0.217 & 0.89 & 5.1 & 44.34 & 23.3 & 43.7 & 0.44 & 38.84 & 1.77 & 0.44 & 1.77 + 257 & sst24 & 171239.64 + 584148.3 & 0.3 & 0.166 & 0.86 & 11.5 & 44.45 & 35.8 & 43.65 & 1 & 38.98 & 1.88 & 0.39 & 1.91 + 258 & sst24 & 171308.57 + 601621.0 & 0.9 & 0.334 & 0.66 & 5.2 & 44.74 & 13.5 & 43.86 & 0.6 & 39.38 & 2.17 & 0.60 & 2.31 + 259 & sst24 & 171316.55 + 583235.5 & 0.8 & 0.080 & 0.52 & 9.5 & 43.72 & 9.7 & 42.43 & 0.71 & 38.15 & 1.15 & -0.83 & 1.08 + 260 & sst24 & 171414.89 + 585221.9 & 0.7 & 0.166 & 0.59 & 7.1 & 44.24 & 1.4 & 42.24 & 0.33 & 38.48 & 1.67 & -1.02 & 1.41 + 261 & sst24 & 171446.54 + 593400.3 & 0.8 & 0.130 & 0.62 & 14.7 & 44.34 & 3.3 & 42.4 & 1.06 & 38.76 & 1.77 & -0.86 & 1.69 + 262 & sst24 & 171447.38 + 583805.9 & @xmath103 & 0.258 & 0.55 & 6.7 & 44.62 & & & 0.56 & 39.11 & 2.05 & & 2.04 + 263 & sst24 & 171447.76 + 593509.6 & 1.5 & 0.204 & 0.75 & 6.1 & 44.36 & 24.8 & 43.68 & 0.5 & 38.88 & 1.79 & 0.42 & 1.81 + 264 & sst24 & 171513.93 + 594639.2 & 0.7 & 0.247 & 0.37 & 4.1 & 44.37 & 19.4 & 43.74 & 0.42 & 38.94 & 1.80 & 0.48 & 1.87 + 265 & sst24 & 171542.00 + 591657.4 & 0.5 & 0.119 & 0.54 & 12 & 44.17 & 6.6 & 42.62 & 1.5 & 38.83 & 1.6 & -0.64 & 1.76 + 266 & sst24 & 171555.23 + 601348.0 & 1.6 & 0.430 & 0.30 & 2.5 & 44.65 & 1.5 & 43.13 & 0.3 & 39.36 & 2.08 & -0.13 & 2.29 + 267 & sst24 & 171607.21 + 591456.3 & 0.5 & 0.054 & 0.61 & 15.5 & 43.59 & 83.0 & 43.02 & 0.8 & 37.88 & 1.02 & -0.24 & 0.81 + 268 & sst24 & 171614.52 + 595423.4 & 0.4 & 0.153 & 0.58 & 10.4 & 44.34 & 9.4 & 43 & 0.84 & 38.81 & 1.77 & -0.26 & 1.74 + 269 & sst24 & 171630.27 + 601422.5 & 0.6 & 0.107 & 0.63 & 6.3 & 43.80 & 20.9 & 43.02 & 0.39 & 38.15 & 1.23 & -0.24 & 1.08 + 270 & sst24 & 171634.03 + 601443.6 & 1.3 & 0.106 & 0.93 & 5.1 & 43.7 & 68.9 & 43.53 & 0.6 & 38.35 & 1.13 & 0.27 & 1.28 + 271 & sst24 & 171641.09 + 591857.2 & 0.6 & 0.057 & 0.71 & 10.5 & 43.46 & 54.1 & 42.88 & 0.7 & 37.82 & 0.89 & -0.38 & 0.75 + 272 & sst24 & 171711.17 + 602709.4 & 0.7 & 0.110 & 0.53 & 6.8 & 43.86 & 1.3 & 41.86 & 0.43 & 38.22 & 1.29 & -1.40 & 1.15 + 273 & sst24 & 171933.40 + 592743.2 & 1.0 & 0.142 & 0.45 & 9.1 & 44.21 & 2.4 & 42.34 & 1.13 & 38.87 & 1.64 & -0.92 & 1.8 + 274 & sst24 & 171944.93 + 595708.1 & 0.1 & 0.068 & 0.60 & 17.5 & 43.84 & 21.7 & 42.64 & 1.32 & 38.29 & 1.27 & -0.62 & 1.22 + 275 & sst24 & 172025.19 + 591503.5 & 0.8 & 0.303 & 0.69 & 4.7 & 44.6 & 8.7 & 43.58 & 0.4 & 39.11 & 2.03 & 0.32 & 2.04 + 276 & sst24 & 172043.33 + 584026.5 & 0.4 & 0.125 & 0.54 & 9.7 & 44.13 & 5.7 & 42.6 & 0.85 & 38.63 & 1.56 & -0.66 & 1.56 + 277 & sst24 & 172159.50 + 595034.8 & 1.6 & 0.028 & 0.56 & 13.8 & 42.96 & 17.1 & 41.76 & 0.85 & 37.31 & 0.39 & -1.50 & 0.24 + 278 & sst24 & 172220.27 + 590949.7 & 2.5 & 0.180 & 0.70 & 6.4 & 44.27 & 10.7 & 43.20 & 1.2 & 39.11 & 1.7 & -0.06 & 2.04 + 279 & sst24 & 172313.15 + 590533.8 & 0.9 & 0.107 & 0.83 & 8.3 & 43.92 & 36.7 & 43.27 & 0.64 & 38.37 & 1.35 & 0.01 & 1.3 + 280 & sst24 & 172348.13 + 590154.5 & 0.1 & 0.318 & 0.51 & 3.6 & 44.53 & 1.8 & 42.93 & 0.3 & 39.02 & 1.96 & -0.33 & 1.95 + 281 & sst24 & 172355.65 + 601302.0 & 1.6 & 0.175 & 0.66 & 8 & 44.35 & 40.8 & 43.75 & 0.81 & 38.91 & 1.78 & 0.49 & 1.84 + 282 & sst24 & 172356.10 + 594047.8 & 0.8 & 0.284 & 0.47 & 3.1 & 44.37 & 17.0 & 43.81 & 0.2 & 38.75 & 1.80 & 0.55 & 1.68 + 283 & sst24 & 172400.61 + 590228.2 & 0.6 & 0.179 & 0.44 & 6.2 & 44.25 & 6.7 & 42.99 & 1.5 & 39.19 & 1.68 & -0.27 & 2.12 + 284 & sst24 & 172402.14 + 600601.5 & 2.3 & 0.155 & 0.45 & 6 & 44.11 & 22.4 & 43.38 & 0.59 & 38.66 & 1.54 & 0.12 & 1.59 + 285 & sst24 & 172546.84 + 593655.5 & 0.6 & 0.035 & 0.50 & 26.1 & 43.43 & 172.3 & 42.95 & 1.95 & 37.87 & 0.86 & -0.31 & 0.8 + 286 & sst24 & 172551.44 + 601139.4 & @xmath103 & 0.029 & 0.44 & 25.2 & 43.25 & & & 1.03 & 37.43 & 0.68 & & 0.36 + 287 & sst24 & 172611.96 + 592851.8 & 0.2 & 0.073 & 0.56 & 13 & 43.78 & 15.5 & 42.55 & 1 & 38.21 & 1.21 & -0.71 & 1.14 + 1 & mrk 13 & 075614.20 + 601815.0 & 0.0046 & @xmath1 65 & 476.9 & 41.62 & @xmath141.64 & @xmath1 - 0.93 & -1.64 + 2 & mrk 14 & 081059.30 + 724740.0 & 0.0081 & 151 & 883.5 & 42.39 & 42.51 & -0.06 & -0.87 + 3 & mrk 21 & 094830.80 + 575815.0 & 0.0281 & 109 & 304.5 & 43.01 & 43.45 & 0.88 & -0.25 + 4 & mrk 25 & 100351.80 + 592610.0 & 0.0100 & 251 & 817.4 & 42.53 & 42.91 & 0.34 & -0.73 + 5 & mrk 31 & 101942.80 + 572524.0 & 0.0260 & @xmath1 65 & 135.4 & 42.59 & @xmath143.16 & @xmath10.59 & -0.67 + 6 & mrk 44 & 120324.50 + 384848.0 & 0.0227 & @xmath1 65 & 122 & 42.43 & @xmath143.04 & @xmath10.47 & -0.83 + 7 & mrk 52 & 122542.80 + 003422.0 & 0.0071 & 1050 & 2311.2 & 42.69 & 43.24 & 0.67 & -0.57 + 8 & mrk 87 & 082140.30 + 735918.0 & 0.0094 & 118 & 187.2 & 41.84 & 42.53 & -0.04 & -1.42 + 9 & mrk 90 & 083000.00 + 524150.0 & 0.0142 & 156 & 193.9 & 42.22 & 43.01 & 0.44 & -1.04 + 10 & mrk 93 & 083642.30 + 661359.0 & 0.0177 & 467 & 146.6 & 42.28 & 43.68 & 1.11 & -0.98 + 11 & mrk 105 & 092026.30 + 712416.0 & 0.0118 & @xmath1 65 & 119.8 & 41.84 & @xmath142.47 & @xmath1 - 0.1 & -1.42 + 12 & mrk 140 & 101628.20 + 451918.0 & 0.0055 & @xmath1 65 & 737.5 & 41.97 & @xmath141.81 & @xmath1 - 0.76 & -1.29 + 13 & mrk 149 & 103801.60 + 641559.0 & 0.0057 & 194 & 163.3 & 41.34 & 42.31 & -0.26 & -1.92 + 14 & mrk 161 & 110159.00 + 451341.0 & 0.0200 & 400 & 681 & 43.06 & 43.72 & 1.15 & -0.20 + 15 & mrk 165 & 111836.30 + 631650.0 & 0.0108 & @xmath1 65 & 785 & 42.58 & @xmath142.39 & @xmath1 - 0.18 & -0.68 + 16 & mrk 168 & 112546.80 + 470000.0 & 0.0342 & @xmath1 65 & 157.6 & 42.89 & @xmath143.4 & @xmath10.83 & -0.37 + 17 & mrk 185 & 114116.60 + 474127.0 & 0.0104 & 173 & 62.7 & 41.45 & 42.78 & 0.21 & -1.81 + 18 & mrk 190 & 115147.60 + 484059.0 & 0.0033 & 391 & 2060.9 & 41.97 & 42.14 & -0.43 & -1.29 + 19 & mrk 193 & 115528.30 + 573952.0 & 0.0172 & 103 & 331.7 & 42.62 & 43.00 & 0.43 & -0.64 + 20 & mrk 203 & 121812.80 + 441024.0 & 0.0246 & @xmath1 65 & 784.6 & 43.30 & @xmath143.11 & @xmath10.54 & 0.04 + 21 & mrk 206 & 122417.00 + 672624.0 & 0.0044 & 221 & 244.2 & 41.29 & 42.14 & -0.43 & -1.97 + 22 & mrk 282 & 140207.20 + 692859.0 & 0.0202 & @xmath1 65 & 420.8 & 42.86 & @xmath142.94 & @xmath10.37 & -0.40 + 23 & mrk 286 & 141926.70 + 713518.0 & 0.0253 & 704 & 1223.6 & 43.52 & 44.17 & 1.6 & 0.26 + 24 & mrk 307 & 223556.30 + 201920.0 & 0.0185 & 189 & 1403 & 43.31 & 43.33 & 0.76 & 0.05 + 25 & mrk 316 & 231340.50 + 140116.0 & 0.0409 & @xmath1 65 & 439.4 & 43.50 & @xmath143.56 & @xmath10.99 & 0.24 + 26 & mrk 319 & 231838.30 + 251358.0 & 0.0270 & 542 & 330.9 & 43.01 & 44.11 & 1.54 & -0.25 + 27 & mrk 326 & 232806.20 + 233153.0 & 0.0119 & 715 & 506.7 & 42.48 & 43.51 & 0.94 & -0.78 + 28 & mrk 353 & 010316.40 + 222034.0 & 0.0156 & 574 & 265.5 & 42.43 & 43.66 & 1.09 & -0.83 + 29 & mrk 357 & 012240.60 + 231015.0 & 0.0529 & @xmath1 65 & 1164.7 & 44.15 & @xmath143.78 & @xmath11.21 & 0.89 + 30 & mrk 363 & 015058.50 + 215951.0 & 0.0098 & 270 & 772.3 & 42.50 & 42.93 & 0.36 & -0.76 + 31 & mrk 369 & 023728.20 + 210831.0 & 0.0136 & @xmath1 65 & 486.5 & 42.58 & @xmath142.59 & @xmath10.02 & -0.68 + 32 & mrk 373 & 065435.10 + 502109.0 & 0.0197 & 316 & 179.3 & 42.47 & 43.60 & 1.03 & -0.79 + 33 & mrk 391 & 085446.30 + 393219.0 & 0.0133 & 145 & 284.5 & 42.32 & 42.92 & 0.35 & -0.94 + 34 & mrk 401 & 093017.00 + 293224.0 & 0.0057 & 562 & 1224.8 & 42.22 & 42.77 & 0.2 & -1.04 + 35 & mrk 409 & 094941.20 + 321316.0 & 0.0051 & @xmath1 65 & 697.9 & 41.88 & @xmath141.74 & @xmath1 - 0.83 & -1.38 + 36 & mrk 412 & 095759.70 + 321422.0 & 0.0150 & @xmath1 65 & 854.6 & 42.91 & @xmath142.68 & @xmath10.11 & -0.35 + 37 & mrk 413 & 095915.90 + 314200.0 & 0.0386 & @xmath1 65 & 132.3 & 42.92 & @xmath143.5 & @xmath10.93 & -0.34 + 38 & mrk 420 & 110340.30 + 375546.0 & 0.0430 & @xmath1 65 & 429.8 & 43.53 & @xmath143.6 & @xmath11.03 & 0.27 + 39 & mrk 439 & 122436.20 + 392259.0 & 0.0035 & 707 & 3756.1 & 42.28 & 42.45 & -0.12 & -0.98 + 40 & mrk 446 & 125008.70 + 330933.0 & 0.0237 & 288 & 715 & 43.23 & 43.72 & 1.15 & -0.03 + 41 & mrk 465 & 140124.00 + 364800.0 & 0.0090 & 90 & 650.2 & 42.35 & 42.38 & -0.19 & -0.91 + 42 & mrk 480 & 150633.80 + 423829.0 & 0.0180 & 199 & 522.9 & 42.85 & 43.32 & 0.75 & -0.41 + 43 & mrk 489 & 154421.50 + 410511.0 & 0.0317 & @xmath1 65 & 972.2 & 43.62 & @xmath143.33 & @xmath10.76 & 0.36 + 44 & mrk 496 & 161140.70 + 522724.0 & 0.0293 & 1113 & 1158.8 & 43.63 & 44.50 & 1.93 & 0.37 + 45 & mrk 529 & 231422.20 - 024341.0 & 0.0118 & 228 & 580.3 & 42.53 & 43.02 & 0.45 & -0.73 + 46 & mrk 538 & 233614.10 + 020919.0 & 0.0093 & 2850 & 4189.8 & 43.18 & 43.91 & 1.34 & -0.08 + 47 & mrk 545 & 000953.40 + 255526.0 & 0.0152 & 1082 & 966.3 & 42.97 & 43.91 & 1.34 & -0.29 + 48 & mrk 575 & 014833.10 + 123650.0 & 0.0183 & 491 & 108.1 & 42.18 & 43.73 & 1.16 & -1.08 + 49 & mrk 602 & 025950.60 + 024617.0 & 0.0095 & 625 & 157.2 & 41.77 & 43.26 & 0.69 & -1.49 + 50 & mrk 638 & 114137.90 + 375931.0 & 0.0222 & 145 & 392.7 & 42.91 & 43.37 & 0.8 & -0.35 + 51 & mrk 684 & 143104.80 + 281714.0 & 0.0461 & 129 & 1647.1 & 44.18 & 43.96 & 1.39 & 0.92 + 52 & mrk 708 & 094211.50 + 044023.0 & 0.0068 & 765 & 76.9 & 41.17 & 43.06 & 0.49 & -2.09 + 53 & mrk 710 & 095449.50 + 091616.0 & 0.0049 & 426 & 1251.9 & 42.09 & 42.51 & -0.06 & -1.17 + 54 & mrk 711 & 095511.30 + 132546.0 & 0.0192 & 397 & 156 & 42.38 & 43.68 & 1.11 & -0.88 + 55 & mrk 717 & 101040.30 + 242451.0 & 0.0212 & 765 & 217.2 & 42.62 & 44.05 & 1.48 & -0.64 + 56 & mrk 719 & 101559.90 + 045717.0 & 0.0320 & @xmath1 65 & 498.3 & 43.34 & @xmath143.34 & @xmath10.77 & 0.08 + 57 & mrk 732 & 111349.70 + 093511.0 & 0.0292 & @xmath1 65 & 415 & 43.18 & @xmath143.26 & @xmath10.69 & -0.08 + 58 & mrk 743 & 113813.00 + 120643.0 & 0.0033 & 180 & 2982.4 & 42.13 & 41.80 & -0.77 & -1.13 + 59 & mrk 752 & 115243.40 + 014427.0 & 0.0204 & 260 & 989.8 & 43.24 & 43.55 & 0.98 & -0.02 + 60 & mrk 759 & 121037.50 + 160159.0 & 0.0072 & 537 & 4418 & 42.99 & 42.96 & 0.39 & -0.27 + 61 & mrk 764 & 121600.00 + 124114.0 & 0.0661 & @xmath1 65 & 242.1 & 43.66 & @xmath143.98 & @xmath11.41 & 0.40 + 62 & mrk 769 & 122525.50 + 162812.0 & 0.0057 & 1048 & 5770.3 & 42.90 & 43.04 & 0.47 & -0.36 + 63 & mrk 781 & 125350.80 + 094236.0 & 0.0094 & @xmath1 65 & 465.5 & 42.24 & @xmath142.28 & @xmath1 - 0.29 & -1.02 + 64 & mrk 789 & 133224.20 + 110623.0 & 0.0315 & 618 & 451.5 & 43.28 & 44.30 & 1.73 & 0.02 + 65 & mrk 809 & 142233.30 + 134302.0 & 0.0256 & 211 & 332.7 & 42.95 & 43.64 & 1.07 & -0.31 + 66 & mrk 863 & 155625.90 + 090319.0 & 0.0423 & 2442 & 340.6 & 43.42 & 45.16 & 2.59 & 0.16 + 67 & mrk 897 & 210745.80 + 035240.0 & 0.0263 & @xmath1 65 & 249.3 & 42.86 & @xmath143.17 & @xmath10.6 & -0.40 + 68 & mrk 912 & 222815.00 - 025304.0 & 0.0166 & 165 & 81.3 & 41.97 & 43.17 & 0.6 & -1.29 + 69 & mrk 921 & 224830.70 + 314652.0 & 0.0128 & @xmath1 65 & 379.3 & 42.42 & @xmath142.54 & @xmath1 - 0.03 & -0.84 + 70 & mrk 1002 & 013717.50 + 055240.0 & 0.0106 & 787 & 653.6 & 42.49 & 43.46 & 0.89 & -0.77 + 71 & mrk 1012 & 015657.80 - 052410.0 & 0.0163 & 522 & 218.2 & 42.39 & 43.66 & 1.09 & -0.87 + 72 & mrk 1158 & 013459.30 + 350222.0 & 0.0153 & 219 & 402.7 & 42.60 & 43.22 & 0.65 & -0.66 + 73 & mrk 1259 & 103833.80 - 071015.0 & 0.0073 & @xmath1 65 & 2443.7 & 42.73 & @xmath142.05 & @xmath1 - 0.52 & -0.53 + 74 & mrk 1301 & 113550.00 + 352007.0 & 0.0053 & @xmath1 65 & 177.5 & 41.33 & @xmath141.78 & @xmath1 - 0.79 & -1.93 + 75 & mrk 1308 & 115412.30 + 000812.0 & 0.0038 & 205 & 1749.2 & 42.03 & 41.99 & -0.58 & -1.23 + 76 & mrk 1379 & 141740.20 - 072503.0 & 0.0086 & 817 & 2117.7 & 42.83 & 43.31 & 0.74 & -0.43 + positions for sources in table 1 , shown as function of f@xmath4(7.7@xmath0 m ) to illustrate that measured positional differences do not depend on brightness .
crosses are measured differences ; triangles are sources with differences @xmath10 3 ; diamonds are bcd sources without pah features .
median difference is 0.64 . ]
irs spectra for all 25 sources in the @xmath2 10 mjy sample of @xcite which have sdss classifications as starburst ( hii for sdss classification , instead of any agn classification ) .
all spectra are normalized to f@xmath4(7.7@xmath0 m ) = 1 mjy ; uniformity of the pah features among starbursts is illustrated by the small scatter in pah strengths .
primary spectroscopic differences are among the slopes of the dust continuum underlying pah features ; representative fit to dust continuum shown by diagonal solid line
. strong pah emission feature at 7.7 @xmath0 m is feature used in text to measure sfr(pah ) using @xmath3l@xmath4(7.7 @xmath0 m ) . ]
m pah luminosity for all sources with radio detections in table 1 ( asterisks are bcds ) .
horizontal line is the median value of 0.0 , indicating no offset in sfr(pah ) compared to sfr(radio ) .
@xmath3l@xmath4(7.7 @xmath0 m ) is the luminosity at the peak of the 7.7 @xmath0 m feature in ergs s@xmath7 ( log [ @xmath3l@xmath4(7.7 @xmath0m)(l@xmath6 ) ] = log [ @xmath3l@xmath4(7.7 @xmath0m)(ergs s@xmath7 ) ] - 33.59 ) .
sfr(radio ) determined from log [ sfr(radio ) ] = log [ @xmath3l@xmath4(1.4 ghz ) ] - 37.07 for @xmath3l@xmath4(1.4 ghz ) in ergs s@xmath7 and sfr(radio ) in m@xmath6 yr@xmath7 .
sfr(pah ) determined from log [ sfr(pah ) ] = log [ @xmath3l@xmath4(7.7 @xmath0 m ) ] - 42.57 .
error cross in lower right hand corner shows 1@xmath27 uncertainties in both directions ; uncertainty for sfr(pah ) arises from uncertainty in determining transformation from @xmath3l@xmath4(7.7 @xmath0 m ) to the total l@xmath8 which determines sfr ; uncertainty in luminosity arises from @xmath9 10% uncertainty in measurement of f@xmath4(7.7 @xmath0 m ) .
( using empirically determined transformation between @xmath3l@xmath4(7.7 @xmath0 m ) and total infrared luminosity l@xmath8 , log l@xmath8 = log @xmath3l@xmath4(7.7 @xmath0 m ) - [email protected] for l@xmath8 in l@xmath6 and @xmath3l@xmath4(7.7 @xmath0 m ) in ergs s@xmath7 . ) ] m pah luminosity for all sources without radio detections in table 1 using 1 mjy as the limit for f@xmath4(1.4 ghz ) ( asterisk is a bcd ) .
limit is in the sense that sfr(pah)/sfr(radio ) is actually greater than the limit plotted .
horizontal line is the median of 0.0 for the detected sources in figure 3a ; as expected , the median for limits is less than 0.0 . ]
= log [ @xmath3l@xmath4(fuv ) ] - 43.26 for @xmath3l@xmath4(fuv ) in ergs s@xmath7 and sfr(uv ) in m@xmath6 yr@xmath7 .
sfr(pah ) determined from log [ sfr(pah ) ] = log [ @xmath3l@xmath4(7.7 @xmath0 m ) ] - 42.57 .
sfr(pah ) exceeds sfr(uv ) , as would expected for extinction by dust of the ultraviolet continuum ; median value of ratio log [ sfr(pah)/sfr(uv ) ] = 1.67 , indicating that only 2% of the ultraviolet continuum typically escapes a starburst ; ratio increases with infrared pah luminosity according to the linear fit which is shown , of form log [ sfr(pah)/sfr(uv)]= ( [email protected])log [ @xmath3l@xmath4(7.7 @xmath0 m ) ] - [email protected] .
error cross in lower right hand corner shows 1@xmath27 uncertainties in both directions ; uncertainty for sfr(pah ) arises from uncertainty in determining transformation from @xmath3l@xmath4(7.7 @xmath0 m ) to the total l@xmath8 which determines sfr ; uncertainty in luminosity arises from @xmath9 10% uncertainty in measurement of f@xmath4(7.7 @xmath0 m ) .
( using empirically determined transformation between @xmath3l@xmath4(7.7 @xmath0 m ) and total infrared luminosity l@xmath8 , log l@xmath8 = log @xmath3l@xmath4(7.7 @xmath0 m ) - [email protected] for l@xmath8 in l@xmath6 and @xmath3l@xmath4(7.7 @xmath0 m ) in ergs s@xmath7 . ) ] /@xmath26(starburst ) .
solid curve shows fraction as measured by sfr(uv)/[sfr(pah ) + sfr(uv ) ] from equation ( 2 ) ; vertical bars show @xmath5 1@xmath27 range of values using @xmath26/(@xmath16 + @xmath26 ) for ultraviolet - selected galaxies in @xcite .
zero value corresponds to source without any dust extinction . ]
( 25 @xmath0 m ) observed by iras , which is transformed to f@xmath4(7.7 @xmath0 m ) using average infrared spectra for sources with 42 @xmath1 log [ @xmath3l@xmath4(15 @xmath0 m ) ] @xmath1 44 in @xcite , for which f@xmath4(25
@xmath0m)/f@xmath4(7.7 @xmath0 m ) = [email protected] .
error cross in lower right hand corner shows 1@xmath27 uncertainties in both directions ; uncertainty for sfr(pah ) arises from uncertainty in determining transformation from @xmath3l@xmath4(7.7 @xmath0 m ) to the total l@xmath8 which determines sfr combined with uncertainty in assumed ratio f@xmath4(25 @xmath0m)/f@xmath4(7.7 @xmath0 m ) ; uncertainty in luminosity arises from uncertainty in ratio of f@xmath4(25 @xmath0m)/f@xmath4(7.7 @xmath0 m ) .
diamonds correspond to upper limits in ratio for sources detected by galex but not by iras with iras limit taken as 65 mjy ; limit is in the sense that sfr(pah)/sfr(uv ) is smaller than limit shown .
linear fit shown has form log [ sfr(pah)/sfr(uv)]= ( [email protected])log @xmath3l@xmath4(7.7 @xmath0 m ) - [email protected] .
( using empirically determined transformation between @xmath3l@xmath4(7.7 @xmath0 m ) and total infrared luminosity l@xmath8 , log l@xmath8 = log @xmath3l@xmath4(7.7 @xmath0 m ) - [email protected] for l@xmath8 in l@xmath6 and @xmath3l@xmath4(7.7 @xmath0 m ) in ergs s@xmath7 . ) ] | we present a comparison of star formation rates ( sfr ) determined from mid - infrared 7.7 @xmath0 m pah luminosity [ sfr(pah ) ] , from 1.4 ghz radio luminosity [ sfr(radio ) ] , and from far ultraviolet luminosity [ sfr(uv ) ] for a sample of 287 starburst galaxies with z @xmath1 0.5 having @xmath2 irs observations .
the previously adopted relation log [ sfr(pah ) ] = log [ @xmath3l@xmath4(7.7 @xmath0 m ) ] - [email protected] , for sfr in m@xmath6 yr@xmath7 and @xmath3l@xmath4(7.7 @xmath0 m ) the luminosity at the peak of the 7.7 @xmath0 m pah feature in ergs s@xmath7 , is found to agree with sfr(radio ) . comparing with sfr(uv ) determined independently from ultraviolet observations of the same sources with the galex mission ( not corrected for dust extinction ) , the median log [ sfr(pah)/sfr(uv ) ] = 1.67 , indicating that only 2% of the ultraviolet continuum typically escapes extinction by dust within a starburst .
this ratio sfr(pah)/sfr(uv ) depends on infrared luminosity , with form log [ sfr(pah)/sfr(uv ) ] = ( [email protected])log [ @xmath3l@xmath4(7.7 @xmath0 m ) ] - [email protected] , indicating that more luminous starbursts are also dustier . using our adopted relation between @xmath3l@xmath4(7.7 @xmath0 m ) and l@xmath8 , this becomes log [ sfr(pah)/sfr(uv)]= ( [email protected])log l@xmath8 - [email protected] , for l@xmath8 in l@xmath6 .
only blue compact dwarf galaxies show comparable or greater sfr(uv ) compared to sfr(pah ) .
we also find that the ratio sfr(pah)/sfr(uv ) is similar to that in infrared - selected starbursts for a sample of markarian starburst galaxies originally selected using optical classification , which implies that there is no significant selection effect in sfr(pah)/sfr(uv ) using starburst galaxies discovered by @xmath2 .
these results indicate that sfrs determined with ultraviolet luminosities require dust corrections by a factor of @xmath9 10 for typical local starbursts but this factor increases to @xmath10 700 for the most luminous starbursts at z @xmath9 2.5 .
application of this factor explains why the most luminous starbursts discovered by @xmath2 at z @xmath9 2.5 are optically faint ; with this amount of extinction , the optical magnitude of a starburst having f@xmath4(7.7 @xmath0 m ) of 1 mjy should be @xmath11 @xmath9 25.6 . |
although the methods and techniques have been greatly improved since the late nineteenth century , the precision on the measurement of the gravitational constant , g , is still the less accurate in comparison with the other fundamental constants of nature @xcite .
moreover , given the relative uncertainties of most of the individual experiments ( reaching about @xmath0 for the most precise measurements ) , they show an incompatibility which leads to an overall precision of only about 1 part in @xmath1 .
thus the current status of the @xmath2 terrestrial measurements ( see @xcite ) implies either an unknown source of errors ( not taken into account in the published uncertainties ) , or some new physics @xcite . in the latter spirit
, many theories which include extradimensions have been proposed as candidates for the unification of physics . as such
, they involve a coupling between gravitation and electromagnetism ( ge coupling ) , as well as with other gauge fields present .
+ here we show that the discrepancy between the present results of the g - measurements may be understood as a consequence of the ge coupling .
also , this theory predicts a variation of the effective fine structure `` constant '' @xmath3 with the gravitational field , and thus with the cosmological time ( see @xcite ) .
an argument initially from landau and lifshitz @xcite may be applied to the pure kaluza - klein ( kk ) action ( see @xcite ) : the negative sign of the kinetic term of the five dimensional ( 5d ) kk internal scalar field , @xmath4 , leads to inescapable instability . the question to know which of the two conformally related frames ( einstein - pauli frame or jordan - fierz frame ) is the best remains debated in the literature , each frame having its own advantage .
the negative kinetic energy density for the @xmath4-field , and thus the instability , occur in both frames . in the following ,
we perform calculations in the jordan - fierz frame , where , as we show , the discrepant laboratory measurements of g find a natural explanation .
also , it is true that the 5d kk theory may yield a zero kinetic term ( and thus a zero kinetic energy density , which would also be unusual in 4d ) , but this occurs only when the electromagnetic ( em ) field is identically zero everywhere .
this may be relevant for some cosmological solutions , but not for our discussion .
on the contrary , we argue here that the @xmath4-field is tightly related to the em field via the link between the compactified space of the fifth dimension and the u(1 ) gauge group .
since stabilization may be obtained if an external field is present , we assume here a version , kk@xmath5 , which includes an external bulk scalar field minimally coupled to gravity ( like the radion in the brane world scenario ) .
after dimensional reduction ( @xmath6 ) , this bulk field reduces to a four dimensional scalar field @xmath7 and , in the jordan - fierz frame , the low energy effective action takes the form ( up to a total divergence ) @xmath8 \,d^{4}x\ ] ] @xmath9 \,d^{4}x,\ ] ] where @xmath10 is the potential 4-vector of the em field , @xmath11 the em field strength tensor , @xmath12 the self - interaction potential of @xmath5 of the symmetry breaking type and @xmath13 its source term .
the source term of the @xmath5-field , @xmath13 , includes the contributions of the ordinary matter , of the em field and of the internal scalar field @xmath4 . for each
, the coupling is defined by a function ( temperature dependent , as for the potential @xmath12 ) @xmath14 , where the subscript @xmath15 stands for `` matter '' , `` em '' and `` @xmath4 '' .
in addition , the necessity to recover the einstein - maxwell equations in the weak fields limit , implies the following conditions : @xmath16 and @xmath17 , where @xmath18 denotes the vacuum expectation value ( vev ) of the @xmath5-field .
the contributions of matter and @xmath4 are proportional to the traces of their respective energy - momentum tensors .
a contribution of the form @xmath19 accounts for the coupling with the em field . besides ,
though the fit to the data involves @xmath20 , we may infer that @xmath21 is negligibly small at very high temperature ( _ e.g. _ , like in the core of the sun or at the big bang nucleosynthesis ) and even vanishes beyond the critical temperature ( say @xmath22 k ) of the potential @xmath23 as one gets @xmath24 in that case .
following lichnerowicz @xcite , let us interpret the quantity @xmath25 of the einstein - hilbert term , and the factor @xmath26 $ ] of the maxwell term respectively as the effective gravitational `` constant '' and the effective vacuum dielectric permittivity .
the effective vacuum magnetic permeability reads @xmath27^{-1}$ ] , so that the velocity of light in vacuum remains a true universal constant .
both terms depend on the local ( for local physics ) or global ( at cosmological scale ) value of the kk scalar field @xmath4 , assumed to be positive defined .
the least action principle applied to the action ( [ kk action and psi j - f frame ] ) yields the generalized einstein - maxwell equations and the scalar fields equations @xmath28 and @xmath29
let us study the _ spatial _ variation of @xmath4 in the weak fields conditions out of the fields source , but in presence of a static dipolar magnetic field , @xmath30 .
we denote @xmath31 , @xmath32 and @xmath33 respectively the radius from the centre , the azimuth angle and the colatitude .
thus , writing @xmath34 , and taking into account that @xmath35 ( definition of the vev ) , equation ( [ kk scalar eq ] ) simplifies after linearization as @xmath36 \ , ( \,\vec{\nabla } v \,)^{2}.\ ] ] since @xmath37 and @xmath38 identically , the solution of equation ( [ stabilized kk scalar eq stat .
approx ] ) above reads merely @xmath39 \,v^{2}.\ ] ] for our purpose , it is sufficient to limit the expansion of the scalar potential , @xmath40 , to the terms of the legendre function of degree one ( @xmath41 ) and order one ( @xmath42 ) . hence @xmath43 $ ] , where @xmath44 , @xmath45 and @xmath46 are the relevant gauss coefficients , @xmath47 is the earth s radius and @xmath48 denotes its magnetic moment . setting @xmath49 , @xmath50 and @xmath51 , the solution of equation ( [ stabilized kk scalar eq stat .
approx ] ) then reads ( and similarly for @xmath5 by making the substitution @xmath52 ) @xmath53 where we have set @xmath54 thence , one derives the expression of @xmath55 by inserting the solution ( [ stabilized kk scalar sol .
1st approx ] ) above in relation ( [ g eff ] ) .
it is worth noticing that the magnetic potential , @xmath40 , scales as @xmath56 , where @xmath57 is the magnitude of the geomagnetic field at radius @xmath31 . indeed , because of this scaling effect , the small spatial variations of the geomagnetic field will influence significantly the laboratory measurements of g whereas the large local magnetic fields present in the laboratory ( _ e .
g. _ , the magnetic suspension used to support the balance beam , the magnetic damper , etc ... ) will not .
a rough estimate shows that , even using a 30 tesla superconducting magnet , one still needs to gain at least one order of magnitude with the most precise g measuring apparatus avalaible yet .
hence , our prediction is consistent with the earlier conclusion of lloyd @xcite .
there are presently almost 45 results of g measurements published since 1942 @xcite .
because of the too numerous uncontrolled systematic biases , the mine measurements are excluded from the present study ( including them will not change our fit because of their lack of precision , typically less than @xmath58 ) . also , the more discordant laboratory measurement ( high ptb value @xcite ) is excluded , since it may suffer from a systematic effect .
the `` official '' values are presently @xmath59 ( codata 86 , @xcite ) and @xmath60 ( codata 2000 , @xcite ) in mks unit .
in the following , all the measurements are weighted equally in the fit .
fitting the 44 data with these values gives respectively @xmath61 and @xmath62 ( @xmath63 per degrees of freedom ) .
if we forget the official values and try a best fit , assuming an arbitrary constant value of @xmath2 , we obtain @xmath64 with @xmath65 .
the fit to the same sample of 44 measurements ( figure 1 ) , on account of the ge coupling , yields ( in mks units ) with @xmath66 : @xmath67 from the above fit , one derives both estimates of the true gravitational constant @xmath68 and the coupling parameter @xmath69 the latter quantity , expressed in the canonical form @xmath70 , yields a 5d planck scale @xmath71 tev of the order of the value that is invoked in the literature to solve the hierarchy problem . [ cols="^,^,^,^",options="header " , ] + + table 2 : reduced @xmath72 for the two different hypothesis h0 ( hypothesis of a constant @xmath2 ) and h1 ( hypothesis of an effective g ) , and different samples s1 and whole ( except the high ptb value @xcite , see text ) .
+ considering the whole sample , we check the relevance of our result under h1 compared to the best value of @xmath2 under h0 , by applying the f test ( fisher law ) .
this yields @xmath73 , which indicates that , independently of the number of parameters ( two instead of one ) , our fit is better with a significance level greater than @xmath74 @xcite . let us emphasize that the most precise value of g today @xcite contributes to @xmath75 to less than @xmath76 in the above fit . likewise , the last published value of g @xcite would contribute to less than @xmath77 .
this suggests that the agreement may be better than purely indicated by the @xmath78 values .
moreover , if one substitutes @xmath79
@xcite announced by the msl team in june 2002 ( cpem , ottawa , canada ) for the prior @xmath80 published in 1999 @xcite , one would obtain with the sample s1 : h0 : @xmath81 h1 : @xmath82 and with the whole sample of 44 measurements : h0 : @xmath83 h1 : @xmath84
it is worth noticing that the scalar fields under considerations identify neither to the dilaton nor to the inflaton of higher dimensional theories , without further assumptions . in particular
, the computation of both scalar fields , as given by equations ( [ ext . scalar field eq ] ) and ( [ kk scalar eq ] ) , involves only ( see the right hand sides ) quantities related to fields sources , and not to the test bodies .
hence , the effective @xmath85 given by relation ( [ g eff ] ) does not depend on the composition of the test bodies .
besides , let us emphasize that the equation of motion of a neutral point - like particle in the genuine kk theory reduces to the 4d geodesic equation after dimensional reduction @xcite , although the kk scalar field is coupled to the maxwell invariant @xmath86 . in a forthcoming paper @xcite , we address the effect of the varying effective coupling constants on the masses of composit particles . on account of the higgs mechanism of quarks and leptons masses generation , by promoting the yukawa coupling constants to effective parameters ( on an equal footing with @xmath2 or @xmath3 ) that depend on both scalar fields @xmath5 and @xmath4 , we prove that the kk@xmath5 model is actually consistent with the current experimental bounds on the violation of the equivalence principle . hence , we conclude that present laboratory experiments may not measure a true constant of gravitation . instead , in addition to all other possible biases ( _ e .
anelasticity in the wire of torsion pendulum as pointed out by kuroda @xcite , and which since has been generically taken into account in the experiments ) , they may be pointing out an effective one depending on the geomagnetic field at the laboratory position .
j. renner , in _ determination of gravity constant and measurement of certain fine gravity effects _ , y. d. boulanger and m. u. sagitov ( eds . ) , ( national aeronautics and space administration , washington , 1974 ) , pp .
26 - 31 .
g. mller _
et al . _ , ( 1990 ) , geophys . j. int .
* 101 * , 329 ; m. a. zumberge _
et al . _ , ( 1991 ) , phys . rev .
lett . * 67 * , 3051 ; m. oldham , f. j. lowes and r. j. edge , ( 1993 ) , geophys .
* 113 * , 83 ; j. k. hoskins _ et al . _ ,
( 1985 ) , phys . rev .
* d32 * , 3084 ; y. t. chen _
et al . _ , ( 1984 ) , proc .
a394 * , 47 ; r. d. rose _ et al .
_ , ( 1969 ) , phys .
* 23 * , 655 ; g. i. moore _
_ , ( 1988 ) , phys . rev .
* d38 * , 1023 ; f. d. stacey _ et al .
_ , ( 1987 ) , rev .
mod . phys . * 59 * , 157 ; x. yang _
_ , ( 1991 ) , chinese phys .
* 8 * , 329 ; m. s. saulnier and d. frisch , ( 1989 ) , am . j. phys . ,
* 57 * , 417 ; g. j. tuck _ et al .
_ , in _ the fith marcel grossmann meeting _ , d. g. blair and m. j. buckingham ( eds . ) , ( world scientific , singapore , 1989 ) , pp
. 1605 - 1612 ; k. kuroda and h. hirakawa , ( 1985 ) , phys .
* d32 * , 342 . | it is shown that 5d kaluza - klein theory stabilized by an external bulk scalar field may solve the discrepant laboratory g measurements .
this is achieved by an effective coupling between gravitation and the geomagnetic field .
experimental considerations are also addressed . |
understanding the steady state of a system in contact with two heat baths at unequal temperatures is a central question to the theory of non - equilibrium systems@xcite .
with very few exceptions ( such as zero range processes@xcite ) , non equilibrium systems exhibit long range correlations in their steady state@xcite .
these correlations have been studied theoretically by various approaches and measured experimentally . for a one dimensional diffusive system of size @xmath0 , for instance a stochastic lattice gas ,
these long range correlations take the following scaling form@xcite : @xmath1 ( where @xmath2 is an observable at position @xmath3 such as the density or the energy ) when the distances between the positions @xmath4 are macroscopic ( i.e. @xmath5 ) . the macroscopic fluctuation theory developed by bertini
et al.@xcite allows one to write down the general equations satisfied by the scaling functions @xmath6@xcite .
these partial differential equations are usually difficult to solve . the expressions of the @xmath6 are however known in a number of examples@xcite , where they have been obtained either from exact solutions of microscopic models or by integrating the partial differential equations derived from the macroscopic fluctuation theory .
diffusive systems are also known to satisfy fourier s law , meaning that , for large system sizes , the steady state flux @xmath7 of energy through a system of size @xmath0 scales like @xmath8 : @xmath9 where @xmath10 is a function of the temperatures @xmath11 and @xmath12 of the heat baths at the two extremities of the system , which maintain it out of equilibrium ; the function @xmath13 vanishes linearly with the difference @xmath14 so that the flux becomes a gradient . over the last 15 years@xcite it has been realized on the basis of numerical simulations that one - dimensional mechanical systems which conserve momentum do not satisfy fourier s law
instead , they exhibit a power law decay , called anomalous fourier s law , of the average flux @xmath15 with system size : @xmath16 so far , the exponent @xmath17 has not been determined analytically for any microscopic model with non - quadratic interactions ; however , numerical@xcite and analytic@xcite calculations , based on mode coupling theory or on other approaches@xcite , indicate that , depending on the type of the non - linearities and on the accuracy of the simulations , @xmath17 can take values ranging between @xmath18 and @xmath19 .
the main two systems for which this anomalous heat conduction has been observed are : * the fermi - pasta - ulam model , a chain of @xmath20 harmonic oscillators with an additional cubic ( fpu-@xmath17 ) or quartic ( fpu-@xmath21 ) interaction potential@xcite . for the latter case ,
most estimates indicate that @xmath22 ; * a one - dimensional gas of @xmath20 hard point particles with elastic collisions@xcite ( when the particles are identical , the collisions simply exchange the velocities of the incoming particles : hence the system has the same transport properties as an ideal gas , i.e. a ballistic transport with @xmath23 ) . for the collisions to be non - trivial ,
one choice is to consider particles of alternating masses 1 and @xmath24 ( 2-mass model ) . in this case
, it is commonly found that @xmath25@xcite . in comparison with diffusive systems , for which ( [ fourier ] ) holds
, the main feature of these models is that their dynamics conserve momentum , and it is believed that this property plays a key role in the divergence of the heat conductivity@xcite .
note , however , that there is no proof for this behavior : the argument of @xcite is not correct , although its conclusion is consistent with observations . in this article
, we consider a simpler model originally introduced in @xcite , the velocity reversal model , in order to better understand which features could be at the origin of the anomaly .
this model is , like the 1d hard - particle gas , a system of @xmath20 free - moving particles undergoing collisions .
the collisions are , however , not elastic , but instead are given by the following simple rule : * when two particles collide with velocities of opposite signs , @xmath26 , their two velocities get reversed ( @xmath27 , @xmath28 at the collision ) ; * when they collide with velocities of the same sign , the particles simply pass each other ( @xmath29 and @xmath30 if the particles are kept ordered from left to right ) .
in contrast with the fermi - pasta - ulam chain and the hard - particle gas , these dynamics do not conserve momentum .
it should be noted , however , that the absolute values of the velocities themselves are conserved , leading in general to more conserved quantities than in standard momentum - conserving models .
one should also note that , when the velocities are reversed at independent exponential times instead of at collisions , the system becomes diffusive@xcite . here , the @xmath20 particles of the velocity reversal model are in a one - dimensional box of length @xmath31 between two boundaries which play the role of heat baths . whenever a particle hits a boundary , its velocity is changed as if the particle was reinjected instantaneously , with its velocity thermalized by the reservoir .
the reservoirs are thus described by the velocity p.d.f . of these `` reinjected particles '' , @xmath32 and @xmath33 . an obvious choice is to consider maxwellian reservoirs at temperatures @xmath11 and @xmath12 , with velocity densities @xmath34 we will study below this `` maxwellian case '' for @xmath35 and @xmath36 . for a system which thermalizes well
, one would expect other choices of these reservoir velocity densities to only affect small regions near the boundaries , but not to influence the macroscopic behavior of the system .
for the velocity reversal model , however , the conservation of the absolute values of the particle velocities prevents thermalization from taking place : for instance , one can easily see that , for another choice of reservoirs , where particles are always reinjected with velocities @xmath37 from the left reservoir and @xmath38 for the right reservoir , @xmath39 the only possible velocities inside the system are @xmath40 and @xmath41 .
we will study this `` two - speed case '' with @xmath42 and @xmath43 .
in a first part , we show , by studying the steady - state current of systems with @xmath44 particles , that , both in the maxwellian and the two - speed case , the velocity - reversal model exhibits anomalous fourier s law , in contrast to the prediction of the associated boltzmann equation@xcite
. we then present , in a second part , measurements of the steady - state correlation functions of the momentum density for maxwellian and two - speed systems of @xmath45 particles , as well as for a 2-mass hard - particle gas with @xmath46 and @xmath47 for comparison : they indicate a size dependence very different from ( [ diff_corr ] ) .
the main advantage of the velocity reversal model described above , compared to the hard - particle gas for instance , is that its associated boltzmann equation can be solved analytically , thanks to the simpler form of its collision term . by solving this boltzmann equation in the steady state with the appropriate boundary conditions ,
the average current between the reservoirs can be predicted , as done in @xcite .
the simplest case is the two - speed system ( [ 2vitesses ] ) with velocities @xmath37 and @xmath38 , for which the boltzmann equation relates the four densities @xmath48 and @xmath49 of the particles with respective velocities @xmath40 and @xmath41 , with @xmath50 the space coordinate : @xmath51 for the velocity - reversal model with @xmath52 particles , in the maxwellian case with @xmath53 and @xmath36 ( above ) and in the two - speed case with @xmath42 and @xmath43 ( below ) . in both cases , the current decreases as a non - integer power of the system size : over the range of sizes we considered , our data are well - fitted by @xmath54 in the maxwellian case and @xmath55 in the two - speed case . ] and @xmath43 and for system sizes @xmath56 , @xmath57 and @xmath58 ( continuous lines ) . according to the boltzmann equation
, these profiles ( given by @xmath59 and @xmath60 respectively ) should be linear in @xmath61.,title="fig : " ] and @xmath43 and for system sizes @xmath56 , @xmath57 and @xmath58 ( continuous lines ) . according to the boltzmann equation
, these profiles ( given by @xmath59 and @xmath60 respectively ) should be linear in @xmath61.,title="fig : " ] + setting the time derivatives to @xmath62 , these equations can be solved@xcite to determine the steady state of the two - speed system : @xmath63 yielding linear profiles for the four densities as well as a steady - state current @xmath64 satisfying fourier s law([fourier ] ) . by a similar calculation
, the boltzmann equation also predicts that the energy current should also satisfy fourier s law in the maxwellian case .
we studied numerically the maxwellian ( [ gaussien ] ) and the two - speed ( [ 2vitesses ] ) velocity reversal models for @xmath44 particles , with @xmath35 , @xmath36 in the first case and @xmath42 , @xmath43 in the second .
because the true steady state is unknown , our measurements were performed by starting the system in an equilibrium configuration at @xmath65 in the maxwellian case , and by choosing the particle velocities uniformly at random among @xmath40 , @xmath41 in the two - speed case .
we then let the system evolve from that initial state , and sampled the time evolution of @xmath66 , @xmath67 and @xmath68 : @xmath66 was measured by a time average of the instantaneous energy flux , @xmath69 , while @xmath67 and @xmath68 were measured by counting the particles and energy within the boxes @xmath70 for @xmath71 .
we then took the long - time limits of these measured quantities to estimate their values in the steady state .
as shown in figure [ joel_cur ] , our data indicate that @xmath66 decreases like a power law of @xmath20 for both choices of reservoirs for the system sizes we considered .
they are consistent with ( [ anomalous ] ) , with @xmath72 for the maxwellian reservoirs and @xmath73 for the two - speed reservoirs . in the latter case , the density and energy profiles @xmath67 and @xmath68 shown in figure [ 2v_prof ] also differ noticeably from the linear profiles predicted by the boltzmann approach , at least for the sizes we were able to study .
the main assumption in the derivation of the boltzmann equation is that there are no correlations between the velocities of particles entering a collision in which their velocities are reversed . as it fails to predict the anomalous fourier s law of figure [ joel_cur ]
, it is interesting to investigate the form of the steady - state correlations for the velocity reversal model .
we measured numerically the steady - state two - point correlation functions of the momentum field , @xmath74 , for the velocity reversal model in the maxwellian ( fig .
[ g_cv ] ) and in the two - speed ( fig .
[ 2v_cv ] ) cases , for systems of @xmath75 to @xmath76 particles . as in the previous section , we approximated the momentum density function @xmath77 by its discretization over the boxes @xmath78 for @xmath79 .
we took the stationary - state correlations @xmath74 , which we measured for @xmath80 and @xmath81 , to be the long - time limits of their time evolutions for a system started in the arbitrary initial state described in the previous section . of the momentum field @xmath77 for the velocity reversal model in the maxwellian case ( [ gaussien ] ) with @xmath53 and @xmath82 , for @xmath80 ( left ) and @xmath81 ( right).,title="fig : " ] of the momentum field @xmath77 for the velocity reversal model in the maxwellian case ( [ gaussien ] ) with @xmath53 and @xmath82 , for @xmath80 ( left ) and @xmath81 ( right).,title="fig : " ] of the momentum field @xmath77 for the velocity reversal model in the two - speed case ( [ 2vitesses ] ) with @xmath83 and @xmath43 , for @xmath80 ( left ) and @xmath81 ( right).,title="fig : " ] of the momentum field @xmath77 for the velocity reversal model in the two - speed case ( [ 2vitesses ] ) with @xmath83 and @xmath43 , for @xmath80 ( left ) and @xmath81 ( right).,title="fig : " ] for diffusive systems , one would expect @xmath74 to follow a scaling of the form ( [ diff_corr ] ) : @xmath84 , except for @xmath85 with a variance @xmath86 of order 1 . for the velocity reversal models of figures [ g_cv ] and [ 2v_cv ] , we observed @xmath87 to be of order @xmath88 : however , when measuring @xmath74 for @xmath89 and @xmath90 , as in figures [ g_cv ] and [ 2v_cv ] ) , we observed an additional anomalous gowth in @xmath74 for @xmath61 close to @xmath91 in a region growing with system size , but becoming narrower on a macroscopic scale .
the sign of @xmath74 seems to be always negative for @xmath91 closer to the left ( warmer ) reservoir and positive for @xmath91 closer to the right , colder reservoir .
as shown in figures [ g_cv ] and [ 2v_cv ] , where our data are multiplied by the system size @xmath20 , the correlations seem to decay slower than @xmath92 as @xmath20 grows , over a region which seems to grow slower than the system size . in order to compare the behavior of these models with those of usual momentum - conserving models
, we also measured the two - point correlation function of the momentum for a hard - particle gas of particles of alternating masses @xmath88 and @xmath93 between maxwellian reservoirs at @xmath35 and @xmath94 : we observed a similar growth of the correlation functions at intermediate scales ( fig .
[ 2m_cv ] ) . for this hard - particle gas ( fig .
[ 2m_cv ] ) , due to a faster relaxation to the steady state , we were able to simulate much larger system sizes than for the velocity reversal model : at these sizes , @xmath95 , the anomalous growth of @xmath96 of figure [ 2v_cv ] is compatible with a scaling of the form @xmath97 as shown in figure [ scaling ] .
the sizes we could reach for the velocity - reversal model of figures [ g_cv ] and [ 2v_cv ] did not exhibit such a clear scaling form . for the velocity - reversal model as well as for the 2-mass hard particle gas , we also measured the correlations of the energy field @xmath68 , which exhibited a behavior very similar to the one shown above for the momentum field . of the momentum field @xmath77 for the 2-mass hard - particle gas with @xmath93 between reservoirs at @xmath53 and @xmath82 , for @xmath80 ( left ) and @xmath81 ( right).,title="fig : " ] of the momentum field @xmath77 for the 2-mass hard - particle gas with @xmath93 between reservoirs at @xmath53 and @xmath82 , for @xmath80 ( left ) and @xmath81 ( right).,title="fig : " ] , for @xmath80 ( left ) and @xmath81 ( right ) , for the 2-mass hard particle gas with @xmath47 between reservoirs at temperatures @xmath53 and @xmath82 .
, title="fig : " ] , for @xmath80 ( left ) and @xmath81 ( right ) , for the 2-mass hard particle gas with @xmath47 between reservoirs at temperatures @xmath53 and @xmath82 .
, title="fig : " ]
the numerical simulations presented in this work show that the velocity - reversal model exhibits an anomalous fourier s law , although it does not conserve momentum .
its density and energy profiles ( fig . [ 2v_prof ] ) differ noticeably from those predicted by the boltzmann equations . for the sizes we could achieve , however ,
they have not yet converged : this might question whether our data showing the anomalous fourier s law ( fig .
[ joel_cur ] ) correspond to the asymptotic regime . within the system
sizes we could reach , the two - point function of the momentum also looks very different from those of diffusive systems : they seem to be concentrated on a mesoscopic scale , much larger than the microscopic scale , but smaller than the system size . within this mesoscopic region
, the decay of the correlations seems to be anomalous as well , i.e. seems to be a non - integer power of the system size .
we have checked that the pair correlations seem to have a rather similar behavior for the 2-mass hard particle gas .
this leads us to believe that the shape of the correlation functions we have observed is probably another signature of the anomalous fourier s law : it would therefore be interesting to see whether these anomalous correlations are also present for other models known to exhibit an anomalous heat conductivity , such as the fermi - pasta - ulam chain .
looking at higher - order correlations would require major numerical efforts .
they would however be very interesting to measure in order to guess what would replace ( [ diff_corr ] ) for systems exhibiting anomalous fourier s law .
the present work is purely numerical in nature .
it would of course be very interesting to see what the existing theories explaining the anomalous fourier s law would predict for the pair correlation functions we have measured . for diffusive systems , the scaling ( [ diff_corr ] ) of the correlations with system size is closely linked to the scaling form of the large deviation function of the system s density profile@xcite : what such a scaling form would look like for systems exhibiting anomalous fourier s law is also an interesting open issue .
bertini , l. , de sole , a. , gabrielli , d. , jona - lasinio , g. , landim , c. : towards a nonequilibrium thermodynamics : a self - contained macroscopic description of driven diffusive systems .
. phys . * 135 * , 857872 ( 2009 ) | we study by means of numerical simulations the velocity reversal model , a one - dimensional mechanical model of heat transport introduced in 1985 by ianiro and lebowitz .
our numerical results indicate that this model , although it does not conserve momentum , exhibits an anomalous fourier s law similar to the one s previously observed in momentum - conserving models .
this is contrary to what is obtained from the solution of the boltzmann equation ( be ) for this system .
the pair correlation velocity field also looks very different from the correlations usually seen in diffusive systems , and shares some similarity with those of momentum - conserving heat transport models .
+ + keywords : current , anomalous fourier law , hard particle gas |
the spectrum of cosmic microwave background ( cmb ) photons has been measured very accurately and , at the present time , shows no deviations from a blackbody distribution .
these measurements provide tight constraints on any source of energy injection into the primordial medium at redshifts @xmath4 .
the relativistic motion of cosmic strings may continually inject energy into the cosmic medium starting from very early epochs , and certainly much earlier than @xmath5 . depending on the particle physics interactions between the string and the cosmic medium
, the energy deposited could be significant @xcite . however , there is also energy deposition due to the gravitational interactions of photons and strings , and this process is completely generic .
the gravitational scattering of photons by strings can result in gravitational lensing of background structure with angular separation @xmath6 where @xmath1 is newton s constant and @xmath2 is the string s lineal energy density .
however , the bending of light is an elastic process and can not lead to spectral distortions .
photon energies can only be re - distributed by cosmic strings at order @xmath7 and this is the effect we are interested in analyzing .
inelastic scattering by strings is similar to photon pair creation by the time - dependent metric of a cosmic string as discussed in @xcite . indeed
the feynman diagram is the same for both processes except for the reversal of the momentum in an external photon leg .
however , important technical differences arise due to the different kinematics that make the scattering calculation more difficult . to obtain the spectral distortions we also need to evaluate a boltzmann integral which makes the analysis challenging . in sec .
[ setup ] we introduce the formalism and evaluate the scattering amplitude for a general cosmic string loop . to obtain numerical results ,
we specialize to a simple loop trajectory , one that is expected to be close to the most generic loop @xcite , and is called a `` degenerate , kinky loop '' . in sec .
[ kinkyloop ] we numerically evaluate the photon scattering rate in phase space after averaging over all possible orientations and shapes of the degenerate kinky loop . the spectral distortion is shown in fig . [ boltzmannintegral_plot ] and the result is used to estimate the fractional spectral distortions of the cmb in sec . [ conclusions ] . in appendix [ nintegratediscussion ]
we discuss some technical details of the evaluation of the boltzmann integrals .
we will consider the process of photons scattering off a cosmic string taking place in a weakly curved spacetime , @xmath8,\end{aligned}\ ] ] the scattering process is governed by the action @xmath9 where @xmath10 is the action of the photon @xmath11 @xmath12 with @xmath13 ; @xmath14 is the quadratic action of the graviton @xmath15 in de donder gauge @xmath16 , @xmath17 with @xmath18 , and @xmath19 is the nambu - goto action for the cosmic string @xmath20 here , our cosmic string is of length @xmath21 with tension @xmath2 , and @xmath22 is the 4 coordinate vector of a given point @xmath23 on its worldsheet .
the determinant of the induced metric of the latter is @xmath24 ; denoting @xmath25 we have @xmath26 the stress - energies of the photons and the string are , respectively , @xmath27 and @xmath28 for @xmath29 , we have already chosen set of coordinates such that the induced metric on the string is conformally flat and @xmath30 . in the same gauge ,
the solution to the nambu - goto equations of motion from eq . ,
@xmath31 , can always be expressed as a superposition of null left- and right - movers , @xmath32 + x^\mu_-[\sigma_- ] \right)\end{aligned}\ ] ] and @xmath33 where from henceforth we shall switch to using the variables @xmath34 and their associated derivatives @xmath35 instead of @xmath23 . moreover it is advantageous to exploit the periodicity of the trajectory , i.e. @xmath36 = x^\mu[t,\sigma]$ ] , to fourier decompose the stress energy of the cosmic string , eq .
, into the form @xmath37 & = -\frac{i\mu l}{4 { m_\text{pl } } } \sum_{\ell = -\infty}^\infty c_{(+|\ell)}^{\{\mu } c_{(-|\ell)}^{\nu\ } } \nonumber\\ & \qquad \qquad \times ( 2\pi ) \delta\left ( k_0 - \frac{4\pi\ell}{l } \right ) .\end{aligned}\ ] ] ( we denote @xmath38 . ) the @xmath39s contain information about the trajectory of the cosmic string loop and are given by @xmath40 \right ] \partial_\pm x_\pm^\alpha , \end{aligned}\ ] ] where @xmath41 for the scattering problem at hand , @xmath42 , so that the @xmath43s become their complex conjugate under the interchange of @xmath44 : @xmath45 = \overline{c_{(\pm|\ell)}}[p',p]$ ] .
( the overbar denotes complex conjugation . ) for the pair production process described in @xcite , @xmath46 .
let us denote the polarization vector of the incoming photon as @xmath47 $ ] and that of the outgoing photon as @xmath48 $ ] .
in the lorenz gauge @xmath49 , which we shall use here , they satisfy @xmath50 . the stress - energy operator of the photons in eq .
, wick - contracted with the state @xmath51 representing one incoming photon of momentum @xmath52 and with the state @xmath53 representing one outgoing photon of momentum @xmath54 is @xmath55 & = -\bigg ( p^{\{\alpha } p'^{\beta\ } }
\ \epsilon \cdot \bar{\epsilon } ' + p \cdot p ' \ \epsilon^{\{\alpha } \bar{\epsilon}'^{\beta\ } } \\ & \qquad \qquad - p^{\{\alpha } \bar{\epsilon}'^{\beta\ } } \epsilon \cdot p ' - p'^{\{\beta } \epsilon^{\alpha\ } } \bar{\epsilon } ' \cdot p \nonumber\\ & \qquad \qquad - \eta^{\alpha\beta } \left ( p \cdot p ' \
\epsilon \cdot \bar{\epsilon } ' - p \cdot \bar{\epsilon } ' \ p ' \cdot \epsilon \right ) \bigg ) .
\nonumber\end{aligned}\ ] ] the propagator of the graviton that follow from eq .
is @xmath56 at this point , the quantum amplitude of a single photon with initial momentum @xmath52 , scattering off the cosmic string , and acquiring a final momentum @xmath54 is then given by @xmath57 ( because the classical stress - energy tensor of photons is traceless in 4 dimensions , one would find that @xmath58 . )
this amplitude has a diagrammatic representation shown in fig .
[ scatteringdiagram ] .
we proceed to calculate @xmath59 , summed over all initial and final photon polarizations . as a consequence of ward identities obeyed by electromagnetism ,
this amounts to the replacements @xmath60 . from eq .
, we will also face the product @xmath61 and their associated sums over all integers @xmath62 and @xmath63 .
this product is zero unless the arguments of the two @xmath64-functions are simultaneously equal to zero ; this collapses the double sum into one , while the product of @xmath64 functions become @xmath65 define the amplitude - squared per unit time as @xmath66 where @xmath67 suppressing the @xmath62 index on the @xmath68s , we arrive at @xmath69 with @xmath70 where @xmath71 is the same parameter occuring in .
we display this dependence of @xmath72 on @xmath71 so as to compare against the pair production result in eq .
( 26 ) of @xcite .
the primary differences going from the scattering to pair production amplitude are that , @xmath71 changes sign from @xmath73 to @xmath74 , and the @xmath75 in the @xmath68 is replaced with @xmath76 . for the pair production case , since @xmath77 , the sum over @xmath62 only runs over @xmath78 and the positive integers ; whereas for the scattering case , @xmath62 runs over all integers and zero . whichever @xmath79 is being used
, @xmath80 always holds ; we have used this to arrive at eq . .
with the scattering amplitude per unit time , we may insert it into the generic form of the boltzmann equation for the scattering of bosons ( see , for example , chapter 5 of @xcite ) , @xmath81 where @xmath82 is the number of photon scatterings off the cosmic string in question , per unit time , per unit momentum space ; and @xmath83 is the number of background photons per unit volume per unit momentum space .
the first term describes the rate of photons entering the momentum bin @xmath84 due to their scattering from other momenta @xmath85 ; the second term is the rate of leaving the momentum bin @xmath84 due to the same scattering mechanism .
one may check directly using eq . that @xmath86 is invariant under the swap @xmath44 : @xmath87 is invariant , whereas as noted already , @xmath88
= \overline{c_\pm}[p , p']$ ] , so @xmath89 is replaced with @xmath90 ( and @xmath91 with @xmath92 ) , where we have used @xmath93 .
since every product of the @xmath89s and @xmath91s contains even powers of @xmath54 , the minus sign can be dropped ; hence , the swap @xmath44 amounts to the replacements @xmath94 in every term of @xmath86 .
the conclusion then follows if we observe that , the terms in @xmath86 are either real or comes in pairs whose sum is real .
this invariance @xmath95 means the quadratic - in-@xmath96 terms in eq .
cancel .
furthermore , if we ignore the backreaction of this scattering process on the background distribution of photons , so that @xmath83 can be assumed to be time independent ; then @xmath97 , the bose - einstein distribution for a massless spin-1 particle . for simplicity , let us also integrate over all possible final directions of the photon .
altogether , we are now lead to @xmath98 to be clear , our notation for the incoming @xmath52 and outgoing @xmath54 momenta is @xmath99 so that , for instance @xmath100 is the integral over the solid angle of the incoming photon .
note that , because scattering does not change the total number of photons present , integrating the right hand side of eq .
with respect to @xmath101 ( from @xmath78 to @xmath102 ) would yield zero .
we will later use this as a check of our numerical calculations . to summarize what we have found so far : for a given cosmic string loop trajectory ( and the relevant left- and right - movers @xmath103 ) , first determine the modes @xmath68s in eq .
, and afterwards insert them into eq . to obtain @xmath86 . then the rate of photon scattering off the cosmic string per unit energy , is obtained by evaluating the boltzmann integral in eq . .
in @xcite , an analysis was performed for the gravitationally induced pair production rate of photons without specializing to a particular cosmic string loop trajectory .
this is possible because , for the pair production case , there exists asymptotic expansions of eq . by utilizing the large dimensionless phase @xmath104 .
( in other words , one can exploit the large ratio of the astrophysical length of the string to the microscopic wavelength of the photon . ) for instance , one may seek a saddle point of the @xmath105 , i.e. search for configurations @xmath106 where @xmath107 . because @xmath108 and @xmath109 are null vectors , that means @xmath110 must be spacelike or null for a saddle point to exist .
that in turn implies @xmath85 and @xmath84 are collinear , since we require @xmath111 .
a slight deviation ( in , say , the direction vectors @xmath112 and @xmath113 ) from this configuration would fail to satisfy the stationary phase condition and therefore lead to a subdominant pair production rate ; this in turn yields estimates on how very narrow regions ( whose widths scale as some inverse power of @xmath62 ) in angular space provide the dominant contribution to the @xmath105 integrals . in all , what @xcite found was that the energy emitted due to the pair of photons was largely independent of the mode number @xmath62 ; in other words , equal power in photon pairs were emitted at each harmonic @xmath114 .
for the scattering case at hand , such an asymptotic analysis appears to be more difficult .
for example , the stationary phase condition now becomes @xmath115 ; but with the minus sign instead of a plus sign ( as was the case for pair production ) , the requirement that @xmath116 is spacelike or null yields no useful constraint on the configuration @xmath106 .
another important difference is that , for the pair production process , because @xmath117 , the rate at a given @xmath62 , without performing the analogous infinite sum in eq . , is already a physical quantity of interest ; it is the power generated at a given energy @xmath118 . for the scattering process ,
on the other hand , we wish to ask how many net photons are being added or subtracted , at a given @xmath101 , from the background thermal bath .
this means , for a fixed @xmath101 we have to perform both the solid angle integrals and the infinite sum over @xmath62 in eq . in order to obtain a physical quantity
however , at least for the two photon temperatures we have considered , we shall find that the dominant contribution to the infinite sum come from the small @xmath119 terms , specifically @xmath120 .
so as to achieve a concrete estimate of scattering rates of photons off cosmic strings , we shall therefore , in this section , specialize to the `` degenerate '' kinky loop trajectory that was also used in @xcite and @xcite .
specifically , the left- and right - movers of the kinky loop are @xmath121 where @xmath122 and @xmath123 are constant unit vectors .
now , denoting @xmath124 we have , via eq .
, @xmath125 we first integrate over @xmath126 in eq . , using the @xmath64-function to set @xmath127 in the integrand . because @xmath128 , we also need to introduce a step function @xmath129 multiplying the result after integration .
next , for convenience , we set @xmath130 and rescale @xmath131 followed by the definitions @xmath132 with the rescaling in eq . , the length @xmath133 , as @xmath134 is varied from @xmath135 to @xmath74 , ranges over @xmath136 _ collinear limit _ the reader may worry that the presence of @xmath137 in the denominator of eq . indicates the differential scattering rate may be singular in the collinear limit , as @xmath138 .
this is not the case , but rather , @xmath139 the third line of eq .
contains all the remaining dependence on angles and is finite for all @xmath122 , @xmath123 , and @xmath113 , because @xmath62 is an integer and @xmath140 is finite at @xmath141 .
_ @xmath142 can be dropped _
we also note that the @xmath142 term in eq .
can in fact be dropped because , in this limit , the bose - einstein factors on the first line of eq .
goes to zero while the rest of the integrand is finite .
explicitly , the bose - einstein factors , when @xmath143 , can be expanded as @xmath144 ( for later use , we have expanded up to @xmath145 . )
this leads us to the following expansion of the small @xmath62 terms of eq .
, @xmath146 for @xmath147 , this expression is non - singular for all @xmath101 , and therefore zero when the limit @xmath148 is taken .
we also remark that , in the preceding discussion on the collinear limit , if we took @xmath138 followed by @xmath149 , we would find the boltzmann integrand in eq
. would read @xmath150 in other words , the two limits @xmath138 and @xmath148 commute and yield a zero integrand .
_ average over loop configurations _ if we were able to integrate over the solid angles with respect to the incoming and outgoing photon directions ( @xmath112 and @xmath113 ) , the resulting expression must only depend on @xmath122 and @xmath123 through the euclidean dot product @xmath151 , since the scattering rate is a scalar and there are no other vectors in the current problem .
the result would allow one to see how the scattering rate per unit energy @xmath101 depends on the kinky loop configuration .
however , in this paper , we shall be content with obtaining an estimate of the rate of scattering averaged over all possible kinky loop configurations if we assume a uniform distribution , this means we shall now examine @xmath152 notice the boltzmann integrand in eq .
has factorized into three portions : the first line is independent of @xmath122 and @xmath123 ; the second depends solely on @xmath122 and the third solely on @xmath123 .
this means the averaging in eq .
also factorizes into two independent solid angle integrals , one with respect to @xmath122 and the other with respect to @xmath123 .
in fact , the primary integrals we need are @xmath153 by a flip of parity , @xmath154 or @xmath155 , we see that these two integrals are actually the same , @xmath156 .
therefore let us focus on the first .
observe that the only vector in the integrand is @xmath157 , and the integral transforms as a rank 2 tensor under o@xmath158 transformations : for @xmath159 , where @xmath160 o@xmath158 , we have @xmath161 .
therefore we may fix the form of @xmath162 to be @xmath163 where @xmath164 are o@xmath158 scalars .
contracting both sides with @xmath165 and with @xmath166 allow us to solve for @xmath167 and @xmath168 in terms of the two scalar integrals @xmath169 where @xmath170 .
the results are : @xmath171 and @xmath172 the si and ci are the sine and cosine integral respectively . whereas the @xmath173 minus ci terms may also be re - expressed in terms of cin and the euler - mascheroni constant @xmath174 ( see @xcite ) , @xmath175 which is an even power series
this is why we may put an absolute value around the arguments of @xmath173 and ci . with the @xmath176 in eq . ,
and recalling @xmath177 , @xmath178 , the boltzmann integral in eq . averaged over loop configurations now becomes @xmath179 where @xmath180 . even though @xmath112 and @xmath113 appear in the integrand , after all the euclidean dot products are carried out , the integrand in eq .
will only depend on them via the combination @xmath181 . again using o@xmath158 invariance ,
we have reduced the @xmath182 integration to @xmath183 times a one dimensional integral with respect to @xmath184 , after the replacement @xmath185 . after this integral with respect to @xmath184 there will not be any dependence on @xmath112 nor @xmath113 , and therefore we have replaced the solid angle integral with respect to @xmath113 with @xmath186 . _
zero @xmath101 limit _ taking the zero outgoing energy limit , @xmath187 , of eq .
( [ boltzmannintegral_kinkyloop_ii ] ) we get @xmath188 where @xmath189 in this formula denotes the original ( un - rescaled ) photon temperature ( see eq .
( [ rescaling ] ) ) , and we have reinstated @xmath21 .
this non - zero scattering rate indicates that , with a sufficiently high cosmic string density , there could be a significant backreaction on the low energy photon distribution . to get to eq .
we have set @xmath190 , and @xmath191 .
in particular , starting from eq .
, as @xmath187 , @xmath192 as we shall soon see , the rate of scattering ( away from @xmath193 ) will continue to be negative for low energy photons , before becoming positive at higher energies .
the motion of cosmic strings kick photons from lower to higher energies .
_ change - of - variables _ we will now evaluate eq .
numerically for @xmath194 and @xmath195 . for a photon temperature of @xmath196
k , this corresponds to loop sizes of @xmath197 kpc and @xmath198 kpc respectively .
( for comparison , the diameter of the milky way is 30 kpc . )
let us change our variables from @xmath184 to @xmath199 , i.e. @xmath200 to obtain @xmath201 let us understand qualitatively the integrand in eq .
, over the range of integration , @xmath202 .
cin@xmath203 is an even function of @xmath3 , starting from zero at @xmath204 and rising , for large @xmath205 , to go as @xmath206 ( see eq . ) .
consequently , the @xmath207 in eq . is an even function of @xmath62 , peaks at @xmath208 and proceeds to fall off as @xmath209 for @xmath210 .
hence , the second line of eq . grows at most as a cubic polynomial in @xmath211 for @xmath210 . whereas si@xmath203 is an odd function that rises from zero at @xmath204 and asymptotes to @xmath212 for large @xmath213 ; and thus the @xmath214 term asymptotes to @xmath215 for @xmath210 .
we see that the third line of eq .
( inside the @xmath216 ) , grows as a quadratic polynomial multiplied by @xmath217 , which decreases as @xmath218 from below . with the @xmath219 ( on the first line in eq . ) multiplying the entire squared quantity , we therefore expect the integrand to peak close to the lower end of the integration @xmath220 and fall off with increasing @xmath211 .
moreover , for fixed @xmath101 , large @xmath221 terms ought to be suppressed relative to their smaller @xmath222 counterparts .
these observations will aid us in the ensuing numerical integration in appendix [ nintegratediscussion ] we discuss some of the technical issues we face .
we remind the reader that the rescaling in eq .
is in force from equations through , except for eq . .
) described in eq .
, per unit time , per unit energy .
there are actually two curves here , one at temperature @xmath223 and the other at @xmath224 ; because for @xmath225 the rate scales as @xmath226 , they coincide.,width=336 ] in fig . [ boltzmannintegral_plot ] we present the results for eq . , when @xmath227 and @xmath228 , with respective maximum scattering rates per energy ( roughly ) @xmath229 and @xmath230 . that the higher temperature yields a larger rate is what one would expect on physical grounds , since there are more photons available per unit volume at higher temperatures .
( photon number density is proportional to the cube of its temperature . ) since temperature enters the current problem only through the bose - einstein factors in eq . , and since it is these bose - einstein terms that are responsible for the existence of positive and negative scattering regions , we see that the energy scale over which this transition from negative to positive rate takes place is itself of the @xmath231 .
we have calculated the amplitude for inelastic scattering of photons off a cosmic string loop .
we have used the result to calculate the spectral distortions from a loop of string placed in a thermal background of photons .
the shape of the spectral distortion is shown in fig .
[ boltzmannintegral_plot ] .
it is of interest to estimate the fraction of photons that will be inelastically scattered by a distribution of cosmic string loops .
first we note that the scattering rate shown in fig .
[ boltzmannintegral_plot ] , can be estimated as ~(g_)^2 ( l t_)^2 , ~t _ , where in this section we are using @xmath232 to denote the frequency of the photon . moreover , here we are only concerned with the scattering when @xmath233 because this spectral region contains the most number of photons .
the important scaling as @xmath234 has been verified numerically by comparing the two plots in fig .
[ boltzmannintegral_plot ] ; it can also be checked by using rescaling arguments on eq . .
now we can multiply the scattering rate by the loop number density to get the number of scattered photons per unit time per unit volume .
since the scattering rate is proportional to @xmath234 , the largest loops will be most important for this process .
such loops have size comparable to the horizon and so we take @xmath235 , whereas the loop number density is approximately @xmath236 , where @xmath237 is the cosmic time
. then we can integrate over time to get the total change in number density of photons @xmath238 per frequency interval , @xmath239 where @xmath240 is the initial temperature from which we start counting the number of scattered photons .
( we have used the fact that the photon temperature scales inversely as the cosmological scale factor , @xmath241 , assumed @xmath242 goes as some power law , and the initial scale factor is much smaller than its present value @xmath243 . ) to get the fractional change in photon number density at some frequency , @xmath244 evaluated at the present epoch , due to scattering by strings , we divide eq . by the number density of ambient photons at the present time with frequency @xmath245 ,
i.e. @xmath246 . then
~(g_)^2 z^2 where @xmath3 is the redshift after which the scattered photons do not get re - thermalized by compton scattering . we will not go into a detailed calculation here but simply mention that @xmath3 is definitely higher than the recombination redshift ( @xmath247 ) and possibly @xmath248 @xcite . with @xmath249 and using @xmath250 ,
the spectral distortion from inelastic scattering of photons by the gravitational field of strings is at a level @xmath251 .
the distortions are similar to the @xmath252distortions of the cmb in that the total number of photons is unchanged .
however , the shape is different and almost completely resides in the low frequency part of the spectrum .
as in the case of gravitational wave emission @xcite , cosmic strings produce a burst signal in addition to a stochastic signal .
it is possible that cusps and kinks on strings can up - scatter ambient photons in a burst , and that this may provide another observational signature .
we leave the exploration of this possible signature for future work .
we thank levon pogosian and hiroyuki tashiro for comments .
yzc was supported by funds from the university of pennsylvania ; and by the doe while at arizona state university where this work was initiated .
tv was supported by the doe at asu .
some of this work was done while the authors were at the institute for advanced study , princeton , which we thank for its hospitality .
much of the numerical and analytic work in this paper was done with mathematica @xcite .
in this appendix we describe some of the technicalities involved in the numerical evaluation of the boltzmann integral in eq . .
to begin , let us re - express eq . in the following schematic form
: @xmath253 with @xmath254 one may attempt to use the nintegrate function in mathematica to calculate numerically the @xmath255-integral portion of eq .
, for a given @xmath101 , over as many @xmath62s as possible , before summing them up .
one would find that the @xmath256 terms are the dominant ones , but these same terms are such that , the pair @xmath257 yields @xmath255 integrals of nearly equal magnitude but opposite signs .
it is therefore prudent to combine the positive and negative @xmath62 terms , for @xmath258 , before integrating , i.e. @xmath259 where @xmath260 notice we have also changed the summation variable such that now @xmath62 is summed over all non - zero positive integers .
for @xmath261 , it is often useful to expand the bose - einstein factors using the small @xmath62 limit in eq . before applying nintegrate .
specifically , by decomposing the integrand @xmath262 into @xmath62-even plus @xmath62-odd terms , and denoting @xmath263 , we have @xmath264 of course , there are only two distinct integrals here , namely @xmath265 ; the temperature dependent pieces of @xmath261 do not take part in the integration . what we are emphasizing is the cancellations that occur between the @xmath62-odd and @xmath62-even portions of @xmath262 when summing over the originally positive and negative @xmath62 terms , at each order in the @xmath266 expansion .
for all the cases where we used this expansion in eq . , we have found that the sum of the third and fourth lines is always much less than the second line usually much less than a percent and therefore may be neglected . for each @xmath267
, we evaluate @xmath268 for @xmath269 and sum them up .
for @xmath270 , we perform the integrals for ten @xmath62s per decade , for example , @xmath271 and @xmath272 , up to a few orders of magnitude beyond that of @xmath273 .
( for small @xmath274 , due to numerical underflow we do not integrate beyond @xmath275 but we simply stop at some @xmath62 that yields @xmath276 that are many tens of orders of magnitude below @xmath277 . )
then we use the mathematica function interpolation to compute a best fit line to the resulting numerical data and proceed to approximate the summation for @xmath270 by an integral over the fit range , @xmath278 .
we find that , for each @xmath101 , the sum over the first ten @xmath279 is usually responsible for roughly 91 percent of @xmath280 .
when numerically evaluating @xmath276 we have found it useful to not do it over the entire range of @xmath211 , but rather start from integrating from the lower limit @xmath281 to @xmath282 $ ] , where @xmath283 is the upper limit ( for instance , @xmath284 and @xmath285 for @xmath261 ) , then from @xmath281 to @xmath286 $ ] , etc .
, until the error incurred is less than one percent .
this algorithm is justified because , as already explained after eq . , we expect our integrand to always peak near @xmath220 .
practically speaking , this saves mathematica considerable amount of time and effort when sampling the integrand . as a check of our calculation
, we fit a line through the numerical data points of our plots in fig .
[ boltzmannintegral_plot ] and find that the area above the @xmath101-axis is equal to the area below the @xmath101-axis at the sub - percent level .
this is both consistent with our demand for percent - level accuracy in the numerical integrations of @xmath276 , as well as the physical requirement that scattering processes do not change the net photon number .
h. tashiro , e. sabancilar and t. vachaspati , phys .
d * 85 * , 103522 ( 2012 ) [ arxiv:1202.2474 [ astro-ph.co ] ] .
d. a. steer and t. vachaspati , phys .
d * 83 * , 043528 ( 2011 ) [ arxiv:1012.1998 [ hep - th ] ] .
c. j. copi and t. vachaspati , phys .
d * 83 * , 023529 ( 2011 ) [ arxiv:1010.4030 [ hep - th ] ] .
k. jones - smith , h. mathur and t. vachaspati , phys .
d * 81 * , 043503 ( 2010 ) [ arxiv:0911.0682 [ hep - th ] ] .
y. -z .
chu , h. mathur and t. vachaspati , phys .
d * 82 * , 063515 ( 2010 ) [ arxiv:1003.0674 [ hep - th ] ] .
d. binosi and l. theussl , comput .
commun . * 161 * , 76 ( 2004 ) [ hep - ph/0309015 ] .
d. binosi , j. collins , c. kaufhold and l. theussl , comput .
commun . * 180 * , 1709 ( 2009 ) [ arxiv:0811.4113 [ hep - ph ] ] .
e. kolb , m. turner , `` the early universe , '' westview press ( 1994 ) | photons can gravitationally scatter off a cosmic string loop and gain or lose energy .
we consider the spectral distortion induced by cosmic string loops placed in an ambient thermal bath of photons .
the fractional deviation from a thermal spectrum caused by cosmic strings is estimated to scale as @xmath0 , where @xmath1 is newton s constant , @xmath2 is the string tension , and @xmath3 is the cosmological redshift after which spectral distortions can survive .
this effect is large enough to potentially be of observational interest . |
rapid variability is a tell - tale sign of accretion .
the first coherent oscillation in a cataclysmic variable binary was reported by walker as far back as 1956 ( walker 1956 ) .
the strictly periodic 71s brightness oscillations he found in dq her are now understood as being produced by a spinning magnetic white dwarf that is being fed by an accretion disc , and accretes the material along its magnetic field lines . since the dynamical time scales of matter close to the white dwarf are of the order of tens of seconds , it was not until the development of high speed photon - counting photometers in the 70s that periodic signals were discovered in other systems .
the dwarf nova oscillations ( dnos ) are periodic brightness modulations discovered in dwarf nova systems during outburst ( warner & robinson 1972 ) .
the short time scales ( 2 - 40s ) indicate an origin close to the white dwarf . in some systems ,
different periods were found at different times , and the period stability is considerably lower than the coherent oscillations in dq her .
see warner ( 1995 ) for review .
whereas the amplitudes in the optical are usually so low ( @xmath2 ) that the oscillations are only revealed after period analysis , marsh & horne ( 1998 ) detected coherent oscillations at two periods in the hst light curves of oy car with an amplitude of up to 20% towards the end of a super - outburst .
reminiscent of the dnos are the flood of qpos reported in x - ray observations of accreting neutron star binaries ( see van der klis 2000 for review ) .
the properties of these oscillations are in many ways very similar to those of dnos , albeit on shorter time scales since these systems contain a much smaller compact object .
it shows that moderately coherent oscillations appear to be a generic feature of accretion onto compact objects .
v2051 ophiuchi is a short period eclipsing dwarf nova system belonging to the su uma sub - class .
its status as a dwarf nova was contested ( e.g. warner & odonoghue 1987 ) until two super - outbursts were observed in 1998 and 1999 by amateur variable star networks .
spectroscopy reveals double - peaked emission lines that exhibit a rotational disturbance effect during eclipse , corresponding to a pro - grade rotating accretion disc ( cook & brunt 1983 , watts et al .
baptista et al . ( 1998 ) present a photometric model for the binary based on hst and ground based observations . due to the unusual low - state the system was in during their observations ,
the orbital phases when bright spot and white dwarf are eclipsed by the mass donor star could be directly measured .
warner & cropper ( 1983 ) concluded that most of the rapid variability ( flickering ) was produced in the inner disc regions rather than the bright spot or white dwarf .
the only quasi periodic oscillation in this system was reported by warner & odonoghue ( 1987 ) , who detected a very short lived 42s oscillation during outburst .
it was not classified as a dno because of its low coherence , and the absence of a corresponding spike in the power spectrum at that period .
we present high time resolution spectroscopy of the dwarf nova v2051 oph shortly after a normal outburst , obtained using our own data collection hardware equipped on keck ii , and report on the discovery of emission line oscillations at the level of several percent . in the next section we will describe the data acquisition system and novel reduction steps that were used .
section 3 covers the analysis of the data and discusses the properties of the oscillations .
we discuss likely models in section 4 and summarise our conclusions in section 5 .
the observations were obtained as part of a 5 night campaign to study rapid variability of cataclysmic variables and x - ray binaries ( obrien 2000 ) .
the optical data were taken using the low resolution imaging spectrograph ( lris ; oke et al .
1995 ) on the 10-m keck ii telescope on mauna kea , hawaii between ut 7:22 and 7:57 on july 4th 1998 .
the lris was used with a 5.2 arc - second slit masked with aluminized mylar tape to form a square aperture .
the 300/5000 grating used has a mean dispersion of 2.55 / pixel in the range 3600- 9200 .
we used a novel data acquisition system to obtain almost 28,000 2048 pixel spectra of v2051 oph , in the form of a continuous data stream .
each spectrum has a mean integration time of 72.075 msecs and there was no dead - time between individual spectra .
each 2048 pixel spectrum also contains a 25 pixel underscan region and a 75 pixel overscan region .
the noise for a given pixel was calculated using a readout noise of 6.3 e@xmath3 and a gain of 4.7 e@xmath3/adu .
cosmic rays were rejected with a threshold of 7-sigma from the de - biased frames .
a master flat - field image was created by finding the median of 700 individual flat - field frames .
this image showed no deviations above 0.3 % in all but 3 pixels .
we decided that it was therefore not necessary to flat - field individual spectra .
calibration arc spectra were taken through the same slit aperture .
bacground spectra , including light from the sky but dominated by a constant electronic background , were taken at the beginning and end of target observing , so that long time scale variations could be detected .
the mean and variable components of the background spectrum were found by creating a light curve for each pixel and extracting the mean and gradient of this light curve .
these coefficients were then filtered in wavelength , with a running median filter of width 101 pixels .
the gradient of the data was found to be almost zero .
the thus derived background spectrum , which accounts for @xmath4 0.9% of the total flux , was subtracted from each target spectrum .
the arc calibration was done by fitting a second order polynomial to 7 lines in a median spectra of exposures of hg and ar lamps .
arc spectra from the beginning and end of the exposures were used to take into account any drifts in the wavelength scale .
the interpolated wavelength calibration was applied to each target spectrum using the molly spectral analysis package .
the spectrophotometric standard star , feige 67 ( oke 1990 ) was observed using the identical setup .
a low order polynomial fit was found to the median of all the individual flux star spectra and the individual target spectra were flux calibrated using molly .
individual time marks , accurate to 200ms , were placed after every other spectrum using the computer clock . in order to find the absolute times , secondary timestamps at known ut
were placed at regular intervals throughout the observations .
a calibrated ut radio clock was used for this purpose . with the help of these secondary timestamps , the times of each individual spectrum were calibrated , and we were thus able to achieve an absolute timing accuracy of close to 200ms . to monitor seeing conditions and telescope tracking , the auto - guider ccd output was stored .
we used this output to measure the position and fwhm of the guide star image every @xmath42s during our observations .
gaussian fits to the stellar images reveal very stable tracking and seeing conditions throughout our observations .
a conservative estimate of the seeing was derived from the mean fwhm of the guide star images .
this suggests that seeing was better than 0.8 arcseconds , equivalent to a seeing limited fwhm resolution of our spectra of 3.7 ccd pixels .
occasionally , not all pixels of a spectrum were successfully transferred to the acquisition pc .
these incomplete spectra were rejected from the data stream . in order to increase the signal to noise of the remaining spectra and make the data - set more manageable
, individual spectra were binned together in groups of 7 to give a final 3968 spectra with a time resolution of 0.5045 seconds .
since the dynamical timescale close to the white dwarf is of the order of a few seconds , this time resolution is still ample for our purposes .
these spectra cover 40% of the binary orbit including an eclipse .
the linear ephemeris of baptista et al .
( 1998 ) was used to calculate orbital phases throughout this paper .
phase 0.0 corresponds to inferior conjunction of the companion star i.e. mid - eclipse .
figure [ trail ] shows a trailed spectrogram plot of the extracted spectra obtained covering the orbital phases between 0.69 and 1.08 .
the two gaps near phase 0.720 and 0.945 were introduced as the telescope was slewed to sky to allow robust sky subtraction as discussed above .
double - peaked emission lines are clearly visible , with the blue - shifted side of the lines eclipsing first , followed by the red - shifted peak , indicating the presence of a pro - grade rotating accretion disc . a brightening of the object to v=13.2 was reported by the vsnet network on july the 2nd ( vsnet alert # 1934 ) .
the usual quiescent brightness level of this dwarf nova is around v=15.7 .
this suggests a normal outburst was in progress in early july .
unfortunately , no measurements were available for the preceding week that could confirm the exact start of the outburst . at the time of our observations , a v magnitude of 14.8 was reported , and 10 days after the observations the object was back to a quiescent level at v=15.7 .
it appears , therefore , we were observing v2051 oph towards the later phase of decline from normal outburst .
figure [ specs ] ( top ) illustrates the average out of eclipse spectrum obtained by averaging all spectra between orbital phases 0.69 and 0.92 .
strong broad disc emission lines are superposed on the flat continuum .
the rapid drop in the spectral efficiency of the lris setup below 4000 made it difficult to achieve robust flux calibration for bluer wavelengths .
the balmer jump appears to be in emission , however care must be taken with the flux scale in this region .
the balmer lines have a full width half maximum of 2145 @xmath5 20 km s@xmath6 , and the balmer decrement from h@xmath7 to @xmath8 is 1:0.65:0.50:0.41 , suggesting a mix of optically thick and thin emission .
the helium i lines are clearly double peaked and feature a slightly higher fwhm of 2245 @xmath5 30 km s@xmath6 .
the blue peak is consistently stronger during this phase interval .
curiously , only h@xmath7 exhibits a slightly stronger red peak .
this blue asymmetry is also present in the average spectrum of watts et al .
( 1986 ) that covers h@xmath9 and h@xmath10 , but not h@xmath7 .
the h@xmath7 profile in warner & odonoghue ( 1987 ) again shows a stronger red peak .
this curious reversal of the double peak asymmetry seems to be a persistent feature of v2051 oph .
the line wings extend out to at least 3500 @xmath5 200 km s@xmath6 for both the balmer and helium lines .
double gaussians were fitted to the 6 strongest balmer lines simultaneously .
the double - peak separation was measured to be @xmath11 km s@xmath6 , reflecting the radial velocities ( @xmath12 km s@xmath6 ) of the gas on the outer edge of the accretion disc . if the velocity field of the disc is keplerian and taking @xmath13 for the mass and radius of the white dwarf ( baptista et al .
1998 ) , then the orbital radius @xmath14 corresponding to this velocity @xmath15 is given by : @xmath16 this gives an outer disc radius of @xmath17@xmath18 .
although our estimated outer disc radius strictly exceeds the distance to l1 ( @xmath19 ) , tidal interactions will make the disc velocities sub - keplerian and keep the disc within the primary roche lobe .
similarly , the emission line wings indicate that disc emission extends to @xmath4 3500 km s@xmath6 corresponding to an inner disc radius of @xmath20@xmath21 . in order to confirm that the emission lines are indeed formed in the disc , we measured the eclipse behaviour of the emission lines as a function of velocity .
we extracted light curves for each pixel between @xmath5 4250 km s@xmath6 of h@xmath9 and measured the eclipse half depth at ingress and egress ( figure [ zwave ] ) .
we see the blue - shifted side of the outer disc being eclipsed first followed by progressively smaller radii at larger velocities .
the red - shifted disc emission is eclipsed later , with the outer edge disappearing just before mid - eclipse .
the presence of a nearby hei line at + 3800 km s@xmath6 distorts the measurements of the far red wing .
some uneclipsed light is present at low velocities .
it is symmetric around zero velocity and has a fwhm of 1000 km s@xmath6 .
also , the average time of mid - eclipse occurs significantly earlier than white dwarf conjunction .
this has also been reported in ip pegasi during outburst ( steeghs 1999 , morales - rueda et al .
analytical models for the velocity dependent eclipse phases of symmetric keplerian discs are centered around phase 0.0 ( e.g. young , schneider & shectman 1981 ) .
these models are clearly not able to reproduce the observed eclipse behaviour of v2051 oph , which suggests a strong disc asymmetry is present .
the analysis demonstrates that most of the line emission is indeed produced in a pro - gradely rotating disc that extends from close to the white dwarf to the outer regions of the primary roche lobe .
we isolated the mid - eclipse spectrum ( figure [ specs ] , middle ) by averaging all spectra between orbital phases 0.995 and 1.005 .
as the relatively blue continuum from the inner disc is eclipsed , the cool companion star makes a larger fractional contribution , resulting in a slightly reddened continuum slope .
weak emission lines with a very strong blue shifted emission component are present , indicating that the eclipse is not total and some ( outer ) disc emission is still visible around these phases . the fact that the blue - shifted side is much stronger than the red - shifted peak , points to an asymmetry in the outer disc that was already visible as a persistent asymmetry in the out - of - eclipse line profiles .
either there is excess line flux produced in the blue - shifted side of the disc , or part of the red - shifted side is absorbed by a geometrically ( and optically ) thick structure . in order to isolate the contribution from the white dwarf we subtracted the spectra from either side of white dwarf ingress and egress ( figure [ specs ] , bottom ) . to correct , to first order , for the changing contribution from the disc , we performed a linear fit to each wavelength pixel on a short section just before and after ingress .
this fit was used to extrapolate the spectrum just before and after white dwarf ingress , and the white dwarf spectrum was obtained by taking the difference .
the same procedure was applied at white dwarf egress , and the two derived white dwarf spectra , identical within the error bars , were averaged to derive our final white dwarf spectrum as plotted in figure [ specs ] . some emission line contribution from the disc
is still present , although strong absorption cores of the balmer lines are due to the white dwarf absorption line spectrum .
the blue slope reflects the continuum contribution expected from a hot white dwarf .
its relative contribution to the out - of - eclipse spectrum increases from @xmath4 15% at 8000 up to @xmath4 30% at 4000 .
black - body fits to the continuum shape indicates a white dwarf temperature around 15000 @xmath5 2000 k. catalan et al . ( 1998 ) also derived a white dwarf temperature of 15000k from hst fos data . in order to measure the contact phases of the white dwarf
accurately , we applied the conventional method of using the derivative of the continuum light curve ( e.g. wood et al .
the light curve was first filtered with a running mean filter using a width of one third of the expected duration of the ingress and egress features .
the numerical derivative was then calculated , and the ingress ( @xmath22 ) and egress ( @xmath23 ) are defined as the orbital phases of minimum and maximum derivative ( figure [ wdphases ] ) . in order to determine these extrema , the derivative light curve was fitted with a gaussian near the white dwarf features .
the fitted centroids of these gaussians are marked by vertical dashed lines . to exploit the available spectral information
, we also fitted the continuum slope of each spectrum with a first order polynomial .
the ingress and egress features are expected to show up prominently on the continuum slope as the blue white dwarf spectrum is eclipsed / reappears .
figure [ wdphases ] plots the derived continuum slope after filtering using the same filter that was used for the continuum light curve .
the slope is a fit to all continuum sections between 4000 and 6800 , with all the emission line regions masked out .
a positive slope corresponds to a spectrum that increases towards red wavelengths .
indeed , white dwarf ingress and egress are easily identified as continuum slope changes .
we performed a similar analysis to this light curve .
the numerical derivative was calculated and the extrema were fitted with gaussians . from the ingress and egress phase we derived the duration of the white dwarf eclipse ; @xmath24 and the time of mid - eclipse @xmath25 .
our values ( with a precision of @xmath26 ) as well as the values determined by baptista et al .
( 1998 ) are listed in table [ wdtable ] .
the two methods give compatible results , although the white dwarf ingress derived from the continuum slope is slightly later compared to the conventionally derived phase .
we therefore averaged the results of the two methods to derive our final measurements .
our derived phases are systematically earlier compared to baptista et al .
( 1998 ) , indicating that the orbital ephemeris used is off by @xmath412 seconds at the time of our observations .
our expected absolute time accuracy is better than 200ms , and this is therefore a significant difference . we did not attempt to construct a new orbital ephemeris but instead use our own measured white dwarf contact phases whenever needed as an internal calibration of the phasing . we also derive a larger value for @xmath27 , compared to baptista et al .
we remark that the system was observed just after outburst in our case , whereas baptista et al .
( 1998 ) observed v2051 oph during a rare low state .
.wd contact phases [ cols="<,^,^,^,>",options="header " , ] table [ amplitudes ] lists the oscillation amplitude ranges obtained from these sine fits to the balmer lines and the continuum .
although the continuum oscillation is much weaker compared to the lines , both reach maximum relative amplitudes of a few percent .
we have obtained some strong constraints to the origin and location of the continuum and emission line oscillations . in the next section
we review several possible models for the observed oscillations , and see if they are able to satisfy the above constraints .
before reviewing possible explanations for the observed oscillations , let us quickly summarise the main features that a model must satisfy in order to explain the observations . in the continuum
we observe oscillations at 56.12s and its harmonic at 28.06s , that must originate close to the white dwarf or on the compact object itself .
the emission lines oscillate mostly at 29.77s , and require a source at a distance of @xmath412@xmath28 , corresponding to an orbital period of 488s , which is also the beat period between 29.77s and 28.06s .
the emission line oscillation amplitudes modulates regularly on the beat period of 488s and emission is only visible going from maximum blue - shift to red - shift , but not vice - versa . over the years various models
have been proposed that attribute the dnos to oscillations or modes in the disc itself .
possible mechanisms are non - radial pulsations ( e.g. van horn , wesemael & winget 1980 ) , axisymmetric radial pulsations ( kato 1978 ) , vertical oscillations or resonant oscillations due to shock waves ( molteni et al . 1996 ) . in most cases ,
the oscillation period is to first order the local keplerian period .
we can then obviously connect each observed frequency to a particular radius in the disc which gives orbits between 1 and 3 white dwarf radii out for the observed oscillations . for example , @xmath29 corresponds to a kepler orbit at @xmath30 , and its harmonic is at @xmath31 .
however , there is no particular reason why the disc would oscillate at these distinct radii , and not at some other , and no explanation of how it manages to do so coherently . also , the presence of the harmonic of @xmath32 but not @xmath33 is not explained , nor is the connection between @xmath33 , @xmath32 and the observed beating behaviour with a period of 488s .
more seriously , we showed in section 3 that the emission line oscillation originates at @xmath34 , but oscillates with a period of 29.77s , which is not the local kepler period .
although we have established that the emission line oscillation requires an origin in the disc , which is an important result in itself , a model relying on disc modes only seems unsatisfactory .
the emission line oscillation can not originate from the white dwarf itself , but the continuum oscillations at @xmath29 and @xmath35 are indeed expected to be related to the white dwarf .
one possible mechanism that can produce periodicities apart from a fixed structure on a rotating white dwarf , is oscillations of the white dwarf .
papaloizou & pringle ( 1977 ) discussed non - radial oscillations of rotating stars and concluded that oscillations in rotating white dwarfs can indeed produce the oscillations at periods corresponding to dnos .
the periods of these modes with order @xmath36 is @xmath37 , with @xmath38 the spin period of the white dwarf , and in principle any number of orders can be superposed at any given time .
it is thus not obvious how one mode will stand out compared to the others and produce a single , coherent spike in the power spectrum .
the detected power at @xmath29 and @xmath35 could thus , in principle , be interpreted as the incarnation of two orders of a non - radial oscillation on the white dwarf .
however , two equatorial spots on a rotating white dwarf with spin period @xmath29 could explain the observations just as well .
the latter would also explain the observed beating behaviour between the spinning white dwarf and orbiting disc material more naturally .
one way to distinguish between these two models would be the detection of other periods , but not @xmath29 , in this system . in the case of white dwarf oscillations
it is easy to produce power at different orders and modes , however the white dwarf spin period can not be changed on short time scales and should leave a persistent signature .
another class of models that relies on the origin of the dnos close to the white dwarf are differentially rotating surface layers above the white dwarf ( warner 1995 ) , or interaction between the disc and the boundary layer around the white dwarf ( e.g. popham 1999 ) .
the main concern with these models , when trying to understand the behaviour of v2051 oph , is that they can in principle account for the oscillations that originate close to the white dwarf ( @xmath39 ) , but not the emission lines oscillation ( @xmath33 ) .
emission line oscillations have been seen so far only in the cv sub - class of intermediate polars such as dq her ( martell et al .
1995 ) and rxj0558 ( harlaftis et al .
1999 ) . in these systems , the magnetic white dwarf accretes via two hot spots at its magnetic poles as it spins rapidly
. the disc is truncated at the radius where magnetic pressure is comparable to the gas ram pressure , and the gas flow makes a transition from keplerian motion in the plane to a motion along field lines .
the observed line oscillations are produced by disc material that is illuminated by the hot spots on the white dwarf as they sweep by .
however , if we try to interpret the observed oscillations in v2051 oph in terms of this model , it fails for several reasons .
we have established that the emission lines are formed in the disc which extends from very close to the white dwarf out to the outer edges of the primary roche lobe .
the emission line oscillation is produced in the disc at @xmath40 , but the disc is clearly not truncated there . in this model , @xmath29 is the spin period of the white dwarf , which leads to a co - rotation radius of @xmath41 .
apart from the fact that the disc apparently extends down to the white dwarf , the system would be a propeller and any material threading at @xmath42 would be propelled out .
finally , in the intermediate polars we observe the emission lines as they move from red - shift to blue shift and not vice versa as in the case of v2051 oph . if we would interpret the observed line oscillation kinematics ( figures [ fold],[spot ] ) in terms of material moving along magnetic field lines rather than in the disc , we would nt see the blue wing oscillations being eclipsed first followed by the red wing at mid - eclipse ( figure [ balamp ] ) .
the intermediate polar scenario therefore fails in the sense that the emission line oscillations can not originate from the edge of a a truncated disc as in dq her and rxj0558 or from a magnetic curtain .
however , the system may well contain a weakly magnetised white dwarf with two active magnetic poles that is responsible for the oscillations at @xmath32 and @xmath43 .
however , this can not account for the emission line oscillations at 29.77s .
we have reviewed some possible models that could explain the observed set of oscillations in v2051 oph .
the continuum oscillations at @xmath32 and @xmath43 are most naturally explained by a spinning white dwarf with @xmath44 containing two hot spots , that produce a modulation at the spin period and its harmonic .
the origin of these hot - spots may be due to a weak magnetic feld of the primary , which is low enough such that the disc can still extend to close to the white dwarf .
the radiation from these two spots illuminates the disc , and is reprocessed as continuum and line emission at a radius of @xmath45 .
the corresponding emission oscillates at the beat frequency between the spin of the white dwarf and the local kepler period of 488s producing line oscillations at 29.8s . we noted before that we only see line emission going from blue shift to red shift and not vice - versa .
this may be caused by the presence of material close to the white dwarf above the disc plane , that absorbs the line emission when it is on the far side of the disc from the point of view of the observer , but not when it is in front .
one candidate for this is the uneclipsed light .
alternatively , the emission of orbiting material could be beamed an - isotropically , only producing line emission away from the white dwarf .
orbiting material would then only be visible going along the front side .
perhaps the inner side is blasted by the white dwarf and is too hot to produce optical line emission , whereas the outer side is at low enough temperatures to emit balmer and helium emission .
what remains to be explained is why does this processing occur at this specific radius , and why does it modulate its amplitude regularly ? since the orbiting disc material at this radius is beating with the white dwarf spin , it needs to be a localised region , both in terms of radius as well as azimuth , that is mostly responding for it to produce a coherent oscillation .
a localised bulge in the disc at the required radius would orbit the disc every 488s and its oscillation amplitude would modulate because of the intervening gas that prevents us from viewing the bulge when its on the far side .
this obscuring material that was required to explain the blue - to - red pattern in the emission lines , thus also explains the regular modulation of the amplitude every 488s .
popham ( 1999 ) also invoked a non - axisymmetric bulge in the disc to produce dnos . in his model ,
the bulge is formed at the transition radius between the disc and the boundary layer surrounding the white dwarf . in our case , the bulge is clearly too far from the white dwarf to be due to such a boundary layer - disc interaction .
we are now in the familiar , and unfortunate , position again that one needs to invoke some unknown mechanism in order to produce a structure in the disc at this specific radius .
figure [ cartoon ] plots a schematic top view of the binary , with the site of the line oscillation marked as a circle .
we calculated a few characteristic radii for v2051 oph , using the binary parameters of baptista et al .
for example , we could nt connect this radius with a tidal resonance .
however , the circularisation radius for this system is ( warner 1995 ) ; @xmath46 this close correspondence between the emission line source and the circularisation radius suggests that gas stream effects may be of relevance .
lubow ( 1989 ) explored the possibility of stream overflow , and in the case of no disc interaction streamlines typically re - impact in the orbital plane at the radius of closest approach ; @xmath47 the exact nature of the stream - disc interaction is not well understood , but we may expect some stream - disc effects between these two radii .
the attractive feature of such a stream overflow would be the natural way in which a vertically extended structure would be present at roughly the right radius to produce the reprocessed emission .
we also noted in section 3.1 that near mid - eclipse the blue shifted contribution from the disc is persistently stronger compared to the red - shifted emission .
stream overflow would block part of the red - shifted disc emission from view , and this asymmetry could thus be easily accommodated .
the reversal of this asymmetry in the h@xmath7 remains a puzzle , however .
it is clearly not trivial to produce a localised bulge in the disc at this stream impact location , that is able to orbit coherently for quite some time while stream overflow is occuring .
observations of this kind spanning longer intervals , could shed some light on the coherency of such structures .
a more thorough simulation of such a scenario is also required in order to test if such a geometry is a viable way of producing the observed oscillations ( and perhaps other dnos ) .
this is beyond the scope of this paper , but such a scenario does appear to be able to accommodate all the features of the observed oscillations . since we observe the system towards the end of an outburst , the radial density distribution is not expected to be that of a steady - state disc .
disc instability calculations usually show a broad peak in the density at roughly the circularisation radius towards the end of the outburst .
such an empty disc would be in particular susceptible to stream overflow , and again the obvious region for stream - disc interaction would be near the circularisation radius where the density peaks and the stream is flowing towards the orbital plane again .
we detected coherent continuum and emission line oscillations in the dwarf nova v2051 oph on decline from a normal outburst .
accretion disc emission extends from very close to the white dwarf out to the outer parts of the primary roche lobe .
the disc emission lines display a persistent blue to red asymmetry , with the blue peak being stronger in all the helium lines as well as the balmer lines , except h@xmath7 which has a stronger red peak . the eclipse light curves are also highly asymmetric which suggests that the blue side of the disc makes a larger contribution to the emission compared to the red - shifted side .
the continuum oscillations are most likely to originate on the surface of a spinning white dwarf with spin period 56.12s and temperature around 15000k .
the amplitude of the oscillation in the continuum varies between 0 and 4% , and disappears during white dwarf eclipse .
the balmer and helium i emission lines oscillate strongly at a period of 29.8s .
the line kinematics as well as the eclipse constraints indicate these to come from a non - axisymmetric bulge in the disc at a radius of @xmath0 .
the corresponding kepler orbit has a period of 488s , and corresponds well with the observed amplitude variations ( 0 - 4% ) on this period .
the oscillating line emission is observed to go from maximum blue - shift to maximum red - shift every 29.8s , but not vice versa .
this , together with the regular modulation of the oscillation amplitudes can be explained by intervening gas above the orbital plane , close to the white dwarf .
this is supported by an uneclipsed component in both the continuum and lines . in the balmer lines
this component is centered on zero velocity and has a fwhm of 1000 km s@xmath6 .
alternatively , the emission is beamed away from the white dwarf . the close correspondence between the location of the oscillations and the circularisation radius of the system as well as the disc asymmetry may indicate the relevance of stream overflow to the presence of vertically extended bulges in the disc .
more observations of this kind as well as more detailed simulations of such a scenario are required to confirm its feasibility in the light of producing dnos .
predictions of our interpretation are the persistent presence of 56.12s/28.06s oscillations from the white dwarf .
this period should be detected at other epochs , most likely in the uv .
the uv also provides opportunities to measure the spin of the white dwarf directly using the rotational broadening of white dwarf absorption lines ( sion 1999 ) . with a spin period of 56.12s
, the expected surface velocity of the spinning white dwarf works out to be : @xmath48 secondly , we expect the period of the emission line oscillations to change across the outburst cycle as the location and extent of stream overflow and re - impact will depend on the radial density distribution of the disc . for example , if we wish to interpret the reported 42.2s oscillation ( warner & odonoghue 1987 ) in this scenario , it can be identified as the beat between the white dwarf spin and a keplerian orbit at @xmath49 .
the data presented here demonstrate the potential of high speed spectroscopy of cvs using large aperture telescopes .
we detected emission line oscillations in a dwarf nova for the first time , and located their origin in the accretion disc .
a better understanding of the elusive dnos would benefit from a high time resolution study of ( eclipsing ) dwarf novae across their outburst cycle .
data presented herein were obtained at the w.m . keck observatory , which is operated as a scientific partnership among the california institute of technology , the university of california and the national aeronautics and space administration .
the observatory was made possible by the generous financial support of the w.m .
keck foundation .
we thank tom marsh for the use of his molly analysis software , and for many useful discussions .
lars bildsten , graham wynn , phil charles , jeno sokolovski and rob hynes are thanked for their comments and suggestions .
we thank the referee for his / her usefull comments and suggestions .
ds was supported by a pparc fellowship during part of this research .
baptista , r. , catalan , m. s. , horne , k. , zilli , d. , 1998 , mnras , 300 , 233 bruch , a. , 1992 , a&a , 266 , 237 catalan , m.s . , horne , k. , cheng , f.h . , marsh , t.r .
, hubeny , i. , 1998 , asp conf .
series 137 , _ wild stars in the old west _
, p426 cook , m.c . , brunt , c.c . , 1983 , mnras , 205 , 465 harlaftis , e.t . , horne , k. , 1999 , mnras , 305 , 437 kato , s. , 1978 , mnras , 185 , 629 lubow , s.h . , 1989 , ap.j , 340 , 1064 marsh , t.r . , horne , k. , 1998 , mnras , 299 , 921 martell , p.j . ,
horne , k. , price , c.m . , gomer , r. , 1995 , apj , 448 , 380 molteni , d. , sponholz , h. , chakrabarti , s.k .
, 1996 , apj , 457 , 805 morales - rueda , l. , marsh , t.r . ,
billington , i. , 2000 , mnras , 313 , 454 obrien , k. , 2000 , ph.d thesis , university of st.andrews oke , b. , 1990 , aj , 99 , 162 oke , b. , oke , cohen , j. g. , carr , m. , cromer , j. , dingizian , a. , harris , f. h. , labrecque , s. , lucinio , r. , schaal , w. , epps , h. , miller , j. , 1995 , pasp , 107 , 375 popham , r. , 1999 , mnras , 308 , 979 sion , e. , 1999 , pasp , 111 , 532 steeghs , d. , 1999 , ph.d thesis , university of st.andrews van der klis , m. , 2000 , araa , 38 van horn , h.m . , wesemael , f. , winget , d.e . , 1980 , ap.j , 235 , 143 young , p.j . ,
schneider , d.p . ,
shectman , s.a . , 1981 ,
apj , 244 , 259 walker , 1956 , ap.j , 123 , 68 warner , b. , robinson , e. , 1972 , nat.phys.sci .
, 239 , 2 warner , b. , cropper , m. , 1983 , mnras , 203 , 909 warner , b. , odonoghue , d. , 1987 , mnras , 224 , 733 warner , b. , 1995 , _ cataclysmic variable stars _ , cambridge astrophysics series 28 , cambridge university press watts , d.j . ,
et al . , 1986 , a&a , 154 , 197 wood , j.h . ,
horne , k. , berriman , g. , wade , r. , odonoghue , d. , warner , b. , 1986 , mnras , 219 , 629 | we have detected coherent oscillations , at multiple frequencies , in the line and continuum emission of the eclipsing dwarf nova v2051 ophiuchi using the 10 m keck ii telescope .
our own novel data acquisition system allowed us to obtain very fast spectroscopy using a continuous readout of the ccd on the lris spectrograph .
this is the first time that dwarf nova oscillations are detected and resolved in the emission lines .
the accretion disc is highly asymmetric with a stronger contribution from the blue - shifted side of the disc during our observations .
the disc extends from close to the white dwarf out to the outer regions of the primary roche lobe .
continuum oscillations at 56.12s and its first harmonic at 28.06 s are most likely to originate on the surface of a spinning white dwarf with the fundamental period corresponding to the spin period .
balmer and helium emission lines oscillate with a period of 29.77s at a mean amplitude of 1.9% .
the line kinematics as well as the eclipse constraints indicate an origin in the accretion disc at a radius of @xmath0 . the amplitude of the emission line oscillation modulates ( 0 - 4% ) at a period of 488s , corresponding to the kepler period at r=@xmath1 .
this modulation is due to the beating between the white dwarf spin and the orbital motion in the disc .
the observed emission line oscillations can not be explained by a truncated disc as in the intermediate polars .
the observations suggest a non - axisymmetric bulge in the disc , orbiting at @xmath1 , is required .
the close correspondence between the location of the oscillations and the circularisation radius of the system suggests that stream overflow effects may be of relevance .
[ firstpage ] accretion , accretion discs novae , cataclysmic variables stars : individual : v2051 oph stars : oscillations |
there is a stochastic multigrid method and a deterministic one . the stochastic version is used to compute high dimensional integrals in euclidean quantum field theory or statistical mechanics by a monte carlo method which uses updates at different length scales @xcite .
the deterministic version @xcite solves discretized partial differential equations .
one hopes to use both of them in simulations of lattice qcd , for updating the gauge fields and for computing fermion propagators in given gauge fields . in either case
the aim is to beat critical slowing down ( csd ) in nearly critical systems .
our notation is as follows : @xmath0 denotes a given `` fundamental '' lattice @xmath1 of spacing @xmath2 .
coarser ( block ) lattices of increasing spacings @xmath3 are denoted @xmath4 . typically , we chose @xmath5 , and a single point as the last layer @xmath6 .
interpolation operators @xmath7 are introduced to transfer functions on coarser lattices into functions on finer lattices , while restriction operators @xmath8 transfer functions from finer to coarser lattices ( `` variational coarsening '' ) .
a crucial problem is how to define and exhibit smooth functions in the disordered context , i.e. when translation symmetry is strongly violated .
other possible applications besides gauge theories are low lying states of spin glasses , the shape of a lightning , waves on fractal lattices ( with bond percolation ) , or the localization of low lying electronic states in amorphous materials . in the case of deterministic mg
, one wants to solve a discretized elliptic differential equation on @xmath1 : d_0 ^ 0 = f^0 .
[ de ] it might have arisen from an eigenvalue equation @xmath9 by inverse iteration .
if @xmath10 has a small eigenvalue , then local relaxation algorithms suffer from csd .
after some relaxation sweeps on @xmath1 one gets an approximate solution @xmath11 whose error @xmath12 is not necessarily small but is _ smooth ( on length scale @xmath2 ) .
the unknown error @xmath13 satisfies the equation d_0 e^0 = r^0 .
[ dee ] with the residual @xmath14 .
given that @xmath13 is smooth , it can be obtained by smooth interpolation of a suitable function @xmath15 on @xmath16 , e^0 = ^1 e^1 .
[ intp0 ] that is , @xmath17 with @xmath18 which depends smoothly on @xmath19 .
now define a restriction operator @xmath20 such that @xmath21 .
then ( [ intp0 ] ) can be inverted , @xmath22 applying @xmath23 to both sides of ( [ dee ] ) yields an equation for @xmath24 , d_1 e^1 = r^1 , [ cgrideq]with @xmath25 and the effective operator @xmath26 . given @xmath24 , one obtains @xmath13 from ( [ intp0 ] ) , and @xmath27 is an improved solution of ( [ de ] ) .
thus , _ the problem has been reduced to an equation on the lattice @xmath28 which has fewer points . if necessary , one repeats the procedure , moving to @xmath29 etc .
the procedure stops , because an equation on a `` lattice '' @xmath6 with only a single point is easy to solve .
_ _ the iterated interpolation @xmath30 } \equiv \a^1 \a^2 \dots \a^j$ ] from @xmath31 to @xmath1 should yield functions on @xmath0 which are _ smooth on length scale @xmath32 , i.e. which change little over a distance @xmath32 ( in the ordered case ) . for reasons of practicality , one must require @xmath33 unless @xmath19 is near @xmath34 . _
a successful mg scheme , whether deterministic or stochastic , needs smooth interpolation kernels @xmath35 .
thus we may ask : _ which functions are smooth in the disordered situation , for instance in an external gauge field ? _ a ( gauge covariant ) naive answer is _ ( _ , _ ) = ( , - ) _ 0 ( , ) ( with discretized covariant derivatives @xmath36 ) . by definition ,
the lowest eigenvalue @xmath37 of the negative covariant laplacian @xmath38 is not small for disordered gauge fields .
( it is positive and vanishes only for pure gauges . )
therefore there are no smooth functions in this case .
nevertheless there is an answer to the question , assuming a fundamental differential operator @xmath10 is specified by the problem ( in the stochastic case , the hamiltonian often provides @xmath10 ) : _ a function @xmath39 on @xmath40 is smooth on length scale @xmath41 when @xmath42 in units @xmath43 . _ we found that a deterministic multigrid which employs interpolation kernels @xmath44}$ ] from @xmath45 to the fundamental lattice @xmath1 which are smooth in this sense , works for arbitrarily disordered gauge fields .
when there are no smooth functions in this sense at length scale @xmath2 , then @xmath10 has no low eigenvalue , and there is no csd and no need for mg .
the above answer appears natural , and the `` projective mg '' of @xcite is in its spirit . but to obtain kernels @xmath44}_{zx}$ ] which are smooth _ on length scale @xmath32 , one needs approximate solutions of eigenvalue equations d_0 ^[0j]_zx = _
0(x ) ^[0j]_zx [ eva ] since @xmath44 } $ ] is required to vanish for @xmath19 outside a neighbourhood of @xmath46 , the problem involves dirichlet boundary conditions . for large @xmath47 , @xmath44 }
$ ] will have a large support .
if there is no degeneracy in the lowest eigenvalue , one can use inverse iteration combined with standard relaxation algorithms for the resulting inhomogeneous equation .
but this and other standard methods will suffer from csd again .
moreover , in the standard multigrid setup , one uses basic interpolation kernels @xmath48 which interpolate from one grid @xmath49 to the next finer one . in this case ^[0j ] = ^1 ^2
^j , [ factor ] and ( [ eva ] ) becomes a very complicated set of nonlinear conditions .
possible solutions are _
( i ) : : replace ( [ eva ] ) by minimality of a cost functional ( cp . later ) .
use neural algorithms to find kernels @xmath7 which minimize it .
this is still under study .
( ii ) : : give up factorization ( [ factor ] ) and determine independent kernels @xmath44}$ ] as solutions of ( [ eva ] ) by multigrid iteration .
_ this is done successively for @xmath50 one uses already determined kernels @xmath51}$ ] with @xmath52 for updating @xmath44}$ ] .
_ we found that this works very well - cp .
[ performance ] . method ( ii ) will need of order @xmath53 storage space and @xmath54 computational work for a @xmath55-dimensional system of linear extension @xmath56 .
any iteration to solve ( [ de ] ) amounts to updating steps of the form ^0 ^ 0 = ^0 + f^0 .
[ iterlin ] with the iteration matrix @xmath57 whose norm governs the convergence , and @xmath58 . if @xmath59 , the iteration converges with a relaxation time @xmath60 .
_ parameters in the algorithm - such as operators @xmath61 , @xmath62 and @xmath63 - are _ optimal if the cost functional @xmath64 is at its minimum .
_ _ as an example , consider a twogrid iteration in which a standard relaxation sweep on @xmath1 with iteration matrix @xmath65 is followed by exact solution of the coarse grid equation ( [ cgrideq ] ) .
the second step leads to an updating with some iteration matrix @xmath66 , and @xmath67 .
therefore one may estimate @xmath68 with @xmath69 ( fine grid relaxation smoothens the error but does not converge fast - therefore @xmath70 is suppressed whereas @xmath71 is not much smaller than @xmath72 ) and try to optimize the parameters above by minimizing @xmath73 : using the trace norm , @xmath74 , one finds e_1 = volume^-1 _ z , w^0
|_zw|^2 with @xmath75 . prescribing @xmath23 , and determining @xmath76 and @xmath77 by minimizing @xmath73 yields what we call the `` ideal interpolation kernel '' @xmath18 for a given restriction map @xmath20 .
since it has exponential tails instead of vanishing for @xmath19 outside a neighbourhood of @xmath46 , it is impractical for production runs , though @xcite .
a feed - forward artificial neural network ( ann ) @xcite can perform the computations to solve ( [ dee ] ) by mg relaxation .
the nodes ( `` neurons '' ) of the nmg are identified with points of the mg as shown in fig .
[ figur1 ] .
the resulting nmg consists of two copies of the same mg , except that the last layer @xmath78 is not duplicated . in the standard mg approach , the basic interpolation kernels @xmath7 interpolate from one layer @xmath45 to the preceding one , @xmath79 .
each node is connected to some of the nodes in the preceding layer . in the upper half ,
the connection strength from @xmath80 to @xmath81 is @xmath61 . in the lower half ,
node @xmath82 is connected to @xmath80 with strength @xmath83 .
in addition there is a connection of strength @xmath84 between the two nodes which represent the same point @xmath19 in @xmath31 ( @xmath85 ) . according to hebb s hypothesis of synaptical learning
, a biological neural network learns by adjusting the strength of its synaptical connections .
the network receives as input an approximate solution @xmath39 of ( 2.1 ) , from which the residual @xmath86 is then determined .
it computes as output an improved solution @xmath87 .
the desired output ( `` target '' ) is @xmath88 .
@xmath89 is a linear function of @xmath90 . except on the bottom layer
, each node receives as input a weighted sum of the output of those nodes below it in the diagram to which it is connected .
the weights are given by the connection strengths .
our neurons are linear because our problem is linear .
the output of each neuron is a linear function of the input .
the result of the computation is = ( _ 0 d_0 ^ -1 + _ k1 ^[0k ] _ k d_k^-1 r^[k0 ] ) r^0 where @xmath91}=r^k ... r^2r^1 $ ] and @xmath51}= \a^1\a^2 ... \a^k$ ] . the operators @xmath92 and @xmath93 , and the damping parameters @xmath94 @xmath95 are not needed separately since they only enter in the combination @xmath96 .
the fundamental differential operator @xmath97 and its diagonal part @xmath98 are furnished as part of the problem .
the connection strengths ( `` synaptical strengths '' ) @xmath61 , @xmath83 ( and possibly @xmath99 ) need to be found by a learning process in such a way that the actual output is as close as possible to the desired output . in supervised learning @xcite , pairs @xmath100 ( `` training patterns '' )
are presented to an ann . given input @xmath101 , the actual output @xmath102 is compared to the target @xmath103 , and the connection strengths are adjusted in such a way that the cost functional @xmath104 gets minimized .
an iterative procedure to achieve this minimization is called _ learning rule .
_ taking for the sequence @xmath101 a complete orthonormal system of functions on @xmath1 , in the limit @xmath105 , the target is @xmath106 for any input , and the output @xmath107 by ( [ iterlin ] ) .
the learning rule for the resulting cost functional e= _ ^^2= tr _
^2 is our previous optimality condition for multigrid relaxation in sect .
[ criteria ] .
the variant ( ii ) in sect . [ smoothness ] involves a slightly different nmg . instead of the connections between neighbouring layers of the multigrid , we now have connections from @xmath1 to @xmath108 with strength @xmath109}_{xz}$ ] , and from @xmath31 to @xmath1 with strength @xmath44}_{zx}$ ] .
if we adopt variational coarsening , all connection strengths are determined by interpolation kernels @xmath51}$ ] which have to be learned .
the damping factors @xmath110 were set to @xmath72 and @xmath111 is the diagonal part of @xmath112 as before , with d_k = ^[0k]d_0 ^[0k ] . the learning rule ( ii )
requires a process of `` hard thinking '' by the nmg .
nodes which have learned their lesson already - i.e. which have their connection strengths fixed - are used to instruct the rest of the neural net , adjusting the strengths of the next layer of nodes in the nmg .
a variant of this algorithm was tested in 2 dimensions , using su(2)-gauge fields which were equilibrated with standard wilson action at various values of @xmath113 , and @xmath114 .
@xmath37 is the lowest eigenvalue of the covariant laplacian @xmath38 , and @xmath115 .
conventional relaxation algorithms for solving ( [ de ] ) suffer from csd for such @xmath10 , for any volume and small @xmath116 .
it turned out that it was not necessary to find accurate solutions of the eigenvalue equation for the interpolation kernels @xmath44}$ ] .
an approximation @xmath44}_{zx}$ ] to @xmath117 was computed by multigrid iteration .
it does not depend on @xmath116 .
updating @xmath39 at @xmath118 changes @xmath39 by _
z= ^[0j]_zxd_j , x^-1r^j_x , r^j=^[0j]r^0 .
the convergence rate ( in units of mg iterations ) of the @xmath119-iteration is shown in fig .
[ figur2 ] for @xmath120 .
one mg iteration involved one sweep ( in checkerboard fashion ) through each mg layer , starting with @xmath121 .
99 g. mack , in : nonperturbative quantum field theory , eds .
g. t hooft et al .
( plenum , new york 1989 ) .
j. goodman and a.d .
sokal , phys .
d 40 ( 1989 ) 2035 .
w. hackbusch , multi - grid methods and applications ( springer - verlag , berlin , 1985 ) .
brower , k. moriarty , c. rebbi and e. vicari , nucl .
b 20 ( 1991 ) 89 .
a. hulsebos , j. smit and j.c .
vink , nucl .
b 368 ( 1992 ) 379 .
t. kalkreuter , phys .
b 276 ( 1992 ) 485 .
j. hertz , a. krogh and r.g.palmer , introduction to the theory of neural computation ( addison - wesley , redwood city ca , 1991 ) . | we present evidence that multigrid ( mg ) works for wave equations in disordered systems , e.g. in the presence of gauge fields , no matter how strong the disorder .
we introduce a `` neural computations '' point of view into large scale simulations : first , the system must learn how to do the simulations efficiently , then do the simulation ( fast ) .
the method can also be used to provide smooth interpolation kernels which are needed in multigrid monte carlo updates .
h # 1 # 1 |
we consider undirected graphs with no loops and no parallel edges .
all notions and facts on graphs , that are used but not described here , can be found in @xcite . + given graphs @xmath0 and
@xmath15 , an @xmath15-_packing _ of @xmath0 is a subgraph of @xmath0 whose all components are isomorphic to @xmath15 .
an @xmath15-_packing _ @xmath16 of @xmath0 is called an @xmath15-_factor _ if @xmath17 . the @xmath15-_packing problem _
, i.e. , the problem of finding in @xmath0 an @xmath15-packing , having the maximum number of vertices , turns out to be @xmath18-hard for every connected graph @xmath15 with at least three vertices @xcite .
let @xmath1 denote a 3-vertex path .
in particular , the @xmath1-packing problem is @xmath18-hard .
moreover , the problem of packing @xmath19-vertex paths , @xmath20 , is @xmath18-hard for cubic graphs@xcite . also , the 3-vertex path packing problem remains @xmath18-hard even for cubic bipartite planar graphs @xcite .
although the @xmath1-packing problem is @xmath18-hard , i.e. , possibly intractable in general , this problem turns out to be tractable for some natural classes of graphs . it would be also interesting to find polynomial algorithms that would provide a good approximation for the problem .
below ( see * [ eb(g)clfr ] * - * [ 2conclfr ] * ) there are some examples of such results . in each case
the corresponding packing problem is polynomially solvable .
+ let @xmath2 and @xmath3 denote the number of vertices and the maximum number of disjoint 3-vertex paths in @xmath0 , respectively .
obviously , @xmath21 .
a graph is called _ claw - free _ if it contains no induced subgraph isomorphic to @xmath22 ( which is called a _ claw _ ) .
block _ of a graph @xmath0 is either an isolated vertex or a maximal connected subgraph @xmath15 of @xmath0 such that @xmath23 is connected for every vertex @xmath24 of @xmath15 .
a block of a connected graph is called an _ end - block _ if it has at most one vertex in common with any other block of the graph .
let @xmath25 denote the number of end - blocks of @xmath0 .
[ eb(g)clfr ] suppose that @xmath0 is a connected claw - free graph and @xmath26 .
then @xmath27 , and this lower bound is sharp .
[ 2conclfr ] suppose that @xmath0 is a connected and claw - free graph having at most two end - blocks @xmath28 in particular , a 2-connected and claw - free graph@xmath29
. then @xmath30 .
obviously , the claim in * [ 2conclfr ] * on claw - free graphs with exactly two end - blocks follows from * [ eb(g)clfr]*. + in @xcite we answered the following natural question : + [ km ] if @xmath0 is a cubic graph , then @xmath31 .
moreover , there is a polynomial time algorithm for finding a @xmath1-packing having at least @xmath32 components . obviously , if every component of @xmath0 is @xmath33 , then @xmath34 .
therefore the bound in * [ km ] * is sharp .
there are some other connected cubic graphs for which the bound in * [ km ] * is attained .
+ let @xmath35 denote the set of graphs with each vertex of degree at least @xmath36 and at most @xmath37 . in @xcite we answered ( among other results )
the following question : + [ 2,3-graphs ] suppose that @xmath38 and @xmath0 has no 5-vertex components
. then @xmath39 .
obviously , * [ km ] * follows from * [ 2,3-graphs ] * because if @xmath0 is a cubic graph , then @xmath38 and @xmath0 has no 5-vertex components .
+ in @xcite we also gave a construction that allowed to prove the following : [ extrgraphs1 ] there are infinitely many connected graphs for which the bound in * [ 2,3-graphs ] * is attained .
moreover , there are infinitely many subdivisions of cubic 3-connected graphs for which the bound in * [ 2,3-graphs ] * is attained .
the next interesting question is : + + the following bound was recently found in @xcite .
[ cubic - connected ] if @xmath0 is a cubic connected graph and @xmath40 , then @xmath41 .
the next natural question is : [ 2con - must ] how many disjoint 3-vertex paths must a cubic 2-connected graph have ? [ cubic - connected ] _ @xcite _ if @xmath0 is a cubic connected graph and @xmath42 , then @xmath43 .
on the other hand , it is also natural to consider the following [ pr2con ] are there 2-connected cubic graphs @xmath0 such that @xmath44 ? in @xcite
we gave a construction that provided infinitely many 2-connected , cubic , bipartite , and planar graphs such that @xmath44 .
+ the main goal of this paper ( see also @xcite ) is to discuss the following old open problem which is similar to problem * [ pr2con]*. .
[ pr3con ] is the following claim true ? + @xmath4 if @xmath0 is a 3-connected and cubic graph , then @xmath5 .
originally we have we ve become interested in this problem because it is related with non - trivial relaxations of some graph hamiltonicity problems we ve been studying ( see @xcite ) .
problem * [ pr3con ] * is also related with the minimum domination problem in a graph . in @xcite b. reed
conjectured that if @xmath0 is a connected cubic graph , then @xmath45 , where @xmath46 is the dominating number of @xmath0 ( i.e. , the size of a minimum vertex subset @xmath47 in @xmath0 such that every vertex in @xmath9 is adjacent to a vertex in @xmath47 ) .
it turns out that reed s conjecture is not true for connected and even for 2-connected cubic graphs @xcite . if claim @xmath4 in * [ pr3con ] * is true , then from * [ 3-con ] * it follows , in particular , that reed s conjecture is true for 3-connected cubic graphs . in @xcite
we proved that if @xmath0 is a cubic hamiltonian graph with @xmath48 , then @xmath49 . in this paper
we show , in particular , that claim @xmath4 in * [ pr3con ] * is equivalent to some seemingly stronger claims ( see * [ 3-con ] * ) .
in section [ constructions ] we give some notation , constructions , and simple observations .
in section [ 3connected ] we formulate and prove our main theorem * [ 3-con ] * concerning various claims that are equivalent to claim @xmath4 in * [ pr3con]*. we actually give different proofs of * [ 3-con]*. thus , if there is a counterexample @xmath50 to one of the above claims , then these different proofs provide different constructions of counterexamples to the other claims in * [ 3-con]*. additionally , different proofs provide better understanding of relations between various @xmath1-packing properties considered in * [ 3-con]*. in section [ almostcubic ] we describe some properties of @xmath1-factors with respect to 3-edge cut - matchings and triangles in cubic 3-connected graphs . in section [ homomorphism ]
we give two @xmath1-factor homomorphism theorems for cubic graphs .
finally , in section [ cycl6-connected ] we give a construction of cubic cyclically 6-connected graphs showing that certain claims in * [ 3-con ] * are best possible .
we can also prove that many equivalences in * [ 3-con ] * are also true in the class of bipartite 3-connected cubic graphs and that if claim @xmath4 in * [ pr3con ] * is true for bipartite graphs , then reed s conjecture is also true for bipartite 3-connected cubic graphs .
we consider undirected graphs with no loops and no parallel edges unless stated otherwise . as usual ,
@xmath51 and @xmath52 denote the set of vertices and edges of @xmath0 , respectively , and @xmath53
. if @xmath47 is a vertex subset or a subgraph of @xmath0 , then let @xmath54 ( or simply @xmath55 ) denote the set of edges in @xmath0 having exactly one end - vertex in @xmath47 , and let @xmath56 .
if @xmath57 , then @xmath58 is the set of edges in @xmath0 incident to @xmath59 , @xmath60 , @xmath61 is the set of vertices in @xmath0 adjacent to @xmath59 , and @xmath62 . if @xmath63 , then let @xmath64 .
let @xmath65 denote the set of components of @xmath0 and @xmath66 .
we call an edge cut @xmath67 in a cubic graph a _ blockade _ , if @xmath67 is matching
@xmath19-blockade _ is a blockade with @xmath19 edges .
every connected cubic graph non - isomorphic to @xmath33 or @xmath68 has a blockade . for every such graph @xmath0 ,
let @xmath69 denote the size of a minimum blockade in @xmath0 ; the parameter @xmath69 is called the _ cyclic connectivity of _ @xmath0 and @xmath0 is called _
cyclic @xmath19-connected _ , if @xmath70 .
+ let @xmath71 and @xmath72 be disjoint graphs , @xmath73 , @xmath74 , and @xmath75 be a bijection .
let @xmath76 denote the graph @xmath77 .
we usually assume that @xmath78 , @xmath79 , and @xmath80 for @xmath81 ( see fig .
[ faasbb ] ) . ]
we also say that _
@xmath76 is obtained from @xmath72 by replacing vertex @xmath82 by @xmath83 according to @xmath84_. let @xmath72 be a cubic graph and @xmath85 .
let @xmath86 , where @xmath87 , be a graph , @xmath88 be a vertex of degree three in @xmath86 , and @xmath89 .
by using the above operation , we can build a graph @xmath90 by replacing each vertex @xmath24 of @xmath72 in @xmath47 by @xmath91 assuming that all @xmath86 s are disjoint .
we call @xmath92 a _ graph - composition _ , @xmath72 the _ frame _ and each @xmath93 a _ brick _ of this graph - composition .
let @xmath94 .
if all @xmath93 s are isomorphic to some @xmath95 , then instead of @xmath96 we simply write @xmath97 . for each @xmath98 , let @xmath99 be the graph having exactly two vertices @xmath100 , @xmath101 and exactly three parallel edges connecting @xmath100 and @xmath101 . then @xmath102 . if , in particular , @xmath103 and each @xmath86 is a copy of @xmath33 , then @xmath0 is obtained from @xmath72 by replacing each vertex by a triangle .
let @xmath104 . obviously , there is a unique bijection @xmath105 such that if @xmath106 , then @xmath107 is an edge in @xmath0 having one end - vertex in @xmath108 and the other in @xmath109 .
let @xmath16 be a @xmath1-packing in @xmath0 . for @xmath106 , @xmath110 , we write @xmath111 or simply , @xmath112 , if @xmath16 has a 3-vertex path @xmath13 such that @xmath113 and @xmath114 .
let @xmath115 be the union of components of @xmath16 that meet ( i.e. , have an edge in ) @xmath116 .
+ obviously : [ aasbb ] let @xmath19 be an integer and @xmath117 . if @xmath71 and @xmath72 above are @xmath19-connected , cubic , bipartite , and planar graphs , then @xmath118 is also a @xmath19-connected , cubic , bipartite , and planar graph , respectively . from * [ aasbb ]
* we have : [ a(b ) ] let @xmath19 be an integer and @xmath117 .
if @xmath72 and each @xmath86 is a @xmath19-connected , cubic , bipartite , and planar graphs , then @xmath96 is also a @xmath19-connected , cubic , bipartite , and planar graph , respectively .
let @xmath119 , @xmath120 , @xmath121 be three disjoint graphs , @xmath122 , and @xmath123 , where @xmath81 .
let @xmath124 denote the graph obtained from @xmath125 by adding three new vertices @xmath126 , @xmath127 , @xmath128 and the set of nine new edges @xmath129 ( see fig .
[ fy ] ) . ] in other words , if @xmath130 is the complete @xmath131-bipartite graph with @xmath132 and @xmath133 , then @xmath134 is obtained from @xmath72 by replacing each vertex @xmath135 in @xmath47 by @xmath136 so that @xmath137 .
let @xmath138 .
if @xmath16 is a @xmath1-packing of @xmath134 , then let @xmath139 be the union of components of @xmath16 meeting @xmath140 and @xmath141 , @xmath81 .
if each @xmath142 is a copy of the same @xmath143 , then we write @xmath144 instead of @xmath145 .
+ in particular , from * [ a(b ) ] * we have : [ y(g1,g2,g3 ) ] let @xmath19 be an integer and @xmath117 .
if each @xmath146 above is a @xmath19-connected , cubic , and bipartite graph , then @xmath147 _ ( see fig . [ fy ] ) _ is also a @xmath19-connected , cubic , and bipartite graph , respectively .
we will also use the following simple observation .
[ 3cut ] let @xmath71 and @xmath72 be disjoint graphs , @xmath73 , @xmath78 , @xmath74 , @xmath79 , and @xmath148 , where each @xmath80 _ ( see fig .
[ faasbb])_. let @xmath16 be a @xmath1-factor of @xmath0 @xmath28and so @xmath6@xmath29 and @xmath149 be the @xmath1-packing of @xmath0 consisting of the components @xmath283-vertex paths@xmath29 of @xmath16 that meet @xmath150 .
+ @xmath151 suppose that @xmath152 , and so @xmath153 .
then one of the following holds _ ( see fig .
[ a1])_@xmath154 + @xmath155 @xmath149 has exactly one component that has two vertices in @xmath156 , that are adjacent @xmath28and , accordingly , exactly one vertex in @xmath157@xmath29 , + @xmath158 @xmath149 has exactly two components and each component has exactly one vertex in @xmath156 @xmath28and , accordingly , exactly two vertices in @xmath157 , that are adjacent@xmath29 , + @xmath159 @xmath149 has exactly three components @xmath160 , @xmath161 , @xmath162 and one of them , say @xmath160 , has exactly one vertex in @xmath156 and each of the other two @xmath161 and @xmath162 has exactly two vertices in @xmath156 , that are adjacent @xmath28and , accordingly , @xmath160 has exactly two vertices in @xmath157 , that are adjacent , and each of the other two @xmath161 and @xmath162 has exactly one vertex in @xmath157@xmath29 .
+ @xmath163 suppose that @xmath164 , and so @xmath165 .
then one of the following holds _ ( see fig .
[ a2 ] ) _ : + @xmath166 @xmath167 , + @xmath168 @xmath149 has exactly two components , say @xmath160 and @xmath161 , such that one of the them , say @xmath160 , has exactly one vertex in @xmath156 and exactly two vertices in @xmath157 , that are adjacent , and the other component @xmath161 has exactly two vertices in @xmath156 , that are adjacent , and exactly one vertex in @xmath157 , + @xmath169 @xmath149 has exactly three components @xmath160 , @xmath161 , @xmath162 and either each @xmath170 has exactly one vertex in @xmath156 @xmath28and , accordingly , has exactly two vertices in @xmath157 , that are adjacent@xmath29 or each @xmath170 has exactly two vertices in @xmath156 , that are adjacent @xmath28and , accordingly , has exactly one vertex in @xmath157@xmath29 .
the main goal of this section is to prove the following theorem showing that claim @xmath4 in * [ pr3con ] * is equivalent to various seemingly stronger claims .
[ 3-con ] the following claims are equivalent for cubic 3-connected graphs @xmath0 : + @xmath171 @xmath172 @xmath173 @xmath0 has a @xmath1-factor , + @xmath174 @xmath172 @xmath173 for every @xmath7 there is a @xmath1-factor of @xmath0 avoiding @xmath8 @xmath28i.e . , @xmath175 has a @xmath1-factor@xmath29 , + @xmath176 @xmath172 @xmath173 for every @xmath7 there exists a @xmath1-factor of @xmath0 containing @xmath8 , + @xmath177 @xmath172 @xmath173 for every @xmath57 there exists at least one 3-vertex path @xmath13 such that @xmath13 is centered at @xmath59 and @xmath12 has a @xmath1-factor , + @xmath178 @xmath172 @xmath173 for every @xmath57 there exist at least two 3-vertex paths @xmath13 such that @xmath13 is centered at @xmath59 and @xmath12 has a @xmath1-factor , + @xmath179 @xmath172 @xmath173 for every @xmath180 there exist edges @xmath181 such that @xmath182 and @xmath183 have @xmath1-factors , + @xmath184 @xmath172 @xmath173 @xmath9 has a @xmath1-factor for every @xmath10 such that @xmath11 , + @xmath185 @xmath172 @xmath173 @xmath12 has a @xmath1-factor for every 3-vertex path @xmath13 in @xmath0 , + @xmath186 @xmath172 @xmath173 for every 3-edge cut @xmath67 of @xmath0 and @xmath187 , @xmath188 , there exists a @xmath1-factor @xmath16 of @xmath0 such that @xmath189 , + @xmath190 @xmath191 @xmath173 for every @xmath57 there exists @xmath180 such that @xmath192 has a @xmath1-factor , + @xmath193 @xmath191 @xmath173 @xmath192 has a @xmath1-factor for every @xmath180 , + @xmath194 @xmath191 @xmath173 for every @xmath57 there exists a 5-vertex path @xmath195 such that @xmath59 is the center vertex of @xmath195 and @xmath196 has a @xmath1-factor _ ( see also * [ 3cut ] * @xmath158 and fig . [ a1 ] ) _ , + @xmath197 @xmath191 @xmath173 for every @xmath57 and @xmath180 there exists a 5-vertex path @xmath195 such that @xmath59 is the center vertex of @xmath195 , @xmath198 , and @xmath196 has a @xmath1-factor _ ( see also * [ 3cut ] * @xmath158 and fig .
[ a1 ] ) _ , + @xmath199 @xmath200 @xmath173 @xmath201 has a @xmath1-factor for every @xmath57 , + @xmath202 @xmath200 @xmath173 @xmath203 has a @xmath1-factor for every @xmath57 and @xmath7 , + @xmath204 @xmath200 @xmath173 for every @xmath57 there exists a 4-vertex path @xmath205 such that @xmath59 is an inner vertex of @xmath205 and @xmath206 has a @xmath1-factor _ ( see also * [ 3cut ] * @xmath168 and fig .
[ a2 ] ) _ , + @xmath207 @xmath200 @xmath173 for every @xmath57 there exists @xmath180 and a 4-vertex path @xmath205 such that @xmath59 is an inner vertex of @xmath205 , @xmath208 , and @xmath206 has a @xmath1-factor _ ( see also * [ 3cut ] * @xmath168 and fig .
[ a2 ] ) _ , + @xmath209 @xmath200 @xmath173 for every @xmath180 there exists a 4-vertex path @xmath205 such that @xmath210 is the middle edge of @xmath205 and @xmath206 has a @xmath1-factor _ ( see also * [ 3cut ] * @xmath168 and fig .
[ a2 ] ) _ , + @xmath211 @xmath200 @xmath173 for every @xmath212 and every 3-vertex path @xmath213 there exists a 4-vertex path @xmath205 such that @xmath214 , @xmath215 is an end - vertex of @xmath205 , and @xmath206 has a @xmath1-factor _ ( see also * [ 3cut ] * @xmath168 and fig .
[ a2])_. theorem * [ 3-con ] * follows from * [ z1ottoz2 ] * * [ z9ottoz1 ] * below .
in @xcite we have shown that claims @xmath216 - @xmath217 are true for cubic , 3-connected , and claw - free graphs .
in @xcite we have proved the following results related to @xmath218 , @xmath219 , and @xmath220 - @xmath221 .
[ pi , w , y ] let @xmath0 be a cubic 2-connected graph .
then + @xmath151 if @xmath201 has a @xmath1-factor for some vertex @xmath59 in @xmath0 _ ( and so @xmath200 ) _ , then there exists a 4-vertex path @xmath222 in @xmath0 such that @xmath223 has a @xmath1-factor , + @xmath163 if @xmath224 has a @xmath1-factor for some edge @xmath210 in @xmath0 _ ( and so @xmath191 ) _ , then there exists a 5-vertex path @xmath195 and two disjoint 4-vertex paths @xmath225 and @xmath226 in @xmath0 such that @xmath196 and @xmath227 have @xmath1-factors , and + @xmath228
if @xmath200 and @xmath0 has a hamiltonian cycle , then there exists a claw @xmath229 in @xmath0 such that @xmath230 has a @xmath1-factor . from claim @xmath228 in *
[ pi , w , y ] * it follows that if @xmath0 is a cubic hamiltonian graph with @xmath200 , then its domination number is at most @xmath231 . + the remarks below show that if claims @xmath232 , @xmath233 , @xmath234 - @xmath219 , and @xmath235 - @xmath221 in * [ 3-con ] * are true , then they are best possible in some sense .
+ @xmath236 obviously claim @xmath232 is not true if condition `` @xmath11 '' is replaced by condition `` @xmath237 '' .
namely , if @xmath0 is a cubic 3-connected graph , @xmath172 , @xmath47 is a 3-edge cut in @xmath0 , and the two components of @xmath9 have the number of vertices not divisible by 3 , then clearly @xmath9 has no @xmath1-factor .
moreover , for every 3-edge graph @xmath205 with no isolated vertices and distinct from @xmath22 , there exist infinitely many pairs @xmath238 such that @xmath0 is a cubic 3-connected graph , @xmath239 , @xmath237 , @xmath240 $ ] ( the subgraph of @xmath0 induced by @xmath47 is isomorphic to @xmath205 , and @xmath9 is connected and has no @xmath1-factor .
also in section [ almostcubic ] we describe an infinite set of cubic 3-connected graphs @xmath0 such that such that @xmath241 has no @xmath1-factor for every triangle @xmath242 in @xmath0 .
+ @xmath243 there exist infinitely many triples @xmath244 such that @xmath0 is a cubic , 3-connected , bipartite , and planar graph , @xmath172 , @xmath13 is a 3-vertex path in @xmath0 , @xmath245 , and @xmath246 has no @xmath1-factor , and so claim @xmath233 is tight .
indeed , let @xmath247 , @xmath248 , and @xmath249 , and so @xmath250 is a triangle .
since @xmath0 is 3-connected , @xmath59 is not adjacent to @xmath251 .
let @xmath13 be a 3-vertex path in @xmath252 containing @xmath253 .
then @xmath59 is an isolated vertex in @xmath246 , and therefore @xmath246 has no @xmath1-factor .
similarly , let @xmath254 be a 4-cycle in @xmath0 , @xmath255 be an edge in a vertex in @xmath256 , and @xmath257 be a 3-vertex path in @xmath0 . then again @xmath59 is an isolated vertex in @xmath246 , and therefore @xmath246 has no @xmath1-factor .
+ @xmath258 the arguments in @xmath243 also show that there exist infinitely many triples @xmath259 such that @xmath260 has no @xmath1-factor , where @xmath13 and @xmath261 are disjoint 3-vertex paths in @xmath0 , and @xmath0 is a cubic , 3-connected , bipartite , and planar graph , and so claim @xmath233 is tight in this sense as well .
indeed , they can be obtained from the constructions in @xmath243 by replacing vertex @xmath59 by a 3-vertex path in @xmath12 containing @xmath59 . in section
[ cycl6-connected ] ( see * [ g , l , l ] * ) we give a construction providing infinitely many cubic cyclically 6-connected graphs @xmath0 such that on the one hand , @xmath262 has a @xmath1-factor for every 3-vertex path @xmath13 in @xmath0 and every edge in @xmath12 , and on the other hand , @xmath0 has pairs of disjoint 3-vertex paths @xmath13 and @xmath261 such that @xmath260 has no @xmath1-factor .
+ @xmath263 there exist infinitely many triples @xmath264 such that @xmath0 is a cubic , 3-connected , bipartite , and planar graph , @xmath191 , @xmath180 , @xmath265 , and @xmath266 has no @xmath1-factor , and so claim @xmath234 is tight .
+ @xmath267 there exist infinitely many triples @xmath268 such that @xmath0 is a cubic 3-connected graph , @xmath191 , @xmath269 , @xmath270 , @xmath271 , and @xmath272 has no @xmath1-factor , and so claim @xmath234 is not true if vertices @xmath59 and @xmath273 are not adjacent .
+ @xmath274 there exist infinitely many quadruples @xmath275 such that @xmath0 is a cubic 3-connected graph with no 3-cycles and no 4-cycles , @xmath200 , @xmath57 , @xmath276 and @xmath82 are non - adjacent edges in @xmath201 , and @xmath277 has no @xmath1-factor , and so claim @xmath278 is tight .
+ @xmath279 there exist infinitely many triples @xmath280 such that @xmath0 is a cubic 3-connected graph with no 3-cycles and no 4-cycles , @xmath200 , @xmath57 , @xmath13 is a 3-vertex path in @xmath201 , and @xmath281 has no @xmath1-factor ( see claim @xmath235 ) .
+ @xmath282 there exist infinitely many pairs @xmath283 such that @xmath0 is a cubic 3-connected graph , @xmath200 , @xmath284 is a 4-vertex path in @xmath0 , and @xmath285 has no @xmath1-factor ( see claims @xmath220 - @xmath221 ) .
+ @xmath286 there exist infinitely many triples @xmath287 such that @xmath0 is a cubic 3-connected graph , @xmath191 , @xmath195 is a 5-vertex path in @xmath0 , and @xmath196 has no @xmath1-factor ( see claims @xmath218 and @xmath219 ) . + the above discussion in @xmath243 leads to the following question . is the following claim true ?
suppose that @xmath0 is a cubic 3-connected graph with no triangles and no squares @xmath28or moreover , cyclically 5-connected@xmath29 , @xmath8 is an edge in @xmath0 , and @xmath13 is a 3-vertex path in @xmath175 .
then @xmath288 has a @xmath1-factor .
we need the following two results obtained before .
[ y , cmp(pi)<3 ] let @xmath289 _ ( see fig .
[ fy ] ) _ and @xmath16 be a @xmath1-factor of @xmath0 .
suppose that each @xmath146 is a cubic graph and @xmath290 .
then @xmath291 for every @xmath81 . *
* let @xmath81 . since @xmath140 is a matching and @xmath292 consists of the components of @xmath16 meeting @xmath140 , clearly @xmath293 .
since @xmath294 , we have @xmath295 .
it remains to show that @xmath296 .
suppose , on the contrary , that @xmath297 . since @xmath16 is a @xmath1-factor of @xmath0 and @xmath298 , clearly @xmath299 and we can assume ( because of symmetry ) that @xmath300 consists of three components @xmath301 , @xmath302 , and @xmath303 for some @xmath304 then @xmath305 , a contradiction .
@xmath306 [ y , e- ] let @xmath71 be a graph , @xmath307 , and @xmath308 _ ( see fig . [
fy])_. suppose that + @xmath309 @xmath71 is cubic , + @xmath310 @xmath311 , and + @xmath312 @xmath276 has no @xmath1-factor containing @xmath313 .
+ then @xmath172 and @xmath0 has no @xmath1-factor .
( uses * [ y , cmp(pi)<3 ] * ) .
suppose , on the contrary , that @xmath0 has a @xmath1-factor @xmath16 .
by definition of @xmath308 , each @xmath146 is a copy of @xmath71 and edge @xmath314 in @xmath146 is a copy of edge @xmath313 in @xmath71 by * [ y , cmp(pi)<3 ] * , @xmath315 for every @xmath81 .
since @xmath16 is a @xmath1-factor of @xmath0 and @xmath316 , clearly @xmath317 is an edge subset of a @xmath1factor of @xmath146 for every @xmath81 ( we assume that edge @xmath318 in @xmath0 is edge @xmath319 in @xmath146 ) . since @xmath320 does not belong to any @xmath1-factor of @xmath146 for every @xmath81 , clearly @xmath321 .
therefore @xmath322 , and so @xmath16 is not a @xmath1-factor of @xmath0 , a contradiction . @xmath306
[ z1ottoz2 ] @xmath216 @xmath323 @xmath324 .
( uses * [ y(g1,g2,g3 ) ] * and * [ y , cmp(pi)<3 ] * ) . obviously , @xmath216 @xmath325 @xmath324 .
we prove @xmath216 @xmath173 @xmath324 .
suppose , on the contrary , that @xmath216 is true but @xmath324 is not true , i.e. , there is a cubic 3-connected graph @xmath71 and @xmath326 such that @xmath327 and every @xmath1-factor of @xmath0 contains @xmath328 .
let @xmath329 , where each @xmath142 is a copy of @xmath143 and edge @xmath330 in @xmath146 is a copy of edge @xmath328 in @xmath71 ( see fig .
since @xmath71 is cubic and 3-connected , by * [ y(g1,g2,g3 ) ] * , @xmath0 is also cubic and 3-connected . obviously , @xmath331 .
by @xmath216 , @xmath0 has a @xmath1-factor @xmath16 .
since each @xmath332 , by * [ y , cmp(pi)<3 ] * , @xmath333 for every @xmath81 . since @xmath16 is a @xmath1-factor of @xmath0 and @xmath316 , clearly @xmath317 is an edge subset of a @xmath1-factor of @xmath146 for every @xmath81 ( we assume that edge @xmath318 in @xmath0 is edge @xmath319 in @xmath146 ) . since @xmath320 belongs to every @xmath1-factor of @xmath146 for every @xmath81 , clearly @xmath334 .
therefore vertex @xmath126 has degree three in @xmath16 , and so @xmath16 is not a @xmath1-factor of @xmath0 , a contradiction . @xmath306
[ z1ottoz3 ] @xmath216 @xmath323 @xmath335 .
claim @xmath336 is obvious .
claim @xmath337 follows from * [ y , e-]*. @xmath306 [ z1ottoz4 ] @xmath216 @xmath323 @xmath338 .
( uses * [ a(b ) ] * and * [ 3cut ] * @xmath151 ) . obviously , @xmath339 .
we prove @xmath340 .
suppose that @xmath216 is true but @xmath338 is not true , i.e. , there is a cubic 3-connected graph @xmath71 and @xmath341 such that @xmath327 and @xmath276 has degree one in every @xmath1-factor of @xmath71 .
let @xmath0 be the graph obtained from @xmath130 by replacing each vertex @xmath24 of @xmath72 by a copy @xmath109 of @xmath156 ( see fig .
[ fprz1toz4 ] ) .
obviously , @xmath172 and by * [ a(b ) ] * , @xmath0 is a cubic 3-connected graph . by @xmath216 , @xmath0 has a @xmath1-factor @xmath16 .
if @xmath342 , then let @xmath343 denote the component of @xmath16 containing @xmath344 .
let @xmath345 .
since vertex @xmath276 has degree one in every @xmath1-factor of @xmath71 and each @xmath93 is a copy of @xmath346 , by * [ 3cut ] * @xmath151 , we have @xmath347 .
+ @xmath348 suppose that there is @xmath349 such that @xmath350 .
by symmetry of @xmath72 , we can assume that @xmath351 and , by * [ 3cut ] * @xmath159 , @xmath352 , @xmath353 , and @xmath354 ( see fig . [ fprz1toz4 ] ) .
@xmath173 @xmath338 ] let @xmath356 .
since @xmath357 and @xmath358 , clearly @xmath359 and , by * [ 3cut ] * @xmath159 , @xmath360 , @xmath361 . then by * [ 3cut ] * @xmath158 , @xmath362 for @xmath363 , a contradiction .
+ @xmath364 now suppose that @xmath365 for every @xmath349 . by symmetry
, we can assume @xmath366 .
then @xmath367 contradicts * [ 3cut ] * @xmath155 .
@xmath306[z1ottot1 ] @xmath368 .
( uses * [ y(g1,g2,g3 ) ] * ) . if @xmath369 ( see fig . [ fy ] ) and each @xmath146 is cubic and 3-connected , then by * [ y(g1,g2,g3 ) ] * , @xmath0 is also cubic and 3-connected .
+ @xmath348 we first prove @xmath370 .
suppose , on the contrary , that @xmath216 is true but @xmath371 is not true , i.e. , there is a cubic 3-connected graph @xmath71 and @xmath73 such that @xmath372 and @xmath373 has no @xmath1-factor for every vertex @xmath273 in @xmath71 adjacent to @xmath276 .
let each @xmath142 above be a copy of @xmath143 , and so @xmath374 has no @xmath1-factor for every @xmath375 .
obviously @xmath172 . by @xmath216
, @xmath0 has a @xmath1-factor @xmath16 .
then it is easy to see that since each @xmath376 , there are @xmath377 such that @xmath378 and @xmath379 . since @xmath16 is a @xmath1-factor of @xmath0 , clearly @xmath380 is a @xmath1-factor of @xmath381 , a contradiction .
+ @xmath364 now we prove @xmath382 .
suppose , on the contrary , that @xmath371 is true but @xmath216 is not true , i.e. , there is a cubic 3-connected graph @xmath71 such that @xmath327 and @xmath71 has no @xmath1-factor .
let @xmath383 be a copy of @xmath346 for @xmath384 , where @xmath73 , and @xmath385 be a copy of @xmath386 for some cubic 3-connected graph @xmath15 and @xmath387 , where @xmath388 .
obviously @xmath191 .
suppose that @xmath16 is a @xmath1-factor of @xmath389
. then @xmath390 .
since @xmath391 , we have @xmath392 .
now since @xmath16 is a @xmath1factor of @xmath393 and @xmath394 , clearly @xmath395 is an edge subset of a component of a @xmath1-factor of @xmath119 ( we assume that edge @xmath396 in @xmath0 is edge @xmath397 in @xmath119 ) .
therefore @xmath71 has a @xmath1-factor , a contradiction . @xmath306
[ z1ottof1 ] @xmath216 @xmath323 @xmath235 .
( uses * [ y(g1,g2,g3 ) ] * ) . by *
[ y(g1,g2,g3 ) ] * , if @xmath369 ( see fig .
[ fy ] ) and each @xmath146 is cubic and 3-connected , then @xmath0 is also cubic and 3-connected .
+ @xmath348 we first prove @xmath398 .
suppose , on the contrary , that @xmath216 is true but @xmath235 is not true , i.e. , there is a cubic 3-connected graph @xmath71 such that @xmath399 and @xmath156 has no @xmath1-factor for some @xmath73 .
let each @xmath383 above be a copy of @xmath143 .
obviously , @xmath172 .
by @xmath216 , @xmath0 has a @xmath1-factor @xmath16 .
let @xmath292 be the union of components in @xmath16 meeting @xmath140 and @xmath400 . since @xmath16 is a @xmath1-factor of @xmath0 and @xmath401 , clearly each @xmath402 and each @xmath403 . since @xmath136 has no @xmath1-factor , each @xmath404 , and so each @xmath405 .
thus , each @xmath406 , and so each @xmath407 is the center of a 3-vertex path in @xmath16 and each @xmath408 .
now , since @xmath401 , @xmath409 is not a @xmath1-factor of @xmath410 .
therefore @xmath16 is not a @xmath1-factor of @xmath0 , a contradiction .
+ @xmath364 now we prove @xmath411 .
suppose , on the contrary , that @xmath235 is true but @xmath216 is not true , i.e. , there is a cubic 3-connected graph @xmath71 such that @xmath327 and @xmath71 has no @xmath1-factor .
let @xmath383 above be a copy of @xmath143 for @xmath384 and some @xmath73 and @xmath385 is a copy of @xmath386 for some cubic 3-connected graph @xmath15 and @xmath387 , where @xmath412 ( see fig .
then @xmath200 . let @xmath413 .
suppose that @xmath414 has a @xmath1-factor @xmath16 . since @xmath71 has no @xmath1-factor , clearly @xmath415 .
then @xmath16 is not a @xmath1-factor of @xmath201 , and so @xmath235 is not true , a contradiction . @xmath306
[ z4ottot2 ] @xmath338 @xmath323 @xmath234 .
( uses * [ y(g1,g2,g3 ) ] * , * [ z1ottoz4 ] * , and * [ z1ottot1 ] * ) .
let @xmath369 ( see fig .
[ fy ] ) . by *
[ y(g1,g2,g3 ) ] * , if each @xmath146 is cubic and 3-connected , then @xmath0 is also cubic and 3-connected .
+ @xmath348 we first prove @xmath416 .
obviously @xmath417 .
by * [ z1ottoz4 ] * ,
@xmath216 @xmath323 @xmath338 . by * [ z1ottot1 ] * , @xmath368 .
the result follows .
+ @xmath364 now we prove @xmath418 .
suppose , on the contrary , that @xmath338 is true but @xmath234 is not true .
then there exists a cubic 3-connected graph @xmath71 and @xmath326 such that @xmath372 and @xmath419 has no @xmath1-factor .
let each @xmath142 above be a copy of @xmath346 and edge @xmath330 in @xmath146 be a copy of edge @xmath328 in @xmath0 .
obviously @xmath172 .
let @xmath420 , where @xmath421 .
by @xmath338 , @xmath0 has a @xmath1-factor @xmath16 containing @xmath422 for some @xmath81 , say for @xmath423 .
if @xmath424 , then @xmath425 , because @xmath426 has no @xmath1-factor and @xmath427 .
also @xmath428 , because @xmath429 .
then @xmath430 has at least four components each meeting @xmath431 , a contradiction . @xmath306
[ z2toz5 ] @xmath324 @xmath173 @xmath217 .
( uses * [ aasbb ] * and * [ 3cut ] * @xmath151 ) .
suppose , on the contrary , that @xmath324 is true but @xmath217 is not true .
then there exists a cubic 3-connected graph @xmath71 and @xmath73 such that at most one 3-vertex path , centered at @xmath276 belongs to a @xmath1-factor of @xmath71 .
it is sufficient to prove our claim in case when @xmath71 has exactly one 3-vertex path , say @xmath432 , centered at @xmath276 and belonging to a @xmath1-factor of @xmath71 .
let @xmath433 , and so @xmath434 .
let @xmath72 be the graph - skeleton of the three - prism : @xmath345 and @xmath72 is obtained from two disjoint triangles @xmath435 and @xmath436 by adding three new edges @xmath437 , @xmath438 , and @xmath439 .
let each @xmath440 , @xmath349 be a copy of @xmath441 , and so edge @xmath442 in @xmath86 is a copy of edge @xmath433 in @xmath71 , @xmath384 .
we also assume that all @xmath86 s are disjoint .
let @xmath0 be a graph obtained from @xmath72 by replacing each @xmath349 by @xmath443 ( see fig .
[ fprz2toz5 ] ) . @xmath173 @xmath217 ] we recall the following notation from section [ constructions ] : * @xmath445 , * @xmath105 is the bijection such that if @xmath106 , then @xmath107 is an edge in @xmath0 having one end - vertex in @xmath108 and the other in @xmath109 , and * if @xmath16 is a @xmath1-packing in @xmath0 , then for @xmath106 , @xmath110 , we write @xmath111 ( or simply , @xmath112 ) , if @xmath16 has a 3-vertex path @xmath13 such that @xmath113 and @xmath114 . given @xmath349 , let @xmath446 be the set of two edges @xmath447 in @xmath448 such that edge @xmath447 is incident to vertex @xmath449 in @xmath0 , @xmath384 .
we assume that each vertex @xmath24 in @xmath72 is replaced by @xmath109 ( to obtain @xmath0 ) in such a way that + @xmath450 , @xmath451 , @xmath452 , + @xmath453 , @xmath454 , @xmath455 . + in figure [ fprz2toz5 ]
the edges in @xmath456 are marked for every @xmath349 .
+ by * [ aasbb ] * , @xmath0 is a cubic 3-connected graph . since @xmath457 , clearly also @xmath458 . by @xmath324
, @xmath459 has a @xmath1-factor @xmath16 .
we know that @xmath71 has exactly one 3-vertex path @xmath432 centered at @xmath276 and belonging to a @xmath1-factor of @xmath71 and that each @xmath440 is a copy of @xmath460 , and so @xmath461 .
therefore , by * [ 3cut ] * @xmath151 and the assumption on @xmath71 , the @xmath1-factor @xmath16 satisfies the following condition for every @xmath349 : + @xmath462 if @xmath463 , then @xmath464 and @xmath465 , where @xmath466 .
+ obviously , @xmath467 in @xmath468 .
therefore @xmath469 and @xmath470 .
since @xmath471 , by conditions @xmath472 and @xmath473 , @xmath474 . now by * [ 3cut ] * @xmath151 , @xmath475 for some @xmath476 and @xmath477 for some @xmath478 .
+ @xmath348 suppose that @xmath479 .
assume first that @xmath480
. then @xmath390 . by *
[ 3cut ] * @xmath151 , @xmath481 .
this contradicts @xmath482 .
thus we must have @xmath483 .
+ @xmath484 suppose that @xmath485 .
suppose that @xmath486 . then by * [ 3cut ] * @xmath151 , @xmath487 and @xmath488 .
this contradicts @xmath489 .
now suppose that @xmath490 .
this contradicts @xmath491 .
+ @xmath492 suppose that @xmath493
. then @xmath494 .
suppose that @xmath486 . by * [ 3cut ] * @xmath155 , @xmath495 .
then @xmath488 , which contradicts @xmath489 . now suppose that @xmath490
. then @xmath496 , which contradicts @xmath491 .
+ @xmath364 now suppose that @xmath497 .
then @xmath498 , and therefore @xmath390 .
by @xmath499 and * [ 3cut ] * @xmath159 , @xmath366 ( and @xmath488 ) . then by * [ 3cut ] * @xmath158 , @xmath481 , which contradicts @xmath482 .
@xmath306 + since @xmath232 @xmath173 @xmath324 , we obtain from * [ z2toz5 ] * : [ z7toz5 ] @xmath232 @xmath173 @xmath217 .
[ z1ottoz6 ] @xmath216 @xmath323 @xmath500 .
* proof*. obviously , @xmath216 @xmath325 @xmath500 and @xmath501 @xmath500 . as to @xmath216 @xmath173 @xmath500
, it follows from * [ z1ottoz2 ] * and * [ z2toz5]*. @xmath306 [ z7toz8 ] @xmath232 @xmath173 @xmath233 .
* proof 1*. suppose , on the contrary , that @xmath232 is true but @xmath233 is not true .
then there exists a cubic 3-connected graph @xmath71 and a 3-path @xmath502 in @xmath71 such that @xmath311 and @xmath503 has no @xmath1-factor .
let @xmath504 .
let @xmath505 , @xmath384 , be two copies of @xmath506 and @xmath119 , @xmath120 be disjoint graphs , and so @xmath507 in @xmath146 is a copy of @xmath432 in @xmath71 .
let @xmath508 , where @xmath509 is a bijection such that @xmath510 for @xmath511 .
let @xmath0 be the graph obtained from @xmath15 by subdividing edge @xmath512 by a new vertex @xmath513 for @xmath514 and by adding a new edge @xmath515 . obviously , @xmath0 is a cubic 3-connected graph and @xmath172 .
by @xmath232 , @xmath516 has a @xmath1-factor @xmath16 .
since @xmath517 , @xmath16 has a 3-vertex path @xmath518 . by symmetry of @xmath0
, we can assume that @xmath519 .
since @xmath298 , @xmath16 has a 3-vertex path @xmath520 for some @xmath521 .
then @xmath522 is a @xmath1-factor of @xmath523 , a contradiction . @xmath306 + * proof 2 * ( uses * [ 3cut ] * @xmath151 ) .
suppose , on the contrary , that @xmath232 is true but @xmath233 is not true .
then there is a cubic 3-connected graph @xmath71 and a 3-path @xmath502 in @xmath71 such that @xmath311 and @xmath503 has no @xmath1-factor .
let @xmath524 .
let @xmath72 , @xmath525 , and @xmath0 be as in * [ z2toz5 ] * ( see fig .
[ fprz7toz8 ] ) . given @xmath349 ,
let @xmath446 be the set of two edges @xmath447 in @xmath448 such that edge @xmath447 is incident to vertex @xmath449 in @xmath0 , @xmath384 .
we assume that each vertex @xmath24 in @xmath72 is replaced by @xmath86 ( to obtain @xmath0 ) in such a way that + @xmath526 , @xmath527 , @xmath528 , + @xmath453 , @xmath454 , @xmath529 .
+ in figure [ fprz7toz8 ] the edges in @xmath456 are marked for every @xmath349 .
@xmath173 @xmath233 ] + since @xmath458 and @xmath0 is cubic and 3-connected , by @xmath232 , @xmath531 has a @xmath1-factor @xmath16 . by * [ 3cut ] * @xmath151 and the lack of @xmath1-factor in each of @xmath532 , the @xmath1-factor @xmath16 satisfies the following condition for every @xmath349 : + @xmath462 if @xmath463 , then @xmath464 and @xmath465 , where @xmath533 .
+ obviously , @xmath534 in @xmath531 , and so @xmath296 for @xmath535 . since @xmath536 and @xmath537 , by conditions @xmath482 and @xmath491 we have @xmath538 .
now by * [ 3cut ] * @xmath155 , @xmath354 and @xmath539 .
this contradicts @xmath499 . @xmath306
[ t2ottoz7 ] @xmath234 @xmath323 @xmath232 .
( uses * [ z1ottoz4 ] * and * [ z4ottot2 ] * ) .
we first prove @xmath540 .
let @xmath0 be a cubic , 3-connected graph with @xmath172 and @xmath541 , @xmath542 be two distinct edges of @xmath0 .
let @xmath543 be the graph obtained from @xmath0 as follows : subdivide edge @xmath544 by a new vertex @xmath545 and edge @xmath546 by a new vertex @xmath547 and add a new edge @xmath548 .
then @xmath543 is a cubic and 3-connected graph , @xmath549 , and @xmath550 .
by @xmath234 , @xmath551 has a @xmath1-factor .
now we prove @xmath552 . obviously , @xmath553 . by * [ z1ottoz4 ] * , @xmath340 and by * [ z4ottot2 ] * , @xmath418 .
implication @xmath552 also follows from obvious @xmath554 , from @xmath555 , ( * [ z7toz8 ] * ) , and from @xmath418 ( * [ z4ottot2 ] * ) .
@xmath306 + here is a direct proof of @xmath556 .
[ z7tot2 ] @xmath556
. * proof .
* let @xmath0 be a cubic , 3-connected graph , @xmath557 , @xmath180 , @xmath558 , and @xmath559 .
let @xmath560 , @xmath561 , and @xmath562 , where @xmath563 , @xmath564 , and @xmath565 . obviously , each @xmath566 is a cubic graph and @xmath567 .
it is easy to see that since @xmath0 is 3-connected , there is @xmath568 such that @xmath569 is 3connected .
clearly @xmath570 .
by @xmath232 , @xmath571 has a @xmath1-factor .
@xmath306 [ z8ottoz7 ] @xmath233 @xmath323 @xmath232 . * proof .
* obviously , @xmath572 . by * [ z4ottot2 ] * , @xmath418 and by * [ t2ottoz7 ] * , @xmath540 . therefore
@xmath233 @xmath173 @xmath232 . by * [ z7toz8 ] * , @xmath232 @xmath173 @xmath233 .
@xmath306 [ z1ottoz8 ] @xmath216 @xmath323 @xmath233 . * proof . * obviously , @xmath233
@xmath173 @xmath216 . by * [ z1ottoz4 ]
@xmath216 @xmath173 @xmath338 . by * [ z4ottot2 ] * ,
@xmath338 @xmath173 @xmath234 . by * [ t2ottoz7 ] * ,
@xmath234 @xmath173 @xmath232 . by * [ z7toz8 ] * , @xmath232 @xmath173 @xmath233 . therefore @xmath216 @xmath173 @xmath233 .
@xmath306 [ z8tof1 ]
@xmath233 @xmath173 @xmath235 . * proof . *
let @xmath0 be a cubic 3-connected graph , @xmath200 , @xmath57 , and @xmath573 .
let @xmath543 be the graph obtained from @xmath0 by replacing @xmath59 by a triangle @xmath242 with @xmath574 so that @xmath575 , @xmath81 .
since @xmath200 , clearly @xmath576 .
consider the 3-vertex path @xmath577 in @xmath543 .
by @xmath233 , @xmath578 has a @xmath1-factor @xmath149 .
obviously , @xmath579 is a @xmath1-factor of @xmath578 and @xmath580 . @xmath306
[ z8tof2 ] @xmath233 @xmath173 @xmath278 .
( uses * [ z8tof1 ] * ) .
let @xmath0 be a cubic 3-connected graph , @xmath200 , @xmath57 , and @xmath581 .
we want to prove that if @xmath233 is true , then @xmath582 has a @xmath1-factor . if @xmath583 , then @xmath584 , and therefore , by * [ z8tof1 ] * , our claim is true .
so we assume that @xmath585 .
let @xmath586 .
let @xmath543 be the graph obtained from @xmath0 by subdividing edge @xmath587 by a vertex @xmath273 and edge @xmath588 by a vertex @xmath215 and by adding a new edge @xmath253 ( see fig .
[ fprz8tof2 ] ) . @xmath173 @xmath278 ] since @xmath200 , clearly @xmath576 .
since @xmath590 and @xmath0 is cubic and 3-connected , @xmath543 is also cubic and 3-connected .
obviously , @xmath591 is a 3-vertex path in @xmath543 and @xmath592 .
by @xmath233 , @xmath593 has a @xmath1-factor . @xmath306 [ f2tot4 ] @xmath278 @xmath173 @xmath219 .
* proof . *
let @xmath0 be a cubic , 3-connected graph , @xmath557 , @xmath57 , and @xmath573 .
let @xmath543 be the graph obtained from @xmath0 by replacing @xmath59 by a triangle @xmath242 with @xmath594 so that @xmath575 , @xmath81 .
since @xmath191 , clearly @xmath595 . by @xmath278
, @xmath596 has a @xmath1-factor @xmath597 , where @xmath598 . since @xmath599 and @xmath600 are dangling edges in @xmath596 , clearly @xmath601 and @xmath602 .
let @xmath603 and @xmath604 be the components of @xmath597 containing @xmath599 and @xmath600 , respectively
. then @xmath605 induces in @xmath0 a 5-vertex path @xmath606 such that @xmath59 is the center vertex of @xmath606 and @xmath607 . @xmath306
[ t3toz1 ] @xmath218 @xmath173 @xmath216 .
( uses * [ 3cut ] * @xmath159 ) .
suppose , on the contrary , that @xmath218 is true but @xmath216 is not true , i.e. , there is a cubic 3-connected graph @xmath71 such that @xmath327 and @xmath71 has no @xmath1-factor .
let @xmath73 .
let @xmath383 above be a copy of @xmath143 for @xmath384 and let @xmath608 be such that @xmath609 .
let @xmath369 ( see fig .
[ fy ] ) . obviously , @xmath191 . by @xmath218
, @xmath0 has a 5-vertex path @xmath195 such that @xmath128 is the center vertex of @xmath195 and @xmath196 has a @xmath1-factor @xmath16 . obviously , @xmath610 for some @xmath81 .
suppose that @xmath611 .
then @xmath195 has an end - edge in @xmath612 and in @xmath613 . since @xmath71 has no @xmath1-factor , by * [ 3cut ] * @xmath159 , @xmath614 for @xmath384 .
then @xmath16 is not a @xmath1-factor of @xmath196 , a contradiction .
now suppose that @xmath615 . by symmetry
, we can assume that @xmath616 .
then @xmath195 has an end - edge in @xmath612 and in @xmath617 . then by * [ 3cut ] * @xmath159 , @xmath618 , where @xmath160 has an end - vertex in @xmath612 and @xmath161 has an end - edge in @xmath612 . by symmetry
, we can assume that @xmath619 for @xmath384 .
then @xmath620 , where @xmath621 and @xmath127 is of degree one in @xmath16 . then
, taking into account the values of @xmath622 @xmath81 , @xmath16 is not a @xmath1-factor of @xmath196 , a contradiction . @xmath306
[ z8tof4 ] @xmath233 @xmath173 @xmath623 . * proof .
* let @xmath0 be a cubic 3-connected graph , @xmath200 , @xmath57 , and @xmath573 .
let @xmath543 be the graph obtained from @xmath0 by replacing @xmath59 by a triangle @xmath624 with @xmath625 , so that @xmath575 , @xmath81 .
since @xmath200 , clearly @xmath576 .
since @xmath0 is cubic and 3-connected , @xmath543 is also cubic and 3-connected . consider 3-vertex path @xmath626 in @xmath543 , where @xmath598 .
by @xmath233 , @xmath627 has a @xmath1-factor @xmath628 . since @xmath629 is a dangling edge in @xmath627 , clearly @xmath630 and @xmath631 .
let @xmath632 be the 3-vertex path in @xmath628 containing @xmath633 , say @xmath634 .
consider in @xmath0 the 4-vertex path @xmath635 .
then @xmath59 is an inner vertex of @xmath636 , @xmath637 , and @xmath638 is a @xmath1-factor in @xmath639 .
@xmath306 + let @xmath640 be a tree such that @xmath641 and @xmath642 .
let @xmath643 , @xmath81 be three disjoint copies of @xmath640 with @xmath644 .
let @xmath15 be the graph obtained from these three copies by identifying for every @xmath645 three vertices @xmath646 , @xmath647 , @xmath648 with a new vertex @xmath649 .
let @xmath146 , @xmath650 , be a cubic graph , @xmath122 , and let @xmath651 be the graph obtained from @xmath15 by replacing each @xmath649 by @xmath652 assuming that all @xmath146 s are disjoint .
[ f3toz1 ] @xmath220 @xmath173 @xmath216 .
( uses * [ a(b ) ] * ) .
suppose , on the contrary , that @xmath220 is true but @xmath216 is not true , i.e. , there exists a cubic 3-connected graph @xmath71 such that @xmath327 and @xmath71 has no @xmath1-factor .
let @xmath73 .
let @xmath653 be any cubic 3-connected graph with @xmath654 and each @xmath383 , @xmath81 , be a copy of @xmath346 .
let the graph - composition @xmath651 and its frame @xmath15 be as defined above .
then @xmath15 is a cubic 3-connected graph .
since each brick @xmath146 is cubic and 3-connected , by * [ a(b ) ] * , @xmath0 is also cubic and 3-connected .
obviously , @xmath200 .
it is easy to see that @xmath206 has no @xmath1-factor for every 4-vertex path @xmath205 in @xmath0 such that @xmath655 is an inner vertex of @xmath205 .
this contradicts @xmath220 .
@xmath306 + obviously , @xmath623 @xmath173 @xmath220 .
therefore from * [ f3toz1 ] * we have @xmath623 @xmath173 @xmath216 .
below we give a direct proof of this implication .
[ f4toz1 ] @xmath623 @xmath173 @xmath216 . * proof .
* suppose , on the contrary , that @xmath623 is true but @xmath216 is not true , i.e. , there exists a cubic 3-connected graph @xmath71 such that @xmath327 and @xmath71 has no @xmath1-factor .
let @xmath73 .
let @xmath385 be a copy of @xmath143 and let @xmath142 for @xmath384 be copies of @xmath656 , where @xmath72 is a cubic 3-connected graph , @xmath657 , and @xmath74 .
let @xmath369 ( see fig .
[ fy ] ) . obviously , @xmath200 . by @xmath623
, @xmath0 has a 4-vertex path @xmath205 such that @xmath128 is an inner vertex of @xmath205 , @xmath658 , and @xmath206 has a @xmath1-factor , say @xmath16 .
since @xmath121 has no @xmath1-factor , clearly @xmath16 is not a @xmath1-factor of @xmath206 , a contradiction .
@xmath306 + implication @xmath659 follows from obvious @xmath660 @xmath216 and from @xmath398 , by * [ z1ottof1]*. it also follows from obvious @xmath661 @xmath235 and from @xmath662 , by * [ z8tof2]*. below we give a direct proof of this implication .
[ z1ottof6 ] @xmath221 @xmath173 @xmath663 @xmath173 @xmath623 @xmath173 @xmath216 @xmath173 @xmath221 .
( uses * [ f4toz1 ] * and * [ z1ottoz8 ] * ) . obviously , @xmath221
@xmath173 @xmath663 @xmath173 @xmath623 . by * [ f4toz1 ]
@xmath623 @xmath173 @xmath216 . therefore @xmath221 @xmath173 @xmath216 .
it remains to prove @xmath216 @xmath173 @xmath221 . by * [ z1ottoz8 ] * , @xmath216 @xmath173 @xmath233 .
thus it is sufficient to show that @xmath233 @xmath173 @xmath221 .
let @xmath0 be a cubic 3-connected graph , @xmath200 , and @xmath213 is a 3-vertex path in @xmath0 .
let @xmath664 and @xmath543 be obtained from @xmath0 by subdividing edges @xmath253 and @xmath665 by new vertices @xmath666 and @xmath667 , respectively , and by adding a new edge @xmath668 .
then @xmath543 is a cubic 3-connected graph and @xmath576 .
consider the 3-vertex path @xmath669 in @xmath543 .
by @xmath233 , @xmath543 has a @xmath1-factor @xmath149 containing @xmath261 .
since vertex @xmath273 has degree one in @xmath578 , clearly @xmath149 has a 3-vertex path @xmath670 .
let @xmath205 be the 4-vertex path @xmath671 in @xmath0 .
then @xmath672 is a @xmath1-factor in @xmath206 . @xmath306
[ z9ottoz1 ] @xmath673 @xmath323 @xmath216 . * proof .
* obviously , @xmath673 @xmath173 @xmath233 . by the above claims , @xmath216 , @xmath233 , @xmath219 , and @xmath221 are equivalent .
so we can use these claims to prove @xmath673 .
let @xmath0 be a cubic 3-connected graph , @xmath67 a 3-edge cut of @xmath0 , @xmath187 and @xmath188 , and @xmath172 .
if the edges of @xmath67 are incident to the same vertex @xmath59 in @xmath0 ( i.e. , @xmath674 ) , then by @xmath233 , @xmath0 has a @xmath1-factor @xmath16 of @xmath0 such that @xmath189 , and so our claim is true .
so we assume that the edges in @xmath67 are not incident to the same vertex in @xmath0 .
then , since @xmath0 is cubic and 3-connected , clearly 3-edge cut @xmath67 forms a matching .
let @xmath71 and @xmath72 be the two component of @xmath675 . by the above arguments , we have @xmath676 and @xmath677 .
let @xmath678 be the graph obtained from @xmath0 by identifying the vertices of @xmath72 with a new vertex @xmath82 and similarly , let @xmath679 be the graph obtained from @xmath0 by identifying the vertices of @xmath71 with a new vertex @xmath276 , and so @xmath680 and @xmath681
. then @xmath682 , and @xmath683 .
let @xmath684 , where @xmath685 and @xmath686 . obviously , @xmath687 forms a 3-vertex path @xmath688 in @xmath678 and a 3-vertex path @xmath689 in @xmath679 . since @xmath0 is cubic and 3-connected , both @xmath678 and @xmath679 are also cubic and 3-connected . since @xmath172 , there are two possibilities ( up to symmetry ) : + @xmath690 @xmath691 and @xmath692 and + @xmath693 @xmath694 .
+ consider case @xmath690 .
by @xmath233 , @xmath695 has a @xmath1-factor @xmath696 .
by @xmath219 , @xmath679 has a 5-vertex path @xmath195 such @xmath276 is the center vertex of @xmath195 , @xmath697 , and @xmath698 has a @xmath1-factor @xmath699 .
then @xmath700 induces a @xmath1-factor @xmath16 in @xmath0 such that @xmath189 . + now consider case @xmath693 . by @xmath221
, we have : + @xmath701 @xmath678 has a 4-vertex path @xmath702 such that @xmath703 is an end - vertex of @xmath702 , @xmath704 , and @xmath705 has a @xmath1-factor , @xmath696 ; + @xmath706 likewise , @xmath679 has a 4-vertex path @xmath707 such that @xmath708 is an end - vertex of @xmath707 , @xmath709 , and @xmath710 has a @xmath1-factor @xmath699 .
+ then @xmath700 induces a @xmath1-factor @xmath16 in @xmath0 such that @xmath189 .
in this section we consider graph - compositions having @xmath1-factors and with all bricks having the same number of vertices modulo 6 .
it turns out that the @xmath1-factors of such graph - compositions intersect its 3-blockades in a special way and correspond to some special subgraphs in the frame of the composition .
let @xmath92 and @xmath16 be a @xmath1-factor of @xmath0 if any exists .
let @xmath711 be the subgraph in @xmath72 induced by the edge set @xmath712 .
let @xmath713 be the directed graph obtained from @xmath711 by directing each edge @xmath344 in @xmath711 from @xmath100 to @xmath24 , if @xmath111 .
for a @xmath1-factor @xmath16 of @xmath72 , let @xmath714 denote the directed graph obtained from @xmath16 by directing the two edges of each 3-vertex path @xmath13 in @xmath16 to the central vertex of @xmath13 .
we call @xmath714 a @xmath1-difactor of @xmath72
. a subgraph @xmath50 of a graph @xmath15 is called a _
cycle - packing _ in @xmath15 if each component of @xmath50 is a cycle . for a cycle - packing @xmath715 in @xmath72 ,
let @xmath716 denote the directed graph obtained from @xmath715 by directing its edges so that every cycle in @xmath716 becomes a directed cycle .
we call @xmath716 a _ cycle - dipacking _ of @xmath72 .
[ b(t ) ] let @xmath72 be a cubic 3-connected graph with no 3-blockades .
let @xmath0 be the graph obtained from @xmath72 by replacing each vertex by a triangle _
( i.e. , @xmath717 , where @xmath718 , and so @xmath0 is a claw - free graph ) . _
+ @xmath151 let @xmath16 be a @xmath1-factor of @xmath0 .
then + @xmath155 @xmath713 is a cycle - dipacking in @xmath72 , + @xmath158 @xmath719 for every 3-blockade @xmath67 of @xmath0 , and + @xmath159 @xmath720 for every triangle @xmath242 of @xmath0 .
+ @xmath163 let @xmath716 be a cycle - dipacking in @xmath72 .
then @xmath0 has a @xmath1-factor @xmath16 such that @xmath721 .
( uses * [ 3cut ] * ) .
obiously , @xmath722 . so all we need prove @xmath151 .
let @xmath723 and @xmath724 .
let @xmath725 and @xmath726 denote the number of edges coming in @xmath24 and going out of @xmath24 in @xmath727 , respectively . by * [ 3cut ] * and the fact that each vertex has been replaced by a triangle , there are three possibilities for every @xmath728 : + @xmath729 @xmath730 , + @xmath690 @xmath731 and @xmath732 , and + @xmath693 @xmath733 and @xmath734 . + if @xmath134 has no vertices satisfying @xmath693 , then we are done .
so let @xmath100 be a vertex in @xmath727 satisfying @xmath693 .
let @xmath687 be a maximal directed path starting at @xmath100 and ending at @xmath59 . clearly , @xmath735 .
then @xmath59 satisfies @xmath690 , and so @xmath736 .
therefore @xmath727 has an edge @xmath210 going out of @xmath59 . since @xmath687 is a maximal directed path starting at @xmath100 , clearly @xmath737 .
then @xmath738 , a contradiction .
finally , @xmath163 follows from the fact that @xmath717 and that every 2-edge matching covers the vertices of every triangle in @xmath33 .
@xmath306 + below we give a construction that provides a big variety of graph - compositions with the properties similar to that of a special graph - composition in * [ b(t)]*. [ v(a)=2mod6 ] let @xmath72 and each @xmath99 , @xmath739 , be a cubic 3-connected graphs and @xmath740 .
suppose that + @xmath741 @xmath742 and each @xmath743 , and so @xmath744 , + @xmath309 each @xmath745 has no @xmath1-factor for every vertex @xmath59 in @xmath746 adjacent to a vertex in @xmath747 , and + @xmath310 each @xmath748 has a @xmath1-factor for every @xmath749 , and + @xmath312 each @xmath750 has a @xmath1-factor for every 5-vertex path @xmath195 in @xmath99 centered at @xmath101 .
+ then + @xmath151 if @xmath16 is a @xmath1-factor of @xmath0 , then @xmath751 is the edge set of a @xmath1-factor @xmath715 of @xmath72 and @xmath721 , + @xmath163 if @xmath16 is a @xmath1-factor of @xmath0 and if @xmath72 and each @xmath99 have no 3-blockades , then @xmath752 for every 3-blockade @xmath67 in @xmath0 , and + @xmath228 if @xmath716 is a @xmath1-difactor of @xmath72 , then there exists a @xmath1-factor @xmath16 of @xmath0 such that @xmath721 .
( uses * [ 3cut ] * ) .
obviously , @xmath163 follows from @xmath151 .
we prove @xmath151 .
let @xmath753 and @xmath754 be the union of the components ( 3-vertex paths ) of @xmath16 meeting @xmath755 . by @xmath741 , @xmath743 .
therefore @xmath754 satisfies @xmath155 or @xmath158 or @xmath159 in * [ 3cut]*. by @xmath309 , @xmath754 does not satisfy @xmath159 in * [ 3cut]*. thus @xmath756 is the edge set of a @xmath1-factor @xmath149 in @xmath72 and the two edges of the same 3-vertex path @xmath13 in @xmath149 are directed to the center vertex of @xmath13 . finally , @xmath228 follows from assumption @xmath310 and @xmath312 .
@xmath306 + it is also easy to prove the following claim similar to that above when the assumption @xmath743 is replaced by @xmath757 .
[ v(a)=4mod6 ] let @xmath72 and each @xmath99 , @xmath739 , be a cubic 3-connected graphs and @xmath740 .
suppose that + @xmath741 @xmath742 and each @xmath757 , and so @xmath744 , + @xmath309 each @xmath758 has no @xmath1-factor , where @xmath759 is a tree subgraph of @xmath99 containing @xmath101 and such that @xmath760 is a 3-edge matching , + @xmath310 each @xmath761 has a @xmath1-factor , and + @xmath312 each @xmath762 has a @xmath1-factor , where @xmath222 is a 4-vertex path in @xmath99 with @xmath101 being an inner vertex of @xmath222 .
+ then + @xmath151 if @xmath16 is a @xmath1-factor of @xmath0 , then @xmath763 is a cycle - dipacking in @xmath72 , + @xmath163 if @xmath16 is a @xmath1-factor of @xmath0 , then @xmath764 for every @xmath739 , and + @xmath228 if @xmath716 is a cycle - dipacking in @xmath72 , then @xmath0 has a @xmath1-factor @xmath16 of @xmath0 such that @xmath721 .
the proof of @xmath151 and @xmath163 in * [ v(a)=4mod6 ] * is similar to that in * [ b(t)]*. the proof of @xmath228 in * [ v(a)=4mod6 ] * is similar to that in * [ v(a)=2mod6]*. + in the next two claims we give constructions of bricks satisfying the assumptions @xmath741 - @xmath312 in * [ v(a)=2mod6 ] * and * [ v(a)=4mod6 ] * , respectively .
+ let @xmath765 ( see fig .
[ fy ] ) , where @xmath121 is the graph having two vertices @xmath215 , @xmath766 and three parallel edges with the end - vertices @xmath215 , @xmath766 , and so @xmath767 .
let @xmath768 , and so @xmath769 is the blockade with three edges having exactly one end - vertex in @xmath770 , @xmath384 .
+ using * [ 3cut ] * , it is easy to prove the following two claims . [
( z , z)0 ] let @xmath95 be a pair defined above .
suppose that @xmath290 for @xmath384 .
then @xmath771 and + @xmath151 @xmath772 has no @xmath1-factor for every vertex @xmath24 in @xmath773 _ @xmath28see also * [ 3cut ] * @xmath159 and fig .
[ a1]@xmath29 _ , + @xmath163 if @xmath774 and @xmath16 is a @xmath1-factor of @xmath775 , then @xmath776 for @xmath384 , and + @xmath228 if @xmath777 is a 5-vertex path in @xmath205 with @xmath215 being the center vertex of @xmath195 and @xmath16 is a @xmath1-factor of @xmath778 , then @xmath776 for @xmath384 . [ ( z , z)4 ] let @xmath95 be a pair defined above .
suppose that @xmath779 for @xmath384 .
then @xmath780 and + @xmath151 if @xmath242 is a tree subgraph of @xmath205 containing @xmath215 and such that @xmath781 is a 3-edge matching , then @xmath782 has no @xmath1-factor _ @xmath28see also * [ 3cut ] * @xmath169 and fig .
[ a2]@xmath29 _ , + @xmath163 if @xmath16 is a @xmath1-factor of @xmath783 , then @xmath784 for @xmath384 , and if in addition , both @xmath119 and @xmath120 are isomorphic to @xmath33 , then @xmath785 for @xmath384 , + @xmath228 if @xmath222 is a 4-vertex path in @xmath205 with @xmath215 being an inner vertex of @xmath222 and @xmath16 is a @xmath1-factor of @xmath786 , then @xmath787 for @xmath384 , and + @xmath788 if @xmath16 is a @xmath1-factor of @xmath789 , then @xmath790 for @xmath384 . if @xmath95 satisfies the assumptions in * [ ( z , z)0 ] * ,
then by * [ ( z , z)0 ] * , the graph - composition @xmath97 satisfies the assumptions of * [ v(a)=2mod6 ] * , and therefore has the properties @xmath151 , @xmath163 , and @xmath228 in * [ v(a)=2mod6]*. similarly , if @xmath95 satisfies the assumptions in * [ ( z , z)4 ] * , then by * [ ( z , z)4 ] * , the graph - composition @xmath97 satisfies the assumptions of * [ v(a)=4mod6 ] * , and therefore has the properties @xmath151 , @xmath163 , and @xmath228 in * [ v(a)=4mod6]*. notice that if any claim in * [ 3-con ] * is true , then there are infinitely many pairs @xmath95 satisfying the assumptions and , accordingly , the conclusions of * [ ( z , z)0 ] * , as well as of * [ ( z , z)4]*. + from * [ v(a)=2mod6 ] * and * [ v(a)=4mod6 ] * we have , in particular : [ g , p , k ] let @xmath791
. then there are infinitely many cubic 3-connected graphs @xmath0 such that @xmath172 and @xmath792 for every @xmath1-factor @xmath16 of @xmath0 and every 3-edge cut @xmath67 of @xmath0 .
let , as in section [ constructions ] , @xmath740 , @xmath793 , and @xmath794 .
let @xmath795 . as we mentioned in section [ constructions ]
, there is a unique bijection @xmath105 such that if @xmath106 , then @xmath107 is an edge in @xmath0 having one end - vertex in @xmath108 and the other end - vertex in @xmath109 .
suppose that @xmath16 is a @xmath1-factor of @xmath72 .
let @xmath796 be the set of vertices of degree @xmath797 in @xmath16 ( and so @xmath424 ) .
for a 3-vertex path @xmath798 in @xmath16 , let @xmath799 and let @xmath800 be the graph obtained from @xmath109 by adding two edges @xmath107 and @xmath801 together with their end - vertices .
let @xmath802 denote the set of @xmath1-factors of a graph @xmath15 and @xmath803 denote the set of @xmath1-factors @xmath715 of @xmath0 such that @xmath804 .
let @xmath805 denote the cartesian product of sets @xmath47 and @xmath806 ( we assume that if @xmath807 , then @xmath808 . + from * [ v(a)=2mod6 ] * we have : [ gamma ] let @xmath72 and each @xmath99 , @xmath739 , be cubic 3-connected graphs .
suppose that each @xmath809 and each @xmath810 satisfies the following conditions : + @xmath309 @xmath811 has no @xmath1-factor for every vertex @xmath273 in @xmath812 adjacent to a vertex in @xmath813 and + @xmath310 @xmath748 has a @xmath1-factor for every @xmath814 , and + @xmath312 @xmath750 has a @xmath1-factor for every 5-vertex path @xmath195 in @xmath815 centered at @xmath101 .
+ then @xmath154 + @xmath816 @xmath817 for @xmath818 , and @xmath819 , + @xmath820 @xmath821 \bigotimes [ \bigotimes \ { \gamma ( a_2^u(p ) ) : u \in v_2(p)\ } ] $ ] , and + @xmath822 @xmath823 , + and so @xmath0 has a @xmath1-factor if and only if @xmath72 has a @xmath1-factor .
suppose that @xmath824 is a cycle - dipacking in @xmath72 .
if @xmath798 is a directed 3-vertex path in @xmath824 , then let @xmath825 be the graph obtained from @xmath109 by adding two edges @xmath107 and @xmath801 together with their end - vertices , a new vertex @xmath215 , and a new edge @xmath826 .
let @xmath827 denote the set of @xmath1-factors @xmath715 in @xmath0 such that @xmath828 .
let @xmath829 denote the set of cycle - dipackings in @xmath72 .
+ from * [ v(a)=4mod6 ] * we have : [ gamma2 ] let @xmath72 and each @xmath99 , @xmath739 , be cubic 3-connected graphs .
suppose that each @xmath830 and each @xmath810 satisfies the following conditions : + @xmath741 each @xmath757 , and so @xmath742 and @xmath744 , + @xmath309 each @xmath758 has no @xmath1-factor , where @xmath759 is a tree subgraph of @xmath108 containing @xmath101 and such that @xmath760 is a 3-edge matching , + @xmath310 each @xmath761 has a @xmath1-factor , and + @xmath312 each @xmath762 has a @xmath1-factor , where @xmath222 is a 4-vertex path in @xmath99 with @xmath101 being an inner vertex of @xmath222 .
+ then + @xmath816 @xmath831 for @xmath832 , @xmath833 , + @xmath820 @xmath834 \bigotimes [ \bigotimes \ { \gamma ( a^u - a^u ) : u \in v(b ) - v(\vec{c})\ } ] $ ] , and + @xmath822 @xmath835 .
as it is shown in section [ almostcubic ] , there are infinitely many pairs @xmath346 satisfying the assumptions in * [ gamma ] * , as well as in * [ gamma2]*.
in this section we describe a sequence ( mentioned in section [ 3connected ] , remark @xmath243 ) of cyclically 6-connected graphs @xmath0 with two disjoint 3-vertex paths @xmath13 , @xmath261 such that @xmath172 and @xmath836 has no @xmath1-factor .
we call a path @xmath242 in a graph @xmath0 a _ thread _ if the degree of each inner vertex of @xmath242 in @xmath0 is equal to two and the degree of each end - vertex of @xmath242 in @xmath0 is not equal to two .
let @xmath837 be a cycle with @xmath838 vertices , @xmath839 , and let @xmath840 be a @xmath1-factor + of @xmath837 , where @xmath841 .
let @xmath842 be the graph obtained from @xmath837 by adding the set + @xmath843 of @xmath844 new vertices and the set @xmath845 i \in \{1 , \ldots , s\ } , j \in \{1,2,3\}\}$ ] of @xmath838 new edges ( see fig .
[ fr1 ] for an illustration of @xmath846 ) .
let @xmath847 .
let @xmath848 .
then the vertices of 3-vertex paths @xmath170 , @xmath849 , and @xmath850 are of degree two in @xmath851 .
let @xmath852 , @xmath853 , and @xmath854 be the threads in @xmath851 containing @xmath170 , @xmath849 , and @xmath850 , respectively , and so each of these threads has exactly five vertices .
let @xmath855 be the graph obtained from @xmath851 by replacing threads @xmath852 , @xmath853 , and @xmath854 by edges @xmath856 , @xmath857 , and @xmath858 .
then obviously , @xmath855 is isomorphic to @xmath859 .
let @xmath860 be the subgraph of @xmath842 induced by the vertex subset @xmath861 .
then @xmath862 , @xmath863 , and the set @xmath864 of edges having one vertex in @xmath860 and the other vertex in @xmath865 is a 6-blockade in @xmath842 , @xmath866 .
[ g , l , l ] let @xmath842 , @xmath839 , be the graph defined above , and so @xmath868
. then + @xmath151 @xmath869 has no @xmath1-factor for every pair @xmath870 , @xmath871 , + @xmath163 @xmath872 for @xmath866 and @xmath873 @xmath28and for every 5-blockade @xmath874 of @xmath875 one of the two components of @xmath876 is a 5-cycle@xmath29 , and + @xmath228 @xmath877 has a @xmath1-factor for every edge @xmath8 and every 3-vertex path @xmath13 in @xmath878 . *
* the proof will consist of three parts .
+ @xmath348 we prove @xmath151 . by symmetry
, we can assume that @xmath879 and @xmath880 . let @xmath881 .
suppose , on the contrary , that @xmath881 has a @xmath1-factor @xmath16 . then each edge @xmath882 , @xmath883 , is a dangling edge in @xmath0 .
then @xmath884 belongs to a 3-vertex path @xmath885 . by symmetry
, we can assume that @xmath886 .
then @xmath887 is an isolated vertex in @xmath888 ( see , for example , fig .
[ fr1 ] , where @xmath889 ) .
therefore @xmath16 is not a @xmath1-factor of @xmath0 , a contradiction .
+ @xmath364 we prove @xmath163 by induction on @xmath797 .
it is easy to check that @xmath875 has the properties described in @xmath163 .
so let @xmath866 . since @xmath842 has 6-cycles , @xmath890 .
so our goal is to prove that @xmath872 .
let @xmath874 be a minimum blockade in @xmath842 .
obviously , each @xmath842 is a connected graph , and so @xmath891 .
let @xmath892 , and so the vertices of 3-vertex paths @xmath893 , @xmath894 , and @xmath895 are of degree two in @xmath896 .
let @xmath897 , @xmath898 , and @xmath899 be the threads in @xmath896 containing @xmath900 , @xmath901 , and @xmath902 , respectively . then graph @xmath859 can be obtained from @xmath896 by replacing each thread @xmath903 by a new edge @xmath904 , @xmath905 .
let @xmath906 .
then @xmath907 and @xmath908 and the edge set @xmath909 is a 6-blockade .
suppose first that @xmath874 has an edge @xmath912 in @xmath913 .
then @xmath914 , and so @xmath874 contains a subset @xmath915 which is a blockade in @xmath896 separating @xmath276 and @xmath82 . if @xmath916 , then @xmath917 , and we are done . since @xmath914 ,
the sizes of minimum blockades separating @xmath276 and @xmath82 in @xmath896 and @xmath859 are the same .
so if @xmath911 , then by the induction hypothesis , @xmath916 , and we are done .
so let @xmath918 and @xmath919 .
since @xmath873 , @xmath920 is equal to the size of a minimum blockade in @xmath875 , and so @xmath921 .
since no minimum blockade of @xmath875 is a cut in @xmath922 , we have @xmath923 , and so @xmath924 .
now suppose that @xmath874 has no edge in @xmath913 , i.e. , @xmath925 .
then it is easy to check ( using the symmetries of @xmath806 ) that every such blockade has six edges , and so @xmath924 .
+ @xmath926 finally , we prove @xmath228 .
let , as in @xmath364 , @xmath906 and @xmath927 .
obviously , each @xmath860 is isomorphic to @xmath806 .
it is easy to prove the following claims .
_ @xmath928 has a @xmath1-factor for every edge @xmath8 in @xmath806 and for every 3-vertex path @xmath13 in @xmath929 . _
_ let @xmath276 and @xmath82 be distinct vertices of degree two in @xmath806 that belong to different components of @xmath930 and let @xmath931 be the graph obtained from @xmath806 by adding a new edge @xmath912 .
let @xmath687 be a 3-vertex path in @xmath931 containing edge @xmath912 .
then @xmath932 has a @xmath1-factor for every edge @xmath8 in @xmath933 . _
+ let @xmath934 . obviously , @xmath935 is a matching and @xmath936 .
now , if @xmath597 is a @xmath1-factor of @xmath860 , then clearly @xmath937 is a @xmath1-factor of @xmath842 .
therefore from claim 1 we have : + 3 .
_ @xmath877 has a @xmath1-factor for every edge @xmath8 in @xmath842 and every 3-vertex path @xmath13 in @xmath938 .
_ thus , it remains to consider a pair @xmath939 , where @xmath134 is a 3-vertex path containing an edge in @xmath935 and @xmath940 .
since @xmath935 is a matching , @xmath134 has exactly one edge @xmath941 in @xmath935 . by symmetry of @xmath842
, we can assume that @xmath942 and @xmath943 , where @xmath944 and @xmath945 are vertices of degree two in @xmath946 and @xmath947 , respectively , and @xmath948 . for @xmath949 ,
let @xmath950 if @xmath951 and @xmath952 .
obviously , @xmath953 is a function from @xmath954 onto @xmath47 .
+ @xmath955 suppose first that @xmath956 .
let @xmath957 , and so @xmath958 .
let @xmath959 . by symmetry
, we can assume that @xmath960 .
obviously , @xmath874 is a @xmath1-packing in @xmath842 and @xmath961 consists of three vertices @xmath962 , @xmath963 , and @xmath135 and one edge @xmath964 , where @xmath962 and @xmath963 have degree two in @xmath860 and belong to different components of @xmath965 and @xmath966 .
let @xmath967 .
then @xmath968 is a 3-vertex path in @xmath969 . by claim 2
, @xmath970 has a @xmath1-factor @xmath971 for every edge @xmath972 in @xmath973 .
let @xmath974 .
since @xmath975 , we have : @xmath976 is a @xmath1-factor of @xmath842 containing @xmath134 and avoiding @xmath8 .
+ @xmath977 finally , suppose that @xmath978 .
let @xmath979 , and so @xmath980 .
let @xmath981 . by symmetry
, we can assume that @xmath960 .
as in @xmath955 , @xmath874 is a @xmath1-packing in @xmath842 and @xmath961 consists of three vertices @xmath962 , @xmath963 , and @xmath135 and one edge @xmath964 , where @xmath962 and @xmath963 have degree two in @xmath860 and belong to different components of @xmath965 and @xmath966 .
therefore we are done by the same arguments as in @xmath955 .
@xmath306 + * acknowledgment .
* i am grateful to the referees for a very careful reading of the paper and very useful remarks . also , many thanks to one of the referees for drawing my attention to recent papers @xcite .
a. kelmans , constructions of cubic bipartite and 3-connnected graphs without hamiltonian cycles . in _
analiz zadach formirovaniya i vybora alternativ _ , vniisi , moscow * 10 * ( 1986 ) 64 - 72 ( see also _ amer
( 2 ) vol .
* 158 * ( 1994 ) 127 - 140 ) . | a subgraph ( a spanning subgraph ) of a graph @xmath0 whose all components are 3-vertex paths is called an @xmath1-packing ( respectively , an @xmath1-factor of @xmath0 .
let @xmath2 and @xmath3 be the number of vertices and the maximum number of disjoint 3-vertex paths in @xmath0 , respectively .
we discuss the following old + ( a. kelmans , 1984 ) .
_ is the following claim true ? + @xmath4 if @xmath0 is a cubic 3-connected graph , then @xmath0 has a @xmath1-packing that avoids at most two vertices of @xmath0 , i.e. @xmath5 . _ + we show , in particular , that claim @xmath4 is equivalent to some seemingly stronger claims ( see * [ 3-con ] * ) .
for example , if @xmath6 , then the following claims are equivalent : @xmath0 has a @xmath1-factor , for every @xmath7 there exists a @xmath1-factor of @xmath0 avoiding ( containing ) @xmath8 , @xmath9 has a @xmath1-factor for every @xmath10 such that @xmath11 , and @xmath12 has a @xmath1-factor for every 3-vertex path @xmath13 in @xmath0 .
we also show that certain claims in theorem * [ 3-con ] * are best possible .
we give a construction providing infinitely many cyclically 6-connected graphs @xmath0 with two disjoint 3-vertex paths l and l such that @xmath6 and @xmath14 has no @xmath1-factor .
it turns out that if claim @xmath4 is true , then reed s dominating graph conjecture is true for cubic 3-connected graphs .
+ : cubic 3-connected graph , 3-vertex path packing , 3-vertex path factor , domination .
+ + + * university of puerto rico , san juan , puerto rico * + * rutgers university , new brunswick , new jersey * + * the paper was accepted for publication * + * in discrete mathematics in august 14 , 2009 * |
9 u. zeitler , a.m. devitt , j.e .
digby , c.j .
mellor , a.j .
kent , k.a .
benedict , and t. cheng , phys .
lett . * 82 * , 5333 ( 1999 ) , and references therein .
v.m . apalkov and m.e .
portnoi , physica e * 12 * , 470 ( 2002 ) .
k.v . kavokin and m.e .
portnoi , to be published .
g.l . bir and g.e .
pikus , _ symmetry and strain - induced effects in semiconductors _
( wiley , new york , 1974 ) .
falko , phys .
b * 46 * , 4320 ( 1992 ) . , as a function of bare level separation , @xmath28 , at filling factor @xmath29 ( solid and dotted lines ) and for the case of non - interacting electrons ( dashed line ) .
the results for @xmath29 are shown for two different values of the fang - howard parameter ( inverse quantum well width ) @xmath30 and @xmath31 in units of inverse magnetic length @xmath32 .
the absorption rate @xmath33 is in units of @xmath34 ; @xmath35 is in coulomb units ( @xmath36).,width=510 ] at filling factor @xmath37 as a function of interlevel separation @xmath38 for the case of zero zeeman splitting , @xmath39 , for three values of parameter @xmath40 .
the data for @xmath41 1.0 and 1.5 are multiplied by 10 .
the absorption rate @xmath33 is in units of @xmath34 ; @xmath38 is in units of @xmath42.,width=510 ] | we propose to use phonon absorption spectroscopy to study many - body gaps and phases of two - subband heterostructures in the quantum hall regime . implications of the spin - orbit interaction for phonon absorption in this system are considered .
phonon spectroscopy is a powerful experimental technique that has been successfully used in the study of the properties of a two - dimensional ( 2d ) electron gas in the quantum hall regime [ 1 ] .
it allows probing of the states with non - zero momentum , which are not accessible by either cyclotron resonance or photoluminescence methods .
we propose to use acoustic - phonon spectroscopy to study many - body effects in a two - subband quasi-2d system , formed at a single semiconductor heterojunction and subjected to a quantizing magnetic field .
we show that the formation of many - body gaps and the existence of different spin phases in this system result in a striking difference in the magnetic field dependence of the phonon absorption rate at different filling factors
. acoustic - phonon scattering of 2d electrons in a quantizing magnetic field is strongly suppressed due to the dispersionless structure of the landau levels ( lls ) . in a two - subband system , however , a sharp enhancement of the electron - phonon interaction occurs when the two lls corresponding to different size - quantization subbands ( we consider the second ll of the first subband and the first ll of the second subband only ) approach one another , and the separation between these energy levels , @xmath0 , becomes close to @xmath1 . here
@xmath2 is the cyclotron energy , @xmath3 is the intersubband splitting , @xmath4 is the velocity of sound and @xmath5 is the magnetic length .
this electron - phonon interaction enhancement should manifest itself in the magnetic field dependence of the phonon absorption rate and dissipative conductivity . at filling factor slightly greater than two ,
when the electron - electron interaction between the particles belonging to either of the intersecting lls can be neglected , a double - peak structure in the magnetic field dependence of the phonon - mediated transport properties of the two - subband system has been predicted [ 2 ] . with increasing electron density , the repulsion between the levels due to electron - electron exchange interactions opens a gap in the interlevel excitation spectra . at filling factor @xmath6
this gap exists at all values of the bare level splitting @xmath7 and can well exceed @xmath1 . as a result , the double - peak structure transforms into a single - peak one ( see figure 1 ) . by far the most interesting case is when the electron filling factor @xmath8 . when the cyclotron energy is close to the intersubband splitting the system can be mapped onto a four - level electron system with an effective filling factor @xmath9 .
the ground state is either a ferromagnetic state or a spin - singlet state , depending on the values of the interlevel splitting and zeeman energy @xmath10 .
the electron - electron interaction renormalizes the phase boundaries , which results in the existence of a ferromagnetic phase even for @xmath11 .
the excitation spectrum has a gap for any value of the interlevel splitting @xmath7 and zeeman splitting @xmath10 .
this results in strong suppression of the electron - phonon interaction .
the rate of phonon absorption by the considered quasi-2d electron system at @xmath8 again has a double - peak structure as a function of level splitting and a steplike structure as a function of zeeman splitting ( see figure 2 ) . in the absence of the spin - orbit interaction ,
the transitions between levels with different spin indices are forbidden because the electron - phonon interaction conserves the electron spin .
the spin - orbit interaction allows such transitions .
matrix elements of the spin - orbit interaction hamiltonian @xmath12 between levels differing by subband number , ll number , and spin projection on the normal to the structure plane , are given by following expressions [ 3 ] : @xmath13 where @xmath14 \right ) \cdot \mathbf{s}$ ] , @xmath15 $ ] , @xmath16 and @xmath17 numerate subbands , @xmath18 is the ll number , and @xmath19 and @xmath20 are phenomenological spin - orbit constants of the semiconductor [ 4 ] .
the expressions for @xmath21 and @xmath22 are given by @xmath23 where @xmath24 is the quantum - well confinement potential .
these matrix elements cause an admixture of states with opposite spin direction to both the initial and final states for the phonon absorption . as a result ,
the phonon absorption rate at @xmath8 never becomes zero ; in the region between peaks , it is of the order of @xmath25 times its value at the peak of absorption , where @xmath26 is the corresponding matrix element of the spin - orbit interaction ( eq.(1 ) ) , and @xmath27 is the many - body energy gap . in the absence of many - body effects , the inter - subband spin - orbit coupling results in anticrossing of levels with different spin , subband and ll indices .
the resulting energy gaps can be measured by high - resolution spectroscopic techniques ( optical or spin / cyclotron resonance spectroscopy ) , as has been suggested for one - subband systems in tilted magnetic fields [ 5 ] .
this work was supported by the uk epsrc . |
before discussing many - body interactions between disks rotating with an angular frequency @xmath0 , first we describe the behaviour of a dilute limit : behaviours of a single rotating disk and a pair of rotating disks ( supplementary fig . 1 and supplementary note 1 ) .
we characterize the rotation speed of a disk either by the angular frequency @xmath0 or the relevant reynolds number @xmath1 ( @xmath2 : density ; @xmath3 : particle radius : @xmath4 : liquid viscosity ) .
figure [ fig : fig1]a and b show the 2d flow field around a rotating disk and the velocity distribution , respectively .
the latter shows that the rotating velocity linearly increases with the distance from the centre of mass , @xmath5 , inside a disk and decays as @xmath6 in its outside .
this is the characteristic of the so - called rankin vortex , i.e. , a forced vortex in the central core surrounded by a free vortex .
we note that the rankin vortex is known to mimic the 2d flow field of tornado and hurricane @xcite . for two co - rotating ` point ' vortexes ,
the analytical solution is known and the two particles rotate around their center of mass towards the same direction as the rotating direction @xcite . in this case , the radius of rotation is the half of the initial interparticle distance . for particles with a finite size , on the other hand , the direction of rotation and the rotation center are the same as the case of point particles , but the radius of the rotation decreases with an increase in the rotation speed of the particles @xmath0 , or @xmath7 .
this problem has a similarity to viscous interactions of co - rotating vortices @xcite , but the crucial difference arises from the fact that the vortex core is an undeformable solid and not a liquid in our case .
there are a few studies on spinning particles in three dimensions @xcite ; however , we note that there is an essential difference between 2d and 3d problems ( supplementary figure 2 and supplenetary note 2 ) .
we confirm that if we fix the centres of mass of two rotating particles on a fixed line , they always repel with each other by the magnus force @xcite .
the attraction between particles is thus due to interparticle hydrodynamic interactions , as schematically explained in fig .
[ fig : fig1]c .
the flow field generated by a rotating disk makes the other disk follow it , and vice versa .
thus , two particles try to follow each other while rotating around the centre of mass of the pair . in 3d ,
on the other hand , such hydrodynamic interactions are weaken due to the presence of the escape dimension , thus the magnus force wins over the hydrodynamic attractive force , and particles repel with each other . in 2d
, this hydrodynamic attraction overwhelms the repulsion due to the magnus force , which leads to the rotating pair of particles ( fig .
[ fig : fig1]d ) , whose interparticle distance monotonically decreases with an increase in @xmath0 , or @xmath7 ( fig .
[ fig : fig1 ] e ) , as long as the area fraction of disks , @xmath8 , is sufficiently small .
however , it should be noted that this situation is realized in a periodic boundary condition . in relation to the above dimensionality effect on hydrodynamic interactions between rotors , it is worth noting that the hexagonal ordering observed by grzybowski et al.@xcite at finite reynolds numbers is due to the effective hydrodynamic repulsion between the disks floating at an interface in three dimensions ( _ i.e. _ , 2.5 dimensions ) . .
@xmath9 increases linearly with @xmath5 inside the disk and then decays as @xmath6 in its outside , which is the characteristic of the so - called rankin voltex .
* c , * schematic picture of a pair of two disks rotating around the centre of mass .
the two disks are rotating counter - clockwise with the same speed . *
d , * trajectory of coupled rotating particles ( the system size=256@xmath10256 ) .
the initial separation of the particles ( arrowed ) was 30 and each particle rotates counter - clockwise with @xmath11 .
* e , * temporal change in the interparticle separation .
we note that for all the cases including @xmath12 , the interparticle distance eventually reaches a final steady - state value of the separation ( see fig .
[ fig : fig1]d ) , which monotonically decreases with @xmath7 .
* f , * a cluster of disks formed at @xmath13 at @xmath14 .
we confirm that irrespective of the value of @xmath7 , a cluster is always formed in the range of @xmath7 studied .
, width=321 ] now we consider the dynamical behaviour of a system of many disks rotating in the counter - clockwise direction with @xmath0 .
the hydrodynamic attraction between rotating disks leads to the formation of rotating clusters for low @xmath8 ( fig .
[ fig : fig1]f ) .
the higher rotation speed of individual disks leads to the formation of a more compact rotating cluster .
this tendency is basically the same as that for a pair of rotors ( fig .
[ fig : fig1]e ) . for this regime ,
only one cluster is formed in the simulation box and the whole cluster rotates in the counter - clockwise direction , as shown in fig .
[ fig : fig1]f
. a cluster always tends to have a circular shape , but it does not have any particular internal structural order , partly because imperfect matching between the size and the number of particles leads to structural fluctuations : the internal structure is basically controlled by the number of disks in it and @xmath7 , which determine the cluster size , but fluctuating with time . with a further increase in @xmath7 ,
however , this cluster state becomes unstable , since the repulsive magnus force of nonlinear origin eventually wins over the hydrodynamic attraction for high @xmath7 ( see below ) . above a critical @xmath8 ( @xmath15 ) , on the other hand , a system exhibits a re - entrant transition between states as a function of @xmath0 ( fig .
[ fig : fig2]a ) : for low @xmath0 a system is in a disordered liquid state with large fluctuations , but with an increase in @xmath0 it enters into a rather stable hexatic phase where rotating particles are localized on a hexagonal lattice .
the border between the cluster to the hexatic state is rather sharp as a function of @xmath8 . .
* * a , * snapshots of particle configurations together with the velocity fields .
the system size is 256@xmath10256 , the number of particles is 80 , and @xmath16 .
we can see sequential transitions from disordered , hexatic ordered , to disordered states with an increase in @xmath7 .
* b , * upper panel : the direction of the hydrodynamic force @xmath17 measured from the interparticle axis .
lower panels : patterns formed by brownian simulations for the two off - axis forces : left is a liquid state for @xmath18 , whereas the right panel is a hexatic state for @xmath19 .
* c , * @xmath7-dependence of the hexatic order parameter @xmath20 , which clearly shows the re - entrant behaviour for various @xmath8 s ( expressed in % ) . * d , * the degree of fluctuations of @xmath20 as a function of @xmath7 for various @xmath8 s ( expressed in % ) .
we can see that as in a thermodynamic hexatic ordering transition , the order parameter exhibits large amplitude fluctuations near the transition points .
* e , * @xmath7-dependence of the decay of the spatial correlation function of the hexatic order parameter @xmath21 normalized by the radial distribution function @xmath22 around the transition at low @xmath7 for @xmath23 . in the hexatic state , @xmath24 decays with a power law with the exponent of -1/4 , as it should be @xcite . in the disordered states it decays almost exponentially .
* f , * the same as * e * but for around the transition at high @xmath7 . in the hexatic state
, @xmath24 again decays with a power law with the exponent of -1/4 . in the disordered states it decays almost exponentially .
, width=321 ] the transition can be characterized by the nature of interparticle interactions .
here we analyse a point pattern to extract the nature of interparticle interactions .
the point pattern analysis is very useful to determine the overall interparticle interactions @xcite .
here we use what we call @xmath25 function , which is the number of connected regions as a function of the radius of circle whose centre is located at the centre of each disk .
the number of connected regions decreases monotonically , with an increase in the circle radius , from the total number of disks @xmath26 to one .
here we use @xmath25 for the one normalized by @xmath26 . by comparing @xmath25 for a reference system made of randomly distributed disks , i.e. , poisson pattern
, we can judge whether the interaction is repulsive or attractive .
for a system of particles with repulsive interactions , particles form a rather regular pattern and @xmath25 decays slower than that for the corresponding poisson pattern . for a system of particles with attractive interactions , on the other hand , particles form a cluster pattern and
@xmath25 decays faster than that for the corresponding poisson pattern . in fig .
[ fig : fig3]a , we show the @xmath25 function for @xmath13 .
we can clearly see that for the cluster - forming case at @xmath27 the interaction is attractive whereas for the disordered state at @xmath28 the interaction is repulsive . in fig .
[ fig : fig3]b , we show the @xmath25 function for @xmath29 .
we can see that for all the value of @xmath7 the interaction is basically repulsive , but its strength is maximum at @xmath30 , for which the hexatic order is formed .
the transition can be seen even more clearly by the degree of localization of the flow field : for a hexatic state each particle has its own localized rotational flow field , whereas for a cluster state a single vortex is always formed and thus the flow field is strongly delocalized .
however , the precise nature of the cluster - hexatic phase transition , such as whether the transition is continuous or discontinuous and whether a disorder state always exists between the two states or there exists a critical point between the cluster and the hexatic state , is not clear at this moment . to access this problem
, we need to survey the border region of a bigger system size with a high resolution of @xmath8 and @xmath7 .
although this is a very interesting problem , we leave this for future investigation .
we also find that the hexatic state is eventually destabilized and melts by a further increase in @xmath0 and a system becomes disordered again .
we note that for all these states interparticle interactions are basically repulsive , unlike the case of low @xmath8 .
although the torque exerted to each disk is exactly the same , the rotation speed of a disk can in principle depend on the particle configuration around it . here
we show in fig .
[ fig : fig4 ] the normalized variance of the angular frequency @xmath0 of rotating disks as a function of the averaged @xmath0 . for
ordered states , the variance , i.e. , the fluctuations of rotation speed , becomes very small , indicating all the disks rotating with almost the same frequency .
for disordered states , on the other hand , the variance is large , reflecting the large fluctuations of particle environment .
this result indicates a strong negative correlation between the degree of fluctuations of rotational speed of particles and the degree of order in fig .
[ fig : fig2]c , we show the @xmath0-dependence of the hexatic order parameter @xmath20 , which clearly indicates the re - entrant nature of the state transitions .
the transitions can also be characterized by the magnitude of fluctuations of @xmath20 , or the susceptibility ( see fig .
[ fig : fig2]d ) , which is also observed in a thermodynamic hexatic ordering in 2d disks . in the hexatic
ordered state , we confirm the power law decay of the spatial correlation of the hexatic order which is specific to the hexatic phase ( see fig . [ fig : fig2]e and f ) .
we do not see any indication of the positional order since the radial distribution function decays almost exponentially ( see below ) . here
we show a hexatic phase observed in a large system ( lattice size=2048@xmath31 and the number of particles=5120 ) at @xmath29 and @xmath30 .
we can see grain boundaries between hexatic order with different orientations in fig .
[ fig : fig5]a and b. figure [ fig : fig6]a and b show the decay of the correlation function of the hexatic order normalized by the radial distribution function @xmath22 , @xmath24 , and that of @xmath32 , respectively .
we can see @xmath24 decays algebraically in a short distance @xmath33 , but decays faster for long distance .
this is because the size of mono - domain regions is finite ( see fig .
[ fig : fig5]a and b ) . on the other hand
, we can see that @xmath22 decays faster than an algebraic decay even for @xmath33 , where @xmath24 decays algebraically .
this clearly indicates the absence of quasi - long - range translational order in the ordered phase .
thus we conclude that the ordered phase is the hexatic phase .
the appearance of the transitions between dynamical states as a function of @xmath8 and @xmath0 in athermal systems is quite striking , which is reminiscent of phase transitions in a thermal system .
we stress that the interparticle interaction in our system is of purely hydrodynamic origin . here
we consider a quite interesting feature of hydrodynamic interparticle interactions , which are absent in a thermal system .
binary interaction potentials that lead to phase ordering in thermal systems always act along the interparticle axis , i.e. , along the line connecting the centres of mass of two interacting particles .
however , this is not the case for our athermal system .
the nonlinear magnus force acts along the line connecting two particles , whereas the stokes force acts according to the flow direction .
furthermore , linear hydrodynamic interactions are tensorial and act not along the interparticle direction .
thus , the total hydrodynamic force does not act along the interparticle axis .
the direction of the hydrodynamic force as a function of @xmath7 is shown in the upper panel of fig .
[ fig : fig2]b with a schematic explanation . with an increase in @xmath7 , both the magnus and the hydrodynamic force increase .
however , the nonlinear magnus force increases more rapidly than the linear hydrodynamic force .
thus , the direction of the total force , which is repulsive , approaches the interparticle axis .
a strong enough repulsive force acting on particles not so far from the interparticle direction leads to the formation of hexatic order . to verify this scenario ,
we have performed brownian dynamics simulation ( without hydrodynamic interactions ) by changing the angle @xmath17 between the direction of an artificial repulsive force and the interparticle direction .
we find that when @xmath17 is continuously decreased , a system indeed forms hexatic order below a critical value of @xmath17 ( see the lower panels of fig .
[ fig : fig2]b )
. next we discuss the nature of the translational motion of rotating disks in the disordered states . to see this
, we calculate the mean - square displacement of disks @xmath34 ( see fig . [
fig : fig7]a ) .
interestingly , in the two types of disordered states particle motion is apparently diffusional : in the long - time limit we observe the relation @xmath35 , where @xmath36 is the effective diffusion constant and @xmath37 is the time duration . it should be noted that in our system there is no thermal noise and an isolated rotating disk exhibits no motion . for a pair of rotating particles , we also observe their trajectory is very stable and there is no fluctuation ( see fig .
[ fig : fig1]d )
. thus , the apparently diffusional motion should be the consequence of self - generated force noise due to many - body hydrodynamic interactions of both linear and nonlinear origin .
we can also see no diffusion , or non - ergodic behaviour , for the hexatic state .
the random nature of fluctuations may be related to the long - range nature of hydrodynamic interactions , which makes a number of the surrounding particles affecting a particle large enough to provide strong _ stochastic _ spatio - temporal fluctuations . for high @xmath7 ,
the particle diffusion starts to become comparable or faster than momentum diffusion : @xmath38 , where @xmath39 is the momentum diffusion constant or the kinematic viscosity @xmath40 ( see fig . [
fig : fig7]a and b ) .
this implies that hydrodynamic interactions induced by the rotation of a particle can not fully propagate to its neighbouring particles for high @xmath7 .
this weakens the repulsive interactions of particles and eventually leads to the melting of the hexatic state .
thus the re - entrant hexatic ordering as a function of @xmath7 may be explained as follows : the increase of the magnus force of nonlinear origin with an increase in @xmath7 makes the direction of the interparticle force more aligned along the interparticle direction and also increases the strength of the repulsive force .
this leads to stabilization of the hexatic order , as explained above .
the ordered state is stable until the repulsive interaction is weakened by the intrinsically kinetic nature of hydrodynamic interactions : unlike ordinary interparticle interactions which propagate with the speed of light , hydrodynamic interactions propagate much slower in a diffusive manner with the momentum diffusion constant @xmath39 .
this kinetic weakening of the repulsive force eventually destabilizes the hexatic state and leads to the melting into the disordered chaotic state .
here we summarize what we observed in our system and show the state diagram as a function of @xmath8 and @xmath7 ( see fig .
[ fig : fig7]c ) .
we can see the three states , i.e. , disordered fluid , cluster , and ordered hexatic state .
it is quite striking that particles interacting by hydrodynamic interactions alone exhibit such rich phase behaviours .
finally , we show other interesting states formed by rotating particles .
the introduction of size polydispersity to hard spheres is known to lead to the formation of a glass state for a thermal system @xcite .
motivated by this , we introduce the polydispersity in the rotating speed of particles , whose variance is @xmath41 , and indeed find a non - ergodic glassy state of the rotating particles ( fig .
[ fig : fig8]a ) for high enough @xmath41 ( fig .
[ fig : fig8]b ) : liquid - glass transition in an athermal system . reflecting the nearly continuous nature of the liquid - to - hexatic transition @xcite ,
there is no sharp transition from a hexatic to a disordered glassy state .
although disks are rotating around their centres of mass , their positions are frozen in a disordered configuration and thus the system can be regarded as a nonergodic glassy state . here
we show how the distribution of the rotation speed of particles leads to the loss of hexatic order and results in the formation of a glassy state . as shown in fig .
[ fig : fig9]a , the larger deviation from the averaged @xmath7 leads to the larger deviation from the average number of nearest neighbours ( = 6 ) . with an increase in the variance of the distribution @xmath41 ,
more defects are produced and the hexatic order is eventually lost above @xmath42 .
we can also see that the the interparticle distance is larger for a particle rotating with a faster speed because of stronger hydrodynamic repulsive force , i.e. , magnus force ( fig .
[ fig : fig9 ] ) .
both of these disorder effects are responsible for the formation of a glassy state : the number of the nearest neighbours and the average distance to the neighbours are both strongly correlated to the rotational speed of particles , which explains a wide enough distribution of the rotation speed results in the formation of a non - ergodic amorphous state instead of a hexatic ordered state .
this glassy state is non - ergodic if we consider the particle configuration , yet maintains strong flow fields , which makes this state very unique . reflecting the kinetic origin of interparticle interactions , the introduction of disorder in the dynamic quantity , @xmath0 , is essential for avoiding the ordering , which is an interesting point unique to purely kinetic athermal systems .
we also find a plastic crystal - like state in a rotating particle pair ( dumbbell ) system ( fig .
[ fig : fig8]c ) above a critical @xmath7 ( fig .
[ fig : fig8]d ) .
this state is characterized by hexatic ordering of the centre of mass of disk pairs ( dumbbells ) without any orientational order in the axis directions of dumbbells .
so we find for this type of athermal systems made of rotating particles almost all states seen in its thermodynamic counterparts , including a liquid , a hexatic phase , a glass , and a plastic crystal .
the only missing state is a liquid - crystalline state , but the lack of this state is natural consequence of the fact that the system is composed of rotating elements .
we also note that the modification of the rotational direction leads to complex behaviours ; for example , we can introduce phase demixing of particles rotating oppositely ( fig .
[ fig : fig8]e ) , which may be used to separate different types of passive and active rotors . in relation to this
, it is worth noting that phase separation between particles rotating clockwise and anti - clockwise was recently observed even without hydrodynamic interactions @xcite .
it is interesting that such phase separation is observed in both systems with and without hydrodynamic interactions .
it is also worth noting that the presence of regular arrays of vortices , e.g. , the triangle state , are predicted for an active polar film @xcite .
so far we have not seen such a regular state , but the similarity in the physics between the two systems implies that it might exist in a certain parameter range .
this is an interesting problem for future study .
it is remarkable that our athermal system where particles interact only via hydrodynamic interactions exhibits such rich phase ( or more strictly , state ) behaviour and reproduces almost all the physical states observed in its thermal counterpart .
we hope that these phase behaviours will be observed experimentally . for this purpose , a quasi-2d version of the experimental setup used in @xcite
may be suitable , since the 2d nature of hydrodynamic interactions is important .
here we have focused on the bulk behaviour of rotating disks to study the fundamental nature of dynamical phase behaviour from a viewpoint of nonequilibrium statistical physics .
the effects of confinement on such a system are also quite interesting not only from a fundamental viewpoint , but also from a viewpoint of applications to microfluidics @xcite .
such effects on rotors in a fluid have recently been numerically studied by gtze and gompper @xcite .
it is quite interesting to study how such spatial confinements affect all the dynamical phases we reported and the transition between them . here
it is worth stressing that even non - ergodic states of particles such as hexatic , glassy , and plastic crystal states have strong hydrodynamic flow fields , which makes these states quite distinct from their thermal counterparts .
the interesting and unique feature of hydrodynamic self - organization is that structural ordering is the consequence of _ self - organization of flow _ dissipating energy and thus even non - ergodic states are maintained by dynamical flow . in nature , there are many dynamical systems in which crucial interactions between elements are of purely hydrodynamic origin .
the coexistence of linear and nonlinear hydrodynamic interactions , the resulting unconventional off - axis force , the finite propagation speed of the interactions , and the significance of hydrodynamic degrees of freedom even for non - ergodic ( apparently static ) states lead to rich and non - trivial self - organization .
we hope that our study sheds new light on _ hydrodynamic self - organization _ and stimulate further study on this intriguing problem
. * simulation method . *
treating hydrodynamic interactions between colloids is difficult even for a thermal system .
there are several methods such as stokesian dynamics , lattice - boltzmann , smooth - particle , and fluid - particle - dynamics ( fpd ) methods . here
we employ our fpd method @xcite , which has an advantage in its theoretical transparency and its applicability to a high reynolds number ( @xmath7 ) regime .
we can access a high @xmath7 regime rather easily particularly because a torque applied externally makes the flow field inside a fluid disk almost exactly that for a solid disk even at high @xmath7 . here
we briefly explain the fpd method @xcite and the physical concept behind it .
a particle whose centre of mass is located at @xmath43 is represented by a smooth viscosity change as @xmath44 , where @xmath45 is the liquid viscosity and @xmath46 is the viscosity inside a colloid particle .
the summation is taken over all @xmath25 particles .
@xmath47 represents particle @xmath48 as @xmath49 + 1\}/2 $ ] , where @xmath3 is the particle radius and @xmath50 is the interface thickness . then the equation of motion to be solved is @xmath51 with @xmath52 , where @xmath53 is the unit tensor . here @xmath2 is the mass density , and we assume that the density of the liquid is the same as that of particles .
@xmath54 is the velocity field , and the pressure @xmath55 is determined to satisfy the incompressibility condition @xmath56 .
here @xmath57 is the force density due to the interparticle interaction determined as @xmath58 , where @xmath59 is the interparticle potential , @xmath60 , and @xmath61 is the area of each particle .
@xmath62 is the force density due to the torque : @xmath63 , where @xmath64 is the unit angular vector in the counter - clockwise direction . @xmath65
is the strength of the torque and @xmath66 leads to the counter - clockwise rotation of a particle . in our fpd method
the particle rigidity is approximately expressed by introducing the smooth viscosity profile , @xmath67 .
the approximation is better for a larger viscosity ratio @xmath68 and a smaller @xmath69 . by multiplying both sides of eq .
( [ navier - stokes ] ) by @xmath70 and then performing its spatial integration , we can straightforwardly obtain an approximate equation of motion of particle @xmath48 : @xmath71 , where @xmath72 and @xmath73 are the mass and the average velocity of particle @xmath48 , respectively . on the right hand side
, @xmath74 is the force arising from the interparticle interaction and the torque , and @xmath75 the force exerted by the fluid .
here we use the following approximate relation @xmath76 for an arbitrary tensor @xmath77 , where @xmath78 is the surface of particle @xmath48 and @xmath79 is the unit outward normal vector to @xmath80 . in practical numerical calculations , the on - lattice velocity field , @xmath81 ,
is evaluated from the physical quantities at time @xmath37 by eq .
( [ navier - stokes ] ) .
then we move particle @xmath48 off - lattice as a rigid body by @xmath82 , where @xmath83 is the time increment of the numerical integration . in our simulation , the units of length @xmath84 and time @xmath85 are related as @xmath86 , which sets both the scaled density and viscosity of the fluid region to unity .
this @xmath85 is a time required for the fluid momentum to diffuse over a lattice size @xmath84 .
the units of stress and energy are @xmath87 and @xmath88 , respectively .
furthermore , we set @xmath89 , @xmath90 , @xmath91 , and @xmath92 .
we confirmed that this choice yields reliable results by comparing them wit the analytical solution for a single particle rotation ( see , e.g. , fig .
[ fig : fig1]a and b ) .
the simulation box used was typically @xmath93 .
to avoid cumbersome expressions , we will use the same characters for the scaled variables below .
we solve the equation of motion [ eq .
( [ navier - stokes ] ) ] by the marker - and - cell ( mac ) method with a staggered lattice under the periodic boundary condition .
* interparticle potentials . * to mimic rotating hard disks , we employ the weeks - chandler - andersen ( wca ) repulsive potential @xcite : @xmath94 , otherwise @xmath95 , where @xmath96 gives the energy scale , @xmath97 and @xmath98 represents the size of particle @xmath99 .
* characterization of structures . *
the 2d radial distribution function @xmath22 was calculated as @xmath100 which is the ratio of the ensemble average of the number density of particles existing in the region @xmath101 to the average number density @xmath102 . here
@xmath25 is the number of particles in the simulation box , whose side length is @xmath103 , and @xmath104 is the increment of @xmath5 .
we are grateful to akira furukawa for discussion on the origin of hydrodynamic attractive interactions .
we also thank jun russo for a critical reading of the manuscript .
this study was partly supported by grants - in - aid for scientific research ( s ) and specially promoted research from the japan society for the promotion of science ( jsps ) and the aihara project , the first program from jsps , initiated by the council for science and technology policy ( cstp ) . * supplementary figure 2 .
switching of the effective hydrodynamic interaction between 3d and 2d .
* decreasing the size of the third @xmath109 dimension of the quasi-2d simulation box ( @xmath110 in @xmath111 and @xmath112 directions ) from 32 to 16 leads to the switching of the interaction from repulsive to attractive , or the switching of the relaxation mode of vortex from cascade to anti - cascade mode .
accordingly , the two rotating disks can have a stable rotating trajectory with a constant radius for the latter . as we show in the main text ( fig .
1a and b ) our fpd method reasonably describe the basic behaviour of a rotating disk . here
we show an additional example which indicates the validity of our fpd method .
supplementary figure 1 shows the relation between the effective exerted torque @xmath108 and the angular velocity @xmath0 of a disk .
we can see a perfect linear relation between them : @xmath113 . here
@xmath108 is the effective torque for our disk , which is described by the smooth profile function @xmath114 and given by @xmath115 , where @xmath116 is the exerted torque . from this analysis
, we obtain @xmath117 . as briefly discussed in the main text , a pair of rotating disks whose rotating axes are parallel have repulsive hydrodynamic interactions .
so the distance between them monotonically increases with time ( see the red curve in supplementary fig .
2 ) . however , decreasing the size of the third @xmath109 dimension of the quasi-2d simulation box ( @xmath110 in @xmath111 and @xmath112 directions ) from 32 to 16 leads to the switching of the relaxation mode of vortex from cascade to anti - cascade mode . accordingly
, the two rotating disks can have a stable rotating trajectory with a constant radius ( see the green curve in supplementary fig .
this is a consequence of the force balance between the repulsive magnus force and the attractive hydrodynamic force . for 2d ,
thus , a pair of particles can form a stable rotating trajectory with a constant radius , as shown in fig . | * self - organization of moving objects in hydrodynamic environments has recently attracted considerable attention in connection to natural phenomena and living systems .
however , the underlying physical mechanism is much less clear due to the intrinsically nonequilibrium nature , compared with self - organization of thermal systems .
hydrodynamic interactions are believed to play a crucial role in such phenomena . to elucidate the fundamental physical nature of many - body hydrodynamic interactions at a finite reynolds number , here we study a system of co - rotating hard disks in a two - dimensional viscous fluid at zero temperature . despite the absence of thermal noise ,
this system exhibits rich phase behaviours , including a fluid state with diffusive dynamics , a cluster state , a hexatic state , a glassy state , a plastic crystal state and phase demixing.we reveal that these behaviours are induced by the off - axis and many - body nature of nonlinear hydrodynamic interactions and the finite time required for propagating the interactions by momentum diffusion . *
self - organization is the autonomous organization of components into patterns or structures and the underlying mechanism can be grouped into two classes : thermal and athermal origin @xcite . compared to the understanding of self - assembly of a thermal system , which can be explained in terms of the free energy , that of an athermal system is far behind due to the lack of a firm theoretical background .
typical examples for the latter is active matter in hydrodynamic environments , which shows many interesting unconventional pattern formation with exotic dynamical features @xcite .
directional swimmers undergoing translational motion , like fishes , tend to swim together while forming a cluster of particular shape @xcite .
the hexatic ordering of directionally self - propelling particles was also recently studied @xcite .
rotation is another important type of motion .
self - organization of passive rotors was first observed in laboratory experiments by grzybowski , stone , and whitesides @xcite .
such rotation can be induced by applying a torque to a particle by magnetic @xcite and optical fields @xcite .
self - organization of active rotors has also been studied intensively .
for example , it was found that there are interesting stable bound states of spinning volvox algae @xcite .
furthermore , it was shown @xcite that micro - organisms , like spermatozoa , self - organize into dynamic vortices and they form an array with local hexagonal order .
this study indicates that large - scale coordination of cells can be regulated hydrodynamically , and chemical signals are not required .
it was also shown recently that self - assembly of rotors is a generic feature of aggregating swimmers @xcite .
there has also been a theoretical prediction for intriguing self - organization of rotating molecular motors in membranes @xcite .
this and related problems have also been investigated by numerical simulations @xcite . as described above ,
there exist two types of rotors @xcite : rotors driven by external torques are called passive rotors , while those that are internally driven are called active rotors and many such examples can be seen in biological systems
. a realistic realization of a truly active system of self - rotors in biological systems may be one in which the particles are torque dipole with no resultant net torque on the system @xcite . to elucidate the physics of self - organization of flow created by rotors at a finite reynolds number ,
a system of hard disks each of which is rotated by an externally applied torque is an ideal model system .
ordering of this system may be regarded as a nonequilibrium counterpart of thermodaynamic ordering of hard disks .
we note that , unlike a thermodynamic system , where a state is selected solely by free energy , dynamic factors such as hydrodynamic interactions also affects the selection of a state of an out - of - equilibrium system .
for example , it has been now recognized that nonlinear hydrodynamic interactions play crucial roles in self - organization of rotating disks @xcite . in other words ,
the phenomena are beyond stokes approximations , and may even have a link to self - organization of vortices @xcite , which is observed at a high reynolds number .
vortex crystals are very interesting such examples @xcite .
the nonequilibrium , nonlinear , nonlocal , non - instantaneous nature of hydrodynamic interactions makes analytical approaches to this problem very difficult and thus numerical simulations are expected to play a crucial role .
this problem also has a link to self - organization of a point vortex system @xcite , but the finite size and solidity of disks lead to far more rich behaviour .
here we employ a fluid particle dynamics ( fpd ) method @xcite , which we developed for studying hydrodynamic interactions between colloidal particles ( methods ) .
this method treats a solid colloid as an undeformable fluid particle inside which the viscosity is considerably higher than the surrounding liquid .
this approximation makes us free from solid - fluid boundary conditions , which significantly simplifies the computation .
this method can quite naturally deal with many - body hydrodynamic interactions even at a high reynolds number . in this communication , we study self - organization of rotating hard disks in a two - dimensional ( 2d ) incompressible liquid by using the fpd method .
like the ising model for magnetic ordering or the hard sphere system for crystallization , this system may serve as a fundamental model system for studying dynamical phase behaviour caused by hydrodynamic interactions between rotating particles .
the situation is similar to the above - mentioned experimental and theoretical works @xcite .
we make our simulation in two dimensions ( 2d ) to compare with the behaviour of its thermal counterpart , 2d hard disks , whose thermodynamic behaviour is reasonably understood @xcite . here
we report surprisingly rich phase ordering behaviours , such as aggregation , re - entrant order - disorder transition , glass transition , plastic crystal formation , and phase demixing in this class of strongly nonequilibrium systems . |
in 1978 , lutz and rybicki showed@xcite that almost complete amplitude analysis of reactions @xmath21 and @xmath22 is possible from measurements in a single experiment on a transversely polarized target .
more recently it was shown@xcite that amplitude analyses of reactions @xmath23 and @xmath24 are also possible from measurements on transversely polarized target .
the work of lutz and rybicki opened a whole new approach to hadron spectroscopy by enabling to study the production of resonances on the level of spin amplitudes rather than spin - averaged cross - sections . in amplitude analyses of these reactions resonances
are observed in mass distribution of recoil nucleon transversity amplitudes @xmath25 and @xmath26 corresponding to recoil nucleon spin `` up '' or `` down '' relative to the scattering plane . in traditional hadron spectroscopy ,
the nucleon spin is irrelevant and resonance should appear in both spin amplitudes with the same mass , with the same width and with similar mass distributions .
the measurements of reactions @xmath0 , @xmath27 and @xmath28 at cern on polarized targets seem to challenge these expectations .
these cern experiments reveal shapes of resonant mass distributions in amplitudes of opposite transversity that are different and dependent significantly on the nucleon spin@xcite .
an important advantage of amplitude analyses is that new resonances can be detected on the level of spin amplitude that are not visible in the spin - averaged measurements on unpolarized targets .
such is the case of the new scalar resonance @xmath2 found in amplitude analyses of reactions @xmath29 at 17.2 gev / c at low momentum transfers @xmath30 0.0050.2 ( gev / c)@xmath31 and in @xmath32 at 5.98 and 11.85 gev / c at larger momentum transfers @xmath33 0.20.4 ( gev / c)@xmath31@xcite .
using monte carlo method of amplitude analysis , it was found in @xcite that in @xmath34 reaction both solutions for spin `` up '' @xmath3-wave amplitude @xmath4 resonate below 880 mev while both solutions for spin `` down '' @xmath3-wave amplitude @xmath5 appeared nonresonating .
this result is in perfect agreement with @xmath6 minimization method of amplitude analysis used in @xcite ( see fig . 1 and 2 of @xcite ) .
detailed quantitative breit - wigner fits to @xmath4 in @xcite showed the importance of coherent background for determination of the resonance parameters of @xmath2 .
the best fit in @xcite gives @xmath35 mev and @xmath36 mev .
the excellent agreement of these values for @xmath37 and @xmath9 with the ellis - lanik theorem@xcite strongly suggests that @xmath2 is the lowest mass scalar gluonium@xcite .
trnqvist expressed concern@xcite that the low mass and the narrow width of @xmath2 is due to the neglect of interference of @xmath2 with @xmath38 ( 980 ) in our breit - wigner fits which initially extended only to 880 mev in dipion mass using the data set of ref .
the apparent nonresonating behaviour of the amplitude @xmath5 below 880 mev was also puzzling .
to investigate these questions we have performed a new amplitude analysis of @xmath39 reaction at 17.2 gev / c and small momentum transfers @xmath330.0050.20 ( gev / c)@xmath31 extending the dipion mass range up to 1080 mev using the data set of ref . [ 7 ] . in this report
we present the results of our detailed study of resonant structure of @xmath3-wave amplitudes @xmath4 and @xmath5 in the mass region of 5801080 mev .
this work is the first attempt to determine resonance parameters of a resonance from fits to production amplitudes of opposite nucleon transversity .
all previous determinations of resonance parameters used spin - averaged mass distributions or , in the case of ref . [ 18 ] , only one spin amplitude @xmath40 .
the new results for @xmath4 shows a narrow resonance below 880 mev , followed by an enhancement above 900 mev and a dramatic dip at @xmath411000 mev corresponding to @xmath1 .
the amplitude @xmath5 shows a broader structure with a dip at @xmath411000 mev .
first we performed separate fits to @xmath4 and @xmath42 using a coherent sum of @xmath2 and @xmath1 breit - wigner amplitudes and constant background .
we find that @xmath8 contributes to both amplitudes but with different widths .
solutions 1 and 2 of amplitude @xmath4 have a narrow width of 100 and 76 mev , respectively .
solutions 1 and 2 of amplitude @xmath5 have a broader width of 217 and 233 mev . the interference with @xmath1 is crucial in reproducing the enhancements above 900 mev and the dip at 1010 mev . the narrow width in the spin `` up '' amplitude @xmath43 and the broad width in the spin `` down '' amplitude @xmath5 is inconsistent with the traditional view of a resonace as a particle with single mass and width . to see
if we can describe the mass spectra of @xmath4 and @xmath5 with a single @xmath8 with the same resonance parameters , we have performed a simultaneous fit to @xmath43 and @xmath5 hoping that the differences in the resonant shapes can be accounted for by differences in the background .
we obtained an acceptable fit but the @xmath6 per data point @xmath7 was significantly higher than @xmath12 in separate fits .
we thus obtained a poorer fit .
our next step was to admit that a narrow @xmath8 contributes to @xmath4 and a broad @xmath8 contributes to @xmath5 .
but a self - consistency requires that the narrow @xmath8 contributes also to @xmath5 , and the broad @xmath8 contributes also to @xmath4 .
we expected these additional contributions to be small in a simultaneous two - pole fit to @xmath4 and @xmath5 .
the results of the fit were surprising .
our best fit had @xmath12 better than the separate fits to @xmath4 and @xmath5 . and the fit required in all amplitudes the presence of two @xmath8 resonances : higher mass and broader @xmath8 with average mass @xmath44 mev and @xmath45 , and a lower mass and narrower @xmath15 with average mass of @xmath16 mev and @xmath46 mev . in
all @xmath3-wave amplitudes the higher mass @xmath8 dominates the @xmath15 contribution .
since this solution gives a very low @xmath12 , it must be taken seriously .
we propose to identify the @xmath18 and @xmath19 as color electric and color magnetic modes of lowest mass scalar gluonium @xmath20 , respectively , following the gluonium classification scheme of bjorken@xcite elaborated in@xcite .
the paper is organized as follows . in section
ii we present and discuss the results of our new amplitude analysis . in section iii
we introduce the parametrization of unnormalized @xmath3-wave amplitudes with a single @xmath8 pole and present the results of separate fits to amplitudes @xmath4 and @xmath5 .
simultaneous fits to these @xmath3-wave amplitudes using a single @xmath8 pole are presented in section iv . in section
v we introduce the parametrization with two @xmath8 poles and discuss the results of simultaneous fits to @xmath4 and @xmath5 using this parametrization . in section vi we discuss the possible constituent nature of the two @xmath8 states @xmath18 and @xmath19 resulting from the best fit .
the paper closes with a summary in section vii .
the high statistics cern - munich measurement of @xmath47 at 17.2 gev / c on polarized target was reported in four data sets with kinematics given in the following table : | we have performed a new model independent amplitude analysis of reaction @xmath0 measured at cern at 17.2 gev / c on polarized target .
the analysis extends the range of dipion mass from 600880 mev used in the previous study to 5801080 mev .
our purpose is to study the role of @xmath1 interference and to better determine the resonance parameters of @xmath2 resonance from fits to both @xmath3-wave amplitudes @xmath4 and @xmath5 with recoil nucleon transversities `` up '' and `` down '' , respectively .
this work is the first attempt to determine resonance parameters of a resonance from breit - wigner fits to production amplitudes of opposite transversity .
we used several fitting procedures to understand the resonant structure of the @xmath3-wave amplitudes in this mass range .
first we performed separate fits to @xmath4 and @xmath5 using a breit - wigner parametrization for @xmath2 and @xmath1 , and a constant coherent background .
we obtained an excellent fit to the data with a low @xmath6 per data point @xmath7 . however the width of @xmath8 is narrow 11076 mev in @xmath4 while it is broad 217233 mev in @xmath5 .
to decide whether the apparent dependence of @xmath9 on nucleon spin is a real physical effect or an artifact of separate fits , we performed simultaneous fits to @xmath4 and @xmath5 .
first we assumed a common single @xmath8 pole .
we found a good fit with @xmath10 mev , @xmath11 mev but the @xmath12 is larger than @xmath12 from separate fits .
we concluded that the results of separate fits may indicate existence of two @xmath8 poles , one with a narrow and the other with a broad width .
self - consistency requires that both poles contribute to both @xmath3-wave amplitudes .
the simultaneous fit to @xmath4 and @xmath5 with two common @xmath8 poles yields an excellent fit with a lower @xmath12 than the separate fits .
all amplitudes are dominated by a @xmath8 state with @xmath13 mev and @xmath14 mev .
all amplitudes receive a weaker contribution from a narrower @xmath15 state with @xmath16 mev and @xmath17 mev .
we propose to identify @xmath18 and @xmath19 with colour - electric and colour - magnetic modes of lowest mass scalar gluonium @xmath20 . |
in the galactic center ( gc ) region , @xmath2 and @xmath1 found the large - scale thin - thermal plasma with highly ionized atomic lines ( @xcite ) . on the other hand , murakami et al .
( 2000 ; 2001a ; 2001b ) discovered clumps with a neutral iron line and proposed the x - ray reflection nebulae ( xrne ) scenario .
clumps of highly ionized iron line are also discovered with @xmath0 and are inferred to be young snrs ( e.g. @xcite ) . in this paper , we report on new diffuse x - ray structures around the galactic center , the molecular cloud and , and the regions revealed with the excellent spatial resolution of @xmath0 and summarize their characteristics .
we use the @xmath0 archive data of the acis - i array on the radio arc , sgr a@xmath3 , sgr b2 , and sgr c regions .
the field of view ( fov ) of each observation is @xmath4 .
the on - axis position @xmath5 and exposure time are given in table [ abamba - e1_tab : tab1 ] .
.the on - axis position and exposure time in each observation .
[ cols="<,^,^ " , ]
figure [ abamba - e1_fig : fig1 ] shows the 3.08.0 kev band images around sgr b2 ( figure [ abamba - e1_fig : fig1]a ) , radio arc and sgr a@xmath3 ( mosaic ; figure [ abamba - e1_fig : fig1]b ) , and sgr c ( figure [ abamba - e1_fig : fig1]c ) .
many diffuse or filamental structures ( no 120 indicated in figure [ abamba - e1_fig : fig1 ] ) are found .
we performed the spectral fitting for all the diffuse sources ( no 120 ) .
background spectra were taken from around each source .
the background - subtracted spectra were fitted with a model of power - law continuum plus a narrow gaussian line .
then we classified all clumps by the best - fit line profile ; 6.4 kev line clumps , 6.56.7 kev line clumps , and clumps with no significant line .
the fitting results are separately given in table [ abamba - e1_tab : tab2 ] ( a , b , and c ) . p6pcccccc + + source no .
& @xmath6 & center energy & equivalent width & @xmath7 & flux ( 210 kev)@xmath8 + & & [ kev ] & [ kev ] & [ @xmath9 & [ @xmath10 + + + 3 & 1.8 ( @xmath111.05.8 ) & 6.34 ( 6.266.41 ) & 0.9 ( 0.51.3 ) & 33 ( 1751 ) & @xmath12 + 5 & 0.5 ( @xmath110.61.1 ) & 6.43 ( 6.376.55 ) & 1.3 ( 0.81.8 ) & 4 ( 110 ) & @xmath13 + 6 & 1.6 ( @xmath110.76.6 ) & 6.35 ( 6.296.41 ) & 1.6 ( 0.82.4 ) & 12 ( 330 ) & @xmath14 + 7 & 3.5 ( 2.04.5 ) & 5.97 ( 5.856.42 ) & 1.1 ( 0.62.4 ) & 8 ( 512 ) & @xmath15 + 9 & @xmath110.1 ( @xmath111.11.3 ) & 6.41 ( 6.366.45 ) & 0.8 ( 0.51.1 ) & 10 ( 227 ) & @xmath16 + 13 & 1.0 ( 0.21.7 ) & 6.41 ( 6.376.45 ) & 1.7 ( 0.388 ) & 40 ( 2063 ) & @xmath17 + 14 & 0.7 ( @xmath112.04.6 ) & 6.42 ( 6.386.47 ) & 1.9 ( @xmath18 0.08 ) & 20 ( 1243 ) & @xmath19 + 15 ( sgr c)@xmath20 & 2.0(fixed ) & 6.36 ( 6.086.44 ) & 2 ( fixed ) & 16 ( 934 ) & @xmath21 + 16 & 2.7 ( 1.63.7 ) & 6.42 ( 5.627.94 ) & 1.4 ( 0.72.7 ) & 7 ( 412 ) & @xmath22 + 17 & 7.3 ( @xmath182.6 ) & 6.46 ( not determined ) & 1.8 ( @xmath23 5.0 ) & 41 ( 1267 ) & @xmath24 + 19 & 3.7 ( 0.59.1 ) & 6.37 ( 6.306.40 ) & 1.4 ( 0.82.63 ) & 34 ( 1778 ) & @xmath25 + 20 ( arches)@xmath26 & 5.0 ( @xmath182.0 ) & 6.38 ( 6.346.42 ) & 1.7(0.92.2 ) & 34 ( 1359 ) & @xmath27 + + 1 & 3.9 ( 2.27.1 ) & 6.64 ( 6.526.88 ) & 0.5 ( 0.11.2 ) & 57 ( 3871 ) & @xmath28 + 8 & 1.9 ( 0.34.3 ) & 6.83 ( 6.637.01 ) & 2.6 ( 0.44.6 ) & 4 ( 212 ) & @xmath29 + 12 & 9.4 ( @xmath18 5.8 ) & 6.62 ( 6.586.63 ) & 19 ( @xmath18 6.6 ) & 44 ( 2152 ) & @xmath30 + 18 & 3.5 ( 2.54.5 ) & 6.70 ( 6.566.83 ) & 2.3 ( 1.03.6 ) & 8 ( 611 ) & @xmath31 + [ + 3pt ] + 2 & 2.5 ( 1.34.5 ) & & & 11 ( 721 ) & @xmath32 + 4 & 2.5 ( 0.85.0 ) & & & 7 ( 318 ) & @xmath33 + 10 & 1.0 ( @xmath110.96.5 ) & & & 2 ( @xmath23 15 ) & @xmath34 + 11 & 1.3 ( 0.82.0 ) & & & 6 ( 39 ) & @xmath35 + [ + 3pt ] + + + + +
most of the 6.4 kev line clumps show deep absorptions in excess to that toward the gc region , similar to the xrn , sgr b2 ( murakami et al . 2000 ) .
they hence are regarded as new xrn candidates .
figure [ abamba - e1_fig : fig2 ] is the 6.07.0 kev image around sgr c superposed on the cs j=10 contours ( @xcite ) .
the spectrum of sgr c shows the neutral iron line and deep absorption .
the iron line emission is shifted toward the gc side of the molecular cloud , as is found for the first xrn , sgr b2 .
therefore , we fitted the spectrum of sgr c region with an xrn model . since the statistics is limited , we fixed the photon index to 2.0 ( @xcite ) and the equivalent width of the 6.4 kev line to 2.0 kev ( @xcite ) , expected value for xrne .
the fitting is acceptable with the best - fit parameters shown in table [ abamba - e1_tab : tab2 ] .
thus , we confirmed the @xmath1 result that sgr c is an xrn ( @xcite ) .
@xmath1 found the 6.4 kev line emission near at the molecular cloud ( @xcite,@xcite ) .
the @xmath0 observations revealed the x - ray structure has two clumps of no 9 and 10 .
the 6.4 kev lines are mainly coming from no 9 , a filament lying along the southwest half of the cloud , hence may be fluorescence irradiated by a gc side source .
another candidate of xrn , but may be unrelated to sgr a@xmath3 is an @xmath36-ellipse located near the active star forming region , the arches cluster at a distance of 8.5 kpc ( @xcite ) .
the spectral parameters show a strong neutral iron line and deep absorption , supporting the xrn scenario .
the irradiating sources , however are likely active young stars in the arches cluster ( @xcite ) .
we find three x - ray bright stars with the total luminosity of a few @xmath37 , assuming a single temperature plasma .
however the required luminosity to produce the diffuse 6.4 kev flux is a few @xmath38 , 10 times larger than that of the core stars .
since x - rays from young stars are more active and variable in younger age , it is conceivable that big flares and/or bursts were more frequently occurred in the recent past than in the present .
the other 6.4 kev clumps no 3 , 57 are located in between sgr a@xmath3 and the radio arc , and show no clear association with molecular clouds .
hence the origin of these sources is puzzling .
since the 6.56.7 kev line comes from highly ionized iron , we can infer that these clumps consist of hot plasmas .
an x - ray shell no 1 is associated with a non - thermal radio source `` wisp '' ( @xcite ) , suggesting to be a part of the radio snr , ( @xcite ) .
in fact , we see diffuse x - rays filling in the east half of ( dotted circle in figure 1b ) , while the west half including no 1 may be behind the molecular cloud , hence is faint and/or largely absorbed in x - ray .
the x - ray spectrum is well fitted with an nei model of a temperature @xmath39 kev , significantly higher than any other young snrs , but comparable to the diffuse emission prevailing in the whole gc region .
together with the @xmath0 discovery of new x - ray snrs , ( @xcite ) and ( @xcite ) , we suspect that many other young x - ray snrs would be found in the further deep observations .
these putative snrs may significantly contribute to the large scale diffuse x - rays near the gc .
no 18 has a plume - like shape lying at the northwest of sgr a@xmath3 and perpendicular to the galactic plane .
the plasma temperature is lower than that in the gc .
this source may not be a single snr but is likely to be an out - flow from the gc plasma .
an attractive idea is that no 18 is a thermal jet emanating from the giant black hole sgr a@xmath3 .
the presence of non - thermal radio filaments is one of the the most striking features in the gc region .
they are all oriented perpendicular to the galactic plane , and are considered to be synchrotron emission in a strong magnetic field of @xmath401 mg .
@xmath0 discovered non - thermal x - ray filaments from sgr a east ( no 2 ) and near the radio arc ( no 10 ) ( @xcite ) .
a hint of possible association with a weak radio filament is suggested for these x - ray filaments .
we thus infer that the x - ray filaments are due either to the synchrotron emission or inverse compton in a relatively weak magnetic field , where the radio flux is rather faint .
since no 10 is located at the east rim of the molecular cloud co [email protected] , where the expanding cavity is interacting with the cloud ( @xcite ) , the fermi acceleration scenario is likely to produce high energy electrons . the synchrotron energy loss of high energy electrons may not be large , hence can emit even synchrotron x - rays
. if the magnetic field is moderately weak , x - rays are more likely from the inverse compton process as proposed for the x - ray filament at the radio `` thread '' ( @xcite ) .
no 11 is a faint x - ray belt running parallel to the galactic plane and probably extending to the other non - thermal filament no 4 .
1 . many diffuse x - ray clumps with different morphology are found from the gc region .
2 . the x - ray spectra are full of variety in line feature : iron lines of 6.4 kev , 6.56.7 kev , and line - less clumps . 3 . the molecular clouds sgr c and [email protected] are likely xrne , irradiated by the past active sgr a@xmath3 ( murakami et al .
the diffuse 6.4 kev x - rays near the arches cluster would be due to the irradiating of flaring young stars in the cluster .
some of the other 6.4 kev clumps located in between sgr a@xmath3 and the radio arc regions show no clear association with molecular clouds , hence the origin is debatable .
5 . the 6.56.7 kev clumps may be young snrs , but the temperatures are higher than the usual snrs . 6 .
the line - less clumps ( filaments ) would be either synchrotron or inverse compton x - rays in weaker magnetic field than those in the radio bright filaments .
, ho p.t.p .
2000 , apj , 533 , 245 inoue h. 1985 , space science review , 40 , 317 koyama k. , awaki h. , kunieda h. , takano s. , tawara y. 1989 , nature , 339 , 603 koyama k. , maeda y. , sonobe y. , takeshima t. , tanaka y. , yamauchi s. 1996 pasj , 48 , 249 koyama . k. 2001 , `` new century of x - ray astronomy '' , asp conference series , vol . 251 , 50 , eds
. h. inoue and h. kunieda , astroph/0108109 ho p.t.p .
, jackson j.m . , barrett a.h . , armstrong , j.t .
1985 , apj , 288 , 575 maeda y. baganoff f.k .
, feigelson e.d . , morris m. , bautz m.w .
, brandt w.n . , burrows d.n . , doty j.p .
et al . accepted by apj , astroph/0102183 murakami h. , koyama k. , sakano m. , tsujimoto m. , maeda y. 2000 , apj , 534 , 283 murakami h. , koyama k. , tsujimoto m. , maeda y. , sakano m. 2001a , apj , 550 , 297 murakami h. , koyama k. , maeda y. 2001b , apj , 558 , 687 nagata t. , woodward c.e . , shure m. , kobayashi n. 1995 , aj , 109 , 1676 oka t. , hasegawa , t. , sato , f. , tsuboi , m. , miyazaki , a. 2001 , pasj , 53 , 779 senda a. , murakami h. , koyama k. 2002 , accepted by apj , astroph/0110011 tsuboi m. , handa t. , ukita n. 1999 , apjs , 120 , 1 wang q.d .
, gotthelf e.v .
2002 , accepted by nature , astroph/0201070 yusef - zadeh f. , law c. , wardle m. , wang q.d .
, fruscione a. , lang c.c . , cotera a. 2001 , astroph/0108174 | this paper reports the diffuse x - ray features around the galactic center observed with @xmath0 .
we confirm the @xmath1 and @xmath2 discoveries of the large - scale thin - thermal plasma with strong lines in the galactic center region .
in addition , many small clumps of emission lines from neutral ( 6.4 kev line ) to he - like ( 6.7 kev line ) irons are discovered .
the 6.4 kev line clumps would be reflection nebulae , while those of the 6.7 kev line are likely snrs .
we also find emission lines of intermediate energy between 6.56.7 kev , which are attributable to young snrs in non equilibrium ionization .
non - thermal filaments or belts with x - ray spectra of no emission line are found , suggesting the fermi acceleration site in a rapidly expanding shell .
all these suggest that multiple - supernovae or extremely large explosion had occurred around the galactic center region in the recent past . |
in terms of the quark model , one can observe a spontaneous symmetry breaking of the approximate @xmath9 symmetry , resulting in nine pseudo - goldstone bosons , that are usually associated with the light mesons formed as @xmath10 quark - antiquark bound states .
this nave picture is , however , complicated by the fact that the @xmath1 meson has a large mass of the order of 1 @xmath11 .
as early as 1970 , kobayashi and maskawa concluded that the large mass of the @xmath7meson ( formerly known as @xmath12 ) is a serious problem that is difficult to understand in a chiral @xmath9 model with an explicit symmetry breaking term between singlet and octet states @xcite .
they found that the existence of an effective six - quark determinantial vertex is necessary .
as shown by t hooft in 1976 , this vertex is contained in instanton - induced quark interactions @xcite .
an interesting aspect of this kobayashi - maskawa-t hooft or kmt term @xcite is that it can give rise to a flavor mixing in the scalar as well as in the pseudoscalar channels .
the coupling between the pseudoscalar singlet and octet states @xmath13 and @xmath14 arises both from the @xmath15 breaking and the anomaly terms , assuming isospin symmetry .
the physical @xmath16 and @xmath7mesons are given by the mixing of the @xmath14 and @xmath13 modes , and the mass of the @xmath13 singlet state turns out to be sensitive to the strength of the kmt vertex .
an explicit calculation for the general case gives the mixing angle @xmath17 @xcite .
the @xmath16 and @xmath1 mesons change their masses as a function of the temperature @xmath18 , due to both the @xmath18 dependence of the quark condensate , and the possible decrease of the kmt coupling constant with increasing @xmath18 .
the mixing angle @xmath19 also becomes @xmath18 dependent : as the temperature increases , mixing between @xmath16 and @xmath1 approximates the ideal one , and the @xmath13 component in the physical @xmath1 decreases . on the other hand , with increasing @xmath18 , the @xmath13 tends to play the role of the ninth nambu - goldstone boson of the @xmath20 symmetry , and loses its mass rapidly . at low temperatures , the @xmath21 part of the symmetry
is broken by instantons , invoking distinct vacuum states .
tunneling between these vacuum states is only possible `` at a cost '' , giving extra mass to the @xmath7meson .
however , as the transition amplitude is dependent on the strong coupling constant @xmath22 , it follows that the effect of instantons rapidly decreases with increasing energy density .
this is an effective restoration of the @xmath21 symmetry at finite @xmath18 , first suggested in ref .
thus in high energy heavy ion collisions , where a hot and dense medium is created , the @xmath21 symmetry may temporarily be restored @xcite . in refs .
@xcite we reported the first observation of a significant reduction of the @xmath7mass , based on an analysis of phenix and star data @xcite from @xmath23=200 gev central au+au collisions at the bnl relativistic heavy ion collider ( rhic ) .
the subject of the present manuscript is to detail and systematically explore these experimental signatures of a partial @xmath21 symmetry restoration at rhic . in thermal models ,
the production cross sections of the light mesons are exponentially suppressed by the mass .
since the @xmath7mesons are heavy , by default one expects the number of @xmath7mesons to be about two orders of magnitude less than the number of pions .
however , as a consequence of the mass reduction , this suppression would be moderated , and the @xmath1 mesons would show up in an enhanced number . once produced , the @xmath7is expected to be decoupled from other hadronic matter , since its annihilation cross section is very small . at the same time , the low-@xmath24@xmath7mesons are trapped in the medium due to energy conservation reasons , forming a condensate . due to expansion , the medium cools down and freezes out , emitting asymptotic particles that propagate in @xmath25 vacuum . by this time , the @xmath7mesons regain their original mass , hence the enhancement will mostly appear at low @xmath2 @xcite .
it is to be noted that the @xmath7lifetime is much longer than the lifetime of the hot and dense medium , therefore a direct observation of the mass shift seems to be extremely difficult .
a promising channel of observation is the dileptonic decay @xmath26 , because a low-@xmath24@xmath7enhancement would give extra lepton pairs to the low invariant mass region .
the paper of kapusta , kharzeev and mclerran on the return of the prodigal goldstone boson @xcite was in fact motivated by the dilepton enhancement seen in cern super proton synchrotron ( sps ) @xmath27 laboratory energies in s+pb reactions .
recent interpretations of ceres @xcite and na60 data @xcite indicate that the approach to a chiral symmetry restored state could proceed through resonance broadening and eventually subsequent melting , rather than by dropping masses or mass dependency or mass degeneracy between chiral partners @xcite .
recent phenix findings also show a definite excess in the @xmath28 dielectron invariant mass region in @xmath5 au+au collisions @xcite .
unlike at lower beam energies , in this case the contribution from a hot hadronic phase without mass shifts seems to be insufficient to account for the enhancement seen in the data @xcite . in the present work
we search for the @xmath7mass modification and the related restoration of the chiral symmetry in a different channel , using already published like - sign bose
einstein correlation ( bec ) measurements as proposed by ref .
we report on a detailed systematic study of the presently available published data sets .
future high precision data points in the low-@xmath24region will help to reduce the uncertainties of the measurements , and hopefully will provide a more precise estimation of the @xmath7mass reduction , compared to our current analysis .
correlations between pions carry important information about the space - time structure of the medium created in heavy ion collisions .
the widths of @xmath16 and @xmath7are @xmath29 and @xmath30 , corresponding to large decay times : they produce pions at @xmath31 and @xmath32 , which are huge compared to the characteristic hbt radii of 46 fm . among the decay channels of the @xmath7 ,
the @xmath33 channel has the largest branching ratio of about 45% .
furthermore , the @xmath16 mesons decay into charged pions : the @xmath34 and @xmath35 processes together have a branching ratio of approximately 27% @xcite .
einstein correlation function of pion - pion pairs can be expressed in terms of the relative and the mean four - momenta , @xmath36 and @xmath37 , respectively : @xmath38 where are the three - momenta of particles 1 and 2 , @xmath39 and @xmath40 are the one- and two - particle invariant momentum distribution functions .
notation is used here , where @xmath41 and is the three - momentum . ]
it has been shown that the source can be handled in the core
halo picture , where the `` core '' consists of primordially created pions and those ones coming from fast - decaying resonances , while the other pions , coming from more slowly decaying resonances , make the `` halo '' @xcite .
the core region is resolvable by bec , while the halo is not .
in fact , the core halo separation always depends on the experimental two - track resolution .
for example in phenix and star , two tracks are separable with a momentum difference of @xmath42 larger than 4 to 5 mev , corresponding to a spatial separability @xmath43 that dies off at about 4050 fm @xcite .
long tails extending to this region were recently observed by the phenix and na49 collaborations at rhic and cern sps energies @xcite .
note that these long tails are also seen in kaon imaging , while the bulk production is well described with characteristic scales of 46 fm @xcite . in the core
halo picture @xcite the correlation function can be measured with the so called extrapolated intercept parameter @xmath44as @xmath45 where @xmath46 is defined by the fourier transform of the one - pion emission function @xmath47 of the core as @xmath48 and the intercept parameter @xmath44 is derived from the _ extrapolation _ of the correlation value to @xmath49 , as @xmath50 . however , this extrapolation does not include the correlation between halo halo and core halo particle pairs , supposed to be unresolvable by our detectors .. generally , @xmath51 . ] the intercept parameter can also be expressed directly with the fraction of the core pions to the total number of pions : @xmath52 where @xmath53 and it is per se sensitive to the core vs. halo ratio .
( in the above formulae @xmath54 denotes the transverse mass of the @xmath55s , and the @xmath56 s are the @xmath57 distributions of the corresponding decays . )
if one does a one - dimensional investigation characterized by the invariant momentum difference @xmath58 , and assuming a gaussian approximation of the source , one will have eq .
( [ eq : clams ] ) in the form @xmath59 in this case @xmath60 . as @xmath44is defined in eq . ( [ eq : clams ] ) as an _ extrapolation _ of the experimental data to @xmath61 , it is essential that it does not depend strongly on the method of the extrapolation . the gaussian assumption for the shape of the source is a model hypothesis that has been shown not to be precise @xcite . instead , one can choose an experimental procedure independent of theoretical assumptions , and get a description of arbitrary precision with the edgeworth expansion by completing the gauss shape with an infinite series in the space of hermite polynomials @xcite . in this case , @xmath62,\end{aligned}\ ] ] where @xmath63 is an overall normalization parameter , the @xmath64 is a weight factor to the @xmath65th hermite polynomial @xmath66 and @xmath67 and @xmath68 are the edgeworth radius and intercept parameters . in practice , an expansion up to
the sixth order is sufficient @xcite .
note that , as the even order hermite polynomials have non - vanishing values at @xmath69 , @xmath68 is different from @xmath3 here : @xmath70 & .\end{aligned}\ ] ] a straightforward generalization of these approximations to the three ( out , side , long ) dimension case is given in @xcite . in the following
we will show that the phenomenon of a partial @xmath8 restoration and a connected mass reduction of the @xmath1 meson is able to explain the behavior of @xmath3 measured in rhic experiments . a previous analysis , based on the same idea ,
was carried out on na44 data , and showed that a reduction of the @xmath71 mass should result in a dip of the @xmath72 at low-@xmath57 values .
however , no significant signal of such a mass modification was seen in @xmath73 s+pb collisions at cern sps ( fig . 1 of @xcite ) .
in contrast to the sps data , a low-@xmath57dip of the @xmath72 was measured in @xmath74 au+au collisions at rhic by both the star and the phenix experiments @xcite .
the @xmath72 points were extracted from the measured correlation function of like - sign pions , using different methods .
a comparison of the measurements to the fritiof @xcite calculation including variation of the @xmath7mass was presented in fig .
4 . of ref .
it is important to note that , although the gaussian fit typically has 12% error on the @xmath75 value , this does not reflect the real error on the intercept caused by the extrapolation .
for example , exponential fits yield larger values for @xmath75 than the gaussian fits do , and the difference between @xmath76 and @xmath77 is larger than several ( sometimes more than 5 ) standard deviations @xcite .
a reasonable range of errors can be estimated with the help of an edgeworth fit @xcite . in the detailed presentation of ref . @xcite it was suggested to utilize the normalized @xmath78quantity in order to remove sensitivity to the extrapolation technique , and to reduce other systematic errors .
note that @xmath79 is defined as the value of the extrapolated intercept parameter in an @xmath80 region where its value is saturated ; according to simulations as well as the presently used phenix and star measurements , this corresponds to 0.5 gev @xmath81 0.7 gev region .
is the @xmath72 value taken at @xmath82 , with the exception of the star data , where the data point at the highest @xmath83 is considered .
note that the @xmath57dependency of the @xmath72 measurements in the 0.50.7 gev region is very weak .
] one should also note that the @xmath72 data may depend on the goodness of the particle identification too : other particles misidentified as pions will reduce correlation and will push the measured @xmath72 data down .
the phenix @xmath72 data set was derived using the bowler - sinyukov method @xcite , while the star data set was obtained from a sixth order edgeworth fit .
here we apply the method of ref .
@xcite using the values and errors of the gaussian @xmath75 and the @xmath84 edgeworth fit parameters taken from @xcite in order to compute the sixth order @xmath3 values using eq .
( [ eq : lame ] ) .
we use complete error propagation in order to determine the corresponding uncertanities of @xmath3 .
although the gaussian @xmath75 and the edgeworth @xmath68 parameters are significantly different , the extrapolated intercept parameter @xmath44is similar to @xmath75 within errors ( see table [ tab : starlam ] and fig .
[ fig : starlam ] in appendix [ sec : starlam ] ) .
however , the error bars on @xmath3 are significantly larger than the nave errors on @xmath75 . as emphasized before , the edgeworth @xmath3 gives a more realistic estimation of both the value and the error of the intercept parameter , hence of the core - halo ratio .
another data set on @xmath44from star was obtained with a gaussian fit using the bowler - sinyukov method , and it shows a good agreement with the star edgeworth data ( fig . 14 of @xcite ) .
we do not use this data set in our analysis , since the error bars do not include all the relevant systematic effects , being in the order of @xmath85 , which is way smaller than the systematic error coming from the choice of the extrapolation .
phenix preliminary data @xcite are also shown , for comparison purposes only , as their systematic errors are not yet finalized .
the data sets for @xmath44are detailed in fig .
[ fig : datasets ] before and after normalizing with @xmath86 .
each of these data sets indicate the dip of @xmath44 in the low-@xmath24region .
given that we are interested in obtaining final errors that include all relevant systematic effects , we decided to analyze simultaneously the phenix final sinyukov - corrected @xmath44data set together with the star edgeworth @xmath44data set , both normalized to their @xmath86values .
one can note that these two data sets are not in perfect overlap with each other , although they are consistent within their errors , and it is possible to fit both data sets simultaneously with good confidence levels .
the gaussian @xmath87 data set of star has errors that apparently do not include systematics from the gaussian ansatz , thus a quantitative comparison to our model was not reasonable .
however , the best fits to the other data sets qualitatively agree with the gaussian @xmath87 points .
we have also checked that separate analysis of the star and phenix data sets yields results which are consistent with the presented results , as part of the systematic studies .
however , the combined phenix and star data set provided a more precise estimate for the allowed regions of the model parameters . _
( left ) _ and @xmath78_(right ) _ from rhic @xmath74 like - sign pion correlation measurements .
note that the errors on the star gaussian data set ( fig .
14 of ref .
@xcite ) are in the order of 0.001 mev ( not all systematics included ) , and one of the phenix data sets has only preliminary errors @xcite . hence we are left with the remaining two data sets when evaluating the systematic errors on @xmath88 : the published phenix data set of ref .
@xcite , and the star data set calculated from the data shown in fig . 13 of ref .
@xcite using the method of appendix [ sec : starlam].,title="fig : " ] _ ( left ) _ and @xmath78_(right ) _ from rhic @xmath74 like - sign pion correlation measurements .
note that the errors on the star gaussian data set ( fig . 14 of ref .
@xcite ) are in the order of 0.001 mev ( not all systematics included ) , and one of the phenix data sets has only preliminary errors @xcite . hence we are left with the remaining two data sets when evaluating the systematic errors on @xmath88 : the published phenix data set of ref .
@xcite , and the star data set calculated from the data shown in fig . 13 of ref .
@xcite using the method of appendix [ sec : starlam].,title="fig : " ] the difference between the selected phenix and star data sets possibly reflects the systematic error from different experimental conditions , for example , phenix data were measured in the 0 - 30% centrality class , while star edgeworth results were published for the 05% centrality selection , and the particle identification in the two experiments is also different .
these differences resulted in a systematic uncertainty of our analysis , too , however , as we shall demonstrate at the end , these systematic errors are still of the order of the statistical uncertainties and the dominant error in estimation of the in - medium modified mass of the @xmath1 mesons comes from the choice of the resonance model and its parameters .
the relative systematic error from the difference in centrality selection is estimated in appendix [ sec : centrality ] to be not larger than 9.8% .
first simulations of the @xmath8restoration at sps s+pb collisions at @xmath73 bombarding energy had predicted a dip of @xmath72 at @xmath57values below 0.25 @xmath11 @xcite . it was found that the depth of this dip was governed by the value of @xmath88as an input parameter for the simulations . in those simulations ,
the @xmath7mesons from the condensate had been assumed to have no transverse momenta at all , resulting in a very steep hole - like structure of the low-@xmath57part of the @xmath72 @xcite .
such an oversimplification is not adequate for the description of the dip in the rhic data . to improve on it ,
we introduce here an effective thermal spectrum for the @xmath7mesons from the condensate , characterized by an inverse slope parameter @xmath89 .
as we demonstrate below , this parameter @xmath89controls the steepness of the dip of @xmath4 , while @xmath88controls its depth . also , instead of relying on a given model of resonance production ( as on fritiof @xcite in ref .
@xcite ) , here we utilize six different models , in order to estimate the systematic error related to the choice of the theoretical model for resonance production . in particular , the six models we use here are alcor @xcite , fritiof @xcite , kaneta _ et al . _
@xcite , letessier _ et al . _
@xcite , stachel _ et al .
_ @xcite and urqmd @xcite .
( see appendix [ sec : models ] for more details on the individual models . )
we have also considered the ampt 2.11 @xmath90 simulation with string melting @xcite that report on a non - thermal scenario without an @xmath7mass modification .
ampt results show an interesting , although genuinely different dropping structure of @xmath87 at low @xmath57values .
comparison of this model to the data , as tersely overviewed in @xcite , indicate that ampt can not describe these data in a statistically acceptable manner , characterized by a @xmath91 corresponding to @xmath92 .
we attribute the behavior seen in ampt to a lower effective @xmath93 of the high mass halo resonances @xcite . in earlier simulations in @xcite ,
the resonance production was generated by an exponential spectrum @xmath94 , the effective freeze - out temperature defined as @xmath95 , with @xmath96and @xmath97being the freeze - out temperature and the average transverse flow , respectively .
this has also been generalized , and a polynomial prefactor had been introduced in order to achieve a more realistic description of the direct production of resonances .
we have fixed @xmath96and @xmath97to rhic measurements @xcite .
thus the @xmath57distribution will follow the form of @xmath98 where @xmath99 is a normalization constant , and @xmath100 , where @xmath101 is the number of spatial dimensions of the expansion ( hence @xmath102 and @xmath103 falls between @xmath104 and @xmath105 ) @xcite .
the choice of @xmath106 corresponds to the case of ref .
@xcite . in order to compare to rhic data at mid - rapidity , we compute the effective intercept parameter of the @xmath55@xmath55 correlations , @xmath72 , with the definition of eq .
( [ eq : lams ] ) .
all the contributions of eq .
( [ eq : mtcmp ] ) to the @xmath80 distribution were simulated in the mid - rapidity region , given by the pseudorapidity$ ] .
in other manuscripts @xmath107 is usually denoted by the variable @xmath16 which in the current work is reserved for the pseudoscalar mesons @xmath16 and @xmath1 . ] cut of @xmath108 of the phenix acceptance . in the systematic checks we took into account that the star analysis @xcite used a different pseudorapidity cut of @xmath109 . for each model
the fractions of the particles were computed by normalizing the multiplicities to the total density @xmath110 the halo of pion production always had the same ingredients ( from the decays of the @xmath111 , @xmath1 , @xmath16 and @xmath112 ) meson , are checked to give a negligible contribution to the @xmath78ratio , which translates to less than 2% uncertainty to the @xmath88 . ] , while the core was composed of all the other resonances that were available in each particular model .
the charged pion @xmath57spectra were obtained from a complete kinematic simulation of the decays above resonances using jetset v7.4 @xcite . the estimated systematic error arising from assigning each @xmath111 decay product to the halo
is given in sec .
[ sec : results ] .
the mechanism of the partial @xmath8 restoration implies that the @xmath1 would have a decreased effective mass in the hot and dense medium @xcite .
the number of the created @xmath1 particles would then follow eq .
( [ eq : mtdist ] ) with the modified mass and the freeze - out temperature respective to the @xmath1 mesons , and the fraction of @xmath1 mesons in the condensate is modeled with the @xmath113 formula , where the @xmath114 is the temperature of the condensate .
as the escaping @xmath1 bosons are regaining their mass , they must lose momentum in order to fulfill the principle of energy conservation , i.e. , @xmath115 ( in the above equation the quantities with an asterisk denote the properties of the in - medium @xmath7 , while the ones without an asterisk refer to the free @xmath7 . according to the kinematical setup of both phenix and star measurements @xcite , the longitudinal component of the @xmath7momentum is considered to be negligible here . ) as a consequence , while @xmath7bosons with @xmath116 will follow the above distribution ( with the effective mass @xmath117 plugged in ) , the ones moving with a momentum less than this limit will be `` trapped '' in the medium until this medium is dissolved .
afterwards , @xmath7-s from the condensate are given a random transverse momentum , following maxwell - boltzmann statistics with a characteristic temperature @xmath89 : @xmath118 with @xmath119 .
note that @xmath120 is a systematically varied model parameter , and its best value is determined by the analysis of the @xmath72 data .
out of the six parameters of the @xmath7spectrum and @xmath72 simulation , the two most important ones are @xmath89and @xmath88 .
they directly determine the shape of the observed dip ( depth and width , respectively ) .
these parameters were considered as `` fit variables , '' and a fine grid @xmath121 scan was carried out .
the most probable values of the @xmath88 and @xmath89 were looked for by the simulation of the @xmath72 spectrum and then determining the @xmath121 from the fits to the data of the phenix and star measurements at each individual setup .
the other four parameters were treated as `` constants , '' although their variation played a role when determining systematic errors . as a default setup , the @xmath55freeze - out temperature was taken from phenix inverse slope parameter fitand @xmath97 only include the statistical errors .
the systematics were taken into account in our systematic studies that involved a wide range of different predictions on @xmath96 , @xmath122 . ] to be @xmath123 mev @xcite , and the conservative assumption of @xmath124 was made.downward , i.e. , toward `` safety '' for our conclusion . ]
the average transverse flow velocity was measured by phenix to be @xmath125 ( in relativistic units ) @xcite .
this was cross - checked in our simulations with a @xmath126 scan with @xmath127 steps at the best value of @xmath88and @xmath89 , and it was found that the most probable values were around @xmath128 , in agreement with the phenix measurement .
the polynomial exponent @xmath106 corresponds to a 2d expansion @xcite , that had been proved to give a good description of data , used by phenix when obtaining @xmath129 and @xmath96 .
however the @xmath130 ( 1d ) and @xmath131 ( 3d ) cases were also examined as systematic checks .
the ( @xmath89 , @xmath57 ) plane was mapped on a grid of 21 non - equidistant steps was mapped in 21 equidistant steps between 0 and 600 @xmath132 . ] for @xmath133 , and by 10 mev equidistant stepsvalues are given with the resolution of @xmath134 throughout the manuscript . ] for the @xmath135 region .
a good simultaneous description of phenix and star data was obtained by certain values of @xmath89and @xmath57 in the case of five out of the six models , as demonstrated in figures [ fig : alcor][fig : urqmd ] . for each model
, we show the confidence level ( @xmath136 ) map , the best result , the result without the mass modification , and the effect of @xmath88and @xmath89scans around the best result .
the standard deviation contours on ( @xmath89 , @xmath137 ) plain for the best fits of @xmath78 simulations to data are summarized for different models in figs .
[ fig : map_cl ] , and [ fig : map_cla ] for @xmath106 and @xmath131 respectively .
spectrum of the phenix and star data points , in comparison with monte carlo simulations based on the alcor model @xcite at parameters from the best fit ( @xmath138 mev , @xmath139 mev ) , for @xmath106 , @xmath140 mev , @xmath141 mev and @xmath142 .
_ top right : _ corresponding confidence level surface on the ( @xmath89 , @xmath137 ) plain
. only the region with acceptable fits ( @xmath143 ) is shown .
_ bottom left : _
@xmath78 spectra for different @xmath144 values , @xmath145 fixed to its best fit value . _ bottom right : _
@xmath78 spectra for different @xmath145 values , @xmath144 fixed to its best fit value . in the latter two plots
the phenix preliminary and the star sinyukov measurements are also shown for comparison ( as explained in sec .
[ sec : datasets ] ) .
, title="fig : " ] spectrum of the phenix and star data points , in comparison with monte carlo simulations based on the alcor model @xcite at parameters from the best fit ( @xmath138 mev , @xmath139 mev ) , for @xmath106 , @xmath140 mev , @xmath141 mev and @xmath142 .
_ top right : _ corresponding confidence level surface on the ( @xmath89 , @xmath137 ) plain .
only the region with acceptable fits ( @xmath143 ) is shown .
_ bottom left : _
@xmath78 spectra for different @xmath144 values , @xmath145 fixed to its best fit value . _ bottom right : _
@xmath78 spectra for different @xmath145 values , @xmath144 fixed to its best fit value . in the latter two plots
the phenix preliminary and the star sinyukov measurements are also shown for comparison ( as explained in sec .
[ sec : datasets ] ) .
, title="fig : " ] spectrum of the phenix and star data points , in comparison with monte carlo simulations based on the alcor model @xcite at parameters from the best fit ( @xmath138 mev , @xmath139 mev ) , for @xmath106 , @xmath140 mev , @xmath141 mev and @xmath142 .
_ top right : _ corresponding confidence level surface on the ( @xmath89 , @xmath137 ) plain .
only the region with acceptable fits ( @xmath143 ) is shown .
_ bottom left : _
@xmath78 spectra for different @xmath144 values , @xmath145 fixed to its best fit value . _ bottom right : _
@xmath78 spectra for different @xmath145 values , @xmath144 fixed to its best fit value . in the latter two plots
the phenix preliminary and the star sinyukov measurements are also shown for comparison ( as explained in sec .
[ sec : datasets ] ) .
, title="fig : " ] spectrum of the phenix and star data points , in comparison with monte carlo simulations based on the alcor model @xcite at parameters from the best fit ( @xmath138 mev , @xmath139 mev ) , for @xmath106 , @xmath140 mev , @xmath141 mev and @xmath142 .
_ top right : _ corresponding confidence level surface on the ( @xmath89 , @xmath137 ) plain .
only the region with acceptable fits ( @xmath143 ) is shown .
_ bottom left : _
@xmath78 spectra for different @xmath144 values , @xmath145 fixed to its best fit value . _ bottom right : _
@xmath78 spectra for different @xmath145 values , @xmath144 fixed to its best fit value . in the latter two plots
the phenix preliminary and the star sinyukov measurements are also shown for comparison ( as explained in sec .
[ sec : datasets ] ) .
, title="fig : " ] .
note that no values of @xmath145 and @xmath144 provided an acceptable fit , resulting in an empty confidence level plot .
this model was therefore excluded from further studies .
, title="fig : " ] .
note that no values of @xmath145 and @xmath144 provided an acceptable fit , resulting in an empty confidence level plot .
this model was therefore excluded from further studies .
, title="fig : " ] .
note that no values of @xmath145 and @xmath144 provided an acceptable fit , resulting in an empty confidence level plot .
this model was therefore excluded from further studies .
, title="fig : " ] .
note that no values of @xmath145 and @xmath144 provided an acceptable fit , resulting in an empty confidence level plot .
this model was therefore excluded from further studies .
, title="fig : " ] .
best fit is at @xmath146 mev , @xmath147 mev .
, title="fig : " ] .
best fit is at @xmath146 mev , @xmath147 mev .
, title="fig : " ] .
best fit is at @xmath146 mev , @xmath147 mev .
, title="fig : " ] .
best fit is at @xmath146 mev , @xmath147 mev .
, title="fig : " ] .
best fit is at @xmath148 mev , @xmath149 mev .
, title="fig : " ] .
best fit is at @xmath148 mev , @xmath149 mev .
, title="fig : " ] .
best fit is at @xmath148 mev , @xmath149 mev .
, title="fig : " ] .
best fit is at @xmath148 mev , @xmath149 mev .
, title="fig : " ] .
best fit is at @xmath148 mev , @xmath149 mev .
, title="fig : " ] .
best fit is at @xmath148 mev , @xmath149 mev .
, title="fig : " ] .
best fit is at @xmath148 mev , @xmath149 mev .
, title="fig : " ] .
best fit is at @xmath148 mev , @xmath149 mev .
, title="fig : " ] .
best fit is at @xmath148 mev , @xmath150 mev .
, title="fig : " ] .
best fit is at @xmath148 mev , @xmath150 mev .
, title="fig : " ] .
best fit is at @xmath148 mev , @xmath150 mev .
, title="fig : " ] .
best fit is at @xmath148 mev , @xmath150 mev .
, title="fig : " ] , @xmath137 ) plain , obtained from @xmath78 of mc simulations for @xmath106 , @xmath151 mev and @xmath142 , based on different chemical freeze - out models , each fitted to the phenix and star combined data set .
the framed band indicates the theoretically predicted range of @xmath152@xmath153 @xcite , while the horizontal dashed - dotted line at @xmath154 indicates weinberg s lower limit @xcite . ] as expected , a strongly reduced @xmath1 mass results in a dip in the low-@xmath57 part of the @xmath72 distribution ( and the smaller the mass , the deeper the dip is ) , while @xmath145 steers the width of this dip through smearing the momentum , while the @xmath155 parameter influences the overall slope of the @xmath72 spectrum at higher @xmath57 .
such a role of @xmath155 was already observed in @xcite , and it is not detailed here .
the systematic uncertainties resulting from the lack of precise knowledge of model constants were mapped by varying these constants within the range that is given by different measurements : spectra with values of @xmath131 and @xmath130 were investigated as well as @xmath156 mev , @xmath157 mev and @xmath158 mev , representing the boundaries of the theoretically or experimentally acceptable region of these parameters . *
the star edgeworth @xmath44values were extracted from data of 05% centrality , compared to the 030% rapidity range of the phenix sinyukov data set .
we estimate that this relative error does not exceed 9.8% .
see details in appendix [ sec : centrality ] . *
although the @xmath111 may give part of its contribution to the tail at the resolution available in star and phenix , we considered it as part of the halo .
this introduces a relative systematic error of 7% on the measured @xmath44 @xcite .
* moving the pseudorapidity cutoff from @xmath108 to @xmath159 adds another error , that is measured to be 3% .
these errors are considered independent and are added up quadratically . the relative error on @xmath78corresponds to the same amount of relative error on @xmath144 in the worst - case scenario.datapoint
was moved by 10% first up and then down , the fitted @xmath88value changed by the same relative error . ]
the best fit values and their error determinations are summarized in table [ tab : massfits ] , while table [ tab : massbounds ] shows the limits of acceptability , defined on the region , where @xmath143 .
the combined phenix and star data can not be described unless a significant @xmath7mass modification is assumed , @xmath160 .
although this value includes the systematics from the model uncertainties , it is also subject to the uncertainties coming from the other sources listed above . on the basis of model simulations , studies and systematic checks , we find that a description of the combined phenix and star data set is possible with @xmath143 only if an in - medium @xmath7mass modification of @xmath6 is utilized .
in other words , @xmath161 mev at the 99.9% confidence level , corresponding to a more than 5-@xmath162 effect , in the considered broad model class .
.[tab : massfits]fitted values of the modified @xmath7mass on the star+phenix combined data set , for different resonance models and parameters .
the fritiof model has cl@xmath163 and therefore it is not shown here .
the statistical errors are given by the 1-@xmath162 boundaries of the fits , determined only for @xmath88 and for the @xmath106 and @xmath164 simulations .
best @xmath88and @xmath89parameters for various systematic checks are shown in the last four rows . [ cols="<,^,^,^,^,^,^,^,^ " , ] in this section we provide a brief description both general and analysis specific of the six different models used to simulate the number of each important resonance that decays into pions , as well as the number of primordial pions .
the fractions of pions from different sources are compared to the total number of particles in table [ tab : resfrac ] .
these values are fed into our simulations as an input .
note that not all core resonances are listed . with the assumption of a fixed @xmath97 , the @xmath72
is solely determined by the number of the halo resonances ( @xmath111 , @xmath16 , @xmath1 , @xmath112 ) , _ and _ by the total number of pions .
the primordial pions can replace those pions that come from fast decays without any effect on @xmath72 .
generally , if a model contains a large number of exotic short lived resonances , it predicts less primordial pions .
* alcor * @xcite : hadron multiplicities especially for strange particles are calculated in the framework of this algebraic coalescence rehadronization model , which counts for redistribution of quarks into hadrons for relativistic heavy - ion collisions .
resonance ratios were taken from recent calculations for the rhic collisions @xcite . since the mass eigenstates of the strange - antistrange mesons are not directly pinned down , the kaneta - xu @xcite predictions were used to fix the @xmath165 ratio for the alcor model .
* fritiof * @xcite is a monte carlo program that implements the lund string model for hadron - hadron , hadron - nucleus and nucleus - nucleus collisions .
resonance ratios were computed with the fritiof model by simulating of 1000 events using rhic central au+au @xmath166 gev parameters .
fritiof was excluded from examination when drawing the conclusions of this analysis , since it was unable to describe the star @xmath78data set , nor the combined star + phenix data set , to a @xmath143 at any setup of the model parameters .
* kaneta - xu * @xcite : in the case of this model , eq . ( 1 ) . of @xcite was used the following way : @xmath167 where @xmath168 is the mass , @xmath169 is the spin of the particle .
@xmath170 denotes the second order modified bessel function . for the sake of simplicity the temperature of the chemical equilibrium @xmath171 , the strangeness saturation factor @xmath172 , the quark potentials @xmath173 and @xmath174 were considered , all in consistency with the published phenix and star measurements in @xmath5 au+au collisions @xcite .
( the relative minuteness of @xmath175 guarantees that the charge factor is negligible for even those rare particles , where it is not exactly 0 ) .
* letessier - rafelski * @xcite : the model of these two authors studied soft hadron production in relativistic heavy ion collisions in a wide range of reaction energy , @xmath176 , and made predictions about yields of particles using the statistical hadronization model .
particle yields of the central events were taken from table 4 of @xcite . *
stachel _ et al .
_ * @xcite : this statistical `` fireball '' model treats the system as a grand canonical ensemble with the temperature and the baryon chemical potential as free parameters .
particle yields of the central events were taken from sec . 2.2 .
table 1 of @xcite .
* urqmd * @xcite : the ultra - relativistic quantum molecular dynamics model is a microscopic model used to simulate ( ultra)relativistic heavy ion collisons in the energy range from bevalac and heavy ion synchrotron ( sis ) up to alternating gradient synchrotron ( ags ) , sps and rhic .
resonance ratios were computed with the urqmd model by simulating 1000 events using rhic central au+au @xmath166 gev parameters .
http://dx.doi.org/10.1103/physrevd.14.3432[g .
t hooft , phys .
d * 14 * , 3432 ( 1976 ) ] [ erratum - ibid .
d * 18 * , 2199 ( 1978 ) ] ; http://dx.doi.org/10.1016/0370-1573(86)90117-1[phys .
rept . * 142 * , 357 ( 1986 ) . ] m. csand [ phenix collaboration ] , http://dx.doi.org/10.1016/j.nuclphysa.2006.06.097[nucl .
a * 774 * , 611 ( 2006 ) . ]
+ for details , see also the corresponding talk : + http://www.kfki.hu/events/hun/qm2005/parallel/csanad/1500_csanad_qm05.ppt .
z. w. lin , c. m. ko , b. a. li , b. zhang and s. pal , http://dx.doi.org/10.1103/physrevc.72.064901[phys .
c * 72 * , 064901 ( 2005 ) . ] s. s. adler _ et al . _
[ phenix collaboration ] , http://dx.doi.org/10.1103/physrevc.69.034909[phys .
c * 69 * , 034909 ( 2004 ) . ]
a. adare _ et al . _ , http://dx.doi.org/10.1103/physrevc.81.034911[phys
c * 81 * , 034911 ( 2010 ) . ] s. weinberg , http://dx.doi.org/10.1103/physrevd.11.3583[phys .
d * 11 * , 3583 ( 1975 ) . ] s. s. adler _ et al .
_ [ phenix collaboration ] , http://dx.doi.org/10.1103/physrevc.75.024909[phys .
c * 75 * , 024909 ( 2007 ) . ]
t. csrg , http://dx.doi.org/10.1556/aph.15.2002.1-2.1[heavy ion phys .
* 15 * , 1 ( 2002 ) . ]
r. s. hayano and t. hatsuda , http://rmp.aps.org/abstract/rmp/v82/i4/p2949_1[rev .
phys . * 82 * , 2949 ( 2010 ) . ]
lvai , http://dx.doi.org/10.1088/0954-3899/35/4/044041[j .
g * 35 * , 044041 ( 2008 ) . ] k. adcox _ et al .
_ [ phenix collaboration ] , http://dx.doi.org/10.1016/j.nuclphysa.2005.03.086[nucl .
phys . a * 757 * , 184 ( 2005 ) . ] b. i. abelev _ et al . _ [ star collaboration ] , http://dx.doi.org/10.1103/physrevc.80.024905[phys . rev .
c * 80 * , 024905 ( 2009 ) . ] h. beker _ et al .
_ [ na44 collaboration ] , http://dx.doi.org/10.1103/physrevlett.74.3340[phys . rev .
lett . * 74 * , 3340 ( 1995 ) . ] | in high energy heavy ion collisions a hot and dense medium is formed , where the @xmath0 or chiral symmetry may temporarily be restored . as a consequence , the mass of the @xmath1(958 ) mesons may be reduced to its quark model value , and the abundance of @xmath1 mesons at low @xmath2 may be enhanced by more than a factor of 10 .
the intercept parameter @xmath3 of the charged pion bose
einstein correlations provides a sensitive observable of the possibly enhanced @xmath1 abundance .
we have analyzed @xmath4 data from @xmath5 central au+au reactions measured at the bnl relativistic heavy ion collider ( rhic ) , using extensive monte carlo simulations based on six popular models for hadronic multiplicities .
based on the combined star and phenix data set , and on various systematic investigations of resonance multiplicities and model parameters , we conclude that in @xmath5 central au+au reactions the mass of the @xmath1 meson is reduced by @xmath6 , at the 99.9% confidence level in the considered model class .
such a significant @xmath7mass modification may indicate the restoration of the @xmath8symmetry in a hot and dense hadronic matter and the return of the ninth `` prodigal '' goldstone boson . a similar analysis of na44 s+pb data at top cern super proton synchrotron ( sps ) energies showed no significant in - medium @xmath1 mass modification . |
variability in non - magnetic , cataclysmic variable ( cv ) close binaries has been largely attributed to instabilities .
the outbursts in cv recurrent novae ( rn ) and cv classical novae ( cn ) can be caused by thermonuclear runaway on the surface of a white dwarf or by a thermal and viscous instability in the disc .
the latter type of outburst involves cycling from low - to - high thermal and viscous states ( see e.g. , meyer & meyer - hofmeister 1981 and references within ) , and thus cycling from low - to - high mass transfer rates ( see e.g. , lasota 2001 ) , that results in a release of accretion energy as mass is drained onto the white dwarf .
transient x - ray binaries with neutron stars or black hole primaries and cv dwarf novae ( dn ) are also subject to thermal and viscous instabilities .
however , the thermal and viscous instabilities in cv rn are larger in scale than those in cv dn .
in addition to thermal and viscous instabilities , cv dn can experience a tidal instability .
results are cv dn su uma and nova - like ( nl ) outbursts .
the larger amplitude and longer duration superoutbursts involve both a thermal - viscous instability and a tidal instability .
both cv dn su uma and nl outbursts and superoutbursts show characteristic hump - shaped modulations , known as positive superhumps , in their light curves .
these positive superhump modulations have a period that is a few percent longer than the orbital period .
they could be caused by dissipation from spiral shocks that formed from converging gas flows in the middle and outer annuli of the disc ( see e.g. , smith et al .
the source to this spiral shock generation is the @xmath2 resonance ( lubow 1991a , b ) , using the terminology established by frank , king , & raine ( 1992 ) .
a competing theory has also been suggested by bisikalo et al .
( 2004 ) - the superoutburst is associated with the generation of an spiral density wave in inner annuli of the disc , a zone that is denser and gas - dynamically un - perturbed
. this inner spiral density wave contributes to a substantial increase in mass transfer rate in that region .
in addition to positive superhumps , some non - magnetic cv dn systems show negative superhumps in their light curves .
these negative superhump modulations have a period that is a few percent shorter than the orbital period .
some systems light curves only show negative superhumps ( e.g. , cn ori ) , or only show positive superhumps ( e.g. , su uma ) , or show both modulations simultaneously [ e.g. , sdss j210014.12 + 00446.0 ( olech et al . 2009 ) ] , or show one type of modulation only later to be replaced by the other [ e.g. , tt ari ( kim et al . 2009 ) ] .
negative superhumps are suggested to be a consequence of a partially tilted disc ( patterson et al .
1993 ) or are related to a warped accretion disc ( petterson 1977 ; murray & armitage 1998 ; terquem & papaloizou 2000 ; murray et al .
2002 ; foulkes , haswell , & murray 2006 ) . our previous numerical simulation work
( wood , montgomery , & simpson 2000 ; montgomery 2004 , montgomery 2009b ) shows that negative superhumps can be produced by a fully tilted disc , and these modulations have a shape and a period that are similar to those in observational light curves .
accretion discs can warp or tilt via a potpourri of suggested sources .
for example , disc tilt in x - ray binaries can be from gas streaming at an upward angle from the inner lagrange point for one half of the orbit and at a downward angle for the second half of the orbit ( boynton et al .
1980 ) . in some cvs
, a disc tilt can be held constant by gas stream being fed by the magnetic field of the secondary ( barrett , odonoghue , & warner , 1988 ) . also for cv systems
, a disc tilt instability can result from a coupling of an eccentric instability to lindblad resonances ( lubow 1992 ) .
a vertical resonant oscillation of the disc mid - plane can be caused by tidal interactions between a massive secondary and a coplanar primary ( lubow & pringle 1993 ) .
a warping instability can be caused by irradiation from the primary ( pringle 1996 , 1997 ) .
a warping can be caused by direct tidal forces from a secondary that is on an inclined orbit .
a disc warp can be caused by misalignments of the spin axis of a compact and/or magnetized primary and the disc axis .
of these possibilities , murray & armitage ( 1998 ) find that cv dn accretion discs do not seem to tilt out of the orbital plane by instabilities . if a warp is present , the warp may or may not propagate . as reviewed in pringle ( 1999 )
, an inviscid , locally isothermal or locally polytropic fluid disc that has keplerian rotation , that is within a spherically symmetric gravitational potential , and that is wholly tilted by the same finite angle has azimuthal wavenumber @xmath3=1[and thus the fluid disc has ( @xmath3=1)-fold symmetry in the azimuthal direction ] and radial mode number @xmath4=0 ( and thus the fluid disc has @xmath4=0 local minima ) . for this disc ,
the wave mode frequency is zero .
if , however , not all of the inviscid disc is tilted by the same amount , then a wave can radially propagate ( see papaloizou & lin , 1995 and e.g. , lubow , ogilvie , & pringle , 2002 ) .
if the disc has some viscosity , then a disc wave can only propagate if the viscosity is low and the tilt angle is approximately less than @xmath5 where @xmath6 is vertical disc thickness and @xmath7 is disc radius .
if the disc has low viscosity , then the wave is transferred through the disc by wave action . likewise ,
if the disc has high viscosity , then the wave is transferred through the disc by dispersion ( papaloizou & pringle , 1983 ) .
if a quiescent , viscous disc is wholly tilted by the same amount and the tilt angle is greater than @xmath5 , a wave should not propagate .
regardless of how an accretion disc tilts or warps and whether the wave propagates , non - coplanar accretion discs in binary systems precess in the retrograde direction ( i.e. , in the direction opposite to spin and orbital motion ) .
retrograde precessional activity has been suggested for x - ray binaries ( e.g. , katz 1973 , larwood 1998 , wijers & pringle 1999 ) , for common systems that produce jets ( e.g. , livio 1999 ) , for protostars ( e.g. , papaloizou & terquem 1995 , larwood et . al . 1996 , larwood 1997 ) , for cv systems
( bisikalo et al . 2003 , 2004 ) and for quasars and supermassive black holes ( e.g. , romero et al . 2000 , caproni & abraham 2002 ) , to name a few . in many of these binary systems ,
the source to the retrograde precession is often suggested to be due to direct tidal forcing . in montgomery
( 2009a ) , we show that tidal torques by the secondary on a tilted , spinning accretion disc is a source to retrograde precession .
a more familiar example is the tidal torque by the sun and moon on the oblate , spinning , tilted earth that causes the earth s first point of aries to precess retrogradely . of note
is that some find retrograde precession occurring within a disc even though the entire accretion disc remains coplanar with the orbit : in numerical simulations by bisikalo et al .
( 2003 , 2004 ) , a coplanar inner spiral density wave is generated within the inner annuli of a coplanar accretion disc and only the coplanar spiral density wave retrogradely precesses .
based on their numerical simulations , the wave is generated by sheared elliptical orbits and these sheared elliptical orbits ( and thus the wave ) retrogradely precess .
bisikalo et al .
( 2004 ) suggest that the retrograde precession is due to an influence by the mass - losing star by the processes discussed in kumar ( 1986 ) and warner ( 1995 , 2003 ) .
however , kumar ( 1986 ) finds the retrograde precession on a tilted disc s @xmath8 annulus due to the tidal influence of the secondary .
since the spiral density wave is found in inner annuli of the accretion disc , theory is needed on retrograde precession for a wave in this location . in montgomery ( 2009a ) , we suggest that care must be taken in choosing the correct state of matter and the correct geometrical shape of the accretion disc as different shapes can produce different retrograde precessional values .
for example , montgomery ( 2009a ) shows that retrograde precession of a non - magnetic , tilted cv dn gaseous accretion disc can best be described by a retrogradely precessing and differentially rotating tilted accretion disc , and superimposed on the disc is a rotating and precessing ring that is located near the innermost annuli of the disc .
the superimposed ring is likely caused by gas stream overflow and additional particle migration into the innermost annuli of the tilted disc ( montgomery 2009a , b ) . instead of a precessing inner ring in accretion discs , some accretion disc simulations show an inner spiral density wave in the same location : for semidetached binaries , bisikalo et al .
( 2003 , 2004 ) show a one - armed spiral shock wave is generated in inner annuli of cooler discs ( i.e. , @xmath9 ) of binaries with high secondary - to - primary mass ratios .
therefore , a study is needed on whether such a wave can be generated in simulations of lower mass ratio , and can such a wave explain retrograde precession in cv dn systems . in this work ,
we generate theory on retrograde precession of an accretion disc that contains an inner spiral density wave .
we assume the wave is within the plane of a quiescent disc , yet the disc is tilted five degrees relative to the orbital plane .
we wish to establish whether replacing the precessing inner ring found in montgomery ( 2009a ) with an inner spiral density wave has similar effects on disc retrograde precession .
to justify the disc geometry ( i.e .. disc with a coplanar inner spiral density wave ) , we generate a low mass ratio numerical simulation using the grid code of bisikalo et al .
( 2003 ) . in 2 of this paper
, we develop the theoretical expressions to describe a retrogradely precessing , tilted cv dn accretion disc that has an inner spiral density wave in the plane of the disc . in 3 , we describe the two numerical codes that generate retrograde precession in cv dn accretion discs and we present numerical simulation results . in 4 , we compare our theoretical and numerical results with results obtained from observations and with previously obtained results . in 5 , we discuss these comparisons , and in 6 , we provide a summary , conclusions , and future work .
in the montgomery ( 2009b ) smoothed particle hydrodynamic ( sph ) numerical simulations of a differentially rotating , retrogradely precessing , tilted accretion disc , a ring forms near the innermost annuli of the disc due to particle migration into this inner annuli and due to gas stream overflow of the tilted disc edge . instead of a ring , numerical simulations by bisikalo et al .
2004 ) show gas stream build - up in the form of a one - armed spiral density wave .
therefore , we consider a hypothetical disc geometry for non - magnetic cv dn system that exhibits negative superhumps in their light curves .
the hypothetical disc geometry is a combination of the disc geometries in both numerical simulations : we assume the disc is wholly tilted five or more degrees degrees relative to the orbital plane . instead of a denser ring of material in the innermost annuli of the disc , we assume an inner spiral density wave that remains in the plane of the tilted disc . in the following subsections ,
we find the disc s net retrograde precessional value . the retrograde precessional angular frequency of a differentially rotating , circular , tilted disc
is found to be ( see e.g. , montgomery , 2009a ) @xmath10 in this equation , @xmath11 is the mass ratio of the secondary - to - primary masses , @xmath12 is the obliquity angle , @xmath13 is the radius of the disc , @xmath14 is the orbital separation , and @xmath15 is the orbital frequency .
this equation applies to a thin disc that lacks an inner hole or has a small inner hole ( e.g. , non - magnetic , cv dn systems ) .
this disc also does not have any surface features such as spiral density waves or dense rings ( or winds , jets , planets , holes , gaps , or any other feature that may change the angular momentum , and thus moment of inertia , of the system ) .
the shape of the disc is ideally circular although the equation is assumed applicable to non - circular discs as well ( see montgomery 2009a ) .
equation ( 1 ) is valid so long as the disc precesses as a unit and the distance between the stars @xmath14 is much greater than the radius of the primary ( see montgomery 2009a ) . the outer radius of the approximately circular disc is assumed to be @xmath16 as given by paczynski ( 1977 ) .
this equation is valid for mass ratios in the range @xmath17 . as retrograde precession
is associated with tilted discs and as negative superhumps are generated when disc tilt is high enough ( see montgomery 2009b ) , then negative superhumps could be an observational beacon that indicates that the disc may be tilted .
as the negative superhump period @xmath18 and the orbital period @xmath19 are fairly precise clocks , we would like to relate the retrograde precessional period to these precise clocks . from observations ,
we know that the orbital period is slightly longer than the negative superhump period .
both periods are related to the net retrograde precession period by @xmath20 if we assume a primary mass @xmath21=0.8@xmath22 and mass ratio @xmath23 , then the secondary mass @xmath24=0.32@xmath22 , @xmath25 where @xmath26 ( montgomery 2009b ) . from newton
s version of kepler s third law , the orbital period and the rotation period at the disc edge is @xmath27 hours and one hour , respectively . from montgomery ( 2009b ) , the negative superhump period for this mass ratio is @xmath28 3.662 hours and therefore , from equation ( 3 ) , @xmath2919.5@xmath3075 hours , assuming small angle approximation for disc tilt .
notice that the spin rate at the edge of the disc is fast compared to the retrograde precessional rate .
however , this type of precession is not fast precession because the wobble rate is not comparable to the spin rate . instead
, this kind of precession is driven by torques ( see e.g. , montgomery 2009a ) . to find the effects an inner spiral density wave has on retrograde precession , we first review the theory generated for cv dn accretion discs that progradely precess , that have spiral density waves , and that generate positive superhumps . from the location of the spiral density waves in a progradely precessing disc
, we can find the location of the inner spiral density wave .
spiral density waves contribute to retrograde precession through pressure as lubow ( 1992 ) finds for progradely precessing cv dn accretion discs that show positive superhumps in their light curves .
lubow ( 1992 ) finds that the net prograde precession @xmath31 is the net of a dynamical component , a pressure component , and a transient component .
of the three , the dynamical component contributes most to prograde precession .
the second of the three is a retrograde component and is a correction term to the dynamical component .
this pressure component is necessary to more correctly explain the net prograde precession in steady state discs .
the third component is also a correction term and can contribute to the prograde dynamical component or to the retrograde pressure component .
however , this transient component is most effective during the development stage of positive superhumps . if only quasi - steady state discs are considered , then this third component can be ignored .
pressure effects within progradely precessing discs could be due to a two - armed spiral mode .
the source to this spiral shock generation is the @xmath2 resonance ( lubow 1991a , b ) , using the terminology established by frank , king , & raine ( 1992 ) .
dissipation from these spiral shocks are created from converging gas flows , and this dissipation is the source to the observed positive superhump as suggested by smith et al .
the location of these spiral density waves is in the middle and outer annuli of the quasi - steady state disc ( see e.g. , smith et al . , 2007 ) . assuming this disc geometry for prograde precession , an analytical expression for the pressure component is developed in lubow ( 1992 ) .
the standard wentzel - kramers - brillouin ( wkb ) relation for spiral density waves in the tight winding limit is given by @xmath32 where @xmath33 and @xmath34 are the particle angular and radial frequencies , respectively , @xmath35 is the oscillation frequency of the spiral density waves , @xmath36 is the gas sound speed within the disc , and @xmath37 is the radial wavenumber of the spiral arms ( see e.g. , goldreich & tremaine 1979 ) . in a non - precessing disc , @xmath38 , and in a slowly precessing disc , @xmath39 . in the limit where @xmath40 or @xmath34 and applying the previous approximation , lubow ( 1992 ) approximates @xmath35 in the previous equation . then lubow ( 1992 )
couples @xmath35 with the dynamical motion within the disc to find the net prograde precession : @xmath41 we identify the last term in the equation as the pressure term associated with these spiral density waves located in the middle - to - outer portion of the disc .
note that we assume quasi - static discs ; the net prograde precession has only a dynamical prograde component and a retrograde pressure component . in e.g. ,
montgomery ( 2001 ) , we show that the positive superhump period excess is @xmath42^{-1 } - 1 \\ & = & \frac{p_{orb}}{p_{p , net } - p_{orb } } \end{aligned}\ ] ] where we have substituted @xmath43 into equation ( 7 ) to obtain equation ( 8) . in these equations , @xmath44 is the positive superhump period and @xmath45 is the net prograde precessional period .
if we assume @xmath46 , then we obtain the positive superhump period excess @xmath47 the pitch angle @xmath48 of the spiral wave is related to the wavenumber by @xmath49 where @xmath50 is the radial distance to that fluid element within a spiral arm . in montgomery ( 2001 )
, we assume average values for @xmath36 and @xmath51 based on our numerical simulations .
we find the average distance to the two numerically generated spiral density arms and the average pitch angle to be @xmath52 and @xmath53 , respectively , where @xmath14 is the binary separation .
after substitution of @xmath50 , @xmath51 , @xmath37 , @xmath54 , and the smith & dhillon ( 1998 ) secondary mass - period relation into equations ( 5 ) and ( 6 ) , we find @xmath55^{-1 } \\ & \sim & \omega_{orb}\left[\frac{28.4(1+q)^{2/3}}{q^{0.422}}\right]^{-1 } \nonumber.\end{aligned}\ ] ] as reviewed in kato ( 2001 ) , local analytical studies of hydrodynamic accretion discs have found three types of trapped global modes .
one of the three types is inertial acoustic @xmath56-modes ( the other two are gravity @xmath57-modes and vertical@xmath58-modes and are not discussed in this work ) .
for a two - dimensional disc with no vertical modes , the @xmath56-mode is described by @xmath59 where @xmath60 and @xmath61 are the particle angular and radial frequencies , respectively , @xmath62 is the oscillation frequency of an inner spiral density wave , @xmath36 is the gas sound speed within the disc , @xmath63 is the radial wavenumber of the spiral arms , and @xmath3 is an integer that indicates the different azimuthal modes .
notice that this equation is very similar to equation ( 4 ) , however this equation is evaluated at the radius @xmath64 of inner spiral density wave and hence the wave number and particle angular and radial frequencies are different .
also note that we assume the same sound speed throughout the disc .
global spirals in discs are the @xmath65 @xmath56-modes .
for a one - armed spiral wave , we can reduce the previous equation to @xmath66 in a non - precessing disc , @xmath67 , and in a slowly precessing disc , @xmath68 .
if the disk is also geometrically thin and @xmath69 ( i.e. , we have a global oscillation ) , then the previous equation reduces to @xmath70 this result is similar to that found by kato ( 1983 ) .
as kato ( 1983 ) noted , this oscillation at r moves much slower than the angular velocity of the disc rotation .
therefore , the oscillation acts to slow the motion of the gas particles in the disc and hence the sign is negative in this equation . to find the net retrograde precession in the disc
, we need to sum the effects due to the inner spiral density wave with the effects due to the dynamical motions in the smooth , featureless disc : @xmath71 the last term in the equation is identified as the pressure term associated with the inner spiral density wave number @xmath63 and particle angular frequency @xmath60 .
notice that these equations are very similar to equation ( 5 ) and ( 6 ) , however these equations are evaluated at at the radius of the inner spiral density wave @xmath64 . in montgomery ( 2009b )
, we show that the negative superhump period excess is @xmath72^{-1 } \\ & = & \frac{p_{orb}}{p_{r , net } + p_{orb}}\end{aligned}\ ] ] where we have substituted equation ( 3 ) into equation ( 20 ) to obtain equation ( 21 ) . in these equations , @xmath18 is the negative superhump period and @xmath73 is the net retrograde precessional period .
if we assume @xmath46 , then we obtain the net negative superhump period excess @xmath74 as @xmath75 is negative , then @xmath76 and @xmath77 are also negative .
if we assume a secondary mass - period relationship by smith & dhillon ( 1998 ) , then we can find the net negative superhump period excess in terms of orbital period ( not shown ) .
the pitch angle of the inner spiral density wave @xmath78 is related to wavenumber @xmath63 by @xmath79 upon substitution into equations ( 18 ) and ( 19 ) , @xmath80 where in the second version of @xmath62 we have assumed [email protected] for the average outer radius of the inner spiral density wave .
this location is the average outer radius of the inner ring seen in the @xmath23 simulation of montgomery ( 2009a ) and is larger than @xmath820.13d which is the minimum radius of the ring s outer radius .
if @xmath14 is obtained from kepler s third law , @xmath8320 kms@xmath84 and @xmath85 ( montgomery 2001 ) , then only the particle orbital frequency @xmath60 is unknown .
however , we can easily find the particle orbital frequency through kepler s laws : if we solve newton s version of kepler s third law for the orbital period and ratio it with newton s version of kepler s third law for the period of a particle within the disc @xmath86 at any radius @xmath7 , ignoring perturbation effects by the secondary , then we find @xmath87 after substitution for @xmath64 and q , we find @xmath88 or @xmath89 .
we can use the smith & dhillon ( 1998 ) secondary mass - period relation to find @xmath90 where @xmath91 is in units of rad s@xmath84 . upon substitution into equations ( 26 ) or ( 27 ) , we find @xmath92^{-1 } \\ & = & \omega_{orb}\left[13.5\frac{(1+q)^{2/3}}{q^{0.422}}\right]^{-1},\end{aligned}\ ] ] and @xmath93 . as expected from keplerian motion , the inner spiral density wave angular frequency is significantly different from the outer spiral density wave angular frequency .
in this section , we describe two different numerical codes that generate accretion discs like those expected for non - magnetic cv dn . in 3.1 , we summarize a grid code that generates accretion discs with an inner spiral density wave as shown in e.g. , bisikalo et al .
( 2003 , 2004 ) . in this work ,
we present the @xmath23 numerical simulation using this code to see if lower mass ratio systems can also generate an inner spiral density wave . in 3.2 , we summarize the sph code that generates accretion discs with a denser inner ring if the disc is tilted at least a few degrees as shown in montgomery ( 2009a ) .
the details of the three - dimensional , high resolution gas dynamical grid code can be found in bisikalo et al .
( 2003 ) . here
we provide a summary .
for this work , we consider a semi - detached binary consisting of a donor secondary star with mass @xmath94 that fills its roche lobe .
an accretor primary star with mass @xmath95 is separated from the donor secondary by a distance @xmath96 .
the modeling is carried out in a non - inertial reference frame rotating with the binary and on a regular three dimensional grid in cartesian coordinates .
the origin of the coordinate frame coincides with the center of mass of the donor secondary .
since the problem is symmetrical about the equatorial plane , the modeling is carried out in half the space : a symmetrical boundary condition is imposed at the corresponding boundary of the computational domain .
steady - state boundary conditions are specified at the remaining boundaries of the domain : the density is @xmath97 , where @xmath98 = 1.5 x 10@xmath99 g @xmath100 @xmath101 is the matter density at the inner lagrange point @xmath102 ; the equilibrium temperature of the optically thin gas is @xmath103 ( see bisikalo et al .
2003 ) ; and the velocity is @xmath104 .
the white dwarf accretor primary is specified to be a sphere with radius @xmath105 , and a free - inflow condition is imposed at its surface : all the matter that falls in cells occupied by the white dwarf accretor is considered to have fallen onto the star . the force field in the system
is described by the roche potential : @xmath106 , \label{eqroche}\end{aligned}\ ] ] where @xmath107 is the gravitational constant and @xmath108 is the rotational speed of the system .
the gradient of the roche potential is equal to zero at five libration points ( the so - called lagrange points ) : @xmath109 . the equipotential surface that passes through the inner lagrange point @xmath102 forms the roche lobes of the system s components .
the size of the donor secondary , a red dwarf , is specified by the boundaries of its roche lobe .
the following boundary conditions have been selected at its surface : @xmath110 , @xmath111 k , @xmath112 .
the speed at the inner lagrange point @xmath102 is specified to be the local sonic speed : @xmath113=6.6 kms@xmath84 .
the region within the donor secondary is excluded from the calculations .
we describe the flow structure in the binary using a system of gravitational gas dynamic equations , taking into account the radiative heating and cooling of the gas in the optically thin case : @xmath114 \frac{\partial\rho\emph{v}}{\partial t } + \nabla \cdot ( \rho\emph{v}\otimes\emph{v } ) + \nabla p = -\rho \nabla \phi - 2 [ { \omega}\times\emph{v}]\rho\,,\\[3 mm ] \frac{\partial\rho(\varepsilon+v^2/2)}{\partial t } + \nabla \cdot ( \rho\emph{v})(\varepsilon + p/\rho + v^2/2 ) = \\
\indent \indent \indent -\rho\emph{v } \nabla \phi + \rho^{2}m_p^{-2 } \left[\gamma(t , t_{a})-\lambda(t)\right ] .
\end{array } \right .
\label{hdc}\ ] ] here , @xmath115 is density , @xmath116 is velocity , @xmath86 is pressure , @xmath117 is time , @xmath118 the inner energy , @xmath119 is the mass of the proton , and @xmath120 and @xmath121 are the radiative heating and cooling functions ( see bisikalo et al .
the system of the gasdynamic equations is closed with the equation of state for an ideal gas , @xmath122 , with an adiabatic index @xmath123 for a monatomic ideal gas .
we solve the system of equations using the roe - osher - einfieldt method adapted for multiprocessor computers ( roe 1986 ; chakravarthy & osher 1985 ; einfeldt 1988 ) .
this technique makes it possible to obtain high - accuracy solutions of gas dynamical problems .
the simulations are carried out until the system is in the steady - state accretion regime .
figure 1 shows that a spiral arm can be generated in the steady state solution for the lower mass ratio @xmath23 numerical simulation where @xmath95 .
the cartesian coordinates ( x , y , z ) are in fractions of the separation distance @xmath14 .
although in this work we only show half the space , the full three - dimensional simulations also have the same spiral wave as shown in the purely hydrodynamic solutions by bisikalo et al .
( 2008 ) and in the hydrodynamic solutions with magnetic fields by zhilkin et al .
therefore , the spiral density wave is not a consequence of symmetry and not due to magnetic fields . as shown in figure 1
, the spiral density wave does have a small projection relative the disc mid - plane .
figure 2 shows snapshots of the disc over three - quarters of an orbit . in their numerical simulations ,
bisikalo et al .
( e.g. , 2003 , 2004 ) find the wave to be a result of sheared elliptical particle orbits in the innermost annuli of the disc : as the particles travel in elliptical orbits around the primary , where the primary is at one common focus to the elliptical orbits , the apastron points of the elliptical orbits are gravitationally torqued by the secondary to precess in the retrograde direction .
the closer an apastron point is to the secondary , the larger its semi - major axis , and the more torque that apastron point experiences . because velocity flow is very slow at these apastron points and
because flow lines of the gas stream can not cross , gas particles accumulate near these apastron points .
we find in this work that these accumulations have two sources : gas particle migration from outer annuli and , as shown in figure 10 of bisikalo et al .
( 2005 ) , the gas stream that overflows the disc edge and travels radially across the disc face to inner annuli .
the result of an over population of gas particles near apastron points is an elliptical spiral density wave in the innermost annuli . because apastron points are not uniformly spaced in angle due to differential rotation of the disc , accumulation of gas particles is not uniform near these points .
the result is a variation in density along the wave .
the net of the two effects is an asymmetric , elliptical , tilted spiral density wave located in inner annuli as shown in figure 1 . because the center of mass is not in the center of the elliptical spiral and not in the orbital plane , the spiral density wave is subject to torques by the secondary and thus the spiral density wave retrogradely precesses .
we find for the q=0.4 numerical simulations in steady state that as the secondary orbits the center of mass , the secondary returns to its starting position ( relative to the spiral density wave ) at a period that is @xmath813% shorter than the orbital period .
this shorter period indicates that the spiral density wave is in retrograde precession . in @xmath81 33
orbits , the spiral density wave has made one full retrograde turn .
we note that the disc does not retrogradely precess . from figures 1 and 2
, we find that the inner spiral density wave outer radius ranges from @xmath124d .
however , the average outer radius is @xmath810.16d , confirming our assumption that the inner spiral density wave is located near the inner ring seen in the sph simulations montgomery ( 2009a ) . for the numerical simulations in montgomery ( 2009b ) ,
we use smoothed particle hydrodynamics ( sph ) which is a lagrangian method that models highly dynamical , astrophysical fluid flow as a set of interacting particles ( see monaghan 1992 for a review ) .
the code utilizes the simplest sph form - constant and uniform smoothing length @xmath125 and constant and uniform mass particles .
modifications to the original code include increasing particle number to 100,000 and adopting a smith & dhillon ( 1998 ) secondary - to - primary mass - period relation .
the code is discussed further in montgomery ( 2009b ) , however applicable information is also summarized here .
the original code models hydrodynamics assuming an ideal gamma - law equation of state @xmath126 where @xmath86 is pressure , @xmath127 is the adiabatic index , @xmath115 is density , and @xmath128 is specific internal energy .
the sound speed is @xmath129 .
the momentum and internal energy equations per unit mass are , respectively , @xmath130 and @xmath131 in their most general form ( simpson 1995 ) . in these equations
, @xmath132 is the viscous force , @xmath133 is the energy generation due to viscous dissipation , and @xmath134 and @xmath135 are the displacements from stellar masses @xmath21 and @xmath24 , respectively , that are given in units of solar masses .
the momentum and energy equations in sph form for particles @xmath48 and @xmath136 are , respectively , @xmath137 and @xmath138 or , if the last equation fails in preventing negative internal energies , @xmath139 in these equations , @xmath140 is the lattanzio et al .
( 1986 ) artificial viscosity @xmath141 where @xmath142 and @xmath143 is the average sound speed , @xmath144 , @xmath145 , and @xmath146 as shown in simpson & wood ( 1998 ) . for our numerical simulations ,
we choose @xmath147 .
our viscosity is approximately equivalent to a shakura & sunyaev ( 1973 ) viscosity parametrisation ( @xmath148 where @xmath6 is the disc scale height ) @xmath149=0.05 . as smak ( 1999 )
estimates @xmath150 0.1 - 0.2 for dn systems in high viscosity states , our simulations are more for quiescent systems . as only approaching particles feel the viscous force and since neither radiative transfer nor magnetic fields are included in this code , all energy dissipated by the artificial viscosity is transferred into changing the internal energies . as radiative cooling is not included in this code , an adiabatic index @xmath127=1.01 is incorporated to prevent the internal energies from becoming too large . by integrating the changes in the internal energies of all the particles over a specific time interval @xmath4 ,
variations in the bolometric luminosity yield an approximate and artificially generated light curve @xmath151 all simulations start by injecting five particles per major time step , 200 major time steps per orbit , or a total of 1000 particles per orbit .
the injection velocity is based on an injection temperature @xmath152 = 4x10@xmath153 k and is approximately the sound speed that has been scaled to system units .
the code adopts an approximate secondary mass - radius relation @xmath154 that applies for @xmath155 ( warner 2003 ) . in this relation ,
@xmath156 and @xmath24 are in solar radii and solar mass , respectively .
a unique feature of this code is the conservation of disc particle number .
any time a particle is accreted onto the secondary or primary mass or is lost from the system , a new particle is injected through the inner lagrange point @xmath157 .
we build a disc of 100,000 particles and maintain this number throughout the simulation . after building the disc
, we then allow the disc to evolve in the short term to a quasi - equilibrium state where only a few particles are injected per orbit . at orbit 200 ,
we artificially rotate the disc out of the orbital plane 5@xmath158 . for our simulations , we assume a primary mass @xmath21=0.8@xmath22 and we vary the secondary mass @xmath24 such that @xmath159 . for example , if @xmath23 then @xmath24=0.32 @xmath22 and our simulation unit length scales to @xmath160 where we have used the secondary mass - radius relationship shown above and the eggleton ( 1983 ) volume radius of the roche lobe secondary @xmath161 the eggleton ( 1983 ) relation is good for all mass ratios , accurate to better than 1% . taking the radius of the primary to be 6.9x10@xmath162 cm , the scaled radius of the white dwarf is @xmath163 . after building the disc
, we allow the disc to evolve in the short term to a quasi - equilibrium state where particles are injected at the rate they are removed from the system by either being accreted onto the primary or the secondary or lost from the system
. the average net rate of accretion for a quasi - static disc is around 500 particles per orbit as shown in montgomery ( 2009b ) .
as this net rate is approximately half that injected to build the disc , the steady state mass transfer rate reduces to @xmath164 1 x @xmath165/yr or a rate similar to an su uma in quiescence .
we estimate the density of the gas near @xmath157 to be @xmath166 g @xmath101 using @xmath167 . as we maintain 100,000 particles in the steady state disc , then the total mass of the disc is @xmath168= 100,000 x @xmath169 = 3.5x10@xmath170 g or @xmath171 2x10@xmath172 , a negligible value compared to the mass of either star . if we assume the outer radius of the disc is @xmath173 , or twice the circularization radius ( warner 2003 ) @xmath174 where @xmath175 , then for @xmath23 we find @xmath176 and @xmath177 . as @xmath178 , self gravity is neglected in these simulations . as output to the code , we create an artificial light curve for each simulation ( see montgomery , 2009b ) .
a non - linear least squares fit to each fourier transform is computed and the standard deviation to twice the negative superhump frequency @xmath179 is found .
this data in table 1 is from montgomery ( 2009b ) and , along with the calculated negative superhump period , the propagated error and the negative superhump period excess are also listed .
the units for period are hours whereas the units for frequency are orbits@xmath84 to be consistent with the fourier transform figure . to convert to real units , multiply one orbit by the orbital period .
[ cols= " < , < , < , < , < , < , < " , ] + @xmath0nl , @xmath1vy scl / nl , @xmath180ps / nl , @xmath153su uma / dn , @xmath181er uma / dn , @xmath182ip / nl , @xmath183z cam , @xmath162sw sex / nl , @xmath184nova , @xmath185classification uncertain + @xmath186taylor et al . ( 1998 ) , @xmath187huber et al . ( 1998 ) , @xmath188patterson et al . ( 2005 ) , @xmath189kraicheva et al . ( 1999 ) , @xmath190wu et al . ( 2002 ) , @xmath191warner ( 1995 ) , + @xmath192borczyk et al . ( 2003 ) , @xmath193patterson et al . ( 1997 ) , @xmath194arenas ( 2000 ) , @xmath195harvey et al . ( 1995 ) , @xmath196patterson et al . ( 1995 ) , @xmath197thorstensen et al . + ( 1997 ) , @xmath198skillman et al .
( 1999 ) , @xmath199nelemans et al .
( 2001 ) , @xmath200roelofs et al .
( 2006 ) @xmath201retter et al .
( 2003 ) , @xmath202hellier & buckley ( 1993 ) , + @xmath203schoembs ( 1982 ) , @xmath204barrera & vogt ( 1989 ) , @xmath205gao et al .
( 1999 ) , @xmath206patterson et al .
( 2001 ) , @xmath207rx j1643.7 + 3402 - patterson et al .
( 2002 ) , + @xmath208shafter & szkody ( 1983 ) , @xmath209retter et al .
( 1997 ) , @xmath210patterson ( 1998 ) , @xmath211rodriguez - gil et al .
( 2007b ) , @xmath212biro ( 2000 ) , @xmath213araujo - betancor + et al . ( 2003 ) , @xmath214boffin et al .
( 2003 ) , @xmath215dhillon et al .
( 1992 ) , @xmath216hoard & szkody ( 1997 ) , @xmath217 woudt & warner ( 2002 ) , @xmath218kozhevnikov + ( 2004 ) , @xmath219li et al .
( 2004 ) , @xmath220fu et al .
( 2004 ) , @xmath221retter et al .
( 2005 ) , @xmath222sdss j040714.78 - 064425.1 - ak et al .
( 2005 ) , @xmath223sdss j210014.12 + + 004446.0 - tramposch et al .
( 2005 ) , @xmath224woudt et al .
( 2005 ) , @xmath225kozhevnikov ( 2007 ) , @xmath226kato et al .
( 2002 ) , @xmath227rodriguez - gil et al . + ( 2007 ) , @xmath228kang et al .
( 2006 ) , @xmath229gulsecen et al .
( 2009 ) , @xmath230szkody & howell ( 1993 ) , @xmath231olech et al .
( 2007 ) , @xmath232ak et al . ( 2005b ) + the disc geometry in this study is hypothetical as it is a combination of the disc geometries found in numerical simulations of montgomery ( 2009b ) , bisikalo et al .
( 2003 , 2004 ) , and this work .
the numerical simulations of montgomery ( 2009b ) show a tilted disc with a coplanar denser inner ring .
the numerical simulations of bisikalo et al .
( 2003 , 2004 ) and this work for @xmath23 show a coplanar disc with a dense inner spiral density wave that projects at a slight angle to the disc mid - plane .
hence the hypothetical disc geometry for this study is a fully tilted disc around the line of nodes and the disc contains an inner spiral density wave that remains coplanar with the disc .
figure 3 shows a comparison of the numerical negative superhump period excess ( @xmath233 ) to the observational negative superhump period excess ( * ) values listed in tables 1 and 2 , respectively . in this figure
, the width of the symbol can be taken as the error in the orbital period and , to express mass ratio in terms of orbital period , we include the smith & dhillon ( 1998 ) secondary mass - period relation .
as the geometry of the accretion discs for each observational system is unknown , we do not differentiate between different cv types with different symbols even though ips probably have a significant absence of inner disc annuli , nls could have discs with spiral density waves that cause the positive superhump as well as an inner spiral density wave or ring that causes the negative superhump , su uma s could have only the spiral density waves that cause the positive superhumps , etc .
therefore , caution is advised when comparing the observational data with the numerical simulations as the scatter in the observational data is likely due to different geometries of a fluid disc ( see montgomery 2009a ) .
we also note that we have not taken into account other factors that may affect angular momentum and retrograde precession such as mass loss from winds , jets , mass loss to circumbinary rings , etc . in figure 3
, we plot as a dash - dot line a differentially rotating , retrogradely precessing disc [ i.e. , from equation ( 1 ) ] .
the dashed line represents a differentially rotating ring that can be found near the innermost annuli of the disc as discussed in montgomery ( 2009a ) .
the dotted line represents equation ( 31 ) , the retrograde precession due to an inner spiral density wave that is located near the ring s radius .
as shown , the dotted line compares well with the dashed line .
likewise , the net effect of a differentially rotating disc with a spiral density wave ( the short dash line ) compares well with the net effect of a differentially rotating disc with an attached , inner ring ( the solid line ) . in other words , the effects of pressure within the disc due to an innermost spiral density wave or the effects due to an attached innermost ring are necessary to better explain net retrograde precession in cv dn tilted , accretion discs should either be present in inner annuli . for the q=0.4 numerical simulation presented in this work
, we find that the spiral density wave takes @xmath8133 orbits or @xmath815.297d to retrogradely precess .
the negative superhump period is @xmath813% shorter than the orbital period or @xmath813.736 hours .
the negative superhump period excess is @xmath810.03 .
from the results , we can say that non - magnetic cv dn that show negative superhumps in their light curves can be described by a tilted , differentially rotating , precessing disc ( i.e. , the dynamical component ) that has either a retrogradely precessing spiral density wave located in the innermost annuli or a retrogradely precessing ring that is superimposed on the innermost annuli of the disc .
therefore , with the addition of either a retrogradely precessing ring or a retrogradely precessing spiral density wave , the geometry of the disc has changed as well as the net angular momentum and the net retrograde precession due to the change in the moment of inertia .
unknown is whether an inner spiral density wave or an inner denser ring is present .
however , both the ring and the spiral density wave have the same effect on the disc so long as both have nearly the same radial extent and are coplanar with the tilted disc .
if either is not coplanar with the disc , then differential precession would be present , effects not seen in non - magnetic cv dn systems .
we note in figure 3 that we assume a smith & dhillon ( 1998 ) secondary mass - period relation to obtain the negative superhump period excess as a function of orbital period . by assuming this relation , the secondaries are taken to be main - sequence stars .
those observations that agree with our numerical simulations and the solid and dashed lines may have tilted accretion discs with either an innermost precessing ring as discussed in montgomery ( 2009a ) or an innermost spiral density wave ( this work ) .
bisikalo et al .
( e.g. , 2003 , 2004 ) find that their numerically simulated inner spiral density wave retrogradely precesses yet the rest of the disc does not .
we confirm this result in the q=0.4 numerical simulation presented in this work . because this spiral density wave has its center of mass off - center and above the orbital plane of the disc , the spiral density wave is subject to torques by the secondary .
the net result is a retrogradely precessing spiral density wave .
the values for the negative superhump period , retrograde precession , and negative superhump period excess are long but not unreasonable when compared with the values in table 2 .
negative superhumps are found in observational systems with high mass ratios , long orbital periods , and high mass transfer rates . in montgomery ( 2009b ) , we find that the negative superhump can be due to extra light being emitted from inner disc annuli .
the extra light is modulated by gas stream overflow that strikes inner disc annuli and by gas particle migration into inner annuli as the secondary orbits . for this result to be true
, we find that the disc must be fully tilted around the line of nodes . in non - magnetic binary systems , discs that are tilted also retrogradely precess .
patterson et al .
( 1993 ) suggest that retrograde precession in cv dn accretion discs may be by the same source that causes the earth to retrogradely precess . in montgomery ( 2009a )
, we generate theoretical expressions for a disc that is misaligned with the orbital plane to test this suggestion , and we find that spinning , tilted , precessing accretion discs in non - magnetic binary systems can be described by a differentially rotating and precessing disc that has a precessing , denser ring superimposed on disc near the innermost annuli of the disc . in this work , we consider a differentially rotating and precessing accretion disc that has an inner spiral density wave instead of a ring but having the same radial extent as the ring and in the same plane as the disc .
that is , in addition to dynamical effects , we consider pressure effects in our analytical expressions .
we adopt the theory used to describe pressure effects within cv dn discs that progradely precess .
we compare our analytical expressions generated in this work with the analytical expressions generated in montgomery ( 2009a ) and the numerical simulations generated in montgomery ( 2009b ) for the mass ratio range @xmath234 . in montgomery ( 2009a )
, we find that non - magnetic , differentially rotating , tilted accretion discs can be described by a differentially rotating , retrogradely precessing accretion disc that has a superimposed retrogradely precessing ring in the innermost annuli of the disc .
we find in this work that these same non - magnetic , differentially rotating , tilted accretion discs can equally be described by a differentially rotating , retrogradely precessing accretion disc that has an innermost spiral density wave so long as the radial extent of the wave is like the radial extent of the ring .
therefore , we conclude that pressure effects in retrogradely precessing systems would need to be included if negative superhumps are present in the light curve and if the fluid disc has an inner spiral density wave . as part of this work
, we generate a @xmath23 accretion disc with an inner spiral density wave using a 3d , high resolution , gas dynamical grid code .
the inner spiral density wave is not generated due to magnetic fields .
it is a consequence of elliptical particle orbits .
the inner spiral density wave projects at a small angle relative to the orbital plane .
the tilt is due to differing densities along the spiral density wave . as a result of the center of mass being off - center and not in the orbital plane , the spiral density wave retrogradely precesses due to tidal torques , yet the disc does not .
the spiral density wave takes approximately 33 orbits to retrogradely precess , a value that is long but not unreasonable compared to the observations .
d. bisikalo was supported by the russian foundation for basic research and by the federal agency of sci . and innovations .
we would also like to thank the anonymous referee .
99 ak t. , retter a. , liu a. , & esenoglu h.h .
, 2005 , pasa , 22 , 105 ak t. , retter a. , liu a. , & esenoglu h.h . , 2005b , newa , 11 , 147 araugo - betancor a. , knigge c. , long k.s . , hoard d.w . , szkody p. , rodgers b. , krisciunas k. , dhillon v.s . , hynes r.i . , patterson j. , & kemp j. , 2003 , apj , 583 , 437 arenas j. , catalan m.s .
, augusteijn t. , & retter a. , 2000 , mnras , 311 , 135 barrett p. , odonoghue d. , & warner b. , 1988 , mnras , 233 , 759 barrera l.h . & vogt n. , 1989 , a&a , 220 , 99 biro i.b . , 2000 , a&a , 364 , 573 bisikalo d.v . ,
boyarchuk a.a .
, kaigorodov p.v .
, & kuznetsov o.a . , 2003 ,
astron . rep .
47 , 809 bisikalo d.v .
, boyarchuk a.a . ,
kaygorodov p.v . ,
kuznetsov o.a . ,
& matsuda t. , 2004 , arep , 48 , 449 bisikalo d.v .
, kaigorodov p.v . ,
boyarchuk a.a . , &
kuznetsov o.a . , 2005 , arep , 49 , 701 bisikalo d.v . , kononov d.a . ,
kaigorodov p.v . ,
zhilkin a.g . ,
boyarchuk a.a .
, 2008 , astronomy reports , vol .
boffin h.m.j . , stanishev v. , kraicheva z. , & genkov v. , 2003 , aspc , 292 , 297 borczyk w. , schwarzenberg - czerny a. , szkody p. , 2003 ,
a&a , 405 , 663 boyton p.e . ,
crosa l.m .
, deeter j.e . , 1980 ,
apj , 237 , 169 caproni a. & abraham z. , 2002 , rmxac , 14 , 74 chakravarthy s.r . & osher s. , 1985 , aiaa , paper 85 - 0363 dhillon v.s . ,
. , marsh t. r. , & smith r.c .
, 1992 , mnras , 258 , 225 eggleton p.p . , 1983 , apj , 268 , 368 einfeldt b. , 1988 , siam j. numer .
, 25 , 294 foulkes s.b . , haswell c.a . ,
& murray j.r . , 2006 , mnras , 366 , 1399 frank j. , king a.r . , & raine d.j . , 2002 ,
accretion power in astrophysics ( 3rd ed .
; cambridge : cambridge univ . press ) fu h. , li z .- y . , leung k .- c .
, zhang z .- s .
, & gaskell c.m . , 2004 ,
chjaa , 4 , 88 gao w. , li z. , wu x. , zhang z. , & li y. , 1999 , apj , 527 , 55 gulsecen h. , retter a. , liu a. , esenoglu h. , 2009 , newa , 14 , 330 harvey d. , skillman d.r .
, patterson j. , & ringwald f.a . , 1995 ,
pasp , 107 , 551 hellier & buckley d.a.h . , 1993 , mnras , 265 , 766 hoard d.w .
& szkody p. , 1997 ,
apj , 481 , 433 huber m.e .
howell s.b .
, ciardi d.r . , &
fried r. , 1998 , pasp , 110 , 784 kang t.w .
, retter a. , liu a. , & richards m. , 2006 , aj , 131 , 1687 kato s. , 1983 , pasj , 35 , 249 kato s. , 2001 , pasj , 53 , 1 kato t. , ishioka r. , & uemura m. , 2002 , pasj , 54 , 1033 katz j.i .
, 1973 , nature phys .
, 246 , 87 kim y. , andronov i.l .
, cha s.m . ,
chinarova l.l .
, & yoon j.n . , 2009 ,
a&a , 496 , 765 kozhevnikov v.p .
, 2004 , a&a , 419 , 1035 kozhevnikov v.p . , 2007 , mnras , 378 , 955 kraicheva z. , stanishev v. , genkov v. , & iliev l. , 1999 a&a , 351 , 607 kumar s. , 1986 , mnras , 223 , 225 lanzafame g. , 2003 , a&a , 403 , 593 larwood j. , 1997 , mnras , 290 , 490 larwood j. , 1998 , mnras , 299 , l32 larwood j. , nelson r.p .
, papaloizou j.c.b .
, & terquem c. , 1996 , mnras , 282 , 597 lasota j .-
p . , 2001 , new astronomy review , 45 , 449 lattanzio j.c .
, monaghan j.j . , pongracic h. , & schwarz m.p . , 1986 , j. sci .
comput . 7 , 591 li z. , leung k .- c . , &
gaskell c.m . , 2004 , rmxac , 21 , 259 livio m. , 1999 , phr , 311 , 225 lubow s.h . ,
1991a , apj , 381 , 259 lubow s.h .
, 1991b , apj , 381 , 268 lubow s.h . , 1992 , apj , 398 , 525 lubow s.h .
& pringle j.e . , 1993 , apj , 409 , 360 lubow s.h .
, ogilvie g.i . , &
pringle j.e . , 2002 , mnras , 337 , 706 meyer f. & meyer - hofmeister e. , 1981 , a&a , 104 , l10 monaghan j.j . , 1992 , ara&a , 30 , 543 montgomery m.m . , 2001 ,
mnras , 325 , 761 montgomery m.m .
, 2004 , ph.d thesis , florida institute of technology montgomery m.m . , 2009a , apj , 705 , 603 montgomery m.m . ,
2009b , mnras , 394 , 1897 murray j.r .
& armitage p.j .
, 1998 , mnras , 300 , 561 murray j.r . , truss m.r . , &
wynn g.a . , 2002 , aspc , 261 , 416 nelemans g. , steeghs d. , & groot p.j
. , 2001 , mnras , 326 , 621 olech a. , rutkowski a. , & schwarzenberg - czerny a. , 2009arxiv0906.3964 olech a. , rutkowski a. , & schwarzenberg - czerny a. , 2007 , aca , 57 , 331 paczynski b. , 1977 , apj , 216 , 822 papaloizou j. & pringle j.e . , 1983 , mnras , 202 , 1181 papaloizou j. & lin d. , 1995 , apj , 438 , 841 papaloizou j.c.b . &
terquem c. , 1995 , mnras , 274 , 987 patterson j. , 1998 , pasp , 110 , 1132 patterson j. , thomas g. , skillman d.r .
, & diaz m. , 1993 , apjs , 86 , 235 patterson j. , jablonski f. , koen c. , odonoghue d. , & skillman d.r . , 1995 , pasp , 107 , 1183 patterson j. , kemp j. , saad j. , skillman d. , harvey d. , fried r. , thorstensen .r . , ashley r. , 1997 , pasp , 109 , 468 patterson j. , thorstensen j.r . , fried r. , skillman d.r . , cook l.m . , & jensen l. , 2001 , pasp , 113 , 72 patterson j. , fenton w.h .
, thorstensen j.r . , harvey d.a . ,
skillman d.r .
, fried r.e .
, monard b. , odonoghue d. , beshore e. , martin b. , niarchos p. , vanmunster t. , foote j. , bolt g. , rea r. , cook l. , butterworth n. , & wood m. , 2002 , pasp , 114 , 1364 patterson j. , kemp j. , harvey d.a . , fried r.e .
, rea r. , monard b. , cook l. , skillman d.r . , vanmunster t. , bolt g. , armstrong e. , mccormick j. , krajci t. , jensen l. , gunn j. , butterworth n. , foote j. , bos m. , masi g. , & warhurst p. , 2005 ,
pasp , 117 , 1204 petterson j. , 1977 , apj , 216 , 827 pringle j.e . , 1996 , mnras , 281 , 357 pringle j.e . , 1997 , mnras , 292 , 136 pringle j.e . , 1999 , in sellwood j.a . , goodman j. , eds , asp conf .
160 , astrophysical discs , astron .
soc . pac .
, san francisco , p. 53 retter a. , leibowitz e.m .
, & ofek e.o .
, 1997 , mnras , 286 , 745 retter a. , hellier c. , augusteijn t. , naylor t. , bedding t.r . , bembrick c. , mccormick j. , & velthuis f. , 2003 , mnras , 340 , 679 retter a. , liu a. , & bos m. , 2005 , assl , 332 , 251 rodriguez - gil p. ( and 17
others ) , 2007 , mnras , 377 , 1747 rodriguez - gil p. , schmidtobreick l. , & gansicke b.t . , 2007b , mnras , 374 , 1359 roe p.l . , 1986 ,
anrfm , 18 , 337 roelofs g.h.a .
, groot p.j . ,
nelemans g. , marsh t.r . , & steeghs d. , 2006 , mnras , 371 , 1231 romero g.e . , chajet l. , abraham z. , & fan j.h . , 2000 ,
a&a , 360 , 57 schoembs r. , 1982 , a&a , 115 , 190 shafter a.w . &
szkody p. , 1983 , pasp , 95 , 509 shakura n.i .
& sunyaev r.a . , 1973 , a&a , 24 , 337 simpson j.c . , 1995 ,
apj , 448 , 822 simpson j.c . & wood m.a . , 1998 ,
apj , 506 , 360 skillman d.r .
, patterson j. , kemp j. , harvey d.a .
, fried r.e .
, retter a. , lipkin y. , vanmunster t. , 1999 , pasp , 111 , 1281 smak j. , 1999 , acta astronomica , 49 , 391 smith a.j .
, haswell c.a . , murray j.r . ,
truss m.r . ,
& foulkes s.b . , 2007 , mnras , 378 , 785 smith d.a .
& dhillon v.s . , 1998 , mnras , 301 , 767 szkody p. & howell s.b . , 1993 , apj , 403 , 743 taylor c.j . , thorstensen j.r . , patterson j. , fried r.e . , vanmunster t. , harvey d.a . , skillman d.r .
, jensen l. , & shugarov s. , 1998 , pasp , 110 , 1148 thorstensen j.r . ,
taylor c.j . ,
becker c.m . , & remillard r.a .
, 1997 , pasp , 109 , 477 terquem c. & papaloizou j.c.b .
, 2000 , a&a , 360 , 1031 tramposch j. , homer l. , szkody p. , henden a. , silvestri n.m . , yirak k. , fraser o.j . , &
brinkmann j. , 2005 , pasp , 117 , 262 warner b. , 1995 , cataclysmic variable stars ( new york : cambridge u. press ) warner b. , 2003 , cataclysmic variable stars ( new york : cambridge u. press ) wijers r.a.m.j . &
pringle j.e . , 1999 ,
mnras , 308 , 207 wood m.a . , montgomery m.m . , &
simpson j.c . ,
2000 , apjl , 535 , l39 woudt p.a . & warner b. , 2002 , mnras , 335 , 44 woudt p.a .
, warner b. , & spark m. , 2005 , 364 , 107 wu x. , li z. , ding y. , zhang z. , & li z. , 2002 , apj , 569 , 418 zhilkin a.g . & bisikalo d.v . | in montgomery ( 2009a ) , we show that accretion discs in binary systems could retrogradely precess by the same physics that causes the earth to retrogradely precess [ i.e. , tidal torques by the moon and the sun ( or secondary star ) on a tilted , spinning , non - spherical earth ( or e.g. , primary star surrounded by an accretion disc ) ] . in addition
, we show that the state of matter and the geometrical shape of the celestial object could significantly affect the precessional value .
for example , a cataclysmic variable ( cv ) dwarf novae ( dn ) non - magnetic system that shows negative superhumps in its light curve can be described by a retrogradely precessing , differentially rotating , tilted disc . because the disc is a fluid and because the gas stream overflows the tilted disc and particles can migrate into inner disc annuli , coupled to the disc could be a retrogradely precessing inner ring that is located near the innermost annuli of the disc . however , numerical simulations by bisikalo et al .
( 2003 , 2004 ) and this work show that an inner spiral density wave can be generated instead of an inner ring .
therefore , we show that retrograde precession in non - magnetic , spinning , tilted cv dn systems can equally be described by a retrogradely precessing and differentially rotating disc with an attached retrogradely precessing inner spiral density wave so long as the wave appears at the same radius as the ring and within the plane of the tilted disc .
we find that the theoretical results generated in this work agree well with the theoretical results presented in montgomery ( 2009a ) and thus with the numerical simulations and select cv dn systems in montgomery ( 2009b ) that may have a main sequence secondary .
therefore , pressure effects do need to be considered in cv dn systems that exhibit negative superhumps if the accretion discs are tilted and have an inner spiral density wave that is in the plane of the disc . accretion discs with an inner spiral density wave m.m .
montgomery@xmath0 & d.v .
bisikalo@xmath1 _ @xmath0 department of physics , university of central florida , orlando , fl 32816 , usa + @xmath1 institute of astronomy , moscow _
keywords : accretion , accretion discs - binaries : general - binaries : close - stars : dwarf novae - novae , cataclysmic variables - hydrodynamics |
fluctuations and correlations of the conserved charges have been proposed to be sensitive observables to study the qcd phase transition and critical point ( cp ) in relativistic heavy - ion collisions @xcite .
theoretical calculations show that fluctuations and correlations of conserved charges are distinctly different in the hadronic and quark gluon plasma ( qgp ) phases @xcite . in the deconfined state of quarks and gluon ,
the elementary set of conserved charges are given by the corresponding quark flavors : net upness(@xmath6=u-@xmath7 ) , net downess(@xmath8=d-@xmath9)and net strangeness(@xmath10=s-@xmath11 ) @xcite .
in hadronic state the conserved charges are net - baryon ( b ) , net - charge ( q ) and net - strangeness ( s ) .
recently , lattice qcd calculations have shown that various order ( up to fourth order ) net - strangeness ( s ) fluctuations and their correlations with net - charge ( q ) and net - baryon ( b ) are quite sensitive to the quark - hadron phase transition @xcite . in the years 2010@xmath122014
, relativistic heavy ion collider ( rhic ) has completed the first phase of the beam energy scan ( bes ) program @xcite . in this program the the collision energies in au+au collisions were tuned to explore the qcd phase structure at high baryon density @xcite .
the star experiment at rhic has collected au+au collisions data with @xmath0 = 7.7 , 11.5 , 14.5 , 19.6 , 27 , 39 , 62.4 , 200 gev , respectively .
one of the experimental motivation is to measure the correlations between net - baryon and net - strangeness with these experimental data collected to study the qcd phase transition in heavy - ion collisions @xcite . in order to provide baselines for the correlation measurements to study the phase transition , we have calculated these observables in au+au collisions at rhic bes energies with urqmd model and compared them to the model results from lattice qcd and hadron resonance gas model . in addition given at lower energies we expect to see the effect of baryon stopping and associated production of strangeness , we have also studied the contributions of various hadron species to the baryon - strangeness correlations .
the paper is organised as follows . in the next section ,
we briefly review about fluctuations and correlations of conserved charges in hadronic and quark - gluon plasma states and give the corresponding results from hadronic resonance gas model and lattice qcd . in section iii
, we present the mixed - cumulant ratios of net - baryon and net - strangeness and also discuss the contribution of strange baryons to the baryon - strangeness correlations in au+au collisions at rhic bes energies from urqmd model .
then , we give a brief summary in section iv .
finally , we present as an appendix the statistical error calculations for various order mixed - cumulants .
in statistics , the generating function of mixed - cumulants for random variables , @xmath13 , @xmath14 , ... , @xmath15 are defined as : @xmath16 then , the mixed - cumulants of random variables @xmath13 , @xmath14 , ... , @xmath15 can be expressed as @xcite : @xmath17 where @xmath18 runs through the list of all partitions of 1,2, ... n , c runs through the list of all blocks of partitions @xmath18 , @xmath19 is the number of parts in the partition and @xmath20 is the number of parts in the block @xmath21 .
+ in mixed - cumulants analysis , we use @xmath22 to represent the net - baryon number and net - strangeness in one event . the average value over whole event ensemble is denoted by @xmath23 , @xmath24 .
the deviation of b , s from their mean value are expressed by @xmath25 , @xmath26 , respectively .
thus , according to eq .
[ eq : mixed - moments ] , the various order mixed - cumulants of event - by - event distributions of the two random variables b and s are defined as : @xmath27 once we have the definition of mixed - cumulants , the ratio of mixed - cumulants can be calculated . ] . the figure is from @xcite . ] a system in thermal equilibrium ( for a grand - canonical ensemble ) can be characterised by its pressure @xcite .
the dimensionless pressure of a hadron resonance gas is expressed through the logarithm of the qcd partition function @xcite : @xmath28\nonumber\\ & = & \frac{1}{2\pi}\sum_{i\in{x}}g_{i}(\frac{m_{i}}{t})^2k_{2}(\frac{m_{i}}{t})\nonumber\\ & \times & { cosh(b_{i}\hat{\mu}_{b}+q_{i}\hat{\mu}_{q}+s_{i}\hat{\mu}_{s})}.\end{aligned}\ ] ] where @xmath29 is the degeneracy factor for hadrons of mass @xmath30 , and @xmath31 , with @xmath32 , @xmath33 , @xmath34 denote the net - baryon number , net - strangeness and the net - charge , and @xmath35 , @xmath36 , @xmath37 are the corresponding chemical potentials respectively . for simplicity
, we have set the electric charge chemical potential @xmath38 these dimensionless fluctuations and correlations of conserved charges ( net - baryon number @xmath39 , net - strangeness @xmath40 ) are formally equivalent to the taylor expansion coefficients of the pressure with the respective chemical potentials @xcite : @xmath41}{\partial\hat{\mu}_b^m\hat{\mu}_s^n}|_{\mu_{b}=\mu_{s}=0}.\end{aligned}\ ] ] where @xmath42 and @xmath43 are the dimensionless baryon and strangeness chemical potentials , @xmath44 and @xmath45 .
they are also called generalised susceptibilities .
all these derivatives are evaluated at @xmath46 , the expectation values of all net charge numbers vanish , i.e. , @xmath47 @xcite .
theoretically , the mixed - cumulants of these conserved quantities are connected to the corresponding susceptibilities by : @xmath48 where the @xmath49 denotes respectively , the volume and temperature of the system .
+ if above the transition temperature the quarks can be well described by uncorrelated quasiparticles , then any object which carries strangeness necessarily is a strange quark , then carries baryon number which in proportion to strangeness , @xmath50@xcite . and also , from eq .
[ eq : susceptibilites ] , we can derive : @xmath51 where @xmath52 and @xmath53 thus one expects baryon number and strangeness to be correlated stronger in a quark - gluon plasma than in a hadron gas . the hot medium created in relativistic heavy ion collisions
is an expanding system , the spatial volume is changing with time evolution . in order to cancel the effect of the spatial volume dependence to first order ,
the ratios @xmath54 , @xmath55 , @xmath56 , @xmath57 are constructed . in order to quantify this and make these ratios to be unity for system where the relevant degree of freedom are quarks
, we introduce the following normalised ratios in eq .
[ eq : normalized r11]-[eq : normalized r31 ] : @xmath58 at vanishing chemical potentials , the mean values of the conserved charges are vanishing @xmath59 , the above equations can be simplified as : @xmath60 for illustration and discussion purpose , the formula with @xmath35=@xmath36=0 are just to demonstrate a special case , which is not used in the calculations presented subsequently . at high temperature , the basic degrees of freedom are weakly interacting quarks and gluons .
the quark operators @xmath61 represent the net - quark number of up , down and strange quarks .
we expressed @xmath3 , @xmath4 , @xmath5 and @xmath2 in terms of quark degrees of freedom , noting that the baryon number of a quark is @xmath62 and the strangeness of a strange quark is negative one@xcite . in terms of quark flavours , the various order ratios @xmath3 , @xmath4 , @xmath5
, @xmath2 can be written as : @xmath63\langle{s^2}\rangle } { \langle{s^4}\rangle -3\langle{s^2}\rangle^2},\\ r_{31}^{bs}|_{\mu_{b},\mu_{s}=0}&=&\frac{\langle{(u+d+s)^3s}\rangle -3\langle{(u+d+s)s}\rangle\langle{(u+d+s)^2}\rangle}{\langle{s^4}\rangle -3\langle{s^2}\rangle^2}\nonumber\\ & = & 1+\frac{\langle{(u+d)^3s}\rangle + 3\langle{(u+d)^2s^2}\rangle -3\langle{(u+d)^2}\rangle\langle{s^2}\rangle + 3\langle{(u+d)s^3}\rangle } { \langle{s^4}\rangle -3\langle{s^2}\rangle^2}\nonumber\\ & -&\frac{-3\langle{(u+d)s}\rangle[\langle{(u+d)^2}\rangle + 2\langle{(u+d)s}\rangle + 3\langle{s^2}\rangle ] } { \langle{s^4}\rangle -3\langle{s^2}\rangle^2}.\end{aligned}\ ] ] for uncorrelated quark flavours , off - diagonal susceptibilities are vanishing compared to the diagonal susceptibilities@xcite , susceptibilities inclusive of strangeness are smaller than those which involve lighter flavours@xcite , then we have @xmath64 and thus @xmath65 where @xmath66 and m+n=2 , 4 . in the non - interacting massless quark gas all these observables are unity ( shown by the lines at high temperatures in figure [ fig:1 ] ) . at high temperatures ,
the shaded regions indicate the ranges of values for these ratios as predicted for the weakly interacting quasi - quarks from the re - summed hard thermal loop ( htl ) perturbation theory at the one - loop order @xcite .
it is well known that at low temperature and chemical potentials the qcd can be modelled as a gas of uncorrelated hadrons , where interactions are included through resonances @xcite .
these states are taken from particle data book @xcite .
as the interactions between hadrons are suppressed , the contributions of individual hadrons to thermodynamics can be regarded as free boltzmann gas @xcite . for a gas of uncorrelated hadron resonances , from eq .
[ eq : normalized r11]-[eq : normalized r31 ] , the numerator receives contributions from only ( strange ) baryons and anti - baryons , while the denominator receives contributions also from strange mesons @xcite .
[ eq : normalized r11 ] can be written as : @xmath67 ref @xcite gives the lattice qcd results for the appropriate combinations of up to fourth order cumulants of net - strangeness fluctuations and their correlations with net - baryon number and electric charge fluctuations and shown in fig .
[ fig:1 ] .
the ratios of the second order correlations @xmath68 and @xmath69 are much closer to the expectations for weakly interacting quasi - quarks , differing only around t-1.25@xmath70 . while the htl perturbative expansion for ratios involving one derivate of the baryonic / electric charges ( @xmath71 and @xmath72 ) @xcite starts differing from the non - interacting quark gas limit @xcite .
it is the same for those involving higher derivatives of the baryonic /electric charges(@xmath73 and @xmath74 ) .
the ultra relativistic quantum molecular dynamics ( urqmd ) model is a microscopic transport model for simulating heavy - ion collisions in the energy range from sis ( schwerionen heavy - ion synchrotron ) to relativistic heavy ion collider ( rhic ) and even for the heavy - ion collisions at the large hadron collider(lhc ) @xcite . in this model
@xcite , the space - time evolution of the fireball is studied in terms of excitation and fragmentation of color strings and formation and decay of hadronic resonances . in this paper
, we performed our calculations with urqmd model in the version 2.3 for au+au collisions at @xmath0= 7.7 , 11.5 , 19.6 , 27 , 39 , 62.4 , 200 gev and the corresponding event statistics are 72.5 , 105 , 106 , 81 , 133 , 38 and 56 million , respectively @xcite .
the statistical errors are calculated by the formula that are derived from the standard error propagations ( see appendix [ error ] ) . to avoid auto - correlations
, we define the centralities with the charged particles within the pseudo - rapidity @xmath75 and perform the analysis in the pseudo - rapidity range ( @xmath76 ) . for the b - s correlations ,
we include the particles @xmath77 , @xmath78 , @xmath79 , @xmath80 , @xmath81 , @xmath82 , @xmath83 , @xmath84 , @xmath85 , @xmath86 , @xmath87 and their anti - particles .
these particles can be classified into strange baryons ( @xmath81 , @xmath88 , @xmath89 , @xmath90 ) , non - strange baryons ( @xmath77,@xmath78 ) and strange mesons ( @xmath91,@xmath80 ) . in order to study the contributions of various hadron species to the baryon - strangeness correlations
, we consider the b - s correlations with the following combinations of hadrons : a. [ item:1 ] net - p vs. net - k : only @xmath77 , @xmath92 , @xmath79 , @xmath93 are included ;
b. [ item:2 ] net - p vs. net-@xmath81 : only @xmath77 , @xmath92 , @xmath81 , @xmath94 are included ; c. [ item:3]net-@xmath81 vs. net - k : only @xmath81 , @xmath94 , @xmath79 , @xmath93 are included ; d. [ item:4]b - s ( all particles ) : @xmath77 , @xmath78 , @xmath79 , @xmath80 , @xmath81 , @xmath88 , @xmath89 , @xmath90 and their anti - particles are included .
e. [ item:5]b - s ( excl .
strange baryon ): only @xmath77 , @xmath78 , @xmath79 , @xmath80 are included ; f. [ item:6]b - s ( excl .
@xmath81 , @xmath94 ) : only @xmath77 , @xmath78 , @xmath79 , @xmath80 , @xmath88 , @xmath89 , @xmath90 and their anti - particles are included .
g. [ item:7]b - s ( excl .
non - strange baryon ) : only @xmath79 , @xmath80 , @xmath81 , @xmath88 , @xmath89 , @xmath90 and their anti - particles are included .
h. [ item:8]b - s ( excl .
@xmath77 , @xmath92 ) : only @xmath78 , @xmath79 , @xmath80 , @xmath81 , @xmath88 , @xmath89 , @xmath90 and their anti - particles are included .
i. [ item:9]b - s ( excl .
strange meson ) : only @xmath77 , @xmath78 , @xmath81 , @xmath88 , @xmath89 , @xmath90 and their anti - particles are included .
j. [ item:10]b - s ( excl .
@xmath79 , @xmath93 ) : only @xmath77 , @xmath78 , @xmath80 , @xmath81 , @xmath88 , @xmath89 , @xmath90 and their anti - particles are included .
figure [ fig : fig3 ] shows the centrality dependence for the cases [ item:4]-[item:10 ] in the au+au collisions at @xmath0 = 39 gev from urqmd model .
the top panels of fig .
[ fig : fig3 ] displays the two dimension correlations for the cases [ item:4]-[item:10 ] for most central ( 0 - 5% ) au+au collisions . to quantify the correlation strength , we calculated the standard correlation coefficient for each case . in statistics , the correlation coefficient ( @xmath95 ) is a measure of the linear dependence between two variables @xmath96 and @xmath97 , and it is the covariance of the two variables divided by the product of their standard deviations .
it is defined as @xmath98 where @xmath99 is the covariance between random variables @xmath96 and @xmath97 .
the @xmath100 and @xmath101 are the standard deviation of x and y , respectively .
the definition gives a value between + 1 and -1 , where a value of 1 represents strong correlation , value of 0 indicates no linear correlation , and value of -1 gives perfect anti - correlation .
these two dimension plots indicate that most of the cases are anti - correlated and dominated by the strange baryons , such as @xmath81 .
the bottom panels of fig .
[ fig : fig3 ] show the centrality dependence of mixed - cumulants ( @xmath102 , @xmath103 , @xmath104 , @xmath105 , @xmath106 , @xmath107 ) and ratios ( @xmath3 , @xmath4 , @xmath5 , @xmath2 ) for the cases [ item:4]-[item:10 ] in au+au collisions at @xmath0 = 39 gev .
it is found that the @xmath102 , @xmath103 and @xmath105 have negative values , while @xmath104 have positive values . a strong centrality dependence for the baryon - strangeness ( b - s ) correlations for the cases [ item:6]-[item:10 ] is observed . by comparing the case [ item:4 ] with [ item:5 ] , we find that the strange baryons have significant contributions to the baryon - strangeness correlations , and the strange mesons ( @xmath108 and @xmath80 ) contribute significantly to strangeness fluctuations of the case [ item:9 ] @xcite .
when the strange baryons are excluded from the @xmath39-@xmath40 correlations , the mixed - cumulant ratios are much smaller than the case with all particles and has values close to zero . on the other hand ,
when the strange mesons are excluded from the @xmath39-@xmath40 correlations , the values of the ratios ( @xmath3 , @xmath4 , @xmath5 , @xmath2 ) become larger than the case with all particles .
furthermore , the results from the case where the non - strange baryons are not included is very close to the results from the case with all particles included .
this indicates that the non - strange baryons have little contributions to the baryon strangeness correlations . since these ratios show weak centrality dependence , we will only consider the most central ( 0 - 5% ) collision centralities for the energy dependence .
-[item:4 ] for most central ( 0 - 5% ) au+au collisions from urqmd model . ] figure [ fig : fig2 ] shows the centrality dependence of b - s correlations for the cases : [ item:1]-[item:4 ] in au+au collisions at @xmath0 = 39 gev .
the left panels of fig .
[ fig : fig2 ] show the correlations for different cases in the most central ( 0 - 5% ) au+au collisions .
the right panels of fig .
[ fig : fig2 ] show the centrality dependence of mixed - cumulants ( @xmath102 , @xmath103 , @xmath104 , @xmath105 , @xmath106 , @xmath107 ) and ratios ( @xmath3 , @xmath4 , @xmath5 , @xmath2 ) for the cases [ item:1]-[item:4 ] . in the two dimension plots
, we find that the net - baryon versus net - strangeness shows strong anti - correlations .
if the strange baryons are excluded , for eg . the net - proton and net - kaon correlations are almost independent with each other ( @xmath109 0 ) .
it indicates that the strange baryons play an important role and have significant contributions to the @xmath39-@xmath40 correlations . in the @xmath39-@xmath40 correlations with all particles ,
the @xmath102 , @xmath103 and @xmath105 show negative values and monotonically decrease with increasing of the number of participant nucleons in the collisions , while the @xmath104 have large and positive values for all of the cases .
interestingly , the ratios @xmath3 , @xmath4 , @xmath5 and @xmath2 with all particles included show weak centrality dependence and are comparable with the lattice qcd results shown in the fig .
[ fig:1 ] . in the left panels of fig .
[ fig : fig2 ] , the net - proton and net-@xmath81 , net - k and net-@xmath81 show finite positive correlations .
this can be understood in terms of the baryon stopping and associate productions of the @xmath79 and @xmath81 .
their mixed - cumulants are positive values , but the ratios @xmath3 , @xmath4 and @xmath2 are negative . furthermore , the variance of the fluctuation of net-@xmath81 is much smaller than the variance of net - kaon distributions .
this leads to bigger deviation from zero for the proton-@xmath81 than kaon-@xmath81 correlations .
figure [ fig : fig4 ] shows the energy dependence of the mixed - cumulants and ratios for the cases [ item:1]-[item:4 ] in the most central ( 0 - 5% ) au+au collisions from urqmd model . in the @xmath39-@xmath40 correlation including all particles ,
the mixed - cumulants @xmath102 , @xmath103 , @xmath105 are negative and show weak energy dependence , while the @xmath106 , @xmath107 and @xmath104 are positive and monotonically increase with increasing energy . on the other hand ,
the ratios ( @xmath3 , @xmath4 , @xmath5 ) show weak energy dependence but the ratio @xmath2 significantly increases at low energies .
this large increasing trend of the @xmath2 are also observed in the correlations between net - proton and net - kaon , which might originate from the associate production of @xmath79 by the reaction channel @xmath110 , where @xmath111 is nucleon and @xmath97 is hyperon .
it is found that the @xmath39-@xmath40 correlations including all particles are comparable with the results from lattice qcd at low temperature ( lattice qcd results shown in fig . [ fig:1 ] ) .
the trends of these ratios for the correlations between net - proton and net-@xmath81 are opposite to the trends observed in the net - kaon and net-@xmath81 correlations . for the two cases , the mixed - cumulant ratios @xmath3 , @xmath4 and @xmath2 are negative , @xmath5 is positive , but higher order ratios @xmath5 and @xmath2 have larger value . to demonstrate the energy dependence of the second order ratio @xmath3 more clearly , we plot @xmath3 as a function of energy in the bottom panels of fig .
[ fig : fig4 ] .
it is found that when all particles are considered or only proton and lambda are included , the @xmath3 decreases with increasing energy , while we observe different trends for the other two cases . at low energy ,
the higher order ratio @xmath2 becomes large for [ item:1 ] ( net - p vs. net - k ) and [ item:4 ] ( net - b vs. net - s ) , the main reason is probably because that at low energy , the baryon stopping of protons and associate kaon production play important roles .
figure [ fig : fig5 ] shows the energy dependence of various order mixed - cumulants and ratios of the baryon - strangeness correlations for the cases [ item:4]-[item:10 ] for the most central ( 0 - 5% ) au+au collisions from urqmd model . by excluding the non - strange baryons or protons from the @xmath39-@xmath40 correlations , we find that the values of ratios @xmath3 , @xmath4 , @xmath5 , @xmath2 are finite and monotonically increases with decreasing collision energy . for the @xmath3 , the case including all particles are very close to the one excluding non - strange baryons or protons . it means the non - strange baryons have small effects on the @xmath39-@xmath40 correlations .
however , if we exclude the strange baryons or @xmath81 , the values of baryon - strangeness correlations are close to zero at high energies .
it indicates that the strange baryons , especially the @xmath81 baryons , carry significant information and play an important role for the @xmath39-@xmath40 correlations .
especially , the @xmath2 show large increase at low energy .
it probably because that @xmath81 is the lightest strange baryon and contributes most significantly to the baryon - strangeness correlations .
when the @xmath79 , @xmath93 are excluded , there are large impacts on the higher order ratios @xmath5 and @xmath2 at low energies .
these two ratios become negative , which are probably due to the kaon associate production at low energy .
there is a strong correlation between protons and kaons as shown in the fig .
[ fig : fig4 ] .
if we remove the kaons from the @xmath39-@xmath40 correlations , the correlations between proton and @xmath81 will dominate the @xmath39-@xmath40 correlations , which is negative .
we have analysed the centrality and energy dependence of the baryon - strangeness ( @xmath39-@xmath40 ) correlations in au+au collisions at @xmath0 = 7.7 , 11.5 , 19.6 , 27 , 39 , 62.4 and 200 gev from urqmd model .
the @xmath39-@xmath40 correlations are studied via various order mixed - cumulants and corresponding ratios . by excluding different hadron species from @xmath39-@xmath40 correlations
, we have studied ten different cases : [ item:1]-[item:10 ] . in these studies , we observed strong correlations between net - proton , net-@xmath81 and net - kaon at low energies , which could be related to the baryon stopping and associated production of the hyperons and kaons .
the various results of the mixed - cumulants and ratios obtained from urqmd calculations show a weak centrality dependence .
as far as the energy dependence is concerned , we observe that the strange - baryons and strange - meson have significant contributions to baryon - strangeness correlations and various order mixed - cumulant ratios @xmath4 , @xmath4 , @xmath5 and @xmath2 , especially at low energy . when strange - baryons are excluded from the @xmath39-@xmath40 correlations ,
the values are relatively small and close to zero .
it is found that the model results are also comparable with the lattice qcd calculations at low temperature .
these studies can help us in understanding the interplay of baryon stopping and kaon associate production to @xmath39-@xmath40 correlations at different collision energies .
it provides a baselines for the @xmath39-@xmath40 correlations as an observable to search for the signals of the qcd phase transition and critical point in heavy - ion collisions .
the work was supported in part by the most of china 973-project no.2015cb856901 , nsfc under grant no .
bm is supported by dae , dst and serb , govt . of india
. 20 m. stephanov , k. rajagopal , e. shuryak , phys . rev . lett . * 81 * , 4816 ( 1998 ) ; m. stephanov , k. rajagopal , e. shuryak , phys . rev .
d * 60 * , 114028 ( 1999 ) ; m. stephanov , phys . rev .
lett . * 102 * , 032301 ( 2009 ) ; m. stephanov , phys . rev . lett . * 107 * , 052301 ( 2011 ) .
s. jeon and v. koch , phys .
lett . * 83 * , 5435 ( 1999 ) ; s. jeon and v. koch , phys .
lett . * 85 * , 2076(2000 ) .
m. asakawa , u. heinz and b. muller , phys .
lett . * 85 * , 2072 ( 2000 ) ; m. asakawa , s. ejiri and m. kitazawa , phys .
lett . * 103 * , 262301 ( 2009 ) ; m. asakawa and m. kitazawa , arxiv : 1512.05038 .
s. gupta , x. luo , b. mohanty , h. g. ritter , n. xu , science , * 332 * , 1525 ( 2012 ) .
s. borsanyi , proc .
sci . lattice , * 015 * ( 2015 ) , arxiv : 1511.06541m .
et al . _ ,
phys . rev .
d 79 , 074505 ( 2009 ) ; a. bazavov , _ et al .
d * 86 * , 034509 ( 2012 ) .
s. borsanyi , z. fodor , s. d. katz , s. krieg , c. ratti and k. szabo , jhep * 1201 * , 138 ( 2012 ) .
a. bazavov , h. t. ding , p. hegde _
d * 88 * , 094021 ( 2013 ) .
r. v. gavai and s. gupta , phys .
d * 73 * , 014004 ( 2016 ) ; r. v. gavai and s. gupta , phys . rev .
d 78 , 114503 ( 2008 ) .
a. bazavov , h. t. ding , p. hegde _
et al . _ ,
phys . rev .
lett . * 109 * , 192302 ( 2012 ) . m. m. aggarwal _ et al .
_ , arxiv:1007.2613 .
x. luo ( for the star collaboration ) , pos(cpod2014)019 .
[ arxiv:1503.02558 ] .
x. luo , nucl .
a * 956 * , 75 ( 2016 ) . [ arxiv:1512.09215 ] .
m. m. aggarwal _ et al .
_ ( star collaboration ) , phys . rev
105 * 022302 ( 2010 ) ; l. adamczyk _
et al . _ , ( star collaboration ) , phys .
lett . * 113 * , 092301 ( 2014 ) .
l. adamczyk _
( star collaboration ) phys .
112 * , 032302 ( 2014 ) .
a. chatterjee , s. chatterjee , t. k. nayak _ et al .
_ , arxiv : 1606.09573 . h. -t .
ding , s. mukherjee _ et al .
d * 92 * , 074043 ( 2015 ) . c. ahlbach , j. usatine and n. pippenger , arxiv : 1211.0652 [ math.co ] .
a. bazavov _
et al . _ ,
. rev.lett . * 111 * , 082301(2013 ) .
a. bazavov , h .-
ding , p. hegde _
lett . * 113 * , 072001 ( 2014 ) .
_ , phys . lett .
b * 738 * , 305 - 310 ( 2014 ) .
r. bellwied , s. borsanyi , z. fodor _ et al .
lett * 111 * 202302 ( 2013 ) .
d. k. mishra , p. k. netrakanti , and b. mohanty , arxiv : 1606.06538 .
s. mukherjee , arxiv : 1307.6255 .
a. bazavov _
_ [ hotqcd collaboration ] , phys . rev .
d * 86 * , 034509 ( 2012 ) . v. koch , arxiv : 0810.2520 .
p. b .- munzinger _
et al . _ ,
phys . rep . * 621 * ( 2016 ) 76 - 126 . c. r. allton , m. doring , s. ejiri , s. j. hands _
et al . _ , phys .
d * 71 * , 054508 ( 2005 ) .
r. v. gavai and s. gupta , phys .
d * 65 * , 094515 ( 2002 ) .
a. majumder and b. muller , phys .
c * 74 * , 054901 ( 2006 ) .
r. v. gavai and s. gupta , phys .
d * 67 * , 034501 ( 2003 ) .
r. v. gavai and s. gupta , phys .
d * 71 * , 114014 ( 2005 ) .
j. o. andersen , s. mogliacci , n. su and a. vuorinen , phys .
d * 87 * , 074003 ( 2013 ) .
n. haque , a. bandyopadhyay , j. o. andersen _ et al . _ , 10.1007/ jhep * 05 * ( 2014 ) 027 .
p. b .- munzinger , k. redlich and j. stachel , arxiv : 0304013[nucl - th ] .
j. beringer , j .- f .
arguin , r. m. barnett _
et al . _ , [ particle data group ] ,
d * 86 * , 010001 ( 2012 ) . v. koch , a. majumder , and j. randrup phys .
lett . * 95 * 182301 ( 2005 ) , arxiv : 0505052 [ nucl - th ] .
p. b. -munzinger , k. redlich , and j. stachel , in quark gluon plasma 3 ( ref.[2 ] ) , pp .
491 - 599 .
blaizot , e. iancu , and a. rebhan , phys .
b*523 * , 143 ( 2001 ) .
m. bleicher _
et al . _ ,
g : nucl . part .
25 , 1859 ( 1999 ) .
s. a. bass _ et al .
41 , 225 - 370 ( 1998 ) .
h. petersen , m. bleicher , s. a. bass , and h. stocker , arxiv : 0805.0567v1 [ hep - ph ] .
m. bleicher , s. jeon , and v. koch , phys . rev .
c * 62 * , 061902(r ) ( 2000 ) .
s. haussler , h. stocker , and m. bleicher , phys .
c * 73 * , 021901(r ) 2006 .
ji xu , shili yu , feng liu , and xiaofeng luo , phys .
c * 94 * , 024901(2016 ) .
xiaofeng luo , phys .
c * 91 * , 034907 ( 2015 ; j. phys .
g : nucl . part .
39 025008 ( 2012 ) .
g. maurice and m. a. kendall , the advanced theory of statistice , vol .
i , 2nd ed .
( charles griffin & company , london , 1945 ) .
it is well known that the variance of statistic @xmath112 can be expressed in terms of the following @xcite : @xmath113 | fluctuations and correlations of conserved charges are sensitive observables for studying the qcd phase transition and critical point in high - energy heavy - ion collisions .
we have studied the centrality and energy dependence of mixed - cumulants ( up to fourth order ) between net - baryon and net - strangeness in au+au collisions at @xmath0= 7.7 , 11.5 , 19.6 , 27 , 39 , 62.4 , 200 gev from urqmd model . to compare with other theoretical calculations , we normalize these mixed - cumulants by various order cumulants of net - strangeness distributions .
we found that the results obtained from urqmd calculations are comparable with the results from lattice qcd at low temperature and hadron resonance gas model .
the ratios of mixed - cumulants ( @xmath1 , @xmath2 ) from urqmd calculations show weak centrality dependence .
however , the mixed - cumulant ratios @xmath3 and @xmath2 show strong increase at low energy , while the @xmath4 snd @xmath5 are similar at different energies .
furthermore , we have also studied the correlations between different hadron species and their contributions to the net - baryon and net - strangeness correlations .
these models studies can provide baselines for searching for the signals of qcd phase transition and critical point in heavy - ion collisions . |
after the completion of this text , i analyzed somewhat further the potential energy of the fluctuations of the mode density by separating it into two parts : when the soup is modeled by a collection of black holes ( or , in general , entities with horizons ) , then the source generates an inflaton mass of order of magnitude @xmath98 where @xmath99 is the entropy change of the black hole that is induced by the exchange . for schwarzschild black holes this is given by @xmath100 , @xmath101 mass of the black hole , @xmath102 mean energy of modes .
one expects both @xmath15 and @xmath103 to be @xmath104 whereupon one finds @xmath105 .
this is the order of magnitude required by the analysis of present day cosmic fluctuations .
: the above estimate is very crude and can easily change by several orders of magnitude due to its sensitivity on @xmath15 . | in the context of the two - fluid model introduced to tame the transplanckian problem of black hole physics , the inflaton field of the chaotic inflation scenario is identified with the fluctuation of the density of modes .
its mass comes about from the exchange of degrees of freedom between the two fluids .
in addition to all else , the major scientific revolutions of the 20th century , quantum mechanics and relativity , have made it possible to envision cosmogenesis on a rational basis . moreover , the theoretical approach that is envisioned has the potentiality ( which has at present to some extent been realized ) , of making contact with present day cosmological observations .
the full awareness of this potentiality came in the late 70 s and early 80 s .
it was first pointed out @xcite that gravity together with the quantum theory of matter allowed for the creation of matter _
ex nihilo_. at the same time it was shown that creation of matter at a constant rate gave rise to an exponential expansion of the scale factor @xcite , now called inflation and that this cured the perplexing problems of causality @xcite and flatness @xcite posed by the friedman robertson walker adiabatic expansion .
a glance at figure [ fig1 ] is sufficient to give one a rough appreciation of the causality problem and how one envisions its solution in terms of creation _
i.e. _ cosmogenesis .
now one gathers information contained within the backward light cone as drawn . at big bang time @xmath0 ,
this backward extrapolation ceases to make sense , if the space - like surface at @xmath1 contains @xmath2 degree of freedom per planckian cell .
there must be at least @xmath3 such cells since this is the number of degrees of freedom present within the horizon now , and this number is conserved during the adiabatic expansion ( to @xmath4 ) .
though on the way from @xmath5 to now there is some contact among the degrees of freedom due to scattering , this is far from enough to account for the observed homogeneity of the cmbr or galaxy distribution ( see @xcite chapter 8 for details ) .
therefore the homogeneity required at @xmath1 should result from a prior cause .
it is postulated that there is a germ , @xmath6 , from which energy is created and that the region , which contains this energy , expands exponentially .
this will be realized in the time interval between @xmath6 and @xmath5 if the energy density remains approximately constant within the region where the energy is created .
this is inflation .
an approach similar to that of @xcite was also suggested during this early period which was based on vacuum energy in the form of the trace anomaly @xcite .
whereas these early attempts were based essentially on production of kinetic energy , be it in the form of particles @xcite , black holes @xcite or zero point energy @xcite , a new tactic was introduced in the early 80 s which relied on potential energy .
the first idea in this vein was latent heat released in a first order phase transition @xcite .
it was postulated that the universe cooled down from a prior state of high temperature whereupon at some point the matter contained within would condense .
but it was recognized at once @xcite that this approach could not work .
the condensed phase would be delivered in bubbles which would be swept out by the expansion too fast to allow for their coalescence .
in addition i find that such a scenario would be epistemologically dissatisfying .
if one is out to solve the causality problem , then one would hope that the cause in question is the cause of the universe i.e. one would like cosmogenesis and inflation to go together . to push the ultimate problem further backwards in time
is , in a sense , to escape the problem .
but this is merely my subjective opinion .
fortunately , this approach could be modified through the invention of `` slow roll '' mechanisms @xcite .
this removed the problem of non coalescence of bubbles since there would be time enough . moreover in linde s hands @xcite chaotic inflation came into being .
we shall not here enter into details of how linde envisages fluctuations within fluctuations , but simply give that part of the idea which we shall adopt in the foregoing .
( we also refer the reader to @xcite for further consideration of this question of fractalizing the fluctuation ) .
the point of view we shall take here is that there is a state of rest in which there is no universe .
but because of quantum mechanics the `` stuff '' in this space fluctuates .
the stuff is complicated in its nature . only at the long distance scale
( and here we adopt the traditional point of view that the only fundamental scale is planckian ) does usual field theory apply .
the short distance scale is described by a `` planckian soup '' , the present great unknown of physics .
it may involve higher dimensions and/or non - commutative geometry and/or strings and/or black hole fluctuations and/or ... whatever , we shall only suppose that it is described by a density or , if one wishes , a scale which is planckian .
the abovementioned state of rest is interpreted variationally .
the mean density in this state is postulated to be the state of minimum energy density in which there is no expansion of the scale factor in the metric that describes this state . we shall postulate that this space time is flat _
i.e. _ @xmath7 , @xmath8 being the energy momentum tensor of everything .
this quiescent state of things is conceived as metastable .
fluctuations that are small both in amplitude and extension will regress and leave no permanent effects . but now comes this wonderous conjunction of quantum mechanics and relativity . in the hamiltonian formulation of general relativity ,
the total energy density of gravity + matter sums to zero .
instrumental in the realization of this fact is that the kinetic energy carried by the motion of the scale factor , @xmath9 , is negative ( _ i.e. _ in the hamiltonian density there is a term in @xmath10 where @xmath11 is the hubble constant : @xmath12 , where dot is derivative with respect to proper time ) .
thus the positive energy density of a fluctuation of matter density is necessarily accompanied by @xmath13 , so as to keep the energy density equal to zero .
this constraint is strict ; even in the future when fundamental theory will relegate the einsteinian theory of gravity to successful long scale phenomenology it will probably survive .
it is due to the invariance of the total action against reparametrezation of the temporal coordinate and one would be loath to give up such a fundamental principle in whatever guise it will express itself in future theory .
it is herein that opens the possibility of permanent effects of a fluctuation .
if this latter is large enough in amplitude ( to allow for a systematic temporal development ) and in extension ( so as to render spatial gradients sufficiently small in energy as compared to the total energy of the fluctuation ) then it will be seized upon by the scale factor and will expand exponentially immediately after its formation . after a while the fluctuation will nevertheless regress and one postulates its energy is converted to ordinary quanta whereupon the entropy so produced remains constant and frw adiabatic expansion ensues . at the present time , the formal development of the above scenario is executed in terms of a scalar field .
all the aforementioned concepts fall neatly into this development and furthermore , detailed quantitative analysis has shown that it is possible to account for presently observed density fluctuations both of the anisotropy of the cmbr and the distribution of galaxies . for details see @xcite , chapter 8 , and the lectures of lazarides at this summer school and workshop at corfu 2001 .
it is my purpose here to dwell on the possible nature of this fundamental fluctuation , often called the
inflaton
field albeit in highly speculative fashion .
who is he , this inflaton ?
my inspiration has been in great measure drawn from the experience that has been garnered from the theory of black hole physics over the past decades and which finds it expression , in the 2-fluid phenomenology that has been developed in recent years @xcite .
the theory of hawking evaporation @xcite from black holes , ( and we take as the simplest example s - wave emission form schwarzchild black holes ) encounters the transplanckian problem . in the original version of the theory , hawking studied free field propagation in the background schwarzchild metric .
he showed that a vacuum fluctuation in the vicinity of the horizon , as it propagates towards spatial infinity , is possessed of an amplitude which in part becomes that of a physical on massshell quantum due to the changing metric it experiences in its voyage .
the distribution of these quanta is thermal , at least for frequencies of the order of or larger than the inverse horizon distance @xmath14 , where @xmath15 is the mass of the black hole . in the derivation of this remarkable phenomenon one
appeals to configurations of fields which are localized to within fantastically small distances of the horizon @xmath16 ( where @xmath17 ) . here and
from now on we work in planckian units @xmath18 .
correspondingly the energies involved in such configurations are @xmath19 and this , for a macroscopic black hole where @xmath20 , is ridiculously large .
since gravitational interaction grows with the product of the energies of the interacting entities it is seen that the modes which contribute to hawking radiation interact extremely strongly in the regions from whence they come in order to deliver their asymptotic flux .
therefore the picture of free mode propagation breaks down .
unruh @xcite has shown nevertheless that this does not kill the effect .
all that one requires is a condition that had been previously proposed by jacobson @xcite , to wit : at some finite distance from the horizon one still has vacuum conditions .
a likely distance is @xmath21 _ i.e. _ @xmath22 .
unruh showed that if one tinkered with the dispersion relation , @xmath23 , in a manner similar to that of phonons in fluids where @xmath24 cuts off at @xmath25 , then jacobson s condition is met _
i.e. _ hawking radiation is robust .
unruh s analysis was numerical , but it has also been demonstrated analytically @xcite . since then ,
other models have been proposed , one based on a deformed commutator @xcite and another which takes into account scattering between incoming and outgoing modes @xcite .
in these latter cases backward extrapolation of the emitted hawking quantum results in its disappearance once it enters into a region which is localized around the horizon , which one supposes is given on the planckian scale ( at least in @xcite ; in @xcite this specification is more complicated ) . in both cases hawking steady state radiation in asymptotia
is maintained whereupon the picture emerges that the quanta boil off from a horizon region , _
i.e. _ there is some kind of planckian soup which serves as a reservoir of `` cisplanckian '' quanta .
all of this to motivate a two - fluid picture of space - time and the matter in it , wherein the large length scale physics is described by conventional field theory , modes ( and their interactions ) which propagate in a fluid which describes the physics at small length scale .
the scale at which the dichotomy takes place is the cut - off of modes , hence proportional to @xmath26 where @xmath27 is a parameter which describes the density of modes .
we expect @xmath28 . our experience in black hole physics has emboldened us to take this picture seriously .
hawking radiation is , in no small measure , as solid as thermodynamics ( for the eternal black hole this statement is based on the periodic structure of green s functions in imaginary time in schwarzchild coordinates in the region of space - time where this description is appropriate .
it is also a necessary concomitant to the elimination of unphysical singularities in @xmath8 in the region of the horizon ) .
we know that the mode picture must fail once we reach planckian scales owing to mode - mode interactions so that beyond that scale , strong interaction physics take over . and
this we do not how to formulate .
but the one acid test where quantum field theory is confronted with strong effects of gravity , the black hole , has taught us that we can account correctly for a phenomenom which is solidly established , hawking emission , using this cut - off low momentum sector whilst remaining non committal concerning the nature of the planckian soup . and
further from @xcite and @xcite , this soup serves as reservoir for the low momentum modes . what does this have to do about the inflaton and the primordial cosmogenetic germ ?
the answer lies in quantum mechanics .
the collective variable , @xmath27 , appears in our description of the wave function of everything .
therefore it fluctuates .
moreover it fluctuates locally .
it is a field @xmath29 .
our proposition is that at least one piece of the physics that is described phenomenologically by an inflaton field is the fluctuation of @xmath27 .
there may be more to the inflaton than @xmath27 alone , due to the dynamics of matter .
for example supersymmetry and its breaking may have important physical repercussions on how @xmath8 fluctuates .
but , given our present appreciation of gravity in its quantum context , fluctuations of @xmath27 are inevitable . whether this is ultimately determinant , ( i.e. whether the fluctuation of @xmath27 is the germ ) is a quantitative question which requires further research . in the foregoing paragraphs
i shall simply indicate that the fluctuation of @xmath27 has the qualitative attributes that one attributes to the inflaton field .
hopefully more quantitative support ( or negation ) will emerge in subsequent work .
one very important question is that of the cosmological constant @xmath30 .
i am assuming @xmath31 in the initial quiescent universeless state ( at least in the region of space - time that holds our universe ) .
i am therefore waiving the interesting problem posed by the existence of @xmath30 now , which appears to the same order of magnitude of the present energy density @xmath32 .
this seems more like a tuning problem than the fundamental issue we are facing here where by all rights one might have expected @xmath33 .
the answer to why @xmath31 at the beginning may be presumed to be hidden in the variational principle which fixes @xmath34 _ i.e. _ in the quiescent state one has @xmath35 where @xmath36 the mean energy density
. one must find out why @xmath37 .
the question is nigh to impossible to answer .
this is because the energy of space - time in this 2 fluid picture takes the form @xmath38 we only know something about @xmath39 _
e.g. _ the zero point contribution is @xmath40 .
of all the rest we are ignorant .
there are two points of view which one can adopt in setting @xmath31 .
one is the time honored procedure of blind substraction .
this can be couched in somewhat more elegant terms as a sort of aristotelian principle of absolute rest , to wit : gravity does not respond to the configuration of fields which is at the rock bottom of @xmath41 .
it only responds to the fluctuations _
i.e. _ to @xmath42 where @xmath43 is the value of @xmath44 at its minimum .
the alternative possibility is that @xmath31 is a principle , say as fundamental as the principle of equivalence .
this would fix @xmath45 .
the implication is that there is a balancing out of parameters which appear in the various terms of ( [ eq1 ] ) , which arranges things just so .
( using foam models it is easy to make such constructions , but they are on no more nor less fundamental grounds that the more usual _ ad hoc _ subtraction ) .
be that as it may , an expansion about @xmath34 leads to @xmath46 where @xmath47 ; @xmath48 an important point is to realize the difference in consequences of ( [ eq2 ] ) when one compares the cosmological fluid to a stable material fluid where ( [ eq2 ] ) would also apply ( say to a superfluid at zero temperature ) .
the difference lies in the conservation of the elements ( atoms ) which make up the latter and their non conservation in the former .
( for example , in the naive model of foam made up of instantons , in virtue of the very word , these entities come and go ) . at this point
, it is meet to set out an important caveat .
it has been argued in @xcite that @xmath31 in virtue of stability .
this argument is insufficient because it is based on conservation of the number of atoms .
it is worth setting out the argument since it then allows one to appreciate why density fluctuations are massless in the case of material liquids , but massy in the cosmological case .
consider a superfluid at zero temperature .
it is stable _
i.e. _ @xmath49 an observer inside the fluid uses the grand ensemble to describe his physics _
i.e. _ within the volume , @xmath50 , which is within his scope of observation , particles come and go and this is accounted for by making use of the energy in @xmath50 which we will denote by @xmath51 @xmath52 where @xmath53 is the average value of @xmath54 in @xmath50 and @xmath55 = chemical potential ; @xmath56 and at @xmath57 , @xmath58 .
also since @xmath59 in virtue of stability we have @xmath60 , whence @xmath61 the conditions @xmath62 are translated into zero cosmological constant .
the argument is not applicable to the cosmological fluid since @xmath54 is not conserved whence @xmath63 is not a physical variation .
rather @xmath64 at @xmath57
_ i.e. _ as @xmath50 varies , @xmath54 can vary with it to keep the density @xmath27 ( whose average is @xmath65 ) at its optimal value .
in other words , as @xmath50 varies particles can pop in and out of the vacuum so as to keep @xmath65 fixed .
therefore stability does not imply @xmath59 .
the true variational principle in this case is @xmath66 one is generally led to believe that @xmath67 when number is not conserved is a negative statement . if you do nt need it then it is zero . in the present case
, this is also true , but it is a more powerful statement than usual in that ( [ eq6 ] ) fixes the density to be optimal .
it also implies @xmath68 . and
this is a far cry from @xmath69 .
what ( [ eq6 ] ) does imply is that the mean energy of the ground state of the cosmological fluid is delivered in the form of a cosmological constant .
indeed @xmath70 means @xmath71 . in general relativity
this means @xmath72 where in this case @xmath73 is flat .
but , to repeat , one must further to prove @xmath74 , and this is a great problem which continues to be vexing .
let us now continue with ( [ eq2 ] ) and show that it invites the proposition that @xmath75 is a massy scalar field which describes density fluctuations in the cosmological fluid , hence having the property that one associates with the inflaton .
the physics , once more , is brought out by comparison with the material fluid .
because @xmath54 is conserved in the material fluid density fluctuations have no zero fourier component .
thus there must be no physical manifestation of the @xmath76 component of @xmath77 , whence @xmath78 . if this is the terminal point of a continuous spectrum than the mass of the field @xmath77 is zero .
the way this is realized is to compute the pressure gradient that is induced by a density gradient @xmath79 .
this pressure gradient causes acceleration of the fluid element ( _ i.e. _ temporal variations in the fluid current @xmath80 ) . using the equation of continuity , @xmath81 ,
then straightforwardly one produces the wave equation for sound with @xmath82 .
what is important is that all the energy of a fluctuation is in the spatial gradients plus the acceleration that these induce in the form of kinetic energy _
i.e. _ @xmath83 for small fluctuations where @xmath84 is the velocity of sound , @xmath85 . now consider the cosmological fluid . here
the @xmath86 component does exist .
thus one can have a fluctuation profile which looks like figure [ fig2 ] .
this is the form one is talking about in inflation and indeed it is how one thinks about the germ .
and there is no reason in the world why such fluctuations of @xmath27 should not exist .
in addition to the usual energy lodged in gradient terms as in regions @xmath87 , there is present a flat region @xmath88 where the gradient vanishes or is small
. nevertheless its energy exists .
it is pulled out of the quiescent configuration because absolute quiescence is not a quantum concept ; @xmath89 \neq 0 $ ] can come about by changing the density of modes and or soup in a region without recourse to @xmath90 in that region .
thus the term @xmath91 can and does contribute to @xmath92 where @xmath93 and @xmath94 are a scalar field ( the inflaton ) and its mass
. of course gradients will exist in any finite size fluctuation and they will carry energy . on grounds of covariance
their action will be proportional to @xmath95 . to calculate the relative weights of these kinetic terms as compared to the potential term
is a dynamical problem which must be addressed . to assert that the inflaton is @xmath75 is premature .
various models must be tested to see if the parameters one comes by are acceptable .
but it is fairly clear that whatever the inflaton turns out to be , the fluctuation of the fundamental `` stuff '' which we now contemplate as the point of departure for the formulation of quantum gravity must play some rle .
it may be one element of a highly complex process involving any one of the number of candidates presently considered at the foundation of matter and gravity , or for all we know it may be the whole story .
we close this rather unsatisfactory speculative discourse with some additional remarks .
\1 ) since @xmath27 is proportional to the density of modes ( essentially @xmath96 is the cut - off ) , fluctuations of @xmath27 will cause not only fluctuations of mode density .
some of these fluctuations will end up as those which one calls upon to account for present fluctuations of the cmbr and structure .
another part will cause the production of quanta due to the nonadiabatic variation @xmath97 . since @xmath75 regresses , one obtains particle production .
this is in analogy to production of phonons induced by the finite rate of change of density in a solid .
once more details are required to see if this is the source of the entropy we observe .
\2 ) one must not lose sight of alternatives to the inflaton scenario .
hawking has investigated black hole pair fluctuations in vacuum .
it is not impossible that there are rare abundant coherent fluctuations which will be seized upon by the scale factor and which will inflate just as the fluctuation of @xmath93 inflates .
this is but one example . to be sure the scenario of chaotic inflation has to some extent `` won its spurs '' .
but fundamental cosmology is essentially speculative in character .
any argument that works is an argument of sufficiency , not of necessity .
so i conclude , it is too easy to be complacent .
we are probably going to meet no end of surprises before we can reach some consensus of opinion on just what it is that is responsible for the birth of our universe . 99 e.p .
tryon , personal communication , 1976 r. brout , f. englert and e. gunzig , gen .
rel .
grav .
* 10 * ( 1979 ) 1 .
a. casher and f. englert , phys .
lett .
b * 104 * , 117 ( 1981 ) .
r. kolb and m. turner , _ the early universe _ , addison - wesley , 1990 a. a. starobinsky , phys .
lett .
b * 91 * ( 1980 ) 99 .
a. h. guth , phys .
rev .
d * 23 * , 347 ( 1981 ) .
a. albrecht and p. j. steinhardt , phys .
rev .
lett . * 48 * , 1220 ( 1982 ) .
a. d. linde , phys .
lett .
b * 108 * , 389 ( 1982 ) .
a. d. linde , _ chur , switzerland : harwood ( 1990 ) 362 p. ( contemporary concepts in physics , 5)_. g. e. volovik , phys .
rept . * 351 * , 195 ( 2001 ) [ arxiv : gr - qc/0005091 ] .
s. w. hawking , commun .
math .
phys . * 43 * , 199 ( 1975 ) .
w. g. unruh , arxiv : gr - qc/9409008 .
t. jacobson , phys .
rev .
d * 48 * , 728 ( 1993 ) [ arxiv : hep - th/9303103 ] .
r. brout , s. massar , r. parentani and p. spindel , phys .
rev .
d * 52 * , 4559 ( 1995 ) [ arxiv : hep - th/9506121 ] .
s. corley and t. jacobson , phys .
rev .
d * 54 * , 1568 ( 1996 ) [ arxiv : hep - th/9601073 ] .
r. brout , c. gabriel , m. lubo and p. spindel , phys .
rev .
d * 59 * , 044005 ( 1999 ) [ arxiv : hep - th/9807063 ] . c. barrabes , v. frolov and r. parentani , phys .
rev .
d * 59 * , 124010 ( 1999 ) [ arxiv : gr - qc/9812076 ] ; _ ibid _ d * 62 * , 044020 ( 2000 ) [ arxiv : gr - qc/0001102 ] . |
the critical percolation properties of the two models were studied using direct monte carlo simulations in the spirit of dhar and barma @xcite .
we studied the change in behaviour of the probability @xmath54 , of appearance of clusters of size greater than @xmath43 , with bond probabilities @xmath12 and looked for characteristic scaling behaviour expected in the critical region to identify percolation thresholds plotted . below the probability threshold , @xmath55 for large @xmath43 while @xmath56 in the supercritical phase @xcite .
the critical region is characterized by scaling laws with @xmath57 described by a power law dependence at the critical value @xmath26 , which we localized by sweeping through super and sub critical probabilities using an interval bisection method . for model
a , cluster size distribution data was obtained by performing @xmath58 realizations ( per value of probability @xmath12 ) of cluster growth starting from a single node .
model @xmath59 simulations were performed on a fixed @xmath60 square lattice also with @xmath58 realizations for each @xmath12 , where cluster connectivity was identified using a breadth first search algorithm .
the qualitative values of the critical exponents @xmath61 presented in the paper for model b are defined as follows : @xmath62 and @xmath63 a simple method used to obtain these was to directly calculate the slope of the plots of these two functions in log - log scale . since the problem contains two parameters @xmath64 , we chose one of them @xmath65 do be the independent parameter with @xmath66 determined so as to be on the critical line of the model . for the determination of @xmath27 , we considered clusters of size @xmath43 greater than @xmath67 to be `` infinite '' or system spanning clusters on the finite lattice .
alternatively , we also determined @xmath27 from @xmath41 and an auxiliary exponent @xmath68 , which describes the critical behaviour of mean cluster size @xmath69 also readily available from the simulation : @xmath70 the following universal equation , derived from scaling relations , is known to hold for directed ( uncorrelated ) bond percolation @xcite and isotropic percolation @xcite : @xmath71 we assumed the equation to be true for the entire critical plane and obtained results for @xmath27 in agreement with those obtained using the first method .
results obtained in the latter manner are those presented in the paper . | entanglement indcued non additivity of classical communication capacity in networks consisting of quantum channels is considered .
communication lattices consisiting of butterfly - type entanglement breaking channels augmented , with some probability , by identity channels are analyzed .
the capacity superadditivity in the network is manifested in directed correlated bond percolation which we consider in two flavours : simply directed and randomly oriented .
the obtained percolation properties show that high capacity information transfer sets in much faster in the regime of superadditive communication capacity than otherwise possible . as a byproduct , this sheds light on a new type of entanglement based quantum capacity percolation phenomenon . _
introduction .
_ percolation ( see _ e.g. _ @xcite ) is a natural concept that emerges in the description of spreading processes in the presence of medium imperfections .
percolation effects in quantum networks have recently been the subject of increasing interest @xcite .
most of the attention has been restricted to the generation of large scale networks with maximally entangled states between elementary nodes to allow for quantum communication applications , starting from initial imperfect , _ i.e. _ non - maximally entangled state networks .
the interesting central new insight introduced in @xcite is that local quantum operations may be used not only to purify entanglement but , simultaneously , to change the topology of the lattice to a new one with lower percolation probability threshold .
this idea has been developed for different states @xcite and lattice dimensions @xcite . in a different context
, percolation concepts also appeared in quantum information theory ( qit ) in the study of cluster state generation @xcite .
the capacity of a network determines its utility in the domain of communication .
the development of qit has led to the uncovering of interesting quantum effects on channel capacities , _
e.g. _ superadditivity of quantum ( q - type ) channel capacity @xcite .
this result followed the intuition developed in the bound entanglement activation effect @xcite ( where two weak resources activate each other becoming collectively useful for some task ) continued further in refs . @xcite and @xcite .
independently , the first superadditivity effect of classical ( c - type ) capacity in quantum multi - access channels has been described @xcite .
both q - type and c - type superadditivities have been proven even for entanglement breaking channels @xcite . in this paper
, we consider percolation effects in quantum networks from a channel perspective .
in particular , we show that channel superadditivity can be used to enhance percolation of information through networks .
the schemes are based on network models of classical information transfer through quantum multipartite channels ( macs ) , which can be mapped to certain types of _ directed _ bond percolation problems @xcite .
we first consider a simple layered communication scheme ( a ) to demonstrate the basic idea of percolation assisted by superadditive capacities and then describe a more complicated scheme of multidirectional communication ( b ) .
interestingly , the percolation problem in case b does not seem to have been studied elsewhere in the literature .
the basic ingredients of the quantum networks are : a _ passive _ fixed underlying network built up of elementary entanglement breaking macs , and an _ active _ auxiliary incomplete network consisting of randomly generated ( open ) bonds .
the term passive means that no bond in the network allows _ a - priori _ high capacity communication ( hcc ) , whereas elements of the active network are high capacity channels .
importantly , the active channels can serve to activate hcc through the passive channels .
model a - layered communication network : a ) passive network ; users are located at nodes of a square lattice , the lattice is filled with butterfly shaped primitives , information flow is directed from layer @xmath0 ; b ) butterfly primitive consists of two macs ( solid and dashed wedges ) from ref .
@xcite ; node ( user ) @xmath1 can communicate with @xmath2 with maximal capacity @xmath3 ; c ) active network filled randomly by triples of ideal channels allowing hcc ( no hcc from a to b ) ; d ) using entanglement changes the geometry of the hcc network leading to directed communication paths ( black arrows ) with capacities @xmath4 from a to b. ] ] _ model a - layered network communication .
_ the elementary channel of the passive network is chosen in this paper as a 2-sender 1-receiver _ noisy _ quantum mac depicted as the wedge shaped channels in fig .
[ fig : model_a](b ) : the slanted line is a two - qudit ( @xmath5dimensional ) input pertaining to one user ; the vertical line is a single qudit input of the second user .
the single qudit system is modified by one of a chosen set of ( orthogonal ) @xmath5 unitary operations fired by the logical value of the two - qudit system ( see ref .
@xcite for details ) .
this single qudit line is further modified by a depolarizing channel and is the sole output system at the receiver s end . for moderate to large depolarization the working regime considered here this channel has poor classical capacity @xmath6 for any user @xcite : @xmath7 where @xmath8 is the maximal attainable capacity for a single qudit output , corresponding to an identity channel .
communication can be improved by adding a high capacity active channel , _
e.g. _ an ideal channel , along the vertical transmission line .
inputting a two - party maximally entangled state to the ideal channel and vertical line of a wedge channel increases the capacity along the slanted line of the wedge channel to a much higher value @xmath9 ( see eq.([rpab ] ) ) a manifestation of superadditivity of channel capacities @xcite ( for high depolarization when the channel becomes entanglement breaking , @xmath10 @xcite ) . we define hcc as transmission at a rate @xmath11 .
the structure of the wedge channel induces a sense of direction of communication . in the network context
, we first consider the scenario where only forward communication is allowed ( see fig.[fig : model_a ] ) .
the passive channel inputs are shared by nearest neighbour pairs of sites in horizontal layers giving rise to a butterfly shaped fixed network .
the active channels are only placed on vertical bonds .
suppose that these are available with probability @xmath12 one can assume that identity channels are initially available at all vertical bonds , yet due to fragility w.r.t .
noise either remain useful ( ideal ) with probability @xmath12 or become unuseful random operations with probability @xmath13 effectively erasing information . in our scheme
, we consider a setup where a triple of identity channels is the basic active channel @xcite this allows the possibility of simultaneous hcc between a given sender and his vertically placed receiver as well as between the sender s two horizontal nearest neighbours and that receiver ( fig .
[ fig : model_a ] ) .
the question of establishing long range hcc in such a network is a directed percolation problem .
one asks , under what conditions , is it possible for any user to be able to perform directed hcc , through intermediate nodes , with a user or users located at a distance scaling with the length of the network .
one may compare two scenarios - ( a ) entanglement - free or classical , where no entanglement is allowed in the protocol used at any node and ( b ) entanglement - assisted ( ea ) , which takes full advantage of the superadditive effect described earlier .
the former case corresponds to communication only along 1dimensional paths and the percolation threshold probability is @xmath14 , rendering the network useless for hcc for finite loss probability of active channels .
the ea scheme involves changing the geometry of the hcc network and is mapped to a _ correlated _ directed bond percolation problem where with probability @xmath12 three directed bonds ( forming an arrow shape ) are placed on the lattice ( fig.[fig : model_a](e ) ) .
the threshold probability is significantly suppressed due to entanglement induced increased connectivity .
we performed a standard monte carlo simulation ( see @xcite ) and found the percolation threshold to be @xmath15 with accuracy @xmath16 .
we have checked that the studied percolation transition lies , as expected , in the directed percolation ( dp ) universality class , by computing a complete set of critical exponents and found them to be in agreement with those obtained for the uncorrelated directed bond percolation problem .
the important basic characteristic of directed percolation is that the connected clusters of nodes are geometrically highly asymmetrical and are restricted to acute cones ( with cone angle @xmath17 when @xmath18 ) with axes along the vertical lines passing through the starting nodes . near the critical point the clusters are very narrow and essentially quasi 1-dimensional , as the probability of obtaining a connection with a site a large distance away from the source and at an angle @xmath19 from the axis , @xmath20 for @xmath21 ( see _ e.g. _ dhar and barma @xcite ) .
the ea scheme beats the classical scheme in that there is a wide window in @xmath12 for hcc , and that the horizontal extent of connections @xmath22 at a distance @xmath23 away changes from @xmath24 to @xmath25 as @xmath12 increases from @xmath26 to 1 . for percolation process on randomly oriented square lattice ( classical scheme ) and butterfly network ( ea scheme ) .
dots are numerically obtained points . ] _ model b randomly oriented communication .
_ we now move to a general scenario within the described framework of channels ( fig .
[ fig : rys_3_model_b+activation ] ) .
now each node can communicate in _ both _ directions with its nearest neighbours ( main bonds ) on a square lattice as well as all of its next nearest neighbours ( diagonal bonds ) through the passive network .
the configuration leading to this possibility is shown in fig.[fig : rys_3_model_b+activation ] ( a ) and ( b ) .
the active network channels are placed along the main bonds of the square lattice and are tagged for use in a particular direction . ,
@xmath27 for classical and ea schemes of multidirectional oriented percolation . ]
hcc between distant nodes is now a multidirectional bond percolation problem .
we consider the case where upward , left - to - right oriented identities are present with probability @xmath28 while downward , right - to - left oriented identities are placed with probability @xmath29 .
note , that bidirectional bonds appear as the independent choice of two oppositely directed channels on a given bond .
this setup is interesting already for the case when , say , @xmath30 .
using an entanglement free protocol , capacity @xmath31 can already be obtained through the main bonds of the square lattice along the up or right directions , in contrast to model a. this is the well known square lattice directed bond percolation problem for which the percolation threshold @xmath32 ( see _ e.g. _ @xcite ) .
but , in the ea scheme , due to increased and correlated connectivity , the percolation threshold is almost halved w.r.t .
the classical scheme and has been calculated here to be @xmath33 .
both phase transitions lie in the dp universality class . at moderate @xmath34 , the probability of a starting node belonging to an infinite or system spanning anisotropic cluster ( in the thermodynamic limit ) @xmath35 using superadditivity effects while @xmath36 otherwise .
the general multidirectional communication problem is the most interesting setup . in the classical scenario ,
this reduces to square lattice randomly oriented percolation .
problems of this category were studied in the context of random resistor diode networks ( see _ e.g. _ @xcite ) mainly using renormalization group calculations .
however not many results seem to be available in the literature on the @xmath37 percolation problem ( see however @xcite for some analytical properties ) . here , we map out the phase diagram in the @xmath38 plane in fig.[fig : perctresmodelb-3.0 ] using monte carlo simulations ( see appendix ) .
note that the diagram is symmetrical w.r.t .
the line @xmath39 due to system symmetry under the interchange @xmath40 .
this result is to be compared with the multidirectional correlated bond percolation phase diagram for the ea scheme ( fig .
[ fig : rys_3_model_b+activation](d ) ) . as can be seen in fig.[fig : perctresmodelb-3.0 ] , the ea scheme is drastically better than the classical scheme in the whole parameter plane .
an interesting feature of both schemes is that they allow switching of universality classes of phase transitions from the dp class to the isotropic percolation ( ip ) class to which standard bond / site percolation belongs .
this is facilitated by the choice of parameters @xmath37 , and as a result one may accordingly change the properties of the long range clusters . to provide evidence for this ,
we calculate two universal critical exponents ( fig . [ fig4 ] ) @xmath27 and the fisher exponent @xmath41 ( see @xcite ) as one moves along the critical lines ( fig .
[ fig : perctresmodelb-3.0 ] ) of the two models .
recall , that @xmath27 determines how the percolation probability @xmath42 increases above the percolation threshold , while the fisher exponent determines how the probability of obtaining a cluster larger than size @xmath43 decreases with @xmath43 at the critical point @xmath44 : @xmath45 .
note that the values of these exponents in the dp class are @xmath46 @xcite and in ip : @xmath47 @xcite .
first consider square lattice multidirectional percolation in fig.[fig4 ] ( the classical scheme ) .
both exponents show that as one moves along the critical line , the phase transitions lie in the dp universality class until @xmath48 , since this region inherits the exponents at the dp point @xmath49 . at @xmath50 , which is an isotropic symmetry point ,
one obtains values of the exponents corresponding to the ip universality class . in between
, there is a characteristic crossover region between the two types of behaviour . in particular
, this means that for a choice of parameters approaching the isotropic point , one obtains different characteristic growth of clusters ( determined by different critical exponents ) than when in the dp region . secondly the cluster geometrical characteristics change from highly anisotropic to isotropic .
the butterfly network percolation problem ( ea scheme ) basically follows the same pattern ( see fig .
[ fig4 ] ) and can be considered a rescaled version of the classical problem , wherein again lies the quantum advantage of the ea scheme . for this network ,
the isotropy point is found to be located at @xmath51 .
the slightly lower values of @xmath27 in the dp regime as compared to the square lattice problem do not seem to be significant - and are a result of the extreme sensitivity of the calculated exponents on the accuracy of critical probabilities . _ concluding remarks .
_ we have described percolation effects in a channel context showing how directed percolation effects ( not considered before ) emerge in the consideration of quantum networks . in the context of percolation theory , to our knowledge , the multidirectional correlated bond percolation problem has not been studied before .
finally , quite remarkably , our results provide a new entanglement percolation effect in the spirit of ref .
@xcite . keeping everything else unchanged ,
consider the bell measurement ( bm ) channel of fig . 1 of ref .
@xcite instead of the mac used here , with ac ( bc ) playing the role of diagonal ( resp .
vertical ) bond of the square lattice .
then any randomly generated singlet between b and c can be switched , via entanglement swapping , to a singlet on the diagonal ac .
this leads to the same geometry as discussed here but now one asks about the possibility of building a long range network of singlets .
note that singlets are directionless . without using the bm channel
, we obtain a `` classical '' scheme related to square lattice percolation which has threshold @xmath52 .
since directionless percolation must certainly be at least as good as directed percolation , an ea scheme making use of the switching mechanism of the bm channel will have threshold at most equal to the calculated threshold at the isotropic point of model b @xmath53 .
the authors thank j.k .
korbicz for discussions .
p. h. thanks also c. h. bennett and a. grudka for discussions .
the work was supported by eu project qessence and by the polish ministry of science and higher education through grant no .
nn202231937 .
part of the work was done in quantum information centre of gdansk . 40 d. stauffer and a. aharony , introduction to percolation theory , taylor & francis 2003 .
g. grimmett , percolation , springer - verlag 1999 .
a. acin , j. i. cirac , and m. lewenstein , nature phys .
* 3 * , 256 ( 2007 ) .
g. j. lapeyre , jr . , j. wehr , and m. lewenstein , phys .
rev .
a * 79 * , 042324 ( 2009 ) .
s. perseguers , l. jiang , n. schuch , f. verstraete , m. lukin , j. cirac , and k. vollbrecht , phys .
rev .
a * 78 * , 062324 ( 2008 ) ; s. perseguers , d. cavalcanti , g. j. lapeyre , jr . , m. lewenstein , and a. acin , phys
. rev . * a 81 * , 032327 ( 2010 ) .
s. broadfoot , u. dorner , and d. jaksch , europhys .
lett .
88 , 50002 ( 2009 ) .
s. perseguers , phys .
rev .
a 81 , 012310 ( 2010 ) .
a. grudka _
et al . _ ,
manuscript in preparation ( 2011 ) .
k. kieling , j. eisert , , , percolation in quantum computation and communication , in : _ quantum and semi - classical percolation and breakdown in disordered solids _ , pages 287 - 319 ( springer , berlin , 2009 ) .
g. smith , j. yard , science * 321 * , 1812 - 1815 ( 2008 ) .
p. horodecki , m. horodecki , r. horodecki , phys .
rev . lett . * 82 * , 1056 ( 1999 ) .
p. shor , j. a. smolin , and a. v. thapliyal , phys .
rev .
lett . * 90 * , 107901 ( 2003 ) .
w. duer , j. i. cirac , and p. horodecki , phys .
rev .
lett . * 93 * , 020503 ( 2004 ) .
l. czekaj and p. horodecki , phys .
rev .
lett . * 102 * , 110505 ( 2009 ) .
a. grudka and p. horodecki , phys .
rev .
a * 81 * , 060305(r ) ( 2010 ) .
s. r. broadbent and j. m. hammersley , proc .
camb .
. soc . * 53 * , 629 ( 1957 ) .
d. dhar and m. barma , j. phys .
c : solid state phys . , * 14 * , l1 ( 1981 ) .
h. hinrichsen , adv.phys .
* 49 * , 815 ( 2000 ) .
s. redner , j. phys . * a * : math
. gen .
* 14 * , l349 ( 1981 ) ; _ ibid _ * 15 * , l685 ( 1982 ) ; phys . rev .
* b 25 * , 3242 ( 1982 ) .
geoffrey r. grimmett , random structures & algorithms * 18 * 257 ( 2001 ) this is a superadditive effect as the identity is unavailabe to the user of the slanted input line of the wedge channel . for qubit channels , _
e.g. _ this can be physically realized as a process of ideal transmission with probability @xmath12 of a single photon having three frequency states and two polarization degrees of freedom for each frequency , where the dominating noise process is photon loss .
see supplemental material for methods .
* directed percolation effects emerging from superadditivity of quantum networks * ukasz czekaj , ravindra w. chhajlany and pawe horodecki |
a small sample of neptune trojans has been accumulated by a variety of surveys ; however , inferences drawn from this sample about the intrinsic distributions of neptune trojan orbital properties have been limited and generally qualitative . the challenge inherent in extracting meaningful information from this sample is accurately determining the properties of the surveys that discovered them , and the properties of all surveys that , while sensitive to neptune trojans , did not discover any . in this work
, i treat these unknown survey properties as nuisance parameters , and marginalize over them to extract as much useful information about the intrinsic orbital distributions of the neptune trojans as possible . of particular dynamical interest
are the inclination , eccentricity , and libration amplitude distributions .
these distributions encode information about the formation mechanism ( in - situ formation , chaotic capture , or other processes ) and post - formation evolution .
several neptune trojans have remarkably high inclinations ( @xmath7 ) , even though surveys have by - and - large targeted fields near the ecliptic where objects on inclined orbits spend relatively little time .
previous works have noted that this qualitatively indicates the existence of a large , poorly - sampled high - inclination neptune trojan population ( sheppard & trujillo 2006 , sheppard & trujillo 2010a ) . in this work ,
i simultaneously consider the inclination , eccentricity , and libration amplitude distributions , generate synthetic populations of neptune trojans defined by these distributions , then pass these synthetic populations through `` coverage functions : '' simplified observational filters that are treated as independent functions of heliocentric ecliptic latitude @xmath8 , and heliocentric longitudinal separation from the trojan libration centers @xmath9 ( libration centers located roughly @xmath10 from neptune ) , and inclination .
the properties of these observational coverage functions are then marginalized over , effectively marginalizing the unknown properties of the surveys which discovered the trojans . like the jupiter trojans and other trans - neptunian populations ,
the inclination distribution is modeled as a brown s distribution ( @xmath11 ) .
the libration amplitude @xmath12 and eccentricity @xmath13 distributions are both modeled as rayleigh distributions , motivated by the distributions of the jupiter trojans .
the inclination , libration amplitude , and eccentricity distributions are all truncated at upper limits derived from stability constraints , requiring appropriate corrections to their proposal volume and probability density functions .
all statistical analysis is performed in the conceptually simple yet analytically powerful `` approximate bayesian computation '' framework , described in section [ abc ] .
l r r r r + name & @xmath14 & @xmath15 & @xmath16 & @xmath17@xmath18 + 2001 qr@xmath19 & 1.3@xmath20 & 25.5@xmath21@xmath20 & 0.57@xmath20 & 10.46@xmath20 + 2004 up@xmath22 & 1.4@xmath20 & 10.8@xmath23@xmath20 & 0.73@xmath20 & 10.24@xmath20 + 2005 tn@xmath24 & 25.0@xmath20 & 8.7@xmath25@xmath20 & 0.62@xmath20 & 8.51@xmath20 + 2005 to@xmath26 & 5.2@xmath20 & 9.2@xmath27@xmath20 & 1.62@xmath20 & 9.12@xmath20 + 2006 rj@xmath28 & 8.2@xmath20 & 6.3@xmath29@xmath20 & 7.99@xmath20 & 0.58@xmath20 + 2007 vl@xmath30 & 28.1@xmath20 & 14.2@xmath31@xmath20 & 11.25@xmath20 & 9.44@xmath20 + 2008 lc@xmath32 & 27.6@xmath20 & 16.4@xmath33@xmath20 & 2.80@xmath20 & 0.84@xmath20 + 2011 hm@xmath34 & 29.4@xmath20 & 9.8@xmath35@xmath20 & 2.60@xmath20 & 7.72@xmath20 + + @xmath36 : j2000 ecliptic inclination .
@xmath37 : half - peak rms libration amplitude and @xmath38 uncertainty .
@xmath18 : absolute value of j2000 heliocentric ecliptic longitude separation of object and nominal trojan center ; see eqn .
[ lprime ] .
@xmath39 : j2000 heliocentric ecliptic latitude .
the minor planet center ( mpc ) lists nine neptune trojans ( six l4 , three l5 ) , but one of the l5 trojans is unstable and likely a recently captured centaur ( gladman et al .
2012 , horner et al .
this object is therefore not considered to be reflective of the intrinsic inclination distribution of the ( putatively primordial ) neptune trojans . with the addition of the newly discovered l5 trojan 2011 hm@xmath34 ( parker et al .
2013 ) , the 8 known long - term stable trojans have ecliptic inclinations ranging from @xmath40 to @xmath41 , and heliocentric ecliptic latitudes at discovery ranging in amplitude from @xmath42 to @xmath43 .
figure 1 illustrates these properties , and it is clear that trojans have generally been discovered at latitudes significantly lower than their inclinations , even though an object spends roughly 50% of their time at latitudes greater than 70% of their inclination .
only the object 2006 rj@xmath28 was higher than its median latitude at the time of discovery .
this indicates that it is likely that most surveys that discovered neptune trojans targeted the ecliptic , and were therefore strongly biased toward detecting low inclination objects , and yet discovered a surplus of high - inclination objects .
the larger number of known l4 neptune trojans compared to l5 is likely an artifact of the l5 cloud being more poorly surveyed due to its current proximity to the galactic plane .
these neptune trojans were discovered by a variety of surveys , performed at a variety of facilities and under varying conditions , and normally would not represent a sample from which estimating an intrinsic , debiased orbit distribution would be statistically advisable .
however , using appropriate statistical care , we can make a conservative estimate of the range of plausible properties of the neptune trojan orbital distribution by marginalizing over the plausible volume of the unknown characteristics of all discovery surveys .
this _ survey - agnostic _ approach can conceivably be applied to other populations , and since it is performed in a bayesian framework , the outcomes can be meaningfully combined with results from large , monolithic , well - characterized surveys such as des ( eg . , gulbis et al .
2010 ) , cfeps ( eg . , petit et al . 2011 ) and the ongoing _ _ outer solar system origins survey__. because of their small sample size and currently poorly - characterized orbit distributions , i consider the neptune trojans a useful demonstration population .
the libration amplitudes listed in table 1 were generated with the same technique as parker et al .
each object s motion was integrated with _
mercury6 _ ( chambers 1999 ) in the presence of the giant planets for 1 myr .
100 clones of each object were integrated , with initial state vectors centered on the jpl horizons solution , perturbed to populate the cartesian uncertainty manifold generated by fitting all ground - based observations of each object with the _
fit_radec _ and _ abg_to_xyz _ routines developed in association with bernstein & kushalani ( 2000 ) .
libration amplitudes for each clone were measured by assuming that libration is sinusoidal and deriving the sinusoidal half - amplitude from the rms of the @xmath44 samples of the resonant angle over the entire 1 myr integration : @xmath45 which produces the more appropriate half - amplitude for scaling a sinusoidal model than the usual peak - to - peak definition of @xmath46 $ ] .
the rms produces a value that better reflects the mean libration behavior , while defining the amplitude from peak to peak is sensitive to large , single - cycle excursions of the resonant argument . as such , the rms - defined amplitude is always smaller than the peak - to - peak definition .
for all parameter estimation in this work i utilize a likelihood - free rejection sampler specifically , the `` approximate bayesian computation '' rejection ( abcr ) scheme first presented in pritchard et al .
approximate bayesian computation ( abc ) is conceptually simple but statistically powerful , and has the primary advantage of not requiring the computation of any true likelihood value .
i briefly outline this approach below , and refine its description later as merited by the specifics of each application .
the literature on abc is well - developed , with much more sophisticated methods available than those employed here ; for a recent review see marin et al .
( 2011 ) . in the context of this work ,
i utilize abcr in order to approximate the posterior probability density functions of a set of parameters @xmath47 which define the properties of a model @xmath48 that describes an observed population of minor planets . given a metric of similarity @xmath49 that reflects the relative similarity of some set of properties of an observed sample @xmath50 and a synthetic sample @xmath51 , where the synthetic set is generated by a model @xmath52 , and where the parameters @xmath47 of @xmath48 are drawn from prior distribution @xmath53 : 1 .
propose a new set of model parameters @xmath54 , drawing from the prior distribution @xmath55 for each parameter .
2 . generate a synthetic sample of observations @xmath51 from model @xmath48 with parameters @xmath54 .
3 . generate the similarity metric @xmath56 for this synthetic sample compared to the observed sample .
4 . if similarity metric @xmath56 is less than some cutoff value @xmath57 , keep @xmath54 as a successful trial .
repeat steps 14 until sufficient number of trials have been successful .
the distribution of parameters of the retained @xmath54 sample from successful trials approximates the posterior distributions of those parameters in the bayesian framework , given @xmath57 is sufficiently small and the number of trials is sufficiently large . in practice , i set @xmath57 dynamically by sampling many times and saving each sampled @xmath54 , then determining a value for @xmath57 which will retain less than some fraction of the total @xmath54 sample ( in this work , this acceptance rate was set between 0.1%0.01% , depending on the number of dimensions of the parameter space under consideration ) .
the similarity metric @xmath56 is usually defined by some distance on a set of summary statistics of the properties of @xmath50 and @xmath51 , but implementations vary on a case - by - case basis .
the use of summary statistics to define a metric of similarity is computationally advantageous for large datasets where other methods might be computationally prohibitive .
if the summary statistics adopted are sufficient for the underlying distributions , then there is no information lost by using them . in this work
, however , the data set is very small , and in the absence of a clear choice of sufficient summary statistics i instead utilize the two - sample anderson - darling test to determine a similarity metric ( the two - sample ad - test @xmath58 statistic ) . for multiple dimensions
, the adopted metric @xmath56 is the sum of all one - dimensional , two - sample @xmath58 statistics over all dimensions , divided by the number of dimensions tested .
qualitatively , this @xmath56 is small when all distributions are similar to each other , and large when one or more of the distributions are discrepant . in this implementation ,
the critical values of @xmath58 or @xmath56 and sample - size effects are accounted for as the abcr analysis effectively bootstraps these out .
the adoption of a metric based on a collection of 1d 2-sample ad - tests is motivated by its use by the cfeps survey ( eg . , petit et al .
2011 ) , but by utilizing the abcr framework instead of simply adopting the lowest @xmath59-value inferred from each independent ad - test over all dimensions , this work reduces the risk of erroneously identifying a low @xmath59-value in any given application of the metric .
the only other large population of trojan objects in the solar system belongs to jupiter . while evidently less populous than neptune s trojan swarms ( sheppard & trujillo 2006 ) , the jupiter trojans are much better sampled , and are nearly complete to @xmath60 . to motivate the functional forms of inclination , libration amplitude , and eccentricity distributions
i will adopt for the neptune trojans , i consider those of the jupiter trojans as templates . to date
, no population - wide orbit distribution models have been computed for the jupiter trojans . here
i consider only the relatively complete @xmath60 sample , and model the inclination , libration amplitude , and eccentricity distributions .
i find that the inclination distribution is very well - modeled by a truncated brown s distribution ( brown 2001 ) : @xmath61 while both the libration amplitudes are well - modeled by truncated rayleigh distributions defined by widths @xmath62 and @xmath63 and truncations @xmath64 and @xmath65 . for the jupiter
trojans i hold the truncations of each distribution fixed at the maximum observed value of each parameter ; @xmath66 , @xmath67 , and @xmath68 . in the later neptune trojan analysis ,
these thresholds are defined by dynamical constraints . for consistency , i perform parameter estimation in the same abcr framework that i later apply to the neptune trojans , though i only consider one distribution at a time . in the abcr analysis ,
the parameters of interest ( @xmath69 , @xmath62 , and @xmath63 ) are sampled uniformly in their distributions means .
that is to say , for each trial , @xmath70 is selected uniformly , and then mapped to @xmath69 given that the model inclination distribution is a brown s distribution . because the nominal functional form for the inclination , libration amplitude , and eccentricity distributions are truncated at some upper limit , their means are not trivial to compute .
for example , the mean of a non - truncated rayleigh distribution is simply given by @xmath71 , but for a truncated rayleigh the mean is given by @xmath72 where @xmath73 is the truncation value ( see appendix [ raymeanappendix ] for a derivation and more details ) .
the functional form for a truncated brown s distribution is similar ( see appendix [ brownsmeanappendix ] for a derivation and more details ) .
proposing in the distribution mean has several advantages ; chiefly , it translates an infinite prior volume ( @xmath74 ) to a finite one , as the distribution mean is limited to a maximum value by the truncation of the distribution . for a truncated rayleigh distribution ,
the mean asymptotically approaches @xmath75 as @xmath76 .
a truncated brown s distribution has similar asymptotic behavior ( see appendices [ raymeanappendix ] & [ brownsmeanappendix ] for details ) . for the jupiter trojans ,
the uniform - mean prior is not particularly important because the sample is so large and well - defined .
however , when i later consider the more poorly - sampled neptune trojans , the utility of this prior becomes much more apparent . for each distribution
, @xmath77 abcr trials were run , and the 1,000 best trials were retained to generate the posterior pdf for each parameter .
figure [ jtpdf ] illustrates the @xmath78 sample of observed osculating inclination , libration amplitude , and osculating eccentricity , compared to the accepted model distributions , and also illustrates the posterior pdfs for @xmath69 , @xmath62 , and @xmath63 defined by the accepted 1,000 abcr trials .
the observed jupiter trojan sample is very well modeled by the adopted functional forms for the orbit distribution ; a ks - test indicates peak @xmath59-values of 92% , 70% , and 98% for the inclinations , libration amplitudes , and eccentricities , respectively ; indicating that there is no statistically significant evidence for discrepancy between the model distributions and the observed sample and qualitatively indicating good agreement between the two .
because of the lack of consistent information regarding their discovery surveys , a canada - france ecliptic plane survey ( cfeps)-style survey - simulator approach ( jones et al .
2006 ) is not feasible for the neptune trojans .
instead , i adopt a set of basis functions informed by typical trans - neptunian object survey strategies which represent the coverage of the sky by all surveys which _ did discover _ or _ plausibly could have discovered _ neptune trojans
. the parameters of these `` coverage functions '' are allowed to vary , and the discovery circumstances of each known neptune trojan are used to quantify how well a given choice of parameters reproduces them , given a set of assumed orbit distribution parameters . finally , the properties of these coverage functions are marginalized over , and conservative orbit distribution posterior probability density functions are determined . to model the _ a priori _ unknown survey properties ,
i assume that the probability of detecting a given trojan is essentially independent of its orbital phase save for effects of areal coverage in ecliptic latitude @xmath8 and longitude @xmath79 .
i fold all the effects of survey depth ( which couples with the luminosity function ) and areal coverage into two nuisance `` coverage functions '' @xmath80 and @xmath81 .
these represent the probability of any survey detecting a neptune trojan given the trojan s ecliptic latitude and longitude at the time of discovery .
nominally , these coverage functions will ( 1 ) only depend on ecliptic latitude and longitude , ( 2 ) likely be monotonic away from the centers of the trojan clouds , and ( 3 ) have a functional range bracketed by the interval [ 0,1 ] and a domain over the entire sky .
in addition to modeling survey coverage in ( @xmath82 ) , i also include the possibility of a direct bias against high - inclination objects in the mpc database .
the cfeps survey found moderately higher inclination distribution widths @xmath83 for the 3:2 population than previous studies ; because cfeps carefully characterized their tracking biases in a self - consistent way , they suggested that earlier samples of objects with sufficiently accurate orbits to be confidently characterized as members of the 3:2 population may have had an unacknowledged bias against high inclinations .
that such biases might exist is not surprising , and the possible sources of such biases were explored in jones et al .
( 2006 ) . because the surveys that contributed to this inclination - biased sample were also sensitive to neptune trojans
, it is likely that a similar inclination bias also exists in the observed sample of neptune trojans .
( top , dotted regions ) and a gaussian inclination coverage function ( bottom , dotted regions ) , with parameters selected by procedure in text .
the resulting biased model distributions ( solid regions ) excellently match the observed distribution , indicating that the general form of eqn .
[ latcov ] is a good model for the shape of the amalgamated latitude coverage functions of all previous surveys .
1- and 2-@xmath84 ranges illustrated by dark and light regions , respectively.,scaledwidth=50.0% ] and @xmath85 after abcr analysis , and marginalized posterior pdfs for each parameter . , scaledwidth=50.0% ] .
middle panel : inclination distribution width @xmath86 .
bottom panel : inclination coverage function width @xmath87 .
vertical dashed lines indicate approximate distribution modes ( @xmath88 , @xmath89 , and @xmath90 ) , estimated using a gaussian kernel density estimator .
, scaledwidth=50.0% ] in order to determine the shape of the ecliptic latitude coverage function , i use the well - characterized 3:2 resonant population as a tracer .
it is likely that similar ecliptic latitude and inclination biases apply to the 3:2 population as to the neptune trojan population ; most surveys that discovered 3:2 objects were capable of discovering neptune trojans as well , as both populations share broad inclination distributions ( and therefore both occupy a broad swath of ecliptic latitudes ) , similar on - sky rates of motion , and probably share similar luminosity functions . the intrinsic properties of the 3:2 population have been measured by the carefully calibrated cfeps survey ( petit et al .
2011 , gladman et al .
2012 ) . by considering
the _ observed _ sample of ( non - kozai ) 3:2 population members ( provided by buie , private communication ) , and adopting the intrinsic inclination distribution measured by cfeps , it is possible to derive the longitudinally - averaged ecliptic coverage function that applied to the amalgam of all surveys that reported 3:2 discoveries to the minor planet center .
i begin with an ansatz of an appropriate functional form for the ecliptic latitude coverage function @xmath80 : @xmath91 this form has substantial flexibility provided by three free parameters , but the 3:2 observations constrain them to be strongly correlated . in general , its behavior for small @xmath8 is driven by index @xmath92 and half - width @xmath93 , while for large @xmath8 the function flattens to a constant probability floor set by @xmath94 .
the function is truncated above @xmath95 , where @xmath96 is the highest ecliptic latitude ( at discovery ) of any observed 3:2 object , since beyond this point the population provides little constraint on the properties of the coverage function .
i then consider the functional form of a possible inclination bias . in this case
, there is useful information in literature to inform our functional ansatz : namely , that cfeps and other surveys all found that the _ form _ of the 3:2 population s inclination distribution was well - modeled by a brown s distribution @xmath97 but they found significantly different widths . in that case
, i determine that the plausible form of the inclination bias is a gaussian : @xmath98 so @xmath99 where @xmath100 .
to map the best - fit cfeps inclination width @xmath101 to the best - fit des inclination width @xmath102 ( gulbis et al .
2010 ) , a gaussian inclination coverage function requires a width of @xmath103 .
gladman et al . ( 2012 ) notes that of 24 3:2 objects detected by cfeps , four had inclinations over @xmath104 , while no des 3:2 objects had inclinations so high ; adopting a gaussian inclination coverage function width @xmath105 , the probability of the observed sample retaining an otherwise observable object on an @xmath106 orbit is less than 40% . at @xmath107 ( the highest 3:2 inclination observed by cfeps ) , this drops to roughly 7% .
the goal of the following analysis is to utilize the 3:2 population as a `` backlight '' for the neptune trojans , illuminating the biases that likely apply to both populations .
specifically , i do this by using the 3:2 population to inform the functional form of the coverage functions @xmath80 and @xmath108 , and to derive priors for the parameters of these coverage functions . to do this
, i used the sample of all objects in the mpc which are confidently identified as non - kozai 3:2 objects by the des classification scheme ( marc buie , private communication ) .
the observables i model are their inclination @xmath109 and their ecliptic latitude at discovery @xmath110 .
i model their inclination distribution as a brown s distribution with width @xmath83 , and adopt the functional forms for @xmath80 and @xmath108 outlined above , and use abcr to back out the pdfs of @xmath83 , @xmath93 , @xmath92 , @xmath111 , and @xmath87 .
given a sample of synthetic inclinations @xmath112 drawn from a model @xmath52 , synthetic ecliptic latitudes are generated by @xmath113 , where @xmath114 is a uniform random variate over @xmath115 . to define the prior on @xmath83 , i adopt the inclination distribution results for the plutino population from the cfeps analysis in gladman et al . (
2012 ) ; i approximate the posterior probability distribution for the brown s distribution @xmath83 as a triangular pdf with minimum value @xmath116 , modal value @xmath117 , and maximum value @xmath118 .
this triangular pdf reproduces the quoted mode from gladman et al .
( 2012 ) , 2.5% of its integrated likelihood below their quoted 95% lower confidence limit and 2.5% of its integrated likelihood above their 95% upper confidence limit .
the parameter @xmath93 is drawn from a uniform prior over the 5th and 95th percentiles of @xmath110 ( @xmath119 ) .
the parameter @xmath92 is drawn from a uniform prior over @xmath120 $ ] .
the parameter @xmath111 was drawn from a uniform prior over @xmath121 $ ] .
the parameter @xmath87 was defined as @xmath122 , where @xmath123 was drawn from a uniform prior over @xmath124 this effectively defines the prior on @xmath87 such that the value of @xmath125 is drawn from a uniform prior . given these priors , the model for generating @xmath126 and @xmath127 , and the observed samples of @xmath128 and @xmath110
, i perform @xmath129 abcr trials with @xmath56 defined as the mean of @xmath130 and @xmath131 , with equal sample sizes for the observed and model samples . the 1,000 trials which produced the smallest @xmath56
are reserved , and the distribution of @xmath132 , @xmath93 , @xmath92 , @xmath111 , and @xmath87 which defined these trials defines the pdfs of each parameter .
given the simple population model , the proposed functional forms for @xmath80 and @xmath108 work remarkably well .
figure [ plut_model ] illustrates the quality with which the model inclinations and ecliptic latitudes at discovery reproduce the observed distributions .
figures [ plut_modelas ] and [ plut_modelothers ] illustrate the marginalized pdfs of each parameter for the plutino coverage function model that produced the excellent match illustrated in figure [ plut_model ] .
these pdfs ( except for that of @xmath133 ) will become the priors for the coverage functions applied to the model neptune trojans in the next section . using a similar procedure to determine the form of the longitude coverage function is less informative , as there may be a bias present for targeting the center of the trojan clouds that is _ not _ present in the discovered and reported 3:2 population .
since i am considering two trojan clouds , but am making the implicit assumption that both clouds have identical properties ( aside from some relative occupation normalization , which does not affect this model ) , the observable derived from longitude i consider is the absolute value of the separation from the mean trojan centers ( defined for zero libration amplitude ) , or @xmath134 i adopt a rollover function , normalized such that @xmath135 : @xmath136 this functional form also has two free parameters ( a turnover width @xmath137 and a turnover location @xmath138 ) . to consider coverage ranging from effectively uniform in longitude to sharply biased toward the center of the trojan swarms , i adopt a uniform distribution as the prior on @xmath137 , allowing it to range from @xmath139 ( uniform coverage ) to @xmath140 ( strongly favoring detection toward the core of the trojan clouds ) .
@xmath138 is drawn from a uniform prior over @xmath141 $ ] , roughly twice the range of all observed neptune trojan @xmath9 .
the functional forms adopted for the inclination and libration amplitude distributions are motivated by the distributions of the jupiter trojans and other trans - neptunian populations .
the intrinsic inclination @xmath142 distribution is taken to be a truncated brown s distribution ( brown 2001 ) , @xmath143 similarly , the intrinsic libration amplitude distribution is taken to be a truncated rayleigh distribution , as was adopted for the distribution of libration amplitudes for the jupiter trojans ) , @xmath144 finally , the intrinsic eccentricity distribution is also taken to be a truncated rayleigh distribution , as was adopted for the eccentricity distribution of jupiter trojans .
@xmath145 these distributions are clipped by regions of dynamical instability . for the inclination distribution , there are no stable orbits above @xmath146 ( zhou , dvorak & sun 2009 ) , and in the following analysis the inclination threshold is set at @xmath147 . for the libration amplitude distribution , the truncation is placed at @xmath148 as larger libration amplitudes are unstable ( nesvorn & dones , 2002 ) .
eccentricities are conservatively limited to values below @xmath149 ( zhou , dvorak & sun 2010 ) .
besides the physically - motivated fixed truncation locations , all of these these distributions are defined with one free parameter each . for a variety of reasons ,
the eccentricity distribution of the neptune trojans is not likely to be strongly biased by observational effects .
the eccentricity of neptune trojans are limited by stability constraints to relatively low values , @xmath150 ; additionally , by definition , all neptune trojans have nearly identical semi - major axes .
these two facts conspire to produce a relatively weak pericenter bias .
i demonstrate this by considering the relative detection rate of the eccentricity extrema of the population using a simple monte carlo test : draw two large samples of @xmath151-magnitudes from a power - law luminosity function with slope @xmath92 and maximum value @xmath152 .
one sample i preserve , and take it to represent the apparent magnitudes of objects on circular orbits @xmath153 . to the other sample i add a quantity @xmath154 where @xmath155 is drawn from a sample of keplerian orbits with unit semi - major axis , uniformly distributed mean anomaly , and non - zero , finite eccentricity this sample represents the apparent magnitudes of objects on eccentric orbits @xmath156 .
i then define a detection probability function of similar form to real survey sensitivities , @xmath157,\ ] ] where @xmath158 for typical ground based surveys .
amalgams of multiple surves tend to have broader turnovers still .
i select a value for @xmath159 and set @xmath160 .
the potential for bias is estimated by determining the ratio of the detection rate of objects on eccentric orbits to the detection rate of objects on circular orbits .
the results of this simple analysis are illustrated in fig .
[ ecc_bias ] ; for reasonable power - law slopes , the expected preference for eccentric orbits over circular ones never exceeds 20% over unity and for the most eccentric _ known _ neptune trojan , the preference never exceeds 10% .
in addition , the neptune trojan luminosity function has been observed to break to very shallow slope at @xmath161 ( sheppard & trujillo 2010a ) .
if a survey reaches fainter than this break , it effectively `` punches through '' the entire trojan population and is complete over all heliocentric distances occupied by them .
the majority of discovered neptune trojans were found in surveys that reached substantially fainter than this ( six from sheppard & trujillo 2006 & 2010b , one from parker et al .
2013 ) . for the purposes of the present analysis , then , i proceed assuming that the eccentricity bias in the detected sample is weak and can be neglected .
[ ecc_bias ] : black lines illustrate ratio of detection rates for @xmath162 orbits ( roughly the maximum stable neptune trojan eccentricity ) over circular orbits , while gray lines illustrate ratio of detection rates for @xmath163 orbits ( the maximum observed eccentricity ) over circular orbits .
dashed line corresponds to a survey flux - sensitivity rollover width of @xmath164 magnitudes , and the solid line to a broader @xmath165 magnitude rollover width .
, title="fig:",scaledwidth=50.0% ] as with the jupiter trojans , for the main parameters of interest ( @xmath69 , @xmath62 , and @xmath63 ) i propose uniformly in their distributions means .
this becomes particularly important for the neptune trojans because of the inversion of an infinite prior volume to a finite one , allowing the consideration of very large widths in a meaningful way and without imposing an artificial cutoff to the proposal volume .
the latitude and inclination coverage function parameters are drawn randomly from the retained sample of 1,000 plutino model trials .
this preserves all correlations present in the posteriors of these parameters when informed by the observed plutino population .
the longitude coverage function s inverse rollover width @xmath137 is drawn from a uniform prior over [ @xmath139 , @xmath140 ] , while @xmath138 is drawn from a uniform prior over @xmath167 $ ] , containing the range of all observed neptune trojan @xmath9 .
given the orbit distributions and priors described in the previous sections , synthetic samples of neptune trojans can be generated .
the model i use to generate the synthetic properties and discovery circumstances for the neptune trojans is described in detail in appendix b ; in brief , it takes as input the orbit distribution model parameters @xmath69 , @xmath62 , @xmath63 , @xmath168 , @xmath169 , and @xmath170 , and the coverage function forms and parameters , and returns a sample of synthetic ( @xmath171 ) that reflects both the model intrinsic distributions and the model biases simulated by the coverage functions .
[ resultstable ] c c c c c c + parameter & mode & lower 68% & upper 68% & lower
95% & upper 95% + @xmath172 & 10@xmath20 & 9@xmath20 & 16@xmath20 & 7@xmath20 & 26@xmath20 + @xmath63 & 0.044 & 0.039 & 0.070 & 0.033 & 0.125 + + @xmath69 & 21@xmath20 & 17@xmath20 & 46@xmath20 & 13@xmath20 & 91@xmath20 + + @xmath69 & 17@xmath20 & 14@xmath20 & 26@xmath20 & 11@xmath20 & 46@xmath20 + parameter estimation was performed in the abcr framework . for each case under consideration , @xmath173 abcr trials were run , and the 10,000 best trials from each were retained to generate the marginalized posterior pdfs for each parameter under consideration .
two cases were considered : in case ( 1 ) , i included the effect of the inclination coverage function informed by the plutinos . in case
( 2 ) , the inclination coverage function was ignored .
this allows the effect of the putative inclination bias on the results to be quantified .
range of accepted models .
, scaledwidth=50.0% ] range of accepted models . , scaledwidth=50.0% ] , top panel illustrating the distribution when allowing a gaussian inclination bias , and bottom panel when such a bias is not allowed for .
shaded regions illustrate 68% and 95% confidence intervals determined directly from distribution of accepted abcr samples ; solid vertical line marks mode estimated by a gaussian kernel density estimator , dashed vertical line marks distribution median .
, title="fig:",scaledwidth=50.0% ] , top panel illustrating the distribution when allowing a gaussian inclination bias , and bottom panel when such a bias is not allowed for .
shaded regions illustrate 68% and 95% confidence intervals determined directly from distribution of accepted abcr samples ; solid vertical line marks mode estimated by a gaussian kernel density estimator , dashed vertical line marks distribution median . ,
title="fig:",scaledwidth=50.0% ] , but for libration amplitude width @xmath62 .
no distinction between models including or ignoring inclination bias , so only one pdf is illustrated.,scaledwidth=50.0% ] , but for eccentricity width @xmath174 .
no distinction between models including or ignoring inclination bias , so only one pdf is illustrated.,scaledwidth=50.0% ] figure 7 illustrates the allowed intrinsic distributions of neptune trojan orbit properties and observables , while figure 8 illustrates the allowed apparent distributions of these same properties after the survey coverage functions are applied .
the most drastic effect of these coverage functions is apparent in the distribution of latitude at discovery ; while the intrinsic distribution is substantially different from the observed sample , the biased distribution matches the observed sample very well ( much as it did for the 3:2 sample ) . with the inclination coverage function drawn from the plutino priors ,
the width of the neptune trojan inclination distribution is allowed to climb to very large values ( @xmath175 ) within the 95% confidence interval , effectively allowing solutions approaching a uniform @xmath176 distribution .
the lower limit on the 95% confidence interval is @xmath177 , and the mode of the posterior pdf ( illustrated in figure 9 ) is @xmath178 .
this is admittedly a large range ; but even so it is a formal definition of the uncertainty when adopting a brown s distribution for the inclination distribution , and can be utilized for studying the origins of the population .
while their overall distribution is reasonably well - reproduced by the models adopted here , the observed neptune trojan inclinations appear somewhat bi - modal , perhaps suggesting structure more complex than the simple single brown s distribution used .
preliminary tests did not indicate any statistical evidence for a more complex two - component brown s distribution ( similar to the superimposed hot and cold classical kuiper belt ) , but such a model can not be ruled out with the current sample .
the libration amplitude distribution is somewhat better constrained at its upper end ; the maximum allowed value of @xmath179 within the 95% confidence interval is @xmath180 , which is nine degrees shy of the truncation limit of the libration amplitudes , indicating a strong requirement for the presence of the gaussian component of the rayleigh distribution .
widths as low as @xmath181 are allowed , and the peak of the distribution , @xmath182 is very similar to that of the jupiter trojans ( @xmath183 ) .
the posterior pdf for the accepted libration amplitude distribution @xmath179 parameter is illustrated in figure 10 .
the eccentricity distribution is a more moderate case ; if a slightly lower eccentricity truncation were adopted ( eg .
, 0.08 , nesvorny & vokrouhlicky 2009 ) the distribution would be consistent with @xmath184 . with the adopted truncation of @xmath185 , the maximum distribution width allowed in the 95% confidence interval is @xmath186 .
the lower limit is @xmath187 , and the peak falls at @xmath188 , similar again to the jupiter trojans ( @xmath189 ) .
the posterior pdf for the accepted eccentricity distribution @xmath63 parameter is illustrated in figure 11 .
when the inclination coverage function is set to unity over all inclinations , and all other priors remain as before , the entire confidence interval for the inclination distribution width is shifted to lower widths .
the upper limit of the 95% confidence interval becomes better constrained , at @xmath190 , significantly lower than the truncation inclination and indicating evidence for the presence of the gaussian component of the brown s distribution .
the lower limit drops by a few degrees to @xmath191 , and the mode of the pdf ( illustrated in the lower panel of figure 9 ) decreases to @xmath192 , similar to the nominal width of the plutinos ( @xmath117 , gladman et al .
2012 ) and the jupiter trojans ( @xmath193 ) .
all other distributions were unaffected , and retained identical posterior pdfs .
cccccccccccc + i d & @xmath194 & @xmath195 & @xmath196 & @xmath197 & @xmath198 & capture eff . & @xmath199 captured & @xmath200 & @xmath62 & peak @xmath142 prob . &
peak @xmath12 prob . + 1 & 26 & 17.820 & 10@xmath20 & 0.05 & 1@xmath201 & 3.20@xmath202 & 8 & @xmath203 & @xmath204 & 0.335 & 0.935 + 2 & 26 & 17.820 & 10@xmath20 & 0.1 & 1@xmath201 & 2.52@xmath205 & 63 & @xmath206 & @xmath207 & 0.544 & 0.972 + 3 & 26 & 17.820 & 10@xmath20 & 0.15 & 1@xmath201 & 1.74@xmath205 & 87 & @xmath208 & @xmath209 & 0.745 & 0.158 + 4 & 26 & 17.820 & 10@xmath20 & 0.15 & 1@xmath201 & 8.80@xmath202 & 22 & @xmath210 & @xmath211 & 0.984 & 0.425 + 5 & 26 & 17.820 & 10@xmath20 & 0.2 &
1@xmath201 & 6.00@xmath202 & 15 & @xmath212 & @xmath213 & 0.658 & 0.841 + 6 & 26 & 17.820 & 15@xmath20 & 0.15 & 1@xmath201 & 1.00@xmath205 & 25 & @xmath214 & @xmath215 & 0.739 & 0.837 + 7 & 26 & 17.820 & 2@xmath20 & 0.15 & 1@xmath201 & 3.32@xmath205 & 83 & @xmath216 & @xmath217 & 0.861 & 0.144 + 8 & 26 & 17.820 & 5@xmath20 & 0.15 & 1@xmath201 & 2.52@xmath205 & 63 & @xmath218 & @xmath219 & 0.926 & 0.461 + 9 & 28 & 19.191 & 10@xmath20 & 0.0086 & 1@xmath201 & 2.67@xmath220 & 2 & & & & + 10 & 28 & 19.191 & 10@xmath20 & 0.05 & 1@xmath201 & 1.20@xmath205 & 30 & @xmath221 & @xmath222 & 0.993 & 0.985 + 11 & 28 & 19.191 & 10@xmath20 & 0.10 & 1@xmath201 & 4.00@xmath220 & 1 & & & & + 12 & 28 & 19.191 & 10@xmath20 & 0.15 & 1@xmath201 & 5.40@xmath202 & 54 & @xmath223 & @xmath224 & 0.366 & 0.813 + 13 & 28 & 19.191 & 10@xmath20 & 0.15 & 1@xmath201 & 3.20@xmath202 & 8 & @xmath225 & @xmath226 & 0.912 & 0.608 + 14 & 28 & 18.531 & 10@xmath20 & 0.1 & 1@xmath201 & 5.20@xmath202 & 26 & @xmath227 & @xmath228 & 0.964 & 0.852 + 15 & 28 & 18.531 & 10@xmath20 & 0.2 & 1@xmath201 & 2.13@xmath202 & 16 & @xmath229 & @xmath230 & 0.273 & 0.884 + 16 & 28 & 17.870 & 10@xmath20 & 0.15 & 1@xmath201 & 1.60@xmath205 & 40 & @xmath231 & @xmath232 & 0.325 & 0.879 + 17 & 28 & 19.191 & 15@xmath20 & 0.15 & 1@xmath201 & 2.00@xmath202 & 10 & @xmath233 & @xmath234 & 0.896 & 0.705 + 18 & 28 & 19.191 & 2@xmath20 & 0.15 & 1@xmath201 & 7.20@xmath202 & 18 & @xmath235 & @xmath236 & 0.779 & 0.998 + 19 & 28 & 19.191 & 5@xmath20 & 0.15 & 1@xmath201 & 7.48@xmath202 & 56 & @xmath237 & @xmath238 & 0.940 & 0.599 + 20 & 28 & 18.531 & 8@xmath20 & 0.2 & 1@xmath201 & 8.00@xmath220 & 8 & @xmath239 & @xmath240 & 0.396 & 0.969 + 21 & 28 & 19.191 & 10@xmath20 & 0.15 & 1@xmath241 & 5.60@xmath202 & 28 & @xmath242 & @xmath243 & 0.378 & 0.381 + 22 & 28 & 19.191 & 2@xmath20 & 0.15 & 1@xmath241 & 3.60@xmath202 & 27 & @xmath244 & @xmath245 & 0.539 & 0.784 + 23 & 28 & 19.191 & 5@xmath20 & 0.15 & 1@xmath241 & 7.60@xmath202 & 19 & @xmath246 & @xmath247 & 0.619 & 0.839 + 24 & 28 & 19.191 & 10@xmath20 & 0.15 & 5@xmath201 & 5.40@xmath202 & 27 & @xmath248 & @xmath249 & 0.467 & 0.690 + 25 & 28 & 19.191 & 10@xmath20 & 0.15 & 1@xmath201 & 6.67@xmath220 & 5 & @xmath250 & @xmath251 & 0.616 & 0.764 + [ captureresults ] in order to determine the implications of the measured properties of the neptune trojan orbit distribution , i performed an exploratory suite of neptune trojan capture simulations .
these simulations were similar to those of lykawka et al .
( 2009 ) , except that i include an initial eccentric epoch for neptune and consider a pre - excited disk .
these simulations and their outcomes are detailed in the following section .
these simulations were performed with a version of _ mercury6 _ ( chambers 1999 ) that includes an artificial , user - defined force designed to drive semi - major axis migration and eccentricity damping ( wolff et al . 2012 ) .
the semi - major axis evolution followed an exponential form @xmath252 where @xmath253 is a given planet s semi - major axis at time @xmath254 in the integration , @xmath255 and @xmath256 represent the planet s initial and final semi - major axes , respectively , and @xmath198 is a characteristic timescale for migration .
eccentricity damping was performed in a slightly more general way than has been implemented in previous works .
i use an identical functional form as for semi - major axis migration , @xmath257 where @xmath258 is a given planet s eccentricity at time @xmath254 in the integration , @xmath259 and @xmath260 represent the planet s initial and final eccentricities , respectively , and @xmath198 is a characteristic timescale for eccentricity damping .
this allows me to damp eccentricity asymptotically to some final , finite value .
because i am interested in the long - term behavior and stability of captured trojans , i wish to replicate the final architecture of the giant planets orbits as closely as possible . as such
, this ability to drive eccentricities to values near their current vaues ( instead of to zero ) is important .
the initial configuration of the giant planets was motivated by the nice model and by the recent work of wolff et al .
( 2012 ) and dawson & murray - clay ( 2012 ) , which argued that in order to preserve the cold classical kuiper belt while populating the hot classical kuiper belt , neptune must have had a significant eccentricity ( @xmath261 ) at the start of its final epoch of smooth outward migration .
this initial eccentricity must have damped away relatively quickly , with a characteristic timescale less than a few times @xmath262 years .
they argue that the characteristic timescale for semi - major axis must have been longer than the characteristic timescale for eccentricity damping . in these simulations ,
neptune is started with a moderate initial eccentricity . however , in order to avoid early scattering interactions with uranus that would add an unwanted stochastic element to the outcome of these simulations , for many initial configurations neptune s eccentricity is not set as high as dictated by the limits of dawson & murray - clay ( 2012 ) given neptune s initial semi - major axis . the disks that neptune migrates into are composed of massless tracer particles .
the disks are started with a brown s distribution for inclination , a rayleigh distribution for eccentricity , and a power - law semi - major axis distribution with index @xmath263 ( approximately producing a @xmath264 surface density distribution ) defined over the range of 22 au @xmath265 30 au .
the width of the eccentricity distribution is scaled from of the width of the inclination distribution , such that @xmath266 . if any object is generated with a pericenter lower than saturn s initial semi - major axis , it is re - drawn .
initial inclination widths between @xmath267 and @xmath268 were explored .
the smallest disk considered was populated with 25,000 particles , while the largest was populated with 100,000 particles .
two initial semi - major axes were considered ( @xmath26926 au and 28 au ) , and the initial semi - major axis of uranus was set in each case to sweep through a desired range of neptune : uranus period ratios . since uranus and neptune are not trapped in their 2:1 mmr today , it is unlikely that they ever passed through it . as such , due to their current proximity to the 2:1 , the parameter volume from which the system could migrate through _ decreasing _ period ratios is very small , and i only consider initial configurations where the neptune : uranus period ratio increased ( or remained constant ) with time . at one extreme , uranus is placed as close to its current location as possible without causing early - time interactions with the initial state of neptune . for the @xmath270 au simulation ,
this places uranus at its current semi - major axis . at the other extreme
, uranus was placed such that over the course of the simulation , the neptune : uranus period ratio did not evolve . for equal @xmath198 ,
this is trivially achieved by setting @xmath271 .
several intermediate cases were explored as well .
all giant planets other than neptune experienced no eccentricity migration ; their initial eccentricities were set at values representative of their current mean values . for neptune ,
regardless of initial eccentricity , the eccentricity damping timescale @xmath272 was set to @xmath273 years .
the integrations were performed using _ mercury _ s hybrid integrator , and run for @xmath274 ; at this point , the disk particles were down - selected to only those with semi - major axes in proximity to neptune s ( semi - major axis within one neptune hill radius ) .
the integrations were then continued with only the giant planets and the down - selected disk particles until 100 myr total time had elapsed .
for two small simulations , the integration was continued for 1 gyr .
after the prime integration was complete , a short 100,000 year integration was performed where each particle s state vectors were saved very frequently . from this 100,000 year integration , the properties of the remaining objects were determined ; if they were librating around the l4 or l5 points of neptune with libration amplitudes less than @xmath275 and had eccentricities less than 0.12 , they were identified as stable neptune trojans .
the number and properties of the objects identified as stable neptune trojans are compiled for each simulation and modeled with the same distributions used to model the observed sample in the preceding sections .
table 3 outlines the initial conditions of each simulation , the number of trojans captured and the implied capture efficiency , and the properties of the inclination and libration amplitude distributions of the captured sample .
capture efficiencies are inferred from the structure of the primordial disk imposed ; since neptune may only sweep up trojans from a small radial extent within the disk , these efficiencies could have strong spatial variance and simply using a different radial disk structure could produce substantially different capture efficiencies .
these efficiencies ( @xmath276 ) , are often sufficiently high that trojan populations captured through the processes modeled here could be substantially reduced without being problematically low compared to the extant population size ( nesvorn & vokrouhlick 2009 ) .
figure [ all_sim_results ] illustrates the initial conditions and capture results of all 25 simulations . in general
, it is clear that for most simulations , particles prefer to become captured onto trojan orbits with inclinations similar to their primordial inclinations , and the mean primordial inclinations of particles that end up becoming neptune trojans generally closely match their mean inclinations as neptune trojans .
this will be discussed further in the following section , and individual simulations will be highlighted .
in all the simulations run , no captured trojan population was found to be significantly asymmetric between the l4 and l5 swarms .
figure [ l4l5ratio ] illustrates the number captured into each cloud by each simulation , and no points fall outside the 95% confidence region .
inclinations of captured trojan particles were found to change little for objects on initially inclined orbits . in general , most simulations resulted in the preferential capture of particles with higher initial inclinations , resulting in final widths higher than initial disk widths due in large part not to stirring or scattering of particles through the capture process , but rather due to this preference for capturing _ initially _ inclined objects .
this produces a trend whereby the captured trojan width is nearly identical to the initial width of _ just those particles that would go in to become trojans _ ; see figure [ sigmain ] .
this indicates that given the rates of neptune migration explored , the planetesimal disk must have had a significant component of high - inclination objects _ before _
neptune s arrival in order to produce the broad inclination widths that we see today ; indeed , initial widths of @xmath277 are required to produce widths consistent with the lower limit of the 95% confidence interval for the neptune trojan inclination width .
, while circles indicate widths of brown s distributions fit to the initial inclinations of _ just those particles that would go on to become trojans , _ and vertical location indicates the final width of the captured trojans.,scaledwidth=45.0% ] ) and initial giant planet configurations , where only the migration timescale @xmath198 was varied from @xmath77 to @xmath129 years .
points indicate the individual simulations and estimates on the uncertainty of captured trojan @xmath69 ; contours indicate acceptable regions spanned by linear models of @xmath69 vs. @xmath198 ( light gray indicates 68% confidence contour , dark gray indicates 95% confidence contour.),scaledwidth=45.0% ] in two integrations ( 15 and 20 ) , the captured trojan width is substantially higher than the initial disk width ; in these cases , neptune s initial eccentricity was set to 0.2 , and the uranus : neptune period ratio was allowed to sweep through a moderate range of values .
other simulations with the same initial giant planet architecture , but colder initial disks ( @xmath278 ) did not produce any captured trojans and are not listed in table 3 . however , as in most other simulations , in these two integrations the captured trojans were not heated by the capture process ; on average , they ended up with nearly identical final inclinations and initial inclinations .
this indicates that the capture process did not heat the trojans ; instead , the capture process _ preferred _ _
pre_-heated objects . as such , in order to produce large observed widths , a primordial disk must already be stirred to large widths to have sufficient high - inclination objects to populate the high inclination trojan orbits ; if the disk is too cold , and few objects are of sufficiently high inclination to be captured , the mean neptune trojan capture probability plummets . in both of these simulations ,
the lowest initial inclination of a captured object was @xmath279 .
thus , while these integrations seem to suggest that very broad inclination distributions can be generated from initially modest disk widths , they _ still _ require that the initial disk have a significant population of high - inclination objects in order to populate the high - inclination trojan orbits . in several integrations with cold ( @xmath280 ) primordial disks ( eg . , 18 and 22 )
, several particles are visible in figure [ all_sim_results ] that attained captured inclinations substantially higher than their primordial values .
however , the frequency of this occurrence in these simulations was low enough that the mean inclination was not substantially increased .
longer migration timescales tended to produce larger captured inclination widths , as illustrated in figure [ wvt ] . again , this increase in width seems to be largely due to an increasing preference for capturing inclined objects , instead of from stirring during the capture process .
the increase in captured trojan inclination width is relatively slow , increasing by only @xmath28120% over an order of magnitude change in migration timescale .
recent arguments that the migration of neptune may have occurred on timescales much longer than those explored here ( morbidelli et al .
2014 ) may provide an alternative explanation .
if the relatively slow increase in captured population width with increasing migration timescale changes slope beyond the regimes explored here , the long migration timescales suggested in morbidelli et al .
( 2014 ) could produce broad captured trojan inclination distributions from a cold initial disk .
naive extrapolation from the simulations performed here suggest that capture efficiencies in such a scenario would be low , and may require problematically - massive initial disks if reduced by several orders ; however , as previously noted , the capture efficiencies determined by the simulations presented here could be very sensitive to the assumptions of the primordial disk s radial structure and extent .
investing in simulations of capture simulations with much longer migration timescales than those considered here would be worthwhile .
clones of two integrations ( 3 and 12 ) were extended beyond the nominal 100 myrs , and continued out to 1 gyr ( 4 and 13 ) to explore the long - term evolution of the captured populations . in both cases ,
the captured population decayed by factors of a few , and the nominal inclination distribution widths decreased marginally but the change was not statistically significant . in both cases , the nominal libration amplitudes decreased ; for integration 4 , the libration amplitude width decrease was statistically significant , dropping from @xmath282 to @xmath283 . for integration 13 , the nominal libration amplitude width decrease was smaller and not statistically significant given the captured trojan sample sizes .
the neptune trojans have a high mean inclination ; this work has demonstrated that if the population s inclination distribution is well - modeled by a truncated brown s distribution like other resonant minor planet populations , then the width of that distribution must be greater than @xmath284 under conservative assumptions . by simulating the capture of neptune trojans from a dynamically - excited disk by a migrating , eccentric neptune
, i have shown that the inclination widths generally do not change drastically between the initial disk and the captured trojan population ; what change there is does not come in large part from stirring during the capture process , but rather from preferential capture of inclined objects .
this indicates that objects with excited orbits today were likely excited prior to neptune s arrival .
a mechanism that might be responsible for this excitation is not yet clear .
migration of planets driven by interactions with excited disks has not yet been explored in great detail .
the evidence presented in this work indicates that if neptune migrated quickly , the disk neptune migrated into must have been heated prior to neptune s arrival ; further work investigating precisely how neptune s migration would have proceeded when driven by such a disk is merited .
the computations in this paper were run on the odyssey cluster supported by the fas division of science , research computing group at harvard university .
i thank matthew holman , rebekah dawson , and konstantin batygin for their input regarding computational techniques and processes involved in planet migration and resonant capture , and wesley fraser for useful discussions on statistical techniques for minor planet population synthesis .
much of this manuscript was written in the _ 3 little figs _ cafe in somerville , massachusetts , and i thank the staff and owners for their coffee , lavender biscuits , and patience .
bernstein , g. & kushalani , b. 2000 .
aj 120 , 3323 .
brown , m. 2001 .
aj 121 , 2804 - 2814 .
chambers , j. 1999 .
mnras 203 , 793 - 799 .
dawson , r. & murray - clay , r. 2012 .
apj 750 , 43 .
gladman , b. , and 10 co - authors .
apj 144 , 23 .
gulbis , a. , elliot , j. , adams , e. , benecchi , s. , buie , m. , trilling , d. , & wasserman , l. 2010 .
aj 140 , 350 - 369 .
horner , j. & lykawka , p. 2012 .
mnras in press .
jones , r. & 16 co - authors . 2006 .
icarus 185 , 508 - 522 .
lykawka , p. , horner , j. , jones , b. , & mukai , t. 2009 .
mnras 298 , 1715 - 1729 .
marin , j .-
m . , pudlo , p. , robert , c. , & ryder , r. 2011 .
arxiv:1101.0955v2 morbidelli , a. , gaspar , h. , nesvorny , d. 2014 .
icarus 232 , 81 - 87 .
nesvorn , d. & vokrouhlick , d. 2009 .
aj 137 , 5003 - 5011 .
parker , a. & 22 co - authors .
aj 145 , 96 . petit .
j .- m . and 15 co - authors .
aj 142 , 131 .
pritchard , j. , seielstad , m. , perez - lezaun , a. , & feldman , m. 1999 .
molecular biology and evolution 16 , 1791 - 8 .
sheppard , s. and trujillo , c. 2006 .
science 313 , 511 .
sheppard , s. and trujillo , c. 2010a .
science 329 , 1304 .
sheppard , s. and trujillo , c. 2010b .
apjl 723 , l233 .
wolff , s. , dawson , r. , & murray - clay , r. 2012 .
apj 764 , 171 .
zhou , l .- y . ,
dvorak , r. & sun , y .- s .
mnras 398 , 1217 - 1227 .
the pdf of the rayleigh distribution is given by @xmath285 if this distribution is allowed to be continuous over all positive reals , then the mean is given by @xmath286 . however , in the case where there is a finite upper limit @xmath168 applied , the mean is not as straightforward .
the mean of such a truncated distribution with known pdf @xmath287 and cdf @xmath288 ( in the case where the pdf and cdf are defined for arguments greater than 0 ) is given by @xmath289 in the case of a rayleigh distribution , this evaluates to @xmath290 since as @xmath84 grows large , the truncated distribution approaches @xmath291 , the mean of this truncated rayleigh distribution has the asymptotic value of @xmath292 as @xmath76 .
[ raymean ] is a transcendental equation , so in order to propose a @xmath70 as done in the text , eqn .
[ raymean ] must be numerically solved for @xmath84 . in my implementation , this was accomplished with the _ scipy _ implementation of brent s method .
as long as the truncation value @xmath168 is greater than @xmath84 , then the mode of the truncated rayleigh distribution remains @xmath84 .
the median of the truncated rayleigh distribution is trivial to compute , but i include it for completeness here : @xmath293 which , as @xmath294 approaches the limit of the un - truncated distribution of @xmath295 . the brown s distribution ( brown 2001 ) is colloquially given by @xmath296 for this work , we will assume that it applies only to the range @xmath297 ( neglecting retrograde orbits ) .
this means that , by nature , the brown s distribution does not have a simple pdf , as it is already truncated . additionally , since integrals of the form @xmath298 do not evaluate to a convenient form , determining the normalization of a pdf or cdf is not trivial . in order to determine the mean value ,
then , we must evaluate @xmath299 to simplify this prospect , we use the taylor expansion of @xmath300 to make these integrals more similar to those we computed for the rayleigh distribution . for truncations up to @xmath301 ,
the taylor series up to order 5 is sufficient for our purposes . @xmath302 where @xmath303 given that as @xmath84 grows large , the truncated distribution approaches @xmath304 , the mean of a truncated brown s distribution has the asymptotic value of @xmath305 as @xmath76 . for the approximation given by eqn .
[ brownmeanapprox ] the astymptotic limit is @xmath306 . for truncations up to @xmath307 ,
these asymptotic values agree to within 0.05% .
as with the truncated rayleigh distribution , this must be solved numerically for @xmath84 given a desired @xmath70 , and this was accomplished with the same numerical method described previously .
given a proposed parametric model for orbit distributions , i generate a synthetic sample of neptune trojan orbits and observational circumstances in the following way .
first , inclinations @xmath142 , libration amplitudes @xmath12 , and eccentricities @xmath13 are drawn from their respective proposed distributions ( semi - major axes are assumed to be identical ) .
libration phases @xmath308 are uniformly sampled over @xmath309 , and the current resonant arguments are generated by assuming sinusoidal libration behavior : @xmath310 .
since i am defining the distribution with respect to the centers of the l4 and l5 clouds , i set @xmath311 instead of @xmath10 , where @xmath312 is a small , positive angle derived from the empirical libration properties of the synthetic neptune trojans generated by the capture simulations in this work .
this offset was found to depend on @xmath12 , @xmath13 , and ( weakly ) on @xmath142 .
true majority of the variation in mean libration center was found to be well - represented by @xmath313 ; see fig .
[ liboffset ] .
longitudes of ascending node @xmath314 are uniformly sampled over @xmath309 . instead of proposing an argument of pericenter and
mean anomaly , then solving kepler s equation , i use a small - eccentricity approximation to probabilistically determine an offset from the object s mean longitude caused by the object s eccentricity ( the epicyclic component of motion ) . for eccentricities @xmath315 , a sample of approximate offsets between true anomaly and mean anomaly can be drawn from @xmath316 where @xmath114 is a uniform random variate drawn over @xmath317 , and @xmath318 is in radians .
the two other observational parameters needed are the longitudinal angle between the trojan and the mean trojan libration center , @xmath320 , and the absolute value of the heliocentric latitude @xmath321 . from the cartesian
coordinates , now that @xmath142 , @xmath9 , and @xmath323 are available , they are passed through their respective coverage functions ; these functions represent the probability that a given object will be observed given its @xmath142 , @xmath9 , or @xmath323 .
`` synthetic observed ''
trojans are randomly sampled given the product of their coverage functions .
the biased , `` synthetic observed '' sample of ( @xmath324 ) is returned to be compared directly to the real neptune trojan properties .
libration centers found for synthetic neptune trojans .
dashed line indicates trend of @xmath313 , the adopted model used to produce synthetic neptune trojan observation circumstances.,scaledwidth=50.0% ] | the present - day orbit distribution of the neptune trojans is a powerful probe of the dynamical environment of the outer solar system during the late stages of planet migration . in this work ,
i conservatively debias the inclination , eccentricity , and libration amplitude distributions of the neptune trojans by reducing _ a priori _ unknown discovery and follow - up survey properties to nuisance parameters and using a likelihood - free bayesian rejection sampler for parameter estimation .
using this _ survey - agnostic _
approach , i confirm that the neptune trojans are a dynamically excited population : at @xmath095% confidence , the neptune trojans inclination width must be @xmath1 . for comparison and motivation purposes , i also model the jupiter trojan orbit distributions in the same basis and produce new estimates of their parameters ( jupiter trojan @xmath2 , @xmath3 , and @xmath4 ) .
the debiased inclination , libration amplitude , and eccentricity distributions of the neptune trojans are nominally very similar to those of the jupiter trojans .
i use these new constraints to inform a suite of simulations of neptune trojan capture by an eccentric , rapidly - migrating neptune from an initially dynamically - hot disk .
these simulations demonstrate that if migration and eccentricity - damping timescales were short ( @xmath5 myr , @xmath6 myr ) , the disk that neptune migrated into _ must _ have been pre - heated ( prior to neptune s appearance ) to a width comparable to the neptune trojans extant width to produce a captured population with an inclination distribution width consistent with that of the observed population . |
we consider recovering a vector @xmath0 in @xmath1 from corrupted measurements @xmath2 , where @xmath3(@xmath4 ) is the coding matrix and @xmath5 is an arbitrary and unknown vector of errors .
obviously , if the fraction of the corrupted entries is too large , there is no hope of recovering @xmath0 from @xmath16 .
however , if the fraction of corrupted measurements is small enough , one can actually recover @xmath0 from @xmath2 .
as the sparsity of @xmath5 is represented by the @xmath17 norm , @xmath18 , one natural way is to find a vector @xmath8 such that the number of terms where @xmath19 and @xmath20 differ is minimized .
mathematically , we solve the following @xmath17-minimization problem : @xmath21 however , ( [ eqn : l0 ] ) is combinatorial and computationally intractable , and one commonly used approach is to solve a closely related @xmath13-minimization problem : @xmath22 where @xmath23 .
( [ eqn : l1 ] ) can be recast as a linear program , thus can be solved efficiently .
conditions under which ( [ eqn : l1 ] ) can successfully recover @xmath0 have been extensively studied in the literature of compressed sensing ( @xcite ) .
for example , @xcite gives a sufficient condition known as the restricted isometry property ( rip ) .
recently , there has been great research interest in recovering @xmath0 by @xmath6-minimization for @xmath24 ( @xcite ) as follows , @xmath25 recall that @xmath26 for @xmath27 .
we say @xmath0 can be recovered by @xmath6-minimization if and only if it is the unique solution to ( [ eqn : lp ] ) .
then the question is what is the relationship between the sparsity of the error vector and the successful recovery with @xmath6-minimization ? ( [ eqn : lp ] ) is non - convex , and thus it is generally hard to compute the global minimum .
however , @xcite shows numerically that we can recover @xmath0 by finding a local minimum of ( [ eqn : lp ] ) , and @xmath6-minimization outperforms @xmath13-minimization in terms of the sparsity restriction for @xmath5 .
@xcite extends rip to @xmath6-minimization and analyzes the ability of @xmath6-minimization to recover signals from noisy measurements .
@xcite also provides a condition for the success recovery via @xmath6-minimization , which can be generalized to @xmath28 case .
both conditions are sufficient but not necessary , and thus are too restrictive in general .
let @xmath29 be an arbitrary and unknown vector of errors on support @xmath30 .
we say @xmath5 is @xmath31-sparse if @xmath32 for some @xmath33 where @xmath34 is the cardinality of set @xmath35 .
our main contribution is a sharp threshold @xmath36 for all @xmath37 such that for @xmath38 , if @xmath39 for some constant @xmath40 and the entries of @xmath41 are i.i.d .
gaussian , then @xmath6-minimization can recover @xmath0 with overwhelming probability .
we provide two thresholds : one ( @xmath42 ) is for the case when @xmath5 is an arbitrary unknown vector , and the other ( @xmath43 ) assumes that @xmath5 has fixed support and fixed signs . in the latter case ,
the condition of successful recovery with @xmath13-minimization from any possible error vector is the same , while the condition of successful recovery with @xmath6-minimization ( @xmath24 ) from different error vectors differs . using worst - case performance as criterion
, we prove that though @xmath6 outperforms @xmath13 in the former case , it is not comparable to @xmath13 in the latter case .
both bounds @xmath42 and @xmath43 are tight in the sense that once the fraction of errors exceeds @xmath42 ( or @xmath44 ) , @xmath6-minimization can be made to fail with overwhelming probability .
our technique stems from @xcite , which only focuses on @xmath13-minimization and the case that @xmath5 is arbitrary .
in this section , we shall give a function @xmath36 such that for a given @xmath10 , for any @xmath45 , when the entries of @xmath41 are i.i.d .
gaussian , the @xmath6-minimization can recover @xmath0 with overwhelming probability as long as the error @xmath5 is @xmath31-sparse .
the following theorem gives an equivalent condition for the success of @xmath6 minimization ( @xcite ) .
[ thm : slp ] @xmath0 is the unique solution to @xmath6 minimization problem @xmath46 for every @xmath0 and for every @xmath47-sparse @xmath5 if and only if @xmath48 for every @xmath49 , and every support @xmath35 with @xmath50 .
one important property is that if the condition ( [ eqn : slp ] ) is satisfied for some @xmath7 , then it is also satisfied for all @xmath51 ( @xcite ) .
now we define the threshold of successful recovery @xmath42 as a function of @xmath10 .
let @xmath52 , @xmath53, ... ,@xmath54 be i.i.d @xmath55 random variables and let @xmath56 , @xmath57, ... ,@xmath58 be the sorted ordering ( in non - increasing order ) of @xmath59 , @xmath60, ...
,@xmath61 for some @xmath62 $ ] . for a @xmath63 , define @xmath64 as @xmath65 . let @xmath66 denote
@xmath67 $ ] , the expected value of @xmath68 .
then there exists a constant @xmath36 such that @xmath69}{s}=\frac{1}{2}$ ] .
let @xmath70 and let @xmath71 .
let @xmath72 denote the p.d.f . of @xmath73 and @xmath74 be its c.d.f .
define @xmath75 .
@xmath76 is continuous and decreasing in @xmath77 $ ] , and @xmath78=\frac{s}{m}$ ] , @xmath79
. then there exists @xmath80 such that @xmath81 , we claim that @xmath82 has the desired property .
. then @xmath84=mg(z^*)$ ] .
since @xmath85 $ ] is bounded by @xmath86 , and @xmath87 , thus @xmath88}{s}=\frac{1}{2}.$ ] [ prop : rho ] the function @xmath36 is strictly decreasing in @xmath10 on @xmath89 $ ] . from the definition of @xmath80 and @xmath36 , we have @xmath90 and @xmath91 where @xmath92 and @xmath93 are the p.d.f . and
c.d.f . of @xmath94 ,
@xmath95 . from the implicit function theorem , @xmath96 from the chain rule ,
we know @xmath97 , thus @xmath98 note the numerator of ( [ eqn : drhodp ] ) is less than 0 from ( [ eqn : zp ] ) , thus @xmath99 .
we plot @xmath42 against @xmath10 numerically in fig .
[ fig : rho ] .
@xmath36 goes to @xmath100 as @xmath10 tends to zero .
note that @xmath101 , which coincides with the result in @xcite . of successful recovery with @xmath6-minimization
] now we proceed to prove that @xmath42 is the threshold of successful recovery with @xmath6 minimization for @xmath10 in @xmath89 $ ] .
first we state the concentration property of @xmath64 in the following lemma .
[ lemma : srho ] for any @xmath62 $ ] , let @xmath52, ...
,@xmath54 , @xmath56, ... ,@xmath58 , @xmath64 and @xmath66 be as above .
for any @xmath102 and any @xmath103 , there exists a constant @xmath104 such that when @xmath105 is large enough , with probability at least @xmath106 , @xmath107| \leq \delta s$ ] .
let @xmath108^t$ ] . if two vectors @xmath109 and @xmath110 only differ in co - ordinate @xmath111 , then for any @xmath10 , @xmath112 since @xmath113 for all @xmath114 $ ] , latexmath:[\[\label{eqn : srho } from the isoperimetric inequality for the gaussian measure ( @xcite ) , for any set @xmath41 with measure at least a half , the set @xmath116 has measure at least @xmath117 , where @xmath118 .
let @xmath119 be the median value of @xmath120 .
define set @xmath121 , then @xmath122\geq 1-e^{-t^2/2}.\ ] ] we claim that @xmath123 implies that @xmath124 . if @xmath125 , then @xmath126 , thus the claim holds as @xmath127 is non - negative .
if @xmath128 , then there exists @xmath129 such that @xmath130 . let @xmath131 for all @xmath111 and let @xmath132 . from hlder s inequality @xmath133 from ( [ eqn : srho ] ) and ( [ eqn : holder ] ) , @xmath134 . since @xmath135 and @xmath129 , then @xmath136 .
thus @xmath137 , which verifies our claim
. then @xmath138\geq pr[d(x , a ) \leq t]\geq 1-e^{-t^2/2}.\ ] ] similarly , @xmath139\geq 1-e^{-t^2/2}.\ ] ] combining ( [ eqn : srholb ] ) and ( [ eqn : srhoub ] ) , @xmath140\leq 2e^{-t^2/2}.\ ] ] the difference of @xmath141 $ ] and @xmath142 can be bounded as follows , @xmath143 dy\\ & \leq & \int_0^\infty 2e^{-\frac{1}{2}y^{\frac{2}{p}}m^{(1-\frac{2}{p } ) } } dy\\ & = & m^{(1-\frac{p}{2})}\int_0^\infty 2e^{-\frac{1}{2}s^{\frac{2}{p}}}ds\\ \ ] ] note that @xmath144 is a finite constant for all @xmath62 $ ] .
as @xmath27 and @xmath145 $ ] , thus for any @xmath146 , @xmath147 when @xmath105 is large enough .
let @xmath148)^{\frac{1}{p}}\sqrt{m}$ ] , from ( [ eqn : srhomrho ] ) with probability at least ( @xmath149)^{\frac{2}{p}}m}$ ] ) , @xmath150 .
thus @xmath107| \leq @xmath106 for some constant @xmath151 .
[ cor : srho ] for any @xmath152 , there exists a @xmath103 and a constant @xmath153 such that when @xmath105 is large enough , with probability @xmath154 , @xmath155 . when @xmath152 , @xmath156&=&e[s_{\rho^*}]-\sum \limits_{i=\lceil
\rho m\rceil+1}^{\lceil \rho^*m \rceil } e[|x_i|^p]\\ & \leq & e[s_{\rho^*}]-(\lceil \rho^*m \rceil-\lceil \rho m\rceil ) e[|x_i|^p]\end{aligned}\ ] ] then @xmath157/s \leq \frac{1}{2}-2\delta$ ] for a suitable @xmath158 as @xmath159 $ ]
. the result follows by combining the above with lemma [ lemma : srho ] .
[ cor : s1 ] for any @xmath160 , there exists a constant @xmath161 such that when @xmath105 is large enough , with probability @xmath162 , it holds that @xmath163 .
the above two corollaries indicate that with overwhelming probability the sum of the largest @xmath164 terms of @xmath165 s is less than half of the total sum @xmath68 if @xmath166 .
the following lemma extends the result to every vector @xmath167 where matrix @xmath3 has i.i.d .
gaussian entries and @xmath168 is any vector in @xmath1 .
[ lemma : slp ] for any @xmath7 , given any @xmath169 , there exist constants @xmath170 , @xmath171 , @xmath146 such that when @xmath172 and @xmath173 is large enough , with probability @xmath174 , an @xmath175 matrix @xmath41 with i.i.d .
@xmath55 entries has the following property : for every @xmath49 and every subset @xmath176 with @xmath32 , @xmath177 . for any given @xmath178
, there exists a @xmath179-net @xmath180 of cardinality less than @xmath181(@xcite ) .
a @xmath179-net @xmath180 is a set of points such that @xmath182 for all @xmath183 in @xmath180 and for any @xmath168 with @xmath184 , there exists some @xmath183 such that @xmath185 .
since @xmath41 has i.i.d @xmath55 entries , then @xmath186 has @xmath105 i.i.d .
@xmath55 entries . applying a union bound to corollary [ cor : srho ] and [ cor : s1 ] , we know that for some @xmath103 and for every @xmath187 , with probability @xmath188 for some @xmath189 , we have @xmath190 and @xmath191 hold for a vector @xmath183 in @xmath180 . taking @xmath192 for large enough @xmath170 , from union
bound we get that ( [ prop : srho ] ) and ( [ prop : s1 ] ) hold for all the points in @xmath180 at the same time with probability at least @xmath174 for some @xmath193 . for any @xmath168 such that @xmath184 , there exists @xmath194 in @xmath180 such that @xmath195 .
let @xmath196 denote @xmath197 , then @xmath198 for some @xmath199 in @xmath180 .
repeating this process , we have @xmath200 where @xmath201 , @xmath202 and @xmath203 . thus for any @xmath49
, we have @xmath204 . for any index set @xmath35 with @xmath32 , @xmath205 @xmath206
thus @xmath207 .
for a given @xmath158 , we can pick @xmath208 and @xmath209 small enough such that @xmath210 .
we can now establish one main result regarding the threshold of successful recovery with @xmath6-minimization . for any @xmath7 , given any @xmath169 , there exist constants @xmath170 , @xmath211 such that when @xmath172 and @xmath173 is large enough , with probability @xmath174 , an @xmath212 matrix @xmath41 with i.i.d .
@xmath55 entries has the following property : for every @xmath213 and every error @xmath5 with its support @xmath35 satisfying @xmath32 , @xmath0 is the unique solution to the @xmath6-minimization problem ( [ eqn : lp ] ) .
lemma [ lemma : slp ] indicates that @xmath214 for every non - zero @xmath168 , then from theorem [ thm : slp ] , @xmath0 is the unique solution to the @xmath6-minimization problem ( [ eqn : lp ] ) .
we remark here that @xmath42 is a sharp bound for successful recovery . for any @xmath215 , from lemma [ lemma : srho ] , with overwhelming probability the sum of the largest @xmath216 terms of @xmath217 s is more than the half of the total sum @xmath68 , then theorem [ thm : slp ] indicates that the @xmath6-recovery fails in this case .
in fact , for any vector @xmath218 , let @xmath219 , and let @xmath35 be the support of the largest @xmath220 terms of @xmath217 s .
if the error vector @xmath5 agrees with @xmath217 on the support @xmath35 and is zero elsewhere , then with large probability @xmath221 is no greater than that of @xmath222 , which implies that @xmath6-minimization can not correctly return @xmath0 .
proposition [ prop : rho ] thus implies that the threshold strictly decreases as @xmath10 increases .
the performance of @xmath223-minimization is better than @xmath224-minimization for @xmath225 in the sense that the sparsity requirement for the arbitrary error vector is less strict for smaller @xmath10 .
in section [ sec : sbd ] , for some @xmath102 , we call @xmath6-minimization successful if and only if it can recover @xmath0 from any error @xmath5 whose support size is at most @xmath47 . here
we only require @xmath6-minimization to recover @xmath0 from errors with fixed but unknown support and signs .
we will provide a sharp threshold @xmath43 of the proportion of errors below which @xmath6-minimization is successful .
once the support and the signs of an error vector is fixed , the condition of successful recovery with @xmath13-minimization from any such error vector is the same , however , the condition of successful recovery with @xmath6-minimization from different error vectors differs even the support and the signs of the error is fixed . here
we consider the worst case scenario in the sense that the recovery with @xmath6-minimization is defined to be `` successful '' if @xmath0 can be recovered from any such error @xmath5 .
we characterize this case in theorem [ thm : wb ] .
note that if there is further constraint on @xmath5 , then the condition of successful recovery with @xmath6-minimization may be different from the one stated in theorem [ thm : wb ] .
[ thm : wb ] given any @xmath226 , for every @xmath213 and every error @xmath5 with fixed support @xmath35 and fixed sign for each entry @xmath227 , if @xmath0 is always the unique solution to @xmath6-minimization problem ( [ eqn : lp ] ) , then @xmath228 .
let @xmath231 for every @xmath111 in @xmath232 , let @xmath233 for every @xmath111 in @xmath234 .
for every @xmath111 in @xmath235 , let @xmath236 satisfy @xmath237 . as @xmath226 , we can pick @xmath236 ( @xmath238 ) with @xmath239 large enough such that @xmath240 thus @xmath241 , @xmath0 is not a solution to ( [ eqn : lp ] ) , which is a contradiction .
the first inequality holds as for each @xmath111 in @xmath245 , @xmath246 has the same sign as that of @xmath236 if not zero ; and for @xmath226 , @xmath247 holds .
the second inequality comes from the assumption that @xmath248>@xmath249 for all @xmath242 .
[ lemma : rhow ] let @xmath52 , @xmath53, ... ,@xmath54 be i.i.d .
@xmath55 random variables and @xmath35 be a set of indices with size @xmath250 for some @xmath102 .
let @xmath29 be any vector on support @xmath35 with fixed signs for each entry .
if @xmath251 , for every @xmath252 , when @xmath105 is large enough , with probability @xmath253 for some constant @xmath254 , the following two properties hold : define a random variable @xmath258 for each @xmath111 in @xmath35 that is equal to 1 if @xmath259 and equal to 0 otherwise . then @xmath260 .
@xmath261=\frac{1}{2}\mu$ ] for every @xmath111 in @xmath35 as @xmath262 . from chernoff bound , for any @xmath160 , there exist @xmath263 and @xmath264 such that lemma [ lemma : rhow ] implies that @xmath272 . applying the similar net argument in section [ sec : sbd ]
, we can extend the result to every vector @xmath167 where matrix @xmath3 has i.i.d .
gaussian entries and @xmath168 is any vector in @xmath1
. then we can establish the main result regarding the threshold of successful recovery with @xmath6-minimization from errors with fixed support and signs . for any @xmath226 , given any @xmath273
, there exist constants @xmath274 , @xmath275 such that when @xmath276 and @xmath173 is large enough , with probability @xmath277 , an @xmath212 matrix @xmath41 with i.i.d .
@xmath55 entries has the following property : for every @xmath213 and every error @xmath5 with fixed support @xmath35 satisfying @xmath32 and fixed signs on @xmath35 , @xmath0 is the unique solution to the @xmath6-minimization problem ( [ eqn : lp ] ) .
we remark here that @xmath44 is a sharp bound for successful recovery in this setup . for any @xmath291 , from lemma [ lemma : rhow ] , with overwhelming probability that @xmath292 , then theorem [ thm : wb ] indicates that the @xmath6-recovery fails for some error vector @xmath5 in this case .
surprisingly , the successful recovery threshold @xmath42 when fixing the support and the signs of an error vector is @xmath11 for all @xmath10 in @xmath12 and is strictly less than the threshold for @xmath293 , which is 1 ( @xcite ) .
thus in this case , @xmath13-minimization has better recovery performance than that of @xmath6-minimization ( @xmath24 ) in terms of the sparsity requirement for the error vector .
the result seems counterintuitive , however , it largely depends on the definition of successful recovery in terms of worse case performance .
the condition of successful recovery via @xmath13-minimization from any error vector on the fixed support with fixed signs is the same , while the condition of @xmath6-minimization from different error vectors differs .
s. foucart and m .- j .
lai , `` sparsest solutions of underdetermined linear systems via @xmath296-minimization for @xmath297 , '' _ applied and computational harmonic analysis _ , vol .
26 , no . 3 , pp . 395 407 , 2009 . | an unknown vector @xmath0 in @xmath1 can be recovered from corrupted measurements @xmath2 where @xmath3(@xmath4 ) is the coding matrix if the unknown error vector @xmath5 is sparse .
we investigate the relationship of the fraction of errors and the recovering ability of @xmath6-minimization ( @xmath7 ) which returns a vector @xmath8 minimizing the `` @xmath6-norm '' of @xmath9 .
we give sharp thresholds of the fraction of errors that determine the successful recovery of @xmath0 .
if @xmath5 is an arbitrary unknown vector , the threshold strictly decreases from 0.5 to 0.239 as @xmath10 increases from 0 to 1 .
if @xmath5 has fixed support and fixed signs on the support , the threshold is @xmath11 for all @xmath10 in @xmath12 , while the threshold is 1 for @xmath13-minimization .
# 1@xmath14**#1**@xmath15 |
analyzing pure quantum states in an infinite dimensional hilbert space it is useful to distinguish a family of _ coherent states _ , localized in the classical phase space and minimizing the uncertainty principle .
these quantum analogues of points in the classical phase space are often considered as classical states . for an arbitrary quantum state
one may pose a natural question , to what extent it is nonclassical in a sense that its properties differ from that of coherent states .
in other words , is there any parameter that may legitimately reflect the degree of nonclassicality of a given quantum state ?
this question was motivated with the first observation of nonclassical features of electromagnetic fields such as sub - poissonian statistics , antibunching and squeezing .
additionally , it is well known that the interaction of ( non)linear devices with quantum states may flip from one state to another ; for instance , nonlinear devices may produce nonclassical states from their interaction with the vacuum or a classical field .
a systematic survey of nonclassical properties of quantum states would be worthwhile because of the nowadays ever increasing number of experiments in nonlinear optics .
an earlier attempt to sheding some light on the nonclassicality of a quantum state was pioneered by mandel @xcite , who investigated the radiation fields and introduced a parameter @xmath0 measuring the deviation of the photon number statistics from the poissonian distribution , characteristic of coherent states . in general , to define a measure of nonclassicality of quantum states one can follow several different approaches @xcite . distinguishing a certain set @xmath1 of states ( e.g. the set of coherent states @xmath2 ) ,
one looks for the distance of an analyzed pure state @xmath3 to this set , by minimizing a distance @xmath4 over the entire set @xmath1 .
such a scheme based on the trace distance was first used by hillery @xcite , while other distances ( hilbert - schmidt distance @xcite or bures distance @xcite ) were later used for this purpose .
the same approach is also applicable to characterize mixed quantum states : minimizing the distance of the density @xmath5 to the set of coherent states is related @xcite to the search for the maximal fidelity ( the hilbert - schmidt fidelity @xmath6 or the bures - uhlmann fidelity @xmath7 ) with respect to any coherent state , @xmath8 . on the same footing
, the monge distance introduced in @xcite may be applied to describe , to what extent a given mixed state is close to the manifold of coherent states . yet another way of proceeding
is based on the generalized ( cahill ) phase space representation @xmath9 of a pure state , which interpolates between the husimi ( @xmath10 ) , the wigner ( @xmath11 ) and the glauber
sudarshan ( @xmath12 ) representations .
the cahill parameter @xmath13 is proportional to the variance of a gaussian function one needs to convolute with @xmath12 representation to obtain @xmath9 @xcite .
in particular for @xmath14 one obtains the q- , w- and p- representations , respectively . by construction
the @xmath10 representation is non - negative for all states , while the wigner function may admit also negative values , and the @xmath12 representation may be singular or may not exist .
the smoothing effect of @xmath9 is enhanced as @xmath13 increases .
if @xmath13 is large enough so that @xmath9 becomes positive definite regular function , thus acceptable as a classical distribution function , then the smoothing is said to be complete .
the greatest lower bound @xmath15 for the critical value was adopted by lee @xcite , as _ nonclassical depth _ of a quantum state and this approach was further developed in @xcite .
the limiting value , @xmath16 , corresponds to the @xmath10 function which is always acceptable as a classical distribution function .
the lowest value , @xmath17 , is ascribed to an arbitrary coherent state because its @xmath12 function is a dirac delta function , so its @xmath18smoothing becomes regular .
the range of @xmath15 is thus @xmath19 $ ] . if the husimi function of a pure state admits at least one zero @xmath20 , then a cahill @xmath9 distribution with a narrower smearing , @xmath21 , becomes negative in the vicinity of @xmath22 .
therefore the classical depth for such quantum states is maximal , @xmath16 @xcite .
the only class of states , for which @xmath10 representation has no zeros , are the squeezed vacuum states , for which @xmath15 is a function of the squeezing parameter @xmath23 . in the limiting case @xmath24 one
obtains the coherent state for which the @xmath25 distribution is a dirac delta function , that is @xmath17 .
a closely related approach to characterizing quantum states is based on properties of their wigner functions in phase space @xmath26 .
one can prove that the wigner function is bounded from below and from above @xcite . in the normalization @xmath27
used later in this work , such a bound reads @xmath28 .
further bound on integrals of the wigner function were derived in @xcite , while an entropy approach to the wigner function was developed in @xcite . in order to interpret the wigner function as a classical probability distribution one needs to require that @xmath11 is non negative . as found by hudson in 1974
@xcite , this is the case for coherent or squeezed vacuum states only .
a possible measure of nonclassicality may thus be based on the negativity of the wigner function which may be interpreted as a signature of quantum interference .
the negativity of the wigner function has been linked to nonlocality , according to the bell inequality @xcite , while investigating the original einstein - podolsky - rosen ( epr ) state @xcite .
in fact bell argued that the epr state will not exhibit nonlocal effects because its wigner function is everywhere positive , and as such will allow for a hidden variable description of correlations .
however , it is now demonstrated @xcite that the wigner function of the epr state , though positive definite , provides a direct evidence of nonlocality .
this violation of the bell s inequality holds true for the regularized epr state @xcite and also for a correlated two - mode quantum state of light @xcite .
it is also worth recalling that the wigner function can be measured experimentally @xcite , including the measurements of its negative values @xcite .
the interest put on such experiments has triggered a search for operational definitions of the wigner functions , based on experimental setup @xcite .
the aim of this letter is to study a simple indicator of the nonclassicality , which depends on the _ volume _ of the negative part of the wigner function . to demonstrate a potential use of such an approach we investigate certain families of quantum states .
the nonclassicality indicator is defined in section 2 .
the schrdinger cat state , being constructed as coherent superposition of two gaussian wave packets , is analyzed in section 3 while section 4 is devoted to fock states and to the squeezed displaced fock states . finally in section 5 , a brief discussion of results and perspectives
the wigner function of a state @xmath3 defined by @xcite @xmath29 satisfies the normalization condition @xmath30 .
hence the doubled volume of the integrated negative part of the wigner function may be written as @xmath31 dqdp = \iint |w_{\psi } ( q , p)| dqdp -1 \ . \label{defclass}\end{aligned}\ ] ] by definition , the quantity @xmath32 is equal to zero for coherent and squeezed vacuum states , for which @xmath11 is non - negative . hence in this work
we shall treat @xmath32 as a parameter characterizing the properties of the state under consideration .
similar quantities related to the volume of the negative part of the wigner function were used in @xcite to describe the interference effects which determine the departure from classical behaviour .
furthermore , a closely related approach was recently advocated by benedict and collaborators @xcite .
their measure of the nonclassicality of a state @xmath3 reads @xmath33 where @xmath34 and @xmath35 are the moduli of the integrals over those domains of the phase space where the wigner function is positive and negative , respectively .
the normalization condition implies @xmath36 , so that @xmath37 leads to @xmath38 . using this notation we may rewrite ( [ defclass ] ) as @xmath39 and
obtain a simple relation between both quantities @xmath40 with @xmath41 .
it turns out that both quantities are equivalent in the sense that they induce the same order in the space of pure states : the relation @xmath42 implies @xmath43 .
however , from a pragmatic point of view there exists an important difference between both quantities . to compute explicitly the quantity ( [ benedict ] )
one faces a difficult task to identify appropriately the domains , in which the integration has to be carried out . on the other hand , knowing the wigner function @xmath44 of a quantum state , it is straightforward to get its absolute value and to evaluate numerically the integration ( [ defclass ] ) .
let us emphasize again that the hilbert space containing all pure states is huge , so one should not expect to characterize the nonclassical features of a quantum state just by a single scalar quantity .
our approach focuses on a particular issue , whether the wigner function is positive and may be interpreted as a classical probability distribution .
therefore , the proposed indicator @xmath32 should be considered as a tool complementary to these worked out earlier and reviewed above .
a quantum state , called _ schrdinger cat _ , is a coherent superposition of two coherent states localized in two distant points of the configuration space , @xmath45 .
the wave function of such a state reads in the position representation @xmath46 \label{wftion}\end{aligned}\ ] ] where @xmath47 from now on atomic units are used @xmath48 . in other words
we measure the size of the product @xmath49 in units of @xmath50 .
the classical limit @xmath51 means the action @xmath49 characteristic of the system is many order of magnitude larger than @xmath50 .
a glance on eq .
( [ plusmin ] ) reveals that the phase , governed by @xmath52 , is of great importance in that it induces oscillations on the wave function as can be seen in fig.1 .
note that the normalization constant @xmath53 depends on the location of the centers @xmath54 of both coherent states that make up the cat state .
therefore one sees that the wigner function may depend not only on the distance @xmath55 between the both states , but also on their momentum , @xmath52 .
so far , the studies on the cat states @xcite were usually restricted to the case of standing cats , @xmath56 . in this letter
we demonstrate that the parameter @xmath52 influences the shape of the wigner function , in particular , if @xmath57 and both packets are not spatially separated . inserting @xmath58 into the wigner function ( [ wdef ] ) one obtains @xmath59 here @xmath60 represent two peaks of the distribution centered at the classical phase space points @xmath61 , while @xmath62 stands for the interference structure which appear between both peaks .
normalizing ( [ wftion ] ) yields @xmath63 making use of the formula ( [ wcats ] ) for the wigner function of the cat state @xmath64 its nonclassicality parameter @xmath65 may be approximated by @xmath66 - 1 \ . \label{approx}\end{aligned}\ ] ] strictly speaking the right hand side of equation ( [ approx ] ) forms an upper bound for @xmath67 , which may be practically used as its fair approximation .
because of the oscillations of the absolute value of cosine , it is difficult to perform the integration analytically . in the special case @xmath68 , the superposition of coherent states ( [ wftion ] ) reduces to a single coherent state and correspondingly ( [ approx ] ) leads to @xmath69 .
2 shows plots of the wigner function of the cat states for several values of the separation @xmath70 and the momentum @xmath52 .
one clearly sees the formation of the quantum interference structure halfway between the two humps as the separation distance @xmath70 increases .
the frequency of the interference structure increases with the separation @xcite . for intermediate separations ( @xmath71 ) ,
the wigner function changes its structure with @xmath52 , see fig.2b and 2c .
however , for a larger separation distance , @xmath72 , the wigner function for @xmath73 may be approximated by the wigner function for the state with @xmath56 translated by a constant vector @xmath74 . in the case of standing cats , ( @xmath56 ) ,
the indicator @xmath32 increases monotonically with the separation @xmath70 , and reflects presence of the interference patterns at @xmath75 - see fig .
the growth of the nonclassicality saturates at @xmath76 , as the interference patterns become practically separated from both peaks , and the parameter @xmath32 tends to the limiting value , @xmath77 . in the limit @xmath78 the oscillations of the cosine term in eq .
( [ approx ] ) become rapid and a crude approximation @xmath79 gives an explicit upper bound @xmath80 .
this picture gets more complicated for the states with @xmath81 , in particular for a small separation distance , @xmath82 . in this case
, @xmath32 exhibits oscillations as shown in fig.4l , 4 m , 4n . to shed some light on this behavior
we have chosen to plot in fig.3 the wigner function for which @xmath83 achieves extremal values .
for instance , @xmath32 at @xmath84 ( fig .
3b ) is smaller than at @xmath85 ( fig .
3a ) or @xmath86 ( fig .
this is due to the the interference structure , which is not symmetric with respect to the reflection @xmath87 , in contrast to the case of cats with @xmath56 . as shown in fig .
4 , the frequency of oscillations increases with @xmath52 , but the limiting value @xmath88 does not depend on the initial momentum @xmath52 .
this can also be demonstrated , investigating the dependence of the quantity @xmath32 as a function of @xmath52 . as follows from eq .
( [ approx ] ) , the indicator @xmath32 displays regular oscillations with the period @xmath89 see fig .
5 . in other words
a non zero separation parameter @xmath70 breaks the translational invariance in momentum and introduces a characteristic momentum scale @xmath90 .
note that the amplitudes of the oscillations decrease fast with @xmath70 , so that for well separated cats with @xmath72 the quantity @xmath32 is practically independent on @xmath52 .
let us consider the squeezed displaced fock state defined by @xmath91 where @xmath92 is the original fock state and @xmath93 .
the displacement @xmath94 and the squeezed @xmath95 operators are defined by @xcite @xmath96 where @xmath97 and @xmath98 are usual photon annihilation and creation operators , respectively .
the complex variable @xmath99 represents the magnitude and angle of the displacement .
similarly , writing the complex number in its polar form , @xmath100 , it is easy to see that the radius @xmath23 plays the role of the squeezing strength while the angle @xmath101 indicates the direction of squeezing . it was shown in @xcite that the displacement operators @xmath94 form a complete set of operators .
thus any bounded operator @xmath102 , ( for which the hilbert - schmidt norm @xmath103 is finite ) , can be expressed in the form @xmath104 in which the weight function @xmath105 is unique and square - integrable .
given that every density operator is bounded ( @xmath106 ) , one may write an arbitrary density operator @xmath107 . here
the weight function @xmath108 is just the expectation value of the displacement operator commonly known as characteristic function .
the complex fourier transform of @xmath109 defines the wigner function @xmath110 one may therefore express @xmath109 in terms of the wigner function by performing the inverse fourier transform as @xmath111 so that upon substitution into the density operator expression above , one gets @xmath112 the operators @xmath113 denote @xmath114 so that the wigner function may be interpreted as a weight function for the expansion of the density operator in terms of the operators @xmath113 @xcite .
these operators are hermitian , @xmath115 , and possess the same completeness properties as the displacement operators @xmath116 .
making use of the parity operator @xmath117 , one finally shows that @xmath118 with @xmath119 being the photon number . in the case of
the squeezed displaced fock states , @xmath120 , the wigner function becomes @xmath121 performing explicitly calculations of matrix elements , one obtains : @xmath122 with @xmath123@xcite . here
@xmath124 denotes the laguerre polynomial of the @xmath125-th order .
the wigner function ( [ wigner ] ) allows us to compute the nonclassicality parameter @xmath126 for a given displaced squeezed fock state @xmath127 . in what follows certain special cases will be investigated such as squeezed displaced vacuum states , pure fock states and squeezed displaced fock states . it will be therefore convenient to represent the complex variable @xmath128 by the position and momentum coordinates , @xmath129 , and treat likewise the displacement operator , @xmath130 . substituting @xmath131 in eq .
( [ wigner ] ) yields the wigner function for the fock state @xmath92 , @xmath132 l_n\left[2(q^2+p^2)\right ] \ . \label{wfstate}\end{aligned}\ ] ] this allows to evaluate analytically the indicator @xmath133 , for @xmath134 @xmath135 since the zeros of the laguerre polynomials are available up to the @xmath136th order . for larger @xmath125 we computed the quantity @xmath133
numerically and plotted in fig .
the indicator @xmath32 grows monotonically with @xmath125 , as the number of zeros of the laguerre polynomial @xmath124 increases with @xmath125 . for
@xmath137 $ ] this dependence may be aproximated by @xmath138 . hence , the larger
the quantum number @xmath125 , the less the wigner function @xmath139 can be interpreted as a classical distribution function . setting @xmath140 in ( [ wigner ] )
one obtains a squeezed coherent state or squeezed vacuum state . choosing the squeezing angle @xmath141
, one sees that the wigner function is a gaussian centered at the displacement vector ( @xmath142 ) with the shape determined by the squeezing parameter @xmath23 , @xmath143 in such a case the wigner function remains everywhere non
negative for any choice of the squeezing and displacement parameters @xcite , so that the nonclassicality indicator vanish , @xmath144 .
note that the displacement of any state in phase space does not change the shape of the wigner function , so the quantity @xmath32 is independent of the displacement operator @xmath94 .
furthermore , the squeezing operator @xmath95 influences the shape of the wigner function , but does not lead to a change in the volume of its negative part .
therefore , the parameter @xmath32 does not also depend on the squeezing .
as an illustration we have chosen the squeezed ( @xmath145 ) displaced ( @xmath146 ) third photon ( @xmath147 ) state , @xmath148 .
the contour plots of the wigner function of such a state are shown in fig . 7 for some values of the squeezing parameter @xmath23 .
the indicator @xmath32 is equal to @xmath149 , irrespective to the squeezing strength .
if squeezing is strong enough , the ring - like wigner function collapses to a quasi one dimensional object with a cigar form .
the squeezed vacuum is often described as a nonclassical state @xcite .
since the quantity @xmath32 does not depend on squeezing , it should not be interpreted as the only parameter which characterizes the nonclassicality . to describe the nonclassical features of the squeezed states one may use , for instance , the nonclassical depth @xcite .
in this work we have proposed a simple indicator of non - classicality which measures the volume of the negative part of the wigner function . although the proposed coefficient @xmath32 is a function of the related quantity @xmath150 , recently introduced by benedict , czirjk et al .
@xcite , it is much easier to compute numerically .
the quantity ( [ defclass ] ) was used to analyze exemplary quantum states , including the schrdinger cat states .
the nonclassicality @xmath32 increases with the separation between the classical points defining the cat state .
this growth saturates , if the separation distance is so large that the quantum interference patterns are well isolated from both main peaks of the distributions .
moreover , for a non zero momentum @xmath151 , the quantity @xmath32 undergoes oscillations until the separation distance becomes so large that both packets are separated from the interference patterns .
asymptotically , if the separation is large enough , the indicator @xmath32 does not depend on @xmath52 and tends to a constant value , @xmath152 .
in the case of fock states @xmath153 , the quantity @xmath32 equals zero for the coherent vacuum state @xmath154 and grows monotonically with the quantum number @xmath125 .
if a quantum state is displaced by the glauber operator @xmath94 , the shape of the wigner function and the nonclassicality parameter do not change .
although the squeezing operator @xmath95 changes the shape of the wigner function , our results obtained for the squeezed fock states show that the nonclassicality @xmath32 does not depend on squeezing .
the results presented in this work were obtained for pure states of infinite dimensional hilbert space with use of the standard harmonic oscillator coherent states .
it is worth to emphasize that our approach is also suited to analyze mixed quantum states .
furthermore , one may study the similar problem for quantum states of a finite dimensional hilbert space , which was originally tackled in @xcite .
in such a case one defines the husimi function with the help of the @xmath155 , spin coherent states , while the wigner functions may be obtained by expanding the density matrix in the complete basis of the rotation operators @xcite .
the wigner function for finite dimensional systems may also be defined in alternative ways - see @xcite and references therein
. studying the volume of the negative part of the wigner function , defined according to any of these approaches , one may get an interesting information concerning the nonclassical properties of the state analyzed .
for instance some recent attempts @xcite try to link the negativity of the wigner function with the entanglement of analyzed quantum states defined on a composed hilbert space , or with the violation of the bell inequalities .
we are indebted to i. biaynicki birula , i. bengtsson , j. burgdrfer , a.r.r .
carvalho , e. galvo , a. miranowicz , a.m. ozorio de almeida , j. m. rost , k. rzaewski and m.s .
santhanam for fruitful discussions , comments and remarks .
we also would like to thank p.w .
schleich for helpful correspondence and several remarks that allowed us to improve the manuscript
. financial support by polish ministry of scientific research under the grant no pbz - min-008/p03/2003 and the vw grant entanglement measures and the influence of noise is gratefully acknowledged .
ak gratefully acknowledges the financial support by alexander von humboldt ( avh ) foundation / bonn - germany , under the grant of research fellowship no.iv.4-kam 1068533 stp .
k. banaszek and k. wdkiewicz , acta.phys.slovaca * 49 * ( 1999 ) 491 k. banaszek and k. wdkiewicz , phys.rev.lett.*82 * ( 1999 ) 2009 d. t. smithey , m. beck , m. g. raymer and a. faridani phys
* 70 * ( 1993 ) 1244 ; t. j. dunn , i. a. walmsley and s. mukamel , phys . rev . lett .
* 74 * ( 1995 ) 884 ; g. breitenbach , s. schiller and j. mlynek , nature * 387 * ( 1997 ) 471 ; k. banaszek , c. radzewicz and k. wdkiewicz , phys . rev . * a 60 * ( 1999 ) 674 ch .
kurtsiefer , t. pfau and j. mlynek , nature * 386 * ( 1997 ) 150 ; g. nogues , a. rauschenbeutel , s. osnaghi , p. bertet , m. brune , j. m. raimond , s. haroche , l.g .
lutterbach and l. davidovich , phys .
a 62 , ( 2000 ) 054101 ; a. i. lvovsky , h. hansen , t. aichele , o. benson , j. mlynek and s. schiller , phys
* 87 * ( 2001 ) 0504021 .
p. lougovski , e. solano , z. m. zhang , h. walther , h. mack and p. schleich , phys .
* 91 * ( 2003 ) 0104011 ; w. e. lamb , phys . today * 22 * ( 1969 ) 23 ; k. wdkiewicz , phys . rev . lett . * 52 * ( 1984 ) 1064 ; a. royer , phys.rev.lett.*55 * ( 1985 ) 2745 ; k. banaszek and k. wdkiewicz , phys .
* 76 * ( 1996 ) 4344 . | a measure of nonclassicality of quantum states based on the volume of the negative part of the wigner function is proposed .
we analyze this quantity for fock states , squeezed displaced fock states and cat - like states defined as coherent superposition of two gaussian wave packets .
e - mail : [email protected] [email protected] |
fisher information is the maximum amount of information that can be extracted about an unknown parameter ( @xmath3 in our case ) from a random variable @xmath10 .
it is calculated as the variance of the score .
the score is a measure of the sensitivity of the probability distribution @xmath79 with respect to the parameter @xmath3 , and it is given by @xmath80 in our case , we have two different probability distributions , coming from two detectors , @xmath81\,e^{-q^2/2\sigma^2},\ ] ] where @xmath82 .
the mean value of events collected by each detector is given by @xmath83 and @xmath84 is the total number of measurements , which ( in a lossless system ) is equal to the number of prepared events .
note that the mean of the score vanishes , @xmath85dq\nonumber\\ & = & \frac{\partial}{\partial g}\left[\int p_i(q;g)\,dq\right]=0,\end{aligned}\ ] ] so the final expression for the fisher information of @xmath3 takes the form @xmath86 where @xmath87 ^ 2dq=2n\sigma^2\left[1\pm\left(1 - 4g^2\sigma^2\right)e^{-2g^2\sigma^2}\sin\epsilon\right].\ ] ] the result in eq .
( [ fisher ] ) is general for any value of @xmath3 and @xmath15 , meaning that it is always possible to recover all of the fisher information by using both detectors .
the shift in the difference signal is obtained only under the weak interaction approximation , @xmath88 . to induce the anomalous amplification
is also required the almost - balanced condition , @xmath33 . in this sense ,
our proposed technique lies in the regime of @xmath89 .
now we compare the result of eq .
( [ fisher ] ) to the fisher information when using the wva technique .
the intensity in the dark port under the weak interaction approach is given by @xmath90 and after a straightforward calculation we get the number of detected events and the fisher information , @xmath91\approx n\sin^2(\epsilon/2)+o[g]^2,\\ f^{wva}_g&=&2n\sigma^2\left[1+(1 - 4g^2\sigma^2)e^{-2g^2\sigma^2}\cos\epsilon\right]\approx 4n\sigma^2\cos^2(\epsilon/2)+o[g]^2 , % 4n\sigma^2\cos^2\left(\frac{\epsilon}{2}\right)+6n\sigma^4\cos^2\left(\frac{\epsilon}{2}\right)\cot^2\left(\frac{\epsilon}{2}\right)g^2+ ... \end{aligned}\ ] ] respectively .
an external cavity 795 nm wavelength diode laser and an acoustic optical modulator ( aom ) are used to prepare pulses with an intensity profile @xmath92 .
a glan - taylor calcite polarizer prepares the polarization state in @xmath93 , and a piezo - driven half wave plate ( hwp ) rotating in time by an angle @xmath58 away from the @xmath94 axis changes the polarization of the pulse .
we are interested in estimating the angular velocity @xmath61 of the rotating hwp ( see fig . 2 in the main document ) .
a wollaston prism was used to separate the horizontal ( @xmath95 ) and the vertical ( @xmath96 ) polarization components , which were independently measured using a segmented quadrant detector .
the sum and difference signals take the form @xmath97 assuming that @xmath98 .
these pulses were recorded and the angle @xmath62 and the time shift @xmath99 obtained from numerically fitting the distributions .
we estimated @xmath62 and @xmath61 by using eqs .
( 5 ) and ( 6 ) of the main document as well , nevertheless both estimators gave us identical results within the experimental errors .
a triangle function in voltage with frequency @xmath100 , peak - to - peak voltage @xmath101 and @xmath102 duty cycle as an input for the piezo actuator induced a polarization rotation ramp given by @xmath103 , where @xmath104 is the hwp - mounted piezo response .
gaussian pulses with a peak power of @xmath105w were sent in phase with the ramp .
[ fig : results ] shows the measurement results of @xmath61 for three different values of @xmath100 as a function of @xmath101 .
the technique s estimates show an almost perfect linear response and small error bars even without the need of pulse averaging .
the angle @xmath62 was @xmath63 mrad on average .
the values for the pulse length @xmath106 obtained from numerically fitting the measured sum intensities are 5.377(2 ) ms for 20 hz , 10.695(3 ) ms for 10 hz , and 21.398(7 ) ms for 5 hz . in fig .
[ fig : results ] we use the errors propagated from the time shift measurements ( error of the mean ) which were obtained from the statistics of 60 measured time shifts for each point .
the experimental standard deviations of the measurements of fig .
[ fig : results ] were estimated to be between 20 and 37 times the shot - noise ( calculated from eq .
( 8) in the main document ) . as a function of the peak - to - peak voltage applied to the piezo actuator .
the estimated averaged number of collected photons are kept nearly constant , @xmath107 , by measuring one single pulse when @xmath1085 hz , averaging two pulses when @xmath10810 hz , and averaging four pulses when @xmath10820 hz .
the averaged signal - to - noise ratios are between 29 for @xmath109 v up to 190 for @xmath110 v.,scaledwidth=47.0% ] we were limited by the resolution of the hwp .
we measured @xmath7530 @xmath111rad ( @xmath109 v ) as the smallest reliable total rotation induced by the hwp on the pulses , so we boosted the effective sensitivity of our setup for measuring @xmath61 by decreasing the driving frequency on the piezo , @xmath100 . keeping the peak - to - peak voltage on the piezo at 15 v we set @xmath112 mhz , @xmath113 s and @xmath114 mrad , and measured our lowest reported angular velocity of @xmath115 nrad / s .
our reported net rotation of @xmath116rad in a rotation time of 200 seconds was obtained by averaging sets of five pulses and having 20 statistical points .
our measurements for these long periods were affected by long - term drifts due to temperature and air pressure fluctuations , so 100 consecutive pulses ( out of 360 collected ) were selected such that the angle @xmath62 , the peak laser power and the measured time shift remained relatively constant .
we prepare the system in the state @xmath117 is the peak electric field of the gaussian pulse .
we use a classical description for the long non - fourier - limited pulses of light used in our experiment , but the theory can also be applied to coherent pulses of light or single photons . after passing through the piezo - driven hwp , the polarization is rotated as @xmath118 ^ 2\,e^{-t^2/2\tau^2}\\ & = & \frac{i_0}{2}\left[1-\sin\left(4\phi+4\omega_0t\right)\right]\,e^{-t^2/2\tau^2},\,\mbox{and}\\ i_2&=&\frac{i_0}{2}\left[1+\sin\left(4\phi+4\omega_0t\right)\right]\,e^{-t^2/2\tau^2}.\end{aligned}\ ] ] finally , calculating the sum and difference signals we get @xmath119 where @xmath98 .
note that if we change the basis in equation ( [ initialstate ] ) from @xmath95 and @xmath96 to circularly polarized states , @xmath120 this representation defines an exact mapping of the theoretical protocol presented in fig . 1 of the main document , where both states ( @xmath121 and @xmath122 ) are separated into the two different paths of the interferometer , and there is a relative phase of @xmath123 between both arms .
in fact , the state after the light pulse passes trough the piezo - driven hwp takes the form ( eq .
[ [ finalstate ] ] ) , @xmath124\otimes e_0\,e^{-t^2/4\tau^2}.\ ] ] under this representation , the left and right circularly polarized components of light split into the two arms of the interferometer .
if fourier - limited pulses were used , the quantity @xmath125 could be understood as a relative laser frequency shift between the two paths . | we present a parameter estimation technique based on performing joint measurements of a weak interaction away from the weak - value - amplification approximation .
two detectors are used to collect full statistics of the correlations between two weakly entangled degrees of freedom . without discarding of data ,
the protocol resembles the anomalous amplification of an imaginary - weak - value - like response .
the amplification is induced in the difference signal of both detectors allowing robustness to different sources of technical noise , and offering in addition the advantages of balanced signals for precision metrology .
all of the fisher information about the parameter of interest is collected .
a tunable phase controls the strength of the amplification response .
we experimentally demonstrate the proposed technique by measuring polarization rotations in a linearly polarized laser pulse .
we show that in the presence of technical noise the effective sensitivity and precision of a split detector is increased when compared to a conventional continuous - wave balanced detection technique .
[ [ introduction ] ] introduction + + + + + + + + + + + + anomalous amplification @xcite has been shown to be advantageous for precision metrology .
such an amplification provides a way to increase a signal while decreasing or retaining the technical - noise floor @xcite . as a result
, the sensitivity and precision of measurements limited by technical noise can be effectively improved , facilitating the saturation of the standard quantum limit .
anomalous amplification was first proposed for metrology with the introduction of the weak value ( wv ) of an observable @xcite , and parameter estimation protocols defined after it are usually known as weak - value - amplification ( wva ) techniques .
the wv of an observable is obtained by post - selecting the state of a system after a weak interaction with a meter system . in wva ,
such measurements in the system induce a discarding of data counts in the measurements of the meter .
in addition to the notion that the state of the system is post - selected after the weak interaction , we consider post - selection as the process of selecting and processing desired events , which , for wva , results in discarding data in the meter . due to the interference of the pre- and post - selection states of the system
the wv can take large complex values outside the eigenvalue spectrum of the observable , which defines the anomalous amplification in wva .
discussion about the quantum interpretation of such a phenomenon can be found in refs .
@xcite .
many recent applications of wva for metrology have been done in classical optics , where the interference can be understood using standard wave mechanics @xcite .
strong postselection is necessary for anomalous amplification in wva techniques , but discarding data counts has been the target of criticism , and even considered harmfulfor metrology @xcite .
however , it has been shown , theoretically and experimentally , that the statistical information collected by the measurements is insignificantly reduced because the amplified signal can compensate for the reduced detection flux resulting from postselection @xcite . such a result is possible under an almost orthogonal pre- and post - selection procedure in the system allowing one to collect
_ nearly all _ of the fisher information using only a small subensemble of measurement counts in the meter @xcite .
for example , the shot - noise limit defined by the _ input _ number of photons used in measurements of the small velocity of one of the mirrors of a michelson interferometer can be reached using wva @xcite .
moreover , it was recently shown that by postselecting only 1@xmath0 of the photons , when measuring small optical deflections constrained to intrinsic electronic detector noise , 99@xmath0 of the total available fisher information can be recovered @xcite .
signal amplification while avoiding detector saturation sparked interest in wva as a precision metrological technique several years ago @xcite , and was recently shown to be essential for the technical - noise mitigation advantages @xcite .
we will introduce the possibility of inducing anomalous amplification without the need of discarding data and without any loss of fisher information .
strbi and bruder @xcite recently proposed a precision measurement technique for measuring time delays of light by carrying out a full measurement of the two weakly correlated degrees of freedom : frequency and polarization .
they concluded that even for low - resolution detectors , the scheme is robust against systematic errors and fluctuations in alignments of the experimental setup .
we show that by subtracting the readouts of the two detectors in such a measurement , a wva - like response is obtained in the difference signal .
this anomalous amplification behaviour clarifies and extends the results reported in ref .
@xcite , since besides having the amplification response typical of the wva approach ( without discarding data ) , our protocol adds the benefits of balanced detection for precision metrology .
we show that this technique allows us to recover _ all _ of the fisher information of the estimated parameter .
in addition , it permits the removal of systematic error in the measurement of the shift in the difference signal by tracking the unshifted sum signal as well .
[ [ theoretical - framework ] ] theoretical framework + + + + + + + + + + + + + + + + + + + + + following the formalism as in wva , we describe the unitary evolution of a two - party system as @xmath1 , where @xmath2 is a binary degree of freedom ( qubit ) controlling the encoding of the information about the interaction parameter @xmath3 in the continuous ( meter ) observable @xmath4 , where @xmath5 is the instantaneous interaction parameter between the two parties , and @xmath6 is the averaged measured parameter , where @xmath7 is the interaction time . ] .
in contrast to wva , where a small set of measurements for @xmath4 are taken into account due to postselection , the operator @xmath4 after the interaction is always measured and conditioned to one of two detectors . the proposed procedure is done by preparing the initial global state as the product state @xmath8 , with @xmath9 , and by tracking @xmath10 conditioned to the measurement basis @xmath11 on the qubit / system . here @xmath12 and @xmath13 are the eigenvectors of @xmath2 , where @xmath14 , and projections to the two detectors are labelled by 1 and 2 . the small phase @xmath15 defines the measuring basis on the equatorial plane of the bloch sphere , where the initial prepared state for the qubit @xmath16 also lies .
the probability distributions measured on the detectors take the form @xmath17p(q),\end{aligned}\ ] ] where @xmath18 .
we will use a gaussian state @xmath19 with variance @xmath20 for the preparation in @xmath10 . as an example ,
fig .
[ fig : cartoon ] shows a schematic of the measuring technique for an optical setup , where the observable @xmath2 is represented as the which - path degree of freedom in an interferometer , and the phase @xmath15 controls the interference . under the assumption of a weak interaction ,
i.e. @xmath21 , we can express @xmath22 in an optical setup .
the states @xmath12 and @xmath13 correspond to the two paths in the interferometer , a small unbalacing phase @xmath15 controls the interference , and estimation of @xmath3 is obtained from the sum and difference of the two measured distributions .
the red solid lines represent the directly measured distributions @xmath23 and @xmath24 , and the blue dashed lines are the distribution components as in eq .
( [ distributions]).,scaledwidth=45.0% ] the peak value of the distribution of eq .
( [ approximation ] ) is smaller by a factor @xmath25 and the position of the peak is shifted by an amount @xmath26 with respect to @xmath27 .
the sum and difference of the distributions take the form to refer to post - processing probability distributions .
the distributions @xmath23 and @xmath24 are directly obtained from measurements . ]
@xmath28 after a large number @xmath29 of independent measurements of @xmath10 , @xmath30 on detector 1 and @xmath31 on detector 2 , the sum distribution ( [ sumanddiff1 ] ) reproduces the quantum probability distribution of the input state , and the difference distribution ( [ sumanddiff2 ] ) has a shifted attenuated peak similar to wva .
note that the measured shift is in the difference probability distribution and not in the wave - function @xmath19 itself , as it is the case in wva .
in fact , the weak values for the measurements are given by @xmath32 for @xmath33 , and no ( anomalous ) large weak value is induced .
estimations of averaged values for @xmath15 and @xmath3 under the weak interaction approximation can be obtained as @xmath34\tan\epsilon}{2\sigma^2}\label{g},\end{aligned}\ ] ] where @xmath35 , and @xmath20 is the measured variance of @xmath36 . by making @xmath33 and preparing a large variance input state , a large shift is induced and small values of @xmath3 resolved .
this behaviour is similar to the amplification of the wva technique with an imaginary weak value @xcite , and the quadratic response of the wva postselection probability with respect to @xmath15 , @xmath37 , is replaced with a ( larger ) linear response of the difference signal , @xmath38 in eq .
( [ sumanddiff2 ] ) .
the maximum amount of information about the parameter @xmath3 that can be extracted from the measurements is given by adding the fisher information for both detectors @xcite , @xmath39 where @xmath40 .
eq .
( [ g ] ) is the efficient estimator for @xmath3 , which saturates the fisher information ( [ fisher ] ) in absence of noise .
the inverse of the fisher information is known as the cramer - rao bound or crb .
this bound is the smallest possible variance of an unbiased estimation of the parameter @xmath3 .
the smallest possible standard deviation for measurements of @xmath3 is given by @xmath41 , and any source of noise would increase it .
this result shows the standard quantum limit or shot - noise dependence with respect to @xmath42 , characteristic of @xmath29 independent measurements . [
[ comparison - to - other - techniques ] ] comparison to other techniques + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we now consider two alternative approaches to measure the parameter @xmath3 , and compare them to our proposed protocol .
we note that all three techniques are upper bounded by the same fisher information , @xmath43 . the first approach is formulated by noticing that the unitary evolution corresponds to a translation of @xmath3 in the canonical momentum of @xmath10 , and that the ancillary system @xmath2 is not necessary .
the protocol consists on preparing @xmath29 copies of @xmath19 , applying the evolution @xmath44 , and performing measurements of the conjugate momentum @xmath45 ( instead of @xmath10 ) .
the visibility of the shift , i.e. the ratio of the shift and the standard deviation of the measured state , is given by @xmath46 .
this approach is known as the standard technique when similar comparisons to wva are introduced .
the visibility of our technique takes the form @xmath47 , giving an advantage of @xmath48 for @xmath33 when measuring small values of @xmath3 .
this amplification is an important key on how the wva and our protocol are superior in technical - noise - limited experiments .
the second approach is the conventional wva technique , where measuring @xmath10 , using only one detector , is conditioned to the postselection @xmath49 .
this procedure is equivalent to removing the balancing phase @xmath50 in fig .
[ fig : cartoon ] and tracking only the dark port of the interferometer .
the probability distribution for such a measurement is given by @xmath51 .
so , even though the shift of the peak of our protocol is half the shift of the peak in the wva technique for a given @xmath15 ( for @xmath33 ) , the peak value of @xmath52 is much larger than in @xmath53 since @xmath54 .
this result plus the background noise subtraction characteristic of differencing signals allow us to experimentally induce smaller possible values of @xmath15 than in wva , which offers technically advantageous larger amplification than in wva . the fisher information for the wva technique , under the weak interaction approximation , is given by @xmath55 @xcite .
wva measurements _ asymptotically _ recover all of the fisher information , given by @xmath43 , only in the anomalous weak value regime ( @xmath56 ) @xcite , but the proposed differencing technique collects always _ all _ of the information . wva and the almost - balanced weak values technique offer similar amplification behaviour , and noise mitigation advantages will depend strongly on the metrological specific experimental task .
for example , wva could still be a favourable technique in situations where detector saturation is the predominant limiting factor .
finally , note that both of the above techniques ( standard and wva ) require prior measurements of the unshifted peak and variance of the input distribution @xmath27 .
ours does not , since the sum distribution @xmath36 offers these measurements simultaneously .
this is one of the advantages of using two detectors instead of just one .
[ [ experimental - implementation ] ] experimental implementation + + + + + + + + + + + + + + + + + + + + + + + + + + + as a proof - of - principle of the technique in the optical domain , we performed measurements of small polarization changes in a linearly polarized pulse of laser light . a piezo - driven half - wave plate ( hwp ) played the role of the interferometer in fig . [ fig : cartoon ] , where we used polarization instead of the which - path degree of freedom , and time as the variable @xmath10 ( with @xmath57 ) .
the piezo actuator rotated the hwp in time by an angle @xmath58 , which defined the tunable phase as @xmath59 , and the angular velocity of the rotating hwp as the parameter of interest , @xmath60 ( see fig . [
fig : setup2 ] and supplemental material for experimental details ) . of a piezo - driven half wave plate ( hwp ) .
the angular rotation induces changes of polarization in the laser field.,scaledwidth=45.0% ] the estimates for @xmath61 using a split detector showed an almost perfect linear response and small standard deviations even without the need of pulse averaging .
the angle @xmath62 was @xmath63 mrad on average .
this unbalanced phase is _ significantly smaller _ than previously reported postselection angles using wva techniques , proving the possibility of larger amplification under our proposed protocol . besides the two ( standard and wva ) experimental techniques mentioned before , we compare here our protocol to a conventional balanced optical technique for measurements of @xmath61 , where no ancillary system nor measurements of peak shifts in a distribution are required . by replacing the gaussian pulse with a continuous wave ( c.w . )
beam , the perfectly balanced ( @xmath64 ) signal takes the form @xmath65 and the value @xmath61 can be recovered .
the measured intensities are very small compared to our proposed technique , since @xmath66 .
thus , the technical advantages of our technique rely principally on time shift measurements of a gaussian profile with controllable amplitude ( @xmath67 ) , instead of measuring very small voltage amplitudes as it is the case of the c.w . conventional balanced technique . for example , in order to obtain a visible signal ( signal - to - noise ratio slightly larger than one ) using a c.w
. beam , 1024 periods of integration time were required .
our technique gives a better signal - to - noise ratio ( @xmath68 29 ) even without the need of pulse averaging @xcite .
the best possible variance of our measurements are defined by the crb for @xmath61 . following equation ( [ fisher ] ) , @xmath69 where @xmath70 is the number of averaged pulses .
the experimental standard deviations of the measurements with the split detector were estimated to be between 20 and 37 times the calculated ones using eq .
( [ crb ] ) .
such deviation is small considering that no frequency filters , lock - in amplifiers or any other electronic processing device was used . in order to have good photon number statistics and a clearer raw signal , the split detector of fig .
[ fig : setup2 ] was replaced with two single photon counting modules ( spcm ) .
the input power was attenuated before the acoustic optic modulator ( aom ) so that the peak detected photon rate was slightly smaller than @xmath71 counts / s on each spcm .
fig .
[ fig : photoncounting ] shows the standard deviation on the estimation of @xmath61 as a function of the number of averaged pulses @xmath70 , and its comparison to eq .
( [ crb ] ) .
the system follows the @xmath72 expected shot - noise behaviour , and it has ( inset of fig .
[ fig : photoncounting ] ) a precision close to the shot - noise limit .
there is a slight decrease in the relative precision ( @xmath73 ) with the increase on the number of averaged pulses , which is due to long - term drift @xcite .
fig .
[ fig : photoncounting ] shows results for a total collection time of up to 5000 sec ( 50@xmath74100 pulses ) and a factor of no larger than @xmath751.6 away from the shot - noise limit , in comparison to a factor of @xmath751.2 for shorter acquisition times . as a function of the number of pulses averaged , @xmath70 .
pulses with @xmath76 s and @xmath77 were used . the driving signal on the piezo - mounted hwp was 60 v peak - to - peak at @xmath78 hz . each point ( blue ) was obtained from the statistics of 100 measurements , and the solid line ( red ) is given by the shot - noise ( eq . [ crb ] ) .
the inset shows the ratio between the experimental standard deviation and the shot - noise.,scaledwidth=40.0% ] [ [ conclusions ] ] conclusions + + + + + + + + + + + we have introduced a metrological technique for measuring small interaction parameters based on _ almost _ balancing both signals of a joint measurement . even though the protocol does not require the postselection of wva techniques , anomalous amplification is recovered . a small phase controlling the unbalancing of the difference signal plays a similar role to the postselection angle in wva . by subtraction of the signals ,
background noise floor is removed and also a linear response with respect to the small phase @xmath15 is obtained . such linear response is larger than the quadratic response of wva techniques , and it is the principal advantage of our technique by allowing larger amplification ( smaller @xmath15 ) in technical - noise - limited experiments than in wva techniques . in addition
, no prior knowledge of the peak and the variance of the distribution of the input state is required , so systematic error - free estimation of the parameter of interest is possible .
we believe our technique will find interesting applications on precision metrology , within and outside optical systems .
the use of post - selection required for wva has been challenging in experimental set - ups outside of optics .
our proposed almost - balanced weak values technique overcomes the necessity of post - selection and might have direct applications in non - optical experiments .
anomalous amplification in our differencing technique allows us to collect all of the fisher information that wva techniques collect only in the optimal asymptotic limit of a large / anomalous weak value .
this result proves that anomalous amplification can be obtained from the weak system - meter interaction assumption alone , where strong measurements in the system condition the reading in the meter .
technical advantages of our proposed technique and wva over standard techniques depend upon such a coupling to an ancillary system .
a recent related work reported experimental advantages of the original proposal of ref .
@xcite with respect to the wva protocol when estimating ultra - small phases @xcite .
the authors acknowledge andrew jordan and justin dressel for valuable suggestions .
j.m.r . thanks bethany little for a thorough reading of the manuscript .
this work was supported by the army research office , grant no .
w911nf-12 - 1 - 0263 , and by the national natural science foundation of china grant no .
11374368 and the china scholarship council .
29ifxundefined [ 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.60.1351 [ * * , ( ) ] link:\doibase 10.1103/physrevx.4.011031 [ * * , ( ) ] link:\doibase 10.1103/physreva.92.032127 [ * * , ( ) ] link:\doibase 10.1103/physrevd.40.2112 [ * * , ( ) ] link:\doibase 10.1103/physreva.76.044103 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.113.200401 [ * * , ( ) ] link:\doibase 10.1103/physreva.91.032116 [ * * , ( ) ] @noop ( ) , link:\doibase 10.1103/physreva.81.033813 [ * * , ( ) ] http://stacks.iop.org/1367-2630/15/i=7/a=073022 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.112.040406 [ * * , ( ) ] link:\doibase 10.1103/physreva.89.052117 [ * * , ( ) ] @noop ( ) , link:\doibase 10.1103/physreva.80.041803 [ * * , ( ) ] link:\doibase 10.1364/ol.38.002949 [ * * , ( ) ] link:\doibase 10.1103/physreva.85.062108 [ * * , ( ) ] link:\doibase 10.1007/s40509 - 014 - 0002-x [ * * , ( ) ] link:\doibase 10.1103/physreva.91.062107 [ * * , ( ) ] link:\doibase 10.1126/science.1152697 [ * * , ( ) ] , link:\doibase 10.1103/physrevlett.102.173601 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.010405 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.110.083605 [ * * , ( ) ] @noop @noop ( ) , * supplemental material * |
the effect of spin - orbit coupling ( soc ) on strongly correlated electron systems has demonstrated to be the origin of novel phases @xcite , like spin liquids @xcite , unconventional magnetism @xcite , and topological phases @xcite , among others .
also , it has been demonstrated that for an intermediate strength regime , soc can work together with coulomb interactions to enhance the electron correlations , leading to spin - orbit driven mott insulators @xcite .
the key point is that small changes in these interactions can lead to an enormous variety in the electronic and magnetic behaviors . specifically in @xmath12 transition metals like ir
the soc is significant , and comparable to the atomic interactions like crystal electric field @xmath13 and the on - site coulomb interaction @xmath14 .
examples for this scenario are the pyrochlore iridates , which show novel phases like weyl semimetals @xcite , topological mott insulators @xcite , and axion insulators with unusual electromagnetic responses @xcite .
the most common oxidation states of iridium are 3 + and 4 + . on the other hand oxidation states of 5
+ and 6 + are rare and sometimes poorly characterized @xcite .
the double perovskite structure r@xmath0mm@xmath15o@xmath1 may host iridium on the m and/or m@xmath15 position ; ir will be in octahedral coordination in this structure type .
hence , the octahedral crystal electric field splits the @xmath12 levels in a @xmath16 triplet and a @xmath17 doublet .
then , for ir@xmath18 ( 5d@xmath19 electronic configuration ) the large soc acts on the @xmath16 levels , splitting them into an effective @xmath20 doublet and an effective @xmath21 quartet @xcite . according to the strong spin - orbit coupling j@xmath4 model , a non - magnetic @xmath22 ground state
is expected for double perovskites with ir@xmath2 ( 5d@xmath3 ) as the formal oxidation state .
that is the case for the monoclinic sr@xmath0yiro@xmath1 , synthesized for the first time by wakeshima et al .
@xcite in 1999 with the aim of studying the magnetic properties of @xmath12 transition metal oxides in which the metal ions are in an unusual oxidation state . as expected , no transition into a long - range ordered state was observed .
however , recently a report by cao et al . on this system
@xcite reports that a strong non - cubic crystal field ( due to the flattening of the iro@xmath1 octahedra ) together with an `` intermediate - strength '' spin - orbit coupling , leads to a different ground - state configuration , i.e. an antiferromagnetic ground state witht@xmath23 = 1.3 k. the emergence of magnetic ordering was evidenced in magnetization studies at 7 t as well as in specific heat measurements in different applied magnetic fields . for the latter characterization ,
a small anomaly in the low temperature region was interpreted as a signature of long - range magnetic ordering .
recent first - principles calculations @xcite on both sr@xmath0yiro@xmath1 and ba@xmath0yiro@xmath1 compounds gave another explanation for this behavior .
in ref @xcite the authors argue that the breakdown of the j = 0 state in sr@xmath0yiro@xmath1 would be due to a band structure effect rather than the non - cubic crystal field effect . however , a posterior report by pajskr et al .
@xcite combining several numerical and semianalytical methods to study the band structure of both compounds contradicts the conclusions of refs . @xcite and @xcite .
they found that monoclinic sr@xmath0yiro@xmath1 and cubic ba@xmath0yiro@xmath1 are quite similar , and that both exhibit no tendency towards a magnetic phase transition at low temperatures .
more recent works on polycrystalline samples of the series ba@xmath5sr@xmath6yiro@xmath1 @xcite did not observe any signature for long - range magnetic order in this material in their magnetic susceptibility characterization , in strong contrast with cao et al @xcite . on the contrary ,
these reports reinforce the notion of a non - magnetic ground state dominated by strong soc . in both reports @xcite the magnetic characteristics of the system do not change significantly across the series , and the effective magnetic moment per ir is much less than the value expected for a s = 1 system as reported by cao et al .
@xcite , demonstrating a strongly soc dominated ground state . in this work , in order to settle the dispute described above , we present a combined structural , magnetic and thermodynamic characterization of sr@xmath0yiro@xmath1 single crystals .
structure and composition of the crystals were thoroughly characterized by single crystal x - ray diffraction ( xrd ) , synchroton powder xrd , energy dispersive x - ray spectroscopy ( edx ) and scanning electron microscopy ( sem ) .
the single crystal xrd data show evidence of a cubic supercell .
we particularly analyze the specific heat of this material , in order to study the anomaly reported in @xcite . in our studies
no long - range magnetic order was found even in fields up to 9 t. the magnetic contribution to the specific heat was extracted , finding a schottky anomaly due to a small amount of magnetic impurities .
further analysis suggests non - negligible spin correlations , which nonetheless , are not associated with long - range magnetic ordering .
single crystals of sr@xmath0yiro@xmath1 were flux - grown using srco@xmath24 ( alfa aesar , 99.994% ) , y@xmath0o@xmath24 ( alfa aesar , 99.99% ) and iro@xmath0 ( alpha aesar , 99.99% ) as the precursor materials .
anhydrous srcl@xmath0 ( alpha aesar , 99.5% ) was used as the flux .
the stoichiometric amount of starting materials and the flux were mixed using a nutrient to solvent weight ratio 1:13 .
all constituents were put in a platinum crucible ( 50 cubic ml ) with a platinum lid .
the crucible edge and lid were squeezed to semi - seal the crucible assembly .
the set - up was heated to different high temperatures ( 1200@xmath25c , 1250@xmath25c and 1300@xmath25c ) , held for 24 h and then cooled to 900@xmath25c with different cooling rates ( 0.5@xmath25c / h , 1@xmath25c / h and 2@xmath25c / h ) . the furnace was switched off afterwards and cooled to room temperature .
the crystals were found at the bottom of the crucible and were separated by dissolving the flux in water .
after that , in order to gain a better understanding of the growth , the sr@xmath0yiro@xmath1 crystals were examined by optical microscopy .
as depicted in figure s1 @xcite , crystals have different morphologies : cubic - like or irregular polyhedral .
the cubic ones correspond to the pure sr@xmath0yiro@xmath1 phase and the irregular ones seem to be cubic crystals inter - grown with each other due to additional nucleation sites .
all our attempts to avoid the co - crystallization of irregular shaped crystals by varying the hold temperatures , cooling rates and material to flux ratio have been failed . for the following studies
, we used the samples grown by cooling down from 1200@xmath25c to 900@xmath25c with a rate of 2@xmath25c / h and material to flux ratio of 1:13 .
a bunch of single crystals of sr@xmath0yiro@xmath1 with different morphologies were selected and mounted on the tip of a glass fiber for x - ray diffraction ( xrd ) .
intensity data were collected at room temperature using @xmath26 scans on a stoe imaging plate diffraction system ( ipds - ii ) with graphite - monochromatized mo k@xmath27 radiation ( @xmath28 = 0.71069 ) operating at 50 kv and 40 ma using a 34 cm diameter imaging plate .
individual frames were collected with a 4 min exposure time and a 0.8@xmath29 rotation .
x - area , x - red , and x - shape software packages @xcite were used for data collection , integration , and analytical absorption corrections , respectively .
shelxl @xcite and jana2006 @xcite software packages were used to solve and refine the structure .
high - resolution synchrotron powder diffraction data were collected at the beamline p02.1 at the storage ring petra iii ( desy , hamburg , germany ) using an average wavelength of 0.2067 (@xmath30 60 kev ) , with a relative energy bandwidths @xmath13e / e of the order of 10@xmath31 .
the repeatability of the detector translations along the beam was tested in a series of exposures of a lab@xmath1 standard ( nist 660a ) filled into a 0.8 mm - diameter capillary . during this measurement ,
the detector was moved forward and backward ( with and without backlash ) over the entire travel range in steps of 125 mm @xcite .
the synchrotron powder xrd pattern was collected on crushed single crystals of sr@xmath0yiro@xmath1 sealed in a 0.8 mm - diameter capillary .
data were fitted by the rietveld method @xcite using fullprof in the winplotr program package @xcite .
the microstructural and compositional analysis was performed using scanning electron microscopy ( sem , zeiss evoma15 ) along with an electron microprobe analyzer for semi - quantitative elemental analysis in the energy dispersive x - ray ( edx ) mode ( x - maxn20 detector from oxford instruments with a aztecenergy advanced acquisition and edx analysis software ) .
the magnetization as a function of temperature ( in the range 0.43 - 300 k ) and magnetic field was obtained for randomly oriented single crystals of sr@xmath0yiro@xmath1 ( @xmath30 92.5 mg ) using a superconducting quantum interference device ( squid ) magnetometer from quantum design , equipped with an _
ihelium3 _ option .
a thorough background subtraction was performed for all the curves .
specific heat measurements were performed on 29 single crystals(@xmath30 5.6 mg ) between 0.4 k and 10 k using a heat - pulse relaxation method in a physical properties measurement system ( ppms ) from quantum design .
the heat capacity of the sample holder ( addenda ) was determined prior to the measurements for the purpose of separating the heat capacity contribution of the sample from the total heat capacity .
sr@xmath0yiro@xmath1 orders in the double perovskite type structure with the general formula r@xmath0mm@xmath15o@xmath1 , wherein r denotes an alkaline - earth ( or rare - earth ) metal and m and m@xmath15 are d - block elements ( or other metals ) .
double perovskites crystallize in a cubic , tetragonal , or monoclinic symmetry with interpenetrating m and m@xmath15 face - centered cubic ( fcc ) sublattices .
we started the analysis of sr@xmath0yiro@xmath1 single crystal xrd data using the reported monoclinic structural model @xcite , with the space group p2@xmath32/n and lattice parameters:@xmath8 = 5.7826(5 ) , b = 5.7830(5 ) , c = 8.1746(7 ) , @xmath33 = 90.036(7)@xmath25 .
the single crystal xrd refinement leads to the agreement factors of r@xmath34 = 4.17 % ( for i @xmath35(i ) ) andr@xmath36 = 9.19 % .
( color online ) structural details for sr@xmath0yiro@xmath1 crystals : ( a ) perspective view of the double perovskite structure of the monoclinic subcell along the crystallographic b axis .
( b ) reconstructed precession photographs for sr@xmath0yiro@xmath1 crystal in the @xmath37 layers show extra reflections , marked with red circles , required to describe a @xmath38 supercell .
( c ) perspective view of the double perovskite structure of the @xmath39 cubic supercell along the crystallographic c axis .
( d ) comparison between bond lengths and bond angles of the iro@xmath1 and yo@xmath1 octahedra , monoclinic vs cubic . ] after an anisotropic refinement of all atoms , a difference electron density fourier map calculated with phases based on the final parameters shows maximum and minimum residual peaks of + 2.03 and -2.58 e@xmath40 , respectively .
a detailed examination of reconstructed precession images of the @xmath37 layer ( figure [ structure](b ) ) shows strong diffraction maxima due to the small subcell and additional weak peaks forming distinctive groups of four reflections around the allowed peaks .
all those extra peaks can be indexed with the @xmath41 supercell , where @xmath8 , @xmath9 and @xmath10 are the unit cell dimensions of the monoclinic subcell .
the integrated data with the @xmath42 supercell suggests that sr@xmath0yiro@xmath1 crystallizes in the cubic space group @xmath43 with lattice parameter , @xmath8 = 8.1773(7 ) .
the crystal structure of the supercell based on our refinement , shown in figure [ structure](c ) , is built up of corner - sharing iro@xmath1 and yo@xmath1 octahedra units , which feature an ordered rock - salt like arrangement , and the sr placed in between .
please note that the ir@xmath2 ions in this structure build up a fcc network , a general feature of the double perovskite type structure .
the single crystal xrd refinement by taking into account the supercell shows excellent agreement factors of r@xmath34 = 1.65 % ( for i @xmath35(i ) ) and r@xmath36 = 3.79 % and less electron density difference after an anisotropic refinement of all atoms , i.e. maximum and minimum residual peaks + 1.06 and -0.98 e@xmath40 , respectively . the parameters for data collection and the details of the structure refinement are given in table [ table1 ] .
atomic coordinates , thermal displacement parameters ( u@xmath44 ) and occupancies of all atoms are given in table [ table2 ] .
anisotropic displacement parameters ( adps ) and selected bond lengths and angles are given in the supplemental material , tables s1 and s2 @xcite .
.[table1]crystal data and structure refinement for sr@xmath0yiro@xmath1 ( a ) subcell and ( b ) supercell at 293 k. [ cols="^,^,^",options="header " , ] in the sr@xmath0yiro@xmath1 subcell , the monoclinic distortion is very small , as presented in figure [ structure](d ) , with @xmath33 angles very close to 90@xmath25 ( @xmath45 ) and unequal ir - o bond lengths with a small difference of 0.03 , while @xmath46o - ir - o deviates from 90@xmath25 by 2.4@xmath25 , and the iro@xmath1 octahedra are rotated / tilted with @xmath46ir - o - y of 162.42@xmath25 . for our sr@xmath0yiro@xmath1 supercell , the iro@xmath1 octahedra are completely regular with equal ir - o bond lengths , @xmath46o - ir - o= 90@xmath25 and @xmath46ir - o - y = 180@xmath25 .
our experimental results on single crystal xrd are in good agreement with the theoretical paper reported by pajskr et al .
@xcite , which claims that the non - cubic crystal field of sr@xmath0yiro@xmath1 is found to be rather weak and that sr@xmath0yiro@xmath1 is quite similar in terms of structure and low - energy properties to the cubic ba@xmath0yiro@xmath1 analog @xcite .
the results of our rietveld refinement , agreement factors , and refined lattice constants of synchrotron powder xrd ( @xmath28 = 0.2067 ) patterns of the sr@xmath0yiro@xmath1 sample studied in this work are shown in the supplemental material @xcite .
figure s2 shows the experimental synchrotron xrd data , the simulated and residual intensities as well as the corresponding bragg positions using the monoclinic structure model . a tiny trace of unreactedy@xmath0o@xmath24 ( @xmath30 1.06 % ) is observed .
the analysis of the structural data reveals that sr@xmath0yiro@xmath1 crystallizes in monoclinic p2@xmath32/n space group with lattice parameters : @xmath8 = 5.7870(7 ) , b = 5.7912(6 ) , c = 8.1805(1 ) , @xmath33 = 90.214(6)@xmath25 and r@xmath47 = 2.11 , r@xmath48 = 2.01 and reduced @xmath49 = 2.32 .
the result of the rietveld refinement of the synchrotron xrd data using the cubic structural model discovered by single crystal xrd is depicted in figure s3 .
please note that both the monoclinic and the cubic model describe the synchrotron powder xrd data reasonably well with a slightly better fit by the monoclinic model ( bragg r - factor 2.11 for the monoclinic vs 3.46 for the cubic one ) . however , this difference is not significant since the larger number of fit parameters for the monoclinic model will yield better fit results as such .
hence , is very difficult to decide which model , monoclinic or cubic , is correct using powder diffraction data only .
we emphasize that single crystal diffraction data is required to find the cubic @xmath50 supercell .
figure [ sem ] illustrates an sem image of the sr@xmath0yiro@xmath1 single crystal , collected in backscattered electron ( bse ) mode .
edx analysis was performed on various spots and on several crystals .
the analysis suggests that the crystals are chemically homogeneous with a stoichiometry close to the target composition , within the accuracy of the edx method .
sem image of a typical sr@xmath0yiro@xmath1 single crystal in the backscattered electron ( bse ) mode . ]
( color online ) .
the magnetic susceptibility of sr@xmath0yiro@xmath1 as a function of temperature in an external magnetic field @xmath51h = 0.5 t , along with its fit according to a curie - weiss ( cw ) law in the temperature range 20 - 300 k ( left axis ) . on the right axis , the inverse susceptibility ( after subtracting @xmath52 obtained from the cw fit ) is plotted . ]
the magnetic susceptibility in an external magnetic field of @xmath51h = 0.5 t is shown in fig .
[ chixt ] in the temperature range 1.8 -300 k , with no evidence of long - range magnetic order .
the data obeys the curie - weiss law @xmath53 in the temperature range 20 - 300 k , which gives a temperature independent susceptibility contribution @xmath54 emu / mol oe , an effective magnetic moment @xmath55/ir , and a curie - weiss temperature@xmath56 k. due to the @xmath57 electronic configuration of sr@xmath0yiro@xmath1 , a van vleck contribution to the susceptibility is expected : @xmath58 , where @xmath59 comes from the diamagnetic contribution of the core levels and @xmath60 is the van vleck paramagnetic susceptibility .
the diamagnetic susceptibility can be obtained by adding the diamagnetic contribution of all the individual ions , leading to @xmath61 emu / mol oe , and finally @xmath62 emu / mol oe , which is in agreement with similar @xmath57 compounds @xcite .
it is worth to notice that the obtained effective magnetic moment @xmath55/ir is much smaller than the value 2.38@xmath63/ir expected for a conventional s = 1 5d - electron system as mentioned by cao et al .
@xcite , but it is in close agreement with the value @xmath64/ir reported for polycrystalline sr@xmath0yiro@xmath1 @xcite . as will be shown below , the paramagnetic signal can be attributed to a small percentage of paramagnetic centers ( less than 1% ) , so it is likely that the curie - tail in the magnetic susceptibility stems from paramagnetic impurities in the sr@xmath0yiro@xmath1 sample , only .
( color online ) .
temperature dependence of the magnetic susceptibility in the low temperature region for different magnetic fields ( zfc results ) .
the inset shows a zoom of the 5 t data down to 0.43 k. ] in view of the antiferromagnetic transition at t@xmath23 = 1.3 k in 7 t magnetic field which has been observed in ref .
@xcite , we performed further magnetization studies at low temperatures and at high magnetic fields . from our experimental data no magnetic transition is detected down to 0.43 k , even for applied magnetic fields up to 5 t ( see fig.[mxt ] ) @xcite .
( color online ) .
isothermal magnetization curve of sr@xmath0yiro@xmath1 at t = 0.48 k. the red line shows the fit according to the brillouin function ( for details see text ) . ] the magnetization of sr@xmath0yiro@xmath1 as a function of applied magnetic field is shown in fig .
[ brillouin ] for t = 0.48 k. in agreement with the temperature dependent measurements , the magnetic behavior at lowest temperatures is predominantly paramagnetic , and the curve can be fitted with the modified brillouin function : @xmath65 in this expression @xmath66 represents a scaling factor which gives the percentage of paramagnetic ir in the sample , @xmath67 is the avogadro constant , @xmath68 the land factor , @xmath63 the bohr magneton , @xmath69 the total angular momentum , and @xmath70 the boltzmann constant . from the fit the parameters g = 2.1 , and @xmath71 emu / mol oe are found , together with a low percentage of paramagnetic centers ( @xmath72 ) of @xmath73%
. these paramagnetic centers could be defects in the sample stemming from oxygen vacancies , chemical disorder ( y / ir site mixing ) , off - stoichiometry , etc .
a possible scenario could be the presence of ir@xmath18 or ir@xmath74 ions created by intermixing and/or off - stoichiometry @xcite .
the temperature independent susceptibility value @xmath52 is slightly larger than the value obtained from the curie - weiss fitting fig .
[ chixt ] .
this discrepancy could be due to non - negligible spin correlations at low temperature which are not included in eq .
[ brillouin ] , or to a temperature - dependent contribution to the van vleck susceptibility .
it is worth to notice that no metamagnetic features arise in the field dependence of the magnetization in contrast to ref .
@xcite , where a sharp metamagnetic transition has been observed at the critical field h@xmath75 2.6 k at t = 0.5 k. ( color online ) .
temperature dependence of the specific heat of sr@xmath0yiro@xmath1 for different magnetic fields , shown in the same scale as @xcite .
the inset shows a zoom into the low - temperature region . ] the low temperature specific heat in different magnetic fields is shown in fig .
[ cp_all ] .
a small anomaly is observed below t @xmath30 5 k in all the curves , including the zero - field data , and which becomes more pronounced in applied magnetic fields of 1 , 3 , and 5 t. this peak shifts to higher temperature and broadens with increasing magnetic fields , which in principle suggests its magnetic origin
. however , its magnitude is small , and instead of being a sharp , @xmath28-like anomaly , pointing towards a second - order phase transition , it resembles more a hump - like feature .
this makes it difficult to identify the anomaly as a real magnetic phase transition . in ref .
@xcite , cao et al .
detect a similar anomaly for their sr@xmath0yiro@xmath1 single crystals , in qualitatively and semi - quantitative agreement with the one observed here . in @xcite ,
the anomaly in the specific heat was interpreted as the onset of long - range magnetic order , due to its proximity to the ordering temperature t@xmath23 = 1.3 k observed as a sharp peak in the magnetization as function of temperature . in order to settle this open question , we performed additional specific heat measurements on the non - magnetic cubic analog compound ba@xmath0yiro@xmath1 @xcite .
since we observe striking similarities in our specific heat data for sr@xmath0yiro@xmath1 and ba@xmath0yiro@xmath1 @xcite , i.e. a broad anomaly below @xmath76 @xmath30 5 k , it is rather improbable that the specific heat anomaly in sr@xmath0yiro@xmath1 marks a transition into a magnetically long - range ordered state .
( color online ) .
zero - field specific heat of sr@xmath0yiro@xmath1 .
the red line represents the lattice specific heat obtained using a power - law fitting in the range of 6 k @xmath77
t @xmath77 10 k marked by arrows ; for details see text . ] to gain a deeper insight into the origin of this anomaly , the lattice contribution was modelled for 6 k @xmath77 t @xmath77 10 k using the approximation @xmath78 in order to disentangle possible magnetic contributions in sr@xmath0yiro@xmath1 ( fig .
[ lattice ] ) .
the extracted magnetic contribution shows a broad anomaly . from the fit parameters @xmath79 j / mol k@xmath3 and @xmath80 j / mol k@xmath81 ,
the debye temperature @xmath82 k was calculated .
this value is comparable with those obtained for other double - perovskite materials @xcite and iridium - based compounds @xcite . considering that the fitting range was small and chosen to follow the best - fitting procedure in combination with sufficient statistics ,
the extracted magnetic contribution to the specific heat has only a semi - quantitative character , i.e. the assumption @xmath83 = 0 for@xmath76 @xmath84 6 k has been applied .
thus , in order to model the magnetic contribution to @xmath85 in the following a subtraction of the zero - field data from the corresponding field data was performed , yielding the field - dependent magnetic specific heat .
the subtracted data are shown in fig .
[ schottky](a ) .
the shape of the @xmath86 curves resembles a two - level schottky anomaly , shifting to higher temperature with increasing magnetic field .
( color online ) .
( a ) temperature dependence of the low temperature specific heat divided by temperature for different magnetic fields , after subtraction of the zero - field data .
( b ) experimental data from ( a ) represented with a constant offset of 0.03 j / mol k@xmath87 for clarity .
the red lines represent the fits according to a two - level schottky anomaly ( see text ) . ] consequently , the magnetic contribution shown in fig . [ schottky](a ) was modelled as the subtraction of two schottky anomalies , i.e. @xmath88 . following this reasoning ,
the peaks were fitted in the range 0.4 k @xmath77 t @xmath77 6 k with the following expression : @xmath89,\ ] ] where @xmath66 is the concentration of paramagnetic impurities , @xmath90 is the energy separation between the two levels in zero magnetic field , and @xmath91 represents the zeeman splitting in an applied magnetic field @xmath92 .
the fits are shown in the fig .
[ schottky](b ) . according to the fit
, a value @xmath90 = 0.25 k was found for the gap at zero field together with @xmath93 % , indicating that the anomaly is only based on a small amount of paramagnetic impurities , instead of being due to a real long - range magnetic ordering of all iridium atoms in sr@xmath0yiro@xmath1 .
while the fits reasonably describe our data for h @xmath77 5 t , a slightly different behavior is found for h @xmath84 7 t with a lacking peak at low temperatures in our experimental data .
this suggests that even when a schottky anomaly can explain the observed peaks for h @xmath77 5 t ( low fields ) , for higher fields the magnetic contribution is altered . following the discussion about the brillouin fitting in fig .
[ brillouin ] , this could point towards field - induced changes , possibly due to non - negligible spin correlations from correlated magnetic impurities which , nonetheless , are not associated with long - range magnetic ordering , but with short - range correlations only .
the possibility of magnetic correlations due to impurities was already pointed out for polycrystalline ba@xmath5sr@xmath6yiro@xmath1 @xcite and ba@xmath0yiro@xmath1 single crystals @xcite .
( color online ) .
magnetic field dependence of the energy gap between the ground state doublet for paramagnetic impurity spins in sr@xmath0yiro@xmath1 . the red line and the @xmath68 value are the fitting results using @xmath94 . ]
we emphasize the fact that both the paramagnetic susceptibility and the schottky two - level anomaly can be ascribed to the presence of correlated impurity spins ir@xmath18 or ir@xmath74 . with this in mind ,
the zeeman splitting between the two lowest energy levels obtained from the former fits can be plotted as a function of applied magnetic field ( see fig . [ gap ] ) . from @xmath95
the land @xmath68 factor was found to be @xmath68 = 2.1(4 ) , which is consistent with the @xmath68 factor obtained from the brillouin fit , @xmath68 = 2.1(0 ) .
( color online ) .
zero - field magnetic specific heat plotted as c@xmath96/t vs t ( left scale ) together with the magnetic entropy ( right scale ) . ] finally , we can evaluate the magnetic entropy @xmath97 for our samples by integrating the zero - field specific heat data from fig .
[ lattice ] according to the expression : @xmath98 the resulting curves for c@xmath96/t and s@xmath99(t ) are shown in fig .
[ entropy ] . the small value @xmath100 j / mol k is of the same order of magnitude as reported by cao et al .
@xcite(@xmath101 j/ mol k ) .
this non - zero value again points to the presence of paramagnetic impurities , which need to be taken into account . for a @xmath102 = 1/2 ground state of sr@xmath0yiro@xmath1 as claimed in ref .
@xcite , a total entropy @xmath103 j / mol k is expected , which is evidently far from the experimental observations . nevertheless ,
considering the small and finite number of @xmath66 impurities in our samples with presumably @xmath72 contributing to the magnetic entropy , we have @xmath100 j / mol k = @xmath104 , and therefore @xmath105 .
overall , our magnetic and thermodynamic characterization is consistent with a non - magnetic ground state of sr@xmath0yiro@xmath1 with the presence of a small amount of correlated @xmath106 paramagnetic centers . from our results , it is clear that the low temperature specific heat anomalies are not related with the onset of long - range magnetic order .
following a recent matter of debate as to the evolution of magnetism in monoclinic sr@xmath0yiro@xmath1 , we have grown single crystals of sr@xmath0yiro@xmath1 using anhydrous srcl@xmath0 flux .
our single crystal xrd results shows for the first time the presence of a @xmath107 supercell , where @xmath8 , @xmath9 and @xmath10 are the unit cell dimensions of the monoclinic subcell , highlighting the cubic supercell of this compound .
the magnetic susceptibility revealed a predominantly paramagnetic behavior in the temperature range 0.4 k @xmath108 t @xmath77 300 k with an effective magnetic moment @xmath55/ir and non - negligible but small spin correlations . from the brillouin fit for the isothermal magnetization curve at t = 0.48
k a low percentage of @xmath72 impurities were identified as the source for the observed paramagnetism .
no long - range magnetic ordering was observed down to 430 mk .
a thorough study of the specific heat was carried out , in which a broad anomaly in the low temperature region t @xmath77 5 k was identified as a schottky anomaly caused by paramagnetic impurities present in the sample , which are of the same order of magnitude ( @xmath109 % ) as obtained from magnetization measurements .
our results are in strong contrast with the reported ones by cao et al .
@xcite , but in full agreement with recent reports on polycrystalline samples @xcite , pointing towards a non - magnetic ground state with the presence of a small amount of correlated @xmath72 impurities .
the authors would like to thank m. vojta , d. efremov , j. van den brink and j. wosnitza for fruitful discussions .
this work has been supported by the deutsche forschungsgemeinschaft dfg under sfb 1143 and the emmy - noether program ( grant no .
wu595/3 - 3 ) .
10 j. g. rau , e. kin - ho lee and hae - young kee , annu .
rev . condens .
. 7:2:1 - 2.27 ( 2016 ) w. witczak - krempa , g. chen , y. baek kim and l. balents , annu . rev .
matter phys .
5:57 - 82 ( 2014 ) a. nag , s. middey , s. bhowal et al .
116 , 097205 ( 2016 ) s. chikara , o. korneta , w. p. crummett , l. e. delong , p. schlottmann , and g. cao , phys .
b 80 , 140407(r ) ( 2009 ) m. ge , t. f. qi , o. b. korneta , d. e. de long , p. schlottmann , w. p. crummett and g. cao , phys . rev .
b 84 , 100402(r ) ( 2011 ) y. chen and h. y. kee , phys .
b 90 , 195145 ( 2014 ) g. jackeli and g. khaliullin , phys .
102 017205 ( 2009 ) d. pesin and l. balents , nat .
phys . 6 , 376 ( 2010 ) x. wan , a. m. turner , a. vishwanath , s. y. savrasov , phys . rev .
b 83 , 205101 ( 2011 ) w. witczak - krempa and y. b. kim , phys . rev .
b 85 , 045124 ( 2012 ) a. a. burkov and l. balents , phys .
107 , 127205 ( 2011 ) w. witczak - krempa , t. p. choy , y. b. kim , phys . rev .
b 82 , 165122 ( 2010 ) m. kargarian , j. wen , g. a. fiete , phys . rev .
b 83 , 165112 ( 2011 ) x. wan , a. vishwanath , s. y. savrasov , phys .
108 , 146601 ( 2012 ) a. go , w. witczak - krempa , g. s. jeon , k. park , y. b. kim , phys .
109 , 066401 ( 2012 ) n. n. greenwood and a. earnshaw , _ chemistry of the elements _ , elsevier - butterworth - heinemann , amsterdam ( 2012 ) g. chen , r. pereira and l. balents , phys . rev .
b 82 , 174440 ( 2010 ) b. j. kim , h. jin , s. j. moon , j .- y . kim et al . ,
lett . 101 , 076402 ( 2008 ) m. wakeshima , d. harada , y.j .
hinatsu , j alloys compd .
287 , 130 ( 1999 ) g. cao , t. f. qi , l. li , j. terzic et al . ,
112 , 056402 ( 2014 ) s. bhowal , s. baidya , i. dasgupta and t. saha - dasgupta , phys .
b 92 , 121113(r ) ( 2015 ) k. pajskr , p. novk , v. pokorn , j. koloren , r. arita and j. kune , phys . rev .
b 93 , 035129 ( 2016 ) b. ranjbar , e. reynolds , p. kayser and b. j. kennedy , inorg .
54 , 10468 ( 2015 ) b. f. phelan , e. m. seibel , d. badoe jr . ,
w. xie , r. j. cava , solid state commun .
236 , 37 ( 2016 ) see supplemental material at [ url will be inserted by publisher ] x - area ; x - shape ; x - red stoe & cie gmbh : darmstadt , germany ( 2004 ) sheldrick , g. m. shelxtl , version 5.1 , university of gttingen : gttingen , germany , ( 1997 ) petek , v. ; duek , m. ; palatinus , l. jana , the crystallographic computing system ; institute of physics : prague , czech republic , ( 2006 ) a .- c .
dippel , h .-
liermann , j. t delitz et al . ,
journal of synchrotron radiation , 22 , 3 , 675 - 687 ( 2015 ) h. m. rietveld , j. appl .
cryst . 2 , 65 ( 1969 ) t. roisnel , j. rodriguez - carjaval , mater .
forum 378 - 381 , 118 ( 2001 ) t. dey , a. maljuk , d. v. efremov , o. kataev , et al .
, phys . rev .
b 93 , 014434 ( 2016 ) m. bremholm , s. e. dutton , p. w. stephens , and r. j. cava , j. solid state chem .
184 , 601 ( 2011 ) note that the tiny humps in the 5 t curve are due to technical artifacts which are arising from the combination of the large number of data points and negligible changes in the susceptibiity in the shown temperature regime .
p. kayser , m. j. martnez - lope , j. a. alonso , m. retuerto , m. croft , a. ignatov , and m. t. fernndez - daz , inorg .
52 , 11013 ( 2013 ) a brillouin function with @xmath110 also allows to fit the data .
however , this fit leads to non - consistent , too low g values .
y. du , z.x .
cheng , s. x. dou , x. l. wang , h. y. zhao and h. kimura , app .
97 , 122502 ( 2010 ) c.g.f .
blum , a. holcombe , m. gellesch et al .
, j. cryst .
growth 421 , 39 ( 2015 ) y. cai , y. li and j. cheng , `` synthesis , crystal structure , and physical properties of the perovskite iridates '' in _ perovskite materials - synthesis , characterisation , properties , and applications _ , intech ( 2016 ) | recently , the iridate double perovskite sr@xmath0yiro@xmath1 has attracted considerable attention due to the report of unexpected magnetism in this ir@xmath2 ( 5d@xmath3 ) material , in which according to the j@xmath4 model , a non - magnetic ground state is expected . however , in recent works on polycrystalline samples of the series ba@xmath5sr@xmath6yiro@xmath1 no indication of magnetic transitions have been found .
we present a structural , magnetic and thermodynamic characterization of sr@xmath0yiro@xmath1 single crystals , with emphasis on the temperature and magnetic field dependence of the specific heat . here , we demonstrate the clue role of single crystal x - ray diffraction on the structural characterization of the sr@xmath0yiro@xmath1 double perovskite crystals by reporting the detection of a @xmath7 supercell , where @xmath8 , @xmath9 and @xmath10 are the unit cell dimensions of the reported monoclinic subcell . in agreement with the expected non - magnetic ground state of ir@xmath2 ( 5d@xmath3 ) in sr@xmath0yiro@xmath1 , no magnetic transition is observed down to 430 mk .
moreover , our results suggest that the low temperature anomaly observed in the specific heat is not related to the onset of long - range magnetic order .
instead , it is identified as a schottky anomaly caused by paramagnetic impurities present in the sample , of the order of @xmath11 % .
these impurities lead to non - negligible spin correlations , which nonetheless , are not associated with long - range magnetic ordering . |
we consider proteins with general energy landscapes not necessarily funneled and derive the conditions under which folding occurs reliably . generally , energy landscapes contain multiple minima , possibly separated by large energy barriers .
thus folding is not necessarily an equilibrium process and misfolds can occur .
below we consider the dynamics of the folding process and its effect on reliable folding .
multiple minima in the energy landscape lead to multiple minima in the free energy . in this case
we argue that there is a basic kinetic mechanism that determines whether folding is reliable .
we illustrate this kinetic mechanism by considering a transition from state @xmath7 to state @xmath8 on a non - funneled energy landscape .
although we will assume that the transition is driven by a reduction of temperature , the same arguments can be applied when a change of denaturant concentration or another parameter induces folding . in fig .
[ schematicfreeenergy ] schematic illustrations of the free energy are plotted at four temperatures @xmath9 .
we will assume that a transition from @xmath10 is induced by decreasing the temperature at a constant rate @xmath0 such that @xmath11 as a function of time @xmath12 .
initially at @xmath13 the protein resides in state @xmath7 . as temperature
is reduced to @xmath14 an equilibrium transition to state @xmath8 can occur with folding time proportional to @xmath15 , where @xmath16 is the rate at which conformations are explored . at @xmath17
a third state @xmath18 has free energy equal to that of @xmath7 . as temperature
is further reduced to @xmath19 , the minimum corresponding to state @xmath7 no longer exists and the activation barrier @xmath20 grows .
dynamics are important in determining transitions between states @xmath7 and @xmath8 .
if the time that it takes for the temperature to decrease from @xmath14 to @xmath17 is less than the folding time , the protein can fall into the metastable state @xmath18 .
this sets a bound on @xmath0 : if @xmath21 then the protein is likely to populate the state @xmath18 .
note that we use units where boltzmann s constant @xmath22 .
for a misfold to occur , the escape probability from the metastable state must be sufficiently small .
if the protein populates state @xmath18 at time @xmath23 , the probability that it has escaped at time @xmath12 is given by @xmath24 for a maximum waiting time @xmath25 the protein always escapes the metastable state for @xmath26 and rarely escapes for @xmath27 .
the crossover between frequently escaping from and being trapped in state @xmath18 occurs when @xmath28 . using @xmath29
we find that when the rate @xmath30 the probability to become trapped in the metastable state @xmath18 is significant and misfolds occur that satisfies @xmath31 .
note that the limiting rates @xmath32 and @xmath33 can also be determined for any functional form @xmath34 ( which we assumed to be linear ) and maximum waiting time @xmath25 . ] . [
cols="^,^ " , ] from these basic considerations it is apparent that protein folding transitions on non - funneled energy landscapes are influenced by multiple minima in the free energy and the rate @xmath0 at which external parameters are varied to induce folding . to determine whether reliable folding occurs
we must address two important questions : ( i ) can the protein conformation reside in a metastable local minimum ? and ( ii )
is it likely that the protein conformation becomes trapped in that local minimum ?
the answers to these questions define the limiting rates @xmath32 and @xmath33 .
the transition @xmath10 occurs reliably if @xmath0 obeys one of the inequalities , @xmath35 or @xmath36 . in the case that @xmath36 the protein
is given sufficient time to sample all states and the transition @xmath10 occurs reliably as an equilibrium process .
if @xmath37 the protein conformation becomes trapped in the state @xmath8 without fully exploring phase space and the transition occurs reliably , but out of equilibrium . if @xmath38 and @xmath39 then the protein does not transition between @xmath7 and @xmath8 reliably . in the previous section we identified a kinetic mechanism that influences transitions on non - funneled landscapes . in this section
we use this mechanism to formulate a general framework for understanding folding .
we begin by partitioning the energy landscape into basins associated with particular protein topologies , proceed to define the free energy reaction path that describes how the protein transitions from one topology to another , and then use the kinetic mechanism described above to determine whether folding is reliable . as a way to understand complex folding dynamics
, the energy landscape of an arbitrary protein can be partitioned into basins surrounding each local minimum , analogous to the inherent structure formalism for liquids and glasses @xcite . in particular
, the infinite number of protein conformations can be uniquely associated with a finite number of topologies , defined as protein conformations that correspond to local minima of the internal energy .
we denote a topology as @xmath40 , where @xmath41 is an index that contains sufficient information to fully describe the conformation ( _ e.g. _ number , type and arrangement of bonds ) .
the set of conformations @xmath42 associated with each topology @xmath40 is the basin of attraction for that topology .
the basin of attraction is defined such that all conformations that belong to @xmath42 relax to the topology @xmath40 when thermal fluctuations of the protein are suppressed .
thus the infinite number of possible protein conformations is represented by a finite number of topologies and a free energy @xmath43 can be defined for the set of protein conformations @xmath42 .
formally the partition function @xmath44 for conformations constrained to lie in @xmath42 is given by @xmath45 where integration is over all coordinates @xmath46 in the basin @xmath42 and @xmath5 is the internal energy as a function of @xmath46 .
the free energy for a protein constrained to @xmath42 can then be written in terms of the topology @xmath40 as @xmath47 where @xmath48 is the internal energy of topology @xmath40 and @xmath49 is its associated entropy @xcite , given by @xmath50/t \right)\ , d \gamma .\ ] ] the random coil state @xmath51 with zero internal energy has the largest entropy and is therefore the global minimum of free energy at sufficiently large temperature .
given a protein with an energy landscape that has been partitioned into basins of attraction , we define the free energy reaction path as the ordered sequence of topologies that the protein adopts as temperature is reduced in the equilibrium limit .
that is , if the rate @xmath0 is sufficiently small , the protein will come to equilibrium at all temperatures and proceed through the basins of attraction for a reproducible set of topologies @xmath52 .
each transition occurs at the temperature where the free energy of two topologies is equal , _
e.g. _ the transition @xmath53 occurs at the temperature @xmath54 where @xmath55 . in this way , for any energy landscape
, the free energy reaction path encodes the path taken through conformation space when folding occurs as an equilibrium - quasistatic process . to determine whether folding is reliable
, we apply the analysis introduced in the previous section to each transition in the free energy reaction path .
if we label the transitions by @xmath56 then limiting rates @xmath57 and @xmath58 can be determined for each transition by measuring properties of the free energy .
there are then three distinct folding scenarios : ( 1 ) if @xmath59 for all @xmath60 then the protein does not become trapped in metastable conformations and folding occurs reliably in equilibrium ; ( 2 ) if @xmath61 for a single transition @xmath60 then the protein falls out of equilibrium at transition @xmath60 , but reliably folds to the topology @xmath62 ( since the condition @xmath63 guarantees that the protein does not fall into a different metastable state ) .
note that if there exist multiple transitions with @xmath64 then the protein will reliably fold to the topology with the smallest value of @xmath65 for which this condition holds .
finally , ( 3 ) if @xmath66 and @xmath67 for any @xmath60 , and condition ( 2 ) does not hold for a smaller value of @xmath60 , then the protein will not fold reliably . from our analysis
we deduce that there are two types of reliable folding , equilibrium and non - equilibrium .
while reliable equilibrium folding brings the protein to the global minimum of free energy , reliable non - equilibrium folding can target local minima .
the free energy reaction path provides a useful framework to classify the relevant transitions since , depending on the rate @xmath0 , a protein will either ( 1 ) pass through all topologies on the free energy reaction path and arrive at the topology with the smallest free energy , ( 2 ) target an intermediate topology along the free energy reaction path and reliably fold to a local minimum of free energy , or ( 3 ) misfold and deviate from the free energy reaction path . to test the predictions of the previous section we perform off - lattice brownian dynamics simulations of a model protein with a single attractive energy scale .
we model the protein as a polymer chain containing both attractive ( green ) and non - attractive ( white ) spherical monomers of size @xmath68 .
interactions between non - adjacent green monomers are attractive with energy depth @xmath69 , while interactions between non - adjacent pairs of green - white or white - white monomers are purely repulsive .
this model is a variant of the `` hp '' model @xcite . thermal fluctuations of the protein at temperature @xmath1 are included using brownian dynamics simulations with solvent viscosity @xmath70 .
we observe that as the parameter @xmath71 increases from zero the polymer chain transitions from a random coil to a folded conformation . to test the predictions of the theory
we simulate a specific sequence of green and white monomers , pictured in fig .
[ energylandscape ] . in this article
we present results for two dimensions in order to simplify identification of the multiple topologies that the polymer chain adopts .
we have also conducted simulations in three dimensions and these results are included in the supporting information . in fig .
[ energylandscape ] we plot the energy landscape of the polymer chain as a function of two reaction coordinates : the radius of gyration @xmath72 and the end - to - end distance @xmath73 , each normalized by the monomer diameter @xmath68 . in terms of these two reaction coordinates ,
three energy minima exist and are separated by energy barriers .
the minima correspond to three distinct topologies that are pictured in fig .
[ energylandscape ] .
we find a total of four relevant topologies for this simple system , containing either zero @xmath51 , three @xmath74 , four @xmath75 , or five @xmath76 bonds between attractive green monomers .
energy barriers exist between @xmath74 , @xmath75 and @xmath76 because , in order to transition between the topologies , it is necessary to first break a bond and then rearrange the chain conformation .
note that four green particles is the minimum number needed to ensure multiple energy minima in two dimensions , while seven are required in three dimensions .
including additional green particles introduces additional minima and more complicated energy landscapes we treat only the simplest case here .
given the non - funneled energy landscape of the simulated protein we now determine the associated free energy reaction path .
measurements of free energy @xmath77 , normalized by temperature , as a function of @xmath78 and end - to - end distance @xmath73 are shown in fig .
[ figfreeenergya ] for a sequence of @xmath79-values that corresponds to the sequence of schematic plots in fig .
[ schematicfreeenergy ] . in fig .
[ figfreeenergya](a ) we plot @xmath77 for a small value of @xmath80 and observe that the random coil state @xmath51 is the only free energy minimum . in fig .
[ figfreeenergya](b ) @xmath79 is increased to @xmath81 and there are multiple local minima in the free energy , including the topologies @xmath51 , @xmath82 , @xmath74 , and @xmath76 .
the free energies of @xmath51 and @xmath76 are equal in fig .
[ figfreeenergya](b ) . at a slightly higher value @xmath83 ,
[ figfreeenergya](c ) exhibits three minima and the free energy of @xmath51 and @xmath74 are equal . finally at @xmath84 , the free energy plotted in fig .
[ figfreeenergya](d ) exhibits a deep minimum at topology @xmath76 . from the plots in fig .
[ figfreeenergya ] we conclude that the first and only transition in the free energy reaction path is @xmath85 where the protein folds to its native conformation .
although other local minima exist in the free energy and misfolds are possible for @xmath86 , @xmath87 is the global minimum of free energy for @xmath88 .
this simple polymer chain does not exhibit any intermediate states on the free energy reaction path , which prevents us from testing whether proteins can fold reliably to metastable minima .
however we will test all other predictions of the theory . in the _ materials and methods _
section we calculate the limiting rates @xmath89 and @xmath90 for the single transition on the free energy reaction path , where @xmath91 is the simulation time - unit .
c * ( a ) * + + * ( b ) * + + * ( c ) * + now that we have determined the free energy reaction path and calculated the limiting rates , we conduct dynamic simulations of folding . to induce folding in the polymer chain @xmath79 is increased linearly in time at rate @xmath0 ( @xmath92 ) , starting from the topology @xmath51 at @xmath93 . in fig .
[ figfolding](a ) the energy of the polymer chain is plotted as a function of @xmath79 for three different values of @xmath0 , with the final state labeled by its topology . from this figure
we clearly see that small @xmath0 targets the native state @xmath76 whereas larger @xmath0 leads to misfolding . in fig .
[ figfolding](b ) we plot the probability to fold to the native state @xmath76 as a function of @xmath94 , averaged over many folding trajectories studied for each @xmath0 .
the protein folds reliably for small rates .
the modern theory of protein folding requires funneled energy landscapes for reliable folding @xcite .
the simple protein model we consider here provides a contradiction to this viewpoint since it does not possess a funneled landscape but nevertheless folds reliably at small @xmath0 .
the free energy reaction path theory predicts that reliable folding can occur on non - funneled landscapes and provides a means to quantitatively determine the limiting rate below which folding is reliable .
given the values of @xmath32 and @xmath33 quoted above , the free energy reaction path theory predicts reliable folding for @xmath95 . in fig .
[ figfolding](b ) we have measured that reliable folding occurs for normalized rates less than @xmath96 .
the theory therefore makes a correct quantitative prediction of the simulation results .
additionally , the values of @xmath32 and @xmath33 indicate that there is a range of rates @xmath97 where reliable folding to @xmath76 occurs out of equilibrium .
we test this prediction by measuring energy fluctuations for rates at which folding is reliable , as plotted in fig .
[ figfolding](c ) . for @xmath98 fluctuations
are large at the transition point @xmath99 since the protein is sampling both folded and unfolded conformations as it remains in equilibrium . for @xmath100 fluctuations
remain small near the transition point since the protein becomes trapped in the folded state and reliable folding is a non - equilibrium process .
levinthal was the first to realize that the exponential number of collapsed conformations preclude a protein from finding its native state via random sampling .
the experimental observation that proteins fold reliably to a reproducible native state therefore requires an explanation .
the modern view is that protein sequences have evolved to favor energy landscapes with a single funnel and can therefore fold reliably .
we have demonstrated that proteins with non - funneled energy landscapes can also fold reliably , as long as the external parameters that induce folding are adjusted slowly enough . we have identified two reliable folding processes on non - funneled landscapes : equilibrium and non - equilibrium . even though it is possible that in experimental and biological settings the rate at which external parameters are varied to induce folding
is too large to access the equilibrium limit , reliable folding can occur out of equilibrium .
if this is the case , the native state should be regarded as a reliably targeted local minimum on the free energy reaction path that remains metastable over timescales sufficient for biological function .
the importance of the free energy reaction path and the necessity of using small rates to vary external parameters presents challenges for protein folding simulations .
reliable protein folding is especially difficult to study in all - atom simulations where , due to the long time scales and large number of atoms , extremely rapid rates are used to induce folding @xcite . from our results ,
reliable folding on non - funneled landscapes depends on rate , thus simulation studies that argue that funneled energy landscapes are necessary for reliable folding @xcite must be carefully interpreted if only large rates are considered .
our predictions can be tested in experiments by studying folding over a range of rates , using methods such as ultrafast mixing or laser pulsing @xcite .
some progress has been made in this direction @xcite and the observation of `` strange kinetics '' @xcite after rapid temperature jumps is consistent with our predictions . in three dimensions
the limiting rates are proportional to @xmath101 , where @xmath102 is the hydrodynamic radius .
this implies that investigating folding in a variety of solvents with different viscosities @xmath70 can greatly increase the range of experimentally accessible rates .
moreover , due to the inverse dependence on @xmath1 , folding by changing temperature will give different limiting rates than folding by reducing denaturant concentration .
finally , it is intriguing to speculate about folding _ in vivo_. given that the folded state of a protein is dependent on rate at which external parameters are varied to induce folding , and that local minima in free energy can be targeted by adjusting this rate , it is possible that protein sequence has evolved along with the biological environment in which it folds . since the folding process is determined by protein sequence and rate , both are likely used in nature to ensure robust folding .
simulations are performed on polymer chains of spherical monomers , each with diameter @xmath68 .
we include two types of monomers attractive ( green ) and non - attractive ( white ) .
interactions depend on the separation @xmath103 between monomers @xmath60 and @xmath104 , and it is convenient to define the normalized distance @xmath105 . interactions between adjacent monomers are chosen to prevent the polymer chain from breaking , while interactions between non - adjacent monomers are either purely repulsive ( for green - white or white - white interactions ) or attractive ( for green - green interactions ) . more specifically , monomers that are adjacent on the polymer chain experience a piecewise continuous potential @xmath106 that is comprised of a purely repulsive lennard - jones ( rlj ) potential @xcite for separations @xmath107 and a fene potential @xcite for separations @xmath108 : @xmath109 where @xmath110 sets the energy scale and @xmath111 . this potential has a minimum of zero at @xmath112 and diverges at @xmath113 to prevent adjacent monomers from unbinding .
green - green interactions are described by a lennard - jones ( lj ) potential @xmath114 with energy depth @xmath69 at @xmath112 , whereas green - white and white - white interactions obey a rlj potential @xmath115 that provides a repulsive force when particles overlap and no force when they do not overlap .
thermal fluctuations are included using off - lattice brownian dynamics simulations @xcite .
the vector position @xmath116 of each monomer @xmath60 is determined at each time - step by the attractive and repulsive forces arising from the potentials in eqs .
[ interaction1]-[interaction3 ] and random forces arising from thermal fluctuations .
the equation of motion for monomer @xmath60 is @xmath117 , \label{newteqn}\ ] ] where @xmath118 is a gaussian random force and @xmath119 a damping force , with @xmath120 denoting the velocity of monomer @xmath60 and @xmath70 the solvent viscosity .
the gaussian random force has zero mean and a standard deviation proportional to @xmath121 .
we solve eq .
[ newteqn ] using standard numerical integration techniques @xcite in the limit that monomer mass @xmath122 .
folding simulations are conducted by starting with @xmath123 and decreasing @xmath124 linearly in time with rate @xmath0 at constant @xmath125 . in the supporting information
we include two movies from our simulations .
these show the behavior of a two dimensional polymer chain at @xmath126 where folding occurs reliably ( `` slowrate.mov '' ) and at @xmath127 where a misfold occurs ( `` fastrate.mov '' ) . the energy landscape in fig . [ energylandscape ]
is created by running @xmath128 separate folding simulations at each of five rates @xmath129 .
each simulation explores the range @xmath130 and the energy landscape is obtained by constructing a histogram over all observed states .
we believe that the landscape is sufficiently sampled since we observe that there is very little difference at small @xmath73 and @xmath72 between the energy landscape pictured in fig . [ energylandscape ] and ones measured using only data from the smallest @xmath0 .
the free energies in fig . [ figfreeenergya ] are measured by slowly ramping to the desired @xmath79-value with @xmath131 , and then calculating a histogram of the probability @xmath132 to have energy @xmath5 and end - to - end distance @xmath73 over @xmath133 time - steps for each @xmath79-value reported .
the free energy @xmath134 is determined ( within an additive constant ) from the probability via the relation @xmath135 .
the limiting rates can be determined using equations similar to those in eqs .
[ tempfall ] and [ tempstuck ] , @xmath136 these equations are derived for the simulation protocol where @xmath137 increases linearly in time to induce folding , with @xmath1 constant .
the maximum waiting time is taken to infinity .
we first calculate @xmath32 . the data in fig .
[ figfreeenergya ] gives @xmath81 and @xmath138 .
the free energy barrier @xmath139 is determined by preparing the protein in topology @xmath76 at @xmath140 and measuring the amount of time @xmath141 required to transition to topology @xmath51 , averaged over @xmath142 trials .
the free energy barrier is related to the transition time by @xmath143 .
we measure @xmath144 , where @xmath91 is the fundamental unit of time in the simulations . inserting these numbers into eq .
[ equationrf ] yields @xmath145 . the rate @xmath33 is determined by preparing the protein in topology @xmath74 and measuring the average time @xmath146 required to transition to the native topology @xmath76 .
we average @xmath146 over @xmath142 trials for each @xmath79-value and it is plotted in fig .
[ figfreeenergyb ] . since @xmath147
we calculate @xmath148 by direct integration of @xmath149 , according to eq .
[ equationrs ] .
contributions to the numerical value of @xmath33 from @xmath150 are negligible . financial support from nsf grant numbers cbet-0348175 ( gl , jb ) , dmr-0448838 ( gl , cso ) , and yale s institute for nanoscience and quantum engineering ( gl ) is gratefully acknowledged .
we also thank yale s high performance computing center for computing time .
100 c. levinthal ( 1968 ) _ j. chim .
phys . _ * 65 * , 44 - 45 ; c. levinthal ( 1969 ) in _ mossbauer spectroscopy in biological systems _ , eds .
p. debrunner , j. c. m. tsibris and e. munck ( university of illinois press , urbana ) , 22 - 24 .
r. zwanzig , a. szabo and b. bagchi ( 1992 ) _ proc .
sci . _ * 89 * , 20 - 22 .
k. a. dill and h. s. chan ( 1997 ) _ nature struct .
biol . _ * 4 * , 10 - 19 . s. s. plotkin and j. n. onuchic ( 2002 ) , _ q. rev .
biophys . _ * 35 * , 111 - 167 .
d. j. wales ( 2003 ) _ energy landscapes _
( cambridge university press , cambridge ) .
h. s. chan and k. a. dill ( 1998 ) _ proteins _ * 30 * , 2 - 33 .
r. l. baldwin ( 1995 ) _ j. biomolec
. nmr _ * 5 * , 103 - 109 .
j. d. bryngelson and p. g. wolynes ( 1987 ) _ proc .
_ * 84 * , 7524 - 7528 ; p. e. leopold , m. montal and j. n. onuchic ( 1992 ) _ proc . natl
* 89 * , 8721 - 8725 ; j. n. onuchic , z. luthey - schulten and p. g. wolynes ( 1997 ) _ annu .
chem . _ * 48 * , 545 - 600 .
t. kiefhaber ( 1995 ) _ proc .
sci . _ * 92 * , 9029 - 9033 ; q .- x .
hua , s. n. gozani , r. e. chance , j. a. hoffman , b. h. frank and m. a. weiss ( 1997 ) _ nature struct .
* 2 * , 129 - 138 ; t. pan and t. r. sosnick ( 1997 ) _ nature .
* 4 * , 931 - 938 ; j .- y .
chang , l. li and p .- h .
lai ( 2001 ) _ j. biol .
chem . _ * 276 * , 4845 - 4852 ; h. i m , m .- s .
woo , k. y. hwang and m .- h .
yu ( 2002 ) _
_ * 277 * , 46347 - 46354 .
i. e. t. iben _ et .
( 1989 ) _ phys .
lett . _ * 62 * , 1916 - 1919 ; r. d. young _ et .
( 1991 ) _ chemical physics _ _
158 _ , 315 - 327 .
f. h. stillinger and t. a. weber ( 1982 ) _ phys .
rev . a _ _ 25 _ , 978 - 989 .
k. a. dill ( 1990 ) _ biochemistry _ * 29 * , 7133 - 7155 .
f. h. stillinger , t. head - gordon and c. l. hirshfeld ( 1993 ) _ phys .
e _ * 48 * , 1469 - 1477 ; s. schnabel , m. bachmann and w. janke ( 2007 ) _ phys .
lett . _ * 98 * , 048103 .
shea and c. l. brooks iii ( 2001 ) _ annu .
chem . _ * 52 * , 499 - 535 ; c. d. snow , e. j. sorin , y. m. rhee and v. s. pande ( 2005 ) _ annu .
struct . _ * 34 * , 43 - 69 .
a. sali , e. shakhnovich and m. karplus ( 1994 ) _ nature _ * 369 * , 248 - 251 ; a. sali , e. shakhnovich and m. karplus ( 1994 ) _ j. mol .
_ * 235 * , 1614 - 1636
. w. a. eaton , v. munoz , p. a. thompson , c .- k .
chan and j. hofrichter ( 1997 ) _ curr .
_ * 7 * , 10 - 14 ; w. a. eaton , v. munoz , s. j. hagen , g. s. jas , l. j. lapidus , e. r. henry and j. hofrichter ( 2000 ) _ annu
* 29 * , 327 - 359 .
r. m. ballew , j. sabelko and m. gruebele ( 1996 ) _ proc .
_ * 93 * , 5759 - 5764 ; s. takahashi , s .-
yeh , t. k. das , c .- k .
chan , d. s. gottfried and d. l. rousseau ( 1997 ) _ nature struct .
* 4 * , 44 - 50 ; t. p. causgrove and r. b. dyer ( 2006 ) _ chem .
phys . _ * 323 * , 2 - 10 .
j. sabelko , j. ervin and m. gruebele ( 1999 ) _ proc .
sci . _ * 96 * , 6031 - 6036 .
m. p. allen and d. j. tildesley ( 1987 ) _ computer simulation of liquids _
( oxford university press , oxford ) , 257 - 269 .
r. m. jendrejack , m. d. graham and j. j. de pablo ( 2000 ) _ j. chem .
phys . _ * 7 * , 2894 - 2900 .
in the manuscript , simulation results were presented for a two dimensional model protein . here
we include results for three dimensions .
these results exhibit similar behavior and support the theoretical predictions .
we perform off - lattice brownian dynamics simulations in three dimensions to simulate the folding process .
we study the model protein pictured in fig .
[ 3denergylandscape ] that consists of @xmath151 monomers , seven of which are attractive . in fig .
[ 3denergylandscape ] we plot the protein energy landscape as a function of the radius of gyration @xmath72 and end - to - end distance @xmath73 , each normalized by the monomer diameter @xmath68 .
there are two minima at small @xmath72 and @xmath73 , corresponding to the topologies @xmath152 and @xmath153 pictured in the figure . as in the two dimensional case ,
non - funneled energy landscapes promote misfolding if the rate that the attractive strength @xmath154 is increased to induce folding is sufficiently large . in fig . [ 3dfigfolding](a ) we plot the energy as a function of @xmath155 . for small rates the simulated protein folds to the global energy minimum @xmath153 . for larger rates the system misfolds to the local minimum @xmath152 .
in fig . [ 3dfigfolding](b ) we plot the probability to fold to the native state @xmath153 as a function of rate .
the protein folds reliably below a normalized rate of @xmath156 . the limiting rate below which folding is reliable can be predicted by measurements of free energy . in fig .
[ figfreeenergy ] we plot the free energy as a function of end - to - end distance @xmath73 and normalized energy @xmath78 for many different values of @xmath79 . in fig .
[ figfreeenergy](a ) the random coil state @xmath51 is the only minimum in the free energy . for @xmath157 , fig .
[ figfreeenergy](b ) demonstrates that @xmath153 and @xmath158 have equal free energies . in fig .
[ figfreeenergy](c ) the random coil @xmath51 , native state @xmath153 , and metastable state @xmath152 basins of attraction are present . at this value of @xmath159
, topology @xmath152 has a free energy equal to that of @xmath51 . for larger @xmath79 fig .
[ figfreeenergy](d ) demonstrates that the protein has an increasing probability to populate the basin of attraction for @xmath153 , although the basin of attraction for @xmath152 is still visible . from this series of free energy plots ,
it is apparent that the simulated protein possesses a single equilibrium transition at @xmath140 from @xmath51 to @xmath153 , and misfolds to @xmath152 are possible for @xmath86 .
the rate @xmath32 is calculated using the values @xmath160 and @xmath161 , along with the transition time @xmath141 from @xmath153 to @xmath51 at @xmath157 .
we measure @xmath162 , averaged over one hundred trials . given these values
we calculate @xmath163 .
the rate @xmath33 is calculated by measuring the transition time @xmath146 between topologies @xmath153 and @xmath152 , which is shown in fig .
[ freeenergybarrier ] . directly integrating this data for @xmath86 yields @xmath164 . given the values of @xmath32 and @xmath33 we expect the protein to fold reliably for @xmath165 ,
which is consistent with the data in fig . [
3dfigfolding](b ) .
in contrast to the two dimensional simulations , we find @xmath166 and thus this particular protein can only fold in equilibrium . generally we believe that the ordering of @xmath32 and @xmath33 can depend on the length , sequence and energy scales of the protein . | a theoretical framework is developed to study the dynamics of protein folding .
the key insight is that the search for the native protein conformation is influenced by the rate @xmath0 at which external parameters , such as temperature , chemical denaturant or ph , are adjusted to induce folding .
a theory based on this insight predicts that ( 1 ) proteins with non - funneled energy landscapes can fold reliably to their native state , ( 2 ) reliable folding can occur as an equilibrium or out - of - equilibrium process , and ( 3 ) reliable folding only occurs when the rate @xmath0 is below a limiting value , which can be calculated from measurements of the free energy .
we test these predictions against numerical simulations of model proteins with a single energy scale . under appropriate conditions ,
proteins spontaneously fold from an extended one - dimensional chain of amino acids to a unique three - dimensional native conformation . how this occurs on timescales accessible to experiment and relevant to biological function is a question that has intrigued scientists for the past forty years .
levinthal @xcite was the first to recognize the importance of timescales and point out that , assuming a random search of conformation space , proteins would not fold in a person s lifetime .
this argument has come to be known as levinthal s paradox since proteins must fold for human life to exist in the first place . of course
conformation space is not sampled randomly and levinthal s paradox has been resolved by applying statistical mechanics to the protein folding problem @xcite .
each protein conformation has a free energy that determines its probability to be sampled at temperature @xmath1 . while the free energy @xmath2 generally comprises a sum of many enthalpic and entropic terms , it is convenient to express it as @xmath3 , where @xmath4 is the conformational entropy of only the protein degrees of freedom and @xmath5 is the `` internal energy '' that includes all other contributions to the free energy ( from both protein and solvent ) .
the functional dependence of @xmath5 on all protein degrees of freedom is called the energy landscape @xcite , which in general contains many minima . at @xmath6
only the energy landscape is relevant and the protein resides in a local ( or global ) minimum , corresponding to a compact conformation .
as @xmath1 increases the conformational entropy smooths out the minima in the energy landscape and the protein adopts more extended states with larger @xmath4 . in the `` new view '' of protein folding @xcite statistical fluctuations on an energy landscape
give rise to an ensemble of folding pathways . often associated with
the new view is the hypothesis that energy landscapes have the shape of a multi - dimensional funnel @xcite .
proponents argue that in order to fold reliably ( transition to the native state with probability one ) the energy landscape must contain a single low - lying minimum to which all conformations are channeled .
if multiple funnels exist , separated by large enough energy barriers , then at low temperature or denaturant concentration a protein can become trapped in a local minimum of energy that does not correspond to its native conformation . while the existence of a single funnel is a sufficient condition for reliable protein folding , the number of proteins with a single funnel is expected to be small and the observation of kinetic traps @xcite and glassy behavior @xcite in biologically relevant proteins indicates that not all proteins fold on smooth funneled landscapes .
here we address the open question : is a funneled energy landscape necessary for reliable folding ? by formulating a statistical theory that includes the dynamics of folding , we find that a funneled landscape is not necessary for reliable folding .
the important insight is that the rate @xmath0 at which temperature or chemical denaturant concentration is decreased to induce folding affects the final conformation of the protein .
for sufficiently small @xmath0 the protein always folds to its native conformation , whereas for larger @xmath0 it can become trapped in a metastable state .
this leads to new predictions that can be tested in experiments and simulations .
first , proteins with non - funneled energy landscapes can fold reliably to their native state if the rate @xmath0 is below a limiting value .
second , reliable folding can occur as an equilibrium - quasistatic or non - equilibrium process .
third , in a non - equilibrium folding process , a protein can reliably fold to a local ( instead of global ) minimum of the energy landscape .
we conduct off - lattice simulations of model proteins with non - funneled energy landscapes and verify these predictions . |
recent progress in the nano - fabrication of quantum devices enables us to study electron transport through quantum dots(qds ) in a controllable way @xcite since a wide range of energy can be achieved here by continuously changing the applied external potential in contrast with real atoms . a qd- array regarded as
an artificial crystal@xcite also becomes reliable due to the recent development of nanotechnology .
when the size of structures is small with respect to the coherence length , the transmission probability plays an essential role in the quantum transport .
for example , we can obtain the conductance of the devices from the transmission probability with the help of the laudauer - b@xmath3ticker formula .
various methods have been developed for numerical calculations of the transmission probability @xcite .
the quantum transport through a multi - terminal system@xcite can be analyzed in terms of all transmission probabilities between any pair of terminals in principle .
b@xmath3ticker et al.@xcite derived a scattering matrix for a three - terminal junction consisting of homogeneous wires based on the unitary condition and the assumption that the scattering matrix is real .
it was shown that this matrix is energy independent and the coupling parameter has to be determined from phenomenological arguments@xcite .
an energy - dependent scattering matrix for three - terminal junction was found later @xcite .
spintronics has become a new branch of condensed matter physics and material science where not only the charge but also the spin degree of freedom of electrons play an essential role@xcite .
it is of fundamental importance to study the mechanism to generate and control the spin - polarized currents in spintronic devices which may promise practical applications in quantum computing and information@xcite . in recent yeas
the spin - polarized transports in quantum dots have attracted considerable attentions both theoretically@xcite and experimentally@xcite , where the spin currents are generated either by magnetic material electrodes or by the rotating - magnetic field which results in the spin - flip .
most of the studies are concentrated on the one or two - qd devices with only one output electrode and the spin - polarized transport in qd - array consisting of arbitrary number of qds with multi - terminal has not yet be investigated . in this paper
we study the quantum transport through a @xmath0-shape qd - array with three electrodes in which the spin - polarized current is induced by static magnetic fields applied on the three arms of the @xmath0-shape qd - array .
we calculate the transmission probability in terms of the scattering matrix which is spin - polarization dependent in the presence of magnetic fields and leads to the spin - polarized currents .
it is the main goal of the present study to show how the spin currents in output terminals can be generated and controlled .
particularly two spin currents of opposite - spin - polarization flowing simultaneously out to electrodes are obtained and seem extremely useful in the design of quantum- logic - gate devices .
the system consisting of three perfect wires ( electrodes ) attached to a @xmath0-shape qd - array of arbitrary number of qds is explained schematically in fig.1 .
the horizontal qd - array is denoted by the lattice sites @xmath4 @xmath5 @xmath6 and the vertical qd - array by the lattice sites @xmath7 @xmath8 @xmath9 where @xmath10 , @xmath11 , and @xmath12 are the numbers of qds on the left , right and vertical arms respectively .
three perfect wires are also described in terms of discrete lattice sites .
the first magnetic field in the z - direction is assumed to be applied , for example , on site @xmath13 , the second magnetic field on the site @xmath14 is along a direction of angle @xmath15 with the z - axis , while the third one on the site @xmath16 has a direction angle @xmath17 .
the latter two magnetic fields not in the same direction as the first one result in the spin - flip of electrons , which is the mechanism generating the spin - polarized transport in the qd - array ( see the hamiltonian below ) .
we remark that it is not necessary to have the three fields to apply on the certain single qd and what we choose the three qds is only for the convenience of analysis
. it may be more easy in practical experiments to superpose the external fields on the three arms of the @xmath0-shape qd - array .
the hamiltonian of the system includes five parts @xmath18 where @xmath19 and @xmath20 denote the hamiltonians of the horizontal and vertical qd - arrays respectively with @xmath21 and @xmath22 being the energy eigenvalues of corresponding qds .
@xmath23and @xmath24 ( with @xmath25 ) are the creation operators of the electrons with spin index @xmath26 on the qds , where @xmath27 denotes the opposite spin polarization with respect to @xmath26 .
the zeeman terms induced by the external fields on the sites @xmath14 , and @xmath16 result in the spin - flip .
it should be emphasized that the @xmath15 and @xmath28 dependent terms describing the spin - flip in the hamiltonian are the key mechanism of the spin - polarized transport in our system .
the matrix elements defined by @xmath29 denote the hopping integrals between the nearest neighbors of qds and are independent of spin polarization , except for @xmath30 , @xmath31 , and @xmath32 , where zeeman energy induced by the external magnetic fields leads to the energy - level splitting and therefore the imbalance of tunnel couplings between spin - up and spin - down electrons .
the hopping matrix elements in connection with the qds @xmath14 , @xmath13 and @xmath16 are assumed to be @xmath33 .
the hamiltonians of horizontal and vertical electrodes are written as @xmath34 @xmath35 where @xmath36 , @xmath37 and @xmath38 are the on - site energies in the right - hand - side , the left - hand - side and the vertical electrodes respectively . @xmath39 and @xmath40 are the creation operators of electrons in the electrodes . finally the hopping of electrons between the qd and electrode is described by @xmath41 where @xmath42 , @xmath43 , @xmath44 respectively denote the tunnel couplings between the three special qds , where the magnetic fields are applied , and corresponding electrodes . the main quantities which we have to calculate in our formulation are the transmission probabilities @xmath45 of electrons from electrode-@xmath46 to electrode-@xmath47 , and the reflection probabilities @xmath48 in the electrode-@xmath46 which is considered as input terminal . for our three - terminal system
we have @xmath49 ( the left electrode ) and @xmath50 ( the right and vertical electrodes ) with @xmath51 denoting the transmission probability from the left to right electrodes and @xmath52 from the left to vertical electrodes .
however the situation is not simple for the evaluation of transmission and reflection probabilities in the multi - terminal system since the electron transport between two electrodes is by no means an isolated process but affected by other electrodes . to this end
we begin with the stationary shr@xmath53dinger equation @xmath54 which possesses a general solution of the form @xmath55 inserting eq.(3 ) into the shr@xmath53dinger equation ( 2 ) yields @xmath56 @xmath57 @xmath58{l}% $ ( e-(\varepsilon_{n_{r},0}+\sigma g\mu_{b_{1}}b_{1}\cos\theta_{1}% ) ) c_{(n_{r},0),\sigma}+b_{1}\sin\theta_{1}c_{(n_{r},0),-\sigma}$\\ $ + t^{\prime}c_{(n_{r}-1,0),\sigma}+v_{r,\sigma}c_{(n_{r}+1,0),\sigma}=0,$% \end{tabular } \tag{6}%\ ] ] @xmath58{l}% $ ( e-(\varepsilon_{0,n_{u}}+\sigma g\mu_{b_{2}}b_{2}\cos\theta_{2}% ) ) c_{(0,n_{u}),\sigma}+b_{2}\sin\theta_{2}c_{(0,n_{u}),-\sigma}$\\ $ + t^{\prime}c_{(0,n_{u}-1),\sigma}+v_{u,\sigma}c_{(0,n_{u}+1),\sigma}=0.$% \end{tabular } \tag{7}%\ ] ] notice that in the input electrode-@xmath59 the wave function is a superposition of incoming plane wave of unit amplitude and a reflection wave with amplitude @xmath60 , while in the output electrodes-@xmath61 , @xmath62 the outgoing plane waves possess transmission amplitudes @xmath63 and @xmath64 respectively , we have @xmath65 @xmath66 @xmath67 where , @xmath68 , @xmath69 , and @xmath70 are the wave vectors in the left- , right- , and vertical - electrode respectively . using these wave functions , we can eliminate all the coefficients , @xmath71 , in eg.@xmath72 and eg.@xmath73 and obtain @xmath74 @xmath75 @xmath76 @xmath58{l}% $ ( e-(\varepsilon_{n_{r},0}+\sigma g\mu_{b_{1}}b_{1}\cos\theta_{1}% ) ) c_{(n_{r},0),\sigma}+b_{1}\sin\theta_{1}c_{(n_{r},0),-\sigma}$\\ $ + t^{\prime}c_{(n_{r}-1,0),\sigma}+v_{r,\sigma}c_{(n_{r}+1,0),\sigma}=0,$% \end{tabular } \tag{13}%\ ] ] where @xmath77 is the effective energy on site @xmath78 with @xmath79{c}% .\\ \text { \ \ \ \ \ \ \ \ } .\\
\text { \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ } .\\
\text { \ \ \ \ \ \ \ \ \ \ \ \ } \frac{(t^{\prime})^{2}}{e-\varepsilon _ { 0,n_{u}}-{\sum\nolimits^{u}}(e)}. \end{array } } } } \tag{14}%\ ] ] here @xmath80 denotes the self - energy due to the coupling with the horizontal qd - array , and @xmath81 is the green function of the lattice - site @xmath16 satisfying the recursive equation @xmath82^{-1 } , \tag{15}%\ ] ] the solution of which is seen to be @xmath83^{\frac12}\}. \tag{16}%\ ] ] we can see that , the eq .
( 13 ) does not include any of the expansion coefficients , @xmath71 , associated with the vertical lattice - site .
thus it implies that the three - terminal device reduces to an effective two - terminal system and the effect of the vertical qd - array appears as a self - energy @xmath84 which has now been included in the effective on - site energy .
now we rewrite eq .
( 13 ) as a matrix equation @xmath85{c}% c_{(n+1,0),\sigma}\\ c_{(n,0),\sigma}\\ c_{(n+1,0),-\sigma}\\ c_{(n,0),-\sigma}% \end{array } \right ] = m[(n,0),e]\left [ \begin{array } [ c]{c}%
c_{(n,0),\sigma}\\ c_{(n-1,0),\sigma}\\ c_{(n,0),-\sigma}\\ c_{(n-1,0),-\sigma}% \end{array } \right ] , \ ] ] where @xmath86 $ ] is called the transfer matrix which links the expansion coefficient vector @xmath87 to the vector @xmath88 and thus is defined by @xmath89 & = \left [ \begin{array } [ c]{cccc}% -\frac{e-\varepsilon_{n,0}^{\prime}}t & -1 & 0 & 0\\ 1 & 0 & 0 & 0\\ 0 & 0 & -\frac{e-\varepsilon_{n,0}^{\prime}}t & -1\\ 0 & 0 & 1 & 0 \end{array } \right ] = \left [ \begin{array } [ c]{cc}% q_{1 } & 0\\ 0 & q_{2}% \end{array } \right ] , \
n\neq n_{r}\ \nonumber\\ m[(n_{r},0),e ] & = \left [ \begin{array } [ c]{cccc}% -\frac{e-\varepsilon_{n_{r},0}^{\prime}}{v_{r,\sigma } } & -\frac{t^{\prime}% } { v_{r,\sigma } } & \frac{g\mu_{b_{1}}b_{1}\sin\theta_{1}}{v_{r,\sigma } } & 0\\ 1 & 0 & 0 & 0\\ \frac{g\mu_{b_{1}}b_{1}\sin\theta_{1}}{v_{r,\sigma } } & 0 & -\frac { e-\varepsilon_{n_{r},0}^{\prime}}{v_{r,\sigma } } & -\frac{t^{\prime}% } { v_{r,\sigma}}\\ 0 & 0 & 1 & 0 \end{array } \right ] , \ \ ( n = n_{r } ) .
\tag{17}%\end{aligned}\ ] ] where @xmath90 is the effective on - site energy given by @xmath91 from eqs . @xmath92 and
@xmath93 , it can be shown that the transmission amplitude @xmath63 is related to the incident coefficients via the equation @xmath94{c}% t_{21}(\sigma)\\ 0\\ t_{21}(-\sigma)\\ 0 \end{array } \right ] = t(e)\left [ \begin{array } [ c]{c}% 1\\ r_{11}(\sigma)\\ 1\\ r_{11}(-\sigma ) \end{array } \right ] , \tag{19}%\ ] ] where the transfer matrix @xmath95 for the whole system is given by @xmath96{cccc}% e^{-ik_{\sigma}^{r}(n_{r}+1)a } & 0 & 0 & 0\\ 0 & e^{ik_{\sigma}^{r}(n_{r}+1)a } & 0 & 0\\ 0 & 0 & e^{-ik_{-\sigma}^{r}(n_{r}+1)a } & 0\\ 0 & 0 & 0 & e^{ik_{-\sigma}^{r}(n_{r}+1)a}% \end{array } \right ] \times\nonumber\\ & \left [ \begin{array } [ c]{cccc}% e^{ik_{\sigma}^{r}a } & e^{-ik_{\sigma}^{r}a } & 0 & 0\\ 1 & 1 & 0 & 0\\ 0 & 0 & e^{ik_{-\sigma}^{r}a } & e^{-ik_{-\sigma}^{r}a}\\ 0 & 0 & 1 & 1 \end{array } \right ] ^{-1}% % tcimacro{\dprod \limits_{n = n_{r}+1}^{-(n_{l}+1)}}% % beginexpansion { \displaystyle\prod\limits_{n = n_{r}+1}^{-(n_{l}+1 ) } } % endexpansion m[(n,0),e]\nonumber\\ & \times\left [ \begin{array } [ c]{cccc}% e^{ik_{\sigma}^{l}a } & e^{-ik_{\sigma}^{l}a } & 0 & 0\\ 1 & 1 & 0 & 0\\ 0 & 0 & e^{ik_{-\sigma}^{l}a } & e^{-ik_{-\sigma}^{l}a}\\ 0 & 0 & 1 & 1 \end{array } \right ] \left [ \begin{array } [ c]{cccc}% e^{-ik_{\sigma}^{l}(n_{l}+1)a } & 0 & 0 & 0\\ 0 & e^{ik_{\sigma}^{l}(n_{l}+1)a } & 0 & 0\\ 0 & 0 & e^{-ik_{-\sigma}^{l}(n_{l}+1)a } & 0\\ 0 & 0 & 0 & e^{ik_{-\sigma}^{l}(n_{l}+1)a}% \end{array } \right ] .
\tag{20}%\end{aligned}\ ] ] thus from eq.@xmath97 we can obtain the reflection amplitudes @xmath98 and @xmath99 as @xmath100 @xmath101 and the transmission amplitudes @xmath102 and @xmath103@xmath104 @xmath105 the spin - polarized reflection and transmission probabilities are seen to be @xmath106 @xmath107 and the total transmission and reflection probabilities are defined by @xmath108 @xmath109 the spin current can be derived from the difference between the two spin - polarized transmission probabilities @xmath110 the transmission probability , @xmath111 , from electrode-@xmath59 to electrode-@xmath62 can be simply obtained by replacing the on - site energy @xmath112 in the matrix @xmath113 $ ] with @xmath114 where @xmath115{c}% .\\
\text { \ \ \ \ \ \ \ \ } .\\
\text { \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ } .\\
\text { \ \ \ \ \ \ \ \ \ \ \ \ } \frac{(t^{\prime})^{2}}{e-\varepsilon _ { n_{r},0}-{\sum\nolimits^{r}}(e ) } , \end{array } } } } \tag{30}%\ ] ] and @xmath116 is the self - energy due to the presence of the right electrode , and @xmath117^{-1 } , \tag{31}%\ ] ] is the green@xmath1s function of the lattice - site @xmath118 .
moreover the effective on - site energy @xmath119 of eq .
@xmath120 should be replaced by @xmath121 then the transmission probability @xmath111 can be evaluated from @xmath122 and finally we obtain the total and spin transmission probabilities @xmath123 , @xmath124 in the same way as for the calculation of @xmath125 and @xmath126 .
the conductance of the device can be obtained from the transmission probability with the help of the laudauer - b@xmath3ticker formula @xmath127 where @xmath128 is the electron charge , @xmath129 the planck s constant , and @xmath0 the transmission probability of the device .
in the numerical evaluation the lattice constant @xmath130 is chosen as unit of length and hopping constant @xmath131 as the unit of energy .
the on - site energies of all qds , as well as @xmath132 , @xmath133 , @xmath134 are set to the same value of @xmath135 .
the coupling parameters between the qds and the three electrodes are assumed to be @xmath136 and @xmath137 first of all we consider the simplest case that there is only one - qd on each arm of the @xmath0-shape array besides the central dot , i.e. @xmath138 figs . @xmath139 and @xmath140 ,
@xmath141 display the energy - dependence ( spectrum ) of the spin - up and the spin - down transmission probabilities to the electrode-@xmath61 and the electrode -@xmath62 respectively ( for @xmath142 , @xmath143 , @xmath144 , @xmath145 ) .
the total transmissions @xmath125 , @xmath123 are shown in figs . @xmath146 and @xmath147 .
it can be seen that in the presence of the external magnetic fields , the transmitted electrons split into spin - up and spin - down components with different spectra .
@xmath148 is the plot of the relative transmission probabilities between two spin components @xmath126 ( solid line ) and @xmath124 ( dotted line ) versus the energy respectively .
we can see that in some electron - energy ranges , for example from @xmath149 to @xmath150 or from @xmath151 to @xmath152 , the spin transmission probability to the electrode-@xmath62 , @xmath153 is positive ( i.e. the spin current with spin - up polarization ) while the spin transmission probability to electrode-@xmath61 , @xmath154 is negative ( namely the spin - down current ) .
however the situation would be opposite in other energy region , for example from @xmath155 to @xmath156 , where @xmath126 is positive and @xmath124 is negative .
@xmath157 show the spin transition probabilities @xmath126 ( solid line ) and @xmath124 ( dotted line ) as a function of energy for the multi - qd case that @xmath158 and @xmath159 with the magnetic field direction angles @xmath160 . in the electron - energy range from @xmath161 to @xmath162 or from @xmath163 to @xmath164 the spin transmission probability @xmath124 is positive ( i.e. the spin - up current flowing to electrode-@xmath62 ) while @xmath126 is negative ( spin - down current flowing to electrode-@xmath61 ) .
opposite spin currents are observed in the energy region from @xmath165 to @xmath166 , where @xmath124 is positive and @xmath167 is negative .
@xmath168 are the results of @xmath126 ( solid line ) and @xmath124 ( dotted line ) for the direction angles @xmath169 while @xmath170 we see that the spin currents are sensitive to the fields .
it may also be worthy to point out that in the energy region between @xmath171 and @xmath172 , @xmath126 is negligibly small while @xmath124 is positive .
in other words the spin current can exist only in one output electrode . comparing with previous case ( one - qd on each arm )
more resonant peaks of the spectra appear with the increasing number of qds in the same way as the usual quantum transport in the qd - array without spin - polarization .
the generation of spin currents in the two output electrodes in our model can be understood that a force applied on the spin-@xmath173 which is induced by the spatially inhomogeneous magnetic fields , namely @xmath174 ( where * * * * @xmath175 is the bohr magneton , @xmath176 is the spin g - factor ) , removes the degeneracy of spin - polarization in the transport and thus leads to the relative shifts of resonant peaks of the transmission spectra between the spin - up and spin - down components of electron . * * * * the spin current appears when the transmission spectrum of one spin component reaches a maximum while the spectrum of the other spin component is a minimum**. * * the homogeneous dot - arry is just a theoretical idealization and disorder is necessary to be taken into account in practical systems .
it is interesting to see the disorder effect of the qd - array on the spin - polarized quantum transport . to this end
we consider the disorder following refs.@xcite that the on - site energies of qds are alternated with @xmath177 and @xmath178 , where @xmath179 is the probability distribution parameter .
when @xmath180 , the system reduces to the non - disordering one . in the @xmath0-shape qd - array on -
site energies @xmath181 ( @xmath182 , on the horizontal arm ) and @xmath183 ( @xmath184 , on vertical arm ) are set to @xmath177 for odd-@xmath185 and @xmath186 for even-@xmath185 . fig .
( 8) is a comparison of plots of @xmath124 for the probability distribution parameter @xmath187 ( dotted line ) , @xmath188 ( solid line ) respectively with @xmath158 and @xmath159 .
we can see that the disorder can shift the resonant peaks of the transmission probability @xmath124 slightly , and suppress the height of some peaks but no more then 20% while increase the height of other peaks no more then 18% . in other words
the spin splitting is robust against the disorder .
the latest advances in nanotechnology make it possible to fabricate qd - array@xcite and the model of spin - splitter proposed in this paper may be realizable experimentally . in practical experiment
it is more easy to apply the external fields on three arms of the @xmath0-shap qd - array instead of three qds .
the relative shifts of resonant peaks of the transmission spectra between the spin - up and spin - down components of electron can be modulated by the magnetic fields and on - site energies .
in this paper we demonstrate theoretically a method to generate and control spin currents in a three - terminal system .
the output spin currents are tunable in a wide range of magnitudes and various output configurations by adjusting the energy and the direction angles of the magnetic fields as well
. moreover it is demonstrated by the numerical evaluation that the spin currents remain in the presence of the disorder of qd - array .
particularly the spin currents with opposite spin polarizations in the two output electrodes can be generated .
thus this device is , as a matter of fact , a spin splitter similar with the light beam polarimeter in optics .
this observation may have practical application in the fabrication of spintronic devices for logic gates with spin - up and spin - down outputs regarding as qubits .
this work was supported by national natural science foundation of china under grant nos . | we in this paper study theoretically the spin - polarized quantum transport through a @xmath0-shape quantum dot - array by means of transfer - matrix method along with the green@xmath1s function technique .
multi - magnetic fields are used to produce the spin - polarized transmission probabilities and therefore the spin currents , which are shown to be tunable in a wide range by adjusting the energy , and the direction - angle of magnetic fields as well . particularly the opposite- spin- polarization currents separately flowing out to two electrodes can be generated and thus the system acts as a spin splitter .
@[email protected]
pacs numbers : 73.25.-b ; 73.23.-b ; 73.21.la ; 73.50.bk |
it has been known for a long time @xcite that the two photon exchange ( tpe ) mechanism can generate the single - spin normal asymmetry ( ssna ) of electron scattering due to a nonzero imaginary part of the tpe amplitude @xmath0 , @xmath1 where symbol @xmath2 denotes the imaginary ( absorptive ) part .
the one - photon - exchange amplitude @xmath3 is purely real due to time - reversal invariance of electromagnetic interactions .
the first calculations of the tpe effect on the proton @xcite predicted the magnitude of beam ssna at the level of a few parts per million ( ppm ) .
the effect appears to be small due to two suppression factors combined : @xmath4 , since the effect is higher - order in the electromagnetic interaction , and the electron mass @xmath5 arising due to electron helicity flip .
the predictions of ref.@xcite which correspond only to elastic intermediate proton state are in qualitative agreement with experimental data from mit / bates @xcite and they were reproduced later in ref.@xcite . however , the main theoretical problem in description of the tpe amplitude on the proton is a large uncertainty in the contribution of the inelastic hadronic intermediate states . in ref.@xcite
the beam ssna at large momentum transfers was estimated at the level of one ppm , using the partonic picture developed in ref.@xcite for tpe effects but not related to the electron helicity flip .
current experiments designed for parity - violating electron scattering allow to measure the beam asymmetry with a fraction of ppm accuracy @xcite and may also provide data on the parity - conserving beam ssna .
in fact , such measurements are needed because beam ssna is a source of systematic corrections in the measurements of parity - violating observables .
it was noted in @xcite that while considering excitation of inelastic intermediate hadronic states , the beam ssna ( eq.(11 ) of ref.@xcite ) , after factoring out the electron mass , has an enhancement when at least one of the photons in the tpe loop diagram is collinear to its parent electron .
it is interesting that this effect did not appear for the target ssna .
similar behaviour of the beam ssna in the nucleon resonance region was observed also in @xcite where authors used a phenomenological model ( maid ) for single - pion electroproduction . in this paper
we demonstrate that collinear photon exchange in the tpe amplitude results in single- and double - logarithmic enhancement of the beam ssna , whereas such enhancement does not take place for the target ssna ( with unpolarized electrons ) and spin correlations caused by longitudinal polarization of the scattering electrons . for large electron energies and small scattering angles , we use an optical theorem to relate the nucleon compton amplitude to the total photoproduction cross section and
obtain a simple analytic formula for the beam ssna in this kinematics .
first , we write the formula for ssna in terms of rank-3 leptonic and hadronic tensors which appear in the interference between the born and tpe amplitudes as shown in fig.[diagrams ] .
@xmath6 where @xmath7 , @xmath8 is the 3-momentum ( energy ) of the intermediate on - mass - shell electron in the tpe box diagram , @xmath9 and @xmath10 are the 4-momenta of the intermediate photons , @xmath11 .
the factor @xmath12 in eq.([2 ] ) is due to the squared born amplitude , namely , @xmath13\big(f_1 ^ 2+\frac{q^2}{4m^2 } f_2 ^ 2\big ) , \end{aligned}\ ] ] where @xmath14 is the dirac ( pauli ) proton form factor , @xmath15 is the proton mass and @xmath16 our sign convention for the beam asymmetry follows from the definition of the normal vector with respect to the electron scattering plane : @xmath17 using the above notation , we have @xmath18 and @xmath19 where @xmath5 is the electron mass , @xmath20 is the polarization 4-vector of the electron beam ( proton target ) , @xmath21 , and @xmath22 is in general a non - forward proton compton tensor that describes any possible hadronic intermediate states in the tpe amplitude . in accordance with eq.(2 )
the single - spin normal asymmetry probes the imaginary part of contraction of the leptonic and hadronic tensors defined by eqs.([4 ] ) and ( [ 5 ] ) , respectively .
these tensors satisfy the conditions @xmath23 separately for spin independent and spin dependent parts , as follows from gauge invariance of electromagnetic interactions . after some algebra
we arrive at the following expression for the model - independent leptonic tensor @xmath24 where the spin
independent part is @xmath25_{\alpha\beta}+ \frac{1}{2}g_{\alpha\beta}(q_1 ^ 2k_{2\mu}+q_2 ^ 2k_{1\mu}-q^2k_{\mu})+ \\ & & \frac{1}{2}q^2 ( g_{\mu\alpha}k_{\beta}+g_{\mu\beta}k_{\alpha } ) + k_{2\mu}(kk_1)_{\alpha\beta } + k_{1\mu}(kk_2)_{\alpha\beta}\ , \nonumber \\ & & [ ab]_{\alpha\beta } = a_{\alpha}b_{\beta } - a_{\beta}b_{\alpha}\ , \ \ ( ab)_{\alpha\beta } = a_{\alpha}b_{\beta } + a_{\beta}b_{\alpha}\ , \nonumber\end{aligned}\ ] ] and the spin
dependent part is given by @xmath26\ , \\ & & ( abcd)\equiv\epsilon_{\nu\lambda\rho\sigma}a_{\nu}b_{\lambda}c_{\rho } d_{\sigma}\ , \nonumber\end{aligned}\ ] ] where the on - shell condition @xmath27 was used for the intermediate electron 4-momentum .
if one of photon in the box diagram is collinear to its parent electron , for example , @xmath28 where @xmath29 is the squared invariant mass of the intermediate hadronic system , leptonic tensor can be written as @xmath30 the tensor @xmath31 coincides with the born one of elastic electron proton scattering and @xmath32 in the case of longitudinal polarization of the electron beam @xmath33 the tensor @xmath34 is zero .
therefore , we expect no contribution from considered kinematics to the target ssna or to longitudinal - spin correlations because any gauge invariant hadronic tensor has to give zero after contracting with @xmath35 ( see eq.(6 ) ) . in the case of the normal polarized electron beam @xmath36}}\ .\ ] ] tensor @xmath34 is not zero and the considered collinear photon kinematics contributes with essential logarithmic enhancement . therefore , conservation of the electromagnetic current that follows from gauge invariance ( eq.([6 ] ) ) is the reason why the collinear intermediate photons appear in the tpe contribution to the beam ssna , but not to the target ssna . by analogy ,
we do not anticipate contributions from collinear - photon exchange in unpolarized electron - proton scattering , parity - conserving and parity - violating asymmetries due to longitudinal electron polarization the normal polarization of leptons is not involved .
let us consider the hadronic tensor .
small values of @xmath37 correspond to the forward limit of nucleon virtual compton amplitude @xmath22 . on the other hand , because @xmath38 and @xmath39 are also small in the collinear photon limit , we can relate the forward compton amplitude to the total photoproduction cross section by real photons .
a general form of the compton tensor @xmath22 in terms of 18 independent invariant amplitudes that are free from kinematical singularities and zeros was derived in ref.@xcite . among these amplitudes
we choose the ones that contribute at the limit @xmath40 and @xmath41 it automatically constrains virtuality of the second photon to @xmath42 .
there is only one structure that contains the tensor @xmath43 and does not die off under the considered conditions .
it reads @xcite @xmath44 a(q_1 ^ 2,q_2 ^ 2,q^2,w^2 ) , \\ & & \bar{p}=\frac{1}{2}(p_1+p_2 ) , \ \
\bar{q } = \frac{1}{2}(q_1+q_2)\ .\nonumber\end{aligned}\ ] ] it can be verified that @xmath22 defined by the above equation satisfies the conditions @xmath45 the normalization convention is chosen such that the imaginary part of the quantity @xmath46 is connected with the inelastic proton structure function @xmath47 by the following relation @xmath48 @xmath49 and @xmath50 , in turn , defines the total photoproduction cross section @xcite as @xmath51 keeping in mind that the main contribution to the beam ssna arises from collinear photon kinematics , we can combine relations ( 5 ) , ( 14 ) and ( 15 ) and write hadronic tensor in the following form @xmath52_{\alpha\beta}}{w^2-m^2}\big)\ , \ ] ] @xmath53 when writing this last expression we neglect the terms containing @xmath54 which lead to the contribution of the order @xmath55 in the beam ssna .
it may seem at first that in the considered limiting case of very small @xmath37 one can omit all terms proportional to @xmath56 in the hadronic tensor .
but such approximation is valid only for the symmetric part of @xmath22 with respect to the indexes @xmath57 and @xmath58 the reason is that the corresponding symmetric part of the leptonic tensor ( see eq.([9 ] ) ) contains the momentum transfer @xmath56 , and keeping it in the symmetric part of hadronic tensor leads after contraction to additional small terms of the order at least @xmath59 on the other hand , the antisymmetric part of leptonic tensor contains terms which do not include the momentum @xmath60 therefore , the antisymmetric part in eq.([17 ] ) has to be retained because it contributes at the same order with respect to @xmath59 note , however , that this antisymmetric part of the hadronic tensor is not related to the polarized nucleon structure functions , but it comes about as a consequence of the gauge - invariant structure of eq.([14 ] ) even for a spinless hadronic target .
to compute the contraction of tensors in eq.(2 ) we use the relations @xmath61,\ ] ] @xmath62_{\alpha\beta}p_{1\mu}l^{(pol)}_{\mu\alpha\beta } = im_e(u - s)(p_1 q q_1 \xi^e),\ ] ] @xmath63 , which are valid for the normal beam polarization @xmath64 , and the explicit form of 4vector @xmath65 given by eq.([13 ] ) .
then we arrive at @xmath66 where @xmath67 is the total photoproduction cross section with the transverse virtual photons .
when integrating with respect to @xmath29 we take @xmath68 and assume @xmath69 to be constant with energy ( @xmath70 0.1 mb , according to ref.@xcite ) . taking into account eqs .
( 2 ) and ( 18 ) one can write the beam ssna at small values of @xmath37 as @xmath71 @xmath72 here and further we use notation where @xmath73 are the energies ( 3-momenta ) of the initial and intermediate electron , respectively . the angular integration in eq . ( 19 )
can be done by introducing the feynman parameter @xmath74
^ 2}\ , \ ] ] @xmath75 @xmath76 integration over @xmath77 and feynman parameter @xmath78 is straightforward , leading to @xmath79 where @xmath80 , and @xmath81 we extended the upper limit up to @xmath82 because the difference between the value of @xmath83 at inelastic threshold of pion production ( when @xmath84 ) and @xmath85 is negligible at large @xmath86 to calculate the integral in eq.([20 ] ) , we note first that the region where @xmath87 does not contribute because of the factor of @xmath88 . for this reason we can change integration with respect to @xmath83 by integration over @xmath89
then we divide the integration region into the following two parts , @xmath90 and @xmath91 , and choose the auxiliary parameter @xmath92 in such a way that @xmath93 in the first region we can neglect @xmath83 as compared with @xmath85 and write the corresponding contribution in the form @xmath94 we see that the contribution from the first region is negligible , therefore we can choose zero for the lower limit of integration in @xmath95 . in the second region
the quantity @xmath96 that enters @xmath88 is small and we have @xmath97 the integration in eq.([23 ] ) gives @xmath98 in the sum @xmath99 we arrive at the following master formula that defines the beam ssna for small values of @xmath37 and takes into account contributions from intermediate collinear photons in the tpe box diagrams @xmath100 one can see that at fixed values of @xmath37 the beam ssna does not depend on the beam energy if the total photoproduction cross section is energy - independent .
this remarkable property of small - angle beam ssna follows from unitarity of the scattering matrix and does not rely on a specific model of nucleon structure .
the master formula for beam ssna eq.([24 ] ) neglects possible @xmath37 dependence of the invariant form factor of the nucleon compton amplitude , which was taken in its forward limit during the derivation . in numerical calculations
, we estimate additional @xmath37 dependence by introducing an empirical form factor that was measured experimentally in the compton scattering on the nucleon in the diffractive regime ( see @xcite for review ) . in the following , we use an exponential suppression factor for
the nucleon compton amplitude @xmath101 , choosing the parameter @xmath102=8 gev@xmath103 that gives a good description of the nucleon compton cross section from the optical point to @xmath104 0.8 gev@xmath105 ( see table v of ref.@xcite ) .
the predictions of eq.([24 ] ) combined with the above described exponential suppression are presented in fig.[fig : thdep ] for the electron scattering kinematics relevant for the e158 experiment at slac @xcite .
we choose fit 1 of ref.@xcite for the total photoproduction cross section in eq.([24 ] ) .
exact numerical loop integration of eq.([2 ] ) and the analytic results of eq.([24 ] ) agree with each other with accuracy better than 1% .
contributions from the resonance region ( @xmath106 4 gev@xmath105 ) were estimated at 10 - 20% at beam energies of 3 gev , but rapidly decreasing below 1% at higher energies .
we also tested sensitivity of our results to @xmath107 dependence of the electroproduction structure function @xmath50 ( eq.([15 ] ) ) , taking various empirical parameterizations for it .
we found no sensitivity for slac e158 kinematics and only moderate sensitivity ( @xmath70 10% ) when we extend our calculation to lower energies ( @xmath703 gev ) and higher @xmath1080.5 gev@xmath105 . for beam energies of 45 gev
, numerical integration shows that more than 95% ( 80% ) of the result for beam ssna comes from the upper 1/2 ( 3/4 ) part of the @xmath29-integration range .
based on the results of numerical analysis , we conclude that the formula ( [ 24 ] ) gives a good description of beam ssna at small @xmath37 and large @xmath109 above the resonance region . for different beam energies .
notation is as in fig.[fig : thdep].,width=264 ] we also calculated the contribution of the elastic intermediate proton state to the beam ssna for high energies and small electron scattering angles using the formalism of ref.@xcite and found it to be highly supressed compared to the inelastic excitations . for the kinematics of slac e158 @xcite ,
this supression is a few orders of magnitude due to different angular and energy behavior of these contributions . shown in fig.[fig : q2dep ]
are the calculations for beam ssna as a function of @xmath37 for different energies of incident electrons .
one can see that at small @xmath37 , the asymmetry follows @xmath110 behavior desribed by eq.([24 ] ) , while at higher @xmath37 the asymmetry turns over and starts to decrease due to the introduced exponential form factor @xmath101 .
it can be seen that at fixed @xmath37 the magnitude of beam ssna is predicted to be approximately constant , as follows from slow logarithmic energy dependence of the total photoproduction cross section .
= 0.05 gev@xmath105 for different parameterizations of the total photoproduction cross section .
see fig.[fig : sigtot ] for notation.,width=264 ] the latter feature is demonstrated in fig.[fig : sqsdep ] , showing the calculated beam ssna at fixed @xmath37 in a wide energy range up to @xmath111=500 gev , where we used several parameterizations for the total photoproduction cross section on a proton from refs.@xcite , shown in fig.[fig : sigtot ] .
the physical reason for the almost constant photoproduction cross sections at high energies is believed to be soft pomeron exchange @xcite , therefore the beam ssna in the considered kinematics is sensitive to the physics of soft diffraction .
the predicted @xmath37 and energy dependence of beam ssna , along with its relatively large magnitude , is quite different from the model expectations assuming that no hadronic intermediate states are excited in the tpe amplitude .
our unitarity - based model of small - angle electron scattering predicts the magnitude of the beam ssna to reach a few ppm in a wide range of beam energies .
the good news is that it makes beam ssna measurable with presently reached fraction - of - ppm precision of parity - violating electron scattering experiments @xcite . on the other hand ,
the experiments measuring parity - violating observables need to use special care to avoid possible systematic uncertainties due to the parity - conserving beam ssna .
fortunately , these effects can be experimentally separated using different azimuthal dependences of these asymmetries . in the present paper
we calculate the beam ssna for small values of @xmath37 and provide physics arguments for the dominance of contributions from collinear photons in the tpe mechanism . for electron energies above the nucleon resonance region and
small @xmath37 the contribution of collinear virtual photons leads to the beam ssna that is negative and has the order of @xmath112 , where @xmath113 is the total photoproduction cross section on the proton .
this quantity is multiplied by the factor of the order unity that includes a single - logarithm term .
the fact that the beam ssna does not decrease with the beam energy at fixed @xmath37 makes it attractive for experimental studies at higher energies , for example , the energies to be reached at jefferson lab after the forthcoming 12-gev upgrade of cebaf .
this work was supported by the us department of energy under contract de - ac05 - 84er40150 .
n.m . acknowledges hospitality of jefferson lab , where this work was completed .
barut and c. fronsdal , phys . rev . *
120 * , 1871 ( 1960 ) .
f. guerin , c. a. picketty , nuovo cim . * 32 * , 971 ( 1964 ) ; j. arafune , y. shimizu , phys .
d * 1 * , 3094 ( 1970 ) ; u. gnther , r. rodenberg , nuovo cim . a * 2 * , 25 ( 1971 ) ;
a. de rujula , j.m .
kaplan and e. de rafael , nucl .
b * 35 * , 365 ( 1971 ) ; t.v .
kukhto et al . , preprint jinr
e292556 , feb .
1993 .
k. hagiwara _ et al .
_ [ particle data group collaboration ] , phys . rev .
d * 66 * , 010001 ( 2002 ) .
t. h. bauer , r. d. spital , d. r. yennie and f. m. pipkin , rev .
* 50 * , 261 ( 1978 ) [ erratum - ibid .
* 51 * , 407 ( 1979 ) ] . | the parity - conserving single - spin beam asymmetry of elastic electron - proton scattering is induced by an absorptive part of the two - photon exchange amplitude .
we demonstrate that this asymmetry has logarithmic and double - logarithmic enhancement due to contributions of hard collinear quasi - real photons .
an optical theorem is used to evaluate the asymmetry in terms of the total photoproduction cross section on the proton , predicting its magnitude at a few parts per million for high electon beam energies and small scattering angles . at fixed 4-momentum transfers ,
the asymmetry is rising logarithmically with increasing electron beam energy , following the high - energy diffractive behavior of total photoproduction cross section on the proton . |
the study of charmed meson decays are important for improving our knowledge of the standard model ( sm ) . in particular
, the study of leptonic and semileptonic decays allow us to measure the ckm matrix elements @xmath4 and @xmath5 to a high level of precision .
leptonic and semileptonic decays are also used to test theoretical predictions describing the strong interaction ( qcd ) in heavy quark systems .
charmed meson decays are a good laboratory to test lattice qcd ( lqcd ) predictions , which can then be applied with confidence to bottom meson decays .
improved theoretical predictions will not only lower the uncertainty in the ckm matrix elements mentioned above but also @xmath6 and @xmath7 .
leptonic decays of heavy mesons involve the annihilation of the constituent quarks into a neutrino and charged lepton via a virtual @xmath8 boson .
the probability for the annihilation is proportional to the wavefunctions of the quarks at the point of annihilation and is incorporated within the decay constant of the meson .
the leptonic decay partial width of the @xmath9 meson within the sm is given by @xcite @xmath10 where @xmath11 is the fermi coupling constant , @xmath0 ( @xmath1 ) is the @xmath12 ( @xmath13 ) decay constant , @xmath14 ( @xmath15 ) is the @xmath12 ( @xmath13 ) mass , @xmath16 is the mass of the charged lepton , and @xmath5 ( @xmath4 ) is the @xmath17 ( @xmath18 ) ckm transition matrix element .
these decays are helicity suppressed and , based on eq .
[ eq : lepdecpartwidth ] , the sm predicts @xmath19 and @xmath20 , with @xmath21 , respectively .
deviations from these predictions may arise from physics beyond the sm .
recent measurements of leptonic decays of the @xmath12 and @xmath13 are discussed in the following subsections .
the bes collaboration reported a measurement of the @xmath23 branching fraction using 33 pb@xmath24 of @xmath25 annihilation data collected on the @xmath26 resonance with the bes ii detector @xcite .
events are selected using the mark iii `` @xmath27-tagging '' method @xcite , which consists of fully reconstructing the @xmath28 meson from @xmath29 events and studying the @xmath27 decay . in (
semi)leptonic decays , the neutrino in inferred from the missing four - momentum in the event . using a sample of 5300
tagged @xmath30 decays ( note that charge conjugation is implied unless otherwise stated ) from 9 tag modes and requiring one additional particle consistent with a muon , 3 candidates events are observed with 0.3 background events .
this leads to a branching fraction of @xmath31 , where the first error is statistical and the second is systematic .
the cleo collaboration reported an updated measurement of the @xmath23 branching fraction @xcite . using a 281 pb@xmath24 data sample collected at the @xmath26 resonance with the cleo - c detector , a sample of 158,000 @xmath30 decays from 6 tag modes
was studied using the @xmath27-tagging method described above .
candidate events are required to have one charged track of opposite charge to the tagged @xmath30 and the track needs to deposit an energy in the electromagnetic calorimeter ( @xmath32 ) @xmath33 mev , which is consistent with a minimum ionizing particle .
events with an isolated photon - like shower in the calorimeter with an energy in excess of 250 mev are rejected .
figure [ fig : cleodtomunu ] shows the missing mass squared distribution for the candidate events , where the missing mass squared is defined as @xmath34 , where @xmath35 is the beam energy , @xmath36 is the momentum of the tagged @xmath30 , and @xmath37 ( @xmath38 ) is the energy ( momentum ) of the muon .
the signal region , defined by @xmath39 gev@xmath40 , contains 50 candidate and 2.8 background events , of which 1.1 events are determined from the sm prediction for @xmath41 .
the result corresponds to @xmath42 .
distribution for @xmath23 decays @xcite .
the inset shows the signal region , defined by the arrows , in more detail .
the background peak at @xmath43 is consistent with @xmath44 decays and is substantially displayed from the signal region . ]
a complementary cleo analysis searched for @xmath45 decays @xcite . in this case , the charged track is required to be consistent with a positron .
no signal is observed and an upper limit of @xmath46 at 90@xmath47 confidence level ( c.l . )
is obtained .
the cleo collaboration has reported on the search for @xmath41 using the same 281 pb@xmath24 and tagged @xmath30 sample @xcite . for minimum ionizing pions with @xmath48 mev , the signal region
is defined as 0.05 @xmath49 0.175 gev@xmath40 , above @xmath23 signal region and below the @xmath44 background ( see fig . [
fig : cleodtomunu ] ) .
pions which hadronize in the calorimeter are studied by requiring the track to have @xmath50 mev , be inconsistent with a positron , and reside in the region -0.05 @xmath49 0.175 gev@xmath40 .
events are rejected in both cases if it contains an isolated photon - like shower with an energy in excess of 250 mev or if the candidate pion is more consistent with being a charged kaon .
no significant enhancement is observed in either case , resulting in an upper limit of @xmath51 ( 90% c.l . ) . using the cleo result for @xmath23 above , an upper limit for the ratio @xmath52 of 1.8 times that of the sm prediction is placed .
the babar collaboration analyzed a 230.2 fb@xmath24 data sample taken at or near the @xmath54 resonance to study the decay process @xmath55 @xcite .
they studied the process @xmath56 , where @xmath57 are 16 fully reconstructed @xmath58 , and @xmath59 decays and @xmath60 , using a method similar to an earlier cleo measurement @xcite .
figure [ fig : babardstomunu ] shows the resultant @xmath61 distribution . with @xmath62 signal events ,
they determine @xmath63 .
using the recent babar result @xmath64 @xcite , they determine a branching fraction of @xmath65 , where the last error arises from the @xmath66 normalization uncertainty .
distribution for @xmath55 decays @xcite .
the solid line is the fitted signal and background distribution , while the dashed line is the background distribution alone . ]
distributions for @xmath53 decays , where @xmath67 @xcite .
figure [ fig : dstomutaunu]i contains tracks with @xmath48 mev , fig .
[ fig : dstomutaunu]ii contains tracks with @xmath50 mev and inconsistent with the positron hypothesis , while fig .
[ fig : dstomutaunu]iii contains tracks which are consistent with the positron hypothesis . ] as part of the cleo - c run plan , the energy region 3.97 - 4.26 gev was scanned to determine the best energy for @xmath68 production .
the energy @xmath69 gev was found to be the location with the highest production rate of @xmath68 mesons , but the @xmath68 mesons are predominately produced in the processes @xmath70 .
the tagging procedure becomes more complicated by the presence of the transition photon from the radiative decay process @xmath71 . using a 195 pb@xmath24 data sample collected near @xmath69 gev ,
the cleo collaboration reported preliminary branching fractions for @xmath53 decays , where @xmath67 , and with the tau decaying via @xmath72 @xcite .
events are selected with 8 @xmath73 tag modes and the detection of the transition photon .
the missing mass squared for @xmath13 is defined as @xmath74 , where @xmath75 is the center - of - mass energy , @xmath76 ( @xmath77 ) is the energy ( momentum ) of the tagged @xmath73 , @xmath78 ( @xmath79 ) is the energy ( momentum ) of the transition photon , and @xmath80 ( @xmath81 ) is the energy ( momentum ) of the candidate track .
the signal side , consisting of one charged track , is reconstructed using the same criteria as the cleo @xmath22 analyses described above , but with the exceptions of increasing the @xmath82 and @xmath83 @xmath84 signal regions to 0.20 gev@xmath40 and increasing the maximum shower energy to 300 mev .
figure [ fig : dstomutaunu ] shows the @xmath84 distributions
. signals are apparent in figs .
[ fig : dstomutaunu]i and [ fig : dstomutaunu]ii , while no signal is observed near @xmath85 for the positron decay channel in fig .
[ fig : dstomutaunu]iii .
the @xmath55 signal region contains 64 candidate and 7.3 background events , of which 5.3 events are determined from the sm prediction for @xmath86 .
using events with candidate tracks inconsistent with the positron hypothesis and outside of the @xmath55 signal region , 36 candidate and 4.8 background @xmath83 events are found .
the preliminary branching fractions are @xmath87 ( 90% c.l . ) , @xmath88 , and @xmath89 . [ cols="<,<,<,<",options="header " , ] + @xmath90preliminary results , not to be averaged .
using the unquenched lqcd prediction for @xmath91 @xcite allows for an experimental measurement of the ckm matrix elements @xmath5 and @xmath4 . the preliminary @xmath4 result from the cleo tagged ( untagged ) analysis is @xmath92 ( @xmath93 ) , where the last uncertainty arise from the theoretical uncertainty in @xmath94 .
the preliminary cleo result for @xmath5 is @xmath95 ( @xmath96 ) .
for comparison , the 2006 global fit values are @xmath97 and @xmath98 @xcite .
the experimental uncertainties for @xmath4 and @xmath5 are at the 2@xmath47 and 4@xmath47 level , respectively , while the theoretical uncertainty is on the order of 10@xmath47 .
when future theoretical predictions achieve higher precision , semileptonic decays will be an ideal environment to precisely measure @xmath5 and @xmath4 .
while the semileptonic @xmath27 decays to vector mesons are an additional method to measure @xmath4 and @xmath5 , they are complicated by the presence of three form factors associated with the three helicity states of the final state meson .
the spectroscopic pole dominance model proposes to parametrize the vector mesons in terms of a vector and two axial vector form factors , which are defined as @xmath100 where @xmath101 = 2.1 gev and @xmath102 = 2.5 gev .
the normalized form factors are defined by two ratios , the vector to first axial vector @xmath103 and the second to first axial vector @xmath104 .
the cleo collaboration reported preliminary results for the decay processes @xmath106 using the 281 pb@xmath24 @xmath26 data sample and the @xmath27-tagging technique @xcite .
they determine the most precise branching fractions of @xmath107 and @xmath108 . using the world average for the @xmath109 lifetimes @xcite , they determine the partial width ratio @xmath110 .
they also performed a simultaneous form factor analysis of the @xmath111 and @xmath12 decays and find @xmath112 and @xmath113 .
this analysis represents the first form factor measurement of a cabibbo suppressed vector decay mode in the charm system . the cleo collaboration , using the 281 pb@xmath24 @xmath26 data sample and the @xmath27-tagging technique , reported a non - parametric form factor analysis using the focus method @xcite for the decay process @xmath115 @xcite .
cleo confirms the presence of the @xmath116-wave interference in the @xmath117 final state and determines results for the form factors to be consist with the focus analysis .
cleo did not observe any evidence of @xmath118- or @xmath119-wave interference in the @xmath117 system .
the recent experimental results for leptonic and semileptonic decays of charm mesons are beginning to improve their accuracy to point where theoretical errors are dominating the uncertainty in the determination of the @xmath4 and @xmath5 ckm matrix elements .
these measurements provide stringent tests of theoretical models , which in turn will improve the models as to lower their uncertainty but are also used to `` fine - tune '' the models so they can be applied with confidence to beauty leptonic and semileptonic decays
. the experimental precision will continue to improve from the final cleo - c @xmath27 and @xmath68 data samples and the beginning of data collection with the upgraded bes - iii and bepc - ii facilities in 2008 .
i wish to thank neville harnew , guy wilkinson , and all of the organizers for an engaging and enjoyable conference .
this work was supported by the u.s
. national science foundation and department of energy .
99 d. silverman and h. yao , phys .
* d38 * ( 1988 ) 214 .
m. ablikim _
et al . _ ,
bes collaboration , phys .
b610 * ( 2005 ) 183 . j. adler _
et al . _ ,
mark iii collaboration , phys .
* 62 * ( 1989 ) 1821 .
m. artuso _
et al . _ ,
cleo collaboration , phys .
* 95 * ( 2005 ) 251801 .
et al . _ ,
cleo collaboration , phys . rev . *
d73 * ( 2006 ) 112005 .
b. aubert _
_ , babar collaboration , arxiv : hep - ex/0607094 , subm . to phys .
m. chadha _
et al . _ ,
cleo collaboration , phys . rev . *
d58 * ( 1998 ) 032002 .
b. aubert _
et al . _ ,
babar collaboration , phys . rev .
* d71 * ( 2005 ) 091104 .
m. artuso _
et al . _ , cleo collaboration , contributed to the 33rd international conference on high energy physics ( ichep06 ) , moscow , russia , arxiv : hep - ex/0607074 . s. stone ( cleo collaboration ) , presented at ichep06 , moscow , russia , arxiv : hep - ex/0610026 . c. aubin _ et al .
_ , fermilab lattice , milc , and hpqcd collaborations , phys .
* 95 * ( 2006 ) 122002 .
et al . _ ,
b624 * ( 2005 ) 31 .
l. lellouch and c .- j .
d. lin , phys .
* d64 * ( 2001 ) 094501 .
d. becirevic _ et al .
_ , phys . rev .
* d60 * ( 1999 ) 074501 .
j. gordes , j. penarrocha , and k. schilcher , j. high energy phys . *
0511 * ( 2005 ) 014 .
s. narison , arxiv : hep - ph/020220 .
d. ebert , r. n. faustov , and v. o. galkin , phys .
b635 * ( 2006 ) 93 .
g. cvetic _
et al . _ ,
b596 * ( 2004 ) 84
. l. a. m. salcedo _
et al . _ ,
* 34 * ( 2004 ) 297 .
j. f. amundson _
et al . _ , phys .
* d47 * ( 1993 ) 3059 .
n. e. adam _ et al .
_ , cleo collaboration , phys .
* 97 * ( 2006 ) 251801 .
et al . _ , j. phys .
* g33 * ( 2006 ) 1 . y. gao ( cleo collaboration ) , presented at ichep06 , moscow , russia .
d. scora and n. isgur , phys .
* d52 * ( 1995 ) 2783 .
s. fajfer and j. kamenik , phys . rev .
* d71 * ( 2005 ) 014020 .
s. eidelman _
et al . _ ,
b592 * ( 2004 ) 1 .
m. ablikim _
et al . _ ,
bes collaboration , phys . lett .
* b597 * ( 2004 ) 39 .
t. e. coan _
et al . _ ,
cleo collaboration , phys .
* 95 * ( 2005 ) 181802 .
l. widhalm _
et al . _ ,
belle collaboration , phys .
* 97 * ( 2006 ) 061804 .
r. poling ( cleo collaboration ) , presented at the 4th flavor physics and cp violation conference ( fpcp06 ) , vancouver , british columbia , canada , arxiv : hep - ex/0606016 . c. aubin _ et al .
_ , fermilab lattice , milc , and hpqcd collaborations , phys . rev .
* 94 * ( 2005 ) 011601 .
d. becirevic and a. b. kaidalov , phys . lett .
* b478 * ( 2000 ) 417 .
t. becher and r. j. hill , phys .
b633 * ( 2005 ) 61 .
r. j. hill , presented at fpcp06 , vancouver , british columbia , canada , arxiv : hep - ph/0606023 .
j. m. link _ et al .
_ , focus collaboration , phys
b607 * ( 2005 ) 233 . g. s. huang _
_ , cleo collaboration , phys .
* 94 * ( 2005 ) 011802 .
b. aubert _
_ , babar collaboration , contributed paper to ichep06 , arxiv : hep - ex/0607077
. j. m. link _ et al .
_ , focus collaboration , phys .
b633 * ( 2006 ) 183 . m. r. shepherd _
_ , cleo collaboration , phys . rev .
* d74 * ( 2006 ) 052001 . | experimental results for the pseudoscalar decay constants @xmath0 and @xmath1 are reviewed .
semileptonic form factor results from @xmath2 and @xmath3 decays are also reviewed . |
tio@xmath0 has been one of the most studied oxides over the past few years .
this is due to the broad range of applications it offers in strategic fields of scientific , technological , environmental , and commercial relevance .
in particular , tio@xmath0 surfaces and nanocrystals provide a rich variety of suitably tunable properties from structure to opto - electronics .
special attention has been paid to tio@xmath0 s optical properties .
this is because tio@xmath0 is regarded as one of the best candidate materials to efficiently produce hydrogen via photocatalysis @xcite .
tio@xmath0 nanostructures are also widely used in dye - sensitized solar cells , one of the most promising applications in the field of hybrid solar cells @xcite . since the first experimental formation of hydrogen by photocatalysis in the early 1980s , @xcite
tio@xmath0 has been the catalyst of choice .
reasons for this include the position of tio@xmath0 s conduction band above the energy of hydrogen formation , the relatively long lifetime of excited electrons which allows them to reach the surface from the bulk , tio@xmath0 s high corrosion resistance compared to other metal oxides , and its relatively low cost @xcite .
however , the large optical band gap of bulk tio@xmath1 ( @xmath2 3 ev ) means that only high energy uv light may excite its electrons .
this effectively blocks most of the photons which pierce the atmosphere , typically in the visible range , from participating in any bulk tio@xmath1 based photocatalytic reaction . on the other hand ,
the difference in energy between excited electrons and holes , i.e. the band gap , must be large enough ( @xmath3 1.23 ev ) to dissociate water into hydrogen and oxygen .
for these reasons it is of great interest to adjust the band gap @xmath4 of tio@xmath1 into the range 1.23 @xmath5 2.5 ev , while maintaining the useful properties mentioned above @xcite . with this aim
, much research has been done on the influence of nanostructure @xcite and dopants @xcite on tio@xmath1 photocatalytic activity . for low dimensional nanostructured materials , electrons and holes
have to travel shorter distances to reach the surface , allowing for a shorter quasi - particle lifetime .
however , due to quantum confinement effects , lower dimensional tio@xmath6 nanostructures tend to have _ larger _ band gaps @xcite . on the other hand , although doping may introduce mid - gap states , recent experimental studies have shown that boron and nitrogen doping of bulk tio@xmath1 yields band gaps _ smaller _ than the threshold for water splitting @xcite .
this suggests that low dimensional structures with band gaps larger than about @xmath7 ev may be a better starting point for doping .
recently , several promising new candidate structures have been proposed @xcite .
these small ( @xmath8 5 ) tio@xmath1 nanotubes , with a hexagonal abc pto@xmath1 structure ( hexabc ) , were found to be surprisingly stable , even in the boron and nitrogen doped forms . this stability may be attributed to their structural similarity to bulk rutile tio@xmath9 , with the smallest nanotube having the same structure as a rutile nanorod .
a further difficulty for any photocatalytic system is controlling how electrons and holes travel through the system @xcite .
for this reason , methods for reliably producing both @xmath10-type and @xmath11-type tio@xmath1 semiconducting materials are highly desirable .
so far , doped tio@xmath1 tends to yield only @xmath10-type semiconductors .
however , it has recently been proposed that @xmath11-type tio@xmath0 semiconducting materials may be obtained by nitrogen doping surface sites of low dimensional materials @xcite . in this chapter
we will discuss in detail the effects of quantum confinement and doping on the optical properties of tio@xmath0 .
tio@xmath0 nanostructures are also one of the main components of hybrid solar cells . in a typical grtzel cell @xcite , tio@xmath0 nanoparticles with average diameters around 20 nm collect the photoelectron transferred from a dye molecule adsorbed on the surface @xcite .
such processes are favoured by a proper energy level alignment between solid and organic materials , although the dynamic part of the process also plays an important role in the charge transfer .
clearly , tio@xmath0 s characteristics of long quasi - particle lifetimes , high corrosion resistance , and relatively low cost , must be balanced with control of its energy level alignment with molecular states , and a fast electron injection at the interface . despite all the engineering efforts ,
the main scientific goal remains to optimize the efficiency of solar energy conversion into readily available electricity .
different research approaches have been devoted to benefit from quantum size properties emerging at the nanoscale @xcite , find an optimal donor - acceptor complex @xcite , mix nanoparticles and one - dimensional structures , such as nanotubes or nanowires @xcite , and control the geometry of the tio@xmath0 nano - assembly @xcite .
a clear theoretical understanding of tio@xmath0 s optoelectronic properties is necessary to help unravel many fundamental questions concerning the experimental results . in particular
, the properties of excitons , photo - injected electrons , and surface configuration in tio@xmath0 nanomaterials may play a critical role in determining their overall behaviour in solar cells . for tio@xmath0 at both the nanoscale and macroscale regime , it is necessary to first have a complete picture of the optical properties in order to clarify the contribution of excitons .
octahedrons ( right ) of rutile ( top ) and anatase ( bottom ) .
the lattice parameters in are denoted @xmath12 , @xmath13 , and @xmath14 , while @xmath15 and @xmath16 are the distances in between a ti@xmath17 ion and its nearest and next - nearest neighbour o@xmath18 ions , respectively . in the case of rutile
the interstitial ti impurity site is shown with a green circle ( see left top ) .
, scaledwidth=80.0% ] despite the clear importance of its surfaces and nanostructures , investigations of tio@xmath0 bulk ( see [ lfig1 ] ) electronic and optical properties have not provided , so far , a comprehensive description of the material .
important characteristics , such as the electronic band gap , are still undetermined .
most of the experimental and computational work has been focused on synthesis and analysis of systems with reduced dimensionality .
the experimental synthesis and characterization of nanostructured materials is in general a costly and difficult task .
however , predictions of a dye or nanostructure s properties from simulations can prove a great boon to experimentalists .
modern large - scale electronic structure calculations have become important tools by providing realistic descriptions and predictions of structure and electronic properties for systems of technological interest . although it will not be treated in this chapter
, it is important to mention the problem of electron localization in reduced titania @xcite .
this will provide a glimpse of the complexity faced , from the theoretical point of view , when studying transition metal atoms .
the localization of @xmath19 electrons makes the accurate description of their exchange
correlation interaction a difficult task @xcite .
the electron localization in defective titania has been an open question from both experimental and theoretical points of view , and caused much controversy during the past few years @xcite .
oxygen vacancies are quite common in tio@xmath0 , and their presence and behaviour can significantly affect the properties of nanostructures .
when an oxygen vacancy is created in tio@xmath0 , i.e. when tio@xmath0 is reduced , the two electrons coming from the removed o@xmath18 ion must be redistributed within the structure .
one possibility is that these two extra electrons remain localized onto two ti ions close to the o@xmath18 vacancy . in this way
a pair of ti@xmath17 ions become ti@xmath20 ions .
another option is for the two extra electrons to delocalize along the whole structure , i.e. they do not localize on any particular ti ion .
finally , an intermediate situation , with one electron localized and the other spread , is also possible . concerning the tio@xmath0 bulk , conventional density functional theory ( dft ) calculations using either local density approximations ( lda ) or generalized gradient approximations ( gga ) for the exchange - correlation ( xc)-functionals show a scenario with both electrons delocalized . on the other hand ,
hybrid functionals and hartree - fock calculations give rise to a situation with both electrons localized .
for gga+u calculations , the results are very sensitive to the value of the u parameter . for certain values of u
both electrons remain localized , while for others there is an intermediate situation @xcite . experimentally , there are electron paramagnetic resonance ( epr ) measurements suggesting that the extra electrons are mainly localized on interstitial ti@xmath20 ions @xcite .
these interstitial ti@xmath20 ions are impurities placed at the natural interstices of the rutile lattice ( see [ lfig1 ] ) and , similarly to the ti@xmath17 ions of the pure lattice , they also form tio@xmath21 octahedrons .
recent stm and pes experiments have shown that the interstitial ti@xmath20 ions play a key role in the localization of the electrons when a bridge oxygen is removed from the tio@xmath0 ( 110 ) surface @xcite .
these experiments concluded the controversial discussion about the localization of electrons in the bridge oxygen defective tio@xmath0 ( 110 ) surface ( see refs .
@xcite for more details ) .
however , the problem remains unresolved for the bulk case . in summary ,
in this chapter we first analyze the full _ ab initio _ treatment of electronic and optical properties in [ dft ] and [ mbpt ] , before applying it to the two most stable bulk phases , rutile and anatase in [ bulk ] .
these are also the phases most easily found when nanostructures are synthesized .
we will focus on their optical properties and excitonic behaviour .
we then explore the possibility of tuning the oxide band gap using quantum confinement effects and dimensionality , by analyzing atomic clusters , nanowires and nanotubes in [ nanostructure ] . a further component whose effect has to be evaluated
is that of doping , which may further tune the optical behaviour by introducing electronic states in the gap , as presented in [ doping].combining the effects of quantum confinement and doping is hoped to produce a refined properties control . in [ cells ] , we report some details on modeling for dye - sensitized solar cells , before providing a summary and our concluding remarks .
dft is a many - body approach , successfully used for many years to calculate the ground - state electronic properties of many - electron systems .
however , dft is by definition a ground state theory , and is not directly applicable to the study of excited states . to describe these types of physical phenomena
it becomes necessary to include many - body effects not contained in dft through green s function theory .
the use of many - body perturbation theory @xcite , with dft calculations as a zero order approximation , is an approach widely used to obtain quasi - particle excitation energies and dielectric response in an increasing number of systems , from bulk materials to surfaces and nanostructures .
we present a short , general discussion of the theoretical framework , referring the reader to the books and the reviews available in the literature for a complete description ( see , for instance , refs .
@xcite and @xcite ) , before applying these methods to rutile and anatase tio@xmath0 .
as originally introduced by hohenberg and kohn ( hk ) in 1964 @xcite , dft is based on the theorem that the ground state energy of a system of @xmath22 interacting electrons in an external potential @xmath23 is a unique functional of the ground state electronic density .
the kohn and sham @xcite formulation demonstrates how the the hk theorem may be used in practice , by self - consistently solving a set of one - electron equations ( ks equations ) , @xmath24\right ] \phi_i({\bf r } ) = \varepsilon_i^{ks } \phi_i({\bf r } ) , \label { eq : ks } \quad\rho({\bf r } ) = \sum_{i}^{n } \vert \phi_i({\bf r } ) \vert^2,\end{aligned}\ ] ] where @xmath25 is the electronic charge density , @xmath26 are the non - interacting ks wavefunctions , and @xmath27 = v_{ext}({\bf r } ) + \
v_h[\rho({\bf r } ) ] + v_{xc}[\rho({\bf r } ) ] $ ] is the effective one - electron potential . here , @xmath28 is the hartree potential and @xmath29 is the exchange correlation potential defined in terms of the xc - functional @xmath30 as @xmath31 , which contains all the many - body effects .
@xmath29 is usually calculated in either lda @xcite or gga @xcite approximations .
however , semi - empirical functionals are also available , called hybrids @xcite , which somewhat correct the deficiencies of lda and gga for describing exchange and correlation .
this is accomplished by including an exact exchange contribution .
other than the highest occupied molecular orbital ( homo ) , ks dft electronic levels do not correspond to the electronic energies of the many electron system .
indeed , the calculated ks band gaps of semiconductors and insulators severely underestimate the experimental ones
. experimentally , occupied states are accessible by direct photoemission , where an electron is extracted from the system ( @xmath32 ground state ) , while unoccupied states are accessible by inverse photoemission , where one electron is added to the system ( @xmath33 ground state ) . for isolated systems with a finite number of electrons
, the electronic gap may be obtained from the dft calculated ground - state energies with @xmath33 , @xmath22 , and @xmath32 electrons @xcite . however ,
for periodic systems , adding or removing an electron is a non - trivial task .
we therefore summarize in the following a rigorous method for describing excitations based on the green s function approach .
this method allows us to properly describe the electronic band structure .
further information and details about the green s function approach may be found elsewhere @xcite .
when a bare particle , such as an electron or hole , enters an interacting system , it perturbs the particles in its vicinity .
in essence , the particle is `` dressed '' by a polarization cloud of the surrounding particles , becoming a so - called quasi - particle . using this concept
, it is possible to describe the system through a set of quasi - particle equations by introducing a non - local , time - dependent , non - hermitian operator called the self - energy @xmath34 .
this operator takes into account the interaction of the particle with the system via @xmath35 \psi_i({\bf r},\omega ) + \int\sigma({\bf r},{\bf r}',\omega)\psi_i({\bf r}',\omega)d{\bf r } ' = e_i(\omega)\psi_i({\bf r } ) , \label{eq : quasi}\ ] ] where @xmath36 is the quasi - particle wavefunction .
since the operator @xmath34 is non - hermitian , the energies @xmath37 are in general complex , and the imaginary part of @xmath34 is related to the lifetime of the excited particle @xcite .
the most often used approximation to calculate the self - energy is the so - called @xmath38 method .
it may be derived as the first - step iterative solution of the hedin integral equations ( see refs .
@xcite , @xcite , and @xcite ) , which link the green s function @xmath39 , the self - energy @xmath34 , the screened coulomb potential @xmath40 , the polarization @xmath41 and the vertex @xmath42 . in practice [ eq :
quasi ] is not usually solved directly , since the ks wavefunctions are often very similar to the @xmath38 ones @xcite .
for this reason it is often sufficient to calculate the quasi - particle ( qp ) corrections within first - order perturbation theory ( @xmath43 ) @xcite .
moreover , the energy dependence of the self - energy is accounted for by expanding @xmath34 in a taylor series , so that the qp energies @xmath44 are then given by @xmath45 where @xmath46 are the quasi - particle renomalization factors described in refs .
@xcite and @xcite . for a large number of materials ,
the @xmath43 approximation of @xmath34 works quite well at correcting the ks electronic gap from dft .
concerning optical properties , the physical quantity to be determined in order to obtain the optical spectra is the macroscopic dielectric function @xmath47@xmath48(@xmath49 ) .
this may be calculated at different levels of accuracy within a theoretical _ ab initio _ approach .
a major component in the interpretation of the optical measurements of reduced dimensional systems are the local - field effects ( lfe ) .
these effects are especially important for inhomogeneous systems . here , even long wavelength external perturbations produce microscopic fluctuations of the electric field , which must be taken into account .
however , lfe are also important for bulk phases such as anatase and rutile tio@xmath0 .
they must be taken into account in the evaluation of optical absorption , and to calculate the screened interaction @xmath40 used in @xmath38 .
the effect becomes increasingly important when going to lower dimensional systems .
it is well known @xcite that for inhomogeneous materials @xmath50 is not simply the average of the corresponding microscopic quantity , but is related to the inverse of the microscopic dielectric matrix by @xmath51 the microscopic dielectric function may be determined within the linear response theory @xcite , the independent - particle picture by the random phase approximation ( rpa ) , and using eigenvalues and eigenvectors of a one - particle scheme such as dft or @xmath38 .
there is also a different formulation which includes lfe in the macroscopic dielectric function .
this becomes useful when the electron hole interaction is included in the polarization function .
this formulation allows us to include , via the bethe - salpeter equation ( bse ) , excitonic and lfe on the same footing . in so doing , inverting of the microscopic dielectric matrix
is avoided .
the complete derivation may be found in appendix b of ref .
@xcite .
so far , in rpa , we have treated the quasi - particles as non - interacting . to take into account the electron
hole interaction , a higher order vertex correction needs to be included in the polarization .
in other words , the bse , which describes the electron hole pair dynamics , needs to be solved . as explained in ref .
@xcite , the bse may be written as an eigenvalue problem involving the effective two - particle hamiltonian @xmath52 the kernel @xmath53 contains two contributions : @xmath54 , which is the bare coulomb interaction with the long range part corresponding to a vanishing wave vector not included and @xmath40 , the attractive screened coulomb electron hole interaction . using this formalism and considering only the resonant part of the excitonic hamiltonian @xcite , the macroscopic dielectric function may be expressed as @xcite @xmath55 in [ eps_exc ] the dielectric function , differently from the rpa approximation , is given by a mixing of single particle transitions weighted by the excitonic eigenstates @xmath56 .
these are obtained by the diagonalization of the excitonic hamiltonian . moreover ,
the excitation energies in the denominator are changed from @xmath57 to @xmath58 .
the electronic levels are mixed to produce optical transitions , which are no longer between pairs of independent particles .
the excitonic calculation is in general , from the computational point of view , very demanding because the matrix to be diagonalized may be very large .
the relevant parameters which determine its size are the number of @xmath59-points in the brillouin zone , and the number of valence and conduction bands , @xmath60 and @xmath61 respectively , which form the basis set of pairs of states .
calculations performed for insulators and semiconductors show that the inclusion of the electron
hole coulomb interaction yields a near - quantitative agreement with experiment .
this is not only true below the electronic gaps , where bound excitons are generally formed , but also above the continuum edge .
the same results apply to the titania - based materials investigated here , as shown in the following section .
despite the importance of its surfaces and nanostructures , the most recent measurements of tio@xmath0 s bulk ( see fig .
1 ) electronic and optical properties were performed in the 1960s , with a few exceptions . here
we aim to review the existing results obtained using a variety of experimental techniques and _ ab initio _ calculations , in order to elucidate the known properties of tio@xmath0 .
previous data will be compared with a complete , consistent _ ab initio _ description , which includes many body effects when describing electronic and optical properties .
most experimental and theoretical data reported refers to the rutile phase , while anatase in general has been less studied .
however , anatase has received more attention recently , due in part to its greater stability at the nanoscale compared to rutile .
as we will see , while a general agreement seems to exist concerning the optical absorption edge of these materials , values for basic electronic properties such as the band gap still have a large degree of uncertainty .
the electronic properties of valence states of rutile tio@xmath0 have been investigated experimentally by angle - resolved photoemission spectroscopy @xcite , along the two high symmetry directions ( @xmath62 and @xmath34 ) in the bulk brillouin zone .
the valence band of tio@xmath0 consists mainly of o 2@xmath11 states partially hybridized with ti 3@xmath19 states .
the metal 3@xmath19 states constitute the conduction band , with a small amount of mixing with o 2@xmath11 states .
this photoemission data was compared to calculations performed with both pseudopotentials and linear muffin - tin orbital ( lmto ) methods , which gave a direct gap of 2 ev in both cases . on the other hand , within the linear combination of atomic orbitals ( lcao ) method , a gap of 3 ev was obtained . from the symmetry of the tio@xmath21 octahedrons ( see [ lfig1 ] ) , @xmath19 states are usually grouped into low energy @xmath63 and high energy @xmath64 sub - bands .
it is important to note that , from ultraviolet photoemission spectroscopy ( ups ) data , it has been deduced that the electronic gap for rutile is at least 4 ev .
this is the observed binding energy of the first states below the fermi energy .
this is in agreement with previous reported data from electron energy loss spectroscopy @xcite , and from other ups @xcite results .
the electronic structure of rutile bulk has also been described using other experimental techniques , such as electrical resistivity @xcite , electroabsorption @xcite , photoconductivity and photoluminescence @xcite , x - ray absorption spectroscopy ( xas ) @xcite , resonant raman spectra @xcite photoelectrochemical analysis @xcite and ups @xcite .
all of these experiments have provided many important details of its electronic properties , in particular concerning the hybridization between ti 3@xmath19 and o 2@xmath11 states .
however , the electronic band gap , corresponding to the difference between the valence band maximum ( vbm ) and the conduction band minimum ( cbm ) , has not been obtained directly from any experimental data .
although the electronic band gap could be measured using combined photoemission and inverse photoemission experiments , such experiments do not appear in the literature .
the same discussion is valid for the anatase crystalline phase .
even though there are several xas measurements concerning its electronic structure @xcite , photoemission data is completely lacking for anatase . in the absence of more recent and refined experimental results for rutile , and due to the lack of results for anatase , we are left with an estimate of 4 ev for the electronic gap of rutile tio@xmath0 .
it is this value which we shall use as a reference in the following discussion .
we will now review the experimental results for optical properties of tio@xmath0 .
such measurements are of great interest for the photocatalytic and photovoltaic applications of this material . from the optical absorption spectra of both phases
@xcite , the room temperature optical band gap is found to be 3.0 ev for rutile , and 3.2 ev for anatase .
the absorption edge has been investigated in detail for rutile by combining absorption , photoluminescence , and raman scattering techniques @xcite .
these techniques provide a value for the edge of 3.031 ev associated to a 2@xmath65 exciton , while a lower energy 1@xmath66 quadrupolar exciton has been identified below 3 ev .
the first dipole allowed gap is at 4.2 ev @xcite according to the combined results of these three techniques .
concerning anatase @xcite , the optical spectrum have been recently re - evaluated @xcite , confirming the 3.2 ev value for the edge .
the fine details of anatase s spectrum have also been recently investigated @xcite .
data on the urbach tail has revealed that excitons in anatase are self - trapped in the octahedron of coordination of the titanium atom .
this is in contrast to the rutile phase , where excitons are known to be free due to the different packing of rutile s octahedra .
@xcite in general , measurements of optical properties can be significantly affected by the presence of defects , such as oxygen vacancies , and by phonons . both defects and phonons
will be present in any experimental sample of the material at finite temperatures .
these observations have to be kept in mind when directly comparing experimental measurements with the theoretical results presented in the following .
moreover , there is a general trend in theoretical - computational studies to compare the theoretical electronic band gaps with the experimental optical gap values @xcite derived from the above mentioned experiments .
it should be remembered that almost by definition , the optical gap is always smaller than the electronic band gap .
this is because the two types of experiments ( photoemission , and optical absorption ) provide information on two different physical quantities .
reverse photoemission experiments involve a change in the total number of electrons in the material ( @xmath67 ) , while optical absorption experiments do not ( @xmath68 ) .
the latter involves the creation of an electron hole pair in the material , with the hole stabilizing the excited electron .
for this reason , comparison between experimental and theoretical data , and the resulting discussion , must take into account the proper quantities .
the theoretical investigations presented in the literature of tio@xmath0 s structural , electronic , and optical properties are at varying levels of theory and thus somewhat inconsistent . a comprehensive description of properties of both phases in the same theoretical and computational framework is still missing .
here we present , in a unified description and by treating with the same method for the two phases , the electronic and optical properties of the two most stable phases of bulk titania . the _ ab intio _ calculations performed yield results @xcite in quite good agreement with the few available pieces of experimental data .
a combination of dft @xcite and many body perturbation theory ( mbpt ) methods is a reliable and well established toolkit to obtain a complete analysis of electronic and optical properties for a large class of materials and structures . in this dft + @xmath38 + bse framework ,
the properties of the two bulk phases of tio@xmath0 may be properly analyzed .
their structural and energetic properties have been calculated @xcite using dft , as a well established tool for the description of ground state properties . however , the dft gap is , as expected , significantly smaller than the experimental gap , with the relative positions of the @xmath66 , @xmath11 , and @xmath19 levels also affected by this description . to address this ,
standard @xmath43 calculations may be applied to obtain the quasi - particle corrections to the energy levels , starting from dft eigenvalues and eigenfunctions .
finally the electron
hole interaction is included , to properly describe the optical response of the system .
the description of ground state properties @xcite performed in the framework of dft are generally quite good , with the structural description of tio@xmath0 systems in reasonable agreement with experiments @xcite .
the lattice constants are within 2% of experiment , while bulk modulii are within 10% of the experimental results @xcite , as is often found for dft . however , dft incorrectly predicts the anatase phase to be more stable than the rutile one , even for a small energy difference , independently of the xc - functional used @xcite . * ) . and
including the @xmath43 correction ( ) .,scaledwidth=80.0% ] even if dft is not an excited state method , the ks wavefunctions are often used to evaluate the band structure along the high symmetry directions ( _ cf . _ [ lfig2 ] for rutile ) , the density of states , and the spatial behaviour of wavefunctions involved in the relevant bonds in the system @xcite .
the ks electronic gap , corresponding to the difference between the vbm and the cbm , is 1.93 ev and 2.16 ev for rutile ( direct gap ) and anatase ( indirect gap ) , respectively .
these are underestimations by almost 2 ev of the available experimental data @xcite .
however , the overall behaviour of the band dispersion of ks levels is reasonable , with valence bands mainly given by o 2@xmath11 states , and ti 3@xmath19 states forming the conduction bands .
the application of standard @xmath38 methods gives gaps of 3.59 and 3.97 ev for rutile and anatase @xcite , respectively .
the value for rutile is again smaller than the one given by the ups estimation @xcite , but still close to the experimental value of 4 ev .
there exist a number of theoretical works , with calculations performed at different levels of dft or including mbpt descriptions , for the electronic gap of rutile and anatase tio@xmath0 @xcite .
therefore in literature it is possible to find for the electronic gap a quite large range of possible values , attributed to the gap of titanium dioxide , which are often erroneously compared with the experimental optical gap .
the dft - gga values calculated @xcite are comparable to the ones obtained with a variety of different dft approaches , with different functionals , by using plane waves or localized basis methods , and all - electron or pseudopotential approaches . only the hybrid pbe0 @xcite and b3lyp @xcite xc - functionals give larger values . from quasi - particle calculations ,
the electronic gap of anatase has been estimated to be 3.79 ev @xcite by @xmath43 .
however the more refined computational approach , because of its inclusion of a self - consistent evaluation of @xmath38 , yields a gap of 3.78 ev for rutile tio@xmath0 .
@xcite on top of gga ( ) , and via the bethe - salpeter equation ( bse ) ( ) . the experimental spectrum ( ) from refs .
@xcite and @xcite is also shown for comparison . ] moving to optical properties , and by applying the rpa method to both ks and qp energies , we obtain spectra ( fig .
[ lfig3 ] ) that do not in overall behaviour agree with experiment .
differences are clear both in absorption edge determination , and in the overall shape of the spectra .
the inclusion of quasi - particle corrections at the @xmath38 level yields a rigid shift of the absorption spectrum , moving the edge at higher energies , due to the opening of the gap .
however , the shape of the absorption is quite unaffected , because the interaction is still treated using an independent quasi - particle approximation .
a substantially better agreement may be obtained @xcite by solving the bse , which takes into account both many body interactions and excitonic effects .
indeed , it produces a good description of absorption spectra and excitons , as shown in fig .
[ lfig3 ] the optical absorption spectra calculated for the two phases , with polarization along the @xmath69-direction of the unit cells , are provided in [ lfig3 ] .
the spectra given by independent - particle transitions present two characteristic features .
first , the band edge is underestimated , due to the electronic gap underestimation in dft .
second , the overall shape of the spectrum is , for both phases , and both orientations , different from the experiment , in the sense that the oscillator strengths are not correct .
the inclusion of the quasi - particle description , which should improve the electronic gap description , does not improve the overall shape of the spectrum .
the absorption edge is , however , shifted at higher energies , even higher than expected from experiments .
the description of optical properties within the two interacting quasi - particle approximation ( by solving the bse ) definitely improves the results .
the absorption edge is now comparable to the experimental one , with the optical gap estimated from our calculations in good agreement with the available data .
further , the shape of the spectrum is now well described , with a redistribution of transitions at lower energies .
the agreement is generally good for both phases @xcite .
the nature of the exciton is still under debate in tio@xmath0 materials @xcite .
the experimental binding energy is of 4 mev and some uncertainty exists for the exact determination of optical edge .
moreover , the exciton is localized in one of the two phases , and delocalized in the other one , at least based on experimental results . however , an explanation for this behaviour is so far missing . refined measurements @xcite give an exciton of 2@xmath65 character at 3.031 ev , and a 1@xmath66 quadrupolar exciton at energy lower than 3 ev . from _
ab initio _ calculations with @xmath69 polarization , two dark ( optically inactive ) or quasi dark excitons are located in rutile at 3.40 and 3.55 ev ( denoted by @xmath70 , @xmath71 in [ lfig4 ] ) . at the same time , the optically active exciton , with 4 mev of binding energy , is located at 3.59 ev ( @xmath72 in [ lfig4 ] ) .
the spatial distribution of the first three excitons is plotted in [ lfig4 ] .
the transitions are from o 2@xmath11 states to ti 3@xmath19 states of the triplet @xmath63 , as expected .
while the first two optically forbidden transitions involve ti atoms farther away from the excited o atom , the optical active transition involves states of the nearest neighbour ti atoms . in this section
we have endeavoured to clarify though the application of a consistent description , the properties of the two main crystalline phases of titania .
particular attention has been taken to how the inclusion of a proper description of exchange and correlation effects can improve the description of both electronic and optical properties of tio@xmath0 .
now that bulk properties are known , from a theoretical point of view , at the level that the state - of - the - art _ ab initio _ techniques allow us to reach , we can attempt to describe how quantum confinement induced by reduced dimensionality , and doping both effect the properties of tio@xmath0 .
our final aim is to demonstrate how we may tune the electronic and optical gap of nanostructures for photocatalytic and photovoltaic applications .
having described in the previous section the electronic structure of the two bulk tio@xmath0 phases , we will now turn our attention to the influence of nanostructure on the energy gap .
this is an area which has received considerable attention in recent years @xcite , in part due to the inherently high surface to volume ratio of nanostructures .
it is hoped that this will allow materials with shorter quasi - particle lifetimes to function effectively for photocatalytic activities , since excitons are essentially formed at the material s surface . however ,
this advantage is partly countered by quantum confinement effects , which tend to increase the energy gap in nanostructures .
these competing factors make the accurate theoretical determination of energy gaps in nanostructured materials a thing of great interest . in ev versus tio@xmath1 structure for 0d ( tio@xmath1)@xmath73 clusters ( @xmath74 ) , 1d tio@xmath1 ( 2,2 ) nanorods , ( 3,3 ) nanotubes , ( 4,4 ) nanotubes , 2d hexabc and anatase layers , and 3d anatase surface , anatase bulk , and rutile bulk phases .
dft calculations using the highest occupied and lowest unoccupied state gaps with the standard gga rpbe xc - functional ( ) and the hybrid b3lyp xc - functional ( ) , are compared with dft @xmath75 calculations ( ) , @xmath43 quasi - particle calculations ( ) , and experimental results ( ) @xcite .
schematics of representative structures for each dimensionality are shown above and taken from ref .
] however , as shown in the previous section , standard dft calculations tend to underestimate electronic band gaps for bulk tio@xmath0 by approximately 2 ev , due in part to self - interaction errors @xcite .
these errors arise from an incomplete cancellation of the electron s coulomb potential in the exchange - correlation ( xc)-functional .
this may be partially addressed by the use of hybrid functionals such as b3lyp @xcite , which generally seem to improve band gaps for bulk systems @xcite
. however , such calculations are computationally more expensive , due to the added dependence of the xc - functional on the electron s wavefunction . moreover ,
b3lyp calculations for tio@xmath0 clusters largely underestimate the gap relative to the more reliable difference between standard dft calculated ionization potential @xmath76 and electron affinity @xmath77 . for isolated systems such as clusters , the needed energetics of charged species are quantitatively described by standard dft .
also , b3lyp and rpbe @xcite calculations provide the same qualitative description of the _ trends _ in the energy gaps for tio@xmath0 , as seen in [ undoped ] .
another methodology is thus needed to describe the band gaps of periodic systems .
@xmath43 is probably the most successful and generally applicable method for calculating quasi - particle gaps . for clusters it agrees well with @xmath75 , and it has been shows to produce reliable results for bulk phases . on the other hand ,
@xmath43 calculations describe an @xmath67 transition where the number of charges is not conserved , rather than the electron hole pair induced by photoabsorption , which is a neutral process .
indeed , a description in terms of electronic gap can not be compared with , or provide direct information on the optical gap , which is the most investigated quantity , due to its critical importance for photocatalytic processes .
[ undoped ] shows that for both 3d and 2d systems , rpbe gaps underestimate the experimental optical gaps by approximately 1 ev . for 1d and 0d systems , there is a much larger difference of about 4 ev and 5 ev respectively , between the rpbe gaps and the @xmath78 and @xmath43 results .
this increasing disparity may be attributed to the greater quantum confinement and charge localization in the 1d and 0d systems , which yield greater self - interaction effects .
the b3lyp gaps also tend to underestimate this effect , simply increasing the rpbe energy gaps for both 0d and 3d systems by about 1.4 ev .
on the other hand , the rpbe gaps reproduce qualitatively the structural dependence of the @xmath75 , @xmath43 , and experimental results for a given dimensionality , up to a constant shift .
this is true even for 3d bulk systems , where standard dft does not correctly predict rutile to be the most stable structure @xcite . to summarize , quantum confinement effects seem to increase the energy gap significantly for both 0d and 1d systems , while 2d and 3d systems may be much less affected .
this suggests that 2d laminar structures are viable candidates for reducing the minimal quasi - particle , while leaving the band gap nearly unchanged .
however , a more accurate description of the photoabsorption properties of these novel nanostructures , perhaps using the methodologies recently applied to bulk tio@xmath0
@xcite , still remains to be found .
the doping of tio@xmath0 nanostructures has received much recent attention as a possible means for effectively tuning tio@xmath0 s band gap into the visible range @xcite .
recent experiments suggest substituting oxygen by boron or nitrogen in the bulk introduces mid - gap states , allowing lower energy excitations
. however , to model such systems effectively requires large supercells , both to properly describe the experimental doping fractions of @xmath79 10% , and to ensure dopant dopant interactions are minimized . per tio@xmath0 functional unit vs. energy @xmath80 in ev for undoped ( ) , boron doped ( ) and nitrogen doped ( ) ( 2,2 ) tio@xmath0 nanorods from standard dft gga rpbe xc - functional and ( inset )
@xmath43 quasi - particle calculations .
the highest occupied states for boron and nitrogen doped ( 2,2 ) tio@xmath81 nanorods are depicted by isosurfaces of @xmath820.05@xmath83@xmath84 in the structure diagrams to the left . ] [ dos ] shows the dft calculated dos and structures for the most stable boron doped and nitrogen doped tio@xmath1 ( 2,2 ) nanorods . the highest occupied state is also shown as isosurfaces of @xmath820.05@xmath83@xmath85 in the side views of the doped structures . as with tio@xmath1 clusters ,
the influence of boron dopants on tio@xmath1 nanorods may be understood in terms of boron s weak electronegativity , especially when compared with the strongly electronegative oxygen .
boron prefers to occupy oxygen sites which are 2-fold coordinated to neighbouring titanium atoms .
however , as with the 0d clusters , boron s relatively electropositive character induces significant structural changes in the 1d structures , creating a stronger third bond to a neighbouring three - fold coordinated oxygen via an oxygen dislocation , as shown in [ dos ] .
this yields three occupied mid - gap states localized on the boron dopant , which overlap both the valence band o 2@xmath86 and conduction band ti 3@xmath87 states , as shown in [ dos ] .
boron dopants thus yield donor states near the conduction band , which may be photocatalytically active in the visible region .
however , the quantum confinement inherent in these 1d structures may stretch these gaps , as found for the @xmath43 calculated dos shown in the inset of [ dos ] .
on the other hand , nitrogen dopants prefer to occupy oxygen sites which are 3-fold coordinated to ti , as was previously found for the rutile tio@xmath1 surface @xcite .
this yields one occupied state at the top of the valence band and one unoccupied mid - gap state in the same spin channel .
both states are localized on the nitrogen dopant but overlap the valence band o 2@xmath86 states , as shown in [ dos ] .
nitrogen dopants thus act as acceptors , providing localized states well above the valence band , as is also found for the @xmath43 calculated dos shown in the inset of [ dos ] .
although nitrogen dopants act as acceptors in tio@xmath1 1d structures , such large gaps between the valence band and the unoccupied mid - gap states would not yield @xmath11-type semiconductors .
this may be attributed to the substantial quantum confinement in these 1d structures .
however , for 2d and 3d systems , it is possible to produce both @xmath11-type and @xmath10-type classical semiconductors , as discussed in ref .
this has recently been shown experimentally in ref .
@xcite , where co - doping of anatase tio@xmath0 with nitrogen and chromium was found to improve the localization of the acceptor states , and reduce the effective optical gap . by replacing both ti@xmath17 and o@xmath18 atoms with dopants in the same tio@xmath21 octahedral ,
it should be possible to `` tune '' the optical band gap to a much finer degree . in ev versus tio@xmath1 structure for 0d ( tio@xmath1)@xmath73 clusters ( @xmath74 ) , 1d tio@xmath1 ( 2,2 ) nanorods , ( 3,3 ) nanotubes , ( 4,4 ) nanotubes , 2d hexabc and anatase layers , and 3d anatase surface and bulk .
the energy gaps for ( a ) boron doped systems from dft calculations using the highest occupied and lowest unoccupied state gaps with the standard gga rpbe xc - functional ( ) are compared with dft @xmath75 calculations ( ) , and @xmath43 quasi - particle calculations ( ) , and ( b ) nitrogen doped systems from dft calculations using the highest occupied and lowest unoccupied state gaps with the standard gga rpbe xc - functional ( ) are compared with dft @xmath75 calculations ( ) , and @xmath43 quasi - particle calculations ( ) , and experimental ( ) results @xcite .
small open symbols denote transitions between highest fully occupied states and the conduction band . ]
whether calculated using rpbe , @xmath78 or @xmath43 , the energy gaps for both boron and nitrogen doped tio@xmath0 nanostructures are generally narrowed , as shown in [ doped](a ) and ( b ) . however , for nitrogen doped ( tio@xmath1)@xmath88 clusters where nitrogen acts as an acceptor ( @xmath89 ) , the energy gap is actually _ increased _ when spin is conserved , compared to the undoped clusters in rpbe .
this effect is not properly described by the @xmath67 transitions of @xmath78 , for which spin is not conserved for these nitrogen doped clusters . on the other hand ,
when nitrogen acts as a donor ( @xmath90 ) the smallest gap between energy levels does conserve spin .
the boron doped tio@xmath91 nanorods and nanotubes have perhaps the most promising energy gap results of the tio@xmath0 structures , as seen in [ doped](a ) .
boron dopants introduce in the nanorods localized occupied states near the conduction band edge in both rpbe ( _ cf . _ [ dos ] ) and @xmath43 ( _ cf .
_ inset of [ dos ] ) calculations . on the other hand ,
nitrogen doping of nanorods introduces well defined mid - gap states , as shown in [ dos ] .
however , to perform water dissociation , the energy of the excited electron must be above that for hydrogen evolution , with respect to the vacuum level .
this is not the case for such a mid - gap state .
this opens the possibility of a second excitation from the mid - gap state to the conduction band .
however , the cross section for such an excitation may be rather low . for boron doping of 2d and 3d structures , the highest occupied state donates its electron almost entirely to the conduction band , yielding an @xmath10-type semiconductor .
thus at very low temperatures , the rpbe band gap is very small .
the same is true for @xmath10-type nitrogen doped bulk anatase . for these reasons the energy gap between the highest
fully occupied state and the conduction band , which may be more relevant for photoabsorption , is also shown .
these rpbe gaps are still generally smaller than those for their undoped tio@xmath0 counterparts , shown in [ undoped ] . in summary , for both boron and nitrogen doped clusters we find
rpbe gaps differ from @xmath75 by about 3 ev , while for nitrogen doped anatase the rpbe gap differs from experiment by about 0.6 ev .
given the common shift of 1 ev for undoped 2d and 3d structures , this suggests that both boron and nitrogen doped 2d tio@xmath1 structures are promising candidates for photocatalysis .
further , the boron and nitrogen doped 1d nanotube results also warrant further experimental investigation .
tio@xmath0 is so far the most widely used solid material in the development of solar cell devices based on hybrid architectures@xcite . in these devices ,
the dye , synthetic or organic , absorbs light , and electrons excited by the phonons are injected into the underlying oxide nanostructure .
the hybrid system must therefore satisfy several requirements : ( 1 ) a proper absorption range for the dye , ( 2 ) a fast charge transfer in the oxide , ( 3 ) a slow back - transfer process , and ( 4 ) an easy collection and conduction of electrons in the oxide . while absorption properties may be easily tuned at the chemical level by changing or adding functional groups , a critical point is to understand , and therefore control , the process at the interface . indeed , simply having a good energy level alignment is not sufficient because the fast electron injection process is dynamical .
since the experimental characterization of complex systems ( networks of nanostructures , with adsorbed dyes , and in solution ) is quite complicated , the theoretical description of such systems can be of fundamental importance in unravelling the processes governing the behaviour of dye - sensitized solar cells ( dssc ) . the most popular technique for studying these dynamical processes
is time dependent dft ( tddft ) . for a detailed discussion of the success and possible limitations of this method
when applied to hybrid systems , we refer the reader to ref .
tddft is a generalization of dft which allows us to directly describe excited states .
for example , it has previously been used successfully to calculate the optical absorption of large organic molecules ( [ n719 ] ) , such as ru - dyes , or indolines . this same technique has also been applied recently to hybrid systems used for photovoltaic applications .
we want here to highlight that the charge injection transfer has also been modelled for molecules adsorbed on tio@xmath0 clusters@xcite , giving estimations as fast as 8 fs for the injection time @xcite .
tddft is therefore a powerful tool to understand dynamical electronic effects in systems as large and complex as hybrid organic - oxide solar cells .
in this chapter we showed how an oxide of predominant importance in nanotechnological and environmental fields , such as photocatalysis and photovoltaics , can be successfully investigates by state - of - the - art _ ab initio _ techniques .
the optical and excitonic properties of the bulk phases can be properly described by mbpt techniques .
we also showed how the electronic properties of tio@xmath0 may be `` tailored '' using nanostructural changes in combination with boron and nitrogen doping . while boron doping tends to produce smaller band gap @xmath10-type semiconductors , nitrogen doping produces @xmath11-type or @xmath10-type semiconductors depending on whether or not nearby oxygen atoms occupy surface sites .
this suggests that a @xmath11-type tio@xmath1 semiconductor may be produced using nitrogen doping in conjunction with surface confinement at the nanoscale .
we also showed that it has been proved how , in the field of photovoltaics , tddft is a powerful tool to understand the mechanism of charge injection at organic - inorganic interfaces .
we acknowledge funding by `` grupos consolidados upv / ehu del gobierno vasco '' ( it-319 - 07 ) , the european community through e - i3 etsf project ( contract number 211956 ) . we acknowledge support by the barcelona supercomputing center , `` red espanola de supercomputacion '' , sgiker arina ( upv / ehu ) and transnational access programme hpc - europe++ .
g - l . acknowledges funding from the spanish mec through the ` juan de la cierva ' program .
l.c . acknowledges funding from upv / ehu through the ` ayudas de especializacin para investigadores doctores ' program .
we also thank h. petek , s. ossicini , m. wanko , m. piacenza , and m. palummo for helpful discussions .
y. gai , j. li , s .- s .
li , j .- b .
xia , and s .- h .
wei , design of narrow - gap tio@xmath92 : a passivated codoping approach for enhanced photoelectrochemical activity , _ phys .
lett . _ * 102 * , 0 03640214 ( 2009 ) .
qu and g .- j .
kroes , theoretical study of the electronic structure and stability of titanium dioxide clusters ( tio@xmath81)@xmath88 with @xmath10 = 19 , _ j. phys
b _ * 110 * , 0 89989007 ( 2006 ) .
zhai and l .- s .
wang , probing the electronic structure and band gap eveolition of titanium oxide clusters ( tio@xmath92)@xmath93 ( @xmath10 = 110 ) using photoelectron spectroscopy , _ j. am .
soc . _ * 129 * , 0 30223026 ( 2007 ) .
d. j. mowbray , j. i. martnez , j. m. garca lastra , k. s. thygesen , and k. w. jacobsen , stability and electronic properties of tio@xmath81 nanostructures with and without b and n doping , _ j. phys
c _ * 113 * , 0 1230112308 ( 2009 ) . s. in , a. orlov , r. berg , f. carca , s. pedrosa - jimenez , m. s. tikhov , d. s. wright , and r. m. lambert , effective visible light - activated b - doped and b , n - codoped tio@xmath81 photocatalysts , _
_ * 129 * , 0 1379013791 ( 2007 ) .
a. ghicov , j. m. macak , h. tsuchiya , j. kunze , v. hublein , s. kleber , and p. schmuki , tio@xmath81 nanotube layers : dose effects during nitrogen doping by ion implantation , _ chem . phys .
lett . _ * 419 * , 0 426429 ( 2006 ) .
y. chen , s. zhang , y. yu , h. wu , s. wang , b. zhu , w. huang , and s. wu , synthesis , characterization , and photocatalytic activity of n - doped tio@xmath81 nanotubes , _
_ * 29 * , 0 245249 ( 2008 ) . w. zhu , x. qiu , v. iancu , x .- q .
chen , h. pan , w. wang , n. m. dimitrijevic , t. rajh , h. m. m. iii , m. p. paranthaman , g. m. stocks , h. h. weitering , b. gu , g. eres , and z. zhang , band gap narrowing of titanium oxide semiconductors by noncompensated anion - cation codoping for enhanced visible - light photoactivity , _ phys .
lett . _ * 103 * , 0 22640114 ( 2009 ) .
m. k. nazeeruddin , a. kay , i. rodicio , r. humphry - baker , e. mller , p. liska , n. vlachopoulos , and m. grtzel , conversion of light to electricity by _
cis_-x@xmath92bis(2,2-bipyridyl-4,4-dicarboxylate)ruthenium(ii ) charge - transfer sensitizers ( x = cl@xmath94 , br@xmath94 , i@xmath94 , cn@xmath94 , and scn@xmath94 ) on nanocrystalline tio@xmath92 electrodes , _ j. am . chem .
_ * 115 * , 0 63826390 ( 1993 ) .
p. wang , s. m. zakeeruddin , j. e. moser , r. humphry - baker , p. comte , v. aranyos , a. hagfeldt , m. k. nazeeruddin , and m. grtzel , stable new sensitizer with improved light harvesting for nanocrystalline dye - sensitized solar cells , _ adv .
* 16 * , 0 18061811 ( 2004 )
. j .- j .
chen , c .- c .
lu , w .- t .
wu , , and j .- s .
chen , performance and electron transport properties of tio@xmath92 nanocomposite dye - sensitized solar cells , _
nanotechnology _ * 19 * , 0 105702105708 ( 2008 ) .
s. h. kang , s .- h .
choi , m .- s .
kang , j .- y .
kim , h .- s .
kim , t. hyeon , and y .- e .
sung , nanorod - based dye - sensitized solar cells with improved charge collection efficiency , _ adv .
mater . _ * 20 * , 0 5458 , ( 2008 ) .
m. adachi , y. murata , j. takao , j. jiu , m. sakamoto , and f. wang , highly efficient dye - sensitized solar cells with a titania thin - film electrode composed of a network structure of single - crystal - like tio@xmath92 nanowires made by the `` oriented attachment '' mechanism , _
soc . _ * 126 * , 0 1494314949 ( 2004 ) .
m. v. ganduglia - pirovano , a. hofmann , and j. sauer , oxygen vacancies in transition metal and rare earth oxides : current state of understanding and remaining challenges , _ surf .
rep . _ * 62 * , 0 219270 , ( 2007 ) .
p. krger , s. bourgeois , b. domenichini , h. magnan , d. chandesris , p. l. fvre , a. m. flank , j. jupille , l. floreano , a. cossaro , a. verdini , and a. morgante , defect states at the tio@xmath92(110 ) surface probed by resonant photoelectron diffraction , _ phys .
_ * 100 * , 0 05550114 ( 2008 ) .
t. minato , y. sainoo , y. kim , h. s. kato , k. ichi aika , m. kawai , j. zhao , h. petek , t. huang , w. he , b. wang , z. wang , y. zhao , j. yang , and j. g. hou , the electronic structure of oxygen atom vacancy and hydroxyl impurity defects on titanium dioxide ( 110 ) surface , _ j. chem . phys . _
* 130 * , 0 124502111 ( 2009 ) .
e. finazzi , c. d. valentin , g. pacchioni , and a. selloni , excess electron states in reduced bulk anatase tio@xmath92 : comparison of standard gga , gga + u , and hybrid dft calculations , _ j. chem .
* 129 * , 0 15411319 ( 2008 ) .
s. wendt , p.
t. sprunger , e. lira , g. k. h. madsen , z. li , j. o. hansen , j. matthiesen , a. blekinge - rasmussen , e. laegsgaard , b. hammer , and f. besenbacher , the role of interstitial sites in the ti3@xmath19 defect state in the band gap of titania , _ science _ * 320 * , 0 17551759 , ( 2008 ) .
l. hedin and s. lundqvist .
`` effects of electron - electron and electron - phonon interactions on the one - electron states of solids , '' in _ solid state physics _ , vol . *
23 * , edited by h. ehrenreich , f. seitz , and d. turnbull ( academic press , new york , 1969 ) pp .
1181 .
j. hardmand , g. n. raikar , c. a. muryn , g. van der laan , p. l. wincott , g. thornton , d. w. bullett , and p. a. d. m. a. dale , valence - band structure of tio@xmath81 along the @xmath42-@xmath62-@xmath96 and @xmath42-@xmath34-@xmath97 directions , _ phys .
b _ * 49 * , 0 71707177 ( 1994 )
. f. m. f. de groot , j. faber , j. j. m. michiels , m. t. czy yk , m. abbate , and j. c. fuggle , oxygen 1@xmath66 x - ray absorption of tetravalent titanium oxides : a comparison with single - particle calculations , _ phys .
b _ * 48 * , 0 20742080 ( 1993 ) .
l. fleming , c. c. fulton , g. lucovsky , j. e. rowe , m. d. ulrich , and j. lning , local bonding analysis of the valence and conduction band features of tio@xmath92 , _ j. appl
. phys . _ * 102 * , 0 03370717 ( 2007 ) .
f. labat , p. baranek , c. doman , c. minot , and c. adamo , density functional theory analysis of the structural and electronic properties of tio@xmath81 rutile and anatase polytypes : performances of different exchange - correlation functionals , _ j. chem . phys . _
* 126 * , 0 154703112 , ( 2007 ) .
l. chiodo , j. m. garca - lastra , a. incomino , h. petek , s. ossicini , and a. rubio . _
ab initio _ electronic and optical description of tio@xmath92 crystalline phases : towards photocatalysis and photovoltaic applications .
( unpublished , 2009 ) .
m. oshikiri , m. boero , j. ye , f. aryasetiawan , and g. kido , the electronic structures of the thin films of invo@xmath100 and tio@xmath81 by first principles calculations , _ thin solid films _ * 445 * , 0 168174 ( 2003 ) .
m. calatayud , p. mori - snchez , a. beltrn , a. martn pends , e. francisco , j. andrs , and j. m. recio , quantum - mechanical analysis of the equation of state of anatase tio@xmath92 , _ phys .
b _ * 64 * , 0 18411319 ( 2001 ) .
h. m. lawler , j. j. rehr , f. vila , s. d. dalosto , e. l. shirley , and z. h. levine , optical to uv spectra and birefringence of sio@xmath92 and tio@xmath92 : first - principles calculations with excitonic effects , _ phys .
b _ * 78 * , 0 20510818 ( 2008 ) .
cai , k. sendt , and j. r. reimers , failure of density - functional theory and time - dependent density - functional theory for large extended @xmath101 systems , _
phys _ * 117 * , 0 55435549 ( 2002 ) .
u. salzner , j. b. lagowski , p. g. pickup , and r. a. poirier , design of low band gap polymers employing density functional theory hybrid functionals ameliorate band gap problem , _
j. comput .
* 18 * , 0 19431953 ( 1997 ) .
f. de angelis , s. fantacci , a. selloni , and m. nazeeruddin , time dependent density functional theory study of the absorption spectrum of the [ ru(4,4-coo@xmath94 - 2,2-bpy)@xmath92(x)@xmath92]@xmath103 ( x = ncs , cl ) dyes in water solution , _ chem .
lett . _ * 415 * , 0 115120 , ( 2005 ) .
m. nazeeruddin , f. de angelis , s. fantacci , a. selloni , g. viscardi , p. liska , s. ito , b. takeru , and m. gratzel , combined experimental and dft - tddft computational study of photoelectrochemical cell ruthenium sensitizers , _ j. am . chem .
soc . _ * 127 * , 0 1683516847 ( 2005 ) .
f. de angelis , s. fantacci , a. selloni , m. k. nazeeruddin , and m. gratzel , time - dependent density functional theory investigations on the excited states of ru(ii)-dye - sensitized tio@xmath92 nanoparticles : the role of sensitizer protonation , _ j. am .
soc . _ * 129 * , 0 1415614157 ( 2007 ) .
f. de angelis , s. fantacci , and a. selloni , alignment of the dye s molecular levels with the tio@xmath92 band edges in dye - sensitized solar cells : a dft - tddft study , _ nanotechnology _ * 19 * , 0 42400217 ( 2008 ) .
w. r. duncan , w. m. stier , and o. v. prezhdo , _ ab initio _
nonadiabatic molecular dynamics of the ultrafast electron injection across the alizarin-tio@xmath92 interface , _
_ * 127 * , 0 79417951 ( 2005 ) .
w. r. duncan and o. v. prezhdo , temperature independence of the photoinduced electron injection in dye - sensitized tio@xmath92 rationalized by _ ab initio _ time - domain density functional theory ,
soc . _ * 130 * , 0 97569762 ( 2008 ) . | titanium dioxide is one of the most widely investigated oxides .
this is due to its broad range of applications , from catalysis to photocatalysis to photovoltaics . despite this large interest , many of its bulk properties
have been sparsely investigated using either experimental techniques or _ ab initio _ theory .
further , some of tio@xmath0 s most important properties , such as its electronic band gap , the localized character of excitons , and the localized nature of states induced by oxygen vacancies , are still under debate .
we present a unified description of the properties of rutile and anatase phases , obtained from _ ab initio _ state of the art methods , ranging from density functional theory ( dft ) to many body perturbation theory ( mbpt ) derived techniques . in so doing ,
we show how advanced computational techniques can be used to quantitatively describe the structural , electronic , and optical properties of tio@xmath0 nanostructures , an area of fundamental importance in applied research .
indeed , we address one of the main challenges to tio@xmath0-photocatalysis , namely band gap narrowing , by showing how to combine nanostructural changes with doping . with this aim
we compare tio@xmath0 s electronic properties for 0d clusters , 1d nanorods , 2d layers , and 3d bulks using different approximations within dft and mbpt calculations .
while quantum confinement effects lead to a widening of the energy gap , it has been shown that substitutional doping with boron or nitrogen gives rise to ( meta-)stable structures and the introduction of dopant and mid - gap states which effectively reduce the band gap .
finally , we report how _ ab initio _ methods can be applied to understand the important role of tio@xmath0 as electron - acceptor in dye - sensitized solar cells .
this task is made more difficult by the hybrid organic - oxide structure of the involved systems . |
over the last decade much progress has been realized towards understanding the qcd dynamics at high energies .
the successful operation of the desy @xmath1 collider hera has opened a new era of experimental and theoretical investigation into the deep structure of the proton and , in general , of hadronic interactions .
some of the most important observations are the striking rise of the proton structure function @xmath2 for small values of the bjorken variable @xmath3 ( @xmath4 ) , the large contribution of diffractive processes in this kinematical range and the geometric scaling ( for a recent review , see e.g. ref .
theoretically , at small @xmath3 , due to the large gluon density , we expect the transition of the regime described by the linear dynamics , where only the parton emissions are considered , for a new regime where the physical process of recombination of partons become important in the parton cascade and the evolution is given by a nonlinear evolution equation .
this regime is characterized by the limitation on the maximum phase - space parton density that can be reached in the hadron wavefunction ( parton saturation ) , with the transition being specified by a typical scale , which is energy dependent and is called saturation scale @xmath5 ( for recent reviews see ref .
although the hera experimental results have a natural interpretation in terms of the saturation physics , due to the kinematical limitations of the experiment none of these phenomena can be taken as a conclusive evidence for a new regime of the qcd dynamics .
currently , the hera ii run has obtained much more precise experimental data , but they are limited at center of mass energies smaller than 300 gev , where deviations between linear and saturation predictions are small .
consequently , our understanding of the correct qcd dynamics at high energies is still an open question . as hera will stop its operations , in principle , in 2007 , and
the qcd dynamics must be known as precisely as possible in order to maximize the discovery potential for new physics at the next generation of colliders , the study of alternatives which could constrain the qcd dynamics is timely and necessary . in this letter
we study the photoproduction of heavy quarks and vector mesons in the coherent proton - nucleus ( @xmath0 ) interactions at rhic and lhc energies and analyze if these processes can be used to determine the qcd dynamics at high energies ( for similar studies on @xmath6 and @xmath7 collisions , see refs .
the main advantage of using colliding hadron and nuclear beams for studying photon induced interactions is the high equivalent photon energies and luminosities that can be achieved at existing and future accelerators ( for a recent discussion see ref .
lets consider the proton - nucleus interaction at large impact parameter ( @xmath8 ) and at ultrarelativistic energies . in this regime
we expect the electromagnetic interaction to be dominant .
when we consider the electromagnetic field associated to the proton and ion , we have that due to the coherent action of all protons in the nucleus , the electromagnetic field surrounding the ion is very larger than the proton one .
this result can be easily understood if we use the weiszcker - williams method to calculate the equivalent flux of photons from a charge @xmath9 nucleus a distance @xmath10 away , which is given by ( for recent reviews see ref .
@xcite ) @xmath11 \ , \label{fluxunint}\end{aligned}\ ] ] where @xmath12 is the photon energy , @xmath13 is the lorentz boost of a single beam and @xmath14 ; @xmath15 and @xmath16 are the modified bessel functions . from the above expression we have that photon spectrum of a nucleus with charge
@xmath9 is proportional to @xmath17 . due to asymmetry in the collision , with the ion being likely the photon emitter , we have that the photon direction is known , which will implicate an asymmetry in the rapidity distribution ( see below ) .
the coherence condition limits the photon virtuality to very low values @xmath18 , which implies that for most purposes , these can be considered as real .
therefore , coherent @xmath0 collisions can be used in order to study photon - proton interactions . the requirement that photoproduction is not accompanied by hadronic interaction ( ultraperipheral collision )
can be done by restricting the impact parameter @xmath10 to be larger than the sum of the proton and the nuclear radius .
therefore , the total photon flux interacting with the target nucleus is given by eq .
( [ fluxunint ] ) integrated over the transverse area of the target for all impact parameters subject to the constraint that the proton and the nucleus do not interact hadronically .
an analytic approximation for @xmath0 collisions can be obtained using as integration limit @xmath19 , producing @xcite @xmath20\ , \label{fluxint}\end{aligned}\ ] ] where @xmath21 and @xmath22 .
the cross section for the photoproduction of a final state @xmath23 in a coherent @xmath0 collisions will be given by , @xmath24 where @xmath25 , @xmath26 and @xmath27 is the c.m.s energy of the proton - nucleus system .
considering @xmath28 collisions at lhc , the lorentz factor is @xmath29 , giving the maximum c.m.s . @xmath30
energy @xmath31 gev .
therefore , while studies of photoproduction at hera are limited to photon - proton center of mass energies of about 200 gev , photon - proton interactions at lhc can reach one order of magnitude higher on energy .
consequently , studies of @xmath32 interactions at lhc could provide valuable information on the qcd dynamics at high energies . in this work
we consider that the produced state @xmath23 represents a @xmath33 pair or a vector meson @xmath34 .
since photon emission is coherent over the entire nucleus and the photon is colorless we expect that the events to be characterized by one ( heavy quark production , with @xmath35 being the remaining of the proton ) or two ( vector meson production , with @xmath36 ) rapidity gaps .
the photon - hadron interaction at high energy ( small @xmath3 ) is usually described in the infinite momentum frame of the hadron in terms of the scattering of the photon off a sea quark , which is typically emitted by the small-@xmath3 gluons in the proton . however , in order to disentangle the small-@xmath3 dynamics of the hadron wavefunction , it is more adequate to consider the photon - hadron scattering in the dipole frame , in which most of the energy is carried by the hadron , while the photon has just enough energy to dissociate into a quark - antiquark pair before the scattering . in this representation
the probing projectile fluctuates into a quark - antiquark pair ( a dipole ) with transverse separation @xmath37 long after the interaction , which then scatters off the proton @xcite .
the main motivation to use this color dipole approach , is that it gives a simple unified picture of inclusive and diffractive processes . in particular , in this approach the heavy quark photoproduction cross section reads as , @xmath38 where @xmath39 is the light - cone wavefunction of the photon @xcite .
the quark and antiquark helicities are labeled by @xmath40 and @xmath41 .
the variable @xmath37 defines the relative transverse separation of the pair ( dipole ) and @xmath42 @xmath43 is the longitudinal momentum fractions of the quark ( antiquark ) .
the basic blocks are the photon wavefunction , @xmath44 and the dipole - target cross section , @xmath45 .
for photoproduction we have that longitudinal piece does not contribute , since @xmath46 , and the total cross section is computed introducing the appropriated mass and charge of the charm or bottom quark . similarly ,
in the dipole picture the imaginary part of the amplitude for vector meson production at zero momentum transfer reads as ( see e.g. refs .
@xcite ) @xmath47 where @xmath48 is the light - cone wavefunction of the vector meson .
the total cross section for vector meson photoproduction is given by @xmath49 ^ 2}{16\pi\,b_v}\,(1+\beta^{2 } ) \ ; \label{totalcs}\end{aligned}\ ] ] where @xmath50 is the ratio of real to imaginary part of the amplitude and @xmath51 labels the slope parameter .
the values considered for the slope parameter are taken from the parameterization used in ref .
@xcite . &
@xmath52 & 17 mb ( @xmath53 ) & 5 mb ( @xmath54 ) + & @xmath55 & 155 @xmath56b ( @xmath57 ) & 81 @xmath56b ( @xmath58 ) + & @xmath59 & & 14 mb ( @xmath60 ) + & @xmath61 & & 95 @xmath56b ( @xmath62 ) + we have that the total cross sections for vector meson and heavy quark production in dipole approach are strongly dependent on the dipole - hadron cross section @xmath45 , which contains all information about the target and the strong interaction physics . in the color glass condensate ( cgc )
formalism @xcite , @xmath45 can be computed in the eikonal approximation , resulting @xmath63\,\,,\end{aligned}\ ] ] where @xmath64 is the @xmath64-matrix element which encodes all the information about the hadronic scattering , and thus about the non - linear and quantum effects in the hadron wave function .
the function @xmath64 can be obtained by solving an appropriate evolution equation in the rapidity @xmath65 .
the main properties of @xmath64 are : ( a ) for the interaction of a small dipole ( @xmath66 ) , @xmath67 , which characterizes that this system is weakly interacting ; ( b ) for a large dipole ( @xmath68 ) , the system is strongly absorbed which implies @xmath69 .
this property is associate to the large density of saturated gluons in the hadron wave function . in our analysis
we will consider the phenomenological saturation model proposed in ref .
@xcite which encodes the main properties of the saturation approaches , with the dipole cross section parameterized as follows @xmath70\ , , & \mbox{for $ \bar{\tau } > 2$}\ , , \end{array } \right .
\label{cgcfit}\end{aligned}\ ] ] where @xmath71 and the expression for @xmath72 ( saturation region ) has the correct functional form , as obtained from the theory of the color glass condensate ( cgc ) @xcite . for the color transparency region near saturation border ( @xmath73 ) , the behavior is driven by the effective anomalous dimension @xmath74 , where @xmath75 is the lo bfkl anomalous dimension at saturation limit .
hereafter , we label this model by cgc
. reactions for lhc energy ( see text ) . ]
the distribution on rapidity @xmath76 of the produced final state can be directly computed from eq .
( [ sigaa ] ) , by using its relation with the photon energy @xmath12 , i.e. @xmath77 .
explicitly , the rapidity distribution is written down as , @xmath78}{dy } = \omega \frac{dn_{\gamma } ( \omega ) } { d\omega } \,\sigma_{\gamma p \rightarrow x y}\left(\omega \right ) .
\label{dsigdy}\end{aligned}\ ] ] consequently , given the photon flux , the rapidity distribution is thus a direct measure of the photoproduction cross section for a given energy . in fig .
[ fig:1 ] we present our results for the heavy quark photoproduction at lhc energies . for comparison
, we also present the predictions from the linear dynamics , denoted collinear hereafter , which is calculated assuming that the collinear factorization is valid and that the gluon distribution can be described by the grv98 parametrization ( for more details see ref .
@xcite ) . in tab .
[ tabhq ] one presents the correspondent integrated cross sections ( event rates ) , using a luminosity of @xmath79 @xmath80s@xmath81 .
we have verified for completeness that for rhic energy the difference between the predictions for the rapidity distribution is small , which is expect due to small value of the photon - proton center of mass energy . however , at lhc we can observe a large difference between the predictions and having high rates .
for instance , for charm quark cgc gives a cross section a factor 3 lower than collinear and a factor 2 for bottom .
this deviation holds even in case of experimental cuts on rapidity .
therefore , photoproduction of heavy quarks should provide a feasible and clear measurement of the underlying qcd dynamics at high energies .
since rhic is obtaining data for @xmath82 interactions and lhc is to be commissioned , these processes could be analyzed in the next years .
the advantages are a clear final state ( rapidity gap and low momenta particles ) and no competing effect of dense nuclear environment if compared with hadroproduction
. reactions at lhc energy ( see text ) . ]
the present results can be directly compared with those obtained in ref .
@xcite , which use only collinear approach .
the orders of magnitude are similar for rhic and lhc , with main deviations coming from different choices for the gluon pdfs and distinct factorization scale .
our estimates considering cgc ( collinear ) for rhic are 142 ( 110 ) @xmath56b for charm and 0.15 ( 0.10 ) @xmath56b for bottom .
the rapidity distributions can be also contrasted with estimations in refs .
@xcite , where the heavy quark photoproduction in @xmath7 and @xmath6 collisions have been computed . to see what is difference in the order of magnitude among them
let us perform a few parametric estimates .
roughly , the photon flux on nuclei is approximately @xmath17 the flux on proton , @xmath83 .
moreover , for heavy quarks have been not verified large nuclear shadowing @xcite such that @xmath84 .
the ratio between @xmath0 production and @xmath6 ( or @xmath7 ) can be estimated as @xmath85 .
therefore , using eq .
( [ dsigdy ] ) , one obtains @xmath86 and the enhancement reaches a factor @xmath87 for heavy nuclei in the comparison between photoproduction on @xmath0 and in energetic protons . on the other hand , one obtains @xmath88 and then @xmath0 is suppressed in relation to @xmath7 by a factor @xmath89
however , the larger @xmath0 luminosity , which is two order of magnitude higher than for @xmath7 , counteracts this suppression for the event rates . concerning coherent meson production , in fig .
[ fig:2 ] one presents predictions for the rapidity distribution on @xmath0 collisions at lhc energy considering both light ( @xmath90 ) and heavy ( @xmath91 ) ions .
these theoretical predictions are original in the literature .
the corresponding integrated cross sections ( event rates ) are shown in tab .
[ tabhq ] , having high rates . for completeness
, we have computed the estimation for typical light meson ( @xmath92 ) and a heavy one ( @xmath93 ) . for rhic ,
the cross section are 13 mb ( 0.8 @xmath56b ) for @xmath92 ( @xmath93 ) production .
these results can be contrasted with those obtained in the refs .
we can do a similar estimative of the ratio as for heavy quarks considering that for light mesons @xmath94 and for heavy mesons @xmath95 @xcite .
therefore , an enhancement of @xmath17 for @xmath96 remains but now @xmath97 , respectively . for pb ions at lhc
this is a factor 35 for @xmath92 and a factor @xmath98 for @xmath93 .
it should be noticed that the coherent production of mesons is currently measured at rhic for the @xmath7 case .
therefore , the present estimation could be tested in @xmath99 collisions with a good experimental feasibility . for comparison ,
we also present in fig .
[ fig:3 ] the prediction from the color transparency limit of the cgc model , which is obtained assuming that @xmath100 is given in all kinematic range by its value for @xmath101 .
it is equivalent to disregard the saturation physics , which allows to estimate the importance of the cgc physics in the process .
we have that the predictions for @xmath93 production are somewhat similar , which is expected , since the heavy vector meson production is dominated by small pair separations , where the saturation physics does not contribute significantly .
on the other hand , the photoproduction of @xmath92 mesons is dominated by physics below saturation scale .
the difference is huge , reaching 3 orders of magnitude in rapidity distribution , which demonstrate the importance of the saturation physics on this process .
finally , lets comment on the uncertainties and model dependence for the results presented here . for the heavy quarks ( hq ) production , the main theoretical uncertainties are the higher order corrections in perturbative expansion and the quark masses . within the perturbative qcd collinear formalism , the higher order corrections for hq photoproduction are quite small ( a few percents ) in contrast to the hadroproduction case and they can be taken into account by a suitable choice for the quark mass @xcite . the color dipole formalism , which we use in our analysis , contains contributions of higher orders by definition as it is equivalent to the semihard factorization ( or @xmath102-factorization ) approach in leading logarithm approximation . the latter is well known to include higher order perturbative qcd diagrams in its formulation .
concerning the dglap results ( collinear formalism ) , the uncertainties for different choices of the gluon distribution are small as it is probed at a relatively large scale @xmath103 ( approximately 9 gev@xmath104 for charm and 81 gev@xmath104 for bottom ) .
those uncertainties are known to take place at low virtualities @xmath105 1 - 3 gev@xmath104 . moreover ,
our results are an extrapolation of the photoproduction cross section for energies beyond energies of current accelerator regime using the physical parameters which describe correctly both the low energy ( fixed target ) and the hera data points .
therefore , the present estimates should be reliable . concerning vector meson production , in general
it is described by quite distinct approaches for light and heavy mesons . in the light meson case ,
the old vector meson dominance ( vmd ) model is often adopted @xcite , whereas for the heavy mesons a leading logarithmic approximation of the collinear approach is considered @xcite . in this case , the hard qcd scale is given by the vector meson mass .
we have used the unique theoretical formalism available describing simultaneously light and heavy vector meson production @xcite . once again , the parameters of the model are fixed in order to describe the experimental low energy and desy - hera datasets . the transition between light and heavy mesons is dynamically introduced by parton saturation effects ( via saturation scale ) in the proton target .
the qcd dynamics at high energies is of utmost importance for building a realistic description of @xmath106 collisions at lhc . in this limit qcd evolution leads to a system with high gluon density .
we have shown if such system there exists at high energies it can be proven in coherent @xmath0 collisions at lhc .
we propose two specific final states ( heavy quarks and mesons ) where the experimental identification could be feasible .
it should be noticed , however , that the predictions presented here were based on specific models / assumptions for the process and on recent phenomenological investigations .
therefore , these assumptions should be tested in future @xmath0 experimental investigations .
e. iancu and r. venugopalan , arxiv : hep - ph/0303204 ; v. p. goncalves and m. v. t. machado , mod .
* 19 * , 2525 ( 2004 ) ; h. weigert , prog .
nucl . phys .
* 55 * , 461 ( 2005 ) ; j. jalilian - marian and y. v. kovchegov , prog .
phys . * 56 * , 104 ( 2006 ) . v. p. goncalves and m. v. t. machado , eur .
j. c * 28 * , 71 ( 2003 ) ; eur .
j. c * 29 * , 37 ( 2003 ) ; eur . phys .
j. c * 31 * , 371 ( 2003 ) ; eur . phys .
j. c * 40 * , 519 ( 2005 ) ; v. p. goncalves , m. v. t. machado and w. k. sauter , eur . phys .
j. c ( _ in press _ ) , arxiv : hep - ph/0509027 .
j. jalilian - marian , a. kovner , l. mclerran and h. weigert , phys .
d * 55 * , 5414 ( 1997 ) ; j. jalilian - marian , a. kovner and h. weigert , phys .
d * 59 * , 014014 ( 1999 ) , _ ibid . _ * 59 * , 014015 ( 1999 ) , _ ibid .
_ * 59 * 034007 ( 1999 ) ; a. kovner , j. guilherme milhano and h. weigert , phys . rev .
d * 62 * , 114005 ( 2000 ) ; h. weigert , nucl . phys . * a703 * , 823 ( 2002 ) . | in this letter we study the photoproduction of heavy quarks and vector mesons in the coherent proton - nucleus ( @xmath0 ) interactions for rhic and lhc energies and analyze if these processes can be used to determine the qcd dynamics at high energies .
the integrated cross section and rapidity distribution are estimated using the color glass condensate ( cgc ) formalism . a comparison with the linear dynamics predictions is also presented .
our results indicate that the nonlinear dynamics can be proven in those reactions , which are well suited for studing saturation effects . |
in robotics fields , caging is a type of grasping where robots capture an object by surrounding or hooking it . thus the caging problem based on the shape of the robots and the object
is addressed on geometrical representation .
its mathematical description has been partially studied with focusing on methodology how to cage an object by certain robots .
though it is rigorous , it is not for arbitrary target objects and robots . in this article , we propose an essential of caging to describe arbitrary target objects and robots from a mathematical viewpoint , and then it naturally leads us to a degree of difficulty of escaping and caging .
it is a novel concept of the caging which is connected with practical approaches .
caging or holding an object has been discussed in mathematics field such as @xcite , and has been applied to robotic manipulation in parallel .
rimon and blake raised a caging by two circular robots driven by one parameter in two dimensional planar space @xcite , and formulated its conditions .
wang and kumar proposed caging by multi - robot cooperation with mathematical abstract formulas @xcite .
more than three dimensional caging problem is formulated by @xcite , although only circular and spherical robots are referred . there studies discuss existence of object s free movable space closed by the robots
. hence path connectivity of the free space for the captured object @xcite is an important matter to investigate whether the object is caged or able to escape . as mentioned above , confinement of caging formation is studied by previous works such as @xcite , particularly for problems of two dimensional caging .
additionally although path connectivity can be examined by using probabilistic search algorithm @xcite , the difficulty of caging constraint is not quantitatively qualified .
therefore this paper aims to describe caging problem in robotics in the following viewpoint : to apply the formulation to arbitrary robots and objects ; to reveal difficulty of caging constraint mathematically . from a mathematical viewpoint ,
we recall the fundamental fact that a compact connected @xmath0-dimensional topological manifold @xmath1 in @xmath0-dimensional euclidean space @xmath2 could be wrapped by the @xmath3-dimensional sphere @xmath4 , which corresponds to holding in the real world . this fact is described well by the homotopy group @xmath5 .
it is based on the mathematical fact that @xmath4 in @xmath2 divides @xmath2 into its inner side and outer side .
the subspace @xmath1 has the configuration space @xmath6 $ ] by the euclidean moves @xmath7 , i.e. , @xmath6=\{\psi(m)\}$ ] .
let us fix the sphere @xmath4 and then it divides @xmath6 $ ] to the part @xmath6_{in}$ ] of the inner side of @xmath4 , that of the outer side , @xmath6_{out}$ ] , and the intersection part @xmath6_{int}$ ] , i.e. , @xmath6=[m]_{in}\coprod [ m]_{out}\coprod [ m]_{int}$ ] and @xmath6_{int}:= \{m ' \in [ m]\ | \
m'\cap s^{n-1}\neq \emptyset\}$ ] .
if @xmath6_{in}$ ] is not empty , it means that @xmath4 holds @xmath8_{in}$ ] . on the other hand ,
caging is to divide the configurations space @xmath6 $ ] using an @xmath9 dimensional topological manifold @xmath10 .
then there arises a problem how @xmath10 with 2-codimension can divide the configuration space @xmath6 $ ] mathematically .
this is a fundamental problem on caging .
there are many mathematical studies on this problem to cage some proper geometrical objects using special @xmath10 as mentioned above .
further fruchard and zamfirescu @xcite geometrically studied the geometrical conditions whether a circle holds a convex object @xmath6 $ ] .
rodriguez , mason and ferry investigated geometrical and topological properties the multifinger cages rigorously @xcite .
however there was no investigation on the above fundamental problem for general @xmath1 and @xmath10 .
though the path connectivity of an object @xmath1 in @xmath11 studied well in @xcite is an essential property of caging , it is not sufficient .
for example if we regard a wire puzzle constituting two pieces as a pair of a robot @xmath10 and a target object @xmath1 , @xmath1 is not caged by @xmath10 because there is a path between @xmath1 in of @xmath10 and that of the outer side of @xmath10 . however a wire puzzle should be considered as an example of caging and holding from a practical viewpoint .
it is related to the probabilistic treatment of the path connectivity @xcite .
however the path space is very complicate in general @xcite and it is very difficult to assign a probabilistic measure in the path space . in order to introduce a degree of escaping mathematically , in this article
, we describe the fundamental problem more mathematically , and then we show that caging is represented in terms of recursion of the simple euclidean moves , i.e. , the piecewise euclidean move defined in definitions [ def : pwem ] and @xcite .
the move means that caging is classified countably and naturally leads us to the degree of difficulty of escaping of @xmath6 $ ] from @xmath10 .
it is closely related to a combinatorial explosion and the wire puzzle .
it means that there might be a difference between practical caging and complete caging . when we capture a complicate object by caging , we propose that the difficulty should be proactively considered from a practical viewpoint .
further if we treat the euclidean moves probabilistically , we could assign a natural measure on the moves .
let us consider the @xmath0-dimensional real euclidean space @xmath2 and the @xmath0-dimensional euclidean group @xmath12 .
the element @xmath13 of @xmath14 acts on each point @xmath15 of @xmath2 by @xmath16 such that for @xmath17 , @xmath18 and for the identity element @xmath19 , @xmath20 @xcite .
the action @xmath21 of @xmath22 on @xmath2 has the matrix representation : for @xmath13 , there are element @xmath23 and @xmath24 such that for @xmath25 , @xmath26 in this article , let us refer a @xmath27-dimensional topological manifold in @xmath2 , _
@xmath27-subspace_. we say that the @xmath27-subspaces @xmath1 and @xmath28 of @xmath2 are _ congruent _ if there is an element @xmath13 of @xmath14 such that @xmath29 .
we denote it by @xmath30 .
let @xmath6 $ ] be the configuration space of @xmath1 ; @xmath6:=\{\phi_g(m)\ | \
g \in \se(n)\}$ ] .
the quotient space of the set of the @xmath0-subspaces divided by @xmath14 is denoted by @xmath31 , which classifies the shape of subspace in @xmath2 .
we call @xmath31 _ moduli of the shapes _ ; @xmath6 \in \fm$ ] .
we consider a continuous map @xmath32 \to \se(n)$ ] , i.e. , @xmath33,\se(n))$ ] with a fixing point @xmath34 , where @xmath35 means the set of @xmath36-valued continuous functions over @xmath37 . for given @xmath33,\se(n))$ ] ,
the action @xmath38 on @xmath2 parameterized by @xmath39 $ ] is called _ orbit by @xmath7_. for an @xmath0-subspace @xmath40 , the action @xmath41 induces a congruent family @xmath42}$ ] for the lie group @xmath14 , there is its lie algebra @xmath43 whose element @xmath44 satisfies @xmath45 ; for the economy of notations , we use the same notation @xmath44 for its matrix representation .
[ lm : liealg ] for @xmath46 where @xmath47 and @xmath48 , @xmath49 where @xmath50 , @xmath51 where @xmath52 is the @xmath53 unit matrix . if @xmath54 is regular , @xmath55 .
the straightforward computation leads @xmath56 and thus we have the result . here
we recall the properties of @xmath57 @xcite : [ lm : liealg_so(n ) ] 1 .
@xmath58 , 2 .
the matrix representation of an element @xmath54 of @xmath57 is given by @xmath59 such that @xmath60 , i.e. , @xmath61 ( @xmath62 ) , 3 .
the maximal rank of the matrix representation of @xmath57 is @xmath3 if @xmath63 and @xmath0 otherwise , and 4 . for the matrix representation @xmath47
, we have the natural decomposition , @xmath64 by considering @xmath65 , i.e. the cokernel @xmath66 agrees with the kernel of @xmath54 .
1 , 2 , and 3 are obvious as in @xcite . using the euclidean inner product ,
let us show @xmath67 which is equivalent with 4 .
let us consider an element @xmath15 of @xmath68 , which means @xmath69 for every @xmath70 , and equivalently @xmath71 vanishes .
the element @xmath15 must belongs to @xmath72 or @xmath73 because of @xmath74 .
since for @xmath75 and @xmath39 $ ] , @xmath76 belongs to @xmath43 , by using this relation between @xmath14 and @xmath43 , we consider a euclidean move @xmath77 , which we call _ the simple euclidean move_. for @xmath78 $ ] and @xmath46 @xmath79 where @xmath80 satisfying @xmath81 if @xmath54 is regular , @xmath82 .
@xmath83 @xmath84 . by replacing @xmath85 with @xmath86 , @xmath87 .
though the action of the euclidean move should be regarded as a rotation in the projective space @xmath88 via sl(@xmath89 ) in @xmath90 , the simple euclidean move is given by the following lemma : [ lmm : rt ] for @xmath78 $ ] and @xmath91 , the orbit of @xmath25 by the simple euclidean move @xmath92 , i.e. , @xmath93 is reduced to @xmath94 where @xmath95 for @xmath96 and @xmath97 .
it means the following : 1 .
if @xmath98 vanishes , @xmath99 means a rotation by @xmath100 , at a center @xmath101 by identifying @xmath102 with @xmath2 as a set .
2 . if @xmath54 vanishes , @xmath99 is the translation @xmath103 3 . if both @xmath54 and @xmath98 do not vanish , it is a mixed move of the rotation and the translation .
since this action satisfies @xmath104 and we have the decomposition @xmath105 where @xmath98 is the kernel of @xmath54 from lemma [ lm : liealg_so(n ) ] , we have ( [ eq : se_se ] ) by by substituting @xmath106 into @xmath107 .
the map @xmath108 from @xmath43 to @xmath14 is surjective
. though it can be directly proved from the lemma [ lm : liealg , lm : liealg_so(n ) ] , it is also obtained from ( [ eq : se_se ] ) by setting @xmath109 .
then its surjectivity is can be directly proved by the triangulation of the matrix @xmath54 though it is a little bit complicate due to the maximal tori in @xmath110 .
[ cor : ert ] for every pair of congruent @xmath0-subspaces @xmath1 and @xmath28 , there is a simple euclidean move @xmath77 of @xmath111 such that @xmath112 .
[ def : pwem ] the euclidean move @xmath113 which is given by a collection of the simple euclidean moves @xmath114\}_{i=1,\ldots , \ell}$ ] , i.e. , @xmath115 for @xmath116 , $ ] where @xmath117 , we call @xmath99 the piecewise euclidean move .
let us consider the subspace @xmath118 whose codimension is two ; @xmath10 is @xmath9-subspace in @xmath2 .
@xmath10 might be decomposed to @xmath119 connected parts , @xmath120 hereafter @xmath10 is fixed .
the caging is to restrict some @xmath0-subspace in @xmath2 using the other subspace @xmath10 whose codimension is two , which is modeled after figures , wires in three dimensional case .
the subspace @xmath10 is modeled by them .
thus from a practical viewpoint , we do not consider wild geometries such as the hilbert curve but we neither exclude them in this article .
for such a wild object , some of the following results might be trivial . for given @xmath10 ,
we let @xmath6_{\ck^c}$ ] be the subset of the configuration space @xmath6 $ ] whose element is disjoint to @xmath10 , i.e. , @xmath6_{\ck^c}:=\{m ' \in [ m]\ | \ m ' \subset \ck^c\}$ ] , and @xmath121 be the family of @xmath6_{\ck^c}$ ] .
let @xmath1 and @xmath28 be congruent @xmath0-subspaces in @xmath122 i.e. , @xmath1 and @xmath28 belongs to @xmath6_{\ck^c } \in \fm_{\ck^c}$ ] , and there exists @xmath123 satisfying @xmath29 .
if there is @xmath33 , \se(n))$ ] such that its congruent family @xmath42}$ ] satisfies the conditions , we say that @xmath1 and @xmath28 are @xmath122-congruent , denoted by @xmath128 , and @xmath38 is a @xmath122-congruent homotopy for @xmath1 and @xmath28 .
if we can not find @xmath129,\se(n))$ ] satisfying the above conditions , we say that @xmath1 and @xmath28 are not @xmath122-congruent , denoted by @xmath130 .
[ def : caging ] if every element @xmath1 of @xmath6_{\ck^c}$ ] is @xmath122-congruent each other , we say that @xmath6_{\ck^c}$ ] is a @xmath122-congruent set and we can not cage @xmath6_{\ck^c}\in \fm_{\ck^c}$ ] for @xmath10 .
if we find a pair @xmath1 and @xmath28 of @xmath6_{\ck^c}$ ] which are not @xmath122-congruent , we call @xmath6_{\ck^c}$ ] a complete @xmath122-caging set , and we say that we can completely cage @xmath6_{\ck^c}\in \fm_{\ck^c}$ ] . if @xmath10 of @xmath134 case is a space - filling curve like the hilbert curve such that it is dense in @xmath2 , @xmath121 itself is the empty set though we are not concerned with such a case .
however it should be noted that if @xmath121 is not empty , @xmath6_{\ck^c } \in \fm_{\ck^c}$ ] has a non - trivial geometrical structure generally .
in fact , fruchard and zamfirescu considered the similar problem in which the @xmath10 is a circle and @xmath6_{\ck^c } ( \subset [ m])$ ] is of a convex object @xcite , though in general @xmath6 $ ] is not convex from a practical point of view .
if @xmath143 is reduced to @xmath27-@xmath14 action but can not be reduced to @xmath144-@xmath14 action , we say that @xmath143 is @xmath27-th type and @xmath1 and @xmath28 are @xmath27-th @xmath122-congruent if @xmath21 is @xmath27-th type .
further if @xmath1 and @xmath28 are congruent but not @xmath122-congruent , or can not be @xmath27-th @xmath122-congruent of finite @xmath27 , we say that @xmath1 and @xmath28 are @xmath145-th @xmath122-congruent . for given an @xmath0-subspace @xmath147 , the configuration space @xmath6_{\ck^c}$ ] of @xmath121
is decomposed to @xmath148_{\ck^c } = \coprod_{\ell=1}^\infty [ m]_{\ck^c}^{(\ell)}\ ] ] where @xmath6_{\ck^c}^{(\ell)}$ ] is the set of subspaces of the @xmath27-th @xmath122-congruence to @xmath1 .
an example is illustrated in figure 1 of @xmath149 case , in which dots mean @xmath10 and a @xmath37-shaped object corresponds to @xmath1 .
they are @xmath122-congruent to the outer one but it is obvious that the @xmath37-shaped object ( a ) in figure 1 is not the first @xmath122-congruent to ( d ) . 1.7 cm and a @xmath37-shaped object corresponds to @xmath1 . (
a ) , ( b ) , ( c ) , and ( d ) show the piecewise euclidean moves of the @xmath37-shaped object .
, title="fig:",width=264 ] ( a ) 1.3 cm ( b ) 1.3 cm ( c ) 1.3 cm ( d ) the decomposition to @xmath132 and @xmath133 is related to the link problem and the knot theory @xcite .
thus @xmath122-congruence is a profound problem .
some of kinds of caging are surely based on the braid group in the knot theory when the fundamental group @xmath152 of @xmath6 $ ] is not trivial .
however from the practical viewpoint as mentioned in introduction , we are concerned with the @xmath27-th @xmath122-congruence rather than the @xmath122-congruence itself .
theorem [ thm : fst ] means that for a @xmath122-congruent configuration @xmath6_{\ck^c}$ ] , there might be @xmath153_{\ck^c}$ ] so that we can not move @xmath1 to @xmath154_{\ck^c}$ ] by the simple euclidean move .
it reminds us of parking in a small garage and the wire puzzle ; it is also closely related to the probabilistic treatment of connected path in @xmath6_{\ck^c}$ ] @xcite . in the some situations ,
the @xmath156-th simple euclidean move in a piecewise euclidean move is restricted , and thus , countable in a certain sense , but @xmath27 is not small .
to find the piecewise euclidean moves connecting @xmath1 and @xmath28 is basically difficult because of the combinatorial explosion .
it is the origin of the difficulty of the wire puzzle .
in other words , caging problem is connected with the difficulty of combinatorial problem .
this fact implies that if we want to restrict some geometrical objects @xmath6\in \fm$ ] by using the figures or something in a daily life , we do not need a complete caging in definition [ def : caging ] but we have to find whether it is not small @xmath27 of @xmath27-th @xmath122-congruence . in other words
, roughly speaking , there is the degree of the difficulty to take @xmath1 in the inside of @xmath10 to the outside of @xmath10 , though the inside and the outside are not rigorous mathematically .
we need not discriminate between the higher type of @xmath122-congruence and a complete @xmath122-caging set in a daily life .
though it is very difficult to introduce the measure of the path space in general , we could measure the difficulty of caging and the probabilistic treatment of connected path in @xmath6_{\ck^c}$ ] @xcite . for given positive integer @xmath157 ,
if we find @xmath0-subspaces @xmath1 and @xmath28 in @xmath122 such that @xmath1 and @xmath28 are congruent but not @xmath27-th @xmath122-congruent for @xmath158 , we say that @xmath10 dissociates @xmath1 and @xmath28 by @xmath157-th caging .
more practically , the first caging is much more important than higher @xmath27-th caging case because the concerned shape @xmath6 $ ] is not so complicate .
even for the first caging , it is not easy to find whether it is the first caging or not because the dimension of @xmath14 is six if @xmath134 .
determination of the first caging means the determination of topological property of @xmath14 .
for example , if we reduce the determination of continuous space to finite problem by expressing concerned area and @xmath10 in terms of the voxels for 100 points per one - dimension , we have to deal with @xmath159 data , which is huge ; we can not deal with it practically in this stage @xcite .
thus in order to avoid the problems , there are several attempts and proposals including the c - closure method .
the second named author has investigated the problem using the c - closure concept @xcite . in another way , intuitive geometrical features such as loop shape @xcite and _ double fork and neck _ @xcite help us derive sufficient conditions for caging constraint .
hiroyasu hamada + national institute of technology , sasebo college , + 1 - 1 , okishin - machi , sasebo , nagasaki 857 - 1193 , japan , + institute of mathematics for industry , kyushu university , + motooka 744 , nishi - ku , fukuoka 819 - 0395 , japan + satoshi makita + national institute of technology , sasebo college , + 1 - 1 , okishin - machi , sasebo , nagasaki 857 - 1193 , japan , + shigeki matsutani : + national institute of technology , sasebo college , + 1 - 1 , okishin - machi , sasebo , nagasaki 857 - 1193 , japan , + institute of mathematics for industry , kyushu university , + motooka 744 , nishi - ku , fukuoka 819 - 0395 , japan + [email protected] + | it is a crucial problem in robotics field to cage an object using robots like multifingered hand .
however the problem what is the caging for general geometrical objects and robots has not been well - described in mathematics though there were many rigorous studies on the methods how to cage an object by certain robots . in this article , we investigate the caging problem more mathematically and describe the problem in terms of recursion of the simple euclidean moves . using the description , we show that the caging has the degree of difficulty which is closely related to a combinatorial problem and a wire puzzle .
it implies that in order to capture an object by caging , from a practical viewpoint the difficulty plays an important role .
* keywords * caging , euclidean move , wire puzzle , robotics * ams : * 51m04 , 57n15 , 57n35 , 70b15 , 70e60 , |
the structure of magnetic field is often an important factor in many energetic processes in the solar corona .
this especially refers to the topological features of magnetic structure such as null points , separatrix surfaces , and separator field lines .
they serve as preferred sites for the formation of current sheets and the corresponding accumulation of the free magnetic energy @xcite .
the magnetic reconnection process induced in the current sheets at some critical parameters allows the accumulated magnetic energy to convert into other forms : thermal , radiative and kinetic energy of plasma and accelerated particles .
this process is considered to be a driving mechanism of many energetic phenomena in the solar atmosphere @xcite . over the last decade
, it also became clear that the geometrical analogs of the separatrices @xcite , the so - called quasi - separatrix layers ( qsls , @xcite ) , have similar properties . there are indications that the qsls are probably more ubiquitous than the true separatrices @xcite .
this increases the significance of the problem of determining qsls in a given magnetic configuration . in comparison with the separatrices
, the determining of qsls requires a more sophisticated technique , which is based on a point - wise analysis of the magnetic field line connectivity .
the basic quantity in this technique is the squashing degree or factor @xmath0 of elemental magnetic flux tubes .
this quantity has previously been defined for the planar geometry @xcite , which provides a good approximation for describing magnetic structures in active regions with the characteristic size smaller than the solar radius @xmath2 .
such an approximation , however , is hardly applicable for a global description of magnetic connectivity in the solar corona including the open magnetic field of the coronal holes .
the corresponding large - scale structure of magnetic fields is also of substantial interest for solar physics , especially , for understanding solar eruptions .
so the respective generalization of the above technique must allow us to determine @xmath0 for the coronal volume bounded by the photospheric and solar - wind surfaces .
this immediately raises technical problems , which do not exist in the case of the planar geometry .
first , both these boundary surfaces are closed , and therefore , none of them admits a global regular system of coordinates . to avoid a coordinate singularity of a spherical - pole type , at least two overlapping coordinate systems ( coordinate charts ) must be used in this case for describing the locations of the field line footpoints on each of the boundries .
second , the solar - wind boundary surface generally can not be a sphere , but some other curvilinear surface whose geometry depends on the coronal magnetic field @xcite .
these two requirements of the technique can be satisfied only by using a covariant approach to the description of @xmath0 with the coordinate systems that are generally different for each of the boundaries .
the derivation of such a covariant expression for the squashing factor is one of the goals of the present work .
the second goal of the work is to make an essential refinement of the squashing factor itself .
the problem is that the large values of @xmath0 may be caused not only by squashing of elemental flux tubes in the volume but also by a projection effect at the boundary surfaces .
the latter occurs at the field lines which are nearly touching the boundary at least at one of the footpoints .
this effect , in particular , takes place in the vicinity of the bald patches ( bps , @xcite ) , which are the segments of the photospheric polarity inversion line ( pil ) , where the coronal field lines touch the photosphere .
when analyzing magnetic connectivity , it is important to discriminate between the projection effect and volumetric squashing . for this purpose ,
we derive a covariant expression for the perpendicular squashing factor @xmath1 , which describes the squashing of elemental flux tubes only in the directions orthogonal to the field lines . sections [ s : q_cov ] and [ s : q_perp ] present the derivations of @xmath0 and @xmath3 and demonstrate on the examples of planar and spherical geometry how to apply these general expressions .
the difference between @xmath0 and @xmath3 is considered in detail in [ s : qvsqp ] by calculating and comparing these quantities for two particular magnetic configurations .
the obtained results are summarized in [ s : s ] .
consider a plasma - magnetic configuration in a finite volume with a smooth boundary of an arbitrary shape .
it may generally consist of two or even more surfaces , for example , the photosphere and the solar - wind surface form a boundary for the entire solar corona .
each of the two footpoints of a given magnetic field line may belong in general to any of these surfaces .
we will use the designations launch " and target " for the footpoints and parts of the boundary surfaces at which the field lines start and end up .
let @xmath4 and @xmath5 be the systems of curvilinear coordinates at the launch and target boundaries , respectively .
the magnetic field lines connecting these boundaries define a mapping @xmath6 determined by some vector - function @xmath7 .
the local properties of this mapping are described by the jacobian matrix @xmath8 .
\label{d } \end{aligned}\ ] ] for each field line , this matrix determines a linear mapping from the tangent plane at the launch footpoint to the tangent plane at the target footpoint , so that a circle in the first plane is mapped into an ellipse in the second plane ( fig .
[ f : f1]a ) .
the aspect ratio of such an ellipse defines the degree of a local squashing of elemental flux tubes , which means that any infinitesimal circle centered at a given launch point is mapped along the field lines into an infinitesimal ellipse with this aspect ratio at the target footpoint .
this generalizes a coordinate - free definition of the squashing factor to the case of curvilinear boundaries , whose tangent planes are generally not the same , as is in the case of plane boundaries considered in @xcite and @xcite . to derive an analytical expression for
the aspect ratio of the above ellipse , let us introduce first a vector function @xmath9 that describes in a three - dimensional ( 3d ) cartesian system of coordinates the locations of the footpoints at the launch boundary .
then the vectors @xmath10 determine at this boundary the covariant vector basis tangent to the @xmath11-coordinate lines .
thus , @xmath12 is the corresponding covariant metric tensor , which determines local lengths and angles at the launch boundary .
the dot here stands for the usual scalar product in 3d euclidean space . using ( [ ek ] ) and the standard gramm - schmidt procedure
, one can construct an orthonormal basis @xmath13 hereafter @xmath14 $ ] is a determinant of the _ covariant _ metric tensor .
now , any point of a circle of unit radius in the plane tangent to the launch boundary is represented by the vector @xmath15 whose angle parameter @xmath16 .
suppose that the vector - function @xmath17 defines the points at the target boundary , then @xmath18 is the field - line mapping image of @xmath19 at the tangent plane of the target boundary , where the corresponding covariant basis vectors , parallel to the @xmath20-coordinate lines , are @xmath21 hereafter a summation over repeating indices with their values running from 1 to 2 is assumed . with varying @xmath22 ,
the vector @xmath23 traces in this plane an ellipse such that @xmath24 , \label{ot2}\end{aligned}\ ] ] where the asterisk indicates that @xmath25 is a result of evaluating @xmath26 at the target footpoint @xmath27 .
after some simple trigonometry and lengthy algebra using equations ( [ ek])([e2 ] ) , equation ( [ ot2 ] ) is reduced to @xmath28 where @xmath29 and @xmath30 are determined by @xmath31 in which @xmath32 and @xmath33 denote the determinants of _ covariant _ metric tensors at the launch and target footpoints , respectively .
the components of the _ contravariant _ metric tensor @xmath34 can be viewed here as elements of the inverted matrix @xmath35^{-1}$ ] of the covariant metric .
the value @xmath36 is simply @xmath22 plus an additional value , which is independent of @xmath22 and whose expression does not matter for the present consideration .
what actually matters is that @xmath37 runs values from @xmath38 to @xmath39 when @xmath40 and @xmath41 are tracing , respectively , the above circle and ellipse .
the minimum @xmath38 and maximum @xmath39 correspond here to the minor and major axes of the ellipse , respectively , so that its aspect ratio is @xmath42 the large values of this ratio do not differ much from its asymptotic value @xmath43 note also that @xmath44 , since inverting equation ( [ are ] ) yields @xmath45 and @xmath46 . therefore equation ( [ q ] ) will be used as a covariant definition of the squashing factor .
it is evident from the derivation of @xmath0 that this value is invariant to the direction of field line mapping .
indeed , the inverse mapping implies locally that @xmath47 is a maximum stretching coefficient and @xmath48 is a minimum shrinking coefficient .
such coefficients will coincide with the lengths of the major and minor axes of the ellipse obtained from a circle of a unit radius due to this inverse mapping .
thus , although this new ellipse has different lengths of axes , their ratio is the same as for the previous one , which proves the statement .
a formal proof of the statement is also not difficult to obtain by using the derived expressions of @xmath29 and @xmath30 in a similar way as in the case of plane boundaries @xcite .
the invariancy of @xmath0 to the direction of field line mapping justifies its status of a correct measure for the magnetic connectivity .
note also that @xmath30 for a given infinitesimal flux tube is a ratio of its cross section areas at the target and launch points .
therefore , since the magnetic flux is conserved along the tubes , this value coincides with the corresponding inverse ratio of the normal field components , so that @xmath49 where @xmath50 and @xmath51 are normal components of the magnetic field to the boundaries at the conjugate launch and target footpoints . in practice ,
the numerical calculation of @xmath30 through this ratio is more precise than that given by equation ( [ dltc ] ) and therefore it should be used for computing @xmath30 in equation ( [ q ] ) .
the above mathematical construction is related to the cauchy - green deformation tensor @xcite known in the theory of elasticity .
it can be written in our notations as @xmath52 where @xmath7 and @xmath53 represent , respectively , a finite deformation and covariant metric tensor of an elastic two - dimensional body . the contraction of the cauchy - green tensor with a pair of orthonormal vectors @xmath54 and @xmath55 yields the tensor @xmath56 such that its eigenvalues coincide with the squared semiaxes @xmath57 and @xmath58 of the above ellipse . the square root of their ratio defines in accordance with ( [ are ] ) and ( [ q ] ) the squashing factor @xmath0 .
it should be emphasized that this analogy is possible only in our general approach , where two independent systems of coordinates are used for describing the location of the conjugate footpoints .
this allows us to apply coordinate transformations only at the launch boundary , while keeping the coordinates at the target boundary unchanged .
with respect to these transformations , the object defined by ( [ c ] ) does behave as a covariant second - rank tensor .
the latter is not valid , however , if one global 3d system of coordinates is used for describing the entire field configuration and so both boundaries are subject then to coordinate transformations .
this has only a methodological meaning and does not exclude , of course , an application of the derived expressions to such particular cases .
for example , consider a closed magnetic configuration in the half space @xmath59 with the global cartesian coordinates @xmath60 and the photospheric boundary plane @xmath61 .
the field line mapping is then given by @xmath62 .
there are no more differences between upper and low indices and contravariant @xmath34 and covariant @xmath63 metrics ; the latter simply turn into kronecker symbols @xmath64 and @xmath65 .
so equations ( [ n2c ] ) , ( [ dltc ] ) and ( [ dlt2 ] ) are reduced to @xmath66 as required in this case @xcite .
consider now a more complicated class of configurations , where both open and closed magnetic field lines are present .
let the configuration be described in one global system of coordinates @xmath67 , where @xmath68 corresponds to the photospheric launch boundary , while @xmath69 represents the target boundary .
for the open field lines reaching the spherical solar - wind boundary of radius @xmath70 , we put @xmath71 , while for the closed ones we take @xmath72 . thus , @xmath73 , @xmath74 and the field line mapping is @xmath75 , which yields @xmath76 = \left ( \begin{array}{cc } r_{*}^{2}\sin^{2}\theta & 0 \\ 0 & r_{*}^{2 } \end{array } \right ) \label{gijs * } \end{aligned}\ ] ] and @xmath77 = \left ( \begin{array}{cc } r_{\sun}^{-2}\sin^{-2}\theta & 0 \\ 0 & r_{\sun}^{-2 } \end{array } \right ) , \label{gkls } \end{aligned}\ ] ] where the contravariant metric @xmath34 at the launch boundary is obtained from the corresponding covariant metric by inverting it simply as a @xmath78 matrix . using these expressions and equation ( [ n2c ] )
, we obtain @xmath79 .
\label{n2sph}\end{aligned}\ ] ] equations ( [ dltc ] ) and ( [ dlt2 ] ) in this case become @xmath80 the obtained expressions for @xmath29 and @xmath30 have seeming singularities at the poles , where they actually reduce in the generic case to resolved indeterminacies with @xmath81 and @xmath82 proportional to @xmath83 .
these indeterminacies are artificial and unrelated to some special properties of the magnetic structure ; they are caused by the pole singularities inherent in the used global spherical system of coordinates .
moreover , their appearance is unavoidable in any other global system of coordinates on a closed sphere - like surface because of its intrinsic topological properties .
therefore this may generally reduce precision of a numerical evaluation of @xmath0 near the pole - like points . to avoid such indeterminacies , at least two overlapping coordinate charts on each of the spherical boundaries
are required . for this purpose , it is sufficient to use two systems of spherical coordinates turned with respect to each other on the right angle in the @xmath84-direction .
switching from one of such systems in its polar regions to the other , as suggested , for example , by @xcite , makes it possible to resolve the problem of the pole indeterminacies .
the required expressions for such calculations of @xmath0 can be obtained again from equations ( [ n2c ] ) and ( [ dltc ] ) with properly modified metric tensors .
to find the perpendicular squashing factor @xmath3 , we need to know the field line mapping between infinitesimal planes orthogonal to the field lines at the conjugate footpoints ( fig .
[ f : f1]b ) .
note first that the projection effect at the boundaries is local , because it depends only on the orientations of the tangent planes at the boundaries with respect to the vectors of magnetic field at the footpoints .
so the required mapping between the indicated orthogonal planes can be obtained from the respective mapping between the tangent planes by correcting it only at such footpoints .
this implies that @xmath3 can be expressed in terms of the same values as @xmath0 and , in addition , the field vectors @xmath85 and @xmath86 , respectively , at the launch and target footpoints .
we will derive @xmath3 by using the same procedure as for @xmath0 while modifying it in accordance with the above comments .
the vector @xmath19 tracing the circle of unit radius is given by the same expression ( [ o ] ) .
however , since it lies now in the plane perpendicular to the field line at the launch point , the corresponding orthonormal basis is chosen to be orthogonal to @xmath85 , so that @xmath87 this vector @xmath19 is mapped along the field lines into a vector @xmath23 lying in the plane perpendicular to the local field @xmath86 at the target footpoint .
the respective mapping @xmath88 can be represented as a composition @xmath89 of three others according to the following diagram : @xmath90 here the mapping @xmath91 projects the vector @xmath19 along @xmath85 onto the plane tangential to the launch boundary to yield the vector @xmath92 which has a vanishing component along the vector @xmath93 perpendicular to such a plane .
then this vector @xmath94 is mapped by the differential of the field line mapping @xmath95 determined by the jacobian matrix ( [ d ] ) into the vector @xmath96 which lies in the plane tangential to the target boundary .
finally , the obtained vector @xmath97 is projected by @xmath98 along @xmath86 at the target footpoint onto the plane perpendicular to @xmath86 to result in @xmath99 eliminating now @xmath97 and @xmath94 from ( [ o_])([otp ] ) and using ( [ o ] ) with the basis from equations ( [ e1p])([e2p ] ) , we express @xmath23 in terms of @xmath100-basis and magnetic field vectors at the conjugate footpoints .
this allows us to calculate @xmath101 and then @xmath3 in a similar way as done before when deriving @xmath0 .
the result is @xmath102 where the asterisk has the same meaning as in equations ( [ ot2 ] ) , ( [ n2c ] ) and ( [ dltc ] ) .
thus , the obtained @xmath3 differs from @xmath0 only in the form of the metric tensors , which are determined now by @xmath103 the asterisk here implies automatically that the corresponding values refer to the target footpoint , therefore the tilde used for indicating this fact in intermediate equation ( [ otp ] ) is omitted in the final expression ( [ gpij * ] ) .
one can also check that @xmath104 = \left[g_{\perp lk}\right]^{-1}$ ] and @xmath105 = \frac{\left(\bm{b\cdot}{{\bm \epsilon}}_{1 } \bm{\times}{{\bm \epsilon}}_{2}\right)^{2 } } { \bm{b}^{2 } } , \label{gp}\end{aligned}\ ] ] so that equation ( [ dltpc ] ) reduces to @xmath106 if , in addition , the magnetic flux conservation is taken into account .
this expression should be used instead of ( [ dltpc ] ) for computing @xmath107 in ( [ qpc ] ) for the same reason that ( [ dlt2 ] ) should be used in ( [ q ] ) .
thus , formulas ( [ qpc])([dltpc2 ] ) completely define the covariant expression for @xmath3 .
let us see now how these formulas work in the case of a closed magnetic configuration described in a global cartesian system of coordinates @xmath108 with the photospheric boundary plane @xmath61 .
the formulas are significantly simplified in this case to yield @xmath109 where the values of the corresponding covariant and contravariant components of vectors and tensors do not differ from each other .
note also that expression ( [ n2pp ] ) apparently diverges near the pil , where the normal field component @xmath110 vanishes .
this is actually not a true singularity but rather an indeterminacy , which is resolved to give a low limit of @xmath111 if the pil has no bps . at the bps , such an indeterminacy
is also resolved but it may generally have different limits of @xmath112 at the left and right sides of bps . thus , @xmath3 may experience a jump when crossing bps , unless the configuration is symmetric as in the example of [ s : hfttwst ] .
similar to
[ s : q_cov ] , consider also a more general class of configurations , where both open and closed field lines are present and bounded by a spherical solar - wind surface of radius @xmath70 and the photosphere of radius @xmath2 as before .
it is convenient in this case to use matrix notations , in which equation ( [ n2pc ] ) is written as @xmath113 where @xmath114 \displaystyle\frac{\partial \theta}{\partial \phi } & \displaystyle\frac{\partial \theta}{\partial \theta } \end{array } \right ) \label{do}\end{aligned}\ ] ] is the jacobian matrix of the field line mapping .
the covariant and contravariant metrics at the target and launch boundaries , respectively , are determined by the following matrices : @xmath115^ { * } \nonumber \\ & & = r^{2 } _ { * } \left ( \begin{array}{cc } \displaystyle \sin^{2}\theta \left(1-\frac{b_{\phi}^{*2}}{{\bm b}^{*2}}\right ) & \displaystyle -\sin\theta \frac{b_{\phi}^ { * } b_{\theta}^{*}}{\bm{b}^{*2 } } \\ [ 8pt ] \displaystyle -\sin\theta\frac{b_{\phi}^ { * } b_{\theta}^{*}}{\bm{b}^{*2 } } & \displaystyle 1-\frac{b_{\theta}^{*2}}{\bm{b}^{*2 } } \end{array } \right ) , \qquad \label{gpo * } \\ [ 8pt ] & & g^{\perp } \equiv \left[g^{\perp lk}\right ] \nonumber \\ & & = r^{-2}_{\odot } \left ( \begin{array}{cc } \displaystyle\sin^{-2}\theta\left(1+\frac{b_{\phi}^{2}}{b_{r}^{2}}\right ) & \displaystyle \frac{b_{\phi}b_{\theta}}{\sin\theta\ , b_{r}^{2 } } \\ [ 10pt ] \displaystyle\frac{b_{\phi}b_{\theta}}{\sin\theta\ , b_{r}^{2 } } & \displaystyle 1+\frac{b_{\theta}^{2}}{b_{r}^{2 } } \end{array } \right ) , \label{gpo}\end{aligned}\ ] ] where @xmath72 if a given footpoint belongs to a closed field line and @xmath71 otherwise .
the matrix @xmath116 has two types of singularities , which actually lead only to indeterminacies in expression ( [ n2po ] ) .
the first indeterminacies take place at the poles of spherical coordinates ( @xmath117 ) .
they are already discussed above in connection with equation ( [ n2sph ] ) .
the second indeterminacies take place at the pil , where @xmath118 .
they are resolved in much the same way as occurred in the previous case of cartesian geometry to give a finite value of @xmath119 .
if the pil has no bps , the length of the field lines vanishes near the pil , so that @xmath120 and @xmath121 , which results in @xmath122 .
the presence of bps implies , however , a strong projection effect in their neighborhood and the corresponding singularities in the jacobian matrix @xmath123 at the bps .
these singularities are exactly of the same type and value as those in @xmath116 but opposite in signs .
so they cancel each other in equation ( [ n2po ] ) to generally provide different limits at the different sides of the bps , as discussed above for the case of the plane boundaries .
it should be also noted that there is one more type of plausible singularities not yet discussed .
these singularities have to appear if a given configuration has null points of magnetic field and the corresponding separatrix field lines . in this case , the elemental flux tubes enclosing the separatrix field lines split at each of the null points to produce singularities in the derivatives of the jacobian matrix .
such singularities are due to volumetric rather than surface properties of magnetic configurations . therefore ,
if present , they appear in both @xmath0 and @xmath3 distributions thereby indicating the existence of magnetic nulls in the corona .
since the numerical derivatives are estimated as a ratio of finite coordinate differences , their absolute values may not exceed the ratio of the coordinate range at the target boundary to a chosen increment of coordinates at the launch boundary .
the presence of this upper bound on possible values of the numerical derivatives prevents an overflow error in computations of the squashing factors at the footpoints of the null - point separatrices .
a more detailed consideration of this type of singularity requires a special study , which goes far beyond the scope of the present work .
to see the difference between @xmath0 and @xmath3 , their distributions are compared below for several magnetic configurations .
an emphasis is made on their potentiality for determining so - called hyperbolic flux tubes ( hfts ) .
they are defined as two intersecting qsls with extended and narrow photospheric footprints characterized by very large values of @xmath0 . neglecting the possible curving and twisting of an hft , its cross section variation in the longitudinal direction can be respresented as @xmath124{ii1 } \label{-x-}\ ] ] in other words , the width of one of the qsls is shrunken to the thickness of the other in the process of the mapping of their cross sections along magnetic field lines .
this is possible because the field lines in hfts exponentially converge in one transversal direction and diverge in the other .
such a property is typical for hyperbolic flows in the theory of dynamical systems @xcite , which was coined in the term hft @xcite .
the examples below demonstrate that @xmath3 is a more accurate quantity than @xmath0 for characterizing hfts , although @xmath0 also provides valuable complementary information on the magnetic structure . the most simple magnetic configuration in which one would expect the presence of an hft is the so - called x - line configuration , whose field @xmath85 is determined in cartesian coordinates @xmath125 by @xmath126 where @xmath127 is a dimensionless field gradient characterizing the strength of the transverse field @xmath128 compared to the longitudinal field @xmath129 on a characteristic length scale @xmath130 .
the right - hand side of ( [ bx ] ) represents the same field in terms of the euler potentials @xmath131 which are constant along the field lines .
the use of @xmath132 and @xmath133 significantly simplifies the calculation of the squashing factors in this configuration .
suppose that its volume is restricted by @xmath134 , so that the planes @xmath135 are the corresponding boundaries .
the constancy of @xmath132 and @xmath133 means that the boundary points @xmath136 and @xmath137 are related by @xmath138 and @xmath139 . in terms of the notations used in expressions ( [ n2 ] ) for the plane boundaries , this means that @xmath140 and @xmath141 .
since @xmath142 everywhere , equation ( [ dlt ] ) reduces simply to @xmath143 , so that the resulting squashing factor is @xmath144 thus , @xmath0 is constant over the entire planes @xmath135 and , hence , it does not determine any qsl in the x - line configuration with the plane boundaries .
@xcite has arrived at the same conclusion by using the norm @xmath145 defined by ( [ n2 ] ) , since @xmath142 , @xmath143 and so @xmath146 in this configuration .
however , they have found with the help of @xmath145 some evidences of the presence of qsls in the x - line configuration , if it is bounded by cubic , hemispheroidal and spherical surfaces . on the other hand ,
it is clear from the general point of view that there should be an hft in such a configuration irrespective of the shape of the boundary surfaces .
indeed , this configuration has in the unbounded space two genuine separatrix planes @xmath147 and @xmath148 , which can be regarded as a limiting case of the bounded configuration with the boundaries moved off to infinity .
taking into account also that @xmath149 , one could expect that the proper measure for qsls must be growing with @xmath150 near the planes @xmath147 and @xmath148 much stronger than in the remaining volume to indicate the corresponding qsls near these planes .
this would provide in the limit of large @xmath150 an expected continuous transition of such bounded configurations with qsls to the unbounded configuration with the genuine separatrices .
the failure of @xmath0 in determining an hft in this simple case looks very surprising in light of its remarkable success in other more complicated field configurations ( see [ s : hfttwst ] ) .
the reason for this failure lies actually in the above - mentioned projection effect , which is extremely large for the chosen type of boundaries .
the transverse component here grows linearly with the distance from the x - line @xmath151 , while the longitudinal component remains constant .
so the farther a given field line meets the boundary from the x - line , the more they become aligned with each other . such an explanation is fully confirmed by calculations of the perpendicular squashing factor @xmath3 in this case . to derive an analytical expression of @xmath3 ,
let us choose @xmath152 as coordinates on both boundary surfaces , so that @xmath153 and @xmath154 in ( [ qpc])([dltpc2 ] ) .
assume for generality that the boundaries are defined by @xmath155 , then according to ( [ xi ] ) and ( [ eta ] ) the vector functions @xmath156 define , respectively , the launch and target boundary surfaces .
these formulas are needed for calculating ( [ ek ] ) , ( [ glk ] ) , ( [ ei ] ) , ( [ n2pc ] ) ( [ gpij * ] ) , ( [ gplk ] ) , ( [ dltpc2 ] ) and ( [ bx ] ) in the chosen coordinates @xmath152 . with the help of such calculations , equation ( [ qpc ] ) yields @xmath157 + h^{2 } \left(x_{+}^{2}+y_{+}^{2}+x_{-}^{2}+y_{-}^{2}\right ) } { \sqrt{\left[1+h^{2}\left(x_{+}^{2}+y_{+}^{2}\right)\right ] \left[1+h^{2}\left(x_{-}^{2}+y_{-}^{2}\right)\right ] } } , \label{qpx}\end{aligned}\ ] ]
in which @xmath158 determine @xmath159 coordinates of the conjugate footpoints at the defined boundaries .
these expressions determine @xmath3 as a function of the euler potentials @xmath132 and @xmath133 .
note also that @xmath160 and @xmath161 at @xmath162 , so expressions ( [ qpx])([ypm ] ) define in addition the distribution of @xmath3 in the plane @xmath162 , to which @xmath3 is mapped along the field lines from the boundaries . more generally , the combination of these expressions with equations ( [ xi ] ) and ( [ eta ] ) , resolved with respect to @xmath163 and @xmath164 as @xmath165 provides a parametrical representation of @xmath3 , with @xmath132 and @xmath133 as parameters , in any plane @xmath166 between the boundaries . in the particular case of plane boundaries
, we have to put @xmath167 in ( [ qpx])([ypm ] ) . using then ( [ xi ] ) , ( [ eta ] ) , ( [ xpm ] ) and ( [ ypm ] ) , expression ( [ qpx ] ) for @xmath3 can be rewritten even as an explicit function of @xmath125 . by comparing ( [ qpx ] ) and ( [ qx ] ) one can see also that the terms containing @xmath168 and @xmath169 are responsible in this case for eliminating the projection effect . as a result of this , the distribution of @xmath3 in the @xmath162 plane shows very pronounced ridges " along the @xmath163 and @xmath164 axes ( fig .
[ f : f2]a ) by revealing an expected hft with a characteristic x - type intersection of qsls along the x - line . according to equations ( [ x ] ) and ( [ y ] ) , such a structure shrinks and expands along these axes exponentially fast with @xmath170 to give a typical hft variation of its cross section ( see diagram [ [ -x- ] ] ) .
with growing @xmath150 , the ridges of the @xmath3 distribution and so the corresponding qsls become thinner and thinner by extending on larger and larger distances from the x - line .
thus , the perpendicular squashing factor @xmath3 defines indeed an hft such that it continuously transforms in the limit of large @xmath150 into the separatrix planes @xmath147 and @xmath148 . however , the peripheral field lines in this hft are still nearly parallel to the boundaries @xmath135 , which causes too strong of a variation of the hft cross section .
aesthetically more pleasing hft in the x - line configuration can be obtained by using the boundaries which are orthogonal to the field lines .
these are iso - surfaces of magnetic potential @xmath171 having the shape of hyperbolic paraboloids .
it is natural to chose them passing through the points @xmath172 , which means that @xmath173 for such iso - surfaces .
this condition using ( [ xpm ] ) and ( [ ypm ] ) yields @xmath174 which is a transcendental equation for the @xmath175 functions entering in expression ( [ qpx ] ) .
this equation is not difficult to solve numerically for given @xmath132 and @xmath133 and to use the respective solution for the calculation of @xmath3 .
an example of the resulting @xmath3 distribution in the plane @xmath162 at @xmath176 is shown in figure [ f : f2]b .
the corresponding hft with the magnetic surface defined by @xmath177 and boundaries @xmath173 is presented in figure [ f : f2]c . for the chosen type of boundaries , @xmath178 , so both squashing factors define the same hft .
the magnetic field for this hft has the simplest analytical form ; the hyperbolic paraboloids are also relatively simple boundary surfaces .
so we believe that this example provides the simplest possible hft relevant for theoretical studies of basic mhd processes , such as magnetic pinching and reconnection , in three dimensions .
the considered x - line configuration with @xmath135 boundaries is an important example , where @xmath3 , in contrast to @xmath0 , succeeds in determining an expected hft .
however , this example is not representative enough to make a general conclusion on the potentialities of @xmath0 and @xmath3 for detecting qsls . because the field lines in such a configuration behave in a rather artificial way over a major part of the boundaries .
a better comparison of @xmath0 and @xmath3 can be done by using a more realistic field . for this purpose
, we have chosen the analytical model of a twisted magnetic field @xcite , hereafter called the t&d model .
it describes approximate equilibria of a circular magnetic flux rope , whose interior force - free field is continuously embedded into a potential background field .
the latter is produced by fictitious subphotospheric sources consisting of two magnetic monopoles of opposite signs and a line current , all located at the axis of symmetry of the rope .
the axis itself is placed some depth below the photospheric plane and the minor radius of the rope @xmath179 is assumed to be much smaller than the major one @xmath180 and the distance between the monopoles @xmath130 . to compare @xmath0 and @xmath3 in detail ,
we have computed their distributions for three sets of parameters which differ only in values of @xmath180 .
two of these values ( @xmath181 and @xmath182 ) and all the remaining parameters are chosen to be exactly the same as in the t&d model . by growing @xmath180 but keeping other parameters fixed
we imitate an emergence of the flux rope from below the photosphere .
in this process , the configuration passes continuously through three distinct topological phases . for sufficiently small @xmath180
, there is a single bp separatrix surface @xcite the configuration with @xmath183 represents one of these topological states .
the corresponding distributions of @xmath0 and @xmath3 are shown in figures [ f : f3]a and [ f : f3]d , respectively . with growing @xmath180
, this bp and the associated separatrix surface bifurcate into two parts to give birth to a bp separator field line @xcite the configuration with @xmath184 represents this second topological phase .
the corresponding distributions of @xmath0 and @xmath3 are shown in figures [ f : f3]b and [ f : f3]e , respectively .
the points @xmath185 and @xmath186 on these figures are the footpoints of the bp separator , while @xmath187 and @xmath188 are its photospheric contact points .
further growing of @xmath180 leads to a complete disappearance of the bifurcated bp and the associated separatrix surfaces the configuration with @xmath189 represents this last phase .
the corresponding distributions of @xmath0 and @xmath3 are shown in figures [ f : f3]c and [ f : f3]f , respectively .
the magnetic field at these parameters becomes topologically trivial , since its field line mapping is continuous everywhere .
the whole structure can be continuously transformed to a simple arcade - like configuration with the help of a suitable photospheric motion .
thus , with growing @xmath180 or emerging of the flux rope the topological complexity of the configuration first increases and then abruptly decreases . on the contrary ,
the geometrical distortion of the field line mapping gradually increases in the configuration during this process .
both squashing factors @xmath0 and @xmath3 continue to grow with @xmath180 in narrow strips of the photospheric plane .
it is clearly seen from the @xmath0 and @xmath3 distributions plotted with the same grayscaling ( fig .
[ f : f3 ] ) that @xmath190 everywhere and the difference between @xmath0 and @xmath3 becomes smaller with growing @xmath180 .
the most significant difference between them is seen at the first two phases near the bps and the footprints of the associated separatrix surfaces .
as previously anticipated , @xmath0 always rises in these regions , while @xmath3 does not , except near the contact points @xmath187 and @xmath188 of the bp separator ( fig .
[ f : f3]e ) .
the value @xmath3 does rise there but only to give birth to a part of hft footprints , which are matured eventually in the third phase ( fig .
[ f : f3]f ) .
as concerned with the indicated rise of @xmath0 , it is mainly caused by the projection effect : the field lines which are close to the bp separatrix surfaces approach the photosphere near the bps at a small angle to the horizontal , which strongly distorts the footprints of the corresponding elemental flux tubes . comparing @xmath0 and @xmath3 at @xmath189
reveals that the latter is valid for the central part of the pil as well .
the discussed features of the distributions become more transparent if one saturates the grey shading in the plots at the values @xmath191 ( fig .
[ f : f4 ] ) .
these new plots show that the qsls based on the @xmath3 distribution are characterized by a thinner and more uniform thickness .
their footprints acquire at @xmath189 a clear fishhook - like shape in each of the photospheric polarities with maximums of @xmath0 reaching @xmath192 .
the qsls rooted at such fishhooks " intersect each other by combining themselves into an hft @xcite .
this structural feature seems to be very robust , because it appears even in twisted configurations which are not in force - free or magnetostatic equilibrium @xcite .
one can see from figure [ f : f5]a that , except for an essential twisting distortion , the cross section of such an hft varies exactly according to diagram ( [ -x- ] ) .
it has yet to be proved , but it seems to be quite natural that this hft is pinched into a vertical current sheet below the flux rope by its upward movement when the kink or torus instability is developed in the configuration at a sufficiently large twist of the field lines in the rope @xcite .
this interpretation is very important for understanding the properties of sigmoidal structures in flaring configurations @xcite .
figure [ f : f5]a suggests that the sigmoids in such configurations are simply pinching hfts illuminated by a hot plasma material which appears there due to the reconnection process in the above - mentioned vertical current sheet .
this seems to be valid at least for the third topological phase of the flux rope emergence , while an additional interaction with the photosphere must be involved at the first and second topological phases , where the bps are present @xcite .
panels ( _ d_)(_f _ ) in figures [ f : f3 ] and [ f : f4 ] demonstrate that the footprints of the bp separatrix surfaces follow very close to the hft footprints emerging gradually with growing @xmath180 .
this implies the corresponding similarity in the shapes of such separatrix surfaces and hfts .
so the explanations of the sigmoids that rely on either the presence of the bps @xcite or the hfts @xcite are not alternative but rather complementary , since they refer to different phases of the flux rope emergence .
it should be noted also that both @xmath0 and @xmath3 distributions ( fig .
[ f : f4 ] ) contain at the border of the flux rope two less pronounced horseshoe - like features with maximums @xmath193 .
the qsl rooted at these horseshoes " has a helical shape ( fig .
[ f : f5]b ) with a slightly varying cross section along the field lines .
so this qsl has a structure qualitatively different from those two which form the hft .
the comparison of @xmath0 and @xmath3 shows that the squashing of the flux tubes in this helical qsl is only in part due to the projection effect .
the @xmath3 distribution demonstrates that the major contribution to @xmath0 comes from the shearing of the twisted field lines in the rope .
thus , both distributions reveal a helical qsl which is a part of the inner border layer of the flux rope . in this respect
, the considered example demonstrates that our squashing factors help identify flux ropes themselves .
it is not a problem , of course , to locate the flux rope in the t&d model , where its parameters are known from the construction of the model .
yet identifying flux ropes in more complicated configurations obtained , for instance , numerically from magne- togram data is a real problem .
if such configurations are topologically trivial , like those in the third phase of our example , the determination of qsls seems to be the only method for identifying flux ropes . as shown above , both squashing factors allow us in the case of the t&d model to determine similar qsls , except that @xmath3 is more advanced than @xmath0 near bps , whenever they appear .
therefore , we think that , in general , if the numerical grid used for computing @xmath3 is fine enough for detecting possible bp separatrix surfaces by sudden spikes in the @xmath3 distribution , the use of only @xmath3 would be sufficient for the structural analysis of configurations . in practice ,
however , the required resolution of the grid can not be always easily foreseen .
also the computational cost for @xmath0 is not really high , since the same input data as for @xmath3 can be used .
so it is sensible to compute both these distributions at a time and compare them in the same way as we did in the considered example .
some redundancy of the information contained in such distributions is not superfluous but useful , especially , in the case of complicated real magnetic configurations .
thus , from the practical point of view , the value @xmath0 should not be considered as obsolete but rather as a complementary characteristic of magnetic connectivity .
we have derived a covariant form of the squashing factor @xmath0 , which enables us to determine quasi - separatrix layers ( qsls ) in both closed and open magnetic configurations with an arbitrary shape of boundaries .
the corresponding expression for @xmath0 assumes that the jacobian matrix of the field line mapping and the metric tensors at the footpoints are known .
the expression admits also that such
input data " can be represented with the help of two different coordinate systems for determining location of the conjugate footpoints on the boundaries .
this provides a firm theoretical basis for a global description of the field line connectivity in the solar corona . to eliminate the projection effect at the field lines which are nearly touching the boundary , the perpendicular squashing factor @xmath3
is also derived in a similar covariant form .
the value @xmath3 defines the degree of squashing of elemental magnetic flux tubes only in the directions orthogonal to the field lines . in the definition of @xmath3 ,
the boundaries enclosing the magnetic configuration constrain only the length of the flux tubes while not affecting their cross sections at the footpoints . for calculating @xmath3 ,
the vectors of magnetic field at the footpoints are required in addition to the same input data as for @xmath0 .
the use of both covariant squashing factors is demonstrated by calculating them for the boundaries with the planar and spherical geometries .
then the properties of @xmath0 and @xmath3 are compared by considering two examples of magnetic configurations .
the first example is a classical x - line configuration of potential magnetic field in a plasma volume restricted by two boundary surfaces .
it is easy to show that for the plane boundaries perpendicular to the x - line , the value @xmath0 is constant .
so the @xmath0 distribution does not allow us to define any qsls in such a configuration .
the reason for this failure lies in the projection effect , which is very strong for the field lines distant from the x - line .
we have also calculated an analytical expression of @xmath3 for the same field but with the boundaries of an arbitrary shape . in the case of the plane boundaries ,
this new value @xmath3 , in contrast to @xmath0 , has a non - uniform distribution , which does reveal the expected two qsls .
these qsls intersect each other by combining themselves into a hyperbolic flux tube ( hft ) .
a more elegant hft is obtained for the x - line configuration with the boundaries orthogonal to the field lines . to make a better comparison of the properties of @xmath0 and @xmath3 , a second magnetic configuration more relevant for solar physics
is considered .
the respective field is defined by using the titov - dmoulin ( 1999 ) model of a force - free flux rope embedded into a potential background field .
contrary to the case of the x - line configuration restricted by the plane boundaries , both the @xmath0 and @xmath3 distributions reveal qsls in the twisted configuration .
these distributions are similar everywhere except near bald patches ( bps ) and footprints of the associated separatrix surfaces , whenever the bps exist . by definition ,
the value @xmath3 is free of the projection effect , so that @xmath3 rises near bps only if the corresponding flux tubes are subject to a volumetric squashing .
this is not the case , of course , for the value @xmath0 , which always rises in such regions of the photosphere .
so , in comparison with @xmath0 , the value @xmath3 shows itself to be again a superior characteristic for analysis of magnetic connectivity .
nevertheless , we have argued that it is more practical in general to compute both squashing factors for analyzing the structure of a given magnetic configuration .
this does not require additional significant effort , while making it easy to discriminate between the volumetric squashing of elemental flux tubes and the surface projection effect at the boundaries . | the study of magnetic connectivity in the solar corona reveals a need to generalize the field line mapping technique to arbitrary geometry of the boundaries and systems of coordinates .
indeed , the global description of the connectivity in the corona requires the use of the photospheric and solar wind boundaries .
both are closed surfaces and therefore do not admit a global regular system of coordinates .
at least two overlapping regular systems of coordinates for each of the boundaries are necessary in this case to avoid spherical - pole - like singularities in the coordinates of the footpoints .
this implies that the basic characteristic of magnetic connectivity the squashing degree or factor @xmath0 of elemental flux tubes @xcite must be rewritten in covariant form .
such a covariant expression of @xmath0 is derived in this work .
the derived expression is very flexible and highly efficient for describing the global magnetic connectivity in the solar corona .
in addition , a general expression for a new characteristic @xmath1 which defines a squashing of the flux tubes in the directions perpendicular to the field lines is determined .
this new quantity makes it possible to filter out the quasi - separatrix layers whose large values of @xmath0 are caused by a projection effect at the field lines nearly touching the photosphere .
thus , the value @xmath1 provides a much more precise description of the volumetric properties of the magnetic field structure .
the difference between @xmath0 and @xmath1 is illustrated by comparing their distributions for two configurations , one of which is the titov - dmoulin ( 1999 ) model of a twisted magnetic field . |
the study of higher dimensional black holes , branes and their interactions is an active field of research in several different areas of modern theoretical physics @xcite . one interesting direction , which has been first introduced by frolov @xcite , is to consider a brane - black hole ( bbh ) toy model for studying merger and topology changing transitions in higher dimensional classical general relativity @xcite , or in certain strongly coupled gauge theories @xcite through the ads / cft correspondence @xcite .
generalizations of the bbh model by studying thickness corrections to the dirac - nambu - goto effective brane action @xcite from higher order curvature terms , have also been studied recently , first by perturbative approaches @xcite , and later within an exact description @xcite .
the results of the perturbative approaches concluded that there is a
symmetry breaking between the two topologically different solution family , as regular perturbative solutions do not exist for minkowskian embedding topologies except in the special case of a 2-brane .
the problem , on the other hand , can be solved regularly for any brane dimensions in the black hole embeddings .
this virtual
symmetry breaking phenomenon obtained a simple resolution in @xcite , where it was pointed out that perturbative thick solutions break down around their thin counterparts for minkowski topologies , because the thin solutions are not analytic at the axis of the system .
motivated by this observation , in @xcite , a general family of thick solutions could be provided for both topologies within a non - perturbative numerical approach for all but one exceptional case .
the exception , mysteriously , turned out to be the same case , where the regular perturbative solution existed , namely the 2-dimensional , topologically flat case .
the above findings of @xcite naturally raised the question : how can a regular perturbative solution exist in the same single case where the regular non - perturbative solution can not be found ? in the present paper
we provide the answer to this question and obtain the so far missing solution of the topologically flat 2-brane in arbitrary number of bulk dimensions for the thick - bbh system . by including this solution
, we complete the full set of solutions of the problem that we started presenting first with a perturbative approach in @xcite and continued with a non - perturbative description in @xcite .
the present work , therefore , is the final part of the series of papers we addressed to the thickness corrected bbh problem and we kindly refer the reader also to @xcite for the more detailed model setups and for all those definitions , notations and results that might be missing here and would make the present paper completely self - contained .
the plan of the paper is as follows . in sec .
ii , we provide a short overview on the thin- and thick - bbh model setups . in sec .
iii we obtain the special case of the 2-brane equation , and in sec .
iv we analyze its regularity conditions for flat topology . in sec .
v we provide the non - perturbative , numerical solution of the problem in the near horizon region , and in sec .
vi we draw our conclusions .
let us overview quickly the most important properties of the thin - bbh system introduced by frolov in @xcite , and its thickness corrected generalization provided in @xcite .
we consider static brane configurations in the background of a static , spherically symmetric bulk black hole .
the metric of an @xmath0-dimensional , spherically symmetric black hole spacetime is @xmath1 where @xmath2 and @xmath3 is the metric of an @xmath4 dimensional unit sphere .
one can define coordinates @xmath5 on this sphere with the relation @xmath6 the explicit form of @xmath7 is not important , it is only assumed that @xmath7 is zero at the horizon @xmath8 , and it grows monotonically to @xmath9 at the spatial infinity @xmath10 , where it has the asymptotic form @xcite , @xmath11 in the zero thickness case the test brane configurations in an external gravitational field can be obtained by solving the equation of motion coming from the dirac - nambu - goto action @xcite , @xmath12 where @xmath13 is the induced metric on the brane @xmath14 and @xmath15 are coordinates on the brane world sheet .
the brane tension does not enter into the brane equations , thus for simplicity it can be put equal to @xmath9 .
it is also assumed that the brane is static and spherically symmetric , and its surface is chosen to obey the equations @xmath16 with the above symmetry properties the brane world sheet can be defined by the function @xmath17 and we shall use coordinates @xmath18 on the brane as @xmath19 the parameter @xmath20 denotes the number of dimensions in which the brane is rotationally symmetric . in this paper
we consider the special case of @xmath21 , i.e. a 3-dimensional brane world sheet , that is a 2-dimensional , axisymmetric brane embedded into the higher dimensional black hole spacetime . with this parametrization
the induced metric on the brane is @xmath22dr^2+r^2\sin^2\theta d\omega_n^2,\ ] ] where , and throughout this paper , a dot denotes the derivative with respect to @xmath23 , and the action ( [ action0 ] ) reduces to @xmath24 where @xmath25 is the interval of time and @xmath26 is the surface area of a unit @xmath20-dimensional sphere . in the case of a thick brane ,
the curvature corrected effective brane action is obtained by carter and gregory in @xcite , and the corrections to the thin dng action are induced by small thickness perturbations as @xmath27,\ ] ] where @xmath28 is the ricci scalar , @xmath29 is the extrinsic curvature scalar of the brane and the coefficients @xmath30 and @xmath31 are expressed by the wall thickness parameter @xmath32 as @xmath33 the parameter @xmath34 is related to the thickness by @xmath35 which originates from a field theoretical domain - wall model where @xmath34 is the mass parameter and @xmath36 is the coupling constant of the scalar field . after integrating out
the spherical symmetric part and the time dependence on the introduced static , spherically symmetric , higher dimensional black hole background , one obtains ( see also @xcite ) @xmath37\ , \label{l } \end{aligned}\ ] ] where we introduced the notations @xmath38 with @xmath39 here @xmath40 is the relevant dynamical length scale of the system which has to be large compared to the thickness parameter @xmath32 in order to ( [ action1 ] ) remain valid .
the explicit expressions of the curvature scalars @xmath29 and @xmath41 are given in ( 35 ) and ( 36 ) of @xcite .
for a detailed introduction of both the thin- and thick - bbh systems , please refer to @xcite .
from this section on we will focus only on the case of the topologically flat or minkowskian 2-brane . in two dimensions topologically flat is synonymous with
topologically minkowskian , and we retain both terms for the sake of elegant variation . in order to obtain the 2-brane equation of motion , first we observe that the thickness corrected dng - brane action is a function of the second derivative of @xmath42 and thus the euler - lagrange equation of the problem has the form ( see for example @xcite ) @xmath43 from ( [ mel ] ) the actual equation of motion becomes @xmath44 where , in the 2-dimensional ( @xmath21 ) case @xmath45\ , \end{aligned}\ ] ] with @xmath46 and @xmath47 is given in the appendix . as it is immediate to see , ( [ fulleq ] ) is a @xmath48-order , highly nonlinear equation , and it is probably impossible to present its solutions in closed , analytic form .
hence , the goal of this paper , is to provide a regular , numerical solution of ( [ fulleq ] ) in arbitrary bulk dimensions for flat topology .
it was pointed out first in @xcite , that the brane solutions obtained by frolov in @xcite are regular but not analytic ( or smooth ) at the axis of the thin - bbh system for the minkowski embedding branch .
in fact they are not even differentiable at that point and thus belong to the class of @xmath49 functions only . according to this property , and
since we found that the perturbative approach broke down around these solutions for the thick case , we concluded that the thick brane solutions must behave significantly differently at the axis of the system , namely we expected them to be smooth there .
it was surprising however that in the single case of the 2-brane , a regular perturbative solution existed .
we gave a detailed analysis of this solution in @xcite . in @xcite
, approaching the problem by a new , non - perturbative numerical method , we looked for the missing solutions of the minkowski branch in the class of analytic functions .
we obtained regular boundary conditions at the axis of the system by considering the series expansion of the exact @xmath48-order equation of motion around @xmath50 . with this method we successfully provided all the missing , topologically minkowskian solutions of the thick - bbh system , except in the curious case of the 2-brane again .
it was obvious , of course , that a perturbative solution can not exist without the existence of a non - perturbative solution , and also since explicitly constructed , field theoretical domain wall solutions @xcite clearly exist in the case of the 2-brane , we suspected that the lack of this solution must lie somewhere in the validity of the applied method . nevertheless , we were so enthusiastic with providing the whole family of the missing solutions in the @xmath51 class at the axis , that it did nt occur in our mind at the time , that the 2-dimensional case might be special in the topologically minkowskian family as being the only one which is non - analytic at the axis . the main result of the present paper is the observation , that the 2-dimensional solution of the thick - bbh system is in fact a special one in the topologically minkowskian family as being @xmath49 ( or as we will soon see maximum @xmath52 ) function at the axis of the system .
all the other dimensional solutions are @xmath51 .
this also explains the fact why it was only the 2-dimensional case where a perturbative solution could exist around the thin solution . in the remaining of this section
we analyze the asymptotic behavior of ( [ fulleq ] ) near the axis of the system .
we obtain necessary boundary conditions from regularity requirements and show that these conditions can always be fulfilled in order to obtain a regular solution . we know from the above considerations that a regular solution of the problem must exist although analytic solution could not be found at the axis of the system
. thus the point @xmath53 on the axis , where @xmath54 , must be a _ regular singular point _ of the differential equation ( [ fulleq ] ) . even though ( [ fulleq ] ) is highly nonlinear , general results from the theory of _ local _ analysis of linear differential equations can be applied , because we know from physical considerations that ( [ fulleq ] ) should not develop any nontrivial singular points in its domain .
hence , if a solution is not analytic at a regular singular point ( see e.g. @xcite ) , its singularity must be either a pole or an algebraic or logarithmic branch point , and there is always at least one solution of the form @xmath55 where @xmath56 is a number called _ indical exponent _ and @xmath57 is a function which is analytic at @xmath53 and has a convergent taylor series . in the general case
@xmath56 can be any number that solves ( [ fulleq ] ) . in our specific case however , the thin solution , around which a regular perturbative solution existed , had the asymptotic form near @xmath53 ( see @xcite ) @xmath58 and similarly , the corresponding perturbative solution ( see @xcite ) near the same point had the asymptotic form @xmath59 where @xmath60 is the perturbation function and @xmath61 and @xmath62 are coefficient functions defined in @xcite .
thus , in order to obtain the non - perturbative thick solution , we also chose @xmath56 to be @xmath63 , in accordance with the perturbative results .
this choice will also have the advantage of naturally fixing the free boundary condition in the next section for a unique numerical solution . with @xmath64 , from ( [ alfaeq ] )
we obtain that the asymptotic form of @xmath65 near the axis is @xmath66 plugging this expression and its derivatives into ( [ fulleq ] ) , one can obtain the following asymptotic behavior @xmath67 where the coefficient functions @xmath68 are polynomial expressions of @xmath69 with dependences @xmath70 in order to obtain a regular solution , we need to require for the @xmath68 coefficient functions to disappear at @xmath53 . from the explicit forms of @xmath68
, one can find that the coefficient @xmath71 can be chosen freely , and once it s fixed , the remaining coefficients @xmath72 , @xmath73 and @xmath74 can be computed from the requirements that @xmath75 . consequently , the solution is uniquely determined by the parameter @xmath53 ( i.e. the minimal distance parameter ) , and the explicit values of the coefficients @xmath72 , @xmath73 and @xmath74 can be immediately obtained from the @xmath75 equations .
( this procedure is of course necessary before the numerical setup . ) for the question of existence , the @xmath75 equations are always soluble , as it turns out that these are linear equations for the @xmath69 coefficients
. the explicit forms of the coefficients @xmath72 , @xmath73 and @xmath74 , as successive functions of @xmath71 and @xmath53 are given in the appendix . in conclusion we found that a regular 2-dimensional solution can always be given for the exact problem , and a unique solution
is completely determined by the regularity requirements once the coefficient @xmath71 is fixed .
since we are free to choose @xmath71 , it can also put to be @xmath76 for example . in this case the asymptotic form ( [ asymp ] ) starts with @xmath77 which solution is a @xmath52 function , however in all other cases the solution of ( [ fulleq ] ) is @xmath49 at @xmath53 .
for illustrating the obtained results , we provide the numerical solution of the 2-dimensional flat problem . with the experiences we gained from the analysis of the thick - bbh system in @xcite ,
it is not too difficult to obtain the numerical solution here , after the initial conditions for this specific case has been clarified .
as we discussed earlier , we are free to choose the coefficient @xmath71 in the asymptotic solution ( [ asymp ] ) . nevertheless , in order to be completely consistent with our previous perturbative results in the flat 2-brane case , we make the choice @xmath78 that was forced upon us by regularity requirements for the perturbations .
having fixed this freedom , the remaining three conditions @xmath72 , @xmath73 and @xmath74 are determined , as discussed in the previous section , and the corresponding numerical solution is unique . in obtaining the numerical solution we used mathematica@xmath79 `
ndsolve ` function .
the integration range went from @xmath53 until @xmath80 to check the accordance with the corresponding perturbative solution .
the configurations of the perturbative thick 2-brane solutions in the near horizon region have been presented in @xcite . since the effects of the nonlinearities are really small even in the gravitationally strong , near horizon region , the overall global picture of the thickness corrected brane configurations remain very similar to the perturbative case .
to make however the small effects visible , we plot on fig . 1 and fig . 2 the difference function @xmath81 of the present thick- and the original thin solutions in 4 and 5 dimensions .
@xmath82 is the exact analog of the perturbation function @xmath83 defined in ( [ fi ] ) .
curves in 4 dimensions with minimum horizon distance range @xmath84 . ] curves as on fig.1 in 5 dimensions . ] comparing the results with the corresponding plots of the perturbative solutions in @xcite , we find that the general behavior of the @xmath82 curves are essentially the same .
this is of course what one expects . on the other hand one
also expects some differences coming from the nonlinear regime of ( [ fulleq ] ) , and those are also present if we enlarge the very near horizon region of the individual curves . the main features of these differences are plotted on fig . 3 and fig . 4 .
curve in 4 dimensions with minimum horizon distance @xmath85 . ] curve in 4 dimensions with minimum horizon distance @xmath86 . ] on fig .
3 an interesting nonlinear effect appears compared to the perturbative solution . instead of monotonically decreasing down to its global minimum ,
as @xmath60 does in the perturbative case , the @xmath82 curve has an extra local pattern near the minimum horizon distance @xmath53 .
the extra pattern is that the curve goes through some additional local extrema in this region before it finally tends to reach the global minimum which is similar to the one in the perturbative case . during these local differences
however , @xmath82 remains negative in this whole region . on the qualitatively similar curve on fig . 4 , the essential difference compared to fig .
3 is that @xmath82 develops a sign change in this region , that is there are several intersections in the very near horizon region between the thick and thin solutions compared to the single intersection that is present in the case of the perturbative solution ( see fig . 6 in @xcite ) . with increasing horizon distance @xmath23 , the above nonlinearities decay quickly , and the solution agrees very well with the perturbative results .
in the present work we studied the problem of a topologically flat 2-brane in a higher dimensional , thickness corrected bbh system . we provided a regular , non - perturbative , numerical solution for this special case based on earlier perturbative considerations @xcite .
the main result of this paper is the observation that the 2-dimensional case of the problem is a special one as being non - analytic at the axis of the system .
this property makes it unique in the family of thick solutions , as in all other dimensions both the minkowski and black hole embedding solutions are analytic in their entire domain .
we analyzed the asymptotic behavior of the solution and obtained that it is at most a @xmath52 , but in the general case it is only a @xmath49 function at the axis , just like the corresponding thin solutions .
the initial conditions of the problem are not uniquely fixed by regularity requirements , thus the solution we provided here is not unique .
it is however perfectly consistent with our earlier perturbative results . with the present paper ,
we have provided the complete set of solutions of the thick - bbh problem in a series of three consecutive papers .
first , in @xcite , we obtained all possible perturbative solutions and later , in @xcite all non - perturbative solutions were given except the case of the flat 2-brane .
present work completes the set .
in @xcite we analyzed the properties of a topology changing , quasi - static phase transition in the thick - bbh system .
the obtained results in the present case however does not change our previous findings and thus we do nt consider the phase transition in this paper . the result , that thickness corrections change the analytic properties of the brane solutions at the axis of the system might have some physical consequences .
infinitely thin brane solutions , naturally , are very important in higher dimensional physics , but considering the present problem , one has the intuition that the thickness corrections , which are in agreement with field theoretic domain wall models , made the thin - bbh system more stable in the sense that regular , analytic solutions could be provided in essentially all cases . since small physical perturbations to any system are usually proportional to the derivatives of the unperturbed solution , it is very possible that the thin - bbh solutions are not entirely stable against small perturbations in the minkowski branch .
( stability properties of the bbh system for the analytic black hole embedding solutions has been studied for example in @xcite . )
this property however has been cured by the thickness corrections and it is somehow in accordance with our physical expectations .
the special case of the 2-brane as remaining non - analytic after thickness corrections , thus , is an unexpected property , which makes it physically interesting .
so much the more that thick 2-branes , i.e. thin walls on black hole backgrounds in standard 4-dimensional general relativity are certainly real , physical objects .
the fact that these solutions are essentially different from the corresponding ones in higher dimensions is remarkable .
most part of the calculations were performed and checked using the computer algebra program mathematica 7 .
the research was supported by the national research foundation of south africa and the hungarian national research fund , otka no .
k67790 grant .
@xmath87,\nonumber\end{aligned}\ ] ] @xmath88\left.\right|_{r_1}\ , \nonumber\\\nonumber\\\nonumber\\ a_2&=&\frac{1}{26880a_1(a+b){\varepsilon}fr^2}\left[\right .
( -14592 ( a + b ) { \varepsilon}+ a_1 ^ 2 r ( 3 { \varepsilon}\dot f r ( -16 ( 29 a + 35 b ) + 8 a_1 ^ 2 ( 5 a + b ) \dot f r^2\\ & - & a_1 ^ 4 ( a + b ) \dot f^2 r^4 ) 24 a_1 ^ 2 f^2 r ( 2 a_1 ^ 2 r^3 + 2 ( 2 a + b ) { \varepsilon}(12 + a_1 ^ 2 r ) + a_1 ^ 2 { \varepsilon}r^2 ( ( 8 a + 3 b ) \dot f + 2 a \ddot f r ) ) \nonumber\\ & + & 4 f (-8 ( 6 ( 31 a + 28 b ) { \varepsilon}+ a_1 ^ 2 r ( 19 a { \varepsilon}+ 7 b { \varepsilon}+ 3 r^2 ) ) + 3 a_1 ^ 2 r^2 ( { \varepsilon}\ddot f r ( -32 a + a_1 ^ 2 ( a + b ) \dot f
r^2 ) \nonumber\\ & + & \dot f ( -48 ( 2 a + b ) { \varepsilon}+ a_1 ^ 2 ( 3 a + b ) { \varepsilon}\dot f r^2 + a_1 ^ 2 r ( 2 a { \varepsilon}+ r^2 ) ) ) ) ) ) \left.\right]\left.\right|_{r_1}\ , \nonumber\end{aligned}\ ] ] @xmath89\left.\right|_{r_1}\ .
\nonumber\end{aligned}\ ] ] a. flachi and t. tanaka , * 95 * , 161302 ( 2005 ) . v. p. frolov , * 74 * , 044006 ( 2006 ) .
d. mateos , r. c. myers and r. m. thomson , * 97 * , 091601 ( 2006 ) .
f. mena , j. natario and p. tod , _ formation of higher dimensional topological black holes _ , annales henri poincar 10 ( 2010 ) , 1359 - 1376 ; b. kol , j. high energy phys .
* 10 * ( 2005 ) 049 .
b. kol , phys .
* 422 * , 119 ( 2006 ) .
d. mateos , r. c. myers and r. m. thomson , j. high energy phys .
05 ( 2007 ) 067 .
j. maldacena , adv .
* 2 * , 231 - 252 ( 1998 ) .
p. a. m. dirac , proc .
a. * 268 * , 57 ( 1962 ) .
j. nambu , copenhagen summer symposium ( 1970 ) , + unpublished . t. goto , prog .
* 46 * , 1560 ( 1971 ) .
v. p. frolov and d. gorbonos , * 79 * , 024006 ( 2009 ) . v. g. czinner and a. flachi , * 80 * , 104017 ( 2009 ) . v. g. czinner , * 82 * , 024035 ( 2010 ) .
f. r. thangerlini , nuovo cimento * 77 * , 636 ( 1963 ) .
b. carter and r. gregory , * 51 * , 5839 ( 1995 ) .
e. t. whittaker , _ analytical dynamics _ , cambridge university press , london , 1937 , 4th ed . , p. 266 .
a. flachi , o. pujols , m. sasaki and t. tanaka , * 73 * , 125017 ( 2006 ) .
a. flachi and t. tanaka , * 76 * , 025007 ( 2007 ) .
c. m. bender and s. a. orszag , _ advanced mathematical methods for scientists and engineers _ , mcgraw - hill , inc .
new york , 1978 , p. 63 .
k. hioki , u. miyamoto and m. nozawa , * 80 * , 084011 ( 2009 ) . | we present a numerical solution for a topologically flat 2-dimensional thick brane on a higher dimensional , spherically symmetric black hole background .
present solution is the last , missing part of the complete set of solutions for the thickness corrected brane - black hole problem in arbitrary number of dimensions .
we show that the 2-dimensional case is special compared to all the higher dimensional solutions in the topologically minkowskian family as being non - analytic at the axis of the system .
we provide the numerical solution in the near horizon region and make a comparison with the infinitely thin case . |
x - ray observations have shown that a very hot ( @xmath10 k ) diffuse intracluster medium pervades the cores of clusters of galaxies @xcite
. without continued heating , this gas will cool with a characteristic cooling time of @xmath11 , where @xmath12 is the gas temperature in units of @xmath13 k , @xmath14 is the gas density in units of @xmath15 @xmath16 , @xmath17 is the hubble constant in units of 50 km s@xmath4mpc@xmath4 and @xmath18 is the hubble time @xcite .
consequently , if gas density is high enough and the temperature low enough , the characteristic cooling time will be less than a hubble time @xcite .
this cooling gas should `` flow ''
quasi - hydrostatically onto the cd / d galaxy at the bottom of the cluster potential well .
typical mass flows inferred from the x - ray data are @xmath19 m@xmath20 yr@xmath4 within the central @xmath21 kpc of the cluster .
as the gas cools , it will emit strongly in various x - ray lines of highly ionized fe and o for @xmath22 k @xcite .
emission lines of oviii ly@xmath23 and fexxiv arising in this hot gas have been detected in the cores of clusters @xcite .
spatially resolved x - ray spectroscopy has confirmed the presence of distributed cooling gas in clusters @xcite .
as the gas cools rapidly through the @xmath24 k regime , it should emit strongly in certain uv resonance lines particularly the ovi 1032,1038 doublet @xcite . finally ,
as the gas cools to @xmath25 k , it should emit strongly in the hydrogen lines ( h@xmath23 , ly@xmath23 ) . optical h@xmath23 emission from cooling flows has been detected toward several clusters @xcite , @xcite has also reported extended ly@xmath23 emission from a number of clusters having h@xmath23 emission .
emission lines from cooling flow gas at temperatures @xmath26 k have yet to be convincingly detected .
@xcite observed several clusters , including a1795 , with the fuv spectrograph on the _ hopkins ultraviolet telescope _ ( hut ) in an attempt to detect emission from ovi .
no detections were made , although in the case of the strongest x - ray cooling flow cluster observed , a1795 , the hut exposure time was only 2000 sec . in efforts to detect gas at temperatures of @xmath27k
, a number of investigators have attempted to detect the fex @xmath16374 coronal line in cooling flow clusters @xcite .
@xcite provided the the most sensitive search reported to date , and found no evidence for the line and set an upper limit for the surface brightness of @xmath28 erg @xmath3 s@xmath4 arcsec@xmath29 .
emission lines from fexvii arising in gas at @xmath30k have also not been detected in a recent _ xmm _ observation of a1795 @xcite .
detection of emission from gas at @xmath31 k would provide the lynchpin to physically connect the previously detected hot and cool gas components in the cores of clusters .
we have therefore begun a program with the _ far ultraviolet spectroscopic explorer _
( fuse ) to search for emission from ovi 1032,1038 in the cores of the x - ray bright , cooling flow clusters a1795 , a2199 and a2597 .
results for a1795 and a2597 are presented in this paper .
fuse has twice the effective area of hut at ovi @xmath11032 , which , when coupled with longer integration times and higher spectral resolution can provide a considerable increase in sensitivity over the hut .
for example , the fuse observations reported here for a1795 are a factor of 15 more sensitive than the hut observations . throughout this paper
, we have used a value of @xmath32 km s@xmath4mpc@xmath4 .
all distances scale as the inverse of @xmath33 , while observed line luminosities and derived cooling flow mass deposition rates scale as @xmath34 .
far - ultraviolet spectra of the cores of a1795 and a2597 have been obtained with the fuse observatory .
a description of the fuse instrument and its in - orbit performance is described in papers by @xcite and @xcite . in brief ,
fuse spectra cover the wavelength range @xmath35 with a spectral resolution of @xmath36 km s@xmath4 ( depending on wavelength ) for point sources .
fuse spectra are recorded in four separate optical channels , two of which employ lif optics covering the wavelength range @xmath37and two which use sic optics covering the @xmath38 range .
dispersed light is focused onto two photon - counting microchannel plate detectors .
all observations were carried out with the lwrs spectrograph entrance aperture ( @xmath39 ) .
for extended emission which fills the lwrs aperture , the instrumental line width is @xmath21 km s@xmath4 , which is narrower than the expected line width of cooling flow gas in the core of a cluster of galaxies .
the data were processed with the fuse calibration software ( _ calfuse _ , version 1.8.7 ) .
since the signal from the source is expected to be weak , extra care was taken in subtraction of the detector background .
photons events with pulse heights less than 4 or greater than 15 were screened out of the data , since these events are predominately background in nature .
furthermore , the standard background subtraction employed by _ calfuse _ , which is adequate for bright objects , was not used . instead
, the background level on each detector was determined from a co - addition of all exposures for each cluster .
the local background in the vicinity of the expected line positions was measured on the 2d spectral image at locations just above and below the spectra , and were then subtracted from the extracted spectra .
we have searched the extracted spectra for emission from ovi @xmath11032 and ciii @xmath1977 .
ovi @xmath11032 is redshifted to 1097 and 1117 in a1795 and a2597 respectively .
these wavelengths fall in the lif channels for both detectors 1 and 2 , and for the sic channel in detector 2 ( for a1795 only ) , and is outside the wavelength range of the sic channel in detector 1 .
the combined effective area of the 2 lif channels at @xmath40 , based on in - flight calibrations , is about 52 @xmath3 .
the sic2 channel at this wavelength has an effective area of only @xmath41 @xmath3 , and hence we have not used the sic2 spectrum in our analysis . the ciii @xmath1977 line is also redshifted to a wavelength covered by both lif and sic channels .
however , the combined effective areas of the lif channels is much greater than the sic channels , and given the fact that the signal is close to the detector background , we have used only the lif spectra from channels 1 and 2 .
terrestrial airglow lines in the far - uv @xcite have also posed a special problem in our search for emission lines from cooling flows .
for example , in a1795 the ciii @xmath1977 line is redshifted to 1038.7 , which is adjacent to an oi airglow line . in a2597 ,
the ciii line is redshifted to 1057.4 , which is potentially contaminated by a terrestrial n@xmath42 emission band .
finally , in a1795 , the ovi @xmath11032 line is redshifted to coincide with a band of niairglow lines .
we have thoroughly investigated all these possible contaminations and discuss the results below .
a1795 was observed on february 23 , 2000 with an integration time of 37,189 seconds .
the cluster has a redshift of 0.0631 @xcite , hence the fuse aperture subtends @xmath43 kpc at the cluster .
the spectrograph aperture was centered on the cd galaxy at coordinates : @xmath44 .
any ovi @xmath11032 emission from a1795 would be redshifted to an observed wavelength of 1097 , which is unfortunately at nearly the same wavelength as a complex of ni terrestrial airglow lines @xcite .
these lines are strong only when looking at the illuminated earth .
however , they are also present in spectra taken near the earth limb , and are faintly visible throughout the daylit portion of the orbit . to avoid the effects of these airglow lines as much as possible , we have restricted our analysis of the ovi line to data obtained during spacecraft orbital night , limiting the integration time to 17 ksec .
the lif1 and lif2 spectra of a1795 were coadded and the resulting spectrum at the expected wavelengths of the redshifted ovi lines is shown in figure 1 , where the data has been binned by 0.2 ( @xmath45 km s@xmath4 ) . a faint continuum from the cd elliptical galaxy
is detected at a level of @xmath46 erg @xmath3 s@xmath4 @xmath4 , in good agreement with the uv continuum flux reported by @xcite at wavelengths @xmath47 from an iue large aperture observation .
this relatively large far - uv flux from an old elliptical galaxy at a redshift of 0.0631 indicates the presence of substantial star formation activity .
@xcite has estimated an ongoing star formation rate of @xmath48 m@xmath20 yr@xmath4 from the iue data .
@xcite and @xcite have reported broad - band uv emission from the center of a1795 , and concluded that a population of early type stars is present , which presumably formed from the cooling flow . in figure 1 , we note possible galactic absorption lines of h@xmath42 at 1094 and 1100 , which makes the spectrum appear somewhat noisier than it truly is .
these absorption features do not correspond to any metal absorption lines at the redshift of a1795 . in any event
, there is no obvious oviemission .
after subtracting a continuum fit to the galaxy spectrum , we compute a flux in the @xmath11032 line of @xmath49 erg @xmath3 s@xmath4 , where we have integrated from @xmath50 to @xmath51 km s@xmath4 about the expected line center .
the uncertainty quoted here is purely a random error .
there is also a systematic uncertainty in the continuum placement , so we estimate a 2@xmath52 upper limit to the flux in @xmath11032 line of @xmath8 erg @xmath3 s@xmath4 , or in terms of the average surface brightness within the fuse lwrs aperture , @xmath53(ovi)@xmath54 erg @xmath3 s@xmath4 arcsec@xmath29 or equivalently , @xmath55 erg @xmath3 s@xmath4 sr@xmath4 .
this then yields an upper limit to the luminosity in the ovi @xmath11032 line of @xmath56ovi@xmath57 erg s@xmath4 .
we have have also searched the a1795 spectrum for ciii @xmath1977 emission using the full 37 ksec integration taken during both orbital night and day . in figure 2 , we display the @xmath58 spectral region from that dataset
. the background subtraction appears to be quite good , since the cii @xmath11036 absorption line from the galactic interstellar medium is black at line center , as expected .
a faint emission line appears to be present adjacent to the oi @xmath11039.2 terrestrial airglow line , and we suggest that this is emission from ciii @xmath1977 in the cooling flow of a1795 , redshifted to 1038.7 .
the line flux is uncertain due to the airglow contamination , but we estimate a flux of @xmath59 erg @xmath3 s@xmath4 ; ie . only a @xmath60 detection .
we also note broad absorption extending from @xmath61 to 1037 in the observed frame .
absorption at the two ends of this range are due to ovi @xmath11032 and cii@xmath11036 from the galactic ism .
the absorption at 1034 corresponds in wavelength to redshifted ly@xmath62 in a1795 .
absorption by neutral hydrogen in a1795 could , in principle , be confirmed by detecting ly@xmath63 absorption .
however , ly@xmath63 is redshifted to 1090 , which is very near the edge of segment b of detector 1 , where calibration is problematical .
this wavelength is not recorded on detector 2 , since it lies in the gap between segments a and b. a2597 was observed on june 26 , 2000 for a total integration time of 43,032 seconds .
the spectrograph aperture was centered on the cd galaxy at coordinates : @xmath64 .
the redshift of the a2597 cd galaxy is 0.0823 @xcite , thereby shifting the ovi @xmath11032 line to to 1116.86 in the observed frame .
this is in a region of the spectrum where there is no observed or expected airglow emission @xcite .
we have therefore used spectra obtained during both orbital day and night in our analysis of a2597 .
the co - added spectrum from the lif1 and lif2 channels is shown in figure 3 , binned to 0.2 . a weak , but clearly detected emission line is evident at the expected position of redshifted ovi @xmath11032 .
the emission line is clearly visible in both the individual lif1 and lif2 spectra , so it is very unlikely to be due to a detector artifact .
indeed , no such emission feature is observed at 1117 in the combined day and night ( 34 ksec ) observation of a1795 .
we therefore conclude that this weak emission line is redshifted ovi @xmath11032 from the cooling flow nebula of a2597 .
we measure a flux in the @xmath11032 line of @xmath65 erg @xmath3 s@xmath4 , where we have integrated from @xmath50 to @xmath51 km s@xmath4 about the expected line center .
the fuse aperture subtends @xmath66 kpc at the cluster , resulting in an average surface brightness within the fuse lwrs aperture of @xmath53(ovi)@xmath67 erg @xmath3 s@xmath4 arcsec@xmath29 or equivalently , @xmath68 erg @xmath3 s@xmath4 sr@xmath4 .
the luminosity in the ovi @xmath11032 line is then @xmath56ovi@xmath69 erg s@xmath4 .
an examination of the spectrum has also revealed an emission line at 1057.4 , the expected position of redshifted ciii @xmath1977 ( see figure 4 ) . @xcite
have observed emission from terrestrial n@xmath42 at this wavelength when pointed at the bright earth . in order to determine
if the observed line in our data is terrestrial in origin , we have also extracted spectra of a2597 obtained during orbital night .
although the night - only data was restricted to 14 ksec , the emission line at 1057 is detected quite easily .
furthermore , the flux in the line is the same during the night and daylit portions of the orbit .
finally , we note that there is no emission line at this wavelength present in the a1795 spectra .
we conclude that the evidence is strong that the emission line at 1057 is not of terrestrial origin , and is in fact redshifted ciii @xmath1977 from a2597 .
the measured line flux is @xmath70 erg @xmath3 s@xmath4 measured within @xmath71 km s@xmath4of the expected line center .
the flux in the ovi line is a function of the temperature of the gas , its distribution in the cluster , and the mass flow rate in the cooling flow . @xcite and @xcite have computed the expected luminosities of ultraviolet emission lines from a gas cooling from a temperature @xmath72k to @xmath6k .
the luminosity in the ovi line is given by @xmath73 where @xmath74 is the mass flow rate , @xmath75 is the mean mass per atom , @xmath76 is the abundance of oxygen , @xmath77 is the fraction of oxygen in the o@xmath78 ionization stage , @xmath79 is the number of particles per h atom , @xmath80ovi ) is the volume emissivity in the ovi line , and @xmath81 is the radiative cooling function @xcite .
the gas cools from t@xmath82 down to t@xmath83 with an evolution that could be isobaric , isochoric or some combination of the two ; @xmath84 is 1 for isobaric cooling and 0 for isochoric . the intracluster medium ( icm )
has abundances that are typically @xmath85 times solar @xcite which can be understood in terms of the icm being composed of galactic ejecta mixed with primordial gas .
note in the above equation that lower metallicity ( of oxygen in the present case ) is offset by a similar decrease in the cooling function , assuming that the abundance of oxygen is not dissimilar to other metals providing the bulk of the cooling .
hence , the line luminosity is rather insensitive to metallicity , as first pointed out by citetcowie81 .
we also note that the assumption made here is that this is the luminosity of an optically thin , extended region , with no self - absorption . @xcite
have calculated uv line luminosities for solar abundances , with t@xmath86 k and t@xmath87 k. @xcite arrive at very similar results . adopting an intermediate case between isobaric and isochoric cooling ,
the predicted luminosity in the ovi @xmath11032 line is @xmath88 erg s@xmath4 , where @xmath89 is the mass cooling flow rate in solar masses per year . for an optically thin plasma , the luminosity of the 1038 line is half of that of the 1032 line .
based on model prediction above , our upper limit to the observed luminosity of a1795 in the ovi @xmath11032 line implies an upper limit to the mass cooling flow rate of only @xmath90 m@xmath20 yr@xmath4 .
@xcite determined the mass flow rates in a large sample of clusters observed with the _ einstein observatory _ and exosat , using a deprojection analysis of the x - ray images .
for a1795 , they found a mass deposition rate of @xmath91 yr@xmath4 within a radius of @xmath92 kpc .
@xcite computed mass flow rates for a sample of 30 of the most luminous x - ray clusters , by applying a similar deprojection technique to rosat hri images , and analyzing spatially resolved spectra of the clusters obtained with asca . within a cooling radius of @xmath93 kpc , @xcite finds a spectroscopic mass flow rate of @xmath94 yr@xmath4 .
however , deprojection of the rosat images yielded @xmath95 yr@xmath4 . a deprojection analysis including intrinsic absorption using the same column densities determined from the spectral analysis of the asca data yielded a corrected value of @xmath96 yr@xmath4 . in the generally adopted cooling flow picture , gas begins cooling from a temperature of @xmath97k at several hundred kpc from the cluster center , and cools as it flows inwards with @xmath98 , where @xmath99 is the integrated mass deposition rate interior to @xmath100 .
for a1795 , the fuse aperture has an effective radius of @xmath101 kpc , or @xmath102 as determined by @xcite . assuming a mass deposition rate proportional to radius , then the x - ray mass cooling rate within the fuse aperture is @xmath103 yr@xmath4 depending on whether one adopts @xmath104 , @xmath105 or @xmath106 from @xcite .
our upper limit to the mass cooling flow rate of @xmath107 yr@xmath4 from the non - detection of ovi @xmath11032 emission is about half of that derived from the asca x - ray spectra , assuming no absorption of the ovi line .
the observed ovi @xmath11032 luminosity in a2597 implies a cooling flow rate of @xmath108 m@xmath20 yr@xmath4 , based on the model described above .
@xcite has derived a cooling rate for the ambient hot gas of @xmath109 yr@xmath4 within a radius of @xmath110 kpc from rosat hri x - ray images .
the effective radius of the fuse aperture at the redshift of a2597 corresponds to @xmath111 kpc , or @xmath112 . scaling the mass deposition rate with radius
, we then find an x - ray mass cooling rate within the fuse aperture of @xmath113 yr@xmath4 , which is almost 3 times the cooling rate of @xmath114 yr@xmath4 inferred from the the luminosity of the oviline .
one possible way to diminish the observed ovi emission is through scattering by dust . @xcite
have reported the existence of significant amounts of dust in the cores of some cooling flow clusters , based on submillimeter observations .
@xcite has investigated the ly@xmath23/h@xmath23 emission line ratios in cooling flow clusters , and found an average excess reddening , above the foreground extinction , of @xmath115 in the range @xmath116 .
@xcite discusses the location of the dust in the cluster , and favors the idea that the bulk of the dust is in the extended cluster component , instead of embedded in the localized cooling flow region .
an extended dust component was found to be in rough agreement with estimates of cluster extinction based on background galaxy counts @xcite , and is supported by the lack of a correlation between @xmath115 with @xmath74 in cooling flow systems .
if this is the case , then regardless of the geometry of the cooling flow region , ovi emission from the core of the cluster will have to pass through the dust in the outer cluster on its way to the observer .
@xcite derived a value of @xmath117 for a1795 ( although take note of the large uncertainties , which are primarily systematic ) . assuming an upper limit to @xmath115 of 0.22
, this translates into @xmath118 magnitudes of extinction in the far - ultraviolet , computed assuming a standard galactic dust extinction curve with @xmath119 .
if we assume that all of this extinction arises in the outer cluster and scatters the ovi emission , then our observed , extinction - corrected line luminosity is @xmath120ovi@xmath121 erg s@xmath4 .
this then raises the limit to the mass cooling rate to @xmath122 yr@xmath4 , which does not constrain the x - ray models . for a2597 ,
@xcite quotes @xmath123 which would result in an extinction corrected mass flow rate derived from the ovi observation of @xmath124 yr@xmath4 , which is in reasonable agreement with the x - ray derived mass flow rate within the fuse aperture , given the uncertainties in the actual extinction value .
@xcite have pointed out that dust in the icm may be unlike galactic dust , and may have different extinction properties .
indeed , it is clear that extinction properties are dependent upon the past processing history and environment in which the dust grains reside @xcite .
however , extinction in the far - uv is dominated by small grains , and these would be subject to destruction in the hot icm by sputtering . in hot gas with density @xmath125 @xmath16 , a grain of size @xmath126 microns will be destroyed in @xmath127 yr @xcite .
so , grains smaller than @xmath128 m in size would be destroyed in less than a hubble time unless the grains were near the outskirts of the cluster where @xmath129 @xmath16 . if intracluster dust at large cluster radii existed and had an extinction law like that of the small magellanic cloud ( ie . containing small grains ) , then the icm extinction could be @xmath130 mag greater at ovi than it would if an average galactic law were assumed
. hot gas near the cluster center should be free of dust grains with sizes @xmath131 m due to sputtering .
so , any clouds that condense out of the cooling flow should also be dust - free
. however , the gas surrounding the central galaxies in many cooling flow clusters appears to be dusty @xcite .
@xcite have reported significant depletion of calcium in cooling flow emission line regions , and this is generally attributed to dust depletion . in the core of a1795 ,
@xcite have resolved a 7 kpc long dust lane , whose morphology matches that of the h@xmath23 emission . @xcite have also resolved an elongated dust lane southeast of the nucleus of a2597 , together with patches of dust near the cluster core .
this dusty gas may have been stripped from a gas - rich galaxy passing through the cluster core , or it might be the remains of a merger of such a galaxy with the central cd galaxy . consequently , these dusty filaments may be very dynamic , and may form and be destroyed continuously throughout the life of the cluster . the effect of dust in the core of the cluster on ovi emission would likely depend on its geometry and/or filling factor .
if the dust were confined to optical emission line filaments having a small filling factor , then the the observed ly@xmath23/h@xmath23 emission line ratios could be altered by this dust , and yet the ovi emission , which arises in the intercloud regions ( or at cloud interfaces ) , could suffer much less extinction than the optical emission lines . the existence of magnetic fields in intracluster gas with strengths of order 1 @xmath53 g have been inferred from observations of faraday rotation of radio sources seen through the icm @xcite .
tangled magnetic fields , or field lines that run perpendicular to the temperature gradient , can prevent thermal conduction in an ionized gas .
recent x - ray observations of two clusters , a2142 @xcite and a3667 @xcite , by _
chandra _ have revealed what appears to be central , cooler intracluster gas separated from the outer , hot gas by a very sharp density discontinuity . @xcite
have pointed out that this sharp temperature discontinuity could not be maintained without a significant suppression of heat conduction .
recent _ chandra _ x - ray images of the core of a1795 reveal an 80 kpc long filament of x - ray emission extending to the sse of the cd galaxy @xcite .
there is a remarkable morphological similarity between this x - ray emission and the h@xmath23 emission .
although the nature of the filament is unclear ( ie . a cooling wake , or ram pressure stripped gas from a moving galaxy , etc ;
see @xcite ) there appears to be a direct physical relationship between the x - ray cooling gas and the optical emission .
although not all of this filamentary structure is within the fuse aperture , the brightest parts of the filament near the cd galaxy are within the aperture ( see figure 5 ) .
if the cool clouds in the filament are small enough , they will be destroyed by thermal evaporation in the hot , ambient gas , unless conduction is quenched by tangled magnetic fields .
the existence of the filament may then indicate that conduction is ineffective in cooling the hot x - ray emitting gas .
the absence of ovi emission in a1795 , the rather weak ovi emission in a2597 , and the absence of observed fexvii lines in cooling flows @xcite could then be explained by the suppression of conduction .
nevertheless , we do observe ovi emission in a2597 at a level consistent with expectations from x - ray observations if we correct the ovi flux for a reasonable amount of scattering by dust .
we are then left to ponder why no x - ray emission has been detected by _
xmm _ from gas with temperatures less than @xmath132k . a possible explanation is provided by turbulent mixing of hot and cool gas @xcite .
@xcite have pointed out that the inner regions of cooling flows are likely to be highly turbulent , with turbulent velocities of order @xmath21 km s@xmath4 . assuming the presence of entrained cool clouds , any relative motion between the hot gas and the cool clouds will excite kelvin - helmholtz instabilities that will grow and create a mixing layer at the interface .
the surface area of the cool material is increased dramatically by this turbulence , and allows efficient diffusion of energy between the hot and cool phases @xcite , resulting in a very short mixing time .
the steady - state temperature of the mixed gas is intermediate between the hot and cool gas with a value of @xmath133 @xcite , where @xmath134 and @xmath135 are , respectively , the cool and hot gas temperatures and @xmath136 is a factor of order unity , and depends on the ratio of the efficiencies with which hot and cool gas deposit mass and energy into the mixing layer . for the case
we envision in cluster cooling flows , if cool gas at temperature @xmath137k and hot gas at temperature @xmath138k are mixed , then the intermediate gas temperature in the mixing layer will be @xmath139k .
this gas is at the peak of the cooling curve , and would rapidly cool and emit ovi @xmath11032 , while being too cool to emit the fexvii x - ray lines .
we have obtained far ultraviolet spectroscopy of the cores of the galaxy clusters a1795 and a2597 , in order to detect emission lines arising in cooling flow gas with temperatures of @xmath140 k. we do not detect ovi @xmath11032 in a1795 , with a @xmath141 upper limit of @xmath142 erg @xmath3 s@xmath4 .
this corresponds to an upper limit to the mass deposition rate of @xmath143 yr@xmath4 within the central 28 kpc , assuming cooling models of @xcite and @xcite .
we report a marginal ( @xmath60 ) detection of emission from ciii @xmath1977 in a1795 .
we have detected ovi @xmath11032 and ciii @xmath1977 emission in a2597 providing direct evidence linking the cool gas at @xmath6k with the ambient hot , @xmath5k gas .
in collisional ionization equilibrium , o@xmath78 and c@xmath144 are formed at temperatures of @xmath145 k and @xmath146 k , respectively .
the observed ovi @xmath11032 line flux in a2597 is @xmath147 erg @xmath3 s@xmath4 and corresponds to mass deposition rate of @xmath148 yr@xmath4 .
the non - detection of ovi @xmath11032 in a1795 could be due to scattering by dust , or to quenching of thermal conduction by intracluster magnetic fields , which prevents the gas from cooling below @xmath149k .
a turbulent mixing model , together with some dust extinction , appears to adequately describe our detection of ovi @xmath11032 emission in a2597 , while also accounting for recent _ xmm _ spectrosocopic results , which indicate a lack of fexvii emission , and hence a lack of cooling flow gas at temperatures of @xmath150k .
this work is based on data obtained for the guaranteed time team by the nasa - cnes - csa fuse mission operated by the johns hopkins university .
financial support to u. s. participants has been provided by nasa contract nas5 - 32985 .
we thank joel bregman and rich mushotzky for informative discussions on cooling flows and the recent _ xmm _ results .
allen , s. 2000 , , 315 , 269 allen , s. & fabian , a. 1997 , , 286 , 583 allen , s. & fabian , a. 1998 , , 297 , l63 anton , k. , wagner , s. , & appenzeller , i. 1991 , , l51 begelman , m. & fabian , a. 1990 , , 244 , 26p canizares , c. , markert & donahue , m. 1988 , in cooling flows in clusters of galaxies , proceedings of the nato advanced research workshop , cambridge , england , june 22 - 26 , 1987 , ( dordrecht : kluwer academic publishers ) , 63 cowie , l. l. 1981 , in x - ray astronomy with the einstein satellite : proceedings of the meeting , cambridge ma , january 28 - 30 , 1980 , ( dordrecht : reidel publishing co. ) , 227 cowie , l. & binney , j. 1977 , , 215 , 723 cowie , l. , hu , e. , jenkins , e. , & york , d. 1983 , , 272 , 29 dixon , w. v. , hurwitz , m. & ferguson , h. 1996 , , 469 , l77 donahue , m. & stocke , j. 1994 , , 422 , 459 donahue , m. & voit , m. 1993 , , 414 , 17 edgar , & chevalier , r. 1986 , , 310 , l7 edge , a. , stewart , g. & fabian , a. 1992 , , 258 , 177 edge , a. , ivison , r. , smail , i. , blain , a. , kneib , j .- p . 1999 , , 306 , 599 ettori , s. & fabian , a. 2000 , , 317 , l57 fabian , a. et al .
2001 , , 321 , l33 fabian , a. , mushotzky , r. , nulsen , p. & peterson , j. 2001 , , 321 , l20 fabian , a. & nulsen , p. 1977 , , 180 , 479 feldman , p. , sahnow , d. , kruk , j. , murphy , e. & moos , h. m. 2000 , , in press forman , w. & jones , c. 1982 , , 20 , 547 gordon , k. & clayton , g. 1998 , , 500 , 816 heckman , t. 1981 , , 251 , l59 heckman , t. , baum , s. , van breugel , w. , & mccarthy , p. 1989
, , 338 , 48 hu , e. 1988 , in cooling flows in clusters of galaxies , ed .
fabian ( dordrecht : kluwer ) , 73 hu , e. 1992 , , 391 , 608 hu , e. , cowie , l. & wang 1985 , , 59 , 447 karachentsev , i. & lipovetskii , v. 1969 , soviet astron , 12 , 909 kim , k. , kronberg , p. & tribble , p. 1991
, , 379 , 80 koekemoer , a. , odea , c. , sarazin , c. , mcnamara , b. , donahue , m. , voit , m. , baum , s. & gallimore , j. 1999 , , 525 , 621 loewenstein , m. & fabian , a. 1990 , , 242 , 120 markevitch , m. et al .
, , 541 , 542 mathis , j. & cardelli , j. 1992 , , 398 , 610 mcnamara .
b. & oconnell , r. 1992 , , 393 , 579 mittaz , j. et al .
2001 , , 365 , l93 moos , h. m. et al .
2000 , , 538 , l1 mushotzky , r. 1992 , in clusters and superclusters of galaxies , ed , a. c. fabian ( dordrecht : kluwer ) , 91 oegerle , w. & hill , j. 1994 , , 107 , 857 peterson , et al .
2001 , , 365 , l104 pinkney , j. et al.1996 , , 468 , l13 sahnow , d. s. et al .
2000 , , 538 , l7 sarazin , c. , burns , j. , roettiger , k. & mcnamara , b. 1995 , , 447 , 559 slavin , j. , shull , j.m . & begelman , m. 1993 , , 407 , 83 smith , e. p. , et al .
1997 , , 478 , 516 sparks , w. , macchetto , d. & golombek , d. 1989 , , 345 , 153 tamura , et al .
2001 , , 365 , l87 taylor , g. , odea , c. , peck , a. & koekemoer , a. 1999 , , 512 , l27 vikhlinin , a. , markevitch , m. & murray , s. 2001 , , 549 , l47 voit , m. , donahue , m. & slavin , j. 1994 , , 95 , 87 voit , m. & donahue , m. 1995 , , 452 , 164 yan , l. & cohen , j. 1995 , , 454 , 44 | we present far - ultraviolet spectroscopy of the cores of the massive cooling flow clusters abell 1795 and 2597 obtained with fuse .
as the intracluster gas cools through @xmath0k , it should emit strongly in the ovi 1032,1038 resonance lines .
we report the detection of ovi @xmath11032 emission in a2597 , with a line flux of @xmath2 erg @xmath3 s@xmath4 , as well as detection of emission from ciii @xmath1977 . a marginal detection of ciii @xmath1977 emission is also reported for a1795 .
these observations provide evidence for a direct link between the hot ( @xmath5 k ) cooling flow gas and the cool ( @xmath6 k ) gas in the optical emission line filaments . assuming simple cooling flow models , the ovi line flux in a2597 corresponds to a mass deposition rate of @xmath7 yr@xmath4 within the central 36 kpc .
emission from ovi @xmath11032 was not detected in a1795 , with an upper limit of @xmath8 erg @xmath3 s@xmath4 , corresponding to a limit on the mass cooling flow rate of @xmath9 yr@xmath4 .
we have considered several explanations for the lack of detection of oviemission in a1795 and the weaker than expected flux in a2597 , including extinction by dust in the outer cluster , and quenching of thermal conduction by magnetic fields .
we conclude that a turbulent mixing model , with some dust extinction , could explain our ovi results while also accounting for the puzzling lack of emission by fexvii in cluster cooling flows . |
this is the first draft to the paper , prepared so far only to appear in the proceedings as the talk of one of the authors ( h.b.f.n .. the contribution presents the main ideas for the fermionization / bosonization procedure in any dimension @xmath3 .
more detailed explanation will appear in the paper .
although many things are not yet strictly proven , the fermionization / bosonization seems , hopefully , to work in any dimensional space - time and also in the presence of a weak background field .
we hope , that the fermionization / bosonization procedure might help to better understand why nature has made of choice of spins , charges and families of fermions and of the corresponding gauge and scalar fields , observed in the low energy regime and why the _ spin - charge - family _
theory @xcite might be the right explanations for all the assumptions of the standard model , offering the next step beyond the _ standard model_.
this talk demonstrates that : * bosonization / fermionization is possible in an arbitrary number of dimensions ( although the fermions theories are non - local due to the anticommuting nature of fermions , while bosons commute ) . * the number of degrees of freedom for fermions versus bosons obeys in our procedure in any dimension the aratyn - nielsen theorem @xcite . *
the number of families in four dimensional space - time is ( a multiple of ) four families . to prove for massless fermions and bosons that the bosonization / fermionization is possible in an arbitrary number of dimensions we use the jacoby s triple product formula , presented by leonhard euler in 1748 [ 20 ] which is a special case of glaisher s theorem @xmath5 let the reader notices that the product on the left hand side runs over 0 and all positive integers , while on the right hand side it runs only over odd positive integers .
we shall connect the left hand side with the fermion degrees of freedom and the right hand side with the boson degrees of freedom , recognizing that for all positive numbers the number of partitions with odd parts equals the number of partitions with distinct parts [ 19 ] .
let us demonstrate this in a special case : + among the 22 partitions of the number 8 there are 6 that contain only odd parts , namely + ( 7 + 1 , 5 + 3 , 5 + 1 + 1 + 1 , 3 + 3 + 1 + 1 , 3 + 1 + 1 + 1 + 1 + 1 , + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 ) .
this is to be connected with the boson ( the right hand side of eq .
[ 11ue ] ) degrees of freedom .
+ if we count partitions of 8 , in which no number occurs more than once , that is with distinct parts , we obtain again 6 such partitions , namely + ( 8 , 7 + 1 , 6 + 2 , 5 + 3 , 5 + 2 + 1 , 4 + 3 + 1 ) .
this is to be correspondingly connected with the fermion ( the left hand side of eq .
[ 11ue ] ) degrees of freedom .
+ for every type of restricted partition there is a corresponding function for the number of partitions satisfying the given restriction .
an important example is q(n ) , the number of partitions of n into distinct parts @xcite .
the generating function for q(n ) , partitions into distinct parts , is given by [ 22 ] @xmath6 the first few values of q(n ) are ( starting with q(0)=1 ) : + ( 1 , 1 , 1 , 2 , 2 , 3 , 4 , 5 , 6 , 8 , 10 , ) the pentagonal number theorem can be applied giving a recurrence for q @xcite : @xmath7 where a(k ) is @xmath8 , if @xmath9 for some integer m , and is 0 otherwise . * bosonization illustrating formula * + below the @xmath10 space dimensional version , @xmath11 , of eq .
[ 11ue ] ( for only a `` quadrant '' ) is presented .
@xmath12 @xmath13 * the idea for the procedure for a proof of the existence of the multidimensional / bosonization formula * the idea for generalizing the bosonization / fermionization procedure from @xmath14dimensional case , where the euler s triple formula [ 11ue ] relates the number of the fock space basis vectors ( states ) for fermions and bosons , respectively , classified after the discretized momenta , where a fermion can have any positive or zero integer momentum @xmath15 - the left hand side of eq . [ 11ue ] - while a boson can have any odd value @xmath16 of the momentum .
both , the left and the right hand side manifest the same degrees of freedom . this can be seen if summing up on the fermionic side possibilities that @xmath15 of @xmath17 is a sum of all different integers , ( example 1 : factor @xmath18 at @xmath19 arises from the fact that @xmath20 ) , while at the boson ( right hand ) side the factor at @xmath21 is the sum of all possibilities that @xmath16 is written as a sum of positive odd integers ( example 1 : factor @xmath18 at @xmath22 arises from @xmath23 ) . we can generalize this bosonization / fermionization procedure from @xmath24 to any dimensional space time @xmath25 as follows : + * \1 .
we devide the whole system of all the discretized momentum vectors into " classes of proportional vectors ( meaning in practice vectors deviating by a rational factor only ) , or rays ( we might call them the rays of the module ) . *
\2 . for each `` class ''
the proof is given by the 1 + 1 dimensional case which means by just using the formula by euler and extending it to both positive and negative integers .
for @xmath26 it is meaningful to require that not only momentum components but also the energy are the same for fermions and bosons . * formulas of bosonization as products over rays / classes
* we correspondingly generalize eq .
[ 11ue ] by generalizing @xmath27 by a product of @xmath28 , ( @xmath29 ) and @xmath30 , the momentum @xmath16 by @xmath31 ( or rather @xmath32 ) , the power of @xmath30 by @xmath33 ( @xmath34 ) on the fermion ( the left hand ) side , and correspondingly the momentum @xmath15 by @xmath35 ( @xmath36 ) , the power of @xmath30 by @xmath37 ( @xmath38 ) on the boson ( the left hand ) side , while the summation runs over the classes .
@xmath39 where @xmath40 runs over the set @xmath41 of the @xmath10-tuples of non - negative integers , that can not be written as such a tuple multiplied by an over all integer factor .
* splitting the fock space into cartesian product factors from each ray @xmath40 * denoting the fock space for the theory - it be a boson or a fermion one - as @xmath42 for the @xmath10-dimensional theory , and by @xmath43 the fock space for the - essentially 1 + 1 dimensional theory associated with the ray/(or class ) @xmath40 describing the particles with momenta being an integer ( though not 0 ) times the representative for @xmath40 , namely @xmath44 - it is suggested that we write the full fock space as the product @xmath45 * introduction of creation and annihilation operators * we shall introduce for a boson interpretation of the hilbert space the fock space @xmath42 : @xmath46 where the integers can be any , except that they must _ not _ all @xmath10 ones be even . similarly for fermions : @xmath47 where now the @xmath48 numbers can be any integers .
* boson operators dividable into rays or classes @xmath40 , also fermion operators dividable into rays or classes @xmath40 , except for one type * we can write any " not all even ( discretized ) momentum @xmath49 as an _ odd _
integer @xmath15 times a representative for a class / ray @xmath40 @xmath50 the boson momentum with a given even / odd combination for its momentum components ( say @xmath51 ) goes to a ray / class @xmath40 with the _ same combination of even / odd - ness_. similarly one can proceed also for fermions with not all momentum components even ; but the fermion momenta that have all components even _ get divided into rays / classes with different even / odd combinations_. there are no rays with the even combination @xmath52 , of course , because a tuple of only even numbers could be divided by 2 .
we have made an important step arriving at a model suggesting how it could be possible to match momenta and energies for a system with either fermions or bosons . to completely show the existence of fermionization ( or looking the opposite way , bosonization ) we should
, however , write down the formula for how the fermion ( boson ) creation and annihilation operators are constructed in terms of the boson ( fermion ) operators , so that it can become clear ( be proven ) that the phase conventions and identification of the specific states with a given total momentum and energy for fermions can be identified with specific states for the boson system .
such a construction is well known for 1 + 1 dimensions , where it looks like @xmath53 in the " position representation , meaning that @xmath54 let us think of the case of making the field operators in position space @xmath55 hermitean by assumming @xmath56 at first one might naively think that - since each of our rays ( or classes ) c functions as the 1 + 1 dimensional system and we can write the whole fermion , as well as the whole boson , space according to ( [ firay ] ) - it would be trivial to obtain the bosonization for the whole system and thereby have achieved the bosonization in the arbitrary dimension , which is the major goal of this article .
however , one should notice that constructing in a simple way a system composed from several independent subsystems such it is the whole system @xmath42 , composed from the subsystems @xmath43 ( for @xmath57 ) , one obtains _ commutation _ between operators acting solely inside one subsystem @xmath40 , say , and operators acting solely inside another subsystem @xmath58 , say .
but we want for the fermions the _ anticommutation _ relations rather than the commutation ones , and thus some ( little ? ) trick is needed to achieve this anticommutation .
first we shall show how this anticommutation can be achieved by means of an ordering of all the rays @xmath59 by some ordering inequality being chosen between these rays : @xmath60 .
but this is a very ugly procedure and we shall develop a slightly more general attempt in which we construct a phase @xmath61 for each pair of rays @xmath40 and @xmath58 .
then we shall go on seeking to make the choice of this phase @xmath62 in a continuous and more elegant way .
since that shall turn out to be non - trivial , we shall develop the ideas by first seeking for such a construction of the phase for the odd dimensional space of @xmath63 , meaning @xmath64 , to learn the idea , although we are most keen on even space - time dimensions , such it is the experimentally observed number of space - time dimensions , @xmath65 .
let us suppose that we have a formal way of constructing the fermion creation and annihilation operators in terms of the boson operators .
we do indeed have such a construction , since we can fourier transform back and forth the construction in the position representation ( [ pos ] ) and the 1 + 1 dimensional bosonization is so well understood . since for the present problem the details of this 1 + 1 dimensional bosonization relations are not so important , we shall just assume that we are able to deduce for each ray or class @xmath40 a series of fermion creation - @xmath66 and @xmath67 - and annihilation - @xmath68 and @xmath69 - operators , that function well as fermion operators _ inside the ray @xmath40 _ , so to speak . @xmath70 and @xmath71 denotes odd and even , respectively .
the only important think is that these operators _ can be expressed in terms of the boson annihilation and creation operators belonging to the same ray @xmath40 _ : @xmath72 for these operators we know form the 1 + 1 dimensional bosonization that we can take them to obey the usual anticommutation rules _ provided we keep to only one ray @xmath40 _ : @xmath73 we have similar anticommutation rules for annihilation and creation operators if exchanging the index @xmath70 ( meaning odd ) by the index @xmath71 ( meaning even ) , but now we should take into account that the fermion operators with zero momentum , i.e. @xmath74 , are not constructed from a single ray @xmath40 .
rather there are - referring to our little problem with the explicit factor @xmath75 in the state counting formulae - not enough degrees of freedom in the 1 + 1 dimensional boson system to deliver a fermion operator with a zero momentum .
we should therefore imagine that we do not have these zero momentum fermion operators attached to our rays either .
this is actually good for our hopes of bosonizing in higher dimensions because the zero momentum fermion operators would have had to be common for the infinitely many rays and we would have had too many candidates for the zero momentum fermion mode .
now instead we totally miss the zero momentum creation and annihilation fermion operators for the many dimensional system .
that is , however , not at all so bad as it would have been to get an infinity of them , because we fundamentally can not expect to produce all fermion operators from boson ones because we can not possibly build up a sector with an odd number of fermions from boson operators acting on say some vacuum with an even number .
therefore one fermion operator must be missing .
this becomes the zero momentum one and that is o.k .. our real problem remains that these naive fermion operators taken for two different rays , @xmath40 and @xmath76 , will commute @xmath77 we could define an @xmath78-operator , where @xmath79 is the fermion number operator .
it sounds at first very easy just to write @xmath80 where the sums run over respectively the odd and the even positive values for @xmath16 for the @xmath70 and the @xmath71 components .
but now this fermion number operator- as taken as a function of the _ naive _ operators - ends necessarily up being an expression in purely boson operators ( from the ray @xmath40 ) , and thus it looks at first as being valid except when the expression @xmath81 , which we are interested in , is equal to @xmath82 on all states that can truly be constructed from boson operators .
if it were indeed so , our idea of using @xmath81 to construct the multidimensional fermion operators , would not be so good . however , there is a little detail that we did not have enough bosonic degrees of freedom to construct the zero momentum fermion operator in 1 + 1 dimensions .
therefore we can not really include in the definition of the " fermion number operator for the ray @xmath40 , @xmath83 , the term coming from @xmath84 .
this term would formally have been @xmath85 , but we decided to leave it out .
this then means that the fermion number operator , for which we decide to use @xmath83 as the number of fermions operator in the ray @xmath40 is _ not the full fermion number operator for the corresponding 1 + 1 dimensional theory , but rather only for those fermions , that avoid the zero momentum state_. to require this avoidance of the zero momentum is actually very attractive for defining a fermion number operator _ for the ray @xmath40 _ as far as the momentum states included in such a ray really must exclude the zero momentum state in a similar way as a ray in a vector space is determined from the set of vectors in the ray not being zero .
but this precise definition avoiding the zero - momentum fermion operator contribution to the fermion number operator @xmath83 leads to the avoidance of the just mentioned problem that this fermion number @xmath83 looked as always having to be even when constructed in terms of boson operators .
now there should namely be enough boson degrees of freedom that one should be able to construct by boson operators all the different possible combinations for fermion states being filled or unfilled ( still not the zero momentum included ) .
thus one does by pure bosons construct both - the even @xmath83 and the odd @xmath83 - states and thus the @xmath83 with the zero momentum fermion state not counted can indeed be a function of the boson operators and can _ take on both even and odd values _ for momentum , depending on the boson system state .
so , we can have - using this leaving out the zero momentum fermion state in the rays - an operator @xmath86 the operator @xmath81 for each ray @xmath40 counts if the number of fermions in the 1 + 1 dimensional system is even , then @xmath87 , or odd , then @xmath88 . we construct the following improved fermion operator ( annihilation or creation ) , @xmath89 the inclusion of this extra operator factor helps to convert the commutation relations between the fermion annihilation and creation operators for different rays into anticommutation relations , as it can easily be seen @xmath90 thus we deduced , for @xmath91 in our in fact at first just chosen ordering of @xmath92 , that the fermion operators do anticommute .
it is not difficult to show similarly also in the case @xmath93 , that the fermion operators anticommute .
the crux of the matter is that when e.g. @xmath94 there is the factor @xmath95 contained in the product @xmath96 , which is attached to @xmath97 in order to correct it into @xmath98 , while there is no analogous factor @xmath99 contained in the factor @xmath100 attached at @xmath101 in order to bring it into @xmath102 . in this way one gets just the one extra minus sign in the product of the fermion operators that makes them anticommute .
it is not difficult to see that the idea of using such an ordering @xmath92 could be slightly generalized to have instead of the factors only minus or plus phase factors of the form @xmath103 @xmath104 it is also not difficult to see that , in order to obtain the anticommutation relations instead of the commutation ones ( which we have for @xmath105 ) , the phases must obey the rule @xmath106 we may in fact seek to evaluate the product of two fermion creation operators with the ansatz ( [ phansatz ] ) @xmath107 where in the last step we used ( [ picon ] ) .
thus we see that in this way we can get - really in infinitely many ways - some algebraicly defined fermion operators that do indeed anticommute as they should .
but it should be had in mind that both these procedures , by choosing @xmath61 and the forgoing proposal with the ordering @xmath92 , are a priori discontinuous and arbitrary .
we expect , however , that the latter method with @xmath61 can be lead to a smooth and attractive scheme in the case of @xmath64 or equivalently @xmath63 . in the case of
@xmath64 we can say that the set of our @xmath41 form a kind of a set of `` rational angles in the sense that each ray specifies an angle modulo @xmath109 ( rather than modulo @xmath110 as for an oriented arrow it would specify ) , but that one only obtains those angles which rationalize tangenses .
but the fact that they are after all implemented as angles - although only modulo @xmath109 , means that they are at least locally ordered as numbers along a real or rather rational axis .
so apart from troubles at the end and beginning we have an ordering and we could attempt to use it even for the implementation of the ordered set of rays method by proposing a ' ' nice
@xmath92 ordering .
however , we think we get a better chance by using the @xmath61 method in this @xmath63 and thus @xmath64 case . we have to think about what topological properties we shall and can achieve for the function @xmath61 depending on a pair of rays @xmath40 and @xmath58 . since the classes or rays are " a kind of rational directions , though without orientation , the topological space of the rays is like the sphere @xmath111 with opposite points identified .
this topological space obtained by the identification of the opposite point on the @xmath111 sphere is actually topologically identical to the projective space of @xmath112 dimensions .
for the case @xmath63 or @xmath64 the topological space rays thus become simply the projective line ( using real numbers ) , but that is topologically just the @xmath113 circle .
had this topological space been naturally orderable we could have used the ordering as the @xmath92 above .
however , it is a circle @xmath113 and not a simple line with plus and minus infinity ; the infinities have so to speak been identified to only one point in the projective line .
this means that using the method to define the fermion fields / operators by means of @xmath92-method would be very non - elegant , and would probably violate almost everything wanted .
let us now think about a slight generalization by using the @xmath61 .
we need to make a choice of a function @xmath114 defined on the cross product of two projective spaces of dimension @xmath115 each .
since it shall obey the condition ( [ picon ] ) , it can not at all be a smooth or continuous function at the points where @xmath116 .
let us , for a while , take care that this method works well for @xmath63 only .
in this @xmath63 case the cross product of the two projective lines becomes topologically simply a two - dimensional torus .
so we face topologically to define @xmath61 on a two - dimensional torus .
however , we are forced to give up having continuity along the `` diagonal - meaning the set of points on this torus with @xmath116 - and it is thus rather a @xmath61 defined as a continuous function on the torus minus its ' ' diagonal , which we must choose / find . this two dimensional torus minus its `` diagonal is rather like a belt .
i.e. , it is topologically like the outer surface of a finite piece of a tube .
it has two separate edges , each being topologically an @xmath113 circle , namely two images of the ' ' diagonal seen from the two sides . in between there is then the two - dimensional bulk area of the topological shape of the surface of the finite piece of a tube .
it is inside this bulk region that we shall attempt to construct @xmath61 to be smooth and " nice .
choosing @xmath117 might be a good choice . here
@xmath118 is the angle that the unoriented direction of slightly complicatedly defined angle - dividing line dividing the angle between the unoriented directions @xmath40 and @xmath58 forms with some coordinate axis ( in momentum space ) . the precise way of defining this @xmath119 is illustrated on the figure .
it consists in the following ( let us remind the reader that we are still in the @xmath120 case ) : * \a .
we introduce a " clockwise rotation orientation in the spatial momentum plane . *
we draw a circle arrow from one of the two `` ends
( half lines ) of which the line @xmath40 ( the ray @xmath40 is basically just a line ) in this clockwise direction , and note the angle between this end of @xmath40 and the first ' ' end
( = half line ) of @xmath58 ( met in the following the circular arrow ) , which measures less than @xmath121 .
we draw a line , that divide this under b. noted angular region into halves .
this line through the ( momentum space ) origo is denoted `` clock average '' ( as marked on the figure ) . * \d .
such an unoriented line as the " clock average defines relative to a coordinate system in spatial momentum space an angle - value modulo @xmath109 .
we call this angle - value @xmath118 and it is as just said defined modulo @xmath109 ( but only modulo @xmath109 , because the line `` clock average '' is unoriented ) . * \e . multiplying this angle - @xmath118 - by 2 its ambiguity
to be only defined modulo @xmath109 becomes instead an ambiguity modulo @xmath110 .
thus our proposed expression ( [ prop ] ) for @xmath61 is defined modulo @xmath110 , and that is what we need , since in our construction we exponentiate @xmath62 after multiplication by @xmath122 and an operator @xmath123 that has only integer eigenvalues .
thus the expression , which we use , @xmath124 becomes well defined even though @xmath125 makes sense only modulo @xmath110 .
* let us see whether this proposal is indeed a good one . * to see that our proposal ( [ prop ] ) is a good one we must first of all check that it obeys ( [ picon ] ) .
that is we must see what happens to the expression when we permute the two independent variables @xmath40 and @xmath58 . since by definition the circular arrow constructed in step b. goes out from the @xmath40-line , the first of the two arguments in @xmath61 , we must draw this circle - arrow after the permutation from @xmath58 instead . therefore the half - angle noted under point b. above will after the permutation differ from the one before the permutation .
this means that the line ( through the origo ) " clock average gets after the permutation perpendicular to its direction before the permutation of @xmath40 and @xmath58 .
therefore @xmath126 , which means that this angle gets shifted modulo @xmath109 with @xmath127 .
after the multiplication by @xmath18 ( point e. ) it means that @xmath128 , which is just ( [ picon ] ) .
thus we got indeed by proposal ( [ prop ] ) the condition ( [ picon ] ) fulfilled
. we can now remark that quite obviously our proposal ( [ prop ] ) is continuous as function of the directions @xmath40 and @xmath58 _ except where @xmath40 and @xmath58 just coincide - what means that it is zero ( mod @xmath109 ) angle between them_. let us note that had we not chosen the clock - wise rule , but instead taken , say , the smallest angle between @xmath40 and @xmath58 and just found the halfening line between those `` ends
, we would have got a discontinuity when @xmath40 and @xmath58 were perpendicular to each other .
but by our precise choice we avoided that singularity .
( for a point close to the diagonal the two arguments , @xmath40 and @xmath58 , are approximately the same ray .
permuting them will for a continuous function @xmath129 make almost no difference , and thus it can not possibly change by @xmath109 , while crossing the ' ' diagonal the function @xmath130 would ask to jump by @xmath109 . )
we propose the generalization of eq .
( [ pos ] ) to an arbitrary dimension , due to our experience with the clifford objects ( apart from some modifications due to whether we choose weyl or majorana fermions for family or for geometrical components ) , by using the relation @xmath131
we started with massless noninteracting bosons or fermions .
but we want to work with the interacting fields .
there are many kalb - ramond fields appearing in our type of fermionizable boson model in higher dimensions and correspondingly it is not easy to see how to make an interacting theory .
there are many ways to come from noninteracting bosonisable ( fermionizable ) fermion ( boson ) fields , which might lead to the fermion fields interacting with the boson fields as it is in the _ spin - charge - family _ theory .
but on the level of our fermionizable ( bosonizable ) boson ( fermion ) model with many kalb - ramond fields we must keep in mind that the conserved charges in the kalb - ramond theories are vectorial and thus one gets very many vectorial conserved quantities .
this makes scattering processes ( unless all the scattering particles are without these vectorial charges ) very non - trivial .
one chance would be to let either fermion or boson fields to interact with gravity . crudely speaking gravity couples to energy and momentum , and since in the free bosonization procedure
we have at least sought to get the total d - momentum be the same in the corresponding states of fermions and bosons there might be a chance that we fermionize a theory with both - the bosons of the kalb - ramond type _ and _ gravity through the vierbein formulation - and correspondingly obtain a theory with both fermions _ and _ bosons , the later would be the gravity degrees of freedom .
this might lead to exactly the theory @xcite that one of us ( n.s.m.b . ) has postulated as the true model for nature beyond the standard model ( the _ spin - charge - family _ theory ) . since our scheme a priori looks to require the majorana fermions to have real fields like the bosons - at least in the simplest version - we only expect to get chiral fermions in those dimensions wherein majorana fermions can simultaneously be weyl ( which means chiral ) @xcite as in d=2,6,10,14 ,
... it is therefore even a slight support for the _ spin - charge - family _ theory that its phenomenologically favoured dimension is just 13 + 1 = 14 .
one should for appreciating this idea of adding gravity without fermionizing it have in mind that one _ does not have to fermionize ( bosonize ) all degrees of freedom _ , but rather can - if one wishes - decide to fermionize ( bosonize ) some degrees of freedom but not all . especially , if the motivation were to make all fermions from bosons because one claims that fermions are not properly local and should not be allowed to exist , then of course it is enough that we start with a purely boson theory as the fundamental one - and then we better only fermionize a part of bosons unless we could identify a purely fermionic theory with nature .
but of course there seemingly are bosons in nature and we thus must end phenomenologically with a theory with _ both _ bosons and fermions . starting from fundamental bosons only that is only achievable by only a _
partial fermionization_. * hope for the progress * the hope is , which is evidently from what we have proposed in this contribution , that we shall construct formulas for the higher dimensional cases by generalizing the formulas we already have for the one dimensional case , generalizing as well the `` classes '' to higher dimensions . in the spirit of seeking to identify the fields characterized by their " odd / even indices with spin components , we hope to derive from the bosonization formula a scheme formally stating the relation between the boson and the fermion second quantized fields , @xmath132 boson field components , while there will be @xmath133 fermion components .
let us try to clarify how the here discussed fermionization procedure is supposed to be , so to speak , the root for a theory beyond the _ spin - charge - family _ theory of norma susana manko bortnik @xcite ( and her collaborators ) , the ( one of ) way(s ) we see as a very promising hope that one could justify this _
spin - charge - family _ theory by the hoped fermionization is as follows : _ we build up a model with only bosons as the fundamental theory in say - 13 + 1 dimensions - in the sense that this 13 + 1 dimensional purely bosonic theory with a series of the kalb - ramond fields _ and _ with usual 13 + 1 dimensional gravity should be the fundamental choice of nature _ ( not necessarily starting in d=13 + 1 ) .
then this theory should be _ partly _ fermionized in the sense that only the series of kalb ramond fields get fermionized , but not the gravity ( bosonic ) degres of freedom .
the latter remain gravitational degrees of freedom hopefully now functioning as gravity for the fermions that came out of the fermionization .
the _ spin - charge - family _ theory will show up out of the kalb - ramond components . 1 .
the first assumption of our new scheme , which might be the pre - scheme of the _ spin - charge - family _ theory , is that _ fermions a priori do not obey proper locality_. the accusation towards all the theories with fermions which are fundamental fermions rather than fermionized bosons is that the axiom of locality in a quantum field theory is for the fermions @xmath134 while true physical locality should have been a _ commutation rule _ like the one obeyed by the boson fields @xmath135 true locality means , one would think , that each little region in space is approximately a completely separate system that only interacts very indirectly with a far away different little region .
if so , the physical operators describing the situation in one little region should commute with those describing the situation in a different little region , and not anticommute as the fermion fields do .
one might like to assume that only products of an even number of fermion fields are considered as proper operators describing the little system region , what satisfies the requirement of getting commutation relations between the field variables describing the situation in different regions .
but such an assumption must be justified as a physical assumption , discussing seriously also odd products of fermion fields .
+ the point of view we suggest here is that we admit that we can not have fermions at all in a truly local way !
this then means that the fundamental physics should be a model _ without fermions _ so that all fermions come from bosons that become fermionized .
since it is not easy to find so terribly many systems of bosons that can be fermionized , and thus if one finds some way of fermionizing , then this way is presumably already likely to be almost the only one possible .
+ at least we expect that the fermionization of a boson system of fields can only be made provided the number of fermions and the number of bosons agree with the theorem which one of us ( h.b.f.n . ) and aratyn @xcite put forward many years ago . + for massless free fermions on the one side and massless free bosons on the other side we obtained that the number of components for the bosons and the fermions counted in the same way with respect to the fields being real or complex , should be in the ratio @xmath136 where the dimension is @xmath137 , or the spatial dimension is @xmath10 .
+ the number of components - at least the number of real counted components - must of course be positive integer or zero .
thus the minimal number of fermion components must be @xmath133 , while the number of boson components must be @xmath132 or the numbers must be an integer multiplum of these numbers .
+ alone this theorem of ours @xcite makes appreciable restriction for when bosonization / fermionization is at all possible .
we are suggesting here the starting point with the bosonic degrees of freedom only , consisting of `` series of the kalb - ramond fields , all the chain through , except for one ( pseudo)scalar '' . by this
we mean that we have as the bosons a series of separate fields @xmath138 with all the values of the number of indexes , antisymmetric with respect to all their indexes .
+ there is a simple way in which one could get the number @xmath139 of boson components , if we arrange to have - by some gauge choice - only spatial values of the indexes @xmath16 , @xmath15 , ... on the a - fields , removing the @xmath140 field with zero indexes .
the number of components become equal to the number of subsets of @xmath10 letters , which is @xmath133 . removing pure scalar
, we get this number @xmath132 , as we want for the theorem of@xcite .
4 . from the @xmath132 bosons represented by the kalb - ramond fields ( with the scalar removed ) ,
then according to the theorem @xcite there must be the @xmath141 components of fermion fields .
this means for the weyl spinor representation of fermion fields in even @xmath142 , with @xmath0 members that there are @xmath143 @xmath0 real fermion fields . to get @xmath141 real weyl spinor representation fermion fields
there must be @xmath144 @xmath145 @xmath146 families .
5 . from the bosonization requirement
we obviously get out that there must exist an even number of families as it also comes out from the _ spin - charge - family _ theory of one of us ( n.s.m.b . )
but now there is a correction due to the components of the kalb - ramond fields with time indices , that is with @xmath147 .
this gives very interesting corrections as we may postpone till later .
there is an interesting hope for that actually our at first free bosons being fermionized to free fermions could be generalized to have an universal coupling to a gravitational field - the bosonic field , which we do not fermionize , keeping it as gravitation , interactiing with the fermions - so that we finally arrive at a theory with several families of fermions and gravity . in this contribution
we wrote down a formula for counting the number of states for the fermion and the boson systems having the same number of fock states with given momentum and energy for the free massless case of our bosonization / fermionization procedure .
we used in reality an infrared cut off that meant that we in fact considered a torus world with for different components different periodicity conditions : some components of fields had antiperiodicity while the others had periodicity property along various coordinate directions .
the wave functions for the momentum eigenstates are as is well known all along taking on only pure phase factor values , i.e. they obey @xmath148 all along .
the number of turns around zero , which they perform when one goes around the torus along the different coordinates , is an integer ( or a half integer depending on the boundary condition ) .
we can consider this number of turns going around the torus in different ways ( along different coordinates ) a topological quantity in the sense that it as an integer can not change under a small deformation .
our main idea is at this point that we in this way can introduce at least a not too strong gravitational field and still have single particle solutions to the equations of motion characterized by the same system of ( topological ) quantum numbers .
that should suggest that we have the same set up for making the in this work studied bosonization in a not too strong gravitational field as in the free case .
we namely should be able to classify the single particle states as functions of the space - time variables @xmath27 on the by gravitational fields deformed torus ( torus due to infra red cut off ) according to a topological classification in terms of the number of times the wave function encircles the value zero in the complex plane when the one follows a closed curve , following , say , the coordinates of the deformed torus .
for the massless theory we have scale - invariance for the matter fields - the series of the kalb - ramond fields or the fermions - so , as long as we consider the gravitational field as a background field , i.e. we ignore the dynamics of the gravitational field itsef - we can scale up the momenta of the single particles by just letting the phase of an eigen - solution be scaled up by a factor .
only the periodicity conditions will enforce such scalings to be by integer factors , just as they must be also in the free flat case .
so we argue that with a background gravitational field , that is with a not too strong field , we have a possible description in terms of a discretized enumeration quite in the correspondence with the one for the flat case . remembering that we obtained the bosonization w.r.t .
state counting in fact class @xmath40 for class , meaning that the momentum eigenstates in the classes corresponding to rays went separately from boson to fermion or oppositely , we may have given arguments at least suggesting that a corresponding bosonization correspondence as the one in the free flat case also applies to the case with some ( may be not too large though ) gravitational field as a background field .
since we now with our expansions in power series seek to guarantee that we shall make the bosonization or fermionizations just in such a way that the d - momentum will be the same for the fermion configuration and the boson configuration corresponding to each other , we might hope that we could formulate the exact correspondence and the interpretations in terms of fields with spin indexes so that indeed the momentum densities would be the same for the fermions and for the corresponding bosons .
if we succeed in that then the action on the gravitational fields which only feel the matter via the energy momentum tensor @xmath149 would be the same for the bosons and the fermions in the corresponding states . in that case
the development of the gravitational fields would be the same for the corresponding fermion and boson configurations .
thus the bosonization / fermionization procedure would truly have been made also in the with gravity interacting models .
just the gravity field itself should _ not be fermionized_. 99 aratyn , h.b .
nielsen , `` constraints on bosonization in higher dimensions '' , nbi - he-83 - 36 , conference : c83 - 10 - 10.2 ( ahrenshoop sympos.1983:0260 ) , p.0260 proceedings .
jacobi , c. g. j. , _ fundamenta nova theoriae functionum ellipticarum _ ( in latin ) , knigsberg : borntraeger ( 1829 ) , isbn 978 - 1 - 108 - 05200 - 9 , reprinted by cambridge university press 2012 euler , l. : _ demonstratio theorematis circa ordinem in summis divisorum observatum _ , novi commentarii academiae scientiarum imperialis petropolitanae * 5 * , 75 ( 1760 ) , translated as bell , j. , _ a demonstration of a theorem on the order observed in the sums of divisors _ [ arxiv : math/0507201 ] ; euler , l. _ evolutio producti infiniti @xmath150 etc . in seriem simplicem _ , acta academiae scientarum imperialis petropolitinae 1780 , pp .
47 - 55 , 1783 .
opera omnia , series prima , vol .
472 - 479 , translated as bell , j. _ the expansion of the infinite product @xmath151 etc . into a single series _ [ arxiv : math.ho/0411454 ] .
abramowitz , milton ; stegun , irene a. , _ handbook of mathematical functions with formulas , graphs , and mathematical tables _ , applied mathematics series 55 ( 10 ed . ) , new york , usa : united states department of commerce , national bureau of standards ; dover publications , page 825 .
glaisher , j.w.l.,_messenger of mathematics _ , * 12 * ( 1883 ) 158 .
manko bortnik , `` the explanation for the origin of the higgs s scalar and for the yukawa couplings by the _ spin - charge - family _ theory '' , _
physics _ * 6 * ( 2015 ) 2244 - 2274 , http://dx.doi.org./10.4236/jmp.2015.615230 [ arxiv:1409.4981 ] .
manko bortnik , `` matter - antimatter asymmetry in the _ spin - charge - family _ theory '' , _ phys .
_ * d 91 * 065004 ( 2015 ) , doi:10.1103 [ arxiv:1409.7791 ] .
manko bortnik , h. b. nielsen , fermions with no fundamental charges call for extra dimensions , _ phys
. lett . _ * b 644 * ( 2007 ) 198 - 202 , hep - th/0608006 . | one purpose of this proceedings - contribution is to show that at least for free massless particles it is possible to construct an explicit boson theory which is exactly equivalent in terms of momenta and energy to a fermion theory .
the fermions come as @xmath0 families and the to this whole system of fermions corresponding bosons come as a whole series of the kalb - ramond fields , one set of components for each number of indexes on the tensor fields . since kalb - ramond fields naturally ( only ) couple to the extended objects or branes
, we suspect that inclusion of interaction into such for a bosonization prepared system - except for the lowest dimensions - without including branes or something like that is not likely to be possible .
the need for the families is easily seen just by using the theorem long ago put forward by aratyn and one of us ( h.b.f.n . ) , which says that to have the statistical mechanics of the fermion system and the boson system to match one needs to have the number of the field components in the ratio @xmath1 , enforcing that the number of fermion components must be a multiple of @xmath2 , where @xmath3 is the space - time dimension .
this " explanation of the number of dimension is potentially useful for the explanation for the number of dimension put forward by one of us ( s.n.m.b . )
since long in the _ spin - charge - family _ theory , and leads like the latter to typically ( a multiple of ) @xmath4 families . and
this is the second purpose for our work on the fermionization in an arbitrary number of dimensions - namely to learn how `` natural '' is the inclusion of the families in the way the _ spin - charge - family _ theory does . |
Subsets and Splits